FGFR1 Regulated Gene-Expression, Cell Proliferation and Differentiation

and 17/2007 TOMI JUKKOLA TOMI Division of Genetics Division University of Helsinki University Faculty of Biosciences Faculty Institute of Biotechnology and Institute of Biotechnology FGFR1 Regulated Gene-Expression, FGFR1 Regulated Gene-Expression, Developing Midbrain and Hindbrain Midbrain Developing Department of Biological and Environmental Sciences Department of Biological and Environmental Cell Proliferation and Differentiation in the and Differentiation Cell Proliferation Helsinki Graduate School of Biotechnology and Molecular Biology of Biotechnology School Helsinki Graduate Dissertationes bioscientiarum molecularium Universitatis Helsingiensis in Viikki Helsingiensis in Dissertationes bioscientiarum molecularium Universitatis

TOMI JUKKOLA FGFR1 Regulated Gene-Expression, Cell Proliferation and Differentiation in the Developing Midbrain and Hindbrain 17/2007 8 and Archaeal Virus SH1 Virus Archaeal 8 and φ lin ﬁ 6 RNA-Dependent RNA Synthesis 6 RNA-Dependent RNA φ Helsinki 2007 ISSN 1795-7079 ISBN 978-952-10-4034-4 Recent Publications in this Series: Yabal 32/2006 Monica Proteins Anchored Membrane Insertion of C-tail 33/2006 Raimo Mikkola and Mechanisms of Effects on Eukaryotic Cells Air Isolated Bacillus Non-Protein Food and Indoor 34/2006 Mikael Niku Generation in Chimeric Cattle Potential and B Lymphocyte Fates of Blood. Studies on Stem Cell Differentiation 35/2006 Minni Koivunen Molecular Details of Phage 36/2006 Kai Fredriksson and Dynamics of Coil-like Molecules by Residual Dipolar Couplings Structure Tsitko 1/2007 Irina Pollutants Organic Actinobacteria Degrading and Tolerating Characterization of 2/2007 Pekka Östman Ionization-Mass Spectrometry Atmospheric Pressure Microchip 3/2007 Anni Hienola Factor Modulation of Growth And HB-GAM in Neural Migration and Differentiation: N-Syndecan Activity in Brain O. Paavilainen Ville 4/2007 Structural Basis of Cytoskeletal Regulation by Twin 5/2007 Harri Jäälinoja Studies of Bacteriophage Cryo-Microscopy Electron 6/2007 Eini Poussu Engineering Applications in Protein Transposition DNA Mu in vitro 7/2007 Chun-Mei Li Gene to Function System of Phytopathogenic Bacterium Pseudomonas syringae: From III Secretion Type 8/2007 Saara Nuutinen Acetylcholine Receptors and alpha7 Nicotinic The Effects of Nicotine on the Regulation Neuronal Intracellular Signalling Pathways Aaltonen 9/2007 Jaakko Analysis during Pharmaceutical Solid Phase to Dissolution Testing: Polymorph Screening From Development and Manufacturing Antikainen 10/2007 Jenni Anchoring and Cell Wall Adhesive Properties of Lactobacillus crispatus: Surface Proteins Jing Li 11/2007 Arabidopsis Disease Resistance Novel Molecular Mechanisms of 12/2007 Piia Salo to Preparative-Layer Detection: From with Ultraviolet and Mass Spectrometric Thin-Layer Chromatography Miniaturized Ultra-Thin-Layer Technique 13/2007 Mikko Sairanen and Morphine Antidepressants Action of Plasticity in the and Neuronal Neurotrophins 14/2007 Camilla Ribacka C Oxidase by Cytochrome Transfer Redox-linked Proton 15/2007 Päivi Ramu PgtE of S. enterica: Role in Salmonella-Host Interaction Outer Membrane Protease/adhesin Alvesalo 16/2007 Joni in the Drug Development Process Application of Genomics and Proteomics and the Drug Discovery Screening for Chlamydia pneumoniae Dissertationes bioscientiarum molecularium Universitatis Helsingiensis in Viikki 17/2007

FGFR1 regulated gene-expression, cell proliferation and differentiation in the developing midbrain and hindbrain

Tomi Jukkola

Institute of Biotechnology and Department of Biological and Environmental Sciences Division of Genetics Faculty of Biosciences and Helsinki Graduate School of Biotechnology and Molecular Biology University of Helsinki Finland

Academic dissertation

To be presented, with the permission of the Faculty of Biosciences of the University of Helsinki, for public criticism, in the auditorium 1041 at the Viikki Biocenter 2, Viikinkaari 5, on August 17th 2007, at 12 o’clock noon. Supervised by Docent Juha Partanen Institute of Biotechnology University of Helsinki Finland Reviewed by Professor Eero Castrén Neuroscience Center University of Helsinki Finland Professor Seppo Vainio Oulu Biocenter University of Oulu Finland Opponent Ph.D. Mark Lewandoski National Cancer Institute Frederick Cancer Research & Development Center U.S.

Custos Professor Jim Schröder Department of Biological and Environmental Sciences University of Helsinki Finland

ISSN 1795-7079 ISBN 978-952-10-4034-4 (paperback) ISBN 978-952-10-4035-1 (PDF, http://ethesis.helsinki.ﬁ ) Edita, Helsinki 2007 If we knew what we were doing we wouldn’t call it science.

Albert Einstein TABLE OF CONTENTS

ABBREVIATIONS List of original publications I. Jukkola T*, Lahti L* Naserke T, Wurst W, Partanen J. FGF regulated gene- expression and neuronal differentiation in the developing midbrain-hindbrain region. Developmental Biology. 2006 Sep 1; 297(1): 141-57. II. Jukkola T*, Sinjushina N*, Partanen J. Drapc1 expression during mouse embryonic development. Gene Expression Patterns. 2004 Oct; 4(6): 755-62. III. Trokovic R*, Jukkola T*, Saarimäki J, Peltopuro P, Naserke T, Weisenhorn DM, Trokovic N, Wurst W, Partanen J. Fgfr1-dependent boundary cells between developing mid- and hindbrain. Developmental Biology. 2005 Feb 15; 278(2): 428- 39. IV. Kala K, Jukkola T, Pata I, Partanen J. Analysis of the midbrain-hindbrain boundary cell fate using a boundary cell-speciﬁ c Cre-mouse line. Manuscript submitted to Genesis.

SUMMARY REVIEW OF THE LITERATURE ...... 1 1. The early brain development ...... 1 1.1. Neural induction and neurulation ...... 1 1.2. Molecular events of neural induction and neurulation ...... 1 2. Mechanisms of brain patterning ...... 2 2.1. Anterioposterior patterning ...... 2 2.2. Isthmic organizer ...... 2 2.2.1. Induction and localization of isthmic organizer ...... 5 2.3. Neuronal development and regionalization of the midbrain ...... 9 2.4. Neuronal development and regionalization of the hindbrain ...... 11 2.4.1. Segmentation and compartmentalization in hindbrain ...... 12 3. Intercellular communication in the midbrain-R1 region ...... 14 3.1. FGF signaling in the midbrain-R1 development ...... 14 3.1.1. The FGF and FGFR molecules ...... 14 3.1.2. Expression patterns of FGFs and FGFRs in the mouse midbrain-R1 ...... 15 3.1.3. FGFRs in the mouse midbrain-R1 development ...... 16 3.1.4. FGFR signal transduction ...... 17 3.2. WNT signaling in the midbrain-R1 development ...... 21 4. Cell proliferation in the developing brain ...... 23 4.1. Interkinetic nuclear migration ...... 23 4.2. Cell cycle regulators ...... 25 4.2.1. Cyclins, Cdks and their inhibitors ...... 25 4.2.2. Cyclins and Cdks in neuronal development ...... 26 4.2.3. Cdk inhibitors in neuronal development ...... 27 4.3. Mechanisms controlling the cell polarity and asymmetric cleavage ...... 28 4.4. Cellular proliferation and regulation in the rhombomeric boundaries ...... 28 AIMS OF THE STUDY ...... 30 MATERIALS AND METHODS ...... 31 RESULTS AND DISCUSSION ...... 34 1. Microarray analysis of Fgfr1 CKO mutants (I-III) ...... 34 2. Identiﬁ cation of FGF regulated genes at the midbrain-R1 region (I, III and unpublished) ...... 43 2.1. Down-regulated genes (I and unpublished) ...... 43 2.1.1. Down-regulation of FGF-FGFR signaling components at E10.5 (I, III) ...... 43 2.1.2. Cnpy1 as a novel regulator of FGF signaling at the midbrain-R1 region (I and unpublished) ...... 45 2.1.3. Down-regulated genes expressed as a narrow stripe near the MHB (I and unpublished) ...... 46 2.1.4. Other down-regulated genes in the Fgfr1 CKO mutants (I) ...... 48 2.2. Up-regulated genes show loss of expression gradients in Fgfr1 CKO mutant embryos (I)...... 48 3. Neuronal progenitor cell populations are affected in Fgfr1 CKO mutant embryos (I) ...... 49 3.1. Neuronal progenitor cell populations are affected in Fgfr1 CKO mutant embryos (I) ...... 49 3.2. Midbrain dopaminergic neuron progenitor pool spreads in Fgfr1 CKO mutants (I) ...... 50 3.3. The most rostral serotonergic neuron precursors are lost in Fgfr1 CKO mutant embryos (I)...... 50 4. Identiﬁ cation of a putative target gene of WNT-β-catenin pathway (II) ...... 52 5. A slowly proliferating narrow population of cells exists at the boundary between the midbrain and R1 (III,IV)...... 54 5.1. Initial changes in gene expression occur close to the midbrain- hindbrain boundary in Fgfr1 mutants (III)...... 54 5.1.1. Initial gene expression changes of FGF-FGFR signaling components at E9.5 (III) ...... 54 5.2. Other FGFRs may mediate FGF signaling at a distance from the midbrain-hindbrain boundary (III) ...... 55 5.3. Isthmic constriction fails to develop in the Fgfr1 CKO mutant embryos (III) . 55 5.4. Expression of cell cycle regulators at the midbrain-R1 region (I, III, IV) ...... 55 5.5. Cell proliferation at the border between midbrain and R1 (II,IV) ...... 56 5.5.1. The location of p21 expressing cells at the border between midbrain and R1 (II,IV) ...... 56 5.5.2. Slowly proliferating cell population at the midbrain-R1 boundary (III,IV) ...... 57 5.5.3. The contribution of the boundary cell population to the midbrain-R1 development (IV) ...... 58 CONCLUSIONS ...... 60 ACKNOWLEDGEMENTS ...... 62 REFERENCES ...... 63 ABBREVIATIONS

