ANALYSIS OF REGULATORY W MUTATIONS AND THE FUNCTION OF THE KIT RECEPTOR IN THE INTESTINE

Michael Kliippel

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Molecular and Medical Genetics University of Toronto

O Copyright by Michael Kliippel, 1998 National Library BibI iotMque nationale du Canada Acquisitions and Acquisitions et Bibliographic Sewices services bibliographiques 395 Wellington Street 395. rue Wellington OttawaON K1AON4 Ottawa ON KIA ON4 Canada Canada

The author has granted a non- L'auteur a accorde me licence non exclusive Licence allowing the exclusive pennettant a la National Library of Canada to Bibliotheque nationale du Canada de reproduce, loan, distribute or sell reproduire, prster, distribuer ou copies of this thesis in microform, vendre des copies de cette these sous paper or electronic formats. la forme de m.icrofiche/fih, de reproduction sur papier ou sur format electronique.

The author retains ownership of the L'auteur conserve la propriete du copyright in this thesis. Neither the droit d'auteur qui protege cette these. thesis nor substantial extracts fiom it Ni la these ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent etre imprimes reproduced without the author's ou autrement reproduits sans son permission. autorisation. Diese Doktorurbeit widme ich meinem Sohn Julian ANALYSIS OF REGULATORY W MUTATIONS AND THE FUNCTION OF THE KIT RECEITOR IN THE INTESTINE,degree of Doctor of Philosophy, 1998, by Michael Kliippel, Graduate Department of Molecular and Medical Genetics, University of Toronto.

The murine W encodes the Kit (RTK). Many structural mutations at the W locus, affecting the Kit coding sequence, lead to pleiotropic defects in hematopoiesis, gametogenesis and melanogenesis. Regulatory W mutations, on the other hand, do not affect the Kit coding sequence and cause cell type specific defects. Moreover, the Kit RTK is expressed in several cell types and tissues with no obvious defects in W mutant mice, raising the question of the function of the Kit molecule in these tissues. In this thesis, I have analyzed the molecular basis and developmental defects associated with two spontaneous regulatory mutations in the mouse W locus. I describe the effects of the ~57and ~banded(bd) mutations on Kit expression during embryogenesis and in the adult animal and show specific defects in melanocyte development in these mutant animals. In addition, I describe the molecular basis of these regulatory W mutations. The Kit RTK is expressed in the adult intestine, but no intestinal defects had been described in W mutant animals. I identlfy the Kit-expressing cells in the small intestine as the interstitial cells of Cajal (ICC) and demonstrate the importance for this cell type in intestinal pacemaking. Moreover, I analyze the developmental origin and lineage relationships of these Kit-expressing cells in the small intestine and investigate the role of Kit in their development. ICC and smooth muscle cells of the intestine are derived from a common precursor. Moreover, a functional Kit receptor is not required as an instructive signal during lineage determination, but is necessary for ICC promeration after birth.

iii These results provide novet information on the multiple roles of the Kit RTK in mouse development, as well as new insights into the regulation of expression of this . Firstly, I would like to thank my supervisor Alan Bemstein for his continual support, his patience and his never-ending attempts to teach me the Art of Science. Alan's influence had a big impact on my scientific as well as my personal development. Thanks for everything, Alan! I also would like to express my appreciation to the members of my committee, Robert Phillips, Manuel Buchwald, and especially Martin Breitman, who will always be remembered. I also want to thank the members of the Bernstein lab, past and present, for their help and friendship. In particular, I wish to thank Derrick Rossi (Leisure Suit Larry [#2]) for his continuous friendship, for discussions, for not executing plan B on my desk, and for many late nights at the bench, surrounded by the roaring silence (the volume dial did go up to 1 1!). In this context, I want to apologize to all other members of the lab for the high impact musical assaults they had to endure during these times. I hope the ringing stopped. I also want to thank Mary E. Brunkow for her friendship, help and support, and Ken Kao, Xianhua Piao, Benny Motro, Fabrice Melet, Klaus-Dieter Fischer, Jonathan Lee, John Abrahamson, Gina Caruana, Adam Hart, Dorit Donoviel, Robert Paulson, Bill Stanford, Mira Puri, Mehran Sam and Beverley Bessey for their help and insightful discussions. I thank Shirley Vessely, Sandra Gardner and George Cheong for their technical assistance. Many other people in the Institute contributed to the stimulating scientific environment. Thanks to all of them! I wish to thank Alicia Maund for her support and her continuous efforts to raise my awareness of social, emotional and psychological issues. It's a long and winding road, but maybe I will be able to reach the end. Most importantly, I would like to thank my mother, father and sister for their never-ending love and support.

Timeo danaos, et dona ferentes or Hey daddio, I don't wanna go, down to the basement- there's something down there!

This doesn't have anyhng to do with the Indians, though, I guess I kind of got side-tracked.

This is funny for so many reasons.

Toronto, October 30, 1997 Michael Kliippel TABLE OF CONTENTS

... Thesis Abstract ...... IIl Acknowledgements ...... v Table of Contents ...... vii ... List of Figures ...... XUI List of Tables ...... xvi List of Abbreviations ...... xvii

CHAPTER 1 INTRODUCTION

ANALYSIS OF THE MOLECULAR BASIS OF MOUSE DEVELOPMENTAL MUTANTS ...... THE MOUSE W LOCUS AND ITS MUTANT ALLELES ...... The mouse W locus encodes the Kit RTK A cluster of coat color mutations on mouse 5 ...... A cluster of receptor tyrosine on mouse chromosome 5 ...... Structural and regulatory mutations at the W locus ...... THE MOUSE SZ LOCUS AND ITS MUTANT ALWS The mouse SI locus encodes the ligand for the Kit RTK ...... Structural and regulatory S1 alleles

vii SIGNAL TIRANSDUCTION THROUGH THE KIT RTK ...... REGULATION OF KIT EXPRESSION ...... Patterns of Kit expression ...... Regulation of Kit expression ...... PHYSIOLOGICAL ROLE OF THE KIT REOR DURING MOUSE DEVELOPMENT ...... Biological roles of the Kit signal transduction pathway ...... Oncogenic potential of the Kit receptor ...... The role of the Kit signaling pathway in hematopoiesis ...... The role of the Kit signaling pathway in melanocyte development ...... The role of the Kit receptor in ear development and hearing ...... The role of the Kit signaling pathway in germ cell development ...... The role of the Kit receptor in tissues with no obvious defects in W mutant mice ...... REFERENCES ......

viii CHAPTER 2 LONG-RANGE GENOMIC REARRANGEMENTS UPSTREAM OF &IT DYSREGULATE TEl3 DEVELOPMENTAL PATTERN OF KIT EXPRESSION m w57 AND w~DEDMICE AND INTERFERE Wrm DISTINCT STEPS IN MELANOCYTE DEVELt OPMENT ......

ABSTRACT ...... INTRODUCrION ...... MAT- AND METHODS ...... Mice and embryos ...... Cell culture ...... How cytometry ...... RNA in situ hybridizations ...... RNA isolation and Northern blot analysis ...... Nuclear run ons ...... Preparation of mouse genomic and high molecular weight genomic DNA, pulsed-field gel electrophoresis and Southern blot analysis ...... Analysis of Kit upstream sequences by Polymerase Chain Reaction (PCR) ...... WILTS ...... Tissue-specific dysregulation of Kit expression in both ~57and wbd adult animals ...... Both ~57and wbd are associated with dysregulation of embryonic Kit expression during mid- and late gestation ...... -. .. Disturbed Kit expression in wbd embryos

during somite differentiation ...... *.....*...... *.... The effects of the ~57and ~bd mutations on morphogenetic behaviour and fate of developing melanoblasts ...... *.**... Both ~57and wbd affect transcriptional initiation of Kit ...... *.... The ~57mutation is associated with a 80 kb deletion 5' of Kit ...... ,,...... * ...... The ~bdmutation is associated with a 2.8 Mb inversion upstream of Kit .....,...... *...... **...... DISCUSSION ...... *...... ~...... *...... ACKNOWLEDGEMENTS ...**...... *...... *...... REFERENCES ....-...... *...... **...... -...... *.....*

CHAPTER 3 THE W/KIT GENE IS REQUIRED FOR THE DEVELOPMENT OF TEIE INTERSTlTIAL CELLS OF CAJAL AND FOR INTESTINAL PACEMAKER ACTIVITY ...... *.*...... *...... *...... *.. 113

ABSTRACT ..- ...... *. .. .. *...... *...... 114 RESULTS AND DISCUSSION ...... -...... *..- 115 REFERENCES ......

C-R 4 DEVELOPMENTAL ORIGIN ANID KIT-DEPENDENT POSTNATAL DEVELOPMENT OF TLLE INTEXSTITIAL CELLS OF CAJAL IN TEE MAMMALIAN SMALL INTESTINE ......

ABSTRACT ......

INTRODUCTION ...... ,., ......

MATERIALS AND METHODS ...... ,...... Mice and embryos ...... RNA in situ hybridization ...... Supra-vital methylene blue staining ...... Elec trophysiology ...... Apoptosis assays ......

RES.ULTS ..... r.r...... r..r...... Analysis of Kit expressing cells in the developing intestine of wild type. ~57and ~bd embryos ...... Continuous restriction of the numbers of Kit-positive cells during development of the outer muscle layers ...... Developmental origin of ICC ...... Analysis of Kit expression in wild type. ~57and ~bd adult small intestine Effects of the ~5~ and ~bdmutations on KC development and intestinal pacemaker activity ...... Kit is not required for embryonic development of ICC ...... The reduction in ICC density negatively correlates with the increase of the surface area of the small intestine ...... ICC apoptosis in ~bdmice ...... DISCUSSION ...... ACKNOWLEDGEMENTS ...... REFERENCES ......

CHAPTER 5 SUMMARY. DISCUSSION AND FUTURE DIRECTIONS ...... ,......

SUMMARY AND DISCUSSION ...... FUTURE WORK ...... REFERENCES ......

xii Lrsr OF FIGURES

Figure 1. Schematic model for the generation of several unlinked clusters of RTK ...... 6 Figure 2. Schematic illustration of structural gain and loss of function mutations in the Kit RTK ...... 10 Figure 3. Pigmentation defects in different W alleles...... 31 Figure 4. Kit-dependent and Kit-independent stages of melanocyte development ...... 32

CHAPTER 2

Figure 1. Tissue and cell type specific perturbances of Kit expression patterns in ~571~57and w6d/wt,d adult mice ...... 71 Figure 2. RNA in situ analysis of Kit and Steel factor expression during embryogenesis...... -. 74 Figure 3. Expression analysis of Tp-2 at dl 1.8 of embryogenesis...... 83 Figure 4. RNA in situ analysis of Trp-2 and Kit expression in melanoblasts...... 85 Figure 5. RNA in situ analysis of Trp-2 expression at d14.5 and dl 8 of embryogenesis...... 86

... Xlll Figure 6. Double labeling RNA in situ analysis at d 14.5 and d 18 of embryogenesis...... 88 Figure 7. Nuclear run on analysis...... 91 Figure 8. PCR analysis of the Kit 5' region in W57 and Wbd mutants...... - - -...... 94 Figure 9. PFGE analysis of wild type and ~57DNA ...... 96 Figure 10. PFGE analysis of wild type and Wbd DNA ...... 99

CHAPTER 3

Figure 1. Histochemical analysis of the small intestine in wild type and W mutant mice...... 119 Figure 2. EM analysis of the small intestine of wild type and W mutant mice ...... 121 Figure 3. Whole-mount RNA in sim analysis...... 123 Figure 4. Action potential generation in wild type and W mutant mice...... *...... 125

CHAPTER 4

Figure 1. RNA in situ analysis of Kit and SMMHC

. * expression m embryonic gut...... Figure 2. Single and double labeling RNA in situ experiments of wild type embryos......

xiv Figure 3. RNA whole mount in situ analysis of Kit expression in adult small intestine...... 147 Figure 4. Electrical activity generated in wild type, ~57and Wbd homozygous mice...... 150 Figure 5. Supra-vital rnethy lene blue staining of wild type and wb4wbd small intestine...... 152 Figure 6. Apoptosis assay of wild type and Wbd/Wbd small intestine ...... 157 Figure 7. The development and lineage relationship of the KCin the mouse small intestine....-...... 163

Figure 1. Effects of the ~57and wbd mutations on melanocyte development ...... 178 Figure 2. Development of KC...... 185 Figure 3. Genomic rearrangements associated with coat color mutations in the vicinity of the W locus on mouse chromosome 5...... -. .- .-. 190 LlST OF TABLES

CHAPTER 1

Table 1. Phenotypes and molecular bases of structural and regulatory W mutations...... I1 Table 2. Phenotypes and molecular bases of structural and regulatory Sl mutations ...... 14

CHAPTER 4

Table 1. Calcuiations of surface area of the small intestine and numbers of KCin wild type and mutant mice...... 154

xvi LlST OF ABBREVIATIONS

Ack-2 Anti c-Kit antibody-2 CM Circular muscle LM Longitudinal muscle KC Interstitial cells of Cajal PCR Polymerase Chain Reaction w Dominant White Spotting SL Steel Slf Steel factor KL Kit ligand SCF Stem cell factor MGF growth factor SMMHC Smooth muscle myosin heavy chain TT-2 Tyrosinase related protein-2 Ph Patch Rw Rump-white rs Recessive spotting mi Microphthalmia Mi tf Mi transcription factor mPs Infantile hypotrophic pyloric stenosis

xvii C-R 1

INTRODUCTION ANALYSIS OF THE MOLECULAR BASIS OF MOUSE DEVELOPMENTAL mms

Our understanding of the molecular control of embryonic and postnatal development in mammals derives from two broad experimental strategies, both of which use mutational analysis to describe the role of a gene in mouse development (Reith and Bernstein, 1991a). The first strategy uses various approaches to analyze gene function from genotype to phenotype. In this strategy, genes are identified and isolated on the basis of several criteria, e.g. homology to genes essential for development in other organisms, importance for cellular proliferation and tumorigenesis and screening techniques based on differential expression patterns. The developmental function of these genes can then be analyzed genetically by generation of gain or loss of function mutations through random or targeted insertion of exogenous DNA into the mouse germ line. A second strategy to understand gene function is the molecular analysis of existing mouse mutants from phenotype to genotype. One advantage of this approach is the existing presence of a mutant developmental phenotype. In addition to spontaneous mutations, the oldest of which have been described for nearly 100 years, novel mutants have been generated by radiation- and chemical-induced mutagenesis and by transgene insertional mutagenesis. The moiecular basis of these mutants is then analyzed either by association with previously cloned genes or by physical mapping and positional cloning, starting with closely linked markers. These two complementary strategies have provided a better understanding of the molecular mechanisms that regulate mammalian development. In this thesis, I have analyzed the molecular basis and developmental defects associated with two spontaneous regulatory mutations in the mouse dominant white sporting (W) locus. In chapter 2, I have described the effects of the ~57and wbanded(bd) mutations on Kit expression during embryogenesis and in the adult animal and showed specific defects in melanocyte development in these mutant animals. In addition, I have described the molecular basis of these regulatory W mutations. In chapter 3, we identified the Kit- expressing cells in the small intestine and demonstrated an important physiological function for this cell type. In chapter 4, I have analyzed the developmental origin and lineage relationships of these Kit-expressing cells in the small intestine and investigated the role of Kit in their development. In this introductory chapter, I wish to give some general information on the genetics of the W locus as well as on aspects of expression and function of the Kit receptor tyrosine kinase. This information will aid in a better understanding of the following chapters.

THE MOUSE W LOCUS AND ITS MUTANT ALLELES

The mouse W locus encodes the Kit RTK Mutations in the mouse Dominant white spotting (W)locus were fust described by Durham in 1908. In addition to the characteristic pigmentation defects, other developmental defects have since been described, including macrocytic anemia, mast cell deficiency and sterility (Silvers, 1979; Russell, 1979). Bone marrow transplantation and embryo chimaera experiments established that the cellular defects in W mutant mice are intrinsic to the progenitor cells that give rise to pigment, germ and blood cells (Silvers, 1979; Russell, 1979). However, the molecular nature of W was not established until 1988, when the Kit proto-oncogene, which encodes a receptor tyrosine kinase (RTK) (Yarden et al., 1987), was mapped to the W locus on mouse chromosome 5 and shown to be responsible for the W mutations (Chabot et al., 1988; Geissler etd., 1988). The Kit RTK, together with the receptors for colony- stimulating factor-1 (Csf-lr, encoded by the Fms gene) and -derived growth factors (Pdgfr), form the type III subfamily of RTKs (Qiu et al., 1988). The members of this RTK subfamily have an extracellular domain containing five immunoglobulin (1g)-like motifs, a single transmembrane domain and a cytoplasmic tyrosine base domain that contains an ATP-binding region and a phosphotransferase domain, separated by a kinase insert (Yarden and Ulrich, 1988; Hanks et al., 1988; Reith and Bernstein, 199 1b).

A cluster of coat color mutations on mouse chromosome 5 The W locus is not the only gene on chromosome 5 affecting coat color; indeed, several different mutations affecting coat color are clustered around the W locus (Searle and Tmslove, 1970). The Patch (Ph) locus involves a chromosomal deletion that removes the Pdgfra gene (Smith et al., 1991; Stephenson et al., 199I), which is located approximately 400 kb proximal to Kit (Brunkow et al., 1995). Animals heterozygous for the Ph mutation have white patches, whereas the homozygous mutation leads to embryonic lethality between days 9 and 16 of embryogenesis (Morris-Graham et al., 1992; OK-Urtreger et al., 1992). Lethality follows the appearance of several morphological abnormalities, including subepidermal blisters and open neural tube, followed by abnormal heart development and defects in non-neuronal derivatives of the cranial neural crest. The rump-white (Rw) mutation also maps close to the W locus and involves a large genomic inversion of the proximal one-third of mouse chromosome 5 (Stephenson et al., 1994). The proximal breakpoint of the Rw inversion maps close to the En-2 gene, whereas the distal breakpoint is within the Kit-Pdgfra intergenic region. Heterozygous Rw animals display a depigmentation of the rump, whereas the mutation in homozygous form causes developmental arrest at d8.5 of embryogenesis and pronounced necrotic changes or resorption at d9.5 (Bucan et al., 1995). Recessive spotting (rs)is another coat color mutation that maps very close to the W locus (Southard and Green, 1971), but the molecular basis for this mutation remains unknown. Animals homozygous for the rs mutation display a mild pigmentation defect with a white belly spot and occasional white head blaze (Dickie, 1966).

A cluster of receptor tyrosine kinases on mouse chromosome 5 The Kit and Pdgfra genes are closely linked and have parallel transcriptional orientations (Brunkow et al., 1995). Interestingly, the genes encoding the FMS and PDGFRB RTK,most closely related to Kit and Pdgfra, respectively, are also linked, on human chromosome 5 (Roberts et al., 1988). It is likely that this gene organization is maintained on mouse chromosome 18, where these two genes map. In addition, the gene for the mouse RTK Flk-1 maps approximately 275 kb distally to Kit, (Brunkow et al., 1995) and the closely related FLT-4 gene maps distally to the FMS gene in humans (Galland et al., 1992). A third RTK gene cluster includes the KIT-related FLK-2IFLT-3 and FLK-1-related FLT-1 genes, located on chromosome 13 in humans and near the distal end of chromosome 5 in the mouse (Rosnet et al., 1993). These data suggest a model whereby, prio r to the divergence of mouse and man, an ancestral RTK gene duplicated and diverged, and the resulting gene cluster subsequently underwent intrachromosornal duplications, leading to the generation of several unlinked clusters of RTK genes (Figure 1). Duplication Divergence

Inter- and intrachromosomal duplication

Pdgfra Kit = m-1 rnouechr. 5 - -i -

FMs - FLT-4 = human chr. 5

Figure 1. Schematic model for the generation of several unlinked clusters of RTK genes. Structural and regulatory mutations at the W locus A large number of W mutations have been described (Silvers, 1979), providing an opportunity to analyze the molecular basis for the developmental defects resulting from germ-line mutations in a RTK. The existing W mutations can be divided into structural mutations, which affect the coding sequence of the Kit gene, and regulatory mutations, which affect Kit (Reith et al., 1990). Animals homozygous for regulatory W alleles display tissue-specific defects, whereas animals carrying two copies of a structural allele usually are severely affected or, in some cases, are not viable and die in utero (Table 1) (Reith et al., 1991b). Most structural W loss-of-function mutations affect Kit kinase activity by either deletion of the whole or parts of the gene or by point mutations in the intracellular kinase domain (Figure 1) (Chabot et al., 1988; Reith et al., 1990; Nocka et al., 1990a; Tan et al., 1990, Reith et al., 1991b), thus leading to defective signal transduction (Nocka et al., 1989; Reith et al., 1991b). Studies of these loss-of-function mutations have identified amino acid residues that are essential for Kit base activity (Figure 1) ( Reith et al., 1991b; Besmer et al., 1993). Interestingly, all loss-of-function point mutations affect amino acids in either the intracellular ATP-binding or phosphotransferase domains, but not in the kinase insert or C-terminal domains (Reith and Bernstein, 1991 b). Dominant negative mutations in the KIT gene have also been demonstrated in humans with piebaldism (Giebel and Spritz, 1991). Piebaldism is an autosornal dominant defect in rnelanocyte development characterized by a white forelock, areas of hypopigmentation on the anterior trunk and extremities, and the absence of defects in other cell lineages (Giebel and Spritz, 1991). Several types of mutations have been reported in piebald patients, including missense and frameshift mutations in the KIT tyrosine kinase domain, a splice-junction mutation as well as mutations in the extracellular domain of the KIT receptor (Fleischman et al., 1991; Giebel and Spritz, 1991; Spritz et al., 1992; Ezoe et al., 1995; Fleischman et al., 1996). The mutation at codon 583 is identical to the strongly dominant negative ~37mutation, and, similar to the extensive coat color effects in mice, affected individuals exhibit unusually extensive depigmentation (Fleischman, 1992). In contrast to heterozygous ~37mice, no anemia is observed in these individuals. In addition to loss-of-function W mutations, several gain-of-function mutations have also been described. The viral v-Kit oncogene, isolated from a feline leukemia virus-associated feline fibrosarcoma in domestic cats, encodes a Kit protein with ligand-independent transforming potential (Besmer et al., 1986; Majumder et al., 1990; Herbst et al., 1995). These activating mutations in v-Kit include deletions of the extracellular and transmembrane domains, deletion of tyrosine-569 and valine- 570, substitution of aspartate at position 761 to glycine, and replacement of the C- terminal 50 amino acids by five unrelated residues (Herbst et d., 1995) (Figure 1). Recently, several gain-of-function mutations in the mouse Kit protein have been described, which result from mutations in the intracellular kinase domain (Furitsu et al., 1993; Tsujimura et al., 1994; Kitayama et al., 1995; Nagata et al., 1995; Piao and Bernstein, 1996; Tsujimura et d.,1996a; Kitayama et al., 1996) and lead to activation of the Kit receptor by either constitutive dimerimtion (Kitayama et al., 1995) or by change of substrate specificity and increased degradation of negative regulators of Kit (Piao et al., 1996) (Figure 2). In contrast to the large number of structural W mutations analyzed on the molecular level, little is known about the molecular nature of regulatory W mutations. The wsash(sh) mutation is associated with a 2.8 Mb inversion, with the distal breakpoint located close to the Tec gene and the proximal breakpoint located between Kit and Pdgfra, affecting tissue-specific expression of the Kit gene (Duttlinger et al., 1993; Duttlinger et al., 1995; Nagle et al., 1995) (Table 1). The WM mutation appears to affect Kit mRNA stability and is associated with a genomic insertion into intron 1 (Geissler et al., 1988) (Table 1). The Patch (Ph) mutation is associated with a deletion of the PdgFa gene and also disturbs the tissue-specific expression of Kit (Stephenson et al., 1991 ;Duttlinger et al., 1995) (see chapter 5, Discussion). Thus, mutations disturbing the tissue-specific expression of Kit provide an experimental approach for determining the molecular mechanisms that regulate the complex temporal and spatial patterns of Kit expression. Loss of function Gain of function

Kit

w 78 aa deletion w3' E+K wjic G+R wJ9 M+I

WV T+N t 814 D I) Y ~42N+D wf R-bW

~41V+M wn A+V

5 unrelated residues

Figure 2. Schematic illustration of structural gain and loss of function mutations in the Kit RTK. The open circles represent Immunoglobulin-like domains in the extracellular domain; the intracellular domain is subdivided into ATP-binding domain (narrow hatches) and a phosphotransferase domain (wide hatches). both of which are seperated by a kinase insert (open box). In addition, the C-terminal region is also represented by an open box. Mutations Hornozygotls Heterozygous MoIecular Type of phenotypes phenotypes lesion mutation

perinatd ventral spot, splice donor structural Durham, 1908 lethal no defects in site mutation, Nocka et d., 1990a hematopoiesis deletion of Hayashi et al., 1991 and fertility TMD, no cell- surface protein white. anemizoat color Little andcloudman,

mast cell ventral spot, reduced kinase dom. neg. Reith et d., 1990 deficiency, mild anemia activity Nocka et al., l99Oa steriIe

perinatal white, anemia' Asp790Asa stnrcturd, GeissIer et al., 198 1 lethal reduced no kinase dom.neg. Tan et al., 1990 fertility activity

eariy lethal ventral spot, 2 cM genomic structural Lyon et al., 1984 no defects in deletion Chabot et al., I988 hematopoiesis Geissler et al.. 1988 and fertility

peri natal mottled fur, Glu582Lys structural, Geissler et al., 198 1 lethal anemia no kina dom. neg. Reith et al., 1990 activity Nocka et al., 1990a

mottled fur, white spots, Val83 1Met structural, Geissler et al., 1981 miId anemia mild anemia reduced kinase weak dom. Reith et al., 1990 activity neg. Nocka et al., 1990a

mostly white, some spotting intronic regulatory Geissler et al., 188 1

reduced insertion, Geissler et al., 1988 fertility, not reduced mRNA anemic levels

Wsh white, mast white sash 2.8 Mb regulatory Lyon et d., I982 celI deficiency in trunk inversion, Tono et al., 1992 dysregulation of Ductlinger et al., 1993 rnRNA expression NagIe et al., 1995 wS7 mottled fur ventral spot lesion unknown, regulatory Reith et al., 1990 mainly in trunk reduced rnRNA region; head spot and protein in mast cells

@d white white band unknown regulatory? Beechey et al., 1986b in trunk

Table 1. Phenotypes and molecular bases of structural and regulatory W mutations. THE MOUSE Sl LOCUS AND ITS MUTANT ALLELES

The mouse SI locus encodes the ligand for the Kit RTK Mutations in the mouse Steel (St) locus on chromosome 10 also affect melanogenesis, hematopoiesis and gametogenesis and lead to phenotypes quite similar to those resulting from mutations in the W locus (reviewed in Reith and Bernstein, 1991 ; Galli et al., 1994) . However, in contrast to the cell-intrinsic defects caused by W mutations, mutations in the Sl locus lead to a defect in the microenvironment in which these cells develop during embryogenesis and function in the adult (Reith and Bernstein, 1991b; Galli et al., 1994). Consistent with the complementary nature of the cellular defects in mutant animals, SL has been shown to encode the ligand for the Kit receptor, alternatively called stem cell factor, mast cell growth factor, Steel factor and Kit ligand (Copeland et al., 1990; Huang et al., 1990; Zsebo et al., 1990). results in the synthesis of of both membrane- bound and soluble forms of Steel factor, suggesting that this ligand can activate Kit signalling either by direct cell-cell contact or by diffusion of the protein to target cells (Anderson et al., 1990; FIanagan and Leder, 1990; Flanagan et al., 199 1). In vitro studies indicate that the membrane-bound form of Steel factor is biologically more potent (Toksoz et al., 1992; Dolci et d., 1991; Matsui et al., 1991; reviewed in Galli et al., 1994); mice carrying two alleles of the ~ldickie(d)mutation, which deletes the transmembrane domain of the membrane-bound form of Steel factor and therefore only produces a functional soluble molecule, display defects in hernatopoiesis, melanogenesis and gametogenesis, suggesting a crucial in vivo role for the membrane-bound form of Steel factor (reviewed in Reith and Bernstein, 1991b; Besmer et al., 1993; Galli et al., 1994). In contrast, treatment of primordial germ cells, hematopoietic stem cells, melanocytes, mast cells and other cell types in culture with soluble Steel factor alone, or in combination with other growth factors, has profound biological effects (reviewed in Galli et al., 1994) (see below).

Structural and regulatory SZ alleles Sl alleles with severe phenotypes affect the coding region for Steel factor, whereas some of the less severe alleles affect the tissue-specific expression of the Steel factor gene (reviewed in Besmer et al., 1993) (Table 2). For example, animals homozygous for both the ~~~anda(~an)and the ~lconfrasfed(con)mutations are black-eyed whites, mildly anemic and females are sterile, while males are fertile (Beechey and Searle, 1983; Beechey et al., 1986a). The defects in female fertility are due to decreased Steel factor mRNA expression in embryonic and adult gonads, which causes sterility by affecting the initiation and maintenance of ovarian follicle development (Bedell et al., 1995). Both the SlWn and the Slcon mutations are associated with large genomic rearrangements 5' of the Steel factor gene, leading to a dysregulation of the temporal and spatial expression of Steel factor (Bedell et al., 1995). The SI~7H mutation consists of a point mutation in the 3' splice acceptor site of intron 7 that causes skipping of exon 8, resulting in a deletion of 23 out of 36 amino acids of the cytoplasmic tail of the Steel factor molecule (Brannan et al.. 1993). Mice hornozygous for this mutation are black-eyed whites and the males are sterile (Peters et al. 1987), raising the interesting possibility that the cytoplasmic tail of Steel factor might play a role in signalling to, or feedback from, Kit-expressing cells, rather than just passively anchoring the molecule within the plasma membrane (Brannan et al., 1992). Mutations Homozygorrs Molecrrlar Type of References phenotypes lesion mutation

SL. sLJ. sLgb, perinatal coat color dil. deletion of structural Sarvella and Russell, 1956 sLSHtslioH lethal on vend side, Steel factor Beechey and SearIe, 1985 occasional coding region Schaible, 196 1, 1963 spotting Copeland et al., 1990 Nocka et al., 1990 Zsebo et al., 1990

white, sterile, coat color dil., intragenic swctLUal, Bernstein, 1960 anemic spotting deletion, lacks cell Brannan et al.. I991 membrane- Hanagan et al., 1991 bound forms Huang et al., 1992

white, mild coat color dil. splice site mut., structural, Peters et al., 1987 anemia, frameshift at aa impaired Brannan et al., I992 males sterile 238 and 265 in function of cytoplasmic - domain bound forms

SPan,SLCon white,mild coat color dil. large genomic regdatory, Beechey and Seade, 1983 anemia, rearrangements dysregulation Beechey et al-, 1986 females sterile of temporal and Huang et al., 1993

TabIe 2. Phenotypes and molecular bases of structural and regulatory SZ mutations. SIGNAL TRANSDUCTION THROUGH THE KIT RTK

Like other RTKs, the Kit signalling cascade is initiated by binding of the dimeric ligand, Steel factor, to the three most N-terminal IG-like motifs in the Kit extracellular domain (Blechman et al., 1993; Lev et al., 1993). Ligand binding leads to rapid dimerization of the receptor molecules (Blume-Jensen et al., 1991; Lev et al., 1992), facilitated by the fourth Ig-like motif (Blechman et al., 1995). These dimerized receptors then undergo trans- at specific tyrosine residues in the intracellular domains (Rottapel et al., 1991). These phosphotyrosine residues and flanking amino acids act as high affinity docking sites for intracellular downstream target molecules, which bind to these phosphotyrosine residues via their Src homology 2 (SH2) domains (reviewed in Pawson, 1988). The SH2 domains associate with phosphotyrosine residues in a sequence context-dependent manner (reviewed in Pawson and Gish, 1992). Several downstream targets of Kit have been identified. For example, phosphatidylinositol 3'-kinase binds to the phosphorylated tyrosine residue 7 19 (Y7 19) in the kinase insert and becomes activated through this interaction (Serve et al., 1994). The activated Kit receptor also associates with Phospholipase Cy1 (PLCyl) (Rottapel et al., 199I), the Tec tyrosine kinase (Tang et al., 1994) and the protein tyrosine 1C (PTPIC or SHP-1) (Yi and Ihle, 1993). Furthermore, Kit activation leads to phosphorylation or activation of Shc (Cutler et al., 1993), Ras (Duronio et al., 1992), Vav (Alai et al., 1992), Raf-1 Lev et al., 1991) and MAP- base (Okuda et al., 1992; Welham and Schrader, 1992). Further downstream in the Kit signalling cascade, transcription of c-myc, c-myb, c-fos and c-jun is induced through Kit activation (Serve et al., 1995). This complex array of downstream signal transduction pathways that are activated by Kit may be a determining factor in the pleiotropic functions of the Kit RTK in development. REGULATION OF KIT EXPRESSION

Patterns of Kit expression The Kit gene is expressed in a variety of tissues and cell types in a highly specific temporal and spatial fashion (Nocka et al., 1989; Orr-Urtreger et al., 1990; Motro et al., 1991; Keshet et d., 1991; Yoshinaga et al., 1991; Motro and Bernstein, 1993; Bernex et al., 1996). The patterns of Kit expression clearly extend beyond those tissues in which phenotypic effects have been classically observed in W animals, and suggest that the Kit signalling pathway may also be important for the development of cells in the placenta, the peripheral and central nervous systems, the gut, the lung, the embryonic kidney and other tissues expressing Kit (On-Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991; Yoshinaga et al., 1991; Motro and Bernstein, 1993; Bernex et al., 1996). Expression during embryonic development is first observed at the two-cell stage. Kit expression has been studied most extensively after implantation. Kit is expressed at day 7.5 (d7.5) in ectodermal cells and the maternal decidua (Orr-Urtreger et al., 1990; Motro et al., 1991). At d8.5, Kit is expressed in the neural tube. the branchial arches, the intestinal tract, the lateral plate mesoderm and in the blood islands of the yolk sac (OK-Urtreger et al., 1990). Between dl0 and dl2 of embryogenesis, Kit is expressed in migrating primordial germ cells and rnelanoblasts, in the neural tube and in the intestinal tract. In addition, the fetal liver, the major hematopoietic organ at these stages of development, expresses high levels of Kit (Om- Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991). At later stages of embryonic development, Kit continues to be expressed in the developing central nervous system as well as in sensory neurons of the peripheral nervous system, in the lung, the gut, the gonads, the embryonic kidney, in melanocytes and mast cells (Orr- Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991). Expression in the liver decreases towards the end of gestation as the hematopoietic stem cells migrate to the bone marrow (Keshet et al., 1991). In the adult animal, Kit is expressed in chondrocytes and osteoblasts of the bone (Bemex et al., 1996) and in approximately 8 % of bone marrow cells, including hematopoietic stem cells, early progenitor cells and mast cells (Okada et al., 1991; reviewed in Bemstein, 1993). Expression in the adult kidney ceases whereas other tissues and cell types that expressed Kit during late embryogenesis continue to do so in the adult (Nocka et al., 1989; OK-Urtreger et al., 1990; Motro et al., 1991). Interestingly, Kit and its ligand Steel factor are expressed in distinct cell types in a complementary fashion throughout development. For example, in the adult ovary, Kit is expressed in oocytes, while Steel factor is expressed in the surrounding granulosa cells (Motro et al., 1991; Keshet et al., 1991). Reciprocal expression patterns are also found in adult testis and brain (Motro et al., 1991). Similarly, Steel factor is expressed along the migratory pathways of Kit-expressing melanoblasts and germ cells during embryonic development (Matsui et al., 1990; Keshet et al., 199 1).