5-HT 5-hydroxytryptophan (serotonin) ANR anterior neural ridge AP anterioposterior APC adenomatosis polysis coli protein BMP bone morphogenetic factor bp base pair(s) BrdU 5-bromo-2-deoxyuridine Cb cerebellum CNPY1 canopy1 CNS central nervous system Cp choroid plexus Cre Cre recombinase cRNA synthetic complementary RNA to mRNA produced from a DNA template DA dopaminergic DAPI 4´,6-diamino-2-phenylindole Di diencephalon DNA deoxyribose nucleic acid DRAPC1 adenomatosis polyposis coli down-regulated 1, APCDD1 protein DV dorsoventral E embryonic day EN engrailed homeobox transcription factor ERM (or ETV5) member of the PEA3 group of ETS domain containing TFs EST expressed sequence tag FGF fi broblast growth factor FGFR fi broblast growth factor receptor FLRT family of glycosylated proteins containing a fi bronectin III domain and leucine rich repeats (LRR) domain FST follistatin protein FZ frizzleds which are members of the seven transmembrane domain cell surface receptors GBX2 gastrulation brain homeobox 2 transcription factor GRG4 (or TLE4) groucho-related gene, transducin-like enhancer of split 4 HH Hamburger Hamilton stage HOX a particular subgroup of homeobox transcription factors hpf hours post fertilization HS heparan sulphate Ic inferior colliculus Ig immunoglobulin IGFBP Insulin-like growth factor binding protein IsO isthmic organizer kDa kilo Dalton (molecular weight) KO knockout LC locus coeruleus MAPK mitogen activated protein kinase Mb midbrain MHB midbrain–hindbrain boundary MKP3 MAPK phosphatase-3 mRNA messenger ribonucleic acid MRP multi-drug resistance protein OTX orthodenticle homolog (Drosophila) p21 (or Cdkn1a) 21 kDa cyclin dependent kinase inhibiting protein PAX paired domain containg homeobox transcription factor PCR polymerase chain reaction PH3 phosphohistone H3 PI3K phosphatidylinositol 3’ kinase R (e.g. R1) rhombomere (e.g. rhombomere 1) RA retinoic acid RNA ribonucleic acid RrF retrorubral fi eld SA serotonergic Sc superior colliculus SEF similar expression to Fgfs SFRP secreted frizzled-related protein SHH sonic hedgehog protein siRNA small interfering RNA used to decrease transcription SNc substantia nigra pars compacta SOX SRY-related HMG-box transcription factor SPRYs sprouty family of proteins Te telencephalon Teg tegmentum TF(s) transcription factor(s) TH tyrosine hydroxylase TRH thyrotrophin releasing hormone TUJ neuron specifi c beta III Tubulin TUNEL terminal deoxynucleotidyl transferase mediated nick labeling Tyr tyrosine residue v vermis of the cerebellum VTA ventral tegmental area VZ ventricular zone WNT(1) wingless-related MMTV integration site 1 ZLI zona limitans intrathalamica LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following three articles and one manuscript as well as unpublished data. In the text they are referred to by their roman numerals.

I. Jukkola T*, Lahti L* Naserke T, Wurst W, Partanen J. FGF regulated gene- expression and neuronal differentiation in the developing midbrain-hindbrain region. Developmental Biology. 2006 Sep 1; 297(1): 141-57.

II. Jukkola T*, Sinjushina N*, Partanen J. Drapc1 expression during mouse embryonic development. Gene Expression Patterns. 2004 Oct; 4(6): 755-62.

III. Trokovic R*, Jukkola T*, Saarimäki J, Peltopuro P, Naserke T, Weisenhorn DM, Trokovic N, Wurst W, Partanen J. Fgfr1-dependent boundary cells between developing mid- and hindbrain. Developmental Biology. 2005 Feb 15; 278(2): 428-39.

IV. Kala K, Jukkola T, Pata I, Partanen J. Analysis of the midbrain-hindbrain boundary cell fate using a boundary cell-speciﬁ c Cre-mouse line. Manuscript submitted to Genesis.

*equal contribution SUMMARY

The neuroectodermal tissue close to the midbrain-hindbrain boundary (MHB) is an important secondary organizer in the developing neural tube. This so-called isthmic organizer (IsO) regulates cellular survival, patterning and proliferation in the midbrain (Mb) and rhombomere 1 (R1) of the hindbrain. Signaling molecules of the IsO, such as fi broblast growth factor 8 (FGF8) and WNT1 are expressed in distinct bands of cells around the MHB. It has been previously shown that FGF-receptor 1 (FGFR1) is required for the normal development of this brain region in the mouse embryo. In the present study, we have compared the gene expression profi les of wild-type and Fgfr1 mutant embryos. We show that the loss of Fgfr1 results in the downregulation of several genes expressed close to the MHB and in the disappearance of gene expression gradients in the midbrain and R1. Our microarray screen identifi ed several previously uncharacterized genes which may participate in the development of midbrain–R1 region. Our results also show altered neurogenesis in the midbrain and R1 of the Fgfr1 mutants. Interest- ingly, the neuronal progenitors in midbrain and R1 show different responses to the loss of signaling through FGFR1. As Wnt1 expression at the MHB region requires the FGF signaling pathway, WNT target genes, including Drapc1, were also identifi ed in our screen. The microarray data analysis also suggested that the cells next to the midbrain- hindbrain boundary express distinct cell cycle regulators. We showed that the cells close to the border appeared to have unique features. These cells proliferate less rapidly than the surrounding cells. Unlike the cells further away from the boundary, these cells express Fgfr1 but not the other FGF receptors. The slowly proliferating boundary cells are necessary for development of the characteristic isthmic constriction. They may also contribute to compartmentalization of this brain region.

Review of the Literature

REVIEW OF THE LITERATURE 1. The early brain development During neurulation the lateral edges of the plate, or neural folds, become 1.1. Neural induction and neurulation prominent and begin to bend toward each other to form a tube. As the neural folds of The complex development of central the plate close, the dorsal midline of the nervous system (CNS) begins already embryo begins to sink somewhat below prior to the gastrulation. The CNS begins the surface, and in making contact, the as a flattened layer of neuroepithelial central depression in the plate becomes cells on the dorsal surface of the embryo. more distinctive. As they become They are induced by the signals from the adjacent, the neural folds fuse, producing surrounding ectoderm and mesenchyme. a hollow neural tube and a continuous The molecular interactions create a dorsal epithelium. After separation from sheet of cells that are forming the neural the overlying epithelium, the neural tube plate. With the onset of neurulation, the thickens differentially as various regions embryo also begins to elongate along begin to form different subregions of the its anterioposterior (AP) axis. At the brain and spinal cord. anterior end of the primitive streak of the gastrulating mouse embryo (approximately 1.2. Molecular events of neural at embryonic (E) day 7.5) lays an important induction and neurulation structure, the node. The inductive and organizing activity of an early node was Recent research has identified specific fi rst reported in amphibians by Spemann molecules that bring about neural and Mangold (reviewed in Hamburger, induction. The complex set of molecular 1988). In these experiments the dorsal interactions commits ectodermal cells blastoporal lip from a newt gastrula stage over the notochord to becoming neural embryo was transplanted into another tissue, being the fi rst step in the formation embryo, which induced the formation of of the nervous system. In amphibians, a secondary embryo with a normal body three secreted molecules, noggin (Nog), pattern. The inductive properties of the follistatin (Fst), and chordin (Chrd), are transplanted tissue indicated that the node among the inductive proteins. It was fi rst (or Spemann’s organizer in amphibians) thought that these molecules directly not only instructed cells in adjacent tissues protect the ectoderm from epidermalizing to acquire new fates, but also provided signals and protect the mesoderm information about the development of AP from ventralizing signals, allowing the axis. Similar transplantation experiments development of a patterned neural plate have shown that the Hensen’s node in (Harland and Gerhart, 1997). However, chick and the node in mouse also act as subsequent research on amphibians has axis-inducing organizers. In addition shown that these inductors act by blocking to the node, inductive signals from the the action of an inhibitors, such as bone perineural ectoderm as well as from the morphogenetic protein 4 (BMP4), in ventral endoderm and axial mesoderm the dorsal ectoderm (reviewed in Lee are thought to infl uence the development and Jessell, 1999). Subsequently, in the of the AP and dorsoventral (DV) polarity absence of BMP4 activity, dorsal ectoderm (Stern, 2001). forms neural tissue as a default state. In

1 Review of the Literature amphibians, neural speciﬁ cation requires experiments have shown that tissue the continuous activity of various BMP explants of the early primitive node can antagonists from blastula through gastrula induce a nervous system with both anterior stages. Due to redundancy of different (head) and posterior (trunk) components, BMP antagonists only the inactivation of whereas explants of older node tissue only three of them (Nog, Fst, and Chrd) has induce trunk characteristics (reviewed been shown to result to a disruption of in Harland and Gerhart, 1997). Earlier dorsal development (Khokha et al., 2005). experiments also showed that certain In birds and mammals, the inactivation of inductors cause the formation of more BMPs is also likely important in neural anterior, and others more posterior, induction, but the role of noggin, follistatin neural structures. For example, in the and chordin as inhibitors of BMPs is less presence of retinoic acid (RA), WNTs, or clear. For example, mice that lack Fst FGFs, the induced neural structures are (Matzuk et al., 1995) or Nog (McMahon posteriorized, and more caudal structures et al., 1998) function do not show obvious form (Kudoh et al., 2002; Storey et al., defects in neural induction although later 1998). In mammals, the anterior hypoblast neural defects are observed in Nog -/- (or anterior visceral endoderm (AVE) in mice. In addition, in mouse, Bmp4 mutant mouse) expresses genes characteristic of embryos do not exhibit an apparent the prechordal plate and probably initiates expansion in neural tissue, as might have head formation (Stern, 2001). Shortly been anticipated from experiments of following the neural induction, additional BMP4 inactivation in amphibians (Winnier signals from the notochord and AVE result et al., 1995). in the expression of the transcription In addition to BMP antagonists, other factor Otx2 in the head region (forebrain/ mechanisms have been demonstrated to midbrain) and Gbx2 in the trunk region decrease the BMP signaling in neural (hindbrain/spinal cord). Both the signaling induction. For example, in the chick, molecules and transcription factors are early FGF expression downregulates Bmp important for the anterior structures. To expression in the prospective neural plate produce a head region, it is necessary to (Wilson et al., 2000). FGF8 signaling is block signals from the RA/WNT/FGF also able to initiate the events of neural signaling pathways as well as the BMP4 induction whereas addition of noggin or signal. chordin to the chick epiblast can alter the position of the neural folds but cannot 2.2. Isthmic organizer induce ectopic neural tissues (Streit et al., 2000). The proper development of the CNS requires a fine-tuned pattern of cell identities along the AP-axis. Further 2. Mechanisms of brain patterning speciﬁ cation and compartmentalization of the different neuronal domains is guided 2.1. Anterioposterior patterning by so called secondary signaling centers (Fig. 1A). The isthmic organizer (IsO) is An important second step is initial such a center which is formed between regionalization of the CNS into broad midbrain (Mb) and the most anterior part anterioposterior regions. Numerous of the hindbrain, rhombomere 1 (R1).

2 Review of the Literature

Figure 1. The role of signaling molecules and transcription factors in the development of midbrain-R1 region. (A) Side view of an E10.5 mouse embryo. (B) Dorsal view of an E10.5 embryo. (C) Midsagittal view of an E12.5 mouse embryo. (D) Midsagittal view of an E18.5 brain. Two secreted factors, FGF8 and WNT1, are required for the development of midbrain- R1 region (A,B). In addition to FGF8 and WNT1, Shh expression in the ventral midline regulates dorsoventral patterning and the expression of Fgf8 at midbrain-R1 region during early development (E8.5-E12)., Otx2 and Gbx2 are transcription factors which act antagonistically to set up the position of the IsO (B). The expression of the En1/En2 transcription factors are required for the maintenance of the FGF8. Also, transcription factors Pax2, Pax5 and Pax8 are expressed in the overlapping patterns, but only Pax2 is required for the induction of the Fgf8 in the IsO (Ye et al., 1998). In addition to the AP and DV patterning, the IsO signals regulate the development of neuronal population such as dopaminergic (DA) neurons in the midbrain and serotonergic (SA) neurons in the R1. Initially, bone morphogenetic proteins (BMPs) expressed in the roof plate and signals from the IsO are regulating the induction of the locus coeruleus (LC) progenitors. The development of III and IV cranial ganglia are also dependent on the IsO signaling. The IsO activity is also required for the ordered speciﬁ cation of inferior and superior colliculi in the dorsal midbrain (mesencephalic structures) as well as for the development of cerebellum in the dorsal R1 (metencephalic structure) (C,D). III, third cranial ganglion (or oculomotor nerve); IV, fourth cranial ganglia (or trochlear nerve); Cb, cerebellum; cp, choroid plexus; Di, diencephalon; Ic, inferior colliculus; mb, midbrain; r1, rhomobomere1; Sc, superior colliculus; Te, telencephalon; Teg, tegmentum; vHb, ventral hindbrain. Adapted from Wurst and Bally-Cuif (2001) and Ang (2006).

3 Review of the Literature axial position (Echevarria et al., 2003; Apart from the IsO, secondary Liu and Joyner, 2001a; Lumsden and organizers have been identifi ed at other Krumlauf, 1996; Wurst and Bally- border regions in the vertebrate neural Cuif, 2001). Transplantation of tissue tube (Echevarria et al., 2003; Lumsden grafts originating from the IsO into the and Krumlauf, 1996; Wurst and Bally- caudal forebrain or hindbrain lead to the Cuif, 2001). For example, the border induction of ectopic midbrain or cerebellar between prethalamus and thalamus called structures, respectively (Martinez et al., zona limitans intrathalamica (ZLI) is an 1991). Thus, the IsO seems to posses the important signaling center for diencephalic organizing properties which are guiding AP patterning (Fig. 1A). Several signaling the development of adjacent brain regions molecules, including SHH, WNTs and (see Figs. 1 and 2). FGFs, have been shown to mediate the The molecular specifi cation of isthmic organizer activity of ZLI (Kiecker and cell fates appears to depend on signaling Lumsden, 2005; Vieira et al., 2005). In molecules that control the pattern and the telencephalon, at least SHH and FGF identity of adjacent cell types. Members of signaling from the anterior neural ridge the fi broblast growth factor (FGF) family (ANR) region are needed for the correct of secreted proteins have a prominent role patterning of anterior fates such as lens in AP patterning of the midbrain and R1. and olfactory placodes (Bailey et al., 2006; Additional signals involving members of Ekker et al., 1995; Ohkubo et al., 2002). the WNT, sonic hedgehog (SHH) and bone It seems that Fgf8 expression in the ANR morphogenetic protein (BMP) families has a similar role to Fgf8 expression in the also contribute to the correct patterning IsO where it is required for AP patterning of midbrain-R1 region (see Fig. 1A,B). of the adjacent tissues (Fig. 2; see below). These planar signals emanated from the IsO are required for the development of the midbrain and cerebellum (Fig. 1C,D).

Figure 2. Inductive properties of the IsO: Tissue from the midbrain-R1 region and FGF8 soaked beads can induce ectopic midbrain and cerebellum development. Transplantation of Mb-R1 tissues to different positions induces Engrailed (En) expression gradients and affects the AP patterning as indicated. Similar induction of En2 and ectopic structure is observed with FGF8 beads. Di, diencephalon; Mb, midbrain; R1, rhombomere 1; Te, telencephalon. Adapted from Liu and Joyner (2001a).

4 Review of the Literature

2.2.1. Induction and localization of loss-of-function (Acampora et al., 1997; isthmic organizer Puelles et al., 2003; Suda et al., 1997) or gain-of-function (Broccoli et al., 1999; The molecules needed for positioning Katahira et al., 2000) manipulations of the IsO are the two homeobox of Otx2 activity, respectively. OTX2 is transcription factors, Otx2 and Gbx2 necessary for maintenance of midbrain which are expressed at E7.5 (0 somite identity through a continued repression on stage) within the anterior and posterior the posteriorizing determinants (Fig. 3). In neuroectoderm, respectively (Fig. 3 and this context, Gbx2 appears to be the major 4; Ang et al., 1994). This juxtaposition is molecular determinant in maintaining the maintained through mutual antagonism, rhombomeric identity by establishing its resulting in a sharp midbrain/R1 border anterior border. On the contrary, Gbx2 by E9.5 with Otx2 transcripts found in inactivation or overexpression shifts the forebrain and midbrain, and Gbx2 Fgf8 expression caudally or rostrally, expression maintained in rhombomeres 1 respectively (Katahira et al., 2000; Millet to 3 (Fig. 3; Li and Joyner, 2001). In mouse et al., 1999; Wassarman et al., 1997). In mutants for Otx2 or Gbx2 genes, cells addition, mice homozygous for a Gbx2 fail to maintain the positional identities hypomorphic allele shows caudal shift in (Acampora et al., 1997; Li and Joyner, expression of Fgf8 and Fgf17 domains 2001; Li et al., 2002; Martinez-Barbera (Waters and Lewandoski, 2006). Otx2 (and et al., 2001). Otx2 -/- mouse embryos Otx1) has also been shown to antagonize die early during embryogenesis and the and deﬁ ne the Shh expression domain to prospective anterior brain regions are lost ventral midbrain. The co-repressor for the (Acampora et al., 1997). Gbx2 function antagonism on Shh expression is thought also appears to be necessary during to be GRG4 (Puelles et al., 2003; Puelles the early development of neural plate. et al., 2004). Although defects are milder than in Otx2 Induction of engrailed transcription mutant embryos, in the absence of Gbx2 factor genes En1 and En2 and paired box function derivatives of rhombomeres 1-3 transcription factor genes Pax2 and Pax5 fail to form (Wassarman et al., 1997). In is regulated by the general AP patterning addition, conditional inactivation of Gbx2 process (Liu and Joyner, 2001a; Olander et after E9.0 has demonstrated the necessity al., 2006). From early-somite stages, En1, of its continued expression for the correct En2, Pax2 and Pax5 become expressed establishment of the isthmus and normal across the Otx2/Gbx2 border in a graded cerebellar midline development (Li et fashion (see Fig. 1B and Fig. 3; Liu and al., 2002). Initially (at E7.5), Otx2 and Joyner, 2001b). The expression of En1 Gbx2 expression defines two distinct and Pax2 is earlier, whereas that of En2 domains, and later these correspond with and Pax5 is later than Fgf8 expression in the compartment boundaries formed at the the midbrain-R1 region (Fig. 4; Crossley border between midbrain and R1 (Trokovic and Martin, 1995). However, neither et al., 2003; Zervas et al., 2004). FGF8 nor WNT1 activities are required Both Otx2 and Gbx2 are showed to for the initial induction of Pax2/5 and exert a dose-dependent activity on Fgf8 En1/2 expression (Olander et al., 2006). expression. A rostral or caudal shift of Later in development, however, FGF8, Fgf8 expression domain is seen both in WNT1, EN1/2 and PAX2/5 molecules

5 Review of the Literature

Figure 3. Schematic of the IsO induction and development. Homeodomain transcription factors Otx2 and Gbx2 position the isthmic organizer in the neural primordium at around E7.5 (A). The co-regulation of Otx2 and Gbx2 is thought to determine the precise location of the MHB and the transcriptional regulation might be involved in the inductive events characteristic of the isthmic organizer. Once the Otx2/Gbx2 interface (positional specifi cation) has been set up, independent signaling pathways (WNT, FGF, RA) converge in their activity to drive organizer function (B). Fibroblast growth factors secreted from the anterior R1 are necessary for, and suffi cient to mimic, organizer activity in patterning the midbrain and anterior hindbrain, and are tightly controlled by feedback inhibition (e.g. OTX1 and 2 indirectly; cytosolic SPRYs directly). Numerous feedback loops maintain appropriate midbrain-R1 region and IsO gene expression (C). A network of other secreted factors such as WNT1 as well as transcription factors, including OTX2, GBX2, EN1/2, PAX2/5/8 and GRG4, further refi nes the expression domain and level of FGF8 at the MHB through opposing effects on PAX2 activity (C). At this stage expression boundaries of other MHB-associated genes become also more restricted. Modifi ed from Hynes and Rosenthal (1991), Wurst and Bally-Cuif (2001), Ye et al. (1998), Raible and Brand (2004) and Guo et al. (2007) and referenced therein.

6 Review of the Literature form a positive regulatory network al., 1999; Shih and Fraser, 1996; Weinstein which maintains the isthmic specific et al., 1994). gene-expression and is necessary for the Expression patterns of Engrailed maintenance of midbrain-R1 identity genes in R1 coincide with parts of the (Wurst and Bally-Cuif, 2001). The mouse neural tube that generate dorsal and ventral En1 and En2 homeobox transcription structures, such as the cerebellum and factors are expressed in a domain pons. In the midbrain, En1 and En2 activity encompassing the posterior midbrain is regulating the development of superior and anterior hindbrain (Davidson et al., and inferior colliculi and ventral midbrain 1988; Davis and Joyner, 1988). Studies nuclei (see below). En1 mutants, which in Xenopus laevis have indicated that die on the day of birth (E18.5), lack the induction of AP neural patterning might cerebellum and inferior colliculus (Wurst involve signals from the dorsal mesoderm et al., 1994). By contrast, En2 mutants are to the overlying ectoderm (Ruiz i Altaba viable and fertile and have minor defects and Jessell, 1993). Work in several species in cerebellar foliation (Millen et al., 1994). suggests that signals from the anterior In the chicken, overexpression of En1 and mesendoderm or notochord might regulate En2 in midbrain showed that engrailed the expression of Engrailed genes in the proteins regulate the anterior-posterior neural plate (Ang and Rossant, 1993; polarity of the optic tectum, the homolog Darnell and Schoenwolf, 1997; Hemmati- of the mammalian superior colliculus Brivanlou et al., 1990; Shamim et al., (Friedman and O’Leary, 1996; Itasaki and 1999). However, in mouse or in zebraﬁ sh Nakamura, 1996; Logan et al., 1996). In which lack the notochord, AP patterning is addition, mouse knockout analysis has normal (Ang et al., 1994; Klingensmith et shown that En1 and En2 are required

Figure 4. The onset of MHB-associated genes during the IsO development. Modiﬁ ed from Hynes and Rosenthal (1991), Wurst and Bally-Cuif (2001) and Raible and Brand (2004) and referenced therein.