Regulation of Kit expression Various and stimuli have been shown to influence Kit expression. Treatment of mast cells with phorbol myristate acetate (PMA) or IgE plus specific reduced levels of Kit mRNA by 50% (reviewed in Galli et al., 1994). Similarly, hematopoietic growth factors such as interleukin-3 (IL-3), erythropoietin (Epo) and - colony-stimulating factor (GM-CSF) also lead to a downregulation of Kit expression in mast cells and hematopoietic progenitor cell lines (reviewed in Galli et al., 1994). These observations suggest that regulation of Kit expression may be closely linked to the molecular mechanisms triggered by specific immunological and stimuli. Analysis of mice with mutations at the microphthalmia (mi) locus also has provided insight into the regulation of Kit expression. Animals homozygous for mutations at mi exhibit microphthalmia, depletion of pigment in hair and eyes, osteopetrosis and mast cell deficiency (reviewed in Silvers et d., 1979; Stechschulte et al., 1987). In addition, milmi mast cells fail to proliferate and mature in response to Steel factor (Ebi et al., 1992). These data suggested a role for the mi gene product in the Kit signaling pathway; indeed, rni/mi mast cells express significantly reduced levels of Kit (Ebi et al., 1992; Isozaki et al., 1994). Subsequently, mi was shown to encode a transcription factor of the basic-helix-loop-helix leucine zipper (bHLH-Zip) , named mi transcription factor (MITF) (Hodgkinson et al., 1993; Steingrimmsson et al., 1994). Analysis of the Kit promoter region in cultured primary mast cells identified a regulatory element at position -349 bp (relative to the transcription initiation site upstream of the first coding exon) which specifically binds MITF, thus leading to an upregulation of Kit expression (Tsujimura et al., 1996b). A mutant form of MITF was unable to bind to this regulatory element and failed to upregulate Kit expression (Tsujimura et al., 1996b). In contrast, promoter deletion studies in IL-3-independent murine mast cell Lines or in a human erythroleukemia cell hehave suggested that the genomic region containing this MITF binding element is not essential for Kit expression. These studies indicated that elements located at positions -105 bp to -44 bp are crucial for positive regulation of Kit expression (Yasuda et al., 1993; Yamamoto et al., 1993). Furthermore, one study also suggested that the region from -992 bp to -604 bp contains negative regulatory elements (Yamamoto et al., 1993). Several consensus binding sequences for known transcription factors have been found in the Kit upstream region, including binding sequences for A P-2, Spi-lIPu. I, and Spl (Yasuda et al., 1993; Yamamoto et al., 1993), but so far no functional relevance for these factors in controlling Kit expression have been demonstrated. Another study demonstrated binding of the c-Myb transcription factor to a approximately 5kb upstream of the fmt Kit coding exon (Vandenbark et d., 1996). Binding of Myb to this site led to a downreguIation of ectopic Kit expression in hematopoietic cell Lines that normally do not express Kit, suggesting a role for c-Myb as a silencer of Kit expression . In addition to the promoter upstream of the first coding exon, another promoter specifically utilized in premeiotic male germ cells has been described (Albanesi et al., 1996). Transcriptional initiation from this promoter, located in Intron 16, leads to the synthesis of truncated Kit transcripts and protein, lacking the extracellular, transmembrane and ATP-binding domains and consisting only of the kinase insert, phosphotransferase domain and 3' tail. The functional significance of this truncated product is unclear but raises the possibility that there is a kinase-independent function for the truncated Kit protein in spermatid development. Interestingly, the Fms gene, which encodes the receptor for colony-stimulating factor4 (CSF-1) and which is closely related to Kit, utilizes two tissue-specific promoters for its expression in monocytes and placental trophoblasts. One promoter is located immediately upstream of the fust Fms coding exon, while the other is located more than 25 kb upstream of the first coding exon (Roberts et al., 1988). In light of the evolutionary relationship between Kit and Fms, and the conservation of the genomic organization of these two genes, it is possible that Kit may also have two or more cell type-specific promoters. As discussed above, the Kit gene has a complex temporal and spatial pattern of expression. The exact phenotypic and genotypic characterization of spontaneous regulatory W mutations which disturb cell type-specific Kit expression could greatly enhance our understanding of the control of Kit expression as well as contribute to our knowledge of tissue- and cell type-specific transcription factors that regulate Kit expression. The wsh mutation has previously been analyzed and shown to dysregulate the spatial pattern of Kit expression during embryogenesis and in adults (Duttlinger et al., 1993; Duttlinger et al., 1995). The adult animals display a pigmentation defect and a mast cell deficiency (Lyon et al., 1982; Stevens and Loutit, 1982; Duttlinger et al., 1993). That is associated with a 2.8 Mb genomic inversion Located 5' of the Kit gene (Nagle et al., 1995). The breakpoints of this inversion are not well defined, and the effects of this inversion on cis-regulatory elements which have been implied in in vitro studies to be important for Kit expression in mast cells and melanocytes have not been determined.

PHYSIOLOGICAL ROLE OF THE KIT REXEPTOR DURING MOUSE DEVELOPMENT

Biological roles of the Kit signal transduction pathway In vitro studies have demonstrated that interactions of Steel factor with the Kit receptor can mediate various biological responses and modulate cell behaviour (reviewed in Galli et al., 1994). Adhesion. The membrane-associated form of Steel factor can mediate the adhesion of Kit-expressing mast cells (Flanagan et al., 1991; Kaneko et al., 1991; Adachi et al., 1992). This effect appears to require solely the extracellular domain of Kit, as mast cells derived from W mutant mice with no Kit tyrosine kinase activity exhibited no impairment of their ability to adhere to Steel factor-expressing fibroblasts (Adachi et al., 1992). Chemotaxis. Several of the cell lineages affected by W mutations, including melanocytes, germ cells and mast cells, follow complex migratory pathways during mouse embryonic development (reviewed in Galli et al., 1994). The expression of Kit in these cell lineages and Steel factor along the migration routes (Keshet et al., 1991) suggests that Kit-expressing cells could exhibit a chemotactic or haptotactic response to a concentration gradient of soluble or membrane-bound Steel factor, respectively. Indeed, in vitro experiments have demonstrated a chemotactic response of endothelid cells transfected with human Kit towards a source of recombinant Steel factor (Blume-Jensen et al., 1991). Moreover, wild type mast cells, but not W mutant mast cells, exhibit a chemotactic response to Steel factor, indicating that chemotaxis to Steel factor requires Kit tyrosine kinase activity as well as the extracellular domain of the Kit receptor (Meininger et al., 1992). Su~ival.Several studies have shown that soluble Steel factor maintains the short-term-survival of mouse primordial germ cells (PGC)(Dolci et al., 1991 ; Matsui et al., 199 l), hematopoietic stem cells (reviewed in Bernstein, 1993; Galli et al., 1994), human melanocytes (Funasaka et al., 1992), mouse and human mast cells (Zsebo et al., 1990; Mitsui et al., 1993) and human NK cell precursors (Carson et al., 1994) in vitro. In addition, the membrane-bound form of Steel factor can support PGC and hematopoiesis for longer periods than soluble Steel factor (Toksoz et al., 1992; Dolci et al., 199 1, Matsui et al., 1991). SUS~~mutant mice, which lack the membrane- associated form of Steel factor. cannot support survival of normal mast cells injected into the dermis (Gordon and Galli, 1990). Transplantation and embryo fusion studies utilizing W or Si mutant mice also indicate that Kit signalling is required for the survival of melanocytes and germ cells (reviewed in Silvers, 1979; Russell, 1979). ADOD~OS~S.Abrahamson et al. (1995) demonstrated that stimuIation of Friend erythroleukemia cells with Steel factor inhibits p53-mediated apoptosis, without affecting the cell cycle. These results further support the idea that the Kit signaling pathway plays an important role for cell survivd. Proliferation. The effects of Steel factor on cell proliferation are most dramatic in mast cells, where Steel factor alone induces proliferation of both immature and mature mast cells (Nocka et aI., 1990b; Tsai et al., 1991a; Tsai et al., 1991b). Steel factor also acts synergistically with IL-3 to promote mast cell proliferation in vitro (Galli et al., 1994). The effects of Steel factor done on proliferation of other cell types is only modest; however, in combination with other growth factors, Steel factor elicits synergistic proliferative responses. For example, Steel factor plus IL-3 promotes the proliferation of erythroid progenitor cells (BFU-E) and primitive rnyeloid cell lineages, and Steel factor plus IL-7 promotes the proliferation of primitive thymocytes (Galli et al., 1994). Steel factor, in combination with leukemia inhibitory factor (LIF) promotes the proliferation of mouse primordial germ cells (Matsui et al., 199 1). Human melanoblasts proliferate in vitro in the presence of S tee1 factor and PMA (Funasaka et al., 1992). Differentiation. Steel factor induces mast cell differentiation and maturation both in vitro and in vivo (Tsai et al., 199 la, 1991 b; Alexander et al., 199 1) an effect that is antagonized by IL-3 (Gurish et al., 1992). In other Kit-expressing cell types, activation of this signalling pathway alone does not promote differentiation (reviewed in Galli et al., 1994). Secretion. Steel factor can also induce Kit-dependent mast cell activation, mediator release, and mast cell-dependent inflammation (Wershil et al., 1992). In addition, Steel factor modulates IgE-stimulated mediator release in mast cells (Columbo et al., 1992). These experiments demonstrate that the Kit signalling pathway can regulate the secretory activity of mast cells. Oncogenic potential of the Kit receptor The first evidence that constitutive activation of the Kit receptor has oncogenic potential came from the analysis of the viral homologue of the Kit gene, v-. v-kit is the transforming gene of the Hardy-Zuckeman-4 strain of feline sarcoma virus, which was isolated from a feline leukemia virus-associated feline fibrosarcoma in domestic cats (Besmer et al., 1986). Leukemic blast cells from patients with acute myeloid leukemia (AML) express high levels of the Kit receptor and Steel factor stimulates their growth (Wang et al., 1991). Ectopic expression of Kit leads to coexpression of Kit and Steel factor in small-cell lung carcinomas (Hibi et al., 1991), gliomas (Stanulla et al., 1995), testicular tumors (Strohmeyer et al., 1995) and breast carcinomas (Hines et al., 1995), suggesting the possibility of an autocrine EWSteel factor signalling loop leading to malignant growth in these tumors. Loss of Kit expression has been observed in the majority of melanomas, whereas normal rnelanocytes do express the Kit receptor (Lassam et al., 1992; Funasaka et d., 1992; Halaban et al., 1992). It is not clear whether the reduced Kit expression in melanoma cells is causally related to the transformation event (Zakut et al., 1993) or is a consequence of dysregulated melanocyte gene expression. In addition, a subset of breast and thyroid carcinomas also do not express Kit anymore (Natali et al., 1992). This might suggest a role for the Kit receptor in maintaining cells in a differentiated state. Loss of Kit function would then cause a loss of differentiation characteristics, leading to malignant transformation. Activating mutations in the Kit receptor have been reported in the rnastocytoma cell Iine P815 (Rottapel et al., 1991). In this cell line, a point mutation in the cytoplasmic base domain of the Kit receptor (Tsujimura et al., 1994) (Figure 2) leads to constitutive signalling through the Kit receptor associated with alterations in substrate specificity and ubiquitination and degradation of a negative regulator of Kit function, the protein tyrosine phosphatase SHPl (Piao et al., 1996). The human mast cell leukemia cell line HMC-1 contains two point mutations in the Kit receptor, one located at the equivalent position as the mutation in the mouse P815 cell line, and the other in the intracellular jwctamembrane domain (Furitsu et al., 1993) (Figure 2). This second mutation leads to spontaneous dimerization of the Kit receptor and constitutive signalling in the absence of exogenous ligand (Furitsu et al., 1993). These examples indicate that the Kit RTK has several possible roles in tumorigenesis: gain-of-function mutations can lead to uncontrolled cell growth, whereas loss-of-function mutations suggest that a functional Kit receptor might also play a role as a tumor supressor. Interestingly, disturbances in Kit expression, either eliminating Kit expression or leading to ectopic Kit expression, are associated with several tumors, demonstrating the importance for a better understanding of the regulatory elements and transcription factors that control Kit expression. The role of the Kit signalling pathway in hematopoiesis Adult hernatopoiesis results from the continued proliferation and hierarchical differentiation of pluripotent stem cells to form the cells of the various hematopoietic lineages (reviewed in Keller, 19%; Bernstein, 1993; Paulson and Bernstein, 1995). The primary hematopoietic defects in W mutant mice are restricted to the erythroid and mast cell lineages, manifested as severe macrocytic anemia and mast cell deficiency. In addition, cells from W mutant bone marrow are unable to compete with wild type bone marrow in the reconstitution of the hematopoietic system of lethally irradiated mice. Thus, it appears that the Kit receptor plays an important role in regulating the function of both primitive pluripotent hematopoietic cells and Lineage- committed progenitor cells (reviewed in Russell, 1979; Bernstein et al., 1993; Paulson and Bernstein, 1995). In agreement with these phenotypic studies, Kit expression in the bone marrow is resmcted to hematopoietic stem cells and early lineage progenitor cells. Expression is downregulated in more mature, Lineage-committed cells with the exception of the mast cell lineage, where Kit is expressed at all stages of differentiation (reviewed in Bernstein, 1993; Galli et al., 1994; Paulson and Bernstein, 1995). In addition, Kit is expressed in early B cells and Steel Factor is a potent co-mitogen for B220+ pre-B cells, but not for more primitive pro-B cells, indicating a role for Kit signalling at specific stages of maturation (Rolink et al., 1991; Rico-Vargas et al., 1994; reviewed in Galli et al., 1994). Injection of an antagonistic monoclonal antibody (named Ack-2 for"anti c-kit") against the extracellular domain of the Kit receptor into adult mice results in a complete loss of rnyeloid progenitors and a severe decrease in CFU-S (colony-forming-units-spleen) activity (Ogawa et al., 1991). These results show that a functional Kit receptor is necessary for the maintenance of primitive pluripotential cells and the production of lineage-committed progenitors. Steel factor can also support the survival of pluripotent progenitor cells in vitro, and, together with other growth factors, can induce proliferation of several hematopoietic lineages (reviewed in Galli et al., 1994). The membrane-bound form of Steel factor supports the survival of hematopoietic stem cells for a longer period than the soluble form, indicating that cell-cell interactions between Kit-expressing hematopoietic stem cells and Steel factor-expressing stromal cells in the bone marrow play an important role in the maintenance of primitive hematopoietic stem cells (Toksoz et al., 1992). During embryonic development, primitive hematopoietic cells derived from the extraembryonic mesoderm are fmt observed in the blood islands of the yok sac starting at around d7.5. At d10, a intraembryonic site of hernatopoiesis is observed in the region around the dorsal aorta, the gonads and the mesonephros, the so-called AGM region. Cells from this region can give rise to all the hematopoietic lineages, suggesting that the AGM region is the first site of definite hematopoiesis (Miiller et al., 1994; Zon, 1995; reviewed in Paulson and Bernstein, 1995). In addition, stem cells from the AGM express the Kit receptor (Sanchez et al., 1996). Soon after, embryonic hematopoiesis, defined by the expression of specific globins and the restriction to the erythroid lineage, switches to fetal hernatopoiesis in the fetal liver, which produces all hernatopoietic lineages and is accompanied by a switch in the expression of specific globins. Later in gestation, adult hematopoiesis commences in the bone marrow, spleen and thymus, again accompanied by a switch in globin expression (reviewed in Paulson and Bernstein, 1995). The Kit RTK is expressed in hematopoietic cells at all stages. However, analysis of W mutant embryos demonstrated that early yolk sac hematopoiesis is not disturbed. Hornozygous W/W embryos only develop anemia at midgestation, shortly after the switch to fetal liver hematopoiesis (reviewed in Paulson and Bernstein, 1995). Injection of the Ack-2 monoclonal anibody at various embryonic stages interfered with fetal, but not yolk sac, hematopoiesis (Ogawa et al., 1991). Together, these results suggest that early yolk sac hematopoiesis is Kit-independent, whereas fetal hematopoiesis in the Liver is Kit-dependent.

The role of the Kit signalling pathway in melanocyte development Defects in skin pigmentation represent the primary identifying characteristic of all W and Sl mutations (Silvers, 1979) (Figure 3). The defect in W mutant animals is intrinsic to melanocytes, whereas the defect in Sl mutant animals affects the dermis and therefore the microenvironment of developing melanocytes (Silvers, 1979). Viable animals homozygous for structural W alleles are black-eyed whites, whereas heterozygous animals usually display patches of depigmentation in the midtrunk region as well as a head blaze (Silvers, 1979). Interestingly, some regulatory W alleles display pigmentation defects that differ significantly from the defects observed in mice carrying a mutated Kit receptor. For example, both WS~and ~bdheterozygous mice have a characteristic sharp band of depigrnentation in the midtrunk region, while mice homozygous for both mutations are black-eyed whites (Lyon et al., 1982; Beechey et al., 1986b) (Figure 3). Mice heterozygous for the ~57mutation are completely black with the exception of a small spot in the ventral trunk region, whereas homozygous ~57animals have mottled fur in the trunk region and a head blaze, resembling some heterozygous structural W mutations (Reith et al., 1990) (Figure 3). These regulatory mutations might affect melanocyte development by different mechanisms than structural W mutations and their analysis might provide new insights into melanocyte development and the role of the Kit RTK in melanogenesis. The melanoblasts that populate the skin and the choroid pigment layer of the eye are derived from the dorsal aspects of the neuroectoderm, called the neural crest, whereas melanoblasts in the pigment layer of the retina are derived from the optic cup. Neural crest-derived melanoblasts migrate out of the dorsal neural tube in an anterior to posterior sequential fashion, starting at around d9.5 of mouse embryogenesis in the cranial neural crest. While rnelanoblasts in the head migrate along several pathways, trunk melanoblasts always migrate on a dorsolaterd pathway along the surface ectoderm towards the somitic dermatome. Subsequently, melanoblasts disperse and proliferate in the ventrally expanding dermis. After formation of the dermis and the overlaying epidermis, melanoblasts migrate into the basal layers of the epidermis and into the hair follicles, where they subsequently proliferate, differentiate and produce pigment (LeDouarin, 1982). Kit is expressed in migrating melanoblasts and mature melanocytes, whereas Steel factor is expressed in the dermatome of the somites and in the dermis (Matsui et al., 1990; Keshet et al., 1990; Motro et al., 1991; Orr-Urtreger et al., 1991). These complementary patterns of Kit and Steel factor expression are consistent with the intrinsic pigmentation defect in W mice, and the microenvironmental defects in Sl mice (Figure 4).

Interestingly, rnelanoblasts from WS~mice expresss normal levels of Kit (Duttlinger et al., 1993). Duttlinger et al. have suggested that the defect in melanogenesis might be due to the ectopic expression of Kit observed in these mice in the dematome of the somites around dl 1. According to this model, ectopically expressed Kit in the dermatome sequesters Steel factor normally expressed in the same structure, leading to a lack of Steel factor available for the Kit-positive melanoblasts migrating towards the dermatome, thereby interfering with the migration or survival of rnelanoblasts. Indeed, wshl~shrnelanoblasts can be found between neural tube and dermatome in a position called the migration staging area (MSA), but seem to disappear shortly thereafter, in accordance with the proposed model (Duttlinger et al., 1993). To determine precisely when Kit signalling is required for important steps in melanocyte development, the Ack-2 monoclonal antibody was injected at various stages of embryonic development (Nishikawa et al., 1991; Yoshida et al., 1996). Injection of the antibody between embryonic days 10.5 and 15.5 resulted in stage- specific, distinct pigmentation defects, whereas antibody treatment before and after this time period had no effect on melanocyte development. These results indicate that a functional Kit receptor may not be important for the initial emigration of melanoblasts from the neural crest. During the migration towards the dermatome these melanoblasts acquire a Kit dependency which is maintained during dermis formation. There appears to be a short period of time before the formation of the epidermis where melanoblast development is not inhibited by the Ack-2 antibody. Migration into the epidermis is Kit-dependent, but the subsequent melanoblast differentiation in hair follicles is Kit-independent. In adult animals, Kit signalhg seems to be important for melanocyte proliferation in hair follicles that is activated during the initial phases of the hair cycle. The membrane-bound and soluble forms of Steel factor appear to be important for different stages in melanocyte development. In SUSl mutant embryos, which do not produce any Steel factor, melanoblasts migrate normally out of the neural crest, but are stalled in their migration in the MSA area between neural tube and dermatome and disappear shortly thereafter. In contrast, in SUS@ embryos, which produce the soluble form of Steel factor, melanoblasts do not halt in the MSA, but continue their migration to the dermatome, where the cells eventually disappear (Wehrle-Haller and Weston, 1995). These results suggest that Kit signalling is not important for the initial emigration of melanoblasts out of the neural tube. They also show that soluble Steel factor is sufficient for the migration of melanoblasts towards the dermatome, whereas the membrane-bound form appears to be necessary for the survival of melanoblasts once they reach the dermatome.

Figure 3. Pigmentation defects in different W alleles. (A) A nonagouti C57BU6J (+I+) mouse. (B)Homozygous ~571~57mouse on a C57BW6J background, displaying extensive, spotted depigmentation in the midtrunk region and a white head blaze. Mice heterozygous for the ~57mutation have a small white belly spot without the head blaze (not shown). (C)

Heterozygous Wban*ed(bd)/+ mouse, on a (C3H X 10l/H)Fl background, displaying a discrete "band" of depigmentation in the midtrunk region. Ln contrast to the wS7/CVS7mouse (see B), no pigmented areas are present within the white "band"; in addition, l&?/+ mice have no white head blaze. (D) a homozygous wb4wbd mouse (on a (C3H X IOl/H)Fl background) is a black-eyed white, resembling the phenotype of some combinations of structural alleles, e.g. W/WY and W/WV (not shown). (E) Heterozygous w~~/+mouse (on C57BU6J background), displaying depigmentation of the trunk and a large white head blaze due to a point mutation in the kinase domain of the Kit RTK. Mice homozygous for this mutation die in utero or shortly after birth from anemia. (F) A heterozygous ~ex@eme(~)/+ mouse (on a (C3WHeH X lOI/H)Fl background) displays extensive depigmentation of the trunk and head. Mice homozygous for this mutation die in utero or shortly after birth from anemia. Note the dilution of the coat color in the pigmented areas in (E) and 0,which represent structural W alleles, whereas this dilution is not observed in the regulatory alleles in

(A), (B)and (C).

late gestatio

migrating melanoblasts Kit-dependent differentiating melanocyte =@+ Kit-independent

Figure 4. Kit-dependent and -independent stages of melanocyte development. Neural crest- derived melanoblasts migrate out of the neural crest on a dorso-lateral pathway towards the MSA in a Kit-independent fashion. The migration from MSA towards the dermatome of the somites is Kit-dependent. During and after the dermis formation, a functional Kit receptor is required for melanoblast survival. After the melanoblasts migrate into the basal layers of the epidermis, their development becomes Kit-independent again. The role of the Kit receptor in ear development and hearing In the developing ear, neural crest-derived Kit-positive rnelanocytes are observed during embryogenesis (Steel et al., 1992). In the adult, Kit-expressing cells are seen in the epithelium of the tympanic membrane and the tuba pharyngovmpanica (Eustachian tube) as well as the stria vascularis of the cochlea in the i~erear (Steel et al., 1992). In adult mice with severe W mutations, rnelanocytes are absent in the stria vascularis, and this cellular defect is associated with an absence of endocochlear potential and hearing impairment (Steel and Barkway, 1989). Studies of different W alleles have demonstrated a good correlation between the presence of melanocytes arid endocochlear potential (Cable et al., 1995).

The role of the Kit signalling pathway in germ cell development Primordial germ cells (PGCs) originate from extraembryonic mesoderm at the base of the allantois adjacent to the caudal primitive streak. They are frst identifiable at d7 of mouse embryogenesis. Over the next four days, the PGCs proliferate and migrate from the hindgut to the gonadal ridge. By d12.5, the gonadal ridge has undergone sexual differentiation into the male or female gonad (reviewed in Galli et al., 1994). Kit is expressed in early and midgestational PGCs, as well as in adult oocytes (Orr-Urtreger et al., 1990; Manova et al., 1990; Manova et al., 1991; Motro et al., 1991; Motro and Bernstein, 1993). In the testis, Kit is expressed in the interstitial Leydig cells and in spermatogonia (Manova et al., 1990; Yoshinaga et al., 1991; Motro et al., 1991). Interestingly, spermatogonia express a truncated form of the Kit protein that lacks extracellular, transmembrane and the intracellular ATP-binding domain (Albanesi et al., 1996). The functional relevance of this soluble Kit molecule in male germ cell development is not clear. Steel factor is expressed in stromal cells dong the migratory route of PGCs (Matsui et al., 1WO), as well as in the granulosa cells surrounding the oocytes (Motro et al., 1991; Keshet et al., 1991). In mice homozygous for severe Wand Sl alleles, germ cells can be identified at d8 of embryogenesis, but subsequently do not increase in numbers and disappear (reviewed in Galli et al., 1994). In vitro and in vivo evidence indicated a role for Kit signalling in the survival and/or proliferation of PGCs (reviewed in Galli et al., 1994). The availability of mutant mice with severe mutations in either the W or SZ loci that are nevertheless compatible with viability have allowed the analysis of the role of Kit signalling in adult gonads. WWor nrYW mice not only have profound reductions in the numbers of germ cells, but also exhibit diminished rates of development of oocytes and spermatogonia (Coulombre and Russell, 1954). The SII 7H mutation affects only male, but not female fertility. SZ~~Hhomozygous females exhibit a significant reduction in primordial and mature oocytes, but are fertile. SZ~~H homozygous males also exhibit greatly reduced numbers of PGCs, but in addition have a cessation of sperm development after the first wave of spermatogenesis and are therefore sterile (Brannan et al., 1992). Accordingly, the splicing defect that leads to an abnormal cytoplasmic tail of the Steel factor molecule in ~117H/S17Hmice appears to impair both embryonic germ cell development and male postnatal germ cell development (Brannan et al., 1992). Kit signalling might also play a role in the estrous reproductive cycle of adult female mice. The cessation of oocyte growth is associated with the cessation of Steel factor expression in the granulosa cells surrounding the oocytes (Motro and Bernstein, 1993). In addition, the cyclic secretion of lutehizing hormone results in an immediate elevation of Steelfactor expression in granulosa cells and decreased levels of Kit expression in stromal-derived theca and interstitial cells, whereas Kit expression in oocytes is not affected by changes in hormone levels during the estrous cycle (Motro and Bernstein, 1993). The role of the Kit receptor in tissues with no obvious defects in W mutant mice After the Kit gene was shown to be allelic with the W locus (Chabot et aI., 1988; Geissler et al., 1988), it was surprising to fmd Kit expression in a wide variety of embryonic and adult tissues that exhibited no obvious defects in W mutant animals. Kit and its ligand Steel factor are expressed contiguously in many areas of the embryonic and adult brain, in sensory neurons of the peripheral nervous system (PNS), in embryonic and adult lung and gut, in embryonic kidney as well as in osteoclasts of the bone (Orr-Urtreger et al., 1990; Matsui et al., 1990; Motro et al., 1991; Keshet et al., 1991; Bernex et al., 1996). Until recently, the role of Kit signalling in these tissues and cell types has remained elusive. There might be several reasons for the unsuccessful search for additional phenotypes in W and S1 mutant mice: 1) Kit might not play a role in the development or function of some of the tissues or cell types it is expressed in. 2) Kit might play some role in these tissues and cell types, but redundancy with other signal transduction pathways, functionally overlapping with the Kit signalling pathway, mask any phenotype. 3) Kit plays a role in these tissues, but the phenotypes are subtle or have not been analyzed by appropriate means, e.g. lack of markers for specific cell types. Some of the Kit-expressing tissues have now been analyzed in some detail in wild type and Wand Sl mutant mice, providing insights into additional developmental and physiological roles of the Kit signalling pathway.

Brain. Kit is widely expressed in the embryonic and adult CNS. However, the numbers of Kit-positive cells are not affected by W or Sl mutations, suggesting that Kit signalling might only have a postmitotic function in cells in the CNS (Motro et al., 1991; Manova et al., 1992). Several studies have described Kit and Steel factor expression in pairs of neurons that form synaptic connections. In the cerebellum, for example, Steelfactor is expressed in hkinje cells of the granular layer, whereas Kit is expressed in basket and stellate cells of the molecular layer (Manova et al., 1992). In the olfactory bulb, tufted cells express Kit, while mitral cells and periglomerular cells express Steel factor (Hirota et al., 1992). Distinct cells in the cerebral cortex and the caudate putamen express Kit, whereas the paracentral and centrolateral nuclei of the thalamus express Steel factor (Hirota et d., 1992). These expression studies have raised the possibility that Kit/Steel factor interactions might play a role in the formation of synaptic connections. The contiguous and high level expression of Kit and Steel factor in the adult murine hippocampus supports a possible role for the Kit/Steel factor signalling pathway in hippocampal-dependent learning and memory. Steel factor is expressed in the hippocampal dentate gym, whose mossy fiber axons form synaptic connections with the Kit-expressing hippocampal CA3 pyramidal neurons (Motro et al., 1991). The hippocampus plays a critical role in configural learning, defined as the processes that establish and store a representation of the relations (e.g. spatial) between more than two stimuli (Sutherland and McDonald, 1990). An increase in synaptic efficiency, called long-term potentiation (LTP), has been suggested to be an important physiological mechanism for configural learning (Davis et al., 1992). W and Sl mutations do not lead to a reduction in the numbers of Kit-expressing cells in the hippocampus (Motro et al., 1991); in addition, the absence of membrane-bound Steel factor in SVSZ~adult mice does not cause detectable changes in granule cell axons and dendrites of the pyramidal neurons (Motro et al., 1996). However, configural learning tasks and electrophysiological analysis of hippocampal synaptic transmission demonstrated that the SVSZ~genotype leads to a deficiency in configural learning, yet does not affect LTP (Motro et al., 1995). These data suggest that impaired Kit signalling can lead to inefficient hippocampal transmission affecting configural leaming without affecting LTP and therefore that configural learning and LTP can be dissociated. PNS. Both Kir and Steel factor are expressed at high levels in adult and embryonic dorsal root ganglia (DRG) (Orr-Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991), which contain the cell bodies of sensory neurons. Steel factor can act as a neurotrophic factor by supporting the survival of a subset of DRG neurons in vitro (Hirata et al., 1993; Camahan et d., 1994). In addition, Steel factor can promote axonal outgrowth from Kit-positive DRG neurons in vitro, suggesting a role for KiVSteel factor interactions in axonal guidance (Hirata et al., 1993). Because no cellular deficits in DRGs of W or SZ mutant mice have been reported, a recent study analyzed several aspects of nerve regeneration in adult WWand SUSZ~mutant mice. The regeneration of damaged axons of specific DRG neurons was deficient in WW and SI/SZ~mutant mice, whereas another form of reparative nerve growth, the collateral sprouting of axonal endings into denervated target areas, was not affected in these animals. Kit expression was upregulated both proximally and distally to the site of nerve damage, suggesting an upregulation of Kit expression as part of a specific repair mechanism that allows damaged axons to regenerate. In addition, the authors also demonstrated reduced cell numbers for several DRG neuronal subpopulations, providing the fust example of a cellular deficit in the nervous system of W and Sl mutant mice (Lourrensen et al., manuscript in preparation).

Sensory organs. Both Kit and Steel factor are expressed in the developing and mature eye. Kit-expressing cells are found in the inner nuclear layer of the retina, whereas Steel factor is expressed in some cells of the outer nuclear layer (M. Kliippel, unpublished results; Bernex et al., 1996). These expression patterns are observed in both wild type and W-mutant mice, suggesting that the development of Kit- and Steel factor-expressing cells in the eye is not impaired by lack of Kit function (M. Kliippel, unpublished results). The digestive svstem. Kit expression is observed in the external muscle layers of the gut as well as in a subset of enterocytes, Paneth cells, Goblet cells and neuroendocrine cells in the gut epithelium (Ward et al., 1994; Bernex et al., 1996). Treatment of newborn mice with the Ack-2 monoclonal antibody to block Kit function led to the disappearance of unidentified Kit-positive cells in the smooth muscle layers of the gut. In addition, this treatment caused changes in gut motility, which, in the BALBIc genetic background, lead to a lethal paralytic ileus (Maeda et al., 1992; Torihashi et al., 1995). These observations raised the possibility that Kit may play a critical role in the development of a component of the intestinal pacemaker system, although these studies did not identify the cellular basis for the observed defects nor did any electrophysiological studies confii that indeed pacemaker activity was affected by anti-Kit antibody treatment.

The res~iratorysvstem. Kit is expressed in the majority of cells in the embryonic and adult lung as well as in the endodermal lining of the respiratory tube. No obvious defects in lung morphology or function have been observed in W or S1 mutant animals (Om-Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991; Bernex et al., 1996).

Bone. A subset of mature and proliferating chondrocytes of several ossification centers express Kit during embryogenesis and in adults (Bernex et al., 19961, but no skeletal defects have been described in W or SI mutant mice.

Teeth. Kit is expressed in the primordia of the tooth buds in embryogenesis, but no functional role of Kit in tooth development has been described (Bernex et al., 1996). Endothelid cells. During mouse embryogenesis, Kit is expressed during the de novo formation of blood vessels, called vasculogenesis, as well as during sprouting of pre-existing blood vessels, called angiogenesis. Endothelial cells continue to express Kit in adult animals, with higher Kit expression in the arterial endothelium and lower expression in veinous endothelium In humans, Kit and Steel factor are co-expressed by endothelial cells of the umbilical vein and aorta (Broudy et al., 1994; Buzby et d., 1994; Weiss et al., 1995; Bernex et al., 1996). No defects in angiogenesis or vasculogenesis have been described so far in W or SZ mutant animals.

Endocrine oreans. Several endocrine tissues express Kit during embryogenesis and in the adult. Expression is found in the thyroid gland, pituitary gland, pineal gland and adrenal gland. These expression patterns are unaltered in W mutant newborn animals, indicating that Kit is not required for embryonic development of endocrine cells (Bernex et al., 1996). The function of Kit in endocrine cell differentiation or function in adult mice remains to be studied.

Urogenital system. Kit is expressed in the embryonic kidney and in the bladder and ureter epithelial cells during embryogenesis and in newborn mice (Orr-Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991; Bernex et al., 1996). No defects in these tissues in W mutant animals have been described.

Early embryogenesis. Kit is first expressed in the 2-cell stage embryo. In blastocystes, scattered Kit-positive cells are seen in the trophectoderm and in the inner cell mass. A high percentage of inner cell mass-derived embryonic stem (ES) cells express Kit (Bernex et al., 1996). ES cells with severe mutations in the Kit gene nevertheless survive and give rise to embryos that develop normally until dl2 of embryogenesis, indicating that Kit is not required for early embryonic development. Extraembrvonic tissues. Both Kit and Steel factor are expressed in the placenta (Motro et al., 1991). At d7.5, Steel factor is expressed in the ectoptacental cone and at d9, both Kit and Steel factor are expressed in the trophoblast cells of the developing placenta. At d14.5, Kit is expressed in the maternal decidua while Steel factor is expressed in the labyrinth trophoblast layer. Because Kit and Steel factor expression patterns are indistinguishable in wild type and W or Sl mutant mice (Motro et al., 1991), the functional role of Kit signalling in the placenta is unknown.

\ The classical view that W and Sl mutations only affect the development of three lineages, melanocytes, primordial germ cells and erythroid cells, has been corrected in recent years by the discovery of additional defects in these animals. Detection of deficiencies in mast cells, in configural learning, in peripheral nerve repair and in some aspect of intestinal pacemaking has demonstrated a functional relevance for the Kit signalling pathway in other systems than the ones described in the classical literature. Thus, these data indicate that Kit signalling has an important function in several tissues and cell types in which this gene is expressed, but that some phenotypes associated with W and Sl mutant animals are either subtle or have not been analyzed appropriately.

The work presented in this thesis extends our understanding of the regulation of Kit expression and the role of the Kit RTK in melanocyte and gut development. In chapter 2, I have analyzed two regulatory W mutations and shown that both are associated with large genomic rearrangements in the vicinity of the Kit gene. These rearrangements cause temporal and spatial dysregulation of Kit expression and lead to distinct defects in melanocyte development. Analysis of one of these regulatory W mutations demonstrated for the first time distinct control mechanisms for Kit expression in subpopulations of neural crest-derived melanoblasts. In chapter 3, I have identified the Kit-expressing cells in the intestine as the interstitial cells of Cajal (KC) and provide biological evidence that these cells are an important part of the intestinal pacemaker system. In addition, I have analyzed the development of the KC and demonstrate in chapter 4 the developmental origin of the KC as well as the lineage relationship to other cell lineages in the developing gut. Abrahamson, J.L., Lee, J.M. and Bemstein, A. (1995). Regulation of p53-mediated apoptosis and cell cycle arrest by steel factor. Mol Cell. Bid. 15,6953-6960.

Adachi, S., Ebi, S.-I., Nishikawa, S.4, Hayashi, M., Yamazaki, T., Kasugai, T., Yamamura, S., Nomura, S. and Kitamura, Y. (1992). Necessity of extracellular domain of W (c-kit) receptors for attachment of murine cultured mast cells to fibroblasts. Blood 79,650.