7 Review of the Literature for the maintenance of the midbrain plate (tectum). Strong Wnt1 expression dopaminergic (DA) neuronal population domain is located also in the most ventral in later stages of embryonic development midbrain (tegmentum). In contrast to (Alberi et al., 2004; Simon et al., 2001; dynamic Fgf8 and Wnt1 expression, the Simon et al., 2004; Simon et al., 2005). expression of En1, En2, Pax2 and Pax5 Another level of molecular control remains relatively broad on both sides of needed for the IsO activity relies on the the Otx2/Gbx2 border between E7.5 and secreted signaling proteins. Like in many E12.5. other CNS regions, the specification of In chicken FGF8 protein has been neural types in midbrain-R1 region is shown to have same midbrain/cerebellum controlled by secreted signals derived inducing properties than the isthmic from local signaling centers (Echevarria et tissue grafts (see Fig. 2; Crossley et al., al., 2003; Lumsden and Krumlauf, 1996; 1996). So far these inducing properties Wurst and Bally-Cuif, 2001). FGF8/17/18 have been demonstrated only for FGF8. and WNT1 are signaling molecules Interestingly, only the FGF8b splice format required for the IsO activity (Figs. 1 and 3). induces ectopic expression in the mouse First Fgf8 and Wnt1 are expressed broadly forebrain of genes normally expressed in in the future midbrain-R1 region but later the isthmus (Olsen et al., 2006). MHB expressions boundaries get juxtaposed specific inactivation of Fgf8 results to (Broccoli et al., 1999). The onset of Fgf8 extensive apoptosis in the midbrain-R1 expression in mouse is around 3-5 somite region at E8.5 (Chi et al., 2003). Similarly stage and is largely limited to the Gbx2 in Wnt1 mutant embryos cells in the positive side of the midbrain-R1 region midbrain-R1 region die apoptotically (Fig. 4; Crossley and Martin, 1995). At around the same stage (McMahon and E9.0, Fgf8 expression is restricted to a Bradley, 1990; McMahon et al., 1992; narrow band at the most anterior part of Thomas and Capecchi, 1990). In concert the R1 whereas the Wnt1-positive cells are with FGF signaling, SHH is thought to located in the midbrain (Wurst and Bally- regulate the dorsoventral patterning of the Cuif, 2001). After the initiation of Fgf8 midbrain-R1 region (Placzek, 1995; Ye et expression, GBX activity is maintaining al., 1998). SHH is expressed ventrally in the Fgf8 expression in the R1, whereas the ﬂ oor plate throughout the midbrain-R1 OTX2 and GBX2/FGF8 regulate each boundary region (see Fig. 1A). other negatively, leading to refinement Although there is vast information of the sharp expression boundaries (see about the complex genetic interactions Fig. 3; Broccoli et al., 1999; Wassarman controlling the localization of isthmic et al., 1997; Millet et al 1999). Initially, organizer, the exact factors responsible Wnt1 expression is detected broadly in for its induction are currently unknown. the prospective midbrain territory, but its In addition to the early regionalization of expression boundaries are also later reﬁ ned the neuroectoderm, it has been suggested (Figs. 3 and 4). At E10, Wnt1 expression is that a transient expression of FGF4 in detected as a narrow band that surrounds the anterior notochord is needed for the the most posterior midbrain, just anterior induction of midbrain-R1 specific gene to the Fgf8 expression domain. Dorsally, expression (Shamim et al., 1999). The Wnt1 expression is also restricted to a induction of Fgf8 expression at the IsO narrow stripe of cells located in the roof might be due to secreted FGF8 from

8 Review of the Literature the adjacent cardiac mesendodermal The midbrain contains the nuclear cells (Crossley et al., 1996). In addition, complex of the oculomotor nerve (also other signals from the mesendoderm are called III cranial nerve) as well as the required for the induction of the IsO- trochlear nucleus (also called IV cranial associated transcription factors such as nerve); these cranial nerves innervate En1 and En2 (Ang and Rossant, 1993). muscles that move the eye and control Recently, the data from the transplantation the shape of the lens and the diameter of experiments in chick embryos suggest that the pupil. The ventral midbrain region convergent WNT and FGF signaling at the form also two distinct nuclei known as the gastrula stage is required to generate the substantia nigra pars compacta (SNc) and cellular properties characteristics for the the ventral tegmental area (VTA) as well IsO (Olander et al., 2006; see Fig. 3). as the retrorubral ﬁ eld (RrF), which are the most prominent sources of dopaminergic 2.3. Neuronal development and (DA) neurons in the CNS (German and regionalization of the midbrain Manaye, 1993). Indeed, midbrain neurons that use dopamine as a neurotransmitter The dorsal part of the midbrain is formed constitute about 75% of all DA neurons by two paired rounded structures, the in the adult brain (Wallen and Perlmann, superior and inferior colliculi. The superior 2003). The location of postmitotic DA colliculus receives input from the retina neurons in the brain is usually determined and the visual cortex and participates in by the immunohistological localization of a variety of visual reflexes, particularly the enzyme tyrosine hydroxylase (TH), the tracking of objects in the contralateral which is the rate-limiting enzyme of visual field. The inferior colliculus, dopamine synthesis. Historically, the DA formed in the caudal midbrain receives neurons in the midbrain of the mouse are both crossed and uncrossed auditory ﬁ bres placed in three cell groups according to and projects upon the medial geniculate TH immunostaining: nucleus A8 cells in body, the auditory relay nucleus of the the RrF, nucleus A9 cells in the SNc and thalamus. nucleus A10 cells in the VTA (Fig. 5).

Figure 5. Midbrain dopaminergic (DA) cell populations during mouse brain development. Signals affecting the early development of DA precursors and patterning of the midbrain-R1 region at E9.5 (A) and the location of TH+ neurons corresponding to A8-A11 groups at E14.5 (B). Di, diencephalon; Ic, inferior colliculi; IsO, isthmic organizer; Mb, midbrain; Sc, superior colliculi; p, prosomere; r, rhombomere; Te, telencephalon. Adapted from reviews by Simon et al. (2003) and Ang (2006). 9 Review of the Literature

Cells of the SNc contain the black receptor for SHH signaling) knockouts the pigment melanin; these synthesize amount of midbrain DA progenitors are dopamine and project to cells of the considerably reduced (Blaess et al., 2006). dorsolateral striatum and caudate putamen The induction and maintenance of DA constituting the nigro-striatal pathway. neurons requires also other factors such as By contrast, neurons of the VTA and RrF TGF-βs (Farkas et al., 2003). Conversely, project to the ventral striatum forming TGF-α is speciﬁ cally needed only for the a part of the mesolimbic system and development of DA neurons of the SNc. establish also additional connections to the The inactivation of TGF-α in mice leads prefrontal cortex (mesocortical system). to 50% reduction of SNc neurons (Blum, DA neurons of the SNc are involved in 1998). the coordination of movements, whereas A specific marker gene for the VTA and RrF neurons are controlling DA progenitors in the midbrain is e.g. mood and rewarding through the retinaldehyde dehydrogenase (also known mesolimbic and mesocortical systems. as Aldh1, Aldh1a1 or Raldh1), which Degeneration of the SNc neurons catalyzes the oxidation of retinaldehyde and their connections to the striatum and into retinoic acid. However, the role of frontal cortex is implicated in several Raldh1 in midbrain DA progenitors is not CNS disorders, including Parkinson’s clear (Prakash and Wurst, 2006). The role disease (Hirsch et al., 1997; Hynes and of WNT signaling in the midbrain DA Rosenthal, 1999). Conversely, over- development has also been established stimulation of VTA neurons has been recently. Prakash et al. (2006) showed that linked to schizophrenia and drug addiction Wnt1 is required for the early development (Chao and Nestler, 2004; Sesack and Carr, of DA neurons through maintenance of 2002). Otx2 transcription factor in the midbrain. The development of the superior and Recent studies have shown that several inferior colliculi and midbrain DA neurons other transcription factors are expressed is regulated by the IsO (Andersson et al., both in DA progenitors and later in 2006a; Brodski et al., 2003; Prakash et immature and mature neurons. Midbrain al., 2006; Trokovic et al., 2003). The DA DA progenitors are expressing a complex progenitors are located rostrally to (or set of transcription factors including within) the IsO as early as E9.0 in mouse, Otx2, Lmx1a, Lmx1b, En1, En2, Msx1, and they differentiate in this region at Msx2, Neurogenin 2 (Ngn2) and Mash1 a time between E9.0-E14.0 (Hynes and (Andersson et al., 2006a; Andersson et Rosenthal, 1999). The first postmitotic al., 2006b; Kele et al., 2006; Puelles et DA neurons can be detected in the ventral al., 2003; Puelles et al., 2004; Simon et midbrain at approximately E10.5 in mouse. al., 2001). These transcription factors as The induction of the DA precursors is well as Nurr1 and Pitx3 are also important regulated by both SHH from the ﬂ oor plate in the later development of DA neurons and FGF8 from the IsO (Ye et al., 1998) (Smidt et al., 2004; Smits et al., 2003; (see Fig. 5). Addition of SHH and FGF8 Wallen et al., 1999; Zetterstrom et al., on embryonic rat neural tube explants 1997). For example, in En1/2 double induced DA neuron development in ectopic mutants, which lack the IsO activity, and locations, and blocking of the two signals thus downregulate Fgf8 by E9.0 stage, DA with antibodies prevented it. Furthermore, neurons are initially generated (Alberi et in Shh knockouts and conditional Smo (a al., 2004; Simon et al., 2001; Simon et al.,

10 Review of the Literature

2003; Simon et al., 2004; Simon et al., development of pons. The cerebellum 2005). This suggests that they are induced is essential for fine motor control of by the earlier, broader Fgf8 expression movement and the body position, and spanning the midbrain-R1 region (at E8.5). its dysfunction disrupts balance and Alternatively, FGF8 from the IsO may not impairs control of speech, limb and eye be needed for the induction, but rather for movements. The cerebellum consists the maintenance and correct patterning of of a medial vermis and two laterally surrounding neuronal populations. On the located hemispheres. Anatomically, the other hand, Nurr1 and Pitx3 are needed cerebellum is dived into anterior, posterior for the maintenance and survival of DA and fl occulonodular lobes. The developing neurons not for the induction of precursors cerebellum consists mainly of three types (Saucedo-Cardenas et al., 1998; Smidt et of neuronal cells: granule cells in the al., 2004). external germinal layer, Purkinje cells, As mentioned above, Otx2 plays an and neurons of the deep nuclei. The pons important role in controlling specifi cation, functions to relay signals from the cortex maintenance and regionalization of the to assist in the control of movement where vertebrate brain. Recently, its role in as the medulla is joined to the spinal cord further development of midbrain neuronal and controls unconscious, yet essential, fates has been addressed by generating body functions such as breathing, conditional mutants inactivating Otx2 swallowing, blood circulation and by Cre recombinase expressed under muscle tone. Both in mammals and fi sh, the control of the En1 gene (Kimmel et embryological manipulations and gene al., 2000). In these conditional mutants, ablation studies have provided evidence the transcriptional repressor Nkx2.2 that normal cerebellar development was ventrally expanded, which in turn depends on formation and function of the leads to the loss of Wnt1 expression isthmic organizer. (Prakash et al., 2006; Puelles et al., 2004). In addition to isthmic signals, such Importantly, the normal generation of as WNT1 and FGF8, BMP signals from midbrain DA neurons in these conditional the dorsal R1 and Math1 expression in Otx2 mutants was not affected as long as the rhombic lip characterize the different Nkx2-2 was also ablated from the ventral cell types that form the complex neuronal midbrain, suggesting that OTX2 function subtypes in cerebellum. Mice lacking in midbrain DA neuron induction is to Math1 expression show a complete repress Nkx2-2 (Prakash et al., 2006). loss of the cerebellar granule-cell layer Thus, Otx2 is required to control identity and hindbrain mossy-fiber afferent and fate of ventral progenitors generating nuclei indicating improper precursor dopaminergic neurons in the midbrain cell differentiation and migration (Ben- through repression of Nkx2-2. Arie et al., 1997). Recent data indicate that BMP signaling pathway regulating 2.4. Neuronal development and Lmx1a transcription factors is required regionalization of the hindbrain for the development of R1 roof plate and its derivatives (Chizhikov et al., 2006). During embryonic development Thus, in addition to the IsO, the roof cerebellum arises from dorsal R1, adjacent plate seems to control both patterning to the fourth ventricle. By contrast, the and cellular proliferation in the cerebellar basal region of R1 will contribute to the primordium.

11 Review of the Literature

The locus coeruleus (LC) is a located in the anterior hindbrain whereas nucleus in the brain stem responsible for groups B1-B4 are located more caudally physiological responses to stress and panic. (reviewed in Hynes and Rosenthal, The LC progenitor cells are born in the 1999). The rostral domain (B5-B9) rostral rhombic lip (Lin et al., 2001), from provides innervation mainly to forebrain where they migrate ventrally within the targets. The regional restriction of 5-HT developing neural tube, and take residency neuron induction and later specifi cation near the fourth ventricle. The mature LC is is believed to result from early signaling localized in the lateral walls of the fourth events that pattern the MHB region and ventricle in R1. This nucleus is the main the more anterior hindbrain. In addition to noradrenergic centre of the brain and the isthmic FGF8 signals and SHH from the main source of noradrenergic innervations fl oor plate, FGF4 is needed for SA neuron in the CNS. The early development of induction in in vitro cultured R1 explants noradrenergic neurons of the LC requires (Ye et al., 1998), but its role in vivo has the establishment of IsO and is regulated not been demonstrated. The homeobox by a BMP induced signaling cascade transcription factor Pet1 has been used as a that specifies LC progenitors. Initially, specifi c marker of developing and adult SA BMP signaling pathway activates Mash1 neurons and it is expressed shortly before transcription factor in LC progenitors 5-HT appears in the mouse hindbrain at of the adjacent rhombic lip (Guo et al., E11.0. Recently, the differentiation of SA 1999). Subsequently, Mash1 activates the neurons in vertebrates has come under homeobox genes Phox2a and Phox2b. In wide examination and has been shown to the absence of Mash1 (Hirsch et al., 1998), involve also other homeobox transcription Phox2a (Morin et al., 1997), or Phox2b factors. Gain-of-function approaches (Pattyn et al., 2000), the LC neurons do not have implicated that homeobox TFs Pet1 form. Phox2a and Phox2b also regulate the (also known Fev), Lmx1b and Mash1 expression of tyrosine hydroxylase (TH) (also known Ascl1) are required for the and dopamine-ß-hydroxylase (DßH), two specifi cation of all SA neurons whereas key enzymes of noradrenergic biosynthesis Nkx2.2 and zinc finger TF Gata3 are required for the LC neuronal activity. needed for the appearance of caudal SA The IsO also regulates differentiation neurons (Ding et al., 2003; Hendricks et of serotonergic neurons (SA) neurons al., 2003; Pattyn et al., 2004). in the anterior hindbrain. These neurons are characterized by the expression 2.4.1. Segmentation and of serotonin transmitter (also known compartmentalization in hindbrain as 5-hydroxytryptamine, 5-HT). The serotonergic system plays a key role in In mouse, the best studied system modulating behaviors, such as appetite for compartmentalization is the and anxiety and has been implicated early embryonic hindbrain which is in many human disorders of mood and characterized by developmental segments mind. Mammalian SA neurons have called rhombomeres. Rhombomere (R)1 been classically grouped into nine cell and R2 (metencephalon) give rise to the populations (B1-B9) which are defi ned by cerebellum and pons. Boundaries between immunohistochemical staining with anti- the rhombomeres coincide with the anterior 5-HT. The more rostral SA groups (B5- boundaries of expression of homeobox- B9 also called the rostral SA domain) are containing regulatory genes. Thus, each of

12 Review of the Literature the eight mouse rhombomeres (R1-8) can evidence that FGF8 can substitute for be characterized by a specifi c combination IsO activity in inducing R1 character of Hox gene expression. However, R1 when expressed in hindbrain tissue. is devoid of Hox gene expression. In The suggestion is that FGF8 is used addition, rhombomeres can be identifi ed in restricting the anterior limit of Hox by the pattern of nerves that emerge from expression by regulating anterior hindbrain them. Importantly, proliferating cells character (i.e. R1). Thus both RA and FGF between the rhombomeres do not mix, as signaling are regulating the Hox expression shown by clonal analysis using injected in the hindbrain. The experiments in frog intracellular dye (Fraser et al., 1990), thus and fi sh suggest that FGF and RA signaling proposing that they are compartments. work independently to pattern distinct In addition, the compartment boundaries regions along the AP axis and that the show specific kinetics of cell cycle Cdx homeobox genes (vertebrate caudal progression (see section 4.4. below). homologs) are responsible for transducing The patterns of Hox gene expression these signals (Isaacs et al., 1998; Shimizu imprint a positional code on the cells and et al., 2006). CDX proteins directly contribute to the individual identities of regulate the expression of the posterior the neural precursors (Guthrie et al., 1991; Hox genes through direct binding to the Guthrie et al., 1992). Segmentation of the cis-regulatory elements of these genes. vertebrate hindbrain into rhombomeres is Therefore the RA and FGF pathways may essential for the AP patterning of cranial be integrated at the transcriptional level to motor nuclei and their associated nerves. control the Hox expression via Cdx genes. For example, a pair of cranial projections Identities of rhombomeres and their originate in single rhombomeric boundaries are also proposed to arise by compartment cooperate to form the locally controlled proliferation coupled functional branchiomotor nerve (Lumsden to cell adhesion changes (Wizenmann and Keynes, 1989). Segmentation refl ects and Lumsden, 1997), which results also to the time course of differentiation, to selective cell mixing. The in vitro where odd-numbered rhombomeres show affi nity experiments by Wizenmann and a delayed maturation compared to the Lumsden (1997) suggest that distinct even-numbered rhombomeres. In addition cell affinity properties restrict cell to Hox genes, the vitamin A derivative, mixing between adjacent rhombomeres. retinoic acid (RA), is an early embryonic Differential expression of cell adhesion signal that specifies rhombomeres. RA molecules is known to be an effective signaling has an essential role in specifying mechanism for cell sorting. Rhombomere- rhombomere identity by acting upstream boundary formation through cell sorting of Hox and other transcription factors, is thought to emerge by repulsive Eph- particularly in the posterior hindbrain ephrin signaling. For example, repulsive (reviewed by Gavalas and Krumlauf, interactions between Eph receptor tyrosine 2000; Glover et al., 2006; Maden, 2006). kinase, EphA4, and cells expressing its These TFs include, for example, Pax2, ephrinB ligands guide the cell sorting Meis1, Krox20 and Hoxb1. that underlies rhombomere-boundary Signaling by FGF8 is also crucial formation (Holder and Klein, 1999; in patterning the R1 and establishing the Mellitzer et al., 1999). Moreover, EphA4 anterior limit of Hox gene expression expression is showed to promote the intra (Irving and Mason, 2000). There is clear rhombomeric cell adhesion in parallel with

13 Review of the Literature the EphA4-dependent repulsion between FGF receptors (FGFR1-4) belong to rhombomeres (Cooke et al., 2005). the receptor tyrosine kinase family. The Apart from Eph-ephrin signaling little extracellular domain of FGFR molecules is known about the importance of other contains (i) a signal peptide, (ii) three (I- cell adhesion molecules to the hindbrain III) immunoglobulin-like (Ig) domains, segmentation. However, for example the (iii) an acidic residues of amino acids cadherin/catenin system might act as a between Ig-like domains I and II. and (iv) reciprocal regulator of proliferation and/or a heparin/heparin sulfate proteoglycans cell-cell adhesion in R1 (Trokovic et al., (HS) binding domain (Fig. 6A,B). The Ig- 2003) as well as in other rhombomeres like domains are needed for the FGF ligand where disruption of Ca2+-dependent binding. In addition, afﬁ nity for heparin/ adhesion abolishes rhombomere- HS is essential for effective activation of specific cell segregation (Wizenmann FGFRs (Ornitz, 2000). FGFRs also contain and Lumsden, 1997). Thus, differential a single transmembrane domain and a split expression of two cadherin molecules is intracellular tyrosine kinase domain. Upon important for restricting cell movement ligand binding, the receptor dimerizes between adjacent rhombomeres. and undergoes a conformational change which leads to trans-autophosphorylation of FGFR monomers as well as 3. Intercellular communication in the phosphorylation of the FGFR associated midbrain-R1 region molecule FRS2 (Fig. 6A,B). The use of an alternative exon 3.1. FGF signaling in the midbrain-R1 encoding the carboxyl terminal half of development Ig domain III is the major determinant of FGF ligand speciﬁ city and FGFR1-3 3.1.1. The FGF and FGFR molecules are present in both IIIb and IIIc isoforms (Fig. 6A,B; Chellaiah et al., 1999; Wang During embryonic development, FGFs et al., 1995). There are 22 Fgfs in the have diverse roles in regulating cell mouse genome (Zhang et al., 2006) which proliferation, migration and differentiation. of Fgf8, Fgf15, Fgf17 and Fgf18 are Most of the FGFs have a classical specifically expressed in the midbrain- amino-terminal signal sequence and are, R1 region during embryogenesis. FGF8, therefore, efficiently secreted via the 17 and 18 are subfamily members which endoplasmic reticulum-Golgi secretory exclusively bind FGFR IIIc splice forms pathway. FGF9/16/20 lack an obvious and FGFR4 in vitro (Fig. 6D; Zhang et amino-terminal signal peptide, but are al., 2006). By contrast, FGF15 subfamily nevertheless secreted. Mouse FGFs 11– shows relatively weak and unspecific 14 are thought to remain intracellular, activity for any FGFR1-3c isoform and secretion of these FGFs has not yet and FGFR4. As mice with Fgfr1IIIb been reported (Ornitz and Itoh, 2001). inactivation are phenotypically normal, Secreted FGFs are molecules which bind it has been proposed that the majority of their cognate receptors leading to signal the FGFR1 signaling is carried out by IIIc transduction in the receiving cells. The isoforms (Partanen et al., 1998).