Alai, M., Mui, A.L., Cutler, R.L., Bustelo, X.R.,Barbacid, M. and Krystal, G. (1992). Steel factor stimulates the tyrosine phosphorylation of the proto-oncogene product, p95vav, in human hemopoietic cells. J. Biol. Chem. 267, 1802 1- 18025.

Albanesi, C., Geremia, Re, Giorgio, M., Dolci, S., Sette, C. and Rossi, P. (1996). A cell- and developmental stage- specific promoter drives the expression of a truncated c-kit protein during mouse spermatid elongation. Development 122, 1291 - 1302.

Alexander, W.S., Lyman, S.D. and Wagner, E.F. (1991). Expression of functional c- kit receptors rescues the genetic defect of W mutant mast cells. EMBO J. 10, 3683.

Anderson, D.M., Lyman, S.D., Baird, A., Wignall, J.M., Eisenman, J., Rauch, C., March, C.J., Boswell, H.S., Gimpel, S.D., Cosman, D. and Williams, D.E. (1990). Molecular cloning of mast cell growth factor, a hematopoietin that is active in both membrane bound and soluble forms. CeII 63,235.

Bedell, M.A., Brannan, C.I., Evans, E.P., Copeland, N.G., Jenkins, N.A. and Donovan, P.J. (1995). DNA rearrangements located over 100 kb 5" of the Steel (So- coding region in Steel-panda and Steel-contrasted mice deregulate Sl expression and cause female sterility by disrupting ovarian follicle development. Genes Dev. 9,455- 470.

Beechey, C.V. and Searle, A.G. (1983). Contrasted, a Steel allele in the mouse with intermediate effects. Genet. Res. Carnb. 42, 183-191. Beechey, C.V.,Loutit, J.F. and Searle, A.G. (1986a). Panda, a new Steel allele. Mouse News Letter 74,92.

Beechey, C.V., Loutit, J.F. and Searle, A. G. (1986b). Banded, a new W allele. Mouse News L. 74,92.

Bernex, F., De Sepulveda, P., Kress, C., Elbaz, C., Delouis, C. and Panthier, J.-J. (1996). Spatial and temporal patterns of c-kit-expressing cells in ~lacZ/+and ~iacZ/WacZmouse embryos. Development 122,3023-3033.

Bernstein, A. (1993). Receptor tyrosine kinases and the control of hematopoiesis. Sem Dev. Biol. 4, 351-358.

Besmer, P, Murphy, J.E., George, P.C., Qui, F., Bergod, P.J., Ledeman, L., Snyder Jr., H.W., Brodeur, D., Zuckerman, E.E. and Hardy, W.D. (1986). A new acute transforming feline retrovirus and relationship of its oncogene v-kit with the gene family. Nature 320,415.

Besmer, P., Manova, K., Dunlinger, R., Huang, E.J., Packer, A., Gyssler, C. and Bacharova, R.F. (1993). The kit-ligand (steel factor) and its receptor c-kit/W pleiotropic roles in gametogenesis and melanogenesis. Development (Suppl.), 125- 137.

Blechman, L.M., Lev, S., Leitner, O., Givol, D. and Yarden, Y. (1993). Soluble Kit protein and anti-receptor monoclonal antibodies confine the binding site of the stem cell factor. J. Bid. Chem. 268,4399-4406.

Blechman, J.M., Lev, S., Barg, J., Eisenstein, M., Vaks, B., Vogel, Z., Givol, D. and Yarden, Y. (1995). The fourth immunoglobulin domain of the stem cell factor receptor couples ligand binding to signal transduction. Cell 80, 103- 113.

Blume-Jensen, P., CIaesson-Welsh, L., Siegbahn, A., Zsebo, K.M.,Westermark, B. and Heldin, C.H. (1993). Activation of the human c-kit product by ligand-induced dimerization mediates circular actin reorganization and chemotaxix. EMBO J. 10, 4121-4128. Brannan, C.I., Bedell, M.A., Resnick, J.L., Eppig, J.J., Handel, M.A., Williams, D.E., Lyman, S.D., Donovan, P.J., Jenkins, N.A. and Copeland, N.G. (1992). Developmental abnormalities in ~teell7Hmice result from a splicing defect in the steel factor cytoplasmic domain. Genes Dev. 6, 1832- 1842.

Broudy, V.C., Kovach, N.L., Bennett, L.G., Lin, E., Jacobsen, F.W. and Kidd, P.G. (1994). Human umbilical vein endothelial cells display high-affinity c-kit receptors and produce a soluble form of the c-kit receptor. Blood 83,2145-2152.

Brunkow, M.E., Nagle, D.L.,Bemstein, A. and Bucan, M. (1995). A 1.8-Mb YAC contig spanning three members of the receptor tyrosine kinase gene family (Pdgfra, Kit, and Flkl) on mouse chromosome 5. Genomics 25,421-432.

Bucan, M., Nagle, D.L., Hough, R.B., Chapman, V.M. and Lo, C.W. (1995). Lethality of Rw/Rw mouse embryos during early postimplantation development. Dev. Biology 168,307-3 18.

Buzby, J.S., Knoppel, E.M. and Cairo, M.S. (1994). Coordinate regulation of Steel Factor, its receptor (Kit), and cytoadhesion molecule (ICAM-1 and ELAM- 1) mRNA expression in human vascular endothelial cells of differing origins. Exp. Hematol. 22, 122-129.

Cable, J., Jackson, IJ. and Steel, K.P. (1995). Mutations at the W locus affect survival of neural crest-derived melanocytes in the mouse. Mech. Dev. 50, 139-150.

Carnahan, J.F., Patel, D.R. and Miller, J.A. (1994). Stem cell factor is a neurotrophic factor for neural crest-derived chick sensory neurons. J. Neurosci. 14, 1433-1440.

Carson, W.E., Haldar, S ., Baiocchi, R.A., Croce, C.M. and Caligiuri, M.A. (1994). The c-kit ligand suppresses apoptosis of human natural killer cells through the upregulation of bcl-2. Proc. Natl. Acad. Sci USA 91,7553-7557.

Chabot, B., Stephenson, D. A., Chapman, V. M., Besmer, P. and Bernstein, A. (1988). The proto-oncogene c-kit encoding a transmembrane tyrosine kinase receptor maps to the mouse W locus. Nature 335,88-89. Cohen, P.S., Chan, J.P., Lipkunskaya, M., Biedler, J.L. and Seeger, R.C. (1994). Expression of stem cell factor and c-kit in human neuroblastoma. The children's group. Blood 84,3465-3472.

Columbo, M., Horowitz, E.M., Botana, L.M., MacGlushan, D.W., Bochner, B.S., Gillis, S., Zsebo, K.M., Galli, S.J. and Lichtenstein, L.M. (1992). The human recombinant c-kit receptor ligand, rhSLF, induces mediator release from human cutaneous mast cells and enhances IgE-dependent mediator release from both skin mast cells and peripheral blood basophils. J. Immunol. 149,599.

Copeland, N.G., Gilbert, D. J., Cho, B. C., Donovan, P. J., Jenkins, N. A., Cosrnan, D, Anderson, D., Lyman, S. D. and Williams, D. E. (1990). Mast cell growth factor maps near the Steel locus on mouse chromosome 10 and is deleted in a number of Steel alleles. Cell 63, 175-183.

Coulombre, J.L. and Russell, E.S. (1954). J. Exp. Zool. 126,277-295.

Cutler, R.L., Liu, L., Damen, J.E. and Krystal, G. (1993). Multiple cytokines induce the tyrosine phosphorylation of Shc and its association with Grb2 in hematopoietic cells. J. Biol. Chem. 268,21463-21465.

Davis, S., Butcher, S.P. and Morris, R.G.(1992). The NMDA receptor antagonist D- 2-amino-5-phosphonopentanoate (D-APS) impairs spatial learning and LTP in vivo at introcerebral concentrations comparable to those that block LTP in vitro. J. Neurosci. 12,21-34.

Dickie, M.M. (1966). Mouse News Letter 35,3 1.

Dolci, S., Williams, D.E., Ernst, M.K., Resnick, J.L., Brannan, C.L., Lock, L.F., Lyman, S.D., Boswell, H.S. and Donovan, P.J. (1991). Requirement for mast cell growth factor for primordial germ cell survival in culture. Nature 352,809-8 11.

Durham, F.M. (1908). A preliminary account of the inheritance of coat color in mice. Rep. Evol. Cum. Roy. Sue. Rep. 4,4 1-53. Duronio, V., Welham, M.J., Abraham, S., Dryden, P. and Schrader, J.W. (1992). p21 ras activation via hernopoietin receptors and c-kit requires tyrosine kinase activity but not tyrosine phosphorylation of p2lras GTPase-activating protein. Proc. Natl. Acad. Sci. USA 89, 1587-1591.

Duttlinger, R., Manova, K., Chu, T. Y., Gyssler, C. Zelentz, A. D., Bachvarova R. F. and Besmer, P. (1993). W-sash affects positive and negative elements controlling c-kit expression: ectopic c-kit expression at sites of kit-ligand expression affects melanogenesis. Development 118,705-7 17.

Duttlinger, R., Manova, K., Berrozpe, G., Chu, T. Y., DeLeon, V., Timokhina, I., Chaganti, R. S. K., Zelenetz, A. D., Bachvarova, R. F. and Besmer, P. (1995). The whand Ph mutations affect the c-kit ex ression profile: c-kit misexpression in embryogenesis impairs melanogenesis in @K and Ph mice. Proc. Natl. Acad. Sci. 92, 3754-3758.

Ebi, Y., Kanakura, Y., Jippo-Kanemoto, T., Tsujimura, T., Furitsu, T., Ikeda, H., Adachi, S ., Kasugai, T., Nomura, S ., Kanayama, Y., Yamatodani, A., Nishikawa, S.I. and Kitamura, Y. (1992). Low c-kit expression of cultured mast cells of mihi genotype may be involved in their defective responses to fibroblasts that express the ligand for c-kit. Blood 80, 1454.

Ezoe, K., Holrnes, S.A., Ho, L., Bennett, C.P., Bolognia, J.L., Brueton, L., et al. (1995). Novel mutations and deletions of the KIT (steel factor receptor) gene in human piebaldism. Am J. Hum. Genet. 56,58-66.

Flanagan, J.G., Chan, D.C. and Leder, P. (1991). Transmembrane form of the kit ligand growth factor is determined by alternative splicing and is missing in the ~ld mutant. Cell 64, 10254035.

Fleischman, R.A., Saltman, D.L.,Stastny, V. and Zneirner, S. (1991). Deletion of the c-kit protooncogene in the human developmental defect piebald trait. Proc. Natl. Acad. Sci. USA 88,10885-10889.

Fleischman, R.A. (1992). Human piebald trait resulting from a dominant negative mutant dele of the c-kit membrane receptor gene. J. Clin Invest. 89, 17 13-17 17. Fleischman, R.A., Gallardo, T. and Mi, X. (1996). Mutations in the ligand-binding domain of the kit receptor: an uncommon site in human piebaldism. J. Invest. Dematol. 107,703-706.

Funasaka, Y., Boulton, T., Cobb, M., Yarden, Y., Fan, B., Lyman, S.D., Williams, D.E., Anderson, D.M.,Zakut, R., Mishima, Y. and Halaban, R. (1992). c-kit kinase induces a cascade of protein tyrosine phosphorylation in normal human melanocytes in response to mast cell growth factor and stimulates mitogen-activated protein kinase but is down-regulated in melanomas. Mol. Biol. Cell 3, 197-209.

Furitsu, T., Tsujimura, T., Tono, T., Ikeda, H., Kitayama, H., Koshimizu, U., Sugahara, H., Butterfield, J.H., Ashman, L.K., Kanayama, Y., Matsuzawa, Y., Kitamura, Y. and Kanakura, Y. (1993). Identification of mutations in the coding sequence of the proto-oncogene c-kit in a human mast cell leukemia cell line causing ligand-independent activation of c-kit product. J. Ch.Invest. 92, 1736- 1744.

Galland, F., Karamysheva, A., Mattei, M.-G.,Rosnet, O., Marchetto, S. and Birnbaum, D. (1992). Chromosomal localization of FLT4, a novel receptor-type tyrosine base gene. Genomics 13,475-478.

Galli, S.J., Zsebo, K.M. and Geissler, E.N. (1994). The kit ligand, stem cell factor. Adv. Immunol. 55, 1-96.

Geissler, E.N., McFarland, E.C. and Russell, E.S. (198 1). Analysis of pleiotropism at the dominant white-spotting (W) locus of the house mouse: a description of ten new W alleles. Genetics 97,337-36 1.

Geissler, EN., Ryan, M.A. and Housman, D.E. (1988). The dominant white spotting (W) locus of the mouse encodes the c-kit proto-oncogene. Cell 55, 185-192.

Giebel, L.B. and Spritz, R.A. (1991). Mutation of the KIT (mastktem cell ) protooncogene in human piebaldism. Proc. Natl. Acad. Sci. USA 88, 8696-8699. Gordon, J. and Galli, S.I. (1990). Phorbol 12-rnyristate 13-acetate-induced development of functionally active mast cells in W/CVV but not SUS@ genetically mast cell-deficient mice. Blood 75, 1637.

Gurish, M.F., Ghildyal, N., McNeil, H.P.,Austen, K-F., Gillis, S. and Stevens, R.L. (1992). J. Exp. Med. 175, 1003-1012.

Halaban, R., FaTB., Ahn, J., Funasaka, Y., Gitay, G.H. and Neufeld, G. (1992). Growth factors, receptor kinases, and protein tyrosine in normal and malignant melanocytes. J. Immunother. 12, 154- 16 1.

Hanks, S.K., Quim, A.M. and Hunter, T. (1988). The protein kinase family: conserved features and deduced phylogeny of the catalytic domains. Science 241,42- 52.

Hayashi, S., Kunisada, T., Ogawa, M., Yamaguchi, K. and Nishikawa, S.-I. (1991). Exon skipping by mutation of an authentic splice site of c-kit gene WNmouse. Nucleic Acid Res. 19, 1367- 1271.

Herbst, R., Munernitsu, S. and Ullrich, A. (1995). Oncogenic activation of v-kit involves deletion of a putative tyrosine-substrate interaction site. Oncogene 10, 369- 379.

Hibi, IS., Takahashi, T., Sekido, Y., Ueda, R., Hida, T., Ariyoshi, Y., Takagi, H. and Takahashi, T. (1991). Coexpression of the stem cell factor and the c-kit genes in small-cell lung cancer. Oncogene 6,229 1-2296.

Hines, S.J., Organ, C., Kornstein, M.J. and Krystal, G.W. (1995). Coexpression of the c-kit and stem cell factor genes in breast carcinomas. Cell Growth Dzfi 6,769-779.

Hirata, T.,Morii, E., Morimoto, M., Kasugai, T., Tsujimura, T., Hirota, S., Kanakura, Y., Nornura, S. and Kitamura, Y. (1993). Stem cell factor induces outgrowth of c-kit- positive neurites and supports the survival of c-kit-positive neurons in dorsal root ganglia of mouse embryos. Development 119,49-56. Hirata, T., Kasugai, T., Morii, E., Hirota, S., Nomura, S., Fujisawa, H. and Kitamura, Y. (1995). Characterization of c-kit-positive neurons in the dorsal root ganglion of mouse. Dev. Brain Res. 85,201-2 11.

Hirota, S., Ito, A., Morii, E., Wanaka, A., Tohyama, M., Kitamura, Y. Nomura, S. (1992). Localization of mRNA for c-kit receptor and its Ligand in the brain of adult rats: an analysis using in situ hybridization histochemistry Brain. Res. Mol. Brain Rex 15,47-54.

Hodgkinson, C.A., Moore, K.J., Nakayama, A., Steingrimsson, E., Copeland, N.G., Jenkins, N.A. and Arnheiter, H. (1993). Mutations at the mouse Microphtalmia locus are associated with defects in a gene encoding a novel basic-helix-loop-helix-zipper protein. Cell 74,395-404.

Huang, E., Nocka, K., Beier, D. R., Chu, T. Y., Buck, J., Lab, H.-W., Wellner, D., Leder, P. and Besmer, P. (1990). The hematopoietic growth factor KL is encoded by the Sl locus and is the ligand of the c-kit receptor, the gene product of the W locus. Cell 63,225-233.

Isozaki, K., Tsujirnura, T., Nomura, S., Morii, E., Koshirnizu, U., Nishimune, Y. and Kitamura, Y. (1994). Cell-type specific deficiency of c-kit gene expression in mutant mice of rnihmi genotype. Am. J. Pathol. 145,827.

Kaneko, Y., Takenawa, J., Yoshida, O., Fujita, K., Sugimoto, K., Nakayama, H. and Fujita, J. (1991). Adhesion of mouse mast cells to fibroblasts: Adverse effects of steel (sl) mutation. J. Cell. Physiol. 147,224.

Keller, G. (1992). Clonal analysis of development in vivo. Curr. Topics Microbial. Imm. 177,41-57.

Keshet, E., Lyman, S. D., Williams, D. E., Anderson, D. M., Jenkins, N. A., Copeland, N. G. and Parada, L. F. (1991). Embryonic RNA expression patterns of the c-kit receptor and its cognate ligand suggest multiple functional roles in mouse development. EMBO J. 10,2425-2435. Kitayama, H., Kanakura, T., Furitsu. T., Tsujimura, K., Oritani, K., Ikeda, H., Sugahara, H., Mitsui, H., Kanayama, Y and Kitamura, Y. (1995). Constitutively activating mutations of c-kit receptor tyrosine base confer factor-independent growth and tumorigenicity of factor-dependent hematopoietic cell lines. Blood 85, 790-798.

Kitayama, H., Tsujimura, T., Matsumura, I., Oritani, K., Ikeda, H., Ishikawa, J., Okabe, M., Suzuki, M., Yamamura, K.4, Matsuzawa, Y., Kitamura, Y. and Kanakura, Y. (1996). Neoplastic transformation of normal hematopoietic cells by constitutively activating mutations of c-kit receptor tyrosine kinase. Blood 88, 995- 1004.

Lassam, N. and Bickford, S. (1992). Loss of c-kit expression in cultured melanoma cells. Oncogene 7,s1-54.

LeDouarin, N. (1982). The neural crest. Cambridge, Cambridge University Press.

Lev, S., Givol, D. and Yarden, Y. (1991). A specific combination of substrates is involved in signal transduction by the kit-encoded receptor. EMBO J. 10, 647-654.

Lev, S., Yarden, Y. and Givol, D. (1992). Dimerization and activation of the Kit receptor by monovalent and bivalent binding of the stem cell factor. J. Bid. Chem. 267, 15970- 15977.

Lev, S., Blechman, J., Nishikawa, S.-I., Givol, D. and Yarden, Y. (1993). Interspecies molecular chimeras of Kit help define the binding site of the stem cell factor. Mol. Cell. Biol. 13,2224-2234.

Little, C.C. and Cloudman, A.M. (1937). The occurrence of a dominant spotting mutation in the house mouse. Proc. Natl. Acad. Sci. USA 23,535-537.

Longley, J., Tyrrell, L., Lu, S.-Z., Ma, Y.-S., Langley, K., Ding, T.-G., Durn, T., Jacobs, P., Tang, L.H. and Modlin, I. (1996). Somatic c-kit activating mutation in urticaria pigmentosa and agressive masteocytosis: establishment of clonality in a human mast cell neoplasm. Nature Gen. 12,3 12-3 14. Lyon, M. F., and Glenister, P. H. (1982). A new allele sash (*h ) at the W-locus and a spontaneous recessive lethal in mice. Genet. Res. 39,315-322.

Maeda, H., Yamagata, A., Nishikawa, S., Yoshinaga, K., Kobayashi, S., Nishi, K. and Nishikawa, S . ( 1992). Requirement of c-kit for development of intestinal pacemaker system. Development 116,369-375.

Majurnder, S., Ray, P. and Besmer, P. (1990). Tyrosine protein base activity of the W%-feline sarcoma virus P80gag-kit-transforming protein. Oncogene Res. 5, 329-

Manova, K. and Bachvarova, R. F. (199 1). Expression of c-kit encoded at the W locus of mice in developing embryonic germ cells and presumptive melanoblasts. Dev. Biol. 146,3 12-324.

Manova, K., Bacharova, R.F., Huang, E.J., Sanchez, S., Pronovost, S.M., Velazquez, E., McGuire, B. and Besmer, P. (1992). c-kit receptor and ligand expression in postnatal development of the mouse cerebellum suggests a function for c-kit in inhibitory interneurons. J. Neurosci. 12,4663-4676.

Matsui, Y., Zsebo, K. M. and Hogan, B. L. M. (1990). Embryonic expression of a haematopoietic growth factor encoded by the Sl locus and the ligand for c-kit. Nature 347,667-669.

Matsui, Y., Toksoz, D., Nishikawa, S., Nishikawa, S.4, Williams, D., Zsebo, K. and Hogan, B.L.M. ( 1991). Effect of Steel factor arid leukemia inhibitory factor on murine primordial germ cells in culture. Nature 353,750-752.

Meininger, C.J., Yano, H., Rottapel, R., Bernstein, A., Zsebo, K.M. and B.R. (1992). The c-kit receptor ligand functions as a mast cell chemoattractant. Blood 79,958-963.

Mitsui, H., Furitsu, T., Dvorak A.M., Irani, A.M., Schwartz, L.B., Inagaki, N., Takei, M., Ishizaka, K., Zsebo, K.M.,Gillis, S. et al. (1993). Proc. Natl. Acad. Sci. USA 90, 735-739. Moms-Graham, K., Schatteman, G.C., Bork, T., Bowen-Pope, D.F. and Weston, J.A. (1992). A PDGF receptor mutation in the mouse (Patch) perturbs the development of a non-neuronal subset of neural crest-derived cell. Development 115, 133- 142.

Motro, B., Van Der Kooy, D., Rossant, J., Reith, A. and Bernstein, A. (1991). Contiguous patterns of c-kit and steel expression: analysis of mutations at the W and Sl loci. Development 113, 120%122 1.

Motro, B. and Bernstein, A. (1993). Dynamic changes in ovarian c-kit and Steel expression during the estrous reproductive cycle. Dev. Dyn. 197,69-79.

Motro, B., Wojtowicz, J.M., Bernstein, A. and Van der Kooy, D. (1995). Steel mutant mice are deficient in hippocampal learning but not long-term potentiation. Proc. Natl. Acad. Sci. USA 93, 1808- 18 13.

Miiller, A.M., Medvinsky, A., Strouboulis, J., Grosveld, F. and Dzierak, E. (1994). Development of hematopoietic stem cell activity in the mouse embryo. Immunity 1, 291-301.

Nagata, H., Worobec, A.S., Oh, C.D., Chowdhury, B.A., Tannenbaum, S., Suzuki, Y. et al. (1995). Identification of a point mutation in the catalytic domain of the protooncogene c-kit in peripheral blood mononuclear cells of patients who have mastocytosis with an associated hematologic disorder. Proc. Natl. Acad. Sci. USA 92, 10560- 10564.

Nagle, D.L., Kozak, C.A., Mano, H., Chapman, V.M. and Bucan, M. (1995). Physical mapping of the Tec and Gabrbl loci reveals that the WS~mutation on mouse chromosome 5 is associated with an inversion. Human Mol. Gen. 4,2073-2079.

Natali, P.G.,Nicotra, M.R.,Sures, I., Mottolese, M.,, Bom, C. and Ullrich, A. (1992). Breast cancer is associated with loss of the c-kit oncogene product. Intern. J. Cancer 52,713-717.

Nishikawa, S., Kusakabe, M., Yoshinaga, K., Ogawa, M., Hayashi, S., Kunisada, T., Era, T., Sakakura, T. and Nishikawa, S. (1991). In utero manipulation of coat color formation by a monoclonal anti-c-kit antibody: two distinct waves of c-kir dependency during melanocyte development. EMBO J. 10,21 11-2 1 18.

Nocka, K., Majumder, S., Chabot, B., Ray, P., Cervone, M., Bernstein, A. and Besmer, P. (1989). Expression of c-.bir gene products in known cellular targets of W mutations in normal and W mutant mice-evidence for an impaired c-kit kinase in mutant mice. Genes Dev. 3.8 16-826.

Nocka, K., Tan, LC., Chiu, T.Y.,Chu, P., Ray, P., Traktman, P. and Besmer, P. (1990a). Molecular bases of dominant negative and loss-of-function mutations at the murine c-kihhite spotting locus: ~37,Wv, and W. EMBO J. 9, 1805- 18 13.

Nocka, K., Huang, E., Beier, D.R., Chu, T.V., Ray, P., Traktman, P. and Besmer, P. (1990b). The hematopoietic growth factor KL is encoded by the Sl locus arid is the ligand of the c-kit receptor, the gene product of the W locus. Cell 63,225-233.

Ogawa, M., Matsuzaki, Y., Nishikawa, S., Hayashi, S.-I., Kunisada, T., Sudo, T., Kina, T., Nakauchi, H. and Nishikawa, S.4. (1991). Expression amd function of c-kit in hematopoietic progenitor cells. J.Exp. Med. 174,63.

Okada, S., Nakauchi, H., Nagayoshi, K., Nishikawa, S., Nishikawa, S.-I., Miura, Y. and Suda, T. (199 1). Enrichment and characterization of murine hematopoietic stem cellsthat express the c-kit molecule. Blood 78, 1706- 17 12.

Okuda, K., Sanghera, J.S., Pelech, S.L., Kanakura, Y., Hallek, M., Griffin, J.D. and Druker, B .J. ( 1992). Granulocyte-macrophage colony-stimulating factor, interleukin- 3, and steel factor induce rapid tyrosine phosphorylation of p42 and p44 MAP kinase. Blood 79,2880-2887.

Orr-Urtreger, A., Avivi, A., Zimmer, Y., Givol, D., Yarden, Y. and Lonai, P. (1990). Developmental expression of c-kit, a proto-oncogene encoded by the W locus. Development 109,9 11-923.

Orr-Urtreger, A., Bedford, M.T., Do, M.-S.,Eisenbach, L. and Lonai, P. (1992). Developmental expression of the a receptor for platelet-derived growth factor, which is deleted in the embryonic lethal Patch mutation. Development 115,289-303. Paulson, R. and Bernstein, A. (1995). Receptor tyrosine kinases and the regulation of hematopoiesis. Sem. Immunol. 7,267-277.

Pawson, T. ( 1988). Non-catalytic domains of cytoplasmic protein-tyrosine kinases: regulatory elements in signal transduction. Oncogene 3,49 1-495.

Peters, J., Ball, S.T. and Loutit, J.F. (1987). A new Steel allele which does not lead to dilution of coat color. Mouse News Letter 77, 125- 126.

Piao, X. and Bernstein, A. (1996). A point mutation in the catalytic domain of c-kit induces growth factor independence, tumorigenicity, and differentiation of mast cells. Blood 87,31 1 7-3 123.

Piao, X., Paulson, R., Van der Geer, P., Pawson, T. and Bernstein, A. (1996). Oncogenic mutations in the Kit receptor tyrosine kinase alters substrate specificity and induces degradation of the protein tyrosine phosphatase SHP-1.Proc. Natl. Acud. Sci. USA 93, 14665- 14669.

Qiu, F., Ray, P., Brown, K., Barker, P.E., Jhanwar, S., Ruddle, F.H. and Besmer, P. (1988). Primary structure of c-kit: relationship with the CSF- l/PDGF receptor kinase family-oncogenetic activation of v-kit involves deletion of extracellular domain and C-terminus. EMBO J. 7, 1003.

Reith, A.D., Rottapel, R., Giddens, E., Brady, C., Forrester, L. and Bernstein, A. (1990). W mutant mice with mild or severe developmental defects contain distinct point mutations in the kinase domain of the c-kit receptor. Genes Dev. 4,390-400.

Reith, A.D. and Bernstein, A. (1991a). Molecular basis of mouse developmental mutants. Genes Dew. 5, 11 15- 1123.

Reith, A.D. and Bernstein, A. (1991b). Molecular biology of the Wand Steel loci. in: Genome Analysis 3: Genes and Phenotypes. 105- 133. Reith, A.D., Ellis, C., Lyman, S.D., Anderson, D.M., Williams, D.E., Bernstein, A. and Pawson, T. (1991). Signal transduction by normal isoforms and W mutant variants of the Kit receptor tyrosine kinase. EMBO J. 10,2451-2459.

Rico-Vargas, S.A., Weiskopf, B., Nishikawa, S. and Osmond, D.G. (1994). c-kit expression by B-cell precursors in mouse bone marrow. Stimulation of B-cell genesis by in vivo treatment with anti-c-kit antibody. J. Immunol. 152,2845-2852.

Roberts, W.M., Look, A.T., Roussel, M.F. and Shen, C.J. (1988). Tandem linkage of human CSF- l receptor (c-fms)and PDGF receptor genes. Cell 55,655-661.

Rolink, A., Streb, M., Nishikawa, S. and Melchers, F. (1991). The c-kit-encoded tyrosine kinase regulates the proliferation of early pre-B cells. Eur. J. Immunol. 21, 2609-26 12.

Rosnet, O., Stephenson, D.A., Mattei, M.-G., Marchetto, S., Shibuya, M., Chapman, V.M. and Birnbaum, D. (1993). Close physical linkage of the FLTI and FLT3 genes on chromosome 13 in man and chromosome 5 in mouse. Oncogene 8, 173- 179.

Rottapel, R., Reedijk, M., Williams, D.E., Lyman, S.D., Anderson, D.M., Pawson, T. and Bemstein, A. (1991). The Steel/' transduction pathway: kit autophosphorylation and its association with a unique subset of cytoplasmic signaling proteins is induced by the Steel factor. Mol. Cell. Biol. 11,3043-305 1.

Russell, E.S.(1979). Hereditary anemias of the mouse: a review for geneticists. Adv. Genetics 20, 357-459.

Sanchez, M.L., Holrnes, A., Miles, C. and Dzierak, E. (1996). Characterization of the first definitive hematopoietic stem cells in the AGM and liver of the mouse embryo. Immunity 5,5 13-525.

Searle, A.G. and Truslove, G.M. (1970). A gene triplet in the mouse. Genetical Res. 15,227-235. Serve, H., Yee, N.S., Stella, G., Sepp, L.L., Tan, J.C. and Besmer, P. (1995). Differential roles of PI3-base and Kit tyrosine 821 in Kit receptor-mediated proliferation, survival and cell adhesion in mast cells. EMBO J. 14,473-483.

Silvers, W.K. (1979). The coat colors of mice. Springer Verlag, Heidelberg.

Smith, E.A., Seldin, ME, Martinez, L, Watson, M.L.,Choudhury, G.G., Lalley, P.A., Pierce, J., Aaronson, S., Barker, J., Naylor, S.L. and Sakaguchi, A.Y. (1991). Mouse platelet-derived growth factor receptor a gene is deleted in ~19Hand Patch mutations on chromosome 5. Proc. Natl. Acad. Sci. USA 88,48 11-48 15.

Southard, J.L. and Greene, M.C. (1971). Mouse News Letter 45,29.

Spritz, R.A., Holrnes, S.A., Ramesar, R., Greenberg, I., Curtis, D. and Beighton, P. (1992). Mutations of the KIT (mastlstem cell growth factor receptor) proto-oncogene account for a continuous range of phenotypes in human piebaldism. Am. J. Hum. Genet. 51, 1058-1065.

Stanulla, M., Welte, K., Hadam, M.R. and Pietsch, T. (1995). Coexpression of stem cell factor and its receptor c-Kit in human malignant glioma cell lines. Acta Neuropathol. 89, 158- 165.

Stechschulte, D.J.R., Sharma, K.N., Dileepan, K.M., Simpson, N., Aggarwal, J., Clancy, J.R. and Jilka, R.L. (1987). Effects of the mi allele on mast cells, basophils, natural killer cells, and osteoclasts in C57BU6J mice. J. Cell. Physiol. 132,565.

Steel, K.P. and Barkway, C. (1989). Another role for rnelanocytes: their importance for normal stria vascularis development in the mammalian inner ear. Development 107,453-463.

Steel, K.P., Davidson, D.R. and Jackson, I.J. (1992). TRP-2DT, a new early melanoblast marker, shows that steel growth factor (c-kit ligand) is a survival factor. Development 115, 11 1 1-1 119.

Steingrimsson, E., Moore, K.J., Lamoreux, L., Ferre-D' Amare, A.R., Burley, S.K., Sanders Zimring, D.C., Skow, L.C., Hodgkinson, C.A., Amheiter, H., Copeland, N.G. and Jenkins, N.A. (1994). Molecular basis of mouse microphtalmia (mi) mutations helps explain their developmental and phenotypic consequences. Nature Gen. 8, 256- 263.

Stephenson, D.A., Mercola, M., Anderson, E., Wang, C., Stiles, C.D., Bowen-Pope, D.F. and Chapman, V.M. (1991). Plarelet-derived growth factor receptor a-subunit gene (Pdgfra) is deleted in the mouse Patch (Ph) mutation. Proc. Natl. Acad. Sci. USA 88,6-10.

Stephenson, D.A., Lee, K., Nagle, D.L., Yen, C.-H.,Morrow, A., Miller, D., Chapman, V.M. and Bucan, M. (1994). Rump-white mutation is associated with an inversion of the proximal portion of mouse chromosome 5. Mammalian Genome 5, 342-348.

Stevens, J. and Loutit, J.F. (1982). Mast cells in mutant mice (W Ph mi). Proc. Roy. Soc. Land. B. 215,405-409.

Strohmeyer, T., Reese, D., Press, M., Ackermann, R., Hartmann, M. and Slamon, D. (1995). Expression of the c-kit proto-oncogene and its ligand stem cell factor (SCF) in normal and malignant human testicular tissue. J. Urol. 153,51 1-5 15.

Sutherland, R.J. and McDonald, R.J. (1990). Hippocampus, amygdala, and memory deficits in rats. Behav. Brain Res. 37,57-79.

Tan, J.C., Nocka, P., Ray, P., Traktman, P. and Besmer, P. (1990). The dominant ~42 spotting phenotype results from a missense mutation in the c-kit receptor kinase. Science 247,209-2 12.

Tang, B., Mano, H., Yi, T. and Me, J.N. (1994). Tec kinase associates with c-kit and is tyrosine phosphorylated and activated following stem cell factor binding. Mol. Cell. Biol. 14,8432-8437.

Toksoz, D., Zsebo, K.M., Smith, K.A., Hu, S., Brankow, D., Suggs, S.V., Martin, F.H. and Williams, D.A. (1992). Support of human hematopoiesis in long-term bone marrow cultures by murine stromal cells selectively expressing the membrane- bound and secreted forms of the human homolog of the Steel gene product, stem cell factor. Proc. Natl. Acud. Sci. USA 89,7350-7354.

Tono, T., Tsujimura, T., Koshimizu, U., Kasugai, T., Adachi, S., Isozaki, K., Nishikawa, S.-I., Morimoto, M., Nishimune, Y., Nomura, S. and Kitamura, Y. (1992). c-kit gene was not transcribed in cultured mast cells of mast cell-deficient ~sh/Wsh mice that have a normal number of erythrocytes and a normal c-kit coding region. Blood 78, 1448- 1453.

Torihashi, S., Ward, S.M.,Nishikawa, S.-I., Nishi, K., Kobayashi, S. and Sanders, K.M. (1995). c-kit-dependent development of interstitial cells and electrical activity in the murine gastrointestinal tract. Cell Tissue Res. 280,97- 11 1.

Tsai, M., Takeishi, T., Thompson, H., Langley, K.E., Zsebo, K.M., Metcalfe, D.D., Geissler, E.N.and Galli, S .J. ( 1991 a). Induction of mast cell proliferation, maturation, and heparin synthesis by the rat c-kit ligmd, stem cell factor. Proc. Natl. Acad. Sci. USA 88,6382.

Tsai, M., Shih, L.-S., Newlands, G.F.J., Takeishi, T., Langley, K.E., Zsebo, K.M., Miller, H.R.P.,Geissler, E.N. and Galli, S.J. (1991b). The rat c-kit ligand, stem cell factor, induces the development connective tissue-type and mucosal mast cells in vivo. Analysis by anatomical distribution, histochemistry and protease phenotype. J. Exp. Med. 174, 125.

Tsujimura, T., Furitsu, T., Morimoto, M., Isozaki, K., Nomura, S., Matsuzawa, Y., Kitamura, Y. and Kanakura, Y. (1994). Ligand-independent activation of c-kit receptor tyrosine kinase in a murine mastocytorna cell line P-815 generated by a point mutation. Blood 83,2619-2626.