14 Review of the Literature

3.1.2. Expression patterns of FGFs and R1 region. Expression of Fgf15 in the FGFRs in the mouse midbrain-R1 midbrain-R1 region is initiated at around E9.0 (Gimeno et al., 2002; Gimeno et al., The onset of Fgf8 expression at the 2003). Transcripts are detected at both midbrain-R1 region is at E8.5 (4-somite sides of the border between midbrain stage), slightly later than the Otx2/Gbx2 and R1. The negative gap between Fgf15 expression boundary is established (Wurst expression domain in midbrain-R1 region and Bally-Cuif, 2001). Until at least E12.5 is at least partially overlapping with the Fgf8 expression domain is detected as a Fgf8 expression domain (Gimeno et al., narrow band in the most anterior R1 (Fig. 2003) and it is detectable until E15.5. 7). The onset and expression of other The location of Fgfr transcripts at the Fgfs at the midbrain-R1 region is also midbrain-R1 region has been described well reported. Fgf17 and 18, belonging in several studies (Blak et al., 2005; to the Fgf8 subfamily, are expressed at Ishibashi and McMahon, 2002; Trokovic the midbrain-R1 region already at E8.5 et al., 2003; expression patterns at E9.5- (Xu et al., 2000). Compared to the Fgf8, E10.5 are summarized in Fig. 7 below). however, their expression domain is Fgfr1 expression is detected broadly narrower. By E10.5, the level of Fgf17 in the primitive streak and head folds at and Fgf18 expressions is quite similar E7.5 whereas Fgfr2-3 are more restricted. to Fgf8, but their expression is detected Between E8.5 and E10.5 Fgfr1 expression also in the most posterior midbrain is detected broadly in the neural tube and (see summarized expression in Fig. 7). at the border between mid- and hindbrain. Between E10.5 and E12.5 expression of Fgfr2 and Fgfr3 are also expressed Fgf17 and 18 is detected in the midbrain- broadly in the neural tube at these stages

Figure 6. Structure and interactions of FGF-FGFR molecules. Different isoforms of FGFR1-3 (A,B). Ligand and heparin/HS binding of FGFRs triggers the intracellular signal transductions through phosphorylation of tyrosine (Tyr) residues in the tyrosine kinase (TK) domain (C). Tyr653 and Tyr654 are needed for the catalytic activity of the activated receptor and subsequent signaling (Mohammadi et al., 1996). Altogether, there are seven Tyr residues in the cytoplasmic TK domain which of three are shown for simpliﬁ cation. Drawing showing the proposed (Zhang et al., 2006) receptor affinities by in vitro binding assay and subgrouping of FGF ligands accordingly (D).

15 Review of the Literature but a gap in expression at the midbrain-R1 3.1.3. FGFRs in the mouse midbrain-R1 region is detected (Blak et al., 2005; Fig. development 7). Interestingly, from at E9.5 onwards Fgfr2 expression in the ventral midbrain The inactivation of Fgfr1 or Fgfr2 results is extending posteriorly close to the in very early lethal phenotypes due to a midbrain-R1 border but do not overlap failure in gastrulation (Arman et al., 1998; with the Fgf8 expression (Blak et al., Deng et al., 1994). By contrast, Fgfr3 2005). Fgfr4 expression has not detected KO mice are viable, but exhibit skeletal at the midbrain-R1 region between E8.5 and inner ear defects leading to a bone and E12.5 (Blak et al., 2005). overgrowth and deafness (Colvin et al., Although earlier in vitro binding 1996; Deng et al., 1996). Because other assays (Zhang et al., 2006) have receptors can compensate for the loss of proposed only a weak affinity of FGF8 FGFR3, the role of FGFR3 in neurogenesis to FGFR1c (Fig. 6D), the interaction was recently studied using the gain-of- of these molecules during midbrain-R1 function approach (Inglis-Broadgate et development was recently studied using al., 2005). In these constitutively active surface plasmon resonance (SPR) assays Fgfr3 knockout mice, the regulation of (Olsen et al., 2006). The SPR analysis proliferation and apoptosis of cortical showed that the FGF8b splice form is progenitors was defective which led to the capable to bind to FGFR1c isoform, increased brain size. which is consistent with genetic analyses Conditional knockout studies of Fgfr1 proposing that FGF8 signals through using En1-Cre and Wnt1-Cre mice have FGFR1 during midbrain-R1 development shown its importance in the development (Chi et al., 2003; Trokovic et al., 2003). of the midbrain-R1 region (Trokovic et al.,

Figure 7. Summarized gene expression patterns of Fgfs and Fgfrs at E9.5-E11.5.

16 Review of the Literature

2003). Conditional inactivation of Fgfr1 clarify the exact role of FGFR signaling in in the midbrain-R1 region using an En1- the midbrain-R1 development. Cre mouse (Kimmel et al., 2000) results in a deletion of the inferior colliculi and 3.1.4. FGFR signal transduction vermis of the cerebellum in the dorsal midbrain-R1 region. By contrast, TH- The phosphorylation events in FGFR positive neurons were found to be present tyrosine kinase domain initiate various in newborn and adult conditional Fgfr1 downstream cascades, such as RAS- mutants, both in the midbrain DA neurons mitogen activated protein kinase (RAS- and in the locus coeruleus (LC) of the MAPK), phosphatidylinositol 3’ kinase R1. In addition, ventral structures such (PI3K/Akt) or phospholipase C gamma 1 as pontine nuclei as well as mesopontine (PLCγ1/PKC) pathways (Mohammadi et nuclei were unaffected in conditional al., 1991; Mohammadi et al., 1996; Ong Fgfr1 mutants. Although TH-positive et al., 2001). These different pathways are cells were present in the LC domain more dependent of docking or adaptor proteins detailed analysis revealed that the LC with distinct Src-homology 2 (SH2) appeared to be disorganized in newborn domains (Ryan et al., 1998). Consequently, Fgfr1 MHB-specifi c mutants. Conditional the recruitment of different types of SH2 Fgfr1 MHB-specific mutants survived domain proteins triggers a specifi c network until adulthood, but they showed impaired of complex signal transduction cascades. motor coordination. The phenotype of Upon activation of a FGFR, the conditional Fgfr1 mutants is clearly phosphorylation of tyrosine residues less severe than the phenotype of the promotes binding of the FRS2 and recruits conditional Fgf8 mutants produced with the GRB2 and SOS adaptor proteins to the same En1-Cre mouse line (Chi et al., the complex (Fig. 8). As a result, SOS 2003). Inactivation of Fgf8 in the MHB catalyzes the conversion of RAS from led to the extensive cell death in the an inactive GDP bound to an active midbrain-R1 region between E8.5 and GTP bound form. RAS then triggers the E10.0. Consequently, the ectopic apoptosis MAPK cascade by recruiting the RAF of MHB cellular progenitors caused the to the membrane and activates it. The deletion of the entire midbrain, isthmus kinase activity of RAS phosphorylates and and cerebellum at E17.5. activates the dual-specifi city kinase MEK, The conditional allele of Fgfr2 was which in turn, phosphorylates the ERK. also recently generated and the effect Phosphorylated forms of ERK molecules of FGFR2 or FGFR3 deficiency on the (pERK1 and 2) can then enter the nucleus development of midbrain-R1 region was where it activates transcription factors, subsequently studied. Interestingly, the including members of the ETS family of analysis of midbrain-R1 specific Fgfr2 transcription factors. mutant mice and conventional Fgfr3 RAS-MAPK pathway is activated by mutant mice showed that both receptors various ligand-receptor systems, including alone are dispensable for the formation RTKs and G-protein-coupled receptors. of a normal phenotype and maintenance Thus, it is striking that FGF signaling of the midbrain-R1 region (Blak et al., seems to be responsible for nearly all 2007). Thus, further analyses of FGFR1- ERK1/2 activity in early frog and fish 3 double and triple mutants are needed to embryos (Christen and Slack, 1999; Tsang

17 Review of the Literature and Dawid, 2004). In addition, expression (siRNA) induced apoptosis in the of activated pERK has been observed in mesenchyme (Kawakami et al., 2003). distinct domains in the mouse neural tube, In the mouse IsO explants the regulation predominantly in the midbrain-R1 region of Mkp3 expression by FGF seems to be overlapping with the Fgf8/Fgf17/Fgf18 induced by PI3K/Akt pathway (Echevarria expression domains (Corson et al., 2003). et al., 2005). These studies suggest that PI3K/Akt pathway also plays roles MKP3 has a role in mediating the anti- in FGF induced cell proliferation and apoptotic signaling through FGF8 induced cell survival. It has been shown that PI3K/Akt pathway. FGF8 induced activation of the PI3K/Akt The PLCγ1/PKC pathway involves pathway results to expression of Mkp3 binding of to phosphorylated tyrosine 766 which has a negative feedback action on residue of FGFR1 (Peters et al., 1992). the RAS-MAPK pathway (Echevarria et Upon binding and activation through al., 2005; Kawakami et al., 2003). During phosphorylation, PLCγ hydrolyzes mouse limb development downregulation phosphatidylinositol-4,5-diphosphate to of Mkp3 by small interfering RNA form two second messengers, inositol-

Figure 8. FGF-FGFR signal transduction pathway. See text for details. Modiﬁ ed from Jukkola et al. (2006).

18 Review of the Literature

1,4,5-triphosphate and diacylglycerol. 5) are thought to be downstream targets of Diacylglycerol is an activator of protein FGF-FGFR signaling (Liu et al., 2003b; kinase C (PKC), whereas inositol-1,4,5- Raible and Brand, 2001; Tsang and Dawid, triphosphate stimulates the release of 2004). The action of these transcription intracellular Ca2+. In mouse, FGFR1 alleles factors is hypothesized to promote the with a single point mutation at residue 766 FGF signaling (e.g. Fgf8 expression) by replacing tyrosine with phenylalanine activating as a positive feedback loop (Y766Fl) has been generated (Partanen (Tsang and Dawid, 2004). The interactions et al., 1998; Peters et al., 1992). It was of ETS transcription factors with the key shown that this mutant receptor failed transcriptional factors such as AP1, SRF to bind PLCγ and the hydrolyzation of and PAX family members further diversify phosphatidylinositol as well as Ca2+ the output of FGF signal transduction, mobilization was abolished after FGF although specific interactions between stimulation (Peters et al., 1992). However, PEA3 subfamily (PEA3, ERM, ER81) mice homozygous for the Y766Fl allele and other key TFs has not been reported were viable and fertile and did not show (reviewed by Sharrocks, 2001). Members obvious brain or craniofacial defects of the PEA3 subfamily have been linked (Partanen et al., 1998). Analysis of these with the specification of neuronal cells mouse mutants showed that PLCγ1/PKC (Arber et al., 2000; Roehl and Nusslein- pathway through FGFR1 activation is Volhard, 2001). Pea3 or Er81 knockout not required for normal pre- or postnatal mice displayed no overt brain phenotypes. development or viability. In Pea3 -/- mice, however, a male sexual A set of genes known to be induced dysfunction was observed, which probably in regions of active Fgf8 signaling and arises owing to neuronal defects (Laing et thought to either transduce or modulate al., 2000). By contrast, inactivation of Er81 the FGF signaling pathway is termed as leads to an improper connection between Fgf8 synexpression group. Among the a subset of sensory and motor neurons Fgf8 synexpression group members are (Arber et al., 2000). One explanation for the transcription factors Pea3 and Erm, these relatively mild phenotypes is that and the inhibitors Mkp3, Sef, Spry1 and there is redundancy between the members Spry2 (Fig. 8; see below). In addition to of the PEA3 subfamily. In the midbrain- being potential transcriptional targets of R1 region, Erm expression is already FGF signaling, PEA3 and ERM function detected at E8.5 whereas the onset of as transcriptional effectors within cells Pea3 expression is around E9.0 (Chotteau- to transduce FGFR-dependent signals. Lelievre et al., 2001). Between E9.5 and By contrast, MKP3, SEF and SPRYs E12.5, Erm and Pea3 expression is seen act within cells as negative feedback across the border between midbrain and inhibitors, modulating and restricting the R1. During development, Er81 expression levels and extent of FGF-FGFR signaling. has not been reported in the midbrain-R1 In RAS-MAPK signaling cascade, region. members of the ETS family are the key Members of the Sprouty (SPRY1- link between the FGF-FGFR signaling 4), MAPK phosphatase (MKP) and SEF and transcriptional regulation. PEA3 (also families are negative modulators of FGF known ETV4, Ets variant gene 4) and signaling (Fig. 8). SPRY is thought to ERM (also known ETV5, Ets variant gene inhibit FGF signaling by sequestering