Tsujimura, T., Morimoto, M., Hashimoto, K., Moriyama, Y., Kitayama, H., Matsuzawa, Y., Kitamura, Y. and Kanakura, Y. (1996a). Constitutive activation of c- kit in FMA3 murine masteocytorna cells caused by deletion of seven amino acids at the juxtarnembrane domain. Blood 87,273-283.

Tsujimura, T., Morii, E., Nozaki, M., Hashimoto, K., Moriyama, Y., Takebayashi, K., Kondo, T., Kanakura, Y. and Kitamura, Y. (1996b). Involvement of transcription factor encoded by the mi locus in the expression of the c-kit receptor tyrosine kinase in cultured mast cells of mice. Blood 88,1225- 1233.

Vandenbark, G.R., Chen, Y., Friday, E., Pavlik, K., Anthony, B., decastro, C. and Kaufman, R.E. (1996). Complex regulation of human c-kit transcription by promoter repressors, activators, and specific myb elements. CeN Growth Dilf. 7, 1383- 13%.

Wagner, E.F. and Alexander, W.S. (1991). Of Kit and mouse and man. Current Bid. 1,356-358.

Wang, C., Koistinen, P., Yang, G.S., Williams, D.E.,Lyman, S.D., Minden, M.D. and McCulloch. (1991). Mast cell growth factor, a Ligand for the receptor encoded by c- kit, affects the growth in culture of the blast cells of acute myeloblastic leukemia. Leukemia 5,493-499.

Ward, S., Burns, A*, Torihashi, S. and Sanders, K. (1994). Mutation of the proto- oncogene c-kit blocks development of interstitial cells and electrical rythmicity in murine intestine. J. Physiol. 480.1,9 1-97.

Wehrle-Haller, B. and Weston, I. A. (1995). Soluble and cell-bound forms of steel factor activity play distinct roles in melanocyte precurser dispersal and swival on the lateral neural crest migration pathway. Development 121,731-742.

Weiss, R.R., Whitakermenezes, D., Longley, J., Bender, J. and Murphy, G.F. (1995). Human dermal endotheliai cells express membrane-associated mast cell growth factor. J. Invesl. Dennatol. 104, 101-106.

Welham, M.J. and Schrader, J.W. (1992). Steel factor-induced tyrosine phosphorylation in rnurine mast cells. Common elements with IL-3-induced signal transduction pathways. J. Imrnunol. 149,2772-2783.

Wershil, B.K., Tsai, M., Geissler, E.N.,Zsebo, K.M.and Galli, S.J. (1992). The rat c- kit ligand, stem cell factor, induces c-kit receptor-dependent mouse mast cell activation in vivo. Evidence that signaling through the c-kit receptor can induce expression of cellular function. 'Exp. Med. 175,245. Yamamoto, K., Tojo, A., Aoki, N. and Shibuya, M. (1993). Characterization of the promoter region of the human c-kit proto-oncogene. Jpn. J. Cancer Res. 84, 1136- 1144.

Yarden, Y., Kuang, W., Yang-Feng, T., Coussens, L., Munemitsu, S., Dull, T.J., Chen, E., Schlessinger, J., Francke, U. and Ullrich, A. (1987). Human proto-oncogene c-kit: a new cell surface receptor tyrosine base for an unidentified ligand. EMBO J. 6,3341.

Yarden, Y. and Ullrich, A. (1988). Growth factor receptor tyrosine kinases. Ann. Rev. Biochem. 57,443.

Yasuda, H., Galli, S.J. and Geissler, E.N. (1993). Cloning and functional analysis of the mouse c-kit promoter. Bioch. Bioph. Res. Corn. 191,893-901.

Yi, T. and Me, J.N. (1993). Association of hematopoietic cell phosphatase with c-Kit after stimulation with c-Kit ligand. Mol. Cell. Biol. 13,3350-3358.

Yoshida, H., Kunisada, T., Kusakabe, M., Nishikawa, S. and Nishikawa, S.I. (1996). Distinct stages of melanocyte differentiation revealed by analysis of nonuniform pigmentation patterns. Development 122, 1207- 12 14.

Yoshinaga, K., Nishikawa, S., Ogawa, M., Hayashi, S.-I., Kunisada, T., Fujimoto, T. and Nishikawa, S.4. (1991). Role of c-kit in mouse spermatogenesis: identification of spermatogonia as a specific site of c-kit expression and function. Development 113, 689-699.

Zon, L.I. (1995). Developmental Biology of hematopoiesis. Blood 86,2876-289 1.

Zsebo, K. M., Williams, D. A., Geissler, E. N., Broudy, V. C., Martin, F. H., Atkins, H. L., Hsu, R. Y., Birkett, N. C., Okino, K. H., Murdock, D. C., Jacobsen, F. W., Takeishi, T., Cattanach, B. M., Galli, S. J. and Suggs, S. V. (1990). Stem cell factor is encoded at the S1 locus of the mouse and is the ligand for the c-kit tyrosine kinase receptor. Cell 63,2 13-224. Chapter 2

Long-range genomic rearrangements upstream of Kit dysregulate the developmental pattern of Kit expression in ~57and ~b~nded mice and interfere with distinct steps in meianocyte development

This is a revised version of a report published by: Michael Kliippel, Deborah L. Nagle, Maja Bucan and Alan Bernstein (1997). Development 124,65-77.

My contributions to the work described in this chapter were as follows: I performed all experiments, with the exception of the comparative mapping of the wbd and WS~mutations, which was done by myself in collaboration with Deborah Nagle and Maja Bucan. Mutations in the murine dominant white spotting (W) locus cause pleiotropic developmental defects that secthematopoietic cells, melanocytes, germ cells and the interstitial cells of Cajal in the gut. W mutations either alter the coding sequence of the Kit receptor tyrosine kinase, resulting in a receptor with impaired kinase activity, or affect Kit expression. Here we describe the molecular and cell type-specific developmental defects of two of the latter class of regulatory W alleles, ~57and ~bande*(bd).In both mutants, the temporal and spatial patterns of Kit expression are dysregulated during embryogenesis and in adult animals. In wbd mice, ectopic expression of Kit in the dermatome of the somites at days 10.8 and 11.8 of development seemed to interfere with melanoblast development. In contrast, the ~57 allele leads to an intrinsic pigmentation defect by downregulating developmental Kit expression in trunk melanoblasts, but not melanoblasts around the otic vesicle. Both mutations affect transcriptional initiation of the Kit gene. The ~57allele is associated with a 80 kb deletion 5' of the Kit coding region while ~bdis associated with a 2.8 Mb genomic inversion of chromosome 5 with the distal breakpoint between Kit and the platelet-derived growth factor receptor alpha (Pdgfra) gene, and the proximal breakpoint between the genes for the GABA receptor beta I (Gabrbl) and the Tec tyrosine kinase, juxtaposing the Kit and Tec tyrosine base genes. Neither ~57nor ~bdaffect genomic sequences previously suggested in in vitro experiments to control cell type-specific expression of Kit. These results link specific mechanisms of cellular and developmental defects to long-range genomic rearrangements which positively and negatively affect Kit transcription in different cell lineages as well as in different subpopulations of the same lineage. INTRODUCTION

Mutations in the dominant white spotting (W) and Steel (Sl) loci on mouse 5 and 10, respectively, result in pleiotropic developmental defects affecting primordial germ cells, hematopoietic cells, rnelanocytes and interstitial cells of Cajal in the small intestine (Reith and Bernstein, 1991; Galli et al., 1994; Huizinga et al., 1995; Ward et al., 1994). The Wlocus encodes the Kit receptor tyrosine kinase (Chabot et al., 1988; Geissler et al., 1988) whereas SZ encodes the ligand for the Kit receptor, Steel factor, also termed Kit ligand, mast cell growth factor and stem cell factor (~ndersonet al., 1990; Copeland et al., 1990; FIanagan et al., 1990; Huang et al., 1990; Williams et al., 1990; Zsebo et al. 1990). Of the numerous W alleles analyzed thus far, the majority affect the Kit coding region, resulting in a mutant receptor with diminished or undetectable kinase activity (Reith et al., 1990; Nocka et al., 1990). In contrast to these structural mutations, there are W alleles which affect the expression, rather than the structure, of the Kit receptor. For example, ~57is a spontaneous mutation that gives rise to an irregular white band in the trunk region, a white head spot and a very mild anemia in homozygotes. Heterozygous animals have a white spot in the ventral trunk region. The ~57allele does not affect fertility. The coding sequence of Kit is not altered by the w~~ mutation (A. Reith and A.B., unpublished results), but both the levels of the Kit protein and Kit kinase activity are reduced to the same extent in mast cells (Reith et al., 1990). Heterozygous ~banded(bd)mice display a white band in the trunk region, while wbd homozygotes are fertile, black-eyed and white, with occasional pigmentation of the ears and the snout. Erythropoiesis is unaffected in wbdmbd hornozygotes (Beechey et al., 1986). The wbd mutation results in a similar phenotype to that seen in ~sash(sh)animals (Lyon and Glenister, 1982), another regulatory allele of W (Duttlinger et al., 1993). Consistent with their multiple functions in many distinct cell types, both Kit and SI are expressed in diverse cell types and tissues during development and in the adult. Furthermore, the Kit and Sl genes, whose protein products interact directly to activate an intracellular signal transduction pathway, are expressed in contiguous cell layers (Motro et al., 1991 ; Keshet et al., 199 1). Biological, genetic and biochemical findings showed that the membrane-bound form of Steel factor is a more potent ligand than the processed soluble protein in activating the Kit receptor (Bernstein, 1993). Together, these observations suggest that the transcriptional mechanisms that regulate Kit and SI expression must ensure the proper spatial, developmental and temporal patterns of expression of both genes in very distinct cell lineages. To understand the molecular mechanisms that regulate Kit expression, we have analyzed the developmental defects and molecular alterations associated with the ~57 and ~bdregulatory mutatims. Here, we show that the ~57and ~bd mutations lead to distinct cellular deficits and dysregulation of Kit expression in adults and during embryogenesis. In particular, we were interested in possible mechanisms of interference with normal melanocyte development in both W mutants. Around dl 1 of embryonic development, neural crest-derived trunk melanoblasts commence their migration along a dorsolateral pathway in a rostra1 to caudal sequence from the dorsal aspect of the neural tube towards the dermatome (Rawles, 1947; Mayer, 1973; Wehrle-Haller and Weston, 1995). Shortly thereafter, melanoblasts migrate ventrally through the developing dermis and, at around 14 to 15 days of embryogenesis, from the dermis into the overlaying epidermis, where they differentiate into mature melanocytes (Mayer, 1973; Rawles, 1947; Serbedzija et al., 1990). Several alternating stages of Kit-dependency and Kit-independency have been described for melanoblasts throughout development (Nishikawa et al., 1991; Yoshida et al., 1996). Here we show that the ~bdallele leads to ectopic expression of Kit in the dermatome of the somites, the mesenchyme around the otic vesicle and the floorplate at embryonic days 10.8 and 11.8, recapitulating the spatial. but not the temporal expression of its Ligand, Steel factor. Early melanoblasts expressed approximately normal levels of Kit but disappeared shortly after emigration out of the neural tube, at the same time when Kit is ectopically expressed in the dermatome. In contrast to wb< the ~57allele does not lead to ectopic Kit expression. kt ~57animals, embryonic Kit expression is downregulated in early trunk melanoblasts, but not melanoblasts around the otic vesicle. This observation suggests that the ~57 mutation leads to an intrinsic defect in the development of a subpopulation of neural crest-derived melanoblasts. These distinct developmental defects in melanogenesis are associated with long-range genomic rearrangements upstream of the Kit coding region that affect transcriptional initiation of the Kit gene. ~57is associated with a 80 kb deletion, whereas genomic DNA from ~bdmice contains an inversion encompassing 2.8 Mb of chromosome 5. Finally, we show that genomic DNA sequences recently implicated in in vitro studies (Yasuda et al., 1993; Tsujimura et al., 1996) in the control of Kit expression in mast cells are not affected in ~57and ~bd mice; nevertheless, Kit expression is severely affected in mast cells derived from both mutants. These observations suggest that the regulation of Kit expression is complex, involving cis-acting elements immediately proximal to Kit, as well as sequences located at some distance upstream of the Kit coding region. MATEIUALS AND Ml3THODS

Mice and embryos C57BW6.J and ~5~~57mice were purchased from the Jackson Laboratory. ~bdhVbd,C3H and 101 mice were provided by Dr. C. Beechey, MRC Radiobiology Unit, Chilton, UK. wbd/'Wbd mice were also kindly provided by Dr. V. Chapman, Roswell Park, Buffalo, N.Y. Embryos were derived from appropriate matings of C57BLJ6J and C3WlOl(Fl) for wild type controls. Mutant embryos were derived from matings of homozygous ~57and wbd and from C3H/lO 1(F1 ) X wbd~bd mice. The noon after vaginal plug was considered day 0.5 p.c.

Cell culture Primary bone marrow-derived mast cell cultures were prepared as previously described (Reith et al. 1990).

Flow cytometry Bone marrow-derived mast cells (BMMC) were fust incubated with ACK2, a monoclonal antibody directed against the extracellular domain of Kit (generously provided by Dr. S. Nishikawa) for 30 min at 4 *C. The cells were then rinsed and stained with fluorescein isothiocyanate (F1TC)-conjugated rabbit anti-rat Ig serum (Cedarlane, Hornby, Ontario). Cells were rinsed and analyzed on a FACScan flow cytometer (Becton Dickinson).

RNA in situ hybridizations RNA in situ hybridizations on sections and whole mounts were performed essentially as previously described (Motro et al. 1991). The Kit, Trp-2 and Sl probes have been described previously (Motro et al. 1991; Steel and Davidson 1992). All experiments were performed on at least three embryos from at least two different litters. Double labeling RNA in situ hybridizations were performed with a digoxigenin- labelled Trp-2 probe and a 3%-labelled Kit probe. Initial experimental procedures were identical to single probe experiments, except that the hybridization solution contained a mix of both probes. After the post-hybridization washes, subsequent steps were performed as previously described (Marks et al. 1992; Miller et al. 1993).

RNA isolation and Northern blot analysis These procedures were basically performed as previously described (Maniatis, 1989).

Nuclear run ons The experiment was performed twice as previously described (Lassarn and Jay 1989) with different preparations of nuclei.

Preparation of mouse genomic and high molecular weight genomic DNA, Pulsed- field gel electrophoresis and Southern blot analysis All techniques and probes have been described previously (Brunkow et al. 1995; Nagle et al. 1994).

Analysis of Kit upstream sequences by Polymerase Chain Reaction (PCR) Genornic DNA fragments from C57BU6J, C3 W 10 1 (He)Fl , ~571~57and wbd/CYbd were amplified using the following primer pairs: Exon l/Intron 1 (+28 to +171, relative to transcriptional start site): KitM (GAGCTCAG- AGTCTAGCGCAGCC)- Kith13' (TGCGCCCACAGAGGGTGCAGT). Exon lhpstream (+23 to -347): Kitup 12 (TGCTACAGCTCTCGCCCAAGTGC)- Kitup 14 (AGGTGGCTGGCCTGTACCTA). Upstream (-278 to -624): Ki tup3 (TAGGT- ACGGGCCAGCCACCT)- Kitup 1 (AGCTGTATTCTTACAGGTTTCGCA). Upstream (-3064 to -3205): Kitup9 (GGGCTATGGGAATGAACCACGTT)-Kitup8 (CAATCACGACAGCATCTAGGGT-TAT). Upstream (-3999 to -422 1 ) : Kitup 15 (GAACCCAGCATGGTGGCACAC)- Kitup 10 (GTTCAAAGCCATCCTCAGCT- GCAT). Primer sequences were derived from published sequences (Yasuda et al., 1993; De Sepuiveda et al., 1995). Cycle parameters were 4 min at 94 OC, followed by 30 cycles of 50 sec at 94 OC, 45 sec at 56 OC-60 QC, 1 min at 72 OC, followed by 2 min at 72 OC. PCR products were analyzed on 8 96 polyacrylamide gels in 1 x TBE and stained with ethidium bromide. RESULTS

Tissue specific dysregulation of Kit expression in both ~57and Wbd adult animals

We first examined the effects of the ~57and Wbd mutations on Kit expression in adult animals by Northern blot analysis on a variety of tissues. ~57hV57mice expressed approximately normal steady state levels of Kit mRNA in brain, gut and testis (Fig. 1A); in contrast, there was a 3-4 fold reduction in the levels of Kit mRNA in ~57~57-derivedmast cells. Transcripts corresponding to the Kit gene were not detectable in lung, a tissue that also normally expresses Kit. We also analyzed the steady state levels of Kit transcripts in various tissues and mast cells from ~bd/Wbdmice. As shown in Fig. lA, Kit expression was undetectable in lung, gut and mast cells, and downregulated in testis; in contrast, the brain expressed approximately wild type Kit RNA levels (Fig. 1A). To investigate if the reductions in Kit expression in mast cells were associated with a corresponding decrease in the levels of Kit protein, we analyzed Kit protein expression in bone marrow derived mast cells (BMMC) by fluorescence activated cell sorting (FACS), using a monoclonal antibody directed against the extracellular domain of the Kit receptor (Nishikawa et al., 1991). Surface expression of the Kit receptor was not detected in wbdl~bdBMMC; in contrast, BMMC from ~57hV-57 expressed Kit at about 30% of wild type levels (Fig. 1B). We used toluidine blue staining of skin and lung sections from adult mice to investigate if this reduction in Kit mRNA and protein levels correlated with a reduction in mast cell numbers; indeed, mast cell numbers in skin and lung of adult ~57/@7 animals were reduced to about 30 %, whereas mast cells were completely absent from the skin and lung of ~bd/~bdmice (data not shown). These data demonstrate that the ~57and wbd mutations exert distinct cell type-specific effects on the levels of Kit transcripts.

Figure 1. Tissue and cell type specific perturbances of Kit expression patterns in ~5~~57 and wbd/Wbd adult mice. (A) Northern blot analysis of Kit mRNA in adult tissues and cell types from wild type, ~57.~57(57) and Wbd/wbd @d) mice. 15 pg of total EWA was loaded per lane; for gut, 3 pg of polyA RNA was loaded per lane. Hybridization of the same blot to a Tubulin cDNA probe was used as a loading control. Sizes of 28s and 18s ribosomal RNAs are indicated on the right. (B) FACS analysis of Kit receptor surface expression on +/+, ~571~57and Wbd/~bd BMMC using the AcM monoclonal antibody. Fluorescence values for levels of expression are indicated in the upper right comers. Omission of the secondary antibody AcM as a negative control resulted in fluorescence values between 0.2 and 0.4 (not shown). Figure 1

brain lung testis

Kit

.I FITC 1000

mast cells gut *L,"i, t .1 FITC 100(

0 -1 FITC lo00 Both ~57and wbd are associated with dysreplation of embryonic Kit expression during mid- and late gestation

We next analyzed Kit expression during embryonic development by RNA in situ hybridization analysis on sectioned embryos at different developmental stages. At 14.5 days (d14.5) of gestation, Kit was expressed in various areas in the brain, including the cortex, midbrain and medulla oblongata, in dorsal root gangiia (drg), olfactory epithelium, digestive tract, gonads, liver, lung, kidney and skin of wild type embryos, as described previously (Orr-Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991) (Figs. 2A and 2D). The levels and patterns of Kit expression in ~57homozygous embryos were indistinguishable from wild type embryos at the level of RNA in situ hybridization (Fig. 2B and 2E). Interestingly, Kit expression in the lung mesenchyme was normal at d14.5 (Fig. 2B), whereas expression was completely absent at dl8 (data not shown) and in adults (Fig. 1A). Similarly, Kit expression in the skin was clearly present at d14.5 (Fig. 2E), but significantly downregulated at d 18 (data not shown). In contrast to the results described above for ~571~57embryos, Kit expression was undetectable in ~bd/Wbdembryos in the lung, the developing external muscle layer of the gut and the skin at both d14.5 (Figs. 2C and 2F) and dl8 (data not shown). Expression in the olfactory epithelium was also severely reduced, whereas expression in the brain appeared only slightly reduced. Expression in drg and Liver was normal (Fig. 2C). The fetal liver is the major site of hernatopoiesis at d14.5; thus, normal Kit expression in this organ in both mutants is consistent with the absence of an obvious anemic phenotype in these mice. In addition, we observed ectopic Kit expression in the heart of ~bd/Wbdembryos at d14.5 (Fig. 2C) and at dl8 (data not shown). Figure 2. RNA in situ analysis of Kit and Steel factor expression during embryogenesis using radiolabelled probes shows a tissue-specific deregulation of Kit expression in ~'d~~bdembryos. Dark-field photomicrographs of sections from d14.5 (A) +/+, (B) ~57~57and (C) ~bd/~bdembryos. (oe) Olfactory epithelium; (co) neocortex; (mb) midbrain; (mo) medulla oblongata. Dark-field picture of Kit expression in the developing skin at d14.5 in (D) +I+, (E) ~57.'~~and (F) wbdhVbd embryos. Note the reduced expression of Kit in olfactory epithelium, lung, gut mesenchyme and skin and the extopic expression of Kit in the heart of wbd/wbd embryos. Expression of Kit in (G) wild type, (H) ~571~57and (I) wbd/Wbd embryos at d10.8. Note the lack of Kit expression in the neural tube (large arrow) and the ectopic expression of Kit in the dermatorne (small arrow) and the floorplate (arrowhead) of WbdhVbd embryos. (J) This ectopic expression recapitulates Steel factor expression in dermatome (small arrow) and floorplate (arrowhead), shown here in an adjacent section to (I), hybridized to a Steel factor probe. (K)Bright-field picture of (I). Expression of Kit in (L) wild type and (N) wbd/wbd embryos at dl 1.8 at forelimb level (main frame) and in the tail (insert). Note apparently normal expression in the neural tube (large arrow) and continued ectopic expression in dermatome (small arrow) and floorplate

(arrowhead) at both axial levels in @d&d embryos. Expression of Steel factor in (M) wild type and (0)wbd/w6d embryos at dl 1.8 in adjacent sections to the sections in (L) and (N), respectively. Note the lack of Steel factor expression in the dermatome at the forelimb level (small arrow), whereas this expression is still present in the tail (small arrow). Scale bar represents 1 mm in A-C, 100 pm in D-K, 200 pm in L-0. Figure 2 Disturbed Kit expression in Mdembryos during somite differentiation

We next compared Kit expression at embryonic days 10.8 and 11.8 in wild type and mutant embryos. At these stages, Kit is normally expressed in early postmitotic neurons in the neural mbe (Figs. 2G and 2L), in the liver as a hematopoietic organ, in the epithelial Lining of the gut, in primordial germ cells and in migrating neural crest derived melanoblasts (data not shown and discussed below) (Om-Urtreger et al., 1990; Motro et al., 1991; Keshet et al., 1991). No qualitative differences in Kit expression between wild type and ~571~57embryos were detectable (Fig. 2H and data not shown), except that the onset of Kit expression in the posterior neural tube appeared to be delayed at dl 1.8 (Fig. 4E). In contrast, at d10.8, Kit was ectopically expressed in ~bdheterozygous (data not shown) and homozygous embryos in the dermatome of the somites, the floorplate (Fig. 21), and the mesenchyme around the otic vesicle (data not shown), in a pattern indistinguishable from that previously reported for the WS~mutation at this stage (Duttlinger et al., 1993). In addition, all three tissues expressed Steel factor normally (Fig. 21 and data not shown) (Matsui et al., 1990; Duttlinger et al., 1993). However, in contrast to the transient ectopic expression of Kit in the floorplate of WS~embryos, evident at d10.5 but not at d11.5 (Duttlinger et al., 1993), the ectopic pattern of Kit expression in the floorplate, as well as in the dermatome and the mesenchyme around the otic vesicle, continued on dl 1.8 in wbd embryos (Figs. 2N and 45; Fig. 4K, expression of Steel factor around the otic vesicle). By d15.5, Kit expression was no longer observed in the floorplate of ~bdembryos (data not shown). This ectopic pattern of Kit expression in the dermatome recapitulates the spatial, but not temporal, expression of Steel factor, the Kit ligand. Steel factor expression ceases during the rostra1 to caudal epithelial-mesenchymal transition of dermatomal tissue at around dl 1 to d12, but reappears at later stages of dermis development (data not shown), (Duttlinger et al., 1993; Matsui et al., 1990; Wehrle-Haller and Weston, 1995). The ectopic expression of Kit does not cease in ~bdembryos during this transition and was observed during the expansion of the developing dermis, at times when Steel factor is not expressed. Fig. 20 shows the downregulation of Steel factor in the dermatome of a dl 1.8 wbd homozygous embryo at the level of the forelimb, while Steel factor expression at more posterior levels, e.g. in the tail, was still apparent (Fig. 20, insert) Ectopic Kit expression in the dermatome was obvious at both levels at this stage (Fig. 2N), and thus appears to be independent of the epithelial-mesenchymal transition. When Steel factor was again expressed in the developing dermis at d14.5, ectopic Kit expression in the dermis ceased (Figs. 2C and 2F). In addition to ectopic Kit expression, the wbd mutation resulted in a delay in the onset of Kit expression in the embryonic neural tube. At d10.8, Kit expression was clearly reduced in the neural tube of wbd/Wbd embryos (Fig. 21) compared to wildtype (Fig. 2G) and ~57/'W57(Fig. 2H) embryos of the same age. At dl 1.8, comparison between wild type and ~bd~~bdembryos showed no difference in Kit expression in the neural tube (Fig. 2L and 2N), suggesting a temporal effect of the ~bdmutation on Kit expression in the developing neural tube.

The effects of the ~57and ~bd mutations on morphogenetic behaviour and fate of developing melanoblasts

To investigate the developmental stages at which the pigmentation defects in ~57and ~bdare established, we analyzed the number and distribution of trunk melanoblasts during development in normal and mutant animals. To follow the development of melanoblasts, we utilized RNA probes specific for Kit and for tyrosinuse-related-protein-2 (Trp-2)as early melanoblast markers (Steel et al., 1992; Wehrle-Haller and Weston, 1995) in whole mount and section RNA in situ experiments. At d10.8, normal numbers of Tp-2-positive melanoblasts were observed in the mesenchyme around the otic vesicle in ~57/+7embryos (data not shown). At dl1.8, we observed Trp-2-positive melanoblasts in the nunk lateral to the neural tube. The temporal differences between various stages of melanoblast migration and development were most distinct at the hindlimb level: in normal and ~57 homozygous embryos, we observed melanoblasts migrating towards the MSA at the base of the tail, followed by their proliferation and migration towards the dermatome at more anterior axial levels (Figs. 3A and 3B). Melanoblast numbers were reduced at the hindlimb level in ~57hV57embryos (Fig. 3B); this reduction was more apparent at the forelimb level (Fig. 3F), compared to wild type embryos (Fig. 3E). The numbers of Trp-2-positive melanoblasts around the otic vesicle were normal (compare Figs. 3K and Fig. 35), as were numbers in other areas of the head (data not shown). In contrast to these observations on ~571~57embryos, Tp-tpositive cells were severely reduced at d10.8 around the otic vesicle (data not shown) and undetectable in the trunk (Fig. 3H), around the otic vesicle (Fig. 3M) and in other areas of the head (data not shown) of ~bd/~bdembryos at dl 1.8. At the base of the tail, a few melanoblasts were found in the MSA over a length of four to five segments (Fig. 3D and 31). Cells at the hindlimb level at dl 1.8 of development represent cells that have recently migrated out of the neural crest towards the migration staging area (MSA) between neural tube and somites (Wehrle-Haller and Weston, 1995). These cells were initially present in wb4Wbd embryos, but disappeared shortly thereafter, as indicated by the absence of melanoblasts at more anterior levels. We also analyzed melanoblast development in heterozygous ~ba+embryos by whole mount RNA in situs. At d10.8, Trp-2-positive melanoblasts were present around the otic vesicle, although the numbers were reduced compared to wild type embryos (data not shown). At dl 1.8, melanoblasts were present at the hindlimb (Fig. 3C) and the forelimb levels (Fig. 3G) and around the otic vesicle (Fig. 31), but again the numbers were reduced. Thus, the ~bdallele exerts a dominant effect early during rnelanocyte development, that also included a modification of the migration characteristics of melanoblasts in w~U+embryos. Melanoblasts normally migrate over the entire surface of the dermatome, not distinguishing between the anterior and posterior part of somites (Loring et al., 1987) (Fig. 3E). In contrast, neural crest cells that give rise to sympathetic ganglia and epinephrine-secreting cells of the adrenal medulla only migrate through the anterior, not the posterior, parts of each somite (Bromer-Fraser, 1986). Trunk melanoblasts in the MSA of the tail region of ~bdheterozygous embryos, although reduced in numbers, had a normal, random distribution (Fig. 3C), whereas melanoblasts at older, more anterior axial levels formed clusters of cells located only over one part of the somites (Fig. 3G). Although we did not determine the anterior and posterior boundaries of the somites, the observed pattern was strikingly similar to the patterns of other neural crest Lineages and different from the normal pattern of melanoblast migration. We do not know the influence this altered migration pattern might have on the development of melanoblasts in @d/+ embryos. To determine if the melanoblasts in mutant embryos express normal levels of Kit. Therefore, we hybridized adjacent sections of dl 1.8 wild-type, w57hVJ7 and wbd.bdembryos to radiolabelled Kit and Trp-2 RNA probes. Trp-Zpositive trunk melanoblasts were readily detectable lateral to the neural tube at the hind limb level of wild type embryos (Fig. 4A); these cells express high levels of Kit (Fig. 4B). Melanoblasts around the otic vesicle of the same embryo also expressed high levels of Kit (Fig. 4C). Tlp-2-positive trunk melanoblasts, albeit reduced in numbers, were also clearly present in ~57~57embryos at this stage (Fig 4D); surprisingly, we did not observe any Kit expression in these trunk melanoblasts (Fig. 4E), whereas melanoblasts around the otic vesicle of the same enbryo expressed apparently normal levels of Kit (Fig. 4F). The insert in Fig. 4E shows strong expression of Kit in the endothelium of the gut and the liver of the same embryo displayed in the main frame, showing that ~57does not lead to a general downregulation of Kit expression. The insets in Figs. 4C and 4F show individual melanoblasts from the main frames at higher magnification: the density of silver grains in these melanoblasts around the otic vesicle from wild type and ~57~57embryos was indistinguishable. Very few Trp-2- positive melanoblasts were present in the hind limb area of ~bd/~bdembryos at dl 1.8; these cells expressed normal levels of Kit (Figs. 4H arid I). Kit-positive melanoblasts around the otic vesicle could not be identified due to the strong ectopic expression of Kit in the mesenchyme (Fig. 45, compare to Steel factor expression in the same area). We showed above that melanoblasts are absent at this stage around the otic vesicle of ~bdhomozygous embryos. Similar results on the differential effects of the ~57and ~bd mutations on Kit expression in trunk melanoblasts were also obtained when using a digoxigenin- labelled Kit probe on dl 1.8 whole mount embryos (data not shown). The surprising result that ~57trunk melanoblasts seemed not to express detectable levels of Kit at dl 1.8 Led us to perform double labelling experiments in which a digoxigenin-labelled Trp-2 probe and a radiolabelled Kit probe were hybridized to the same section. This enabled us to visualize Kit expression in unequivocally identified melanoblasts. Figs. 4L and 4N show melanoblasts from the hind limb region of a wild type embryo at d11.8, labelled with Trp-2. These melanoblasts expressed high levels of Kit, as visualized by the scattered silver grains overlaying the cells (Figs. 4M and 40). In contrast, Kit expression was undetectable in Trp-2-positive melanoblasts from the hind-limb region of a d11.8 ~57/@7 embryo (Fig. 4P), with the density of silver grains barely above background (Fig. 4Q). These results confirm that the ~57mutation downregulates Kit expression in a subpopulation of melanoblasts migrating along the dorsolateral pathway in the W, without affecting Kit expression in melanoblasts around the otic vesicle.

To determine the morphogenetic behaviour of melanoblasts at later stages in development, we analyzed the fate of Trp - 2-positive melanoblasts during organogenesis, at d14.5 and dl8 of embryogenesis. At d14.5, formation of the dermis is complete and melanoblasts begin to migrate from the dennis into the developing epidermis. At this stage, ~57~57embryos had variable numbers of melanoblasts in the trunk. Some embryos had almost normal numbers of Trp-2-positive cells (compare Figs. 5A and 5B), whereas the numbers in other embryos were clearly reduced (Fig. 5C). In all ~57~57embryos analyzed, we detected one or a few melanoblasts in the mid-trunk region per section (Fig. 5F), comparable to the numbers in wild type embryos (Fig. 5E). These observations suggest that the pigmentation defect, i.e. the absence of melanocytes in the mid-trunk region of ~571~57adults, has not yet been established at d14.5, although melanoblast numbers were already reduced from dl 1.8 on. These observations on Tp-2-positive melanoblasts confirm and extend previous findings (Duttlinger et al., 1993) that Kit-positive cells in the epidermis of wild type embryos at midgestation show a graded distribution, with high density in the tail and face, and low density in the trunk region. The number of Tp-2- positive melanoblasts in the developing inner ear as well as other areas of the head were unaffected in ~5~~57embryos at this stage (data not shown). At dl8 of development, large areas in the iower trunk and head were completely devoid of Tp-2-positive melanoblasts in ~5~~57embryos (Fig. 5H and data not shown), corresponding to the location of future depigmentation in the adult. In contrast, wild type embryos at dl8 displayed a continuous layer of Trp-2-positive melanoblasts in the skin of the lower trunk (Fig. 5G) and head (data not shown). Again, the number of Trp-2 positive melanoblasts in the developing inner ear was not affected in W57hV57 embryos (data not shown). We did not observe any Trp-2-positive melanoblasts in wbd~vbdembryos at d14.5 or dl8 (Fig. 5D and data not shown).

To investigate if Kit expression is still downregulated in trunk melanoblasts of ~571~57embryos at d14.5 or d18, we performed double labeling RNA in situ experiments utilizing a digoxigenin-labelled Tp-2 probe and a 3%-radiolabelled Kit probe. This double-labelling technique was utilized to distinguish between two Kit- expressing cell types, melanoblasts and mast cells, present in the developing skin from midgestation onwards. Trp-2-positive melanoblasts in wild type (d14.5, Fig. 6A; d18, Fig. 6E) and ~57~57embryos (d14.5, Fig. 6C; d18, Fig. 6G) expressed approximately equal levels of Kit (Figs. 6B, D, F, H). Figure 3. Expression analysis of Trp-2 at dl 1.8 reveals a reduction in trunk melanoblast numbers in both w~~/w~~and @d mutant embryos. (A) RNA whole mount in situ analysis using a digoxigenin-labelled Trp-2 probe identifies Trp-2-positive melanoblasts in the hindlimb area: single melanoblasts appear first at the base of the tail (right side of picture) in the MSA, followed by proliferation and migration towards the dermatome at more anterior axial levels (left side of picture). (hl) Base of hindlimb. (B) ~5~@7ernbryoshave nod numbers of Trp-2-positive melanoblasts in the MSA, whereas melanoblast numben do not expand to the same degree as in wild type embryos during the proliiferation and migration toward the dermatome. (C) @d/+ embryos display a marked reduction of melanoblasts in the hindlimb area. (D) @d/~bdembryos have a few Trp-2-positive melanoblasts in the MSA (arrows). but the cells disappear shortly after; no celis migrating towards the dermatome can be observed. The asterisk marks the group of melanoblasts magnified in (I). (E) Trp-2-positive melanoblasts at the forelimb level of a wild type embryo. (F) Numbers of Trp-2-positive melanoblasts are severely reduced in w57/WS7embryos. (G) Trp-2-positive melanoblasts are reduced in numbers and are not randomly distributed anymore in the forelimb region of Wbdl+ embryos. The arrow marks an area devoid of melanoblasts. (H) No

Trp-2-positive melanoblas ts are found in the forelimb region of ~bd/Wbdembryos. Normal numbers of Trp-2-positive melanoblasts in the mesenchyme around the otic vesicle are seen in (J) wild type and (K) ~57/CY57embryos. whereas in (L) ~bd/+embryos melanoblast numbers were severely reduced and in (M) W@/w6d embryos melanoblasts are absent. Scale bar represents 150 pm in A-D. 100 pnin E-H and J-M and 40 pm in I. Figure 3 Figure 4. RNA in situ analysis of Trp-2 and Kit expression in melanoblasts (arrows) of wild type, wS7W57 and i@d/wbd embryos at dl 1.8. (A) Trp-2 expression md (B) Kit expression (adjacent sections) in melanoblasts migrating on a dorsolateral pathway at the hindlimb level of a wild type embryo. (C) Kit expression in melanoblasts around the otic vesicle of a wild type embryo. The three melanoblasts marked with the arrow are shown at higher magnification and in bright field in the inlet. (D) Reduced numbers of Trp-2-positive melanoblasts at the hindlimb level of ~5~~5~embryos. (E) In the adjacent section, no Kit expression could be observed in melanoblasts, whereas Kit expression in liver (li) and gut endothelium (arrowhead) appeared normal (inlet). (F) Kit expression in melanoblasts around the oric vesicle appeared to be at wild type levels. The two melanoblasts marked with the arrow are shown at higher mawcation and in bright field in the inlet. We observed only very few Trp-2-positive melanoblasts at the hindlimb level of WbdhVbd embryos (G), which expressed normal levels of Kit (H). (I) shows the melanoblast from (H) at higher mamication and in bright field. (J) wbdNV6d embryos display ectopic Kit expression in the floorplate (filled arrowhead) and in the mesenchyme around the otic vesicle (open arrowhead), which recapitulates (K) Steel factor expression in the same structures. (L-Q) Double labelling RNA in situ analysis of wild type and ~57~~~melanoblasts at the hind limb level. The same section was hybridized to a digoxigenin-labelled Trp-2 probe and a radiolabelled Kit probe. Cells and silver grains are in different focal plains due to a Parlodion coating of the slides before exposure to prevent a color reaction between color product and emulsion. (L,N) Trp-2 positive melanoblasts in wild type embryos express high levels of Kit (M,O), as judged by the density of silver grains overlaying the melanoblasts, whereas Trp-2- positive melanoblasts in ~57/@7embryos show a severe reduction in Kit expression: silver grain density overlaying the melanoblasts is barely above background (Q). (nt, neural tube; ot: otic vesicle; cv: anterior cardinal vein; drg: dorsal root ganglion; dm: dermatome). Scale bar represents 200 pm in A-G, 30 pm in the inlets of Cand F and I, 100 pm in H, J, K, and 15 in L-Q.