19 Review of the Literature

GRB2 and preventing its binding to transmembrane proteins. The intracellular FRS2 (Hanafusa et al., 2002; Minowada domain of SEF is required for the et al., 1999), whereas another inhibitory interaction with FGFR1 and FGFR2 and mechanism involves direct interaction of subsequent inhibition of FRS2 (Tsang et SPRY2 with RAF (Yusoff et al., 2002). al., 2002). Studies in zebrafish, mouse Only Spry1 and Spry2 are expressed at the and human have shown that SEF inhibits midbrain-hindbrain region between E8.5 FGF-induced activation of RAS-MAPK and E10.5 (Minowada et al., 1999). The pathway, and mouse SEF also inhibits similarity of these expression domains FGF-induced activation of protein to Fgf8 expression in various regions kinase B (pkB/Akt), a key protein in the suggested that they are coordinately PI3K pathway (Kovalenko et al., 2003; regulated. Moreover, overexpression (FGF Yang et al., 2003). There are at least two beads or Spry retrovirus infections) and alternatively spliced isoforms of SEF in gene inactivation studies in mouse, chick humans which are not reported in mice. and zebrafish confirmed that vertebrate SEF-a is an isoform that can interact with SPRYs are functionally conserved and the receptors whereas another isoform, act as FGF-induced feedback inhibitors SEF-b, lacks the signal peptide and (Furthauer et al., 2001; Mailleux et al., therefore it is cytoplasmic and suppresses 2001; Minowada et al., 1999). RAS-MAPK pathway downstream of More recently, another negative MEK (Preger et al., 2004). Between E8.5 feedback modulator of FGF-FGFR and E12.5 Sef expression at the midbrain- signaling, named MAPK phosphatase-3 R1 regions overlaps with the Fgf8, (also known MKP3 or DUSP6) has been although its expression domain is broader described in vertebrate limb buds (Eblaghie (Lin et al., 2002). et al., 2003; Kawakami et al., 2003). FLRT3 together with FLRT1 and MKP3 specifi cally dephosphorylates and FLRT2, compose a novel gene family inactivates ERK1/2. In the midbrain- which were fi rst isolated in a screen for R1 region, Mkp3 is expressed already at extracellular matrix proteins expressed in E8.5 in a region overlapping with Fgf8 muscle cells (Lacy et al., 1999). The three expression domain (Echevarria et al., genes encode a highly conserved family of 2005). Later (E9.5-E10.5), when Fgf8 glycosylated proteins (FLRTs) containing expression is defi ned to the most anterior a fi bronectin III domain and leucine rich R1, expression of Mkp3 is seen at both repeats (LRR) domain. Extracellular LRR sides of the border between midbrain repeats are commonly involved in protein- and R1. By siRNA knockdown analysis, protein interactions. In particular, LRRs Echevarria et al. (2005) showed that in are common motifs in the extracellular the midbrain-R1 region the negative region of transmembrane proteins and feedback mechanism of FGFR signaling is also in secreted proteins that are involved modulated by Mkp3 activity. In addition, in ligand–receptor interactions or in cell they proposed that the expression Mkp3 adhesion. The former was supported in the midbrain-R1 region is induced by by in vitro protein-protein interaction PI3K pathway. studies, both in Xenopus laevis (Bottcher SEF (similar expression to Fgfs; et al., 2004) and in mouse (Haines et or Il17rd) and FLRT3 are conserved in al., 2006), suggesting that members of vertebrates and both encode single-pass the FLRT family can directly interact

20 Review of the Literature with the different FGFRs. In addition, essential for normal development of the overexpression studies with Xenopus Flrt3 zebrafi sh tectum and cerebellum. have suggested that FLRTs may promote Constant proliferation of FGFR signaling in vertebrates (Bottcher et neuroepithelial cells in the midbrain-R1 al., 2004). In Xenopus, the role of FLRT3 region causes progressive growth of caudal in promoting FGFR signaling was further midbrain and rostral hindbrain areas, hypothesized by the overlapping Flrt3 and respectively. However, the expression of Fgf8 expression patterns, particularly in the Wnt1 and Fgf8 is constantly restricted midbrain-R1 region (Bottcher et al., 2004). to a band of only a few cell-diameters Recent data suggest, that FLRTs may also wide (e.g. app. 10 cell-layers of Fgf8 mediate homophilic cell adhesion and cell expressing cells in E10.5) at both sides sorting (Haines et al., 2006; Karaulanov of the border between midbrain and R1. et al., 2006). Thus, FLRTs may represent Interestingly, the expression of MAPK yet another family of cell-surface proteins pathway inhibitors such as Mkp3 as well that link cellular adhesion and signaling as FGFR signal transduction activators through transmembrane receptor. such as Cnpy1 is detected in a broader CANOPY1 (CNPY1) is the latest midbrain-R1 region compared to Fgf8. It member of the Fgf8 synexpression group. is thought that the extracellular spreading It was found in a screen for zebrafish of FGF8 in a tissue is due to a diffusion- genes expressed at the isthmus in response based mechanism controlled by an active to signaling between the midbrain and endocytosis of the target cells (Scholpp the isthmus (Hirate and Okamoto, 2006). and Brand, 2004). In these analyses, Cnpy1 was specifi cally expressed in the spreading of FGF8 protein labeled in vitro zebrafi sh midbrain-R1 region. Sequence with Cy3 fl uorophore from a local source analysis revealed that Cnpy1 is one was monitored to determine how the of four structurally related proteins molecule spreads through a target tissue. (CNPY1-4) conserved from mouse to Reducing the rate of endocytosis led to an human. Among Canopy family proteins, extracellular accumulation of FGF8 and only Cnpy1 is specifically expressed resulted in broader domains of FGF-target in the midbrain-R1 region of zebrafish gene expression, suggesting an increased embryos. Human CNPY1 is localized spreading of FGF8. Thus, different levels in the endoplasmic reticulum and it may of endocytosis allow different extent of interact with the extracellular domain spreading of FGF8 protein through the of the FGFR1. CNPY1 has structural target tissue, leading to differential target similarity to saposins, proteins that bind gene response. However, the endogenous sphingolipids, and it was speculated that graded activity of FGF8 or other FGF CNPY1 could infl uence association of the proteins has not been directly visualized receptor with lipid rafts. It was also shown around the IsO. that implantation of FGF8b-soaked beads in the midbrain of zebrafish embryos 3.2. WNT signaling in the midbrain-R1 increased expression of Cnpy1 (Hirate and development Okamoto, 2006). In addition, experiments in which expression of CNPY1 protein In addition to FGF-FGFR signaling was blocked with antisense morpholino WNT signaling has an important role in oligonucleotides showed that Cnpy1 was early embryonic patterning through the

21 Review of the Literature regulation of differentiation, cell death, calcium pathway. The patterning of the tissue polarity and migration. WNTs midbrain-R1 region by WNTs is probably are secreted glycoproteins which can through canonical pathway, as specific inﬂ uence tissue organization and growth inactivation of β-catenin in the MHB by functioning locally or on adjacent cells. mimics the Wnt1 -/- phenotype (Brault et Moreover, they can also act at a distance al., 2001). In the WNT/β-catenin pathway by generating a gradient across a several the activation of DVL leads to the cell layers. Also like FGFs, WNT proteins inactivation of glycogen synthase kinase act on cell surface by binding to receptors 3β (GSK3β) in the axin-based complex. called frizzleds (FZ) which are members This in turn stabilizes the β-catenin and of the seven transmembrane domain cell allows its relocation to the nucleus where surface receptors (Fig. 9). Interactions it promotes the transcription by interacting with complexes between FZ and the LDL with members of the TCF family of receptor-related protein (LRP) molecules HMG box containing proteins. LRPs have been shown to mediate the signaling are especially needed for the canonical activities of WNTs. The WNT-FZ-LRP pathway (Pinson et al., 2000). receptor complex activates dishevelled The most important WNT molecule (DVL) in the cytoplasm. Downstream of for midbrain-R1 development is found to DVL, the WNT pathway diverges into be WNT1 (McMahon and Bradley, 1990; at least three branches: the canonical or Thomas et al., 1991). Although WNT1 WNT/β-catenin pathway, the planar cell does not affect the induction of Fgf8 it polarity (PCP) pathway and the WNT/ is required for the maintenance of FGF8

Figure 9. Canonical WNT pathway. Low density lipoprotein (LDL) receptor-related proteins (LRP5 and 6 in mouse) are required signaling co-receptors for the canonical Wnt pathway (Pinson et al. 2000). Secreted WNT signaling antagonists such as Dickkopf 1 (DKK1) can interact with LRP5/6 proteins. Other secreted antagonists such as Cerberus or Frizzled-related proteins (e.g. FRZB1 or SFRP2) can bind WNTs directly thus presumably inhibiting WNT- FZ interactions. Activation of DVL induces the disassembly of a complex consisting of AXIN, adenomatosis polysis coli (APC), GSK3β and β-catenin. In the absence of WNT activity, the GSK3β of this complex phosphorylates β-catenin, thereby triggering its degradation. Modified from Bouwmeester et al. (1996), Houart et al. (2002) and Leyns et al. (1997).