Figure 5 Figure 5 (previous page). Dark-field photographs of RNA in situ experiments using a radiolabelled Trp-2 probe on sections of d 14.5 and dl8 embryos. Trp-2 positive melanoblasts at the base of the tail of d14.5 (A) wild type, (B), (C) wJ7m'7and (D) @d/w6dernbryos. Note the moderate (B) and severe (C) reduction in melanoblast numbers in two different ~57.57embryos, whereas no Trp-2-positive melanoblasts are present in wbd&d embryos (D). Very few Trp-2-positive melanoblasts are present at the mid-trunk level at d14.5 in both (E) wild type and (F) ~571~57embryos. Melanoblast numbers at the mid- trunk level at d 18 are markedly reduced in (H) ~571~57embryos, compared to (G) wild type embryos, leaving large areas devoid of Trp-2-positive melanoblasts. Arrows mark single melanoblasts. Scale bar represents 150 pm.

Figure 6 (next page). Double labeling RNA in situ analysis on sections of d14.5 and d 18 embryos using a digoxigenin-labelled Trp-2 probe and a radiolabelled Kit probe. Trp-2- positive trunk melanoblasts of dl4.S (A) and dl8 (E) wild type embryos are stained blue and express Kit (B,F). Trp-2-positive melanoblasts at the base of the tail of d 14.5 (C) and dl8 (G)

~57,~57embryos also express Kit at approximately wild type levels (D,H). Scale bar represents 10 pm. Figure 6 Both ~57and ~bd affect transcriptional initiation of Kit

The results presented above demonstrate that the ~57and wbd mutations differentially affect the temporal and spatial expression patterns of Kit, leading to distinct cellular and developmental defects. To understand the molecular basis for these effects on Kit expression, we fmt performed nuclear run-on assays to determine whether these mutations act at the level of transcriptional initiation or mRNA stability. We measured the rates of transcriptional initiation of Kit in homogeneous populations of BMMC from +/+, ~57~57and ~bd/Wbd mice. The results were quantitated by comparing the intensity of the Kit hybridization signal to the corresponding signals for Tubulin and GAPDH, two housekeeping genes whose expression is not expected to be affected by W mutations. While BMMC from wild type mice expressed high levels of Kit rnRNA, transcriptional initiation of Kit was essentially undetectable in ~bd/'Wbdmast cells (6.3 % of wild type levels) and reduced to 29 % of wild type levels in mast cells derived from ~571~57mice (Figs. 7A and 7B). These results demonstrate that both ~57and ~bd affect transcriptional initiation of Kit and that the decrease in the steady state levels of Kit mRNA and protein in mast cells from these mutant mice (see Fig. 1) can be entirely accounted for by effects on transcriptional initiation. Figure 7. (A) Nuclear run on analysis comparing transcriptional initiation of Kit in nuclei from BMMC derived from adult wild type, ~571~57and w6dh@ mice. Controls include the transcriptional initiation of Tubulin and GAPDH, two housekeeping genes, PDGFRa, a gene not expressed in mast cells and the pKS vector, which does not have any transcriptional initiation sites. (B) Quantitation of the results shown in (A) by comparing transcriptional initiation of Kit to that of the two housekeeping genes Tubulin and GAPDH shows that in

~5~~57BMMC, the rate of transcriptional initiation of Kit is only 29 % of that of wild type BMMC, whereas in wbd/wbd~M..~,the rate of transcriptional initiation is only 6.3 % of wild type levels. A second, independent nuclear run on experiment gave similar results. Figure 7

Tubulin The ~57mutation is associated with a 80 kb deletion 5' of Kit

The Kit nucleotide sequence in ~57DNA is identical to the Kit cDNA sequence in C57BU6J DNA (A.D. Reith and A.B., unpublished results) and Southern blot analysis with 6 different restriction failed to reveal any gross structural alteration in ~57and wbd genomic DNA. In addition, we determined by PCR that regulatory sequences located in the first 400 bp upstream of the first coding exon of Kit which have recently been implied in in vitro experiments to play an important role in transcriptional regulation of Kit (Yasuda et al., 1993; Tsujimura et al., 1996) are not affected by either mutation. In wbd,sequences up to -624 bp upstream of Kit are present and linked to the fust coding exon. In ~57,sequences up to -4221 bp are intact (Fig. 8). To determine whether the cellular and regulatory defects in ~57mice might result from long-range genornic rearrangements, we analyzed high molecular weight spleen DNA from wild type, heterozygous w?/+ and homozygous ~571~57mice by pulsed-field gel electrophoresis (PFGE). We have previously established that the genes encoding two other RTKs, Pdgfra and Flkl are closely linked to Kit in the order Pdgfra-Kit-Flk-l with Pdgfra and Kit in the same transcriptional orientation (Brunkow et al., 1995). Therefore, genornic DNAs were analyzed with probes for all three genes. Using four rare cutting restriction enzymes (NotI, Mlul, Nurl. Pmel), we identified novel DNA fragments that were reduced in size by approximately 80 kb in all digests of ~57DNA (Fig.9A). For example, while the Kit cDNA probe hybridized to a 290kb Pmel fragment in wild type DNA, this probe detected both 290kb and 210kb DNA fragments in ~57/+DNA and a 210kb DNA fragment in ~5~~57 DNA. Similarily, the Pdgfra cDNA probe detected a 580kb Not1 fragment in wild type DNA and a novel 500kb fragment in w57~57DNA;NotI fragments of both sizes were present in ~57/+DNA. Only DNA fragments of normal sizes were observed when ~57DNA was analyzed with the Flk-I cDNA probe (data not shown). Using single and double digests (data not shown), we established a long range restriction map of the rearranged ~57chromosome. By comparing this restriction map to the wild type long-range restriction map previously described (Nagle et al., 1994; Brunkow et al., 1995), we conclude that the ~57mutation is associated with a 80 kb deletion, located in a 160 kb PrnellNotl genomic fragment 5' to Kir, between the Kit and Pdgfra genes (Fig. 9B). The Buc3 probe (see Fig. lOC) is not affected by the ~57mutation. The addition of the Pmel restriction sites to the existing genomic map significantly narrows down the genomic region that contains the ~57deletion as well as the genomic sequences that contain the distal breakpoint in the @d and WS~mutations (see below).

Figure 8 (previous page). PCR analysis showing that neither ~57nor Wbd affect previously described regulatory sequences 5' of Kit. (A) Schematic map (not to scale) of the genomic region 5' to Kit and the approximate location of the PCR primers used in (B) (for exact location and sequence of primers, see material and methods). Approximate distances from the transcription initiation site are indicated above the schematic map. Location of the upstream Mi-site and the region immediately upstream of the fvst coding exon (hatched box) are indicated.. (B)PCR amplification of genomic DNA from +/+, ~~71~57and wbd/wbd mice using the primer pairs shown in (A). Control reactions (C) contained no genomic DNA; M indicates the lane of the size marker. The size of the PCR products are indicated on the left of each gel.

Figure 9 (next page). (A) PFGE analysis of high molecular weight DNA from wild type (C57BL/6J), wj7/+ and ~57/~57mice. Restriction digests with NotI, NarI and Mld resulted in novel DNA fragments in ~57DNA when hybridized to a PDGFRa cDNA probe, whereas digests with PmeI and Mlul resulted in novel DNA fragments in ~57when hybridized to a Kit cDNA probe. All novel fragments detected were reduced in size by approximately 80 kb. (B) Long-range restriction map of wild type and rearranged w~~ chromosome. The arrow indicates the position of the 80 kb deletion 5' to Kit in wJ7DNA and the approximate location of the proximal breakpoint of the wbd and WS~inversions. Molecular weights are indicated. Figure 9

I 530 kb, 660kb- ,I St0 kb- .4 0 450 kb'

m Kit POGFFls Kit POGFFla Kit PDGFRe Kit PDGFFle The Wbd mutation is associated with a 2.8 Mb inversion upstream of Kit

The phenotypic similarity of wbd and wsh mice prompted us to compare the Kit-Pdgfra intergenic region in both ~bdand WS~DNA by PFGE. We observed novel DNA fragments of the same size in wbd/~bdand WS~/WS~DNA digested with Mlul or Pmel after hybridization to the Kit probe (Fig.lOA and data not shown), whereas the Pdgfra probe detected novel DNA fragments following digestion with BssHII, Notl, Mlul or Nrul (Fig. LOB and data not shown). To compare further the genornic rearrangements in the ~bdand WS~ alleles, the same PFGE blots were hybridized with cDNA probes for Kit and Tec as well as Pdgfra and Gabrbl. Tec and Gabrbl are located 3 Mb proximal to Kit and known to flank the proximal breakpoint of the WS~inversion (Nagle et al., 1995). The Kit and Tee probes detected a novel DNA fragment of identical size in ~bd/~bdand ~sh4Wsh DNA digested with Pmel (Fig. lOA), while the Pdgfra and Gabrbl probe detected the same altered DNA fragments in wbdhVbd and W~/WS~DNA digested with the enzymes BssHII, NotI, Nncl and NotIZVrul (Fig. 10B). These results establish the structural similarities between the ~bdand WS~alleles and demonstrate that ~bd,an independently isolated allele, is associated with an inversion in the same chromosomal region as WS~.Figure 10C summarizes the physical map of the rearranged ~bdand wsh chromosomes. Our results demonstrate the approximate location of the distal breakpoint of both the wbd and WS~inversions in the same 160 kb Pmel/Notl genomic fragment that also harbours the ~57deletion (Fig. 9B). In both wbd and ~sh,the proximal breakpoint is located in the intergenic region between Tee and Gabrbl . Both inversions juxtapose Kit and Tec as well as Pdgfra and Gabrbl . Figure 10. PFGE analysis showing that the Wbd mutation is associated with a chromosomal inversion. (A) PFGE analysis of the distal breakpoint of the Wbd inversion. Wild type (C3WHe), wshAVsh and w6d/wbd DNA were digested with Pmel. The Kit probe detected novel DNA fragments of identical size in Wbd and WS~DNA in addition, the Kit and Tec probes hybridized to novel DNA fkagments of identical size. (B) PFGE analysis of the proximal breakpoint. With all enzymes tested (BssHIZ,Nrul and NrulflVotl), both the Pdgfra and Gabrbl probes detected novel DNA fragments of identical size in both @d/Wbd and

WSh/wshDNA. Molecular weights (in kb) and a zone of limited mobility (LM) are indicated. (C) Long-range restriction map of the central portion of mouse chromosome 5 of wild type

(C57BU6J and C3H/He), ~bdand WS~chromosomes. The map of the wild type chromosome and ~shchromosome has previously been reported (Nagle et al., 1995). The map indicates sites for the restriction BssHII (B). Partial cutting sites are shown in parentheses. Positions of loci are indicated by solid rectangles above the map. Scale in kb is shown at top of map. The approximate location of the inversion breakpoints are indicated by the zig-zag lines. Figure 10

A pme I Pme I

Kit Pdgfra Kit T"

Pdgfra Gabrbl Pdgfra Gabrbl DISCUSSION

In this paper, we have described the cellular, developmental, and molecular defects associated with two regulatory alleles of the W locus, ~57and @d. ~57and ~bddifferentially affect the temporal and spatial expression of Kit in both the adult and during embryogenesis. Reduction of Kit mRNA expression in BMMC correlated directly with the reduction of Kit surface receptor expression in BMMC and reduction in mast cell numbers in the skin and lung of ~57~57and @dmbd mice.

Embryos homozygous for the ~57mutation displayed an almost wild type Kit expression pattern. Between d14.5 and dl8 of embryogenesis, Kit expression in both the lung and skin was greatly diminished, suggesting that the ~57mutation leads to the dysregulation of proper temporal expression of Kit in the two Kitexpressing cell types normally present in the skin, melanoblasts and mast cells. Since the numbers of mast cells in adult lung and skin are reduced, the loss of Kit expression in embryonic lung and skin of ~bd.+bdembryos between d14.5 and dl 8 most likely reflects a reduction in mast cell numbers; however, it is possible that other cell types in the lung that also express Kit are also affected, as we have shown for melanoblasts in the skin. In contrast to these marked effects on Kit expression in lung and skin, the levels of Kit mRNA in brain, olfactory epithelium, drg, kidney and liver were normal at these stages.

At midgestation, Kit was not expressed in the lung, gut and skin of ~bd~bd embryos. In addition, expression in the olfactory epithelium was greatly reduced, whereas Kit was ectopically expressed in the heart. The effects on Kit expression in skin and lung most likely reflect the absence of melanoblasts and tissue mast cells due to the WM mutation. During somite formation and differentiation at d10.8 and dl 1.8, Kit was ectopically expressed in ~bdembryos in a variety of structures that also express Steel factor, including the dermatome of the sornites, the floorplate and the mesenchyme around the otic vesicle.

To gain insights into the cellular and molecular mechanisms of action of the ~57and W bd alleles, we followed melanoblast development throughout embryogenesis, using Trp-2 as a melanoblast marker. This analysis demonstrated that the pigmentation defect in the mid-trunk of ~57~57animals is a late defect that is not maoifested in all embryos before late gestation, although melanoblast numbers are reduced in the trunk from d1 1.8 onward. At d18, melanoblast numbers are reduced everywhere and large areas in the trunk region, corresponding to the regions of depigmentation in adult ~57i~57animals, are devoid of melanoblasts. This defect in melanoblast development was associated with a marked reduction in Kit expression in trunk melanoblasts, but not melanoblasts around the otic vesicle, at dl 1.8 of embryogenesis, whereas at later stages ~57trunk melanoblasts expressed approximately normal levels of Kit. These data suggest that ~57causes an intrinsic, cell-autonomous defect in a subset of neural crest-derived melanoblasts through the temporal downregulation of Kit expression during early development of trunk melanoblasts. Because melanoblasts display a graded distribution along the embryonic body axis, with highest numbers in the tail and face and lowest numbers in the trunk region (Duttlinger et al., 1993), an overall reduction, but not complete absence, of melanoblasts would lead to selective depigmentation in the trunk region, while tail and face would still have sufficient melanocyte numbers to ensure pigmentation. It has recently been shown that melamblasts undergo several alternating stages of Kit-dependency and Kit-independency between d9.5 and birth (Yoshida et al., 1996). The temporal downregulation of Kit expression in trunk melanoblasts of ~5'@7 embryos might at least partidy overlap with a phase of Kit-independency, therefore causing only a partial depigmentation. Alternatively, some residual expression of Kit, not neccessarily detectable in our assays, might enable some trunk melanoblasts to sunrive through the period of Kit downregulation. ~57~57animals display a white head spot; we showed that this phenotype is established between d14.5 and d18. Although we observed approximately normal expression of Kit in cranio-facial melanoblasts around the otic vesicle at dl 1.8 as well as in trunk and head melanoblasts at d14.5 and d18, we cannot exclude that the ~57 mutation affects Kit expression and therefore development of a subset of cranio-facial rnelanoblasts, leading to depigmentation of parts of the head. Alternatively, cranio- facial melanoblasts might be in general more sensitive to small changes in Kit expression, which we might have not detected. We would argue against this latter point, since Trp-2-positive melanoblast numbers in the developing inner ear appear normal throughout embryogenesis in ~571~57embryos.

Melanoblasts in ~bd/+bdembryos never migrate to the dermatome, although they express apparently normal levels of Kit. During the rostra1 to caudal sequence of dorsolateral melanoblast migration from the neural crest towards the somites, melanoblasts are stalled at the MSA, lateral to the neural tube, and disappear shortly thereafter. Thus, the pigmentation defect is established early, coincidentally at the same time when Kit is ectopically expressed in the dermatome of the somites. Interactions between Kit-expressing rnelanoblasts on the dorsolateral pathway and Sl- expressing dermatome have been suggested to play an important role in early melanoblast migration from the MSA towards the dermatome (Wehrle-Haller and Weston, 1995). Duttlinger et al. (1993, 1995) have previously suggested that the ectopic expression of Kit in the dennatome of WS~and Patch (Ph) embryos might sequester functional soluble Steel factor, thereby reducing the amount of ligand available to the migrating melanoblasts and affecting their survival by a non-cell autonomous mechanism. Interestingly, stimulation of human fetal melanoblasts in vitro with Steel factor affects the expression patterns of several subunits and changes the attachment of those cells to extracellular matrix (ECM) molecules (Scott et al., 1994). The failure of melanoblasts in ~bdand wsh embryos to be exposed to Steel factor from the dermatorne could lead to a change in integrin expression and subsequently to an inability of these cells to recognize and bind to the ECM. In this scenario, melanoblast migration would be stalled at the MSA, as we observed in ~bd/~bdmutant embryos.

[n contradiction to this non-cell autonomous model, Huszar et al. (1991) have shown that transphntation of normal neural crest cells into the amniotic cavity of wsh~~shembryos results in the formation of pigmented areas in the skin, suggesting a cell autonomous defect in WS~-mutantanimals. Thus, it is not clear if the ectopic expression of Kit in the dennatome is causally related to the pigmentation defects in wbd and WS~mice.

The ectopic expression of Kit in the dermatome of ~bdembryos recapitulates the spatial, but not temporal, expression pattern of Steel factor. Steel factor expression is turned off during the epithelial-mesenchymal transition of the dermatome (Wehrle- Haller and Weston, 1995), but is re-expressed at d14.5 in the skin. We observed continuous ectopic expression of Kit in the dermatome of wbd embryos at dl 1.8, even at axial levels where Steel factor expression was turned off during the epithelial- mesenchymal transition. While Steel factor is re-expressed in the skin at d14.5, ectopic Kit expression ceases between dl 1.8 and d14.5. Other structures that express Kit ectopically in WA mutant embryos (Duttlinger et ai., 1993) were also affected in wbd embryos, including the mesenchyme around the otic vesicle and in the floorplate, both structures that normally express Steel factor. However, unlike wsh embryos which cease to express Kit in in the floorplate after d10.5, ~bdembryos sustained ectopic Kit expression in the floorplate on dl 1.8, but not d15.5. In addition, wbd embryos expressed Kit ectopically in the heart at mid- and late gestation, which has not been described for wsk

Both ~57and wbd involve long-range genomic rearrangements only detectable by PFGE. The ~57mutation is associated with a 80kb deletion, located in a 160 kb genomic region between the upstream Pmel site and the S'end of the Kit coding sequence. ~bdis a 2.8 Mb inversion, with one breakpoint 5' to the Kit coding sequence in the same 160 kb fragment that harbours the ~57mutation and one breakpoint between the Tec and the Gabrbl genes. This inversion juxtaposes the Pdgfra and Gabrbl genes on the proximal side as well as Kit and Tec on the distal side of the inversion. Surprisingly, the ~bdinversion appeared indistinguishable from the WS~inversion (this study; Duttlinger et al., 1993; Nagle et al., 1995; Duttlinger et al., 1995), although both are spontaneous and independent mutations (Beechey et al., 1986; Lyon and Glenister, 1982). The Buc3 probe was affected by ~bdand by ~sh, but not by ~57,indicating genornic sequences that are uniquely affected in these mutations. The breakpoints of both the ~57deletion and the ~bdinversion have not yet been cloned; therefore, only an approximate location of their breakpoints can be given. Interestingly, two genomic regions upstream of Kit, a binding site for the Mi transcription factor, encoded by the microphthalmia (mi) locus (Tsujimura et al., personal communication), and the region immediately upstream of the Kit transcriptional initiation site (Yasuda et al., 1993), have recently been shown to be necessary for expression of Kit reporter plasmids in vitro. Neither of these regions are affected by the ~57and wbd mutations. Together, these results demonstrate that important regulatory elements, distinct from the two cis-acting regions described previously and located further upstream, contribute significantly to the complex temporal and spatial pattern of Kit expression in vivo. In both mutations, these large genomic rearrangements differentially affect Kit transcriptional initiation in different cell Lineages during development. The ~57and wbd mutations might disrupt a transcriptional regulatory region which normally drives Kit expression in a cell type-specific manner. Interestingly, Kit expression is only perturbed in trunk melanoblasts in ~571~57embryos, but not in melanoblasts around the otic vesicle, suggesting that distinct cis-acting elements regulate Kit expression in different subpopulations of neural crest-derived rnelanoblasts. The Kit and Sl genes are expressed in cell layers that are immediately contiguous (Motro et al., 1991; Keshet et al., 1991). Thus, the mechanisms that regulate expression of this gene pair must ensure their coordinate expression in space and time during development and in the adult. It is striking, therefore, that the ectopic Kit expression in ~bdanimals recapitulates in part the expression pattern of its ligand, Steel factor. If the mechanisms that regulate Kit and Steel factor expression are related, it is conceivable that negative cis-elements ensure that Kit is not expressed in cells normally expressing its ligand. Disruption of these silencer elements might then result in the ectopic expression of Kit in tissues normally only expressing Sl. Only the spatial, but not the temporal, pattern of Sl expression in the embryonic dermatome was recapitulated by Kit in wbd embryos; thus, the mechanisms involved in ensuring proper Kit expression must have both spatial and temporal aspects, of which only the former is perturbed by the ~bdinversion. The genomic rearrangements in wbd and ~57,respectively, might also bring regulatory sequences not normally located at some distance from Kit into close proximity of this gene, positively or negatively affecting its expression. According to this model, Kit might acquire aspects of the expression pattern of another gene, whose regulatory elements have no influence on Kit transcription in wild type mice.

The Kit RTK is expressed in the early stem cells and lineage-committed progenitor cells of very different cell lineages (Reith and Bernstein, 199 1; Galli et al., 1994). Thus, insights into the mechanisms that govern the regulation of Kit expression should provide insight into the early events involved in lineage specification. The analysis of W mutants which disrupt normal Kit transcription would be an important fust step towards the characterization of factors controlling lineage determination and elaboration.

ACKNOWLEDGEMENTS This paper is dedicated to the memory of Verne Chapman. We thank S. Vesely for help with mast cell preparations, Y. Kitamura for generously providing us with the Ack2 antibody and V. Chapman for generously providing us with ~bd/~bdmice, Diane Poslinski and Debbie Swiatek for animal care, and William Brown and Mary Lyon for providing PFGE plugs with wbd DNA from the Hanvell colony. These studies were supported by grants from the Medical Research Council of Canada (to

A.B .) and the National Cancer Institute of Canada (grant W033 13) (to A.B .), as well as NM grant HD 28410 (to M.B.). M.K. was supported by a long-term Government of Canada Award fellowship and D.L.N. by a fellowship of the IV Drug Abuse Research Center, T32-Da07241. A.B. is an International Research Scholar of the Howard Hughes Medical Institute. REFERENCES

Anderson, D.M., Lyman, S. D., Baird, A., Wignall, J. M., Eisenman, J., Rauch, C., March, C. J., Boswell, H. S., Gimpel, S. D., Cosman, D. and Williams, D. E. (1990). Molecular cloning of mast cell growth factor, a hematopoietin that is active in both membrane bound and soluble forms. Cell 63,235-243.

Beechey, C.V., Loutit, J.F. and Searle, A. G. (1986). Banded-a new W-allele. Mouse News L 74,92.

Bernstein, A. (1993). Receptor tyrosine kinases and the control of hematopoiesis. Sem. dev. Biol. 4,351-358.

Bronner-Fraser, M. (1986). Analysis of the early stages of trunk neural crest migration in avian embryos using monoclonal antibody HNK- I. Dev. Bid. 115,528- 536.

Brunkow, M., Nagle, D., Bernstein, A. and Bucan, M. (1995). A 1.8 Mb YAC contig spanning three members of the receptor tyrosine kinase gene family (Pdgfra, Kit , and Flk-I) on mouse chromosome 5. Genomics 25,421432.

Chabot, B., Stephenson, D. A., Chapman, V. M., Besmer, P. and Bernstein, A. (1988). The proto-oncogene c-kit encoding a transmembrane tyrosine kinase receptor maps to the mouse W locus. Nature 335,88-89.

Copeland, N.G., Gilbert, D. J., Cho, B. C., Donovan, P. J., Jenkins, N. A., Cosman, D, Anderson, D., Lyman, S. D. and Williams, D. E. (1990). Mast cell growth factor maps near the Steel locus on mouse chromosome 10 and is deleted in a number of Steel alleles. Cell 63, 175-183.

De Sepulveda, P., Salaiin, P., Maas, N., Andre, C. and Panthier, 1.-J. ( 1995). SARs do not impair position-dependent expression of a Kit/LacZ transgene. Bioch. Bioph. Res. Comm 211 (3), 735-741. Duttlinger, R., Manova, IS., Chu, T. Y., Gyssler, C. Zelentz, A. D., Bachvarova R. F. and Besmer, P. (1993). W-sash affects positive and negative elements controlling c-kit expression: ectopic c-kit expression at sites of kit-ligand expression affects melanogenesis. Development 118,705-717.

Duttlinger, R., Manova, K., Berrozpe, G., Chu, T. Y., DeLeon, V., Timokhina, I., Chaganti, R. S. K., Zelenetz, A. D., Bachvarova, R. F. and Besmer, P. (1995). The WSh and Ph mutations affect the c-kit ex ression profile: c-kit misexpression in embryogenesis impairs melanogenesis in K and Ph mice. Proc. Natl. Acad. Sci. 92, 3754-3758.

Hanagan, J. G. and Leder, P. (1990). The kit Iigand: A cell surface molecule altered in Steel mutant fibroblasts. Cell 63, 185-194.

Galli, S. J., Zsebo, K. M. and Geissler, E. N. (1994).The kit ligand, Stem Cell Factor. Advances in Immunology 55, 1-96

Geisler, E.N., Ryan, M.A. and Houseman, D.E. (1988). The dominant white spotting (W) locus of the mouse encodes the c-kit proto-oncogene. Cell 55, 185-194.

Huang, E., Nocka, K., Beier, D. R., Chu, T. Y., Buck, J., Lahm, H.-W., Wellner, D., Leder, P. and Besmer, P. (1990).The hematopoietic growth factor KL is encoded by the SI locus and is the ligand of the c-kit receptor, the gene product of the W locus. Cell 63,225-233.

Huizinga, J. D., Thuneberg, L., Kliippel, M., Malysz, J., Mikkelsen, H. B. and Bernstein, A. (1995). W/kit gene required for interstitial cells of Cajal and for intestinal pacemaker activity. Nature 373,347-349.

Huszar, D., Sharpe, A. and Jaenisch, R. (1991). Migration and proliferation of cultured neural crest cells in W mutant neural crest chimeras. Development 112, 13 1- 141.

Keshet, E., Lyman, S. D., Williams, D. E., Anderson, D. M., Jenkins, N. A., Copeland, N. G. and Parada, L. F. (1991). Embryonic RNA expression patterns of the c-kit receptor and its cognate ligand suggest multiple functional roles in mouse development. EMBO J. 10,2425-2435.

Lassam, N., and Jay, G. (1989). Suppression of MHC class I RNA in highly oncogenic cells occurs at the level of transcription initiation. J. of Imrnunol. 143, 3792-3797.

Loring, J. F. and Erickson, C. A. (1987). Neural crest cell migratory pathways in the trunk of the chick embryo. Dev. Biol. 121,220-236.

h Lyon, M. F., and Glenister, P. H. (1982). A new allele sash (I@) at the W-locus and a spontaneous recessive lethal in mice. Genet. Res. 39,3 15-322.

Marks, D. L., Wiema~,J. N., Burton, K. A*, Lent, K. L., Clifton, D. K. and Steiner, R. A. (1992). Simultaneous visualization of two cellular mRNA species in individual neurons by use of a new double in situ hybridization method. Mol. Cel. Neurosci. 3, 395-405.

Matsui, Y., Zsebo, K. M. and Hogan, B. L. M. (1990). Embryonic expression of a haernatopoietic growth factor encoded by the Sl locus and the ligand for c-kit. Nature 347,667-669.

Mayer, T. C. (1973). The migratory pathways of neural crest cells into the skin of mouse embryos. Dev. Biol. 34, 39-46.

Miller, M. A*, Kolb, P. E. and Raskind, M. A. (1993). A method for simultaneous detection of multiple mRNAs using digoxigenin and radioisotopic cRNA probes. J. Histochem. Cytochem 41,174 1- 1750.

Motro, B., Van Der Kooy, D., Rossant, J., Reith, A. and Bernstein, A. (1991). Contiguous patterns of c-kit and steel expression: analysis of mutations at the W and Sl loci. Development 113, 1207- 1221.

Nagle, D. L., Martin-DeLeon, P., Hough, R. B. and Bucan, M. (1994). Structural analysis of chromosomal rearrangements associated with the developmental mutations Ph, W19H and R w on mouse chromosome 5. Proc. Natl. Acad. Sci. 91, 7237-7241.

Nagle, D. L., Kozak, C. A., Mano, H., Chapman, V. M. and Bucan, M. (1995). Physical mapping of the Tee and Gabrbl loci reveals that the WS~mutation on mouse chromosome 5 is associated with an inversion. Hum. Mol. Gen. 4 (ll), 2073-2079.

Nishikawa, S., Kusakabe, M., Yoshinaga, K., Ogawa, M., Hayashi, S., Kunisada, T., Era, T., Sakakura, T. and Nishikawa, S. (199 1). In utero manipulation of coat color formation by a monoclonal anti-c-kit antibody: two distinct waves of c-kit dependency during melanocyte development. EMBO J. 10,21 1 1-2 118.

Nocka, K., Tan, J. C., Chiu, E., Chu, T. Y., Ray, P., Traktman, P. and Besmer, P. (1990). Molecular bases of dominant negative and loss of function mutations at the rnurine c-kit/white spotting locus: w~~,wV,d' and W. EMBO J. 9,1805- 18 13.

OK-Urtreger, A., Avivi, A., Zimmer, Y., Givol, D., Yarden, Y. and Lonai, P. (1990). Developmental expression of c-kit, a proto-oncogene encoded by the W locus. Development 109,9 11-923.

Rawles, M. E. 1947. Origin of pigment cells from the neural crest in the mouse embryo. Physiol. Zool. 20,248-265.

Reith, A. D., Rottapel, R., Giddens, E., Brady, C., Forrester, L. and Bernstein, A. (1990). W mutant mice with mild or severe developmental defects contain distinct point mutations in the base domain of the c-kit receptor. Genes & Dev. 4,390-400.

Reith, A. D. and Bernstein, A. (1991). Molecular biology of the W and Steel loci. Genome analysis Volume 3: Genes and phenotypes. Cold Spring Harbor Laboratory Press, 105-133.

Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press.

Scott, G., Ewing, J., Ryan, D. and Abboud, C. (1994). Stem cell factor regulates human melanocyte-matrix interactions. Pigment Cell Res. 7 (I), 44-5 1. Serbedzija, G. N., Fraser, S. E. and Bronner-Fraser, M. (1990). Pathways of trunk neural crest migration in the mouse embryo as revealed by vital dye labeling. Development 108,605-6 12.

Steel, K. P., Davidson, D. R. and Jackson, I. J. (1992). Trp-2DT, a new early melanoblast marker, shows that steel growth factor (c-kit ligand) is a survival factor. Development 115,111 1-1 119.

Tsujimura, T., Morii, E., Nozaki, M., Hashimoto, K., Moriyama, Y., Takebayashi, K., Kondo, T., Kanakura, Y. and Kitamura, Y. (1996). Involvement of transcription factor encoded by the mi locus in the expression of c-kit receptor tyrosine base in cultured mast cells of mice. Blood 88, 1225- 1233.

Ward, S., Burns, A., Torihashi, S. and Sanders, K. (1994). Mutation of the proto- oncogene c-kit blocks development of interstitial cells and electrical rythmicity in murine intestine. J. Physiol. 480.1,9 1 -97.

Wehrle-Haller, B. and Weston, J. A. (1995). Soluble and cell-bound forms of steel factor activity play distinct roles in melanocyte precurser dispersai and survival on the lateral neural crest migration pathway. Developnzen? 121,73 1-742.

Williams, D. E., Eisenmann, J., Baird, A., Rauch, C., Van Ness, K., March, C. J., Park, L. S., Martin, U., Mochizuki, D. J., Boswell, H. S., Burgess, G. S., Cosman, D. and Lyman, S. D. (1990). Identification of a ligand for the c-kit proto-oncogene. Cell 63, 167-174.

Yasuda, H., Galli, S.J. and Geissler, E.N. (1993). Cloning and functional analysis of the mouse c-kit promoter. Bioch. Bioph. Res. Comm 191 (3), 893-90 1.

Yoshida, H., Kunisada, T., Kusakabe, M., Nishikawa, S. and Nishikawa, S.I. (1996). Distinct stages of melanocyte differentiation revealed by analysis of nonuniform pigmentation patterns. Development 122, 1207- 12 14. Zsebo, K. M., Williams, D. A*, Geissler, E. N., Broudy, V. C., Martin, F. H., Atkins, H. L., Hsu, R. Y., Birkett, N. C., Okino, K. H., Murdock, D. C., Jacobsen, F. W., Takeishi, T., Cattanach, B. M., Galli, S. J. and Suggs, S. V. (1990). Stem cell factor is encoded at the Sl locus of the mouse and is the ligand for the c-kit tyrosine base receptor. Cell 63,2 13-224. Chapter 3

The W/git gene is required for the development of the interstitial cells of Cajal and for intestinal pacemaker activity

This is a revised version of a report published by: Jan D. Huizinga, Lars Thuneberg, Michael Kliippel, John Malysz, Hanne B. Mikkelsen and Alan Bernstein (1995). Nature 373,347-349

My contributions to the work described in this chapter were as follows: I analyzed the expression of Kit mRNA in adult small intestine of wild type and W/WV mice, which demonstrated the expression of Kit in a cell type forming a cellular network between the two external muscle layers of the mouse small intestine in wild type, but not in WWmice. In addition, I was actively involved in the preparation of the manuscript. Methylene blue staining of KC,immunohistochemistry and electron microscopy were done by Lars Thuneberg and Hanne B. Mikkelsen. The electrophysiology was done by Jan D. Huizinga and John Malysz. ABSTRACT

The pacemaker activity in the mammalian gut is responsible for generating anally propagating phasic contractions. The cellular basis for this intrinsic activity is unknown. The smooth muscle cells of the external muscle layers and the innervated cellular network of interstitial cells of Cajd, which is closely associated with the external muscie layers of the mammalian gut, have both been proposed to stimulate pacemaker activity (Thuneberg, 1989; Thuneberg, 1982; Liu et al., 1993; Dude et al., 1983; Liu et d., 1994). The interstitial cells of Cajal were identified in the last century but their developmental origin and function have remained unclear. Here we show that interstitial cells of Cajal express the Kit receptor tyrosine kinase. Furthermore, mice with mutations in the dominant white spotting (W) locus, which have cellular defects in hematopoiesis, melanogenesis and gametogenesis (Reith and Bernstein, 1991) as a result of mutations in the Kit gene (Chabot et al., 1988; Geissler et al., 1988), also lack the network of interstitial cells of Cajal associated with Auerbach's nerve plexus and intestinal pacemaker activity. RESULTS AND DISCUSSION

Injection of antibodies directed against the extracellular domain of the Kit receptor tyrosine base into newborn mice leads to changes in in vitro contraction patterns in the small intestine and absence of Kit messenger RNA in the myenteric plexus area (Maeda et al., 1992). This observation prompted us to to investigate whether the Kit receptor might play a role as a signalling molecule required for the development of the interstitial cells of Cajal (KC) and therefore be essential for intestinal pacemaker activity. To determine whether mutations at the murine W/Kiir locus might affect noddevelopment of the ICC, we fust examined the morphology of the intestines of mutant WWmice and their control litterrnates. Control mice (albino, +I+ and W/+)had normal ICC networks in the small intestine, visualized by selective uptake of of methyiene blue (Figure 1A) and confirmed by electron microscopy (Figures 2A,B). By contrast, the network of KC in the myenteric plexus region was absent in W/W mice (Figure IB), again confirmed by electron microscopy (Figures 2C,D). In the mutants, only scattered methylene blue-positive cells which might be genuine KC were observed (Figure 1B). Using electron microscopy, we counted 5 1 ICC along a length of 1,109 cross sectioned muscle cells bordering the myenteric plexus in W/+ control mice; by contrast only 7 KC were observed along a length of 3,059 muscle cells in WWmice. Quantitatively,'lO% of muscle cells bordering the Auerbach's plexus contacted KC in control mice, compared with only 1.5% in WWmice. We also observed the direct apposition of large stretches of the circular and longitudinal smooth muscle layers in WWmice in the absence of the intermediary network of KC found in normal animals (Figure 2C). Light and electron microscopic analysis of duodenal muscularis revealed no differences in the density, morphology or ultrastrustural features of enteric neurons or glial cells between +/+ and W mutant mice, indicating that defects in the Kit signalling pathway do not affect neuronal or ghal cells in the Auerbach's plexus region. These data indicate that a functional Kit receptor is required for the development of ICC in the Auerbach's plexus region of the small intestine. To determine whether the ICC deficiency in W mutant mice was a cell-autonomous defect, we analyzed Kit RNA and protein expression in W mutant mice and control animals. We used a polyclonal antibody directed against the intracellular domain of the Kit receptor tyrosine base to localize the protein. In wild type and W+,high levels of Kit expression were observed between the longitudinal and circular muscle layers at the level of the Auerbach's plexus (Figure ID). The stained cells surrounded the ganglia and their distribution in the Auerbach's plexus area was identical to that of the methy lene blue-stained ICC (Figure 1C). By contrast, no Kit immunoreactivity was found in or between the muscle layers of W/Wv mice (Figure 1E). We next performed whole-mount RNA in situ experiments on the ileum of wild type and WWmice. The ileum of +I+ mice contained Kit-positive cells whose organization was identical to that of ICC, as revealed by methylene blue staining (Figure 3A). These cells were localized between the longitudinal and circular muscle layers in an identical position to the KC. By contrast, no Kit-positive cells were observed in the muscle layers of WWmice (Figures 3B.D). To determine whether these morphological and cellular differences between +/+ and W mutant mice were associated with functional anomalies, we measured the electrical activity of the small intestine muscle layers. Normal mice displayed slow- wave-type action potentials with an amplitude of 2 1.5 f 5.9 mV,a frequency 32.4 + 1.0 cycles per minute (range 30-36 c.p.m.) and a resting membrane potential -60.0 f 3.0 mV (Figure 4; n=9). By contrast, the ilea of WWmice failed to display any slow-wave-type action potentials (Figure 4). The membrane potential was -44.8 f 1.3 mV (n=14) and at irregular frequency fast spike-like action potential arose from it, singly or in groups of 2-6. Their amplitude was 16.4 f 2.1 mV, the frequency ranged from 4 to 20 c.p.m., and was irregular within one preparation. In 4 out of 14 WW mice, no spontaneous action potentials were observed. The slow-wave compcnent or pacemaker activity of gut smooth muscle cells is insensitive to L-type calcium channel blockers. In the presence of the blockers nifedipine or D600, the slow-wave component of the action potentials in +/+ mice remained unaltered, whereas the electrical activity of WWmice was completely abolished (n=14; Figure 4). These data demonstrate that mutations at the murine W locus lead to the absence of the ICC network in the Auerbach's plexus region and of pacemaker activity in the small intestine, demonstrating an essential role for ICC in gut pacemaker activity. Because the ICC express the Kit receptor, we conclude that the absence of these cells in W mutant mice reflects a direct role for Kit in KC development. The phenotype of W mutant mice was thought to be restricted to cells of the hematopoietic, germ and melanocyte lineages (Reith and Bernstein, 1991), even though Kit and its ligand Steel factor are contiguously expressed in additional anatomical sites, including the small intestine (Motro et al., 1991; Orr-Urtreger et al., 1990). Thus, our experiments extend the range of cell types affected by W mutations to include the interstitial cells of Cajal. Mice with mutations in Steel factor (SVS~~ mutant mice) display abnormalities in the gut similar to those found here in W mutant mice (data not shown). Our data suggest that the functional gut abnormalities and megacolon observed in individuals with piebaldism (Bolognia and Pawalek, 19881, a hypopigmentation disorder that also results from mutations in the Kit proto-oncogene (Giebel and Spritz, 1991; Fleischman et al., 1991) reflect an identical function of the Kit signalling pathway in the development of ICC in humans. Mutations in the RET proto-oncogene are also associated with gut abnormalities in humans (Hirschsprung' s disease) and in gene-targeted mutant mice, as the result of a cellular deficit in neural crest-derived enteric neurons (Romeo et al., 1994; Edery et al., 1994; Schuchardt et al., 1994). Thus, two members of the receptor tyrosine kinase family, Kit expressed in KC and RET expressed in neural crest-derived ganglion cells, are both essential for normal gut function in mammals.

Figure 1. Histochemical analysis of the small intestine in wild type and W mutant mice. A&, Double staining of interstitial cells of Cajal associated with Auerbach's plexus (ICC) (shown as the stained cellular network using vital staining with methylene blue) and (fluorescence; ingestion of fluorescein isothiocyanate FITC-labelled dextran) (Mikkelsen et al., 1988). Whole mounts of the isolated duodenal muscularis externa of a control albino mouse (A) and WxMN mouse) (B); (A) is representative of the normal staining pattern of both cell types in albino mice as well as Wv/+ mice. Normally, ICC together with macrophages form a panid sheath around ganglia and primary fascicles of Auerbach's plexus (AP). In (B), only macrophages are present in normal numbers and are normally organized. C-E,Sectioned jejunal tissue. (C), lpsection of Epon-embedded tissue, processed with preservation of methylene blue, no post-staining. KC are seen as the blue, broken line between the two layers of the muscularis extema (between dashed lines). Arrowheads point to unstained, partidy enveloped elements of AP (compare with A). D.E. Frozen 8psections, showing Kit irnmunoreactivity in ICC in a WYl+ mouse and its absence in the WWmouse(E). Magnification, x225 (A); x320 (B-E);scale bar. 25p. METHODS. Methylene blue was preserved through dehydration and embedding by precipitation as a hexachloroplatinate (Thuneberg, 1992). Immunocytochemistry was performed on unfixed frozen sections of small intestines from WXW mice and their controls. Both the PAP and the biotin-streptavidin and Texas-red methods were used (Mikkelsen et al., 1993). Figure 1 Figure 2. EM analysis of the small intestine of wild type and W mutant mice. (A-D),

Electron micrographs of the jejunal myenteric plexus area in an albino mouse (A), a W+ mouse (B), and a WWmouse (C,D). (A), Normal appearance of ICC in the AP region (1). tightly associated with nerve fascicle (N) of APTbetween the longitudinal (LM) and circular (CM) muscle layers. An abundance of mitochondria (mows) and caveolae are distinctive ultrastructural features of KC. (B). Prestaining with methylene blue assists the differentiation of ICC (1, with increased granularity and electron-density of nuclei and ribosomal areas) from fibroblasts (F), macrophages (M, macrophage process) and pericytes

(C,capillary). (C,D). The overall organization of the muscle layers seemed normal. Although cell types other than KC were counted in normal numbers in W/WV mice, one result of the absence of ICC was a strong increase in the extension of areas of direct apposition of muscle cells of LM and CM (C). Sheath cells around nerves were either absent (D.small tertiary nerve between the muscle layers) or had fibroblast ultrastructure. Scale bars: 2pm (A-C);

l~rm(Dl* METHODS. Tissues were fixed and processed for electron microscopy (Phillips 300 microscope) by routine methods. Figure 2 Figure 3. Whole-mount RNA in situ analysis of Kit expression in the small intestine of wild type and W mutant mice. (A), A network of Kitexpressing cells is present in the isolated external muscle layers in +/+ mice. identical in appearance to the KC network (compare Figure 1A). (B), Kit-positive cells were absent in similar whole-mounts of WWmice. Cross-sectioning of the whole-mount tissues localized the Kit-positive cells in +/+ mice between the longitudinal and circular muscle layers (C, compare Figure 1C at the same magnification) and confiied their absence in WWmice (D). Scale bars: 50pm (A,B);

25p(CP). METHODS. Whole-mount RNA in situ hybridization was performed according to the method of D.He~queand D. Ish-Horowisz (personal communication), with the following modifications. Roteinase K digestion was for 30 rnin at a concentration of 20pg/ml. The hybridization mix contained 5 X SSC, 50% formaaide, 5mM EDTA, 50pg/d yeast tRNA, 0.2% Tween-20, 0.5% CHAPS and 100pg/ml heparin, pH 6.5 with citric acid. After in situ hybridization, tissues were postfixed in 4% paraformaldehyde before wax embedding and sectioning. The Kit riboprobe has been described. Figure 3 Figure 4. Action potential generation in wild type and W mutant mice. The lefi panel depicts the slow-wave-type action potential generated by +/+ mice. Slow-wave-type action potential were generated at a constant frequency. L-type calcium channel blockers did not have any influence on the slow-wave component. Occasionally spikes were superimposed on the slow wave and these were abolished by L-type calcium channel blockers. The right panel depicts typical spike-like action potentials generated by the small intestinal smooth muscle cells of W/Wv mice. Fast spikes occurred as single spikes or in groups os 2-3. These spikes were completely abolished by the action of L-type calcium channel blockers. METHODS. Female mice were killed by cervical dislocation. The small intestine was exposed by a midline abdominal incision and a 3-5 cm segment was removed, 1 crn From the gastroduodenal sphincter. The segment was placed in a dissecting dish filled with oxygenated (95% 02 and 5% C02)Krebs solution and opened flat. It was then pinned to the

Sylgard bottom of a dissecting dish and the mucosa removed by sharp dissection. Muscle strips, 15 mm long, were prepared by cutting them parallel to the longitudinal muscle bundles: The tissue was 1.5 mm wide over a length of 10 mm and gradually widened to 3 rnm at the proximal end At the wide end an area of the tissue (3 X 3 mm) was pinned to the Sylgard bottom of a transfer holder. This transfer holder with the tissue attached was then placed in an organ bath at 36.0-37.0 OC. The end of the tissue that was not pinned was tied to a force transducer to record mechanical con- tion. Intracellular recordings were made using microelectrodes with 30-60 MR tip resistance because microelectrodes with a tip resistance within this range have a tip diameter of -150p, which was appropriate for impaling cells.

Microelectrodes were prepared from 1.2 mm outside-diameter glass capillaries (WPI) and filled with 3 M KCl. A microelectrode was inserted into a microelectrode holder (WPI M700P) connected to an electrometer as a high impedance probe.

Bolognia, J.L. and Pawalek, J.M. (1988). Biology of hypopigmentation. J. Am. Acad. Dennatol. 19,2 17-255.

Chabot, B., Stephenson, D., Chapman, V., Besmer, P. and Bemstein, A. (1988). The proto-oncogene c-kit encoding a transmembrane tyrosine base receptor maps to the mouse W locus. Nature 335,88,

Durdle, N.G., Kingma, Y.J., Bowes, K.L. and Chambers, M.M. (1983). Origin of slow waves in the canine colon. Gastroenterology 84,357-382.

Edery, P., Lyonnet, S., Mulligan, L.M., Pelet, A., Dow, E., Abel, L., Holder, S., Nihoul-Fekete, C., Ponder, B.A. and Munnich, A. (1994). Mutations of the RET proto-oncogene in Hirschsprung's disease. Nature 367, 320-32 1.

Fleischrnan, R.A., Saltman, D.L., Stastny, V. and Zneimer, S. (1991). Deletion of the c-kit protooncogene in the human developmental defect piebald trait. Proc. Natl. Acad. Sci. USA 88, 10885-10889.

Geissler, E., Ryan, M. and Housman, D. (1988). The dominant white spotting (W) locus of the mouse encodes the c-kit proto-oncogene. Cell 55, 185- 194.

Giebel, L.B. and Spritz, R.A. (1991). Mutation of the KIT (masthtem cell growth factor receptor) protooncogene in human piebaldism. Proc. Natl. Acad. Sci. USA 88, 8696-8699.

Huizinga, J.D., Faraway, L. and Den Hertog, A. (1991). Generation of slow-wave- type action potentials in canine colon smooth muscle involves a non-L-type ca2+ conductance. J. Physiol. 442, 15-29.

Liu, L.W. and Huizinga, J.D. (1993). Electrical coupling of circular muscle to longitudinal muscle and interstitial cells of Cajal in canine colon. J. Physiol. 470,445- 461. Liu, L.W., Thuneberg, L. and Huizinga J.D. (1994). Selective lesioning of interstitial cells of Cajal by methylene blue and lights leads to Loss of slow waves. Am. J. Physiol. 266, pG485-496.

Maeda, H., Yamagata, A., Nishikawa, S., Yoshinaga, K., Kobayashi, S., Nishi, K. and Nishikawa, S. (1992). Requirement of c-kit for development of intestinal pacemaker system. Developmenf 116,369-375.

Mikkelsen, H.B., Thuneberg, L. and Wittrup, I.H. (1988). Selective double staining of interstitial cells of Cajal and macrophage-like cells in small intestine by an improved supravital methylene blue technique combined with FITC-dextran uptake. Anat. Embryol. 178, 191-195.

Mikkelsen, H.B ., Huizinga, J.D., Thuneberg, L. and Rumessen, J.J. ( 1993). Lmmunohistochemical localization of a gap junction protein (co~exin43)in the muscularis externa of murine, canine and human intestine. Cell Tissue Res. 274, 249- 256.

Motro, B., Van Der Kooy, D., Rossant, J., Reith, A. and Bernstein, A. (1991). Contiguous patterns of c-kit and steel expression: analysis of mutations at the W and Sl loci. Development 113, 1207- 122 1.

Orr-Urtreger, A., Avivi, A., Zimmer, Y., Givol, D., Yarden, Y. and Lonai, P. (1990). Developmental expression of c-kit, a proto-oncogene encoded by the W locus. Development 109,911-923.

Reith, A. D. and Bernstein, A. (1991). Molecular biology of the W and Steel loci. Genome analysis Volume 3: Genes and pheno~pes.Cold Spring Harbor Laboratory Press, 105-133.

Romeo, G., Ronchetto, P., Luo, Y., Barone, V., Seri, M., Ceccherini, I., Pasini, B., Bocciardi, R., Lerone, M., Kaariainen, H. et al. (1994). Point mutations affecting the tyrosine basedomain of the RET proto-oncogene in Hirschspmg's disease. Nature 367,377-378. Schuchardt, A., D'Agati, V., Larsson-Blomberg, L., Constantini, F. and Pachnis, V. (1994). Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase receptor Ret. Nature 367,319-320.

Thuneberg. L. (1982). Interstitial cells of Cajal: intestinal pacemaker cells? Adv. Anat. Embryol. Cell Bid. 71, 1-130.

Thuneberg, L. (1989). The gastrointestinal sysrem, in Handbook of Physiology. (eds. Schultz, G.S., Wood, J.D. and Rauner, B.B.). 349-386 (American Physiological society, Bethesda).

Thuneberg, L. (1992). Gastroenterology 103, 1388. Chapter 4

Developmental origin and Kit-dependent postnatal development of the interstitial cells of Cajal in the mammalian small intestine

This is a revised version of a report published by: Michael Kliippel, Jan D. Huizinga, John Malysz and Alan Bemstein Developmental Dynamics, in press.

I performed all experiments except the electrophysiological experiments described in Figure 4, which were performed by Jan D. Huizinga and John Malysz. ABSTRACT Interstitial cells of Cajal (KC)form a network of cells between the external longitudinal and circular muscle layers at the level of the parasympathetic Auerbach's plexus in the mammalian small intestine. These cells express the Kit receptor tyrosine kinase and are essential for intestinal pacemaker activity. W mutant mice carrying structural mutations in the Kit gene lack both the network of KC and intestinal pacemaker activity. We were interested in the developmental origin of the cells that make up the network of KC. In addition, the specific stages of ICC development that require a functional Kit receptor have not been characterized-We show that ICC originate from mesenchymal progenitor cells that co-express both Kit and smoorh muscle myosin heavy chain, a marker specific for smooth muscle, during embryogenesis. KC and longitudinal smooth muscle lineages begin to diverge late in gestation. Embryos homozygous for the regulatory ~ba~d~~(~~)mutation do not express Kit in these rnesenchymal progenitor cells. Nevertheless, ~bd/Wbdmice display a normal network of ICC and normal smooth muscle layers at postnatal day 5 (p5). Adult ~bd~bdmice lack a Iilnctional KC network and intestinal pacemaker activity, due to a failure of the ICC to increase in numbers after p5. These data suggest a common developmental origin of the ICC and the longitudinal smooth muscle layers in the mammalian small intestine and show that Kit expression is necessary for the postnatal development and proliferation of ICC, but not for the initial cell lineage decision towards an ICC fate during embryogenesis or for smooth muscle development. INTRODUCTION A century ago, Ramon y Cajal described the specific staining characteristics of the "cellule nerveuse intersti tielle" between the external longitudinal and circular muscle layers of the intestine at the level of the Auerbach's plexus (Cajal, 1893; 19 11 ). For more than 100 years, the function and the developmental origin of these cells remained unclear. While Cajd and others believed that these interstitial cells of Cajal (KC) were primitive nerve cells, it has also been suggested that these cells are specialized smooth muscle cells (Faussone-Pelligrini, 1985a), while others characterized them as fibroblast-like (reviewed by Thuneberg, 1982). Recently, we and others demonstrated that the cells that form the network of ICC express the Kit receptor tyrosine kinase (RTK)and furthermore that W mutant animals with loss-of- function mutations in Kit (Chabot et al., 1988; Geissler et al., 1988) and SZ mutant animals with Ioss-of-function mutations in the Kit Iigand, Steel factor (Copeland et al., 1990; Huang et al., 1990; Williams et al., 1990; Zsebo et al., 1990), lack both the ICC network and intestinal pacemaker activity (Maeda et d., 1992; Ward et d., 1994; Huizinga et al., 1995). In contrast, the enteric nervous system is not affected in these mutant mice. These observations have demonstrated that the ICC are neccessary for the generation of electrical rhythmicity in intestinal muscle. KC also appear to play an equivalent role in humans: infantile hypertrophic pyloric stenosis (IHPS), a common pediatric disorder characterized by projectile vommiting, has been suggested to be caused by a lack of coordination between the movements of the pyloric sphincter and the contractions of the stomach (Hayes and Goldenberg, 1957). It has recently been shown that Kit-positive ICC are absent in the pyloms of MPS patients suggesting that the absence of ICC is responsible for IHPS (Vanderwinden et al., 1996). The findings that KC are necessary for intestinal pacemaker activity and express the Kit RTK raise interesting questions about their developmental origins. The Kit receptor is expressed in diverse cell types, including neural crest-derived melanocytes and a subset of sensory neurons of the peripheral nervous system, cells in the central nervous system, primordial gem cells and hematopoietic cells (Orr- Utregger et al., 1990; Motro et id., 199 1; Keshet et al., 1991). Recently, the neural crest has been excluded as the developmental origin of the ICC in the chick (Lecoin et al., 1996) and in the mouse (Young et al., 1996). Different types of interstitial cells, located either within the circular and longitudinal smooth muscle layers or at the level of the Meisner plexus of the canine intestine, share some morphological and immunohistochemical characteristics with smooth muscle cells (Torihashi et al., 1994). Nevertheless, the different types of interstitial cells in the intestine are morphologically distinct (Thuneberg, 1982; Torihashi et al., 1994) and there is no evidence suggesting that they share a common developmental origin or function. Indeed, the ICC at the level of the Auerbach's plexus are ultrastructurally dissimilar to smooth muscle cells and do not express any of the smooth muscle markers present in some other interstitial cells in the gut (Torihashi et al., 1994). In addition, KC are morphologically distinct from fibroblasts (Torihashi et al., 1994). Moreover, interstitial cells located in the circular muscle layer have a distinct function in mediation of inhibitory neurotransmission (Burns et al., 1996). Thus, ICC of the Auerbach's plexus in adults are functionally, ultrastructurally and irnmunohistochemically distinct from smooth muscle cells, as well as other interstitial cells in the smai1 intestine. Recently, it has been shown that Kit-positive cells in the outer layers of the embryonic gut do not co-express smooth muscle actin or myosin protein (Torihashi et al., 1997). However, these authors did observe cells that co-expressed Kit and Desmin, an intermediate filament protein expressed by smooth muscle cells, in a few cells in the developing longitudinal muscle layer. Thus, they suggested that the few double-positive cells represent smooth muscle progenitor cells, while Kit-positive cells represent KC progenitors. To gain fuaher insight into the developmental origin of the KC, we have analyzed the mRNA expression of Kit and the smooth muscle marker smooth muscle myosin heavy chain (SMMHC)in the gastrointestinal tract throughout embryonic development. To elucidate the role of Kit in KC development, we have studied the effects of two regulatory W mutations, ~57and Wbanded(bd) (Kliippel et al., 1997) on Kit expression, KC development and intestinal pacemaker activity. The results presented here indicate that Kit-positive KC and SMMHC-positive smooth muscle cells develop from a common mesodermally-derived mesenchymal pluripotential precursor that expresses both mRNA markers. Subsequently, we observed a restriction of Kit expression to ICC, and, in parallel, a restriction of SMMHC expression to smooth muscle cells. Thus, these data do not support the hypothesis put forward by Cajal and others, classifying KCas 'primitive nerve cells'. In addition, only the ~bd,but not the ~57mutation, affects Kit expression, ICC development and intestinal pacemaking, demonstrating that the genomic inversion in ~bdDNA (Kliippel et al., 1997) disrupts cis-acting regulatory sequences required for Kit expression in ICC. Interestingly, Kir expression is not necessary for a lineage decision towards the KC fate, but appears to play a role in postnatal proliferation of KC. MATERIALS AND METaODS

Mice and embryos C57BW6J and ~57~57mice were purchased from the Jackson Laboratory. C3H, 101 and ~bd/~bdmice were obtained from the MRC Radiobiology Unit, Chilton, UK. Embryos were derived from appropriate matings of C57BV6J or C3W101 for wild type controls. Mutant embryos were derived from matings of homozygous ~57and ~bd mice. The noon after vaginal plug was considered day 0.5 p-C.

RNA in sihr hybridization RNA in situ hybridization on frozen lopsections and whole mount tissues were performed essentially as previously described (Motro et al., 1991; Huizinga et al., 1995; Kluppel et al., 1997). The Kit cRNA probe has been described previousIy (Motro et al., 1991). Double labelling in situ hybridization was performed as previously described (Kliippel et ai., 1997), using a digoxigenin-labelled smooth muscle myosin heavy chain (SMMHC) cRNA probe (Miano et al., 1994) and a 3%- labelled Kit cRNA probe. Digoxigenin-labelled cells and 3%-labelled cells were in different optical planes due to a Parlodion coating of the slides prior to submersion in emulsion to avoid a color reaction between emulsion and the alkaline phosphatase color product. Therefore, two images were taken of the same frame in different optical planes. Sense probes as controls did not produce any labeling above background.

Supra-vital methylene blue staining Methylene blue staining of the whole gastrointestinal tract was performed as previously described (Mikkelsen et al., 1988). Briefly, mice were killed by cervical dislocation. The gastrointestinal tract 60m stomach to rectum was removed, rinsed in PBS (without Calcium or Magnesium) and submerged in 0.7 mM Lysolecithin in PBS for 1 min. at RT. Tissues were then put in 50 mM Methylene blue in fresh Krebs solution in the dark for 40 min at RT. The solution was bubbled with 95 % OxygedS % CO2 or with air throughout the procedure. After the staining, tissues were briefly rinsed in PBS and submerged in a fmative containing picric acid, e.g. Bouin's fluid. Photographs of the whole mount tissues were taken while the tissues were in the fmative.

Electrophy siology Electrophysiological experiments were performed as previously described (Huizinga et al., 1995).

Apoptosis assays We used the Oncor Apoptaq-Plus Fluorescein apoptosis detection Kit on frozen 10pm sections of small intestines, which were fixed in 4% Paraformaldehyde at 40C for 24 hrs, followed by soaking in 0.5M Sucrose at 40C for 24 hrs, and embedded in TissueTek embedding medium prior to cryosectioning. The assays were performed according to the recommendations of the manufacturer. Counterstaining was performed using the Hoechst 33258 Fluorescent Nuclear Stain. For a positive control, slides were treated with DNaseI prior to the TUNEL procedure; for a negative control, TdT was omitted during the procedure. RESULTS

Analysis of Kit expressing ceUs in the developing intestine of wild type, ~57and ~bdembryos

To gain insight into the developmental origin of the KC, we analyzed Kit expression in the developing intestines of wild type, ~571~57and ~bdhVbd mice throughout embryogenesis by RNA in situ hybridization. At d10.8 and dl 1.8, Kit was only expressed in the epithelium of the gut, but was not yet evident in the mesenchyme of the developing smooth muscle layers in wild type (Motro et al., 1991; Keshet et al., 199 I), ~57or ~bd homozygous mutant embryos (data not shown). At d14.5, both wild type (Figs. 1A and 1B) (Keshet et al., 1991) and ~571~57 (Figs. 1C and 1D) embryos displayed a broad band of strong Kii expression in the developing external muscle layers of the intestine. Occasionally, we observed a restriction of Kit expression, reminiscent of later developmental stages (data not shown and see below). Kit expression was completely absent in the external muscle layers of ~bd,/~bd embryos at d14.5 (Figs. 1E and IF). In contrast, Kit expression in the endothelium was not altered in either ~57or ~bd homozygous embryos (Figs. 1A,C,E). By dl6 of embryonic development, Kit expression became restricted to the developing muscle layers in some areas of wild type intestine and already resembled the very localized adult expression pattern of Kit in the KC (Fig. lG, loop 2), whereas other areas still exhibited a broad band of Kit expression (Fig. 1G, loop 1). The pattern of Kit expression was unaffected by the ~57mutation; in contrast, Kit expression was not observed in the developing muscle layers of ~bd/~bdembryos (data not shown). Interestingly, epithelial expression of Kit was markedly downregulated and barely detectable in all three embryos at this stage (Figs. 1G and data not shown). Two days later, at dl8 of embryogenesis, the patterns of Kit expression in wild type (Figs. lH,I) and ~571++7(data not shown) embryonic gut was restricted to specific groups of cells and single cells in the developing muscle layers in approximately 70% of the intestine, resembling the expression pattern of Kit in the ICC in the adult. For example, figure 11 shows two Kit-positive cells in the muscularis externa of a dl8 wild type embryo. The remaining 30% of the gut at this stage still exhibited a wide distribution of Kit expression in the developing muscle layers (data not shown). Again, there was no expression in ~bd/~bdembryonic gut at this stage (data not shown).

Figure 1. RNA in situ analysis of Kit and SMMHC expression in embryonic gut. (A)Dark field and (B) bright field photograph of wild type d14.5 gut shows strong Kit expression in the developing outer muscle layers (arrow) and in the endothelium (arrowhead). (C and D) An identical pattern of Kit expression was observed in ~57~57embryos. whereas FVbd/Wbd embryos (E and F) expressed Kit only in the endothelium, but not in the outer muscle layers. (G) Dark field photograph of dl6 wild type embryonic gut shows a broad band of Kit-positive cells in loop 1 (mow), while loop 2 already displays a marked reduction of Kit positive cells, resembling the adult pattern of Kit-positive KC. Arrowheads point to the outer developing muscle layers of loop 2. (H)Dark field photograph of dl8 wild type gut shows a restricted expression pattern of Kit, resembling the adult pattern of Kit-positive ICC (arrow). (I) Two Kit-positive cells in the developing muscularis externa at d18. Note the degree of loss of Kit expression in the muscle layers, compared to d14.5 or dl6 embryos.

Scale bar represents 60 p in (A-F), 100 p in (G),200 ~UII in (H) and 50 pn in (I). Figure 1 Continuous restriction of the numbers of Kit-positive cells during development of the outer muscle layers

To exclude the possibility that the reduction in Kit expression in the developing muscle layers was due to the selective proliferation of Kit-negative rnesenchyrnal cells of the muscle layer, we analyzed the numbers of Kir-positive and -negative cells in the developing muscle at different embryonic stages. Only parts of the gut that were sectioned in a round, as opposed to an oval fashion, were used for cell counting in order to ensure equivalent comparison between different developmental stages. We analyzed 5 sections of d14.5 tissue and 6 sections of dl8 tissue from 2 embryos of each stage. At d14.5, there was an average of 516.2 + 38.7 cells/per section in the muscle layers; all of these cells expressed high levels of Kit (as judged by increased grain density due to hybridization of the radiolabelled Kit probe). In contrast, at dl8 the average number of cells in the muscle layer increased to 773.2 f 48.6, but the number of Kit-positive cells decreased to an average of 82.3 f 13.1 per section. These results demonstrate that the numbers of Kit-expressing cells in the developing muscle layers decrease during embryogenesis. Developmental origin of ICC

To investigate the developmental origin and lineage relationships of the KC, we analyzed both Kit expression and expression of a specific marker of smooth muscle cells, smooth muscle myosin heavy chain (SMMHC) (Miano et al., 1994). To determine whether there might be cells that co-expressed both genes, reflecting common cells of origin, we performed single and double labeling RNA in situ experiments, using digoxigenin- and radio-labelled SMMHC and Kit cRNA probes on the same sections and on adjacent sections of wild-type embryos. In a double labelling experiment on d14.5 embryos, cells expressing SMMHC were identified using a digoxigenin-labelled probe and were present in the developing outer muscle layers (muscuiaris externa) of the gut at d14.5 (Fig. 2A). Strong expression was observed in the inner, circular muscle layer, whereas lower, but clearly detectable levels of SMMHC expression were present in the outer, developing longitudinal muscle layer (Figs. 2A,D,G,J). To confii expression of SMMHC in this future longitudinal muscle layer, we hybridized sections adjacent to those used in the double labelling experiment to a radiolabelled SMMHC probe. These experiments confirmed that while the circular muscle layer displayed strong SMMHC expression, the future longitudinal muscle layer expressed lower, but clearly detectable, levels of SMMHC mRNA at d14.5 of embryogenesis (Figs. ZF,I,L). This differential expression of SMMHC reflects the fact that the inner part of the muscularis externa represents the early developing circular muscle layer, while the outer, lightly labelled area represents the future longitudinal muscle layer, which only differentiates postnatally (Faussone- Pellegrini, 1985b; Torihashi et al., 1997). At d14.5 of embryogenesis, we observed three main expression patterns for Kit in double labelling experiments using a radiolabelled Kit probe. Firstly, in a few areas, Kit was uniformly expressed in both the circular and future longitudinal muscle layers (Fig. 2B; compare to SMMHC expression in the same section in Fig. 2A). Secondly, in most areas Kit was strongly expressed in the future longitudinal muscle layer, with occasional labelling of cells of the circular muscle layer (Fig. 2E, compare to SMMHC expression in the same section in Fig. 2D and to SMMHC expression in an adjacent section in Fig. 2F; Fig. 2H, compare to SMMHC expression in the same section in Fig. 2G and to SMMHC expression in an adjacent section in Fig. 21; Fig. 2K,compare to SMMHC expression in the same section in Fig. 21 and to SMMHC expression in an adjacent section in Fig. 2L). We next investigated whether this broad band of Kit expression observed with the radiolabelled probe indicated Kit expression in the majority of cells in the outer layers of the gut. To address this question, we hybridized a sectioned d14.5 embryo with a digoxigenin-labelled Kit probe. Almost all cells of the outer layer were also labelled with this probe, indicating that Kit is expressed in the majority of cells in the developing longitudinal muscle layer (Fig. 2C), as previously demonstrated (Torihashi et al., 1997). These results demonstrate co-expression of Kit and SMMHC rnRNAs in the area of the future longitudinal and some areas of the circular muscle layers at d14.5. Thirdly, in a few areas, Kit expression was markedly downregulated (Figs. 2H.K). By comparing silver grain densities for the Kit and SMMHC radiolabelled probes on adjacent sections, we observed a negative correlation between Kit and SMMHC expression in the future longitudinal muscle layers. In areas of high Kit expression, SMMHC levels were low to intermediate (Figs. 2 H,K),whereas in areas with reduced Kit expression, SMMHC levels were increased (Figs. 2H,K). To determine whether Kit and SMMHC continue to be co-expressed at later stages of embryogenesis, we performed double labeling RNA in situ experiments on dl 8 embryos. SMMHC was expressed in two distinct layers, corresponding to the circular muscle (CM) and the developing longitudinal muscle (LM) layers (Fig. 2M). The cells located between the CM and LM layers did not express SMMHC (Fig. 2M). hterestingly, Kit was no longer expressed in either the CM or LM layers, but was localized in single cells or groups of cells between the muscle layers (Fig. 2N), suggesting that the ICC and smooth muscle lineages have begun to diverge at this stage in development. Embryos homozygous for the ~57or wbd mutations at d14.5 displayed the normal d14.5 SMMHC expression pattern (data not shown), suggesting that the absence of Kit expression in the gut of wbdl~bdembryos does not interfere with SMMHC expression or smooth muscle development.

Figure 2. Single and double labelling RNA in siru experiments of wild type embryos using a digoxigenin (dig)-labelled and radiolabelled (rl) SMMHC and Kit probes. (A to L) d14.5; (M and N), d18. (A,B) Double labelling experiment using a dig SMMHC probe and a rl Kit probe. The borders of the circular (bottom) and the future longitudinal muscle (top) layers are marked by lines. (A) SMMHC was strongly expressed in the circular muscle layer (arrow), while the future longitudinal muscle layer displayed lower levels of SMMHC expression (arrowhead). (B) Taking a picture of the same frame, but in a different optical plane, showed the silver grains indicative of Kit expression. Kit was uniformely expressed in both circular and future longitudinal muscle layers. (C) Single labelling experiment using a dig Kit probe. demonstrating Kit expression in almost all cells of the outer layers of the muscularis extema.

(D,E) Double labelling experiment, using a dig SMMHC probe and a rl Kit probe. (D) SMMHC was strongly expressed in the circular muscle (arrow), and, at lower levels, in the future longitudinal muscle layer (arrowhead). (E) Kit expression became restricted to the future longitudinal muscle layer, while some areas of the circular muscle retained Kit expression (arrow). (F) Section adjacent to the one shown in (D) and (E), hybridized to a rl SMMHC probe. Similar to (D), strong SMMHC expression was found in the circular muscle (arrow), while a uniform layer of silver grains, whose density was clearly above background, was observed in the future longitudinal muscle layer (arrowhead). Double labelling experiment on two different sections, (G,H)and (J,K), using a dig SMMHC and a rl Kit probe. Lines demarcate the inner and outer border of the muscularis externa. In each section, two different loops (labelled 1 and 2) in close proximity are shown (loop 1 as well as loop 2 in (G,H) and (J,K) are not identical). The location of the number indicates the Iocation of the circular muscle layer. In both (G) and (I), strong SMMHC expression was seen in the circular muscle, while lower expression was seen in the future longitudinal muscle layer of both loops. (H) and (K) show strong expression of Kit in the future longitudinal muscle layer of each loop 2, but decreased Kit expression in each loop 1. Hybridization of adjacent sections to a rl SMMHC probe [(I) for (G,H),(L) for (J,K)], which results in silver grains that are easier to quantitate then the digoxigenin color product, demonstrated a negative correlation between Kit and SMMHC expression in the future longitudinal muscle layer. Loop 1 in (I) and (L) showed relatively strong SMMHC expression (arrowheads). These tissues had low Kit expression [as shown in (H) and (K)].Loop 2 in (I) and (L) displayed low (L) to intermediate (I) SMMHC expression in the future longitudinal muscle layers (arrowhead). These tissues had relatively high levels of Kit [as shown in (H) and (K)].SMMHC was highly expressed in the circular muscle layers in both (I) and (L) (arrows). (M.N) Double labelling of dl8 gut. (M) SMMHC labels two layers of developing smooth muscle: the inner circular muscle layer (CM) and the hture longitudinal muscle layer 0. (N) Two Kit-positive cells are located between the two muscle layers (silver grains marked by arrowheads) and do not express SMMHC anymore. Conversely, the SMMHC-positive muscle cells do not express Kit anymore. Scale bar represents 30 p.m. Figure 2 Analysis of Kit expression in wild type, ~57and dJd adult small intestine

Structural mutations in W/Kit lead to an absence of both the network of KC and intestinal pacemaker activity (Ward et al., 1994; Huizinga et al., 1995). To investigate the effects of the two regulatory W mutations, ~57and wbd, on development of the ICC in adult small intestine, we performed whole mount RNA in situ experiments on adult small intestine using a digoxigenin-labelled Kit cRNA probe described previously (Huizinga et al., 1995). After the color reaction, we either separated the external muscle layers from the mucosa and submucosa for better visualization of the KC network or cross-sectioned the tissues. The network of Kit- positive KC in the separated muscle layers of wild type small intestine (Fig. 3A) are located between the external longitudinal and circular muscle layers (Fig. 3D), as shown previously (Thuneberg, 1982; Ward et al., 1994; Huizinga et al., 1995). The small intestine of ~571~57animals displayed an identical pattern of Kit-positive KC (Fig. 3B), located between the external muscle layers (Fig. 3E). In contrast, we did not observe any Kit-positive cells in the separated muscle layers in the small intestines (Fig. 3C) or in the cross-sectioned whole mount tissues (Fig. 3F) of ~bd/Wbd embryos, indicating that Kit expression is either abolished in the ICC or that the network of Kit-expressing ICC is absent in ~bd/Wbdanimals. Figure 3. (A-F)RNA whole mount in situ analysis of Kit expression in adult small intestine. Shown are isolated muscle layers (A,B,C) and sections of whole mount tissues @,E,F). (A) Wildtype small intestine shows a typical network of Kit-positive ICC. (D) The ICC are located between the external circular and longitudinal muscle layers (arrow). The borders of the two muscle layers are indicated by black lines on the left side. (B) ICC in small intestine from w-5'~~~mice express Kit and form a normal network. (E) ICC from W57 ~57mice are located between the two external muscle layers (arrow). (CnNo Kit expression could be detected in the small intestine from @d/wbd mice. The arrow shows the expected location of Kit-positive KC;the black lines on the left indicate the borders of the muscle layers. Scale bar represents LOO pm.

(GI) Supra-vital methylene blue staining of adult small intestine. (G) Wild type small intestine displays the normal network of KC, which was also present in the small intestine of

~57/~57animals (H).(I) The small intestine of wbd/w6d mice lacks a ICC network; some scattered KC are labelled. The picture in (I) shows a relatively densily populated area of WbdiWbd small intestine; other areas are almost completely devoid of KC. Scale bar represents 100 pm-

Effects of the ~57and WM mutations on ICC development and intestinal pacemaker activity

To address whether the ~57or ~bd mutations affect the development of the ICC in the small intestine, we stained the gastro-intestinal tract of 4 month old wild type and mutant mice with methylene blue, which, under certain conditions, labels specifically the ICC (Thuneberg, 1982; Mikkelsen et al., 1988). The small intestines of ~5~~57animals had an apparently normal network of ICC (Fig. 3H, compare to wild type, Fig. 3G), consistent with the normal pattern of Kit expression in these animals (see above). In contrast, the ICC network was absent from wbd/'Wb* mice, and only scattered methylene blue-positive cells were detectable (Fig. 31).

The network of KC is essential for intestinal pacemaker activity in the gut, based on the analysis of W structural mutations (Ward et al., 1994; Huizinga et al., 1995). To investigate the effects of the regulatory ~57and ~bd mutations on intestinal pacemaker activity, we measured electrical activity of the muscle layers of wild type, @7/+7and Wbd/Wbd small intestines. Whereas wild type and ~571~57 small intestines displayed the slow wave-type action potentials characteristic of pacemaker activity, the small intestine of ~bd,~bdanimals failed to display any slow wave-type action potentials (Fig. 4A). The slow waves exhibited a constant frequency of 35.0 + 1.6 cpm in control and 37.0 f 4.0 cpm in ~57~57mice. The resting membrane potentials were somewhat lower in mutant animals (-58.0 f 1.0 mV in ~571'7,-49.3 + 5.2 mV in ~bd/~bd)than in wild type mice (-64.0 + 4.8 mV;Fig. 4A). The muscle of ~bdmbdmice was either quiescent (1 out of 4) or showed fast spike-like action potentials at irregular frequency with an average value of 10.7 + 2.5 cpm (Fig. 4A), resembling the situation in WWmice (Ward et al., 1994; Huizinga et al., 1995). The slow-wave component or pacemaker activity of gut smooth muscle is insensitive to L-type calcium channel blockers (Huizinga et al., 1990). In the presence of the blocker D600, the slow-wave component of the action potentials in wild type and ~571~57mice remained unaltered, whereas the electrical activity of ~bd/CYbd mice was completely abolished (data not shown). These electrical differences in pacemaker activity in the W mutant mice were associated with differences in mechanical contraction patterns (Figure 4B). While isolated ileum from wild type and ~57,@7 animals displayed normal rhythmic, anally propagating contractions, the ilea of wbd/~bdmice failed to produce these rythrnic contractions and instead showed irregular, non-directional contractions.

Figure 4. (A) Electrical activity generated in wild type, ~57and Wbd homozygous mice. A slow wave-type action potential with a constant frequency was generated by wild type and ~57.~7,but not by ~bd~~bd,mice. Wbd/wbd mice generated irregular, fast spike-like action potentials (3 of 4 animals tested) or were completely quiescent (I out of 4). (B) Mechanical contraction patterns of the small intestine of wild type and mutant mice. While wild type and ~57~57mice exhibited a rythrmc and coordinated contraction pattern, w6d/wbd mice displayed an irregular, uncoordinated contraction pattern. Figure 4

A Electrical B Mechanical Kit is not required for embryonic development of ICC

Because mutations at the W locus affect the numbers of ICC in the adult, it was of interest to determine the step during ICC development that requires a functional Kit receptor. To determine whether Kit is necessary for the lineage determination of multipotential progenitor cells towards an KC fate during embryogenesis, we first looked for the network of methylene-blue-positive KC in the small intestines of wbd/Wbd mice after birth. Surprisingly, ICC were present in normal numbers at postnatal day 5 (p5) in these mutant animals (Fig. 5B, compared to wild type animals, Fig. SA), suggesting that the complete absence of Kit expression in the progenitor cells of the external muscle layers during embryogenesis did not interfere with lineage determination and ICC development up to p5. In contrast, by p15, ICC numbers were significantly reduced in ~bdhVbdanimals (Fig. SC), compared to wild type animals (Fig. 5D). Thus, it appears that Kit is not required for lineage determination of ICC during embryogenesis, but rather for postnatal steps in KCdevelopment. The absence of Kit expression leads to a marked reduction in the density of ICC, suggesting that Kit may be involved in some later developmental processes, such as proliferation or swival of these cells.

Figure 5. Supra-vital methyhe blue staining of wild type and wbdMdsmall intestine. At p5, both wild type (A) and WbdhVbd (B) animals displayed equivalent numbers of KC. At p 15, the density of ICC was significantly reduced in @d/wbdmice (D),compared to wild type mice (C). Scale bar represents 100 pm. Figure 5 The reduction in ICC density negatively correlates with the increase of the dacearea of the small intestine

We next asked if the reduction in density of ICC in W mutant mice could be explained by the dramatic increase in surface area of the small intestine that occurs in postnatal development. Indeed, the surface of the small intestine of both wild type and ~bd/~bdmice increased approximately 13-fold between p5 and 6 months of age (Table 1). During the same time period, the density of ICC in ~bdhornozygous animals decreased by approximately the same amount (Table 1); in contrast, the density of ICC in wild type animals decreased only slightly. Using the values in Table 1, we estimated the number of ICC per small intestine of wild type animals increased approximately 11-fold between p5 and 6 months. In contrast, ICC numbers in wbd/~ybdanimals appeared not to increase between p5 and p15 and increased less than 2-fold by 6 months of age. Therefore, in ~bdmutants, the number of ICC throughout the small intestine is not reduced, but appears not to increase as dramatically as in wild type animals.

Table 1. Calculations of surface area of the small intestine and numbers of ICC in wild type and wbdmbd mice at different stages of postnatal development. The approximately 13-fold increase in small intestine surface area between p5 and 6 months of age is matched by an approximately 11-fold increase in the numbers of KC in wild type mice. KC numbers in wad/Wbd mice do not increase significantly during the same time period, leading to an approximate 8-fold reduction in the density of ICC at the age of 6 months. The average surface area of the small intestines was measured from 2 mice for each age; numbers of KC and cell densities were calculated from 2 mice for each age and genotype.

ICC apoptosis in Wbd mice

The presence of wild-type levels of ICC in wbd/~bdnewborn mice, and the failure of these cells to increase in numbers after birth, raised the possibility that the Kit receptor is involved in preventing apoptosis in these cells. In fact, we have shown previously that signaling through the Kit RTK prevents p53-dependent programmed cell death in Friend erythroleukemia cells (Abrahamson et al., 1995). In order to analyze a possible similar role for Kit in ICC development, we asked whether the reduction in the density of KC was associated with increased levels of apoptosis in cells at the levels of the Auerbach's plexus during the first two weeks of life. At p4 and p8 (data not shown), p9 (Figs. 6A,B,C,D), p10 (data not shown) and p15 (Figs. 6E,F,G,H), the levels of apoptosis, as judged by a modified Tunnel assay, were low throughout the tissues examined. More importantly, no differences in the levels of apoptosis between wild type and ~bd~~bdmutant tissues were observed. The numbers of apoptotic nuclei per section of small intestine (approx. lcm distal to pyloric sphincter; a minimum of 5 sections per genotype and age were counted) were as follows: +/+ (p4): 1.6 + 1.8; bad(p4): 1.4 + 0.9; +/+ (p9): 3.3 f 1.8; bdmd (p9): 3.4 -L 6.0; +/+ (p15): 0.3 f 0.5; bdhd (p15): 0.5 + 0.5. Figure 6. Apoptosis assays of wild type and @d/w6d small intestine. Fragmented genomic DNA of apoptotic cells was labelled and subsequently detected using a fluorescein- conjugated antibody. Tissues were counterstained with a Hoechst nuclear stain. At both p9 (A,B,C,D) and p15 (E,F,G,H), rates of apoptosis were very low in both wild type (A,.B;E,F) and &d/w6d (C,D;G,H) small intestine. The arrows mark apoptotic cells located in the two external muscle layers. In E,F,G,H, the borders of the external muscle layers are demarkated.

DNase Etreatment of a control slide (I) led to positive staining in the majority of cells; in a negative control, the Terminal enzyme, used to label 3' DNA ends, was replaced with water (J), leading to a complete absence of signal. Scale bar represents 150p in A,B,C,D,I,J and 300p.m in E,.F,G,H.

DISCUSSION

The ICC located between the two external muscle layers of the vertebrate gut at the level of the Auerbach plexus were first described a century ago by Ramon y Cajal (Cajal, 1893; Cajal, 191 1). The developmental origin of this network of cells has been controversial since they were first described. Based largely on morphologic and ultrastructural evidence, they have been considered to be neuronal (Cajal, 1893), fibroblast-like, glial-like or undifferentiated muscle cells (reviewed in Thuneberg, 1982). Recently, Lecoin et al. (1996), using quail-chick chimeras, demonstrated that the ICC are not derived from the neural crest, the origin of the enteric nervous system. Lecoin et al. ( 1996) suggested that KC might be mesodermally-derived, originating from splancheopleural mesenchyme of the lateral plate along with smooth muscle and connective tissue cells.

In this study, we have taken advantage of recent findings by ourselves and others that the W/Kit locus is required for normal ICC development (Ward et al., 1994; Huizinga et al., 1995) to investigate the developmental origin and lineage relationship of the ICC during mouse embryogenesis. We showed that the number of Kit-positive cells in the muscularis externa becomes continuously restricted between d14.5 and dl8 of embryogenesis. Our experiments also suggested that ICC and smooth muscle cells develop from common mesodemally-derived mesenchymal progenitors, which co-express both Kit and SMMHC at mid-gestation. The three patterns of Kit expression observed around d14.5 indicate a dynamic developmental process of ICC and smooth muscle development. First, both the circular muscle and the future longitudinal muscle layer co-express Kit and SMMHC.Subsequently, Kit expression is eliminated in the circular muscle layer, but both markers are still co- expressed in the future longitudinal muscle layer. Thereafter, the divergence of Kit- positive KC and SMMHC-positive longitudinal smooth muscle cells appeared to be initiated, leading to a reduction in the number of Kit expressing cells and to an increase in SMMHC expression over the next 4 days of embryonic development. There appeared to be spatial differences in the progress through these stages, since we were able to observe all three stages at d14.5. Late in gestation, SMMHC expression is down-regulated in Kit-positive KC and conversely, Kit expression is down-regulated in developing smooth muscle, indicating that both lineages have diverged at that stage. These data provide positive evidence that the putative progenitor cell of ICC and longitudinal smooth muscle co-expresses both Kit and SMMHC,and, therefore, that KC at the level of the Auerbach's plexus and longitudinal smooth muscle cells derive from a common progenitor cell and might be related cell lineages. A previous study analyzing the expression of the Kit receptor and markers of the smooth muscle lineage (smooth muscle actin and myosin) was unable to show expression of smooth muscle actin or myosin in the future longitudinal muscle layer before dl8 and did not observe co-expression of Kit and these markers in the future longitudinal muscle layers during embryogenesis (Torihashi et al., 1997). Here, we have demonstrated SMMHC mRNA expression in the future longitudinal muscle layer at d14.5 and showed that Kit and SMMHC are co-expressed during a limited time during embryogenesis. The discrepancy between these studies might be due to the differences in expression assays used: in their study, Torihashi et al. analyzed protein expression while in the present study, we analyzed mRNA expression. Alternatively, it is possible that Torihashi et al. (1997) might have missed relatively narrow time window in which Kit and SMMHC are co-expressed. It is possible that rnRNA in situ analysis is a more sensitive assay than a protein expression assay, a possibility which might explain why Torihashi et al. (1997) did not observe the low to intermediate levels of SMMHC expression in the future longitudinal muscle layer. Torihashi et al. (1997) suggested that smooth muscle cells are derived from a few cells in the future longitudinal muscle layer which co-express both Kit and Desmin. These authors also suggested that ICC are derived from Kit-positive, Desmin-negative cells in that area. In contrast, our results indicate that the majority of cells in the future longitudinal muscle layer co-express Kit and SMMHC at d14.5 and, therefore, are bipotential progenitors, able to develop along the ICC or smooth muscle lineage. Both this study and Torihashi et al. (1997) show that KC lineage determination and divergence between KC and smooth muscle occurs between d14.5 (dl5 in the Torihashi study) and dl8.

The comparative analysis of the effects of two independent W regulatory alleles, ~57and ~bd, on the ICC lineage, has also provided further insights into the role of the Kit receptor in ICC development. Although the ~bdmutation abrogates the formation of the network of KC in adult homozygous animals, the ~57mutation had no effect on KC development. Accordingly, only mice homozygous for the ~bd mutation failed to exhibit the electrical and mechanical properties of intestinal pacemaking, whereas ~57mice were normal. Mice homozygous for another regulatory W mutation, ~sash(sh),which is phenotypically and genotypically very similar to the ~bdmutation (Duttlinger et ai., 1993; Kliippel et al., 1997),displayed identical defects in ICC development to ~bdmice (data not shown). These observations are consistent with the loss of Kit expression in the gut mesenchyme of wbd hornozygous embryos and normal Kit expression in the gut mesenchyme during embryogenesis in ~57homozygous embryos. These results demonstrate that the ICC defect in wbd/~bdanimals results from the dowe modulation of Kit expression in the mesenchymal progenitor cells and the developing ICC. The ~57mutation is the only W mutation known so far that does not affect ICC development in the homozygous state. Interestingly, the ICC in wbdhVbd mice are present in normal numbers at p5, demonstrating that the Kit receptor is not necessary for either lineage determination of the pluripotent progenitor cell towards an KC fate or for any subsequent steps of lineage differentiation or proliferation of these cells during embryogenesis. Therefore, activation of the Kit signalling pathway appears to play a permissive role, enabling the mesenchymal progenitor cells that are committed to the KC lineage to develop further along this pathway. This conclusion is supported by a recent report, in which the lac2 marker was inserted into the Kit locus by gene targeting, thus disrupting the coding sequence and creating a Kit null allele. Mice homozygous for this targeted Kit locus died shortly after birth, but displayed ZucZ-expressing cells in the outer layers of the small intestine at PO, suggesting that Kit is not essential for embryonic development of KC (Bemex et al., 1996).

Curiously, the reduction in density of ICC in the small intestine of ~bd/Wbd animals between p5 and 6 months is not associated with an actual decrease in overall numbers of ICC. The tissue surface increases about 13-fold during the fust 6 months; ICC of wild type animals expand accordingly, whereas the KC population in W~~IWManimals is not able to expand. Indeed, the reductions in KC density in mice homozygous for various W mutations that affect the Kit coding region (Ward et al., 1994; Huizinga et d., 1995; Torihashi et al., 1995) are also consistent with a stagnation in cell numbers. Without a significant increase in cell numbers, the ICC in wbdhVbd animals are unable to maintain a functional network. This functional ICC network, a requirement for pacemaker activity (this study; Ward et al., 1994; Huizinga et al., 1995), relies on cell-cell contact via long, cellular processes. Thus, if the ICC population cannot maintain a critical density during this increase in the surface area of the small intestine, cell-cell contacts appear to be disrupted due to the increased space between individual cells, leading to an absence of the synchronized electrical pacemaker activity. The post-natal plateau in KCnumbers in WbdhVbd animals appears to reflect a requirement for the Kit receptor during the post-natd expansion of the KC,rather than a role for Kit in preventing apoptosis. However, due to the slow postnatal expansion of ICC, it is possible that apoptosis events were rare at any given time point and could not be detected in our assays. A model of KC and smooth muscle development is depicted in Figure 7.

Figure 7. The development and lineage relationship of the interstitial cells of Cajal (KC) in the mouse small intestine. Kit+/SMMHC+ progenitor cells located in the outer layers of the developing gut at embryonic day 14.5 give rise to two cell lineages. which begin to diverge late in gestation: the Kit+ ICC and the SMMHC+ smooth muscle cells. Progenitor cells and the developing smooth muscle lineage do not require a functional Kit receptor; in addition, the lineage determination of ICC during embryogenesis as well as the formation of the cellular extensions to form the typical network postnatally are also Kit-independent. A lack of functional Kit receptor leads to postnatal defects in ICC development and most likely affects ICC proliferation after pS.

Previous studies have demonstrated that injection of the Kit-antagonizing Ack2 antibody at pO and p2, but not before or after, impairs KC development and intestinal pacemaker activity (Maeda et al., 1992; Torihashi et al., 1995). In contrast, our study indicates effects of the ~bdmutation on the density of KConly after p5. Therefore, there appears to be a delay of several days between the requirement for Kit and any observable phenotype. In this context, it is of interest to note that stimulation of Kit- expressing melanoblasts with Steel factor at early stages of culture induces subsequent responsiveness to the trophic effects of nerve growth factor (NGF) (Langstimm-Sedlak et al., 1996). A similar mechanism could account for the time discrepancy between requirement for Kit and the subsequent reduction in ICC- density. According to this model, Kit signaling between pO and p2 might induce responsiveness of ICC to proliferative signal(s) by other molecules at later stages in development; in the absence of Kit signaling, this responsiveness would not be induced and the KC would therefore be unable to respond. Alternatively, the differences between the antibody studies and our results might be due to technical differences. Indeed, effects of Ack2 treatment in adult mice can differ significantly from the effects of W alleles (Galli et al., 1994).

Our results indicate that signaling through the Kit receptor only becomes important after p5, when the ICC network has formed and is located in close proximity to myenteric neurons between the two muscle layers. These neurons express Steel factor (Torihashi et al., 1997) and thus the close proximity between these two cell types might be important for stimulation of ICC proliferation through functional interactions between the membrane-bound form of Steel factor and the Kit receptor. Indeed, S~Sldickie(d)and ~ld/~ldmutant mice, which lack the membrane- bound form of Steel factor, but still express the soluble form of the ligand, display an absence of the ICC network and intestinal pacemaker activity (Huizinga et al., 1995; Ward et al., 1995; M.K.,unpublished results).

The ~57and wbd mutations exert tissue- and cell type-specific effects on Kit expression and the development of several cell types, including mast cells and melanocytes (Kliippel et al., 1997). Both the ~57and the ~bdmutations are associated with large genomic rearrangements 5' of the Kit coding sequence. The ~5~ chromosome carries an 80 kb deletion 5' to the Kit gene, whereas ~bd,like wsh (Nagle et al., 1995; Kliippel et al., 1997), contains a 2.8 Mb inversion 5' of Kit . Both ~57and ~bd affect transcriptional initiation of Kit in a cell type-specific fashion (Kluppel et al., 1997). Because Kit is normally expressed in ICC and the wbd mutation affects Kit transcription in ICC and their mesenchymd progenitor cells, this mutation will be an important tool to study transcriptional regulation of Kit in the developing KC. The genomic sequences affected by the ~bdinversion might provide insights into the interaction of cis-acting elements and transcription factors which control KC development and establish an ICC-specific program of gene expression.

ACKNOWLEDGEMENTS We thank Dr. Lars Thuneberg for instructions with the methylene blue staining technique and isolation of the external muscles of the small intestine. We thank Ken Harpal for the sectioning of the whole mount tissues and Dr. Joseph Miano for the gift of the SMMHC plasmid. Work in the authors' laboratories is supported by grants from the Medical Research Council of Canada and grant #003313 from the National Cancer Institute of Canada (to A.B.). M.K.was supported by a long-term Government of Canada Award. J.H. is an MRC Scientist. Beechey, C. V., Loutit, J. F. and Searle, A. G. (1986). Banded-a new W-allele. Mouse News Letter 74,92.

Bernex, F., De Sepulveda, P., Kress, C., Elbaz, C., Delouis, C. and Panther, J.-J. (1996). Spatial and temporal patterns of c-kit-expressing cells in ~lacZ/+and wlacZ/wL~~Z mouse embryos. Development 122,3023-3033.

Cajal, S. R. (1893). Sur les ganglions et plexus nerveux de l'intestin. C.R. Soc. Biol. 45,217-223.

Cajal, S.R. (19 11). Histologie du systeme nerveux dell hornme et des vertebres. Maloine, Paris.

Chabot, B., Stephenson, D., Chapman, V., Besmer, P. and Bernstein, A. (1988). The proto-oncogene c-kit encoding a transmembrane tyrosine khase receptor maps to the mouse W locus. Nature 335,88.

Copeland, N.G., Gilbert, D.J., Cho, B.C., Donovan, P.J., Jenkins, N.A., Cosman, D., Anderson, D., Lyman, S.D. and Williams, D.E. (1990). Mast cell growth factor maps near the Steel locus on mouse chromosome 10 and is deleted in a number of Steel alleles. Cell 63, 175- 183.

Faussone Pellegrini, M. S. (1984). Morphogenesis of the special circular muscle layer and of the interstitial cells of Cajal related to the plexus muscularis profundus of mouse intestinal muscle coat. Anat. Embryol. 169, 15 1-158.

Faussone-Pellegrini, M.S . (198Sa). Ultrastructural peculiarities of the inner portion of the circular layer of the colon. II. Research on the mouse. Acta hat. 122, 187- 192.

Faussone Pellegrini, M. S. (198Sb). Cytodifferentiation of the interstitial cells of Caja1 related to the myenteric plexus of mouse intestinal muscle coat. An E.M.study from foetal to adult life. Anat. Ernbryol. 17 1, 163- 169. Geissler, E. N., McFarland, E. C. and Russell, E. S. (1981). Analysis of pleiotropisrn at the dominant white spotting (W) locus of the house mouse: A description of 10 new W alleles. Genetics 97,337-361.

Geissler, E., Ryan, M. and Housrnan, D. (1988). The dominant white spotting (W) locus of the mouse encodes the c-kit proto-oncogene. Cell 55, 185-194.

Hayes, M.A. and Goldenberg, I.S. (1957). The problems of infantile pyloric stenosis. Int. Abstr. Surg. 104, 105-138.

Huang, E., Nocka, K., Beier, D.R., Chu, T.-Y ., Buck, J., Lahm, H.-W., Wellner, D., Leder, P. and Besmer, P. (1990). The hematopoietic growth factor KL is encoded by the SZ locus and is the ligand for the c-kit receptor, the gene product of the W locus. Cell 63,225-233.

Huizinga, J. D., Faraway, L. and Den Hertog, A. (1990). 1. Physiol., Lond. 442, 15- 29.

Huizinga, J. D., Thuneberg, L., Kliippel, M., Malysz, J., Mikkelsen, H. B. and Bernstein, A. (1995). Wait gene required for interstitial cells of Cajal and for intestinal pacemaker activity. Nature 373, 347-349.

Keshet, E., S.D. Lyman, D.E. Williams, D.M. Anderson, N.A. Jenkins, N.G. Copeland and L.F. Parada. (1991). Embryonic RNA expression patterns of the c-kit receptor and its cognate ligand suggest multiple functional roles in mouse development. The EMBO Journal 10,24252435.

Kliippel, M., Nagle, D. L., Bucan, M. and Bemstein, A. (1997). Long-range genornic rearrangements upstream of Kit in the w~~ and wbandedmutants dysregulate the developmental pattern of Kit expression and affect distinct steps of melanocyte development. Development l24,65-77.

Langstimm-Sedak, C.J., Schroeder, B., Saskowski, J.L., Camahan, J.F. and Sieber- Blum, M. (1996). Multiple actions of stem cell factor in neural crest cell differentiation in vitro. Dev. Biol. 174,345-359. Lecoin, L., Gabella, G. and Le Douarin, N. (1996). Origin of the c-kit-positive interstitial cells in the avian bowel. Development 122,725-733.

Liu, L.W.C., Sperelakis, N. and Huizinga, J.D. Pacemaker activity and intercellular communication in the gastrointestinal musculature. In: Pacemaker activity and intercellular communication. AM Arbor Press, 1995, 159- 174. Huizinga, J.D., editor.

Maeda, H., Yamagata, A., Nishikawa, S., Yoshinaga, K., Kobayashi, S., Nishi, K. and Nishikawa, S. (1992). Requirement of c-kit for development of intestinal pacemaker system. Development 116,369-375.

Manova, K. and Bachvarova, R. F. (199 1). Expression of c-kit encoded at the W locus of mice in developing embryonic germ cells and presumptive melanoblasts. Developmental Biology 146, 3 12-324.

Matsui, Y., K.M. Zsebo and B.L.M. Hogan. (1990). Embryonic expression of a haematopoietic growth factor encoded by the Sl locus and the ligand for c-kit. Nature 347,667-669.

Miano, J.M., Cserjesi, P., Ligon, K.L., Periasarny, M. and Olson, E.N. (1994) Smooth muscle myosin heavy chain exclusively marks the smooth muscle lineage during mouse embryogenesis. Circ. Res. 75, 803-812.

Mikkelsen, H.B., Thuneberg, L. and Wittrup, LH. (1988). Selective double staining of interstitial cells of Cajal and macrophage-like cells in small intestine by an improved supravital methylene blue technique combined with FITC-dextran uptake. Anat. Embryol. 178, 191-195.

Motro, B., Van Der Kooy, D., Rossant, J., Reith, A. and Bernstein, A. (1991). Contiguous patterns of c-kit and steel expression: analysis of mutations at the W and S1 loci. Development 113, 1207-1221.

Nagle, D. L., Kozak, C. A., Mano, H., Chapman. V. M. and Bucan, M. (1995). Physical mapping of the Tec and Gabrbl loci reveals that the wsh mutation on mouse chromosome 5 is associated with an inversion. Hum. Mol. Gen. 4 (1 I), 2073-2079. Nishikawa, S., Kusakabe, M., Yoshinaga, K., Ogawa, M.. Hayashi, S., Kunisada, T., Era, T., Sakakura, T. and Nishikawa, S. (1991). In utero manipulation of coat color formation by a monoclonal anti-c-kit antibody: two distinct waves of c-kit dependency during melanocyte development EMBO J. 10,21 1 1-2 11 8.

Ogawa, M., Matsuzaki, S., Nishikawa, S. I., Hayashi, T., Kunisada, T., Sudo, T., Kina, H., Nakauchi, H. and Nishikawa, S. I. (199 1). Expression and function of c-kit in hematopoietic progenitor cells. J. Exp. Med. 174,63-71.

Orr-Urtreger, A., A. Avivi, Y. Zimmer, D. Givol, Y. Yarden and P. Lonai. (1990). Developmental expression of c-kit, a proto-oncogene encoded by the W locus. Development lO9,9 11-923.

Paulson, R. F. and Bernstein, A. (1995). Receptor tyrosine kinases and the regulation of hematopoiesis. Seminars in lmmmology 7, 267-277.

Shah, N.M., Groves, A.K. and Anderson, D,J. (1996). Alternative neural crest fates are instructively promoted by TGFbeta superfamily members. Cell 85, 33 1-343.

Silvers, W.K. (1979). The coat colors of mice. Heidelberg, Berlin, New York, Springer Verlag.

Thuneberg, L. (1982). Interstitial cell of Cajal: intestinal pacemaker cells? Adv. Anat. Embryol. Cell Biol. 71, 1-130.

Torihashi, S., Gerthoffer, W.T., Kobayashi, S. and Sanders, K.M. (1994). Identification and classification of interstitial cells in the canine proximal colon by ultrastructure and immunocytochernistry. Histochemistry 101, 169- 183.

Torihashi, S., Ward, S.M., Nishikawa, S.-I., Nishi, K., Kobayashi, S. and Sanders, K.M. (1995). c-kit-dependent development of interstitial cells and electrical activity in the murine gastrointestinal tract. Cell Tissue Res. 280,97-111.

Torihashi, S., Ward, S.M.and Sanders, K.M.(1997). Development of c-kit -positive cells and the onset of electrical rhythmicity in murine small intestine. Gastroenterology 112, 144- 155. Vandewinden, J.-M., Liu, H., De Laet, M.-H. and Vanderhaeghen, J.4. (1996). Study of the interstitial cells of Cajal in infantile hypertrophic pyloric stenosis. Gastroenterol. 111,279-288.

Young, H.M., Ciampoli, D., Southwell, B.R. and Newgreen, D.F. (1996). Origin of interstitial cells of Cajal in the mouse intestine. Developmental Biology 180.97- 107.

Ward, S., Burns, A., Torihashi, S. and Sanders, K. (1994). Mutation of the proto- oncogene c-kit blocks development of interstitial cells and electrical rythmicity in murine intestine. J. Physiol. 48O.l,9 1-97.

Ward, S.M., Bums, A.J., Torihashi, S., Harney, S.C. and Sanders, K.M. (1995). Impaired development of interstitial cells and intestinal electrical rythmicity in steel mutants. Am. J. Physiol. 269, C1577-C1585.

Williams, D.E., Eisenman, J., Baird, A., Rauch, C., Van Ness, K., March, C.J., Park, L.S., Martin, U., Mochizuki, D.J., Boswell, H.S., Burgess, G.S., Cosman, D. and Lyman, S.D. (1990). Identification of a ligand for the c-kit protooncogene. Cell 63, 167-174.

Zsebo, K.M., Williams, D.A., Geissler, E.N., Broudy, V.C., Martin, F.H., Atkins, H.L., Hsu, R.-Y.,Birkett, N.C., Okino, K.H., Murdock, D.C., Jacobsen, F.W., Langley, K.E.,Smith, K.A., Takeishi, T., Cattanach, B., Galli, S.J. and Suggs, S. (1990). Stem cell factor is encoded at the St locus of the mouse and is the ligand for the c-kit tyrosine kinase receptor. Cell 63, 2 13-224. CHAPTER 5

SUMMARY, DISCUSSION AND FUTURE DIRECTIONS Summary and Discussion

In this thesis, I have investigated two regulatory W mutations and the effect of dysregulation of Kit expression in these mutants on the development of several cell lineages. In addition, this thesis addressed the role for the Kit RTK in gut development and function. Work described in Chapter 2 has demonstrated large genomic rearrangements associated with the ~57and ~bd mutations and the effects of these rearrangements on cell type-specific transcription of the Kit gene. Analysis of these mutations has led to a better understanding of the role of the Kit RTK in rnelanocyte development. The experiments described in Chapters 3 and 4 have provided new insights into the role of the Kit RTK in the development of the interstitial cells of Cajal in the mouse small intestine. In addition, this work has provided definitive evidence that these cells are essential for normal intestinal pacemaker activity. Finally, the work in Chapter 4 has provided insights into the developmental origin of the KCand the role of the Kit RTK in their development.

The analysis of regulatory W mutations provides an experimental approach to investigate the complex temporal and spatial control of Kit expression. In vitro studies suggested several binding sites for transcription factors upstream of the f~stKit coding exon to be important for cell type-specific expression of the Kit gene (Yasuda et al., 1993; Yamamoto et al., 1993; Tsujimura et aI., 1996). I demonstrated in Chapter 2 that the ~57mutation is associated with an 80kb deletion upstream of the fust Kit coding exon, whereas the ~bdmutation is associated with a 2.8Mb inversion upstream of Kit, leading to distinct temporal and spatial changes in Kit expression patterns. Interestingly, neither mutation affects any of the cis-acting elements previously suggested to be important for Kit expression in vitro. In particular, the binding site for the Mi transcription factor, which has been suggested to be involved in the regulation of Kit expression in mast cells (Tsujimura et al., 1996), is not affected by the ~57and wbd genomic rearrangements, although both mutations affect Kit expression in mast cells. These results demonstrate that cis-acting regions, distinct from the cis-acting elements described previously and located further upstream, contribute signif~cantlyto the complex temporal and spatial pattern of Kit expression in vivo. Recently, an additional cis-regulatory element, located approximately 5kb upstream of the first Kit coding exon, has been described. In vitro studies showed that this site binds the Myb transcription factor and that Myb binding is required for the inhibition of Kit expression in several hematopoietic cell types that do not express Kit, indicating that Myb can act as a cell type specific silencer of Kit expression (Vandenbark et al., 1996). I have not examined whether this Myb binding site is affected by either the ~57or ~bd rearrangements. Nevertheless, analysis of Kit receptor surface expression in bone marrow cells derived from ~57and ~bd mice by FACS indicate that, in both mutants, there is a reduction in the number of Kit-positive bone marrow cells (M.K., unpublished results), rather than an increase in the numbers of Kit-positive cells, as one might predict if a regulatory silencer element was affected. These data suggest that the Myb binding site is not affected by the ~57and ~bdrearrangements. Both the ~57and ~bdmutations involve large rearrangements, raising the possibility that these mutations might also affect the regulation or structure of genes other than Kit. Thus, the phenotype of these mutants might, in part, be due to the effects of these large genomic rearrangements on these other putative genes. The ~bd mutation is associated with a 2.8Mb inversion, which disrupts regulatory elements for Kit, but might also disrupt regulatory elements for other genes. My analysis of the expression profile of the Tec gene (data not shown), which is located close to the proximal breakpoint of the ~bdinversion, indicates that the temporal and spatial expression of this gene is not disturbed by the ~bdmutation during embryogenesis. In adults, however, the wbd mutation downregulates Tec mRNA levels in mast cells. Because the absence of mature mast cells in ~bd/~bdmice can entirely be explained by the lack of detectable levels of Kit expression, it is not clear whether this reduction in Tee expression has any effect on mast cell development in @d mice. In a recent collaboration, we showed that Tee expression is downregulated by 50% in melanocytes derived from rs/rs mice (Bennett et al., manuscript submitted; data not shown), indicating that Tec might play a role in melanocyte development. The early and severe melanocyte defect in ~bd/~bdmice cannot be explained by a lack of Kit expression in wbd/Wbd melanoblasts, since they express approximately normal levels of Kit before they disappear. It is conceivable that disturbed Tec expression in early melanocytes might have a severe effect on melanocyte development, thus leading to the pigmentation defects observed in ~bdmice. My analysis of the developmental expression profile of Tee (Kliippel et al., manuscript in preparation; data not shown) indicates that Tec is not expressed at detectable levels in melanoblasts of dl 1.5 wild type or ~bd.bdembryos. Yet, at dl 1.5, all but a few melanoblasts have already disappeared from w~~/wMembryos, suggesting that any effect of the ~bdinversion on Tec expression does not contribute to the early melanoblast defect observed in ~bd/~bdembryos. This does not exclude a role for Tec at later stages of melanocyte development or in the rs mutation, since Tec is expressed in hair follicles at dl8 of embryogenesis and in mature melanocytes in adult animals. The Pdgfra and GABA receptor beta 1 (Gabrbl) genes are also located in close proximity to the breakpoints of the ~bdinversion. PdgFa expression is not disturbed in adult brain and lung of wbd/Wbd animals (M.K.,unpublished result), but a more in-depth analysis of adult and embryonic Pdgfra and Gabrbl expression in wbd/Wbd mice would be neccessary to clarify if the wbd mutation affects the expression of these two genes.

I showed in Chapter 2 that the ~57mutation leads to a temporal downregulation of Kit expression in neural crest-derived melanoblasts in the trunk, but not in neural crest-derived melanoblasts around the otic vesicle at dl 1.5 of embryonic development. Thus, it appears that the ~57deletion affects cis-regulatory elements necessary for Kit expression in a subset of neural crest-derived melanoblasts at this specific stage in development. Despite the lack of Kit expression at that stage, melanoblasts in ~57~57embryos appear to migrate normally from the MSA towards the dematorne of the somites (Figure 1). There are several models that could explain this observation: First, a residual level of Kit expression, not neccessarily detectable by RNA in situ techniques, could contribute to the survival of melanoblasts. Second, I did not determine the exact time window when Kit expression is downregulated and when it comes on again. A short period without Kit expression might enable melanoblasts to survive, perhaps by utilizing previously synthesized Kit receptor molecules. My results showed that ~571~57melanoblasts at d14.5 express approximately normal levels of Kit, and it is conceivable that Kit expression comes on earlier than that. Third, melanoblasts undergo several alternating stages of Kit- dependency and Kit-independency between d9.5 and birth (Nishikawa et al., 1991; Yoshida et al., 1996). The temporal downregulation of Kit expression in trunk melanoblasts of ~57hV57ernbryosmight at least partially overlap with a phase of Kit-independence, therefore causing only a partial depigmentation. Nevertheless, although ~57~57melanoblasts migrate normally during this period of severely reduced Kit expression, their numbers do not increase to the same degree melanoblast numbers increase in wild type embryos. This might indicate that different levels of functional Kit receptor are required for different cellular responses. In this model, reduced levels of functional Kit receptor at dl 1.5 might enable melanoblasts to respond to migratory clues and/or to survive, but the number of functional Kit receptor molecules might be to low to promote normal proliferation. The effect of the ~57mutation on melanoblast proliferation or survival is not only observed during the period of downregulation of Kit expression, but also at later stages of embryonic development. Thus, some homozygous ~57embryos have only slightly reduced numbers of melanoblasts at d14.5, whereas others display a severe reduction at this stage. Nevertheless, at d 18 all ~57~57embryos exhibit a severe reduction of melanocytes in the mid-trunk region, the future location of the mottled depigmentation. It is possible that ~57~57melanoblasts at d14.5 and dl8 express Kit at somewhat reduced levels, which might not have been detected in these assays. Another possibility is that the temporal downregulation of Kit expression around dl 1.5 leads to a permanent proliferative disadvantage of these melanoblasts throughout embryonic development. This latter conclusion is supported by a recent publication which showed that survival of early melanoblasts in vitro is supported by Steel factor alone, but more mature melanoblasts require Steel factor plus any of the neurotrophins (Langstimrn-Sedlak et al., 1996). For formation of pigmented colonies, Steel factor was only required for the fxst haif of the culture period, whereas nerve growth factor (NGF) was only required for the second half of the culture period. Without prior exposure to Steel factor, melanoblasts did not respond to NGF. These observations are consistent with a model in which the exposure of melanoblasts in vitro to Steel factor results in a dependency on neurotrophins at later stages in development. In ~57mice, therefore, the lack of Kit expression at a stage where it might be necessary to induce neurotrophin dependency could result in a continuous decreased responsiveness of melanoblasts to neurotrophin signals and therefore to reduced proliferation or survival throughout the second half of embryonic development. wbdhVbd melanoblasts disappear shortly after they migrate out of the dorsal aspects of the neural tube and reach the migration staging area (MSA), located on the dorsolaterd migration pathway between neural tube and somite (Figure 1). This defect cannot be attributed to a lack of Kit expression, as wbd/'Wbd melanoblasts in the MSA express approximately normal levels of Kit. Interestingly, the melanoblasts disappear at the same time that Kit is expressed ectopically in the dermatome of the somites. This scenario is strikingly similar to previous observations in wsh/~shas well as Phi+ embryos (Duttlinger et al., 1993; Duttlinger et al., 1995; Wehrle-Haller et al., 1996). Interactions between Kit-expressing trunk melanoblasts and Steel factor- expressing dermatome have been suggested to play an important role in early melanoblast migration from the MSA towards the dermatome (Matsui et al., 1990; Manova and Bachvarova, 1991; Duttlinger et al., 1993). In a model previously suggested for the WS~and Ph mutations (Duttlinger et al., 1993; Duttlinger et al., 1995; Wehrle-Haller et al., 1996), the ectopic expression of Kit in the dermatome might sequester functional soluble Steel factor, thereby reducing the amount of of ligand available to the migrating melanoblasts and affecting their survival or migration by a non-cell autonomous mechanism. Interestingly, stimulation of human fetal melanoblasts in vitro with Steel factor affects the expression of several integrin subunits and changes the attachment of those cells to extracellular matrix (ECM) molecules (Scott et al., 1994). Thus, the melanoblast defect in ~bd,WS~ and Ph embryos might be due to insufficient levels of Steel factor, leading to a failure in changing integrin subunit expression and therefore to an inability of these cells to recognize and bind to the ECM. In this scenario, melanoblast migration would be stalled at the MS A, as we observed in ~bd/~bdmutant embryos. late station

migrating melanobIasts Kitdependent ditferentiating melanocyte Kit-independent

Figure 1. The @d mutation affects early melanocyte survival, whereas the ~57mutation appears to affect the proliferation or survival of melanoblasts throughout embryonic development. Interestingly, the wbd mutation in the heterozygous state also affects early melanoblast migration. Melanoblasts normally migrate along the dorsolateral pathway over the whole surface of the dematome (Bromer-Fraser, 1986). In wbd/+ embryos, however, the melanoblasts are located only over one half of the dematome, but not the other, reminiscent of the migration pathways of other neural crest populations which migrate only through the anterior, but not the posterior half of the somites. If the wbd mutation indeed affects integrin expression and attachment of melanoblasts to certain ECM molecules, as discussed above, the wbd/+ melanoblasts might become sensitive to anterior-posterior differences in ECM molecule expression in the somites, which have been shown to exist and have been suggested to play a major role in the choice of migration pathways of neural crest cells (Krull et al., 1995).

The ~bdand WS~ mutations are phenotypically as well as genotypically strikingly similar. Both mutations lead to the same pattern of dysregulation of Kit expression, both mutations lead to identical pigmentation defects, mast cell deficiency, and absence of ICC in the gut, and both mutations appear to be associated with very similar genomic inversions. Nevertheless, there are some data which indicate that ~bdand WS~might not be identical. First, compound heterozygous animals exhibit increased pigmentation in the head region compared to either mutation in the hornozygous state (Beechey et al., 1986). Second, erythroid cells from ~bd/Wbdanimals are more sensitive to ionizing radiation than erythroid cells from wshAVsh animals (Beechey et al., 1986). Third, I observed a prolonged ectopic expression of Kit in the floor plate of wbd/wbd embryos at midgestation compared to that previously described in ~sh/wshembryos. Fourth, Kit was ectopically expressed observed in the embryonic heart of wbd/Wbd embryos at d14.5 and dl6 of embryogenesis, which has not been reported for wsh/wsh embryos. It will be necessary to clone and sequence the inversion breakpoints in both ~bdand WS~ mutants to establish if these mutations are identical or distinct.

In chapter 3, we showed that the Kit RTK is expressed in the interstitial cells of Cajal (KC) in the myenteric plexus area of the intestine. In WWanimals, which carry germ line mutations in the Kit gene, this cell type is absent in the small intestine. This absence of ICC leads to an absence of intestinal pacemaker activity, which is normally characterized by a slow wave-type action potential in the muscularis externa. These data demonstrate a direct requirement for the Kit RTK in KC development and the essential role of ICC in gut pacemaker activity. Thus, these experiments extend the range of cell types affected by W mutations to include the KC. Interestingly, functional gut abnormalities and megacolon are also observed in individuals with piebaldism, suggesting an identical function of the Kit RTK in the development of ICC in humans (Bolognia and Pawalek, 1988). In addition, patients with infantile hypertrophic pyloric stenosis (IHPS), which is characterized by a motility disorder of the pyloric sphincter of the stomach, display a narrowing of the pyloric part of the stomach and hypertrophism of the circular smooth muscle layer (Hayes and Goldenberg, 1957). Infants with this condition cannot transport food appropriately through the stomach and exhibit typical projectile vomitting. This condition is usually treated surgically by a rnyotorny of the pyloric musculature. It has been postulated that IHPS may be caused by a defect of coordination between the movements of the pyloric sphincter and the contractions of the stomach. Impaired enteric innervation of the pyloric stomach musculature has been described in IHPS patients, indicating that changes in the enteric nervous system might account for the functional defects in MPS (Okazaki et al., 1994). Interestingly, a recent study showed that ICC are absent in the pyloric stomach of IHPS patients, suggesting that the lack of KC might contribute to the motility disturbance in DIPS (Vanderwinden et al., 1996). Recently, a functional role for another subtype of ICC, located within the circular and longitudinal muscle layers of the stomach (intramuscular ICC or ICC- IM), has been suggested (Burns et al., 1996). KC-IM are intercalated between the terminals of autonomous nerves and smooth muscle cells and had been suggested to play a role in neurotransmission (Thuneberg, 1982). These cells also express Kit, and, like the ICC in the myenteric plexus area, are affected by mutations in the W locus. WWmice, which lack KC-IM in the stomach, dispiayed greatly reduced NO- dependent inhibitory neurotransmission, suggesting that the cellular mechanism(s) responsible for transducing NO into electrical responses are located in the ICC-IM (Bums et al., 1996). In addition to Kit, mutations in another RTK, the RET proto-oncogene, are also associated with gut abnormalities like megacolon in humans (in 25% of cases of Hirschsprung's disease) and in gene-targeted mutant mice, as a result of of a cellular deficit in neural crest-derived enteric neurons (Romeo et al., 1994; Edery et al., 1994; Schuchardt et al., 1994). Mutations in another membrane receptor, the G-protein- coupled Endothelin B receptor (ETB-R),also lead to cellular deficits in enteric neurons in mice and human (in a subset of cases of Hirschsprung's disease) (Hosoda et al., 1994). Thus, several transmembrane receptors, Kit in ICC and Ret and ETB-R in neural crest-derived enteric neurons, are essential for normal gut function in mammals.

In chapter 4, I analyzed the Lineage relationships and development of the KC and investigated the effects of the regulatory ~57and ~bd mutations on ICC development. The experiments in that chapter suggest that Kit-positive ICC and Kit- negative smooth muscle cells in the gut develop from a common mesenchyrnal precursor, which expresses both Kit and the smooth muscle marker SMMHC at d14.5 of embryonic development. Between d14.5 and d18, the ICC and smooth muscle lineage appear to diverge, giving rise to SMMHC-positivelKit-negative smooth muscle cells and SMMHC-negativelKit-positiveICC. Whereas the ~57mutation does not affect Kit expression in the gut or KC development, the ~bdmutation leads to a complete absence of Kit-expression in the outer layers of the gut in homozygous embryos and to an absence of the ICC network and intestinal pacemaker activity in adults. To determine if Kit is playing an instructive role in lineage determination during embryogenesis or a permissive role supporting aspects of ICC development after the smooth muscle and ICC lineages diverged, I analyzed the presence of ICC after birth in mice carrying the regulatory ~bd/~~bdmutation. KC were present at postnatal day 5 (p5), indicating that ICC development can proceed through the lineage determination steps and early postnatal steps without functional Kit signalling. After p5, KC in wbd/~bdanimals did not expand like wild type KC to maintain a critical density during the growth of the small intestine. Indeed, ICC numbers in wbdhVbd animals stagnated, with approximately equivalent numbers present at p5 and at 6 months of age. There is both in vitro and in vivo evidence to suggest that Steel factor can prevent apoptosis in cells expressing the Kit receptor (Abraharnson et al., 1995). It is possible, therefore, that the absence of ICC, but not smooth muscle cells, in ~bdhVbd and WWmice results from the selective death of ICC during development. My results indicated that this is not the case, suggesting that Kit signaling mainly functions to control KC proliferation after birth. Given the slow rate of expansion of ICC during postnatal development (two-fold increase in numbers between p5 and PIS), it is possible that apoptosis events were not detected. Several recently published studies support some of the conclusions reached in chapter 4. Chick-quail transplantation studies showed that Kit-positive ICC are not derived from the vagd neural crest (Lecoin et al., 1996), which represents the origin of the neural crest cells forming the enteric nervous system. Similarily, transplantation studies in mice have demonstrated that removal and transplantation of pieces of the gut to an ectopic site before enteric neural crest cells have reached the intestine does not affect the development of ICC, indicating that the developmental origins of enteric nervous sysytem and KC are not identical (Young et al., 1996). Based on these very different experimental approaches both groups have argued that the origin of the ICC is not the neural crest and that the ICC might be derived from the gut mesenchyme. Other studies utilized the antagonizing Kit Ack-2 antibody to study the requirement for Kit in KCdevelopment. These studies showed that injection of Ack- 2 at the day of birth and two days later, but not before or after this narrow time window, led to a reduction in KC numbers and a disturbance of gut pacemaker activity by p10 (Maeda et al., 1992; Torihashi et al., 1995). Similarily, introducing a IacZ reporter gene in the Kit locus and thus creating a Kit-null mutation, has led to homozygous mice that die shortly after birth. Nevertheless, lad-positive ICC are present in these animals at pO, indicating that a functional Kit receptor is not required for embryonic development of the KC (Bernex et al., 1996). Figure 2 combines my results from chapter 4 and the results from other studies to depict a model of how ICC development procedes through stages of Kit- dependency and -independency in embryos and adults. Figure 2. Development of the ICC. At embryonic day 14.5, progenitor cells in the developing muscularis externa express both the ICC marker Kit and the smooth muscle marker SMMHC.At late gestation, the ICC and smooth muscle lineage begin to diverge, resulting in SMMHC-positivelKit-negative smooth muscle cells and Kit-positivelSMMHC- negative KC. The ICC, located between the two external muscle layers, start to form a typical network by postnatal day 4. Injection of antagonizing ACK2 antibody at pO and p2, but not before or after, disturbs the development of ICC and gut motility. In W mutant mice, the ICC network appears normal at day 5 after birth, but thereafter, the cells appear to be impaired in their proliferative capacity.

Future work

The ~57and ~bd mutations represent unique in vivo systems to identify cis- regulatory elements that control the developmental patterns of Kit expression. Mapping of the breakpoints of both mutations and comparison of the locations of the rearrangements would provide insights into the location of important positive regulatory elements required for Kit expression in mast cells, melanocytes and KC and negative regulatory elements for silencing Kit expression in heart and dermatome of the somites. Mapping of the breakpoints in the wbd and the WS~mutations would also determine whether these two alleles are identical. Other approaches to locate cis- regulatory elements would be to identify Kit regulatory DNA regions that bind proteins. Several methods which rely on the protection of genomic DNA from either enzymatic digestion or chemical modification by regulatory proteins bound to their cognate cis-elements (e.g. DNaseI footprinting, genomic footprinting or DNaseI hypersensitive site-mapping) could be employed to address this question. Comparing wild type and mutant DNA sequences by this method might give insights into which regulatory elements are utilized in wild type, but not mutant, DNA and therefore help in idenwing the DNA regions that contain the breakpoints.

Ultimately, it will be necessary to clone and sequence the DNA regions affected by the ~57and ~bd mutations. One approach to do so is a genomic walk, using contiguous genomic phage clones of the region in question, to identify the breakpoints of the rearrangements. The genomic region flanked by the breakpoints can then be functionally analyzed by a deletion analysis in transgenic animals. Alternatively, the genomic regions identified could be analyzed by sequence comparison to the homoIogous DNA regions in other species. It has been shown that the sequences of important regulatory regions can be conserved between different species (Koop, 1995). Thus, sequence comparison between different species might idenm several regions of homology, which then could be functionally analyzed in reporter gene-assays in tissue culture systems or in transgenic animals. Both experimental approaches would assist in narrowing down the regions important for binding of tissue- and cell type-specific transcription factors controlling Kit expression. The transgenic approach is often problematic because of position effects due to random integration of the reporter constructs into the genome. This can cause striking differences in transgene expression patterns and can seriously hinder any functional analysis of the regulatory elements in question (De Sepulveda et al., 1995). An alternative approach is the use of homologous recombination in embryonic stem (ES) cells to alter endogenous sequences in the Kit upstream region. This approach eliminates the problem of position effects, since only the endogenous locus is targeted, but requires more effort to carry out a complete analysis of cis-acting sequences. After the identification of smaller genomic regions which are capable of driving either the expression of a reporter gene or the endogenous gene in a subset of Kit-expressing cells, one could sequence these regions and search for interspecies conservation, therefore assisting in narrowing down the genomic regions containing important regulatory elements. Once regulatory elements have been identified, the importance of these elements can be assessed by creating point mutations and analyze their effect on reporter gene or endogenous Kit expression. Importantly, the characterized regions can be used in binding assays to identify transcription factors which control Kit expression in various cell types. This could lead to the identification of transcription factors that are important for proliferation, lineage determination, differentiation as well as stem cell maintanance in hematopoietic, melanocyte-, germ cell- and ICC lineages.

Signaling through the Kit receptor has been implied in regulation of integrin expression, which in turn is important for cellular interactions with the extracellular matrix (ECM) and cell migration (Scott et al., 1994). I showed in chapter 2 that the migration pattern of melanoblasts is altered in wbdl+ embryos. Therefore, it would be interesting to analyze the expression patterns of in these embryos to see if the altered melanoblast migration is associated with a dysregulation of integrin expression, providing an explanation of how melanoblast migration is disturbed in ~bd/+embryos. To my knowledge, no other study has analyzed the migration pathways of melanoblasts in heterozygous W mutants. Therefore, it is possible that the altered melanoblast migration is a feature of most of those W mutations in which melanoblasts survive beyond the migration towards the derrnatome.

Several coat color mutations that map in the Kit-Pdgfa intergenic region lead to a dysregulation of Kit expression in tissues or cell types that are involved in melanocyte development (Figure 3). Examples are the ~57and the ~bd rearrangements as well as the wsh and Ph mutations (Duttlinger et al., 1993; Duttlinger et al., 1995; Wehrle-Haller et al., 1996). The Ph mutation is a deletion that removes the coding sequence of the PdgfTa gene (Smith et d., 1991; Stephenson et al., 199 1), which lies approximately 400 kb proximal to Kit (Brunkow et al., 1995). The exact breakpoints of the Ph deletion have not been determined yet. Nevertheless, it has been shown that the Ph mutation in the heterozygous state leads to ectopic expression of Kit in the dermatome of the somites (Duttlinger et al., 1995; Wehrle-

Haller et al., 1996), similar to the effects of wbd and WS~Therefore, while the embryonic lethality in homozygous Ph mice might be due to the lack of Pdgfra, the pigmentation defect observed in Ph/+ mice might not be due to a 50% reduction in Pdgfra levels, but rather might be caused by a deletion of cis-regulatory elements important for Kit expression. There are several other coat color mutations that also map to this region (Silvers, 1979), but the molecular bases of the pigmentation defect is unknown (Figure 3). The rump-white (Rw) mutation is associated with an inversion of a large part of chromosome 5, and the distal breakpoint maps to the Kit-Pdgfra intergenic region (Stephenson et al., 1994). Yet it is not clear if Rw affects Kit expression and therefore if the pigmentation defect in these animals is due to changes in Kit expression. An analysis of Kit expression in viable Rw/+ embryos throughout development would be necessary to address this point Similarily, the recessive spotting (rs) mutation maps very close to the W locus, but no molecular lesion has been identified yet. In a collaboration, we showed that expression of the Tec gene appears to be downregulated in rs/rs melanocytes (Bennett et al., manuscript submitted), but further analysis is required to idenhfy the molecular lesion leading to the rs phenotype. My analysis of embryonic Tec expression suggests that Tec is not involved in early melanoblast development, but might play a role in later steps of melanocyte differentiation or survival.

Figure 3. Genornic rearrangements associated with coat color mutations in the vicinity of the W locus on mouse chromosome 5.

Kit is required for the postnatal development of the KC in the small intestine. Since my data indicate a role for Kit in ICC proliferation, a comparison of BrdU incorporation into KC, which are identifiable using a Kit antibody, could give insights into whether ICC in WWand wbd~bdanimals have decreased rates of proliferation. Another area of interest is the downstream targets of Kit that are important for ICC development. Several cytoplasmic signalling molecules (e.g. SHP-1, vav, P13'K. PLCy, Tec) are activated upon Steel factor binding to the Kit receptor and are thought to mediate specific cellular responses (Rottapel et al., 1991; Yi and Ihle, 1993; Tang et al., 1994; Serve et al., 1994). Many of the known molecules which bind to Kit are not expressed in the developing muscle layers of the intestine (e.g. Tec, SHP-I), and therefore are not likely to mediate Kit signalling in ICC (Paulson et al., 1996; M.K., unpublished results). The SHP-I phosphatase, a negative regulator of Kit in hematopoietic cells, is encoded by the motheaten (me) locus (Shultz et al., 1993; Tsui et al., 1993). Animals carrying homozygous mutations in both the W and the motheaten locus still lack KC (M.K.,unpublished results), indicating that SHP-1 is not a negative regulator of Kit signaling in ICC. Many cytoplasmic molecules bind to the phosphorylated tyrosine residues of the intracellular domains of activated RTKs via their SH2 domains (Marengere and Pawson, 1994). Thus, an alternative approach to elucidate the ICC-specific signal transduction pathway downstream of Kit is to introduce point mutations in these intracellular binding sites of the Kit receptor by homologous recombination, therefore eliminating the binding of downstream molecules to specific phosphotyrosines. Some point mutations might result in an ICC defect, therefore denimstrating the importance of a particular binding site for binding of molecules important for mediation of ICC- specific functions. Other molecules that are not part of the Kit signalling cascade might also play a crucial role in ICC development. Although W and SI mutations are the only mutations to date that impair ICC development, future analysis of loss-of-function mutations in other molecules will certainly yield more information about functional components of KC development. In addition, differential cDNA screens of ICC vs. smooth muscle or embryonic ICC vs. adult ICC could result in the isolation of genes that are specifically expressed in ICC or in certain stages of ICC development. The development of differential cDNA screens of single cells by PCR could provide a powerful method to identify genes that are differentially expressed in ICC at different developmental stages. The functional relevance of these genes in ICC development can then be assessed by creating null mutations in mice by homologous recombination.

In conclusion, in this thesis I have demonstrated the Kit-dependency of the KC and identified the ICC as an important component of the intestinal pacemaker activity in the mouse small intestine. I analyzed the developmental origin of the ICC and provide evidence for a mesodermal origin of these cells. Moreover, my data suggest a role for Kit in postnatal proliferation of KC. In addition, I have characterized the molecular basis of two regulatory W mutations and described their specific effects on melanoblast development and the developmental expression profile of Kit. These results have provided novel information on the multiple roles of the Kit RTK in mouse development, as well as new insights into the regulation of expression of this gene. REFERENCES

Abrahamson, J.L., Lee, J.M. and Bernstein, A. (1995). Regulation of p53-mediated apoptosis and cell cycle arrest by steel factor. Mol Cell. Biol. 15,6953-6960.

Beechey, C.V., Loutit, J.F. and Searle, A. G. (1986b). Banded, a new W allele. Mouse News L 74,92.

Bennett, D.C., Trayner, I.D., Piao, X., Easty, D.J., Kluppel, M., Alexander, W., Wagner, E.F. and Bernstein, A. recessive spotting: a linked l~custhat interacts with W/Kil but is not allelic. Manuscript submitted.

Bernex, F., De Sepulveda, P., Kress, C., Elbaz, C., Delouis, C. and Panthier, J.-J. (1996). Spatial and temporal patterns of c-kit-expressing cells in ~lacZ/+and ~Iacz~IacZmouse embryos. Development 122,3023-3033.

Bolognia, J.L. and Pawalek, J.M. (1988). Biology of hypopigmentation. J. Am. Acad. Dennatoi. 19,217-255.

Bromer-Fraser, M. (1986). Analysis of the early stages of trunk neural crestcell migration in avian embryos using monoclonal antibody HNK-1. Dev. Bid. 115, 44- 53.

Bmnkow, M.E., Nagle, D.L., Bernstein, A. and Bucan, M. (1995). A 1.8-Mb YAC contig spanning three members of the receptor tyrosine kinase gene family (Pdgfm, Kit, and Flkl) on mouse chromosome 5. Genomics 25,421-432.

Bums, A.J., Lomax, A.E.J., Torihashi, S., Sanders, K. and Ward, S.M. (1996). Interstitial cells of Cajal mediate inhibitory neurotransmission in the stomach. Proc. Natl. Acad. Sci. USA 93, 12008- 12013.

De Sepulveda, P., Salaiin, P., Maas, N., Andre, C. and Panthier, J.-J. (1995). SARs do not impair position-dependent expression of a Kit/LacZ transgene. Bioch. Bioph. Res. Corn. 21,1,735-741. Duttlinger, R., Manova, K., Chu, T. Y., Gyssler, C. Zelentz, A. D., Bachvarova R. F. and Besmer, P. (1993). W-sash affects positive and negative elements controlling c-kit expression: ectopic c-kit expression at sites of kit-ligand expression affects melanogenesis. Development 118,705-7 17.

Duttlinger, R., Manova, K., Berrozpe, G., Chu, T. Y., DeLeon, V., Tirnokhina, I., Chaganti, R. S. K., Zelenetz, A. D., Bachvarova, R. F. and Besmer, P. (1995). The WS~and Ph mutations affect the c-kit ex ression profile: c-kit misexpression in embryogenesis impairs melanogenesis in FfK and Ph mice. Proc. Marl. Acad. Sci. 92, 3754-3758.

Edery, l?, Lyowet, S., Mulligan, L.M., Pelet, A., Dow, E., Abel, L., Holder, S., Nihoul-Fekete, C., Ponder, B.A. and Mu~ich,A. (1994). Mutations of the RET protooncogene in Hirschspnmg 's disease. Nature 367,320-32 1.

Hayes, M.A. and Goldenberg, I.S. (1957). The problems of infantile pyloric stenosis. Int. Abstr. Surg. 104, 105-138.

Hosoda, K., Hammer, R.E., Richardson, J.A., Baynash, A.G., Cheung, LC., Giaid, A. and Yanagisawa, M. (1994). Targeted and natural (piebald-lethal) mutations of endotheh-B receptor gene produce megacolon associated with spotted coat color in mice. Cell 79, 1267- 1276.

Koop, B.F. (1995). Human and rodent DNA sequence comparisons: a mosaic model of genomic evolution. Trends Genet. 11,367-371.

Krull, C.E., Collazo, A., Fraser, S.E. and Bromer-Fraser, M. (1995). Segmental migration of trunk neural crest cells: time-lapse analysis reveals a role for PNA- binding molecules. Development 121,3733-3743.

Langstimrn-Sedlak, C.J., Schroeder, B., Saskowski, J.L., Camahan, J.F. and Sieber- Blum, M. (1996). Multiple actions of stem cell factor in neural crest cell differentiation in vitro. Dev. Biol. 174,345-359.

Lecoin, L., Gabella, G. and Le Douarin, N. (1996). Origin of the c-kit-positive interstitial cells in the avian bowel. Development 122,725-733. Maeda, H., Yamagata, A., Nishikawa, S., Yoshinaga, K., Kobayashi, S., Nishi, IS. and Nishikawa, S. (1992). Requirement of c-kit for development of intestinal pacemaker system. Development 116,369-375.

Manova, K. and Bachvarova, R. F. (1991). Expression of c-kit encoded at the W locus of mice in developing embryonic gem cells and presumptive melanoblasts. Dev. Biol. 146,3 12-324.

Matsui, Y., Zsebo, K. M. and Hogan, B. L. M. (1990). Embryonic expression of a haematopoietic growth factor encoded by the SZ locus and the ligand for c-kit. Natzire 347,667-669.

Nishikawa, S., Kusakabe, M., Yoshinaga, K., Ogawa, M., Hayashi, S., Kunisada, T., Era, T., Sakakura, T. and Nishikawa, S. (1991). In utero manipulation of coat color formation by a monoclonal anti-c-kit antibody: two distinct waves of c-kit dependency during melanocyte development. EMBO J. 10,2 11 1-2 118.

Okazaki, T., Yamataka, A., Fujiwara, T., Nishije, H., Fujimoto, T. and Mijiano, T. (1994). Abnormal distibution of nerve terminals in infantile hypotrophic pyloric stenosis. J. Pediatr. Surg. 29, 655-658.

Paulson, R.F., Vesely, S., Siminovitch, K.A. and Bernstein, A. (1996). Signaling by the W/Kit receptor tyrosine kinase is negatively regulated in vivo by the protein tyrosine phosphatase Shp 1. Nature Gen. 13,309-315.

Romeo, G., Ronchetto, P., Luo, Y., Barone, V., Seri, M., Ceccherini, I., Pasini, B., Bocciardi, R., Lerone, M., Kaariainen, H. et al. (1994). Point mutations affecting the tyrosine kinase domain of the RET proto-oncogene in Hirschspnmg's disease. Nature 367, 377-378.

Rottapel, R., Reedijk, M., Williams, D.E., Lyman, S.D., Anderson, D.M., Pawson, T. and Bernstein, A. (199 1). The SteeUW transduction pathway: kit autophosphorylation and its association with a unique subset of cytoplasmic signaling proteins is induced by the Steel factor. Mol. Cell. Biol. 11, 3043-3051. Schuchardt, A., D'Agati, V., Larsson-Blomberg, L., Constantini, F. and Pachnis, V. (1994). Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase receptor Ret. Nature 367,3 19-320.

Shultz, L. et d. (1993). Mutations at the murine motheaten locus are within the hematopoietic cell protein tyrosine phosphatase (Hcph)gene. Cell 73, 1445-1454.

Scott, G., Ewing, J., Ryan, D. and Abboud, C. (1994). Stem cell factor regulates human melanocyte-matrix interactions. Pigment Cell Res. 7,445 1.

Serve, H., Yee, N.S., Stella, G., Sepp, L.L., Tan, J.C. and Besmer, P. (1995). Differential roles of PI3-base and Kit tyrosine 821 in Kit receptor-mediated proliferation, survival and cell adhesion in mast cells. EMBO J. 14,473-483.

Silvers, W.K. (1979). The coat colors of mice. Heidelberg, Berlin, New York, Springer Verlag.

Smith, E.A., Seldin, M.F., Martinez, L, Watson, M.L., Choudhury, G.G., Lalley, P.A., Pierce, J., Aaronson, S., Barker, J., Naylor, S.L. and Sakaguchi, A.Y. (1991). Mouse platelet-derived growth factor receptor a gene is deleted in wI~Hand Patch mutations on chromosome 5. Proc. Natl. Acad. Sci. USA 88,481 1-4815.

Stephenson, D.A., Mercoia, M., Anderson, E., Wang, C., Stiles, C.D., Bowen-Pope, D.F. and Chapman, V.M. (1991). Platelet-derived growth factor receptor a-subunit gene (Pdgfm) is deleted in the mouse Patch (Ph) mutation. Proc. Natl. Acad. Sci. USA 88,6-10.

Stephenson, D.A., Lee, K., Nagle, D.L., Yen, C.-H., Morrow, A*, Miller, D., Chapman, V.M.and Bucan, M. (1994). Rump-white mutation is associated with an inversion of the proximal portion of mouse chromosome 5. Mammalian Genome 5, 342-348.

Tang, B., Mano, H., Yi, T. and Me, J.N. (1994). Tec base associates with c-kit and is tyrosine phosphorylated and activated following stem cell factor binding. Mol. Cell. Bid. 14,8432-8437. Torihashi, S., Ward, S.M.,Nishikawa, S.-I., Nishi, K., Kobayashi, S. and Sanders, K.M. (1995). c-kit-dependent development of interstitial cells and electrical activity in the murine gastrointestinal tract. Cell Tissue Res. 280,97-111.

Tsui, H., Siminovitch, K., de Souza, L. and Tsui, F. (1993). Motheaten and viable motheaten mice have mutations in the hematopoietic cell phosphatase gene. Nature Gen. 4, 124-129.

Tsujimura, T., Morii, E., Nozaki, M., Hashimoto, K., Moriyarna, Y., Takebayashi, K., Kondo, T., Kanakura, Y. and Kitamura, Y. (1996b). Involvement of transcription factor encoded by the mi locus in the expression of the c-kit receptor tyrosine kinase in cultured mast cells of mice. Blood 88, 1225- 1233.

Thuneberg, L. (1982). Interstitial cell of Cajal: intestinal pacemaker cells? Adv. Anat. Embryol. Cell Biol. 71, 1- 130.

Vandenbark, G.R.,Chen, Y., Friday, E., Pavlik, K., Anthony, B., decastro, C. and Kaufman, R.E. (1996). Complex regulation of human c-kit transcription by promoter repressors, activators, and specific myb elements. Cell Growth Dzz 7, 1383- 1392.

Vanderwinden, J.-M., Liu, H., De Laet, M.-H. and Vanderhaeghen, J.J. (1996). Study of the interstitial cells of Cajal in infantile hypertrophic pyloric stenosis. Gastroenterol. 111,279-288.

Wehrle-Haller, B ., Morrison-Graham, IS. and Weston, J.A. (1996). Ectopic c-kit expression affects the fate of melanocyte precursors in Patch mutant embryos. Dev. Biol. 177,463-474.

Yamamoto, K., Tojo, A., Aoki, N. and Shibuya, M. (1993). Characterization of the promoter region of the human c-kit proto-oncogene. Jpn. J. Cancer Res. 84, 1136- 1144.

Yasuda, H., Galli, S.J. and Geissler, EN.(1993). Cloning and functional analysis of the mouse c-kit promoter. Bioch. Bioph. Res. Corn. 191, 893-901. Yi, T. and Me, J.N. (1993). Association of hematopoietic cell phosphatase with c-Kit after stimulation with c-Kit ligand. Mol. Cell. Bid. 13,3350-3358.

Yoshida, H., Kunisada, T., Kusakabe, M., Nishikawa, S. and Nishikawa, S.I. (1996). Distinct stages of melanocyte differentiation revealed by analysis of nonuniform pigmentation patterns. Development 122, 120%12 14.

Young, H.M., Ciampoli, D., Southwell, B.R. and Newgreen, D.F. (1996). Origin of interstitial cells of Cajd in the mouse intestine. Developmental Biology 180,97407. lMAGE EVALUATION TEST TARGET (QA-3)

APPLIED- IMAGE. lnc -= 1653 East Main Street --,=L Rochester. NY 14609 USA ,== ,== Phone: 716/482-0300 --=-- Fax: 716l288-5989

0 1993. Apprled Image. Inc. N FlrghLs Resenred