22 Review of the Literature expression, possibly through EN1. At E8.0 4. Cell proliferation in the developing (0 somite stage), after the onset of Otx2 brain and Gbx2 expression Wnt1 is detected in a broad domain. Initially, Wnt1 expression In order to generate appropriate number of overlaps partially with the Gbx2 domain, neurons and glia in the correct spatial and but becomes restricted to the Otx2-positive temporal patterns, it is critical to regulate domain in the midbrain at 6 somite stage the number of progenitor cells. In addition (Bally-Cuif et al., 1995). Subsequently, the timing of cell differentiation (i.e. the Wnt1 expression is restricted to a narrow exit from the cell cycle) is crucial for band and its expression at the midbrain-R1 the size, shape and histogenesis of many region is maintained at least until E14.5. tissues, including the brain. A number of β-catenin-TCF target The initial proliferation of neural genes have been identified in diverse progenitors in developing CNS is limited biological systems and it is becoming to the ventricular zone (VZ) that lines the clear that the majority of WNT target lumen of the neural tube. The majority of genes are tissue specific (Ciani and neural progenitors give rise to postmitotic Salinas, 2005). Similarly to FGF-FGFR neurons by asymmetric divisions that signal transduction, numerous WNT target migrate into more apical zones to form genes appear to be positive and negative marginal layers. Although mitotic nuclei regulators of the pathway. These include dynamically change their vertical location positive regulators, such as FZ receptors, within the VZ coincidentally with the LRPs, and TCF/LEF and negative progression of the cell cycle, neural regulators, such as AXIN2 and SFRPs, progenitors maintain a pseudo-stratifi ed which all are activated by the β-catenin- columnar epithelium by linking with each TCF transcription machineries. The best other by adherens junctions. candidates for general downstream targets of β-catenin-TCF are AXIN2 and SP5 4.1. Interkinetic nuclear migration which are expressed in various tissues in vertebrates (Jho et al., 2002; Weidinger et The adult brain is composed of neurons al., 2005). In frog, En2 is a direct target and glia, defi ned by cellular morphology, of WNTs at the midbrain-R1 boundary cell density and the formation of distinct region, but this has not been demonstrated neurites. During development, different in mouse (McGrew et al., 1999). types of cells are generated in the The role of other WNTs in the pseudostratifi ed epithelium that forms the development of midbrain-R1 region has ventricular zone (VZ), a layer of dividing not been established in the mouse. In progenitor cells at the ventricular surfaces zebrafi sh, knockdown studies have shown of the developing neural tube (Bayer et that Wnt1 and Wnt10b provide partially al., 1991; Sauer 1935 in Developmental redundant functions in the development Neurobiology, Jacobson & Rao Eds. of midbrain-R1 region (Buckles et al., 2005). As progenitor cells progress 2004; Lekven et al., 2003). There are also through the cell cycle their nuclei undergo indications that zebrafi sh Wnt8 is required dynamic intracellular migration, called for the early positioning of the IsO (Rhinn interkinetic nuclear migration (Sauer et al., 2005). 1935). During this process (Fig. 10), nuclei move away from the apical surface

23 Review of the Literature

(ventricular side) during G1, reside in the for the balance between cell cycling and basal half of the VZ (pial side) during differentiation. It has been hypothesized S phase and return apically in G2 phase that the self-renewing progenitor cells are so that mitosis occurs at the VZ (Bayer dividing symmetrically (Fig. 10A), in a and Altman, 1991; Sauer 1935). After plane roughly parallel to the ventricular mitosis, a portion of the dividing cells exit surface to produce two identical daughter the cell cycle (post mitotic or quiescent/ cells thus expanding the amount of G0 phase), whereas a complementary stem cells or progenitors (proliferative portion re-enters S phase and sustains pool). By contrast, the progress of the proliferative pool. Thus after a cell neurogenesis (i.e. differentiation), occurs division, a daughter cell can either repeat asymmetrically (Fig. 10B) in a plane or exit the cell cycle. The degree to which roughly perpendicular to the ventricular mitotic cleavage and interkinetic nuclear surface. These self-renewing divisions migration prevails provide a mechanism generate a daughter stem-cell/progenitor

Figure 10. Symmetrical versus asymmetrical cell division. The cells of the neuroepithelium widen the width of the neural tube by dividing symmetrically in the plane of the ventricular surface (yellow arrow in A). Divisions perpendicular to the plane of the epithelium are thought to be asymmetric (red arrow in B) and generate a neural cell or quiescent cell (G0 phase). Modiﬁ ed from the text book: Developmental Neurobiology, Jacobson & Rao Eds. 2005.

24 Review of the Literature and a more differentiated cell type CDKs, either directly or indirectly control such as a non-stem-cell progenitor or the major cell-cycle transitions and phases a neuron (Gotz and Huttner, 2005). As in all eukaryotic organisms. Different neuronal production proceeds during cyclin/CDK interactions are required the development, asymmetrical division at different phases of the cell cycle. The starts to predominate (after E12.5 in association of cyclins with CDKs results mouse). However, toward the end of the to phosphorylation of retinoblastoma neurogenesis, non-stem-cell progenitors susceptibility protein (pRb) and revert back to symmetrical divisions which progression to the next phase of the cell start to predominate, but produce two cycle through transcriptional activation of identical terminally differentiated neurons target genes (Deng et al., 1995; Shtutman (green cell in Fig 10A). et al., 1999). In mammalian cells, there are two classes of CDKs that function at 4.2. Cell cycle regulators the G1/S phase transition (Sherr, 1994). CDK4 and its close relative CDK6 are 4.2.1. Cyclins, Cdks and their inhibitors driven by three D-type cyclins (CyclinD1- D3). By contrast, CyclinE activates The genes that control the number of speciﬁ cally CDK2. CyclinE accumulates cell divisions in the vertebrate CNS are very close to the G1/S phase transition beginning to be understood. Cyclins are and has a role in phosphorylating pRb a group of proteins that show ‘cyclic’ (Ciemerych and Sicinski, 2005). CyclinE changes in their expression levels that expression is more periodic than that of correlate with specific phases of the D-type cyclins (Roberts and Sherr, 2003). cell cycle (Fig. 11A). Cyclins bind their CyclinA also activates CDK2 that plays catalytic partners, cyclin-dependent role in the S phase progression. B-type kinases (CDKs), and it is now clear that cyclins (CyclinB1-B3 in mouse) as well as

Figure 11. The phases of cell cycle and cell cycle gene expression. (A) The diagram shows the period in the cell cycle when each of the cyclins is expressed. Cdks are expressed more or less throughout the cell cycle, but their activation (phosphorylation) is phase speciﬁ c. A cell proliferation marker protein, KI67, distributes to the chromosome periphery during mitosis and nucleolar heterochromatin in the interphase. (B) Regulation of CyclinD-CDK4/6 complex. The association of CyclinDs with CDKs depends, for example, on the Cip/Kip family of CKIs. It is also negatively regulated by INK4 inhibitors. Once assembled CyclinD-CDK4/6 complexes are subjected to inhibition by Cip/kip inhibitors such as p21 and p27. Adapted from Roberts (1999).

25 Review of the Literature

CyclinA together with their kinase partner al., 1999). The INK4 family is composed CDK1 (also called CDC2), represent the of four members: p15, p16, p18 and p19 important controllers of the G2/M phase and they specifically inhibit the CDK4/ progression (Ciemerych and Sicinski, CDK6 activity (Sherr and Roberts, 1999). 2005). CyclinD1 (also called CYCD1 or 4.2.2. Cyclins and Cdks in neuronal CCND1 in MGI) together with associated development kinases (CDK4 and CDK6) phosphorylate the substrate proteins (e.g. retinoblastoma During mouse development D-type cyclins protein, Rb) that lead to release of E2F are expressed in a tissue-specific, but transcription factor. For example pRB greatly overlapping manner (Ciemerych et phosphorylation by CyclinD1/CDK4 al., 2002; Wianny et al., 1998). At E8.5, complex is necessary for progression CyclinD1 is expressed broadly in the from the G1 phase to S phase (DNA midbrain-R1 region whereas CyclinD2 synthesis). In contrast to other cyclins, expression is absent from rhombomeres which are induced periodically during 1 and 2. Later between E9.5 and E11.5 cell cycle progression, the expression of CyclinD1 and CyclinD2 expression is D-type cyclins is controlled largely by the detected strongly at the midbrain-hindbrain extracellular environment (reviewed in region in an overlapping pattern. Mice Roberts, 1999). lacking the individual cyclins (CyclinD1, Regulation of cyclin/CDKs D2 or D3) have been generated and the complexes occurs at multiple levels, observed defects in neurogenesis were including assembly of cyclin and relatively mild. In summary, CyclinD1 CDK subunits, phosphorylation and deﬁ cient mice showed reduced body size dephosphorylation events and association and neurological abnormalities caused of cyclin/CDK complexes by inhibitory by an unknown mechanism (Fantl et al., proteins (Fig. 11B). These regulators 1995; Sicinski et al., 1995). By contrast, include CAKS (CDK activating kinases) CyclinD2 KOs exhibited cerebellar and phosphatases, such as Drosophila and defects which were showed to be due to vertebrate orthologues string and CDC25, a decreased proliferation of granule cell respectively (Kakizuka et al., 1992). In precursors (Huard et al., 1999). In addition, mouse, the activity of one of three CDC25 the adult CyclinD2 knockout mice lack tyrosine phosphatases on CDK1 results in newly born neurons in the dentate gyrus of full activation of CyclinB-CDK1 complex the hippocampus and in the olfactory bulb and cell’s entry into mitosis (Galaktionov (Kowalczyk et al., 2004). Neurological and Beach, 1991). In addition, there defects were not reported in CyclinD3 are number of CDK inhibitors (cyclin- knockouts (Sicinska et al., 2003). Mice dependent kinases inhibitors, CKIs), which lacking the combinations of the D-type bind to CDKs and block their catalytic cyclins (double and triple KOs) represented activity (Sherr and Roberts, 1999). They the sum of the phenotypes observed in the are grouped to two classes called the Cip/ individual CyclinD mutants (Kozar et al., Kip family and the INK4 family. The Cip/ 2004). Thus, these analyses suggested Kip family is composed of three members that the proliferation of several neuronal p21, p27 and p57 which can bind to and cell types could take place in a CyclinD- inhibit all known CDKs (Deng et al., independent fashion (Kozar et al., 2004). 1995; Kukoski et al., 2003; Shtutman et

26 Review of the Literature

The role of other type of cyclins some of the molecules controlling these such as CyclinA or CyclinB in neuronal cellular events, but how the different development has not been reported. These aspects of neurogenesis are integrated cyclins together with their kinase partner into a coherent developmental program CDK1, represents the major regulators remains unclear. One possible mechanism of the G2/M phase progression. Both implicates multifunctional proteins that CyclinA2 and CyclinB1 knockout mice regulate both cell cycle exit and cell showed an early embryonic lethality differentiation. The expression of p21 (Brandeis et al., 1998; Murphy et al., during mouse development correlates 1997) whereas CyclinA1 or CyclinB2 with terminally differentiating cells such mutants were viable, but no overt CNS as olfactory neurons (Deng et al., 1995; phenotypes were observed (Brandeis et al., Parker et al., 1995). However, the fact 1998; Liu et al., 1998). The expression of that p21 knockout mice appears to have A- and B-type of cyclins during the early normal brain development, indicates that brain development has not been reported. any essential role for p21 in terminal In conventional mutant mice defi cient differentiation must be redundant (Deng for the Rb protein, progenitor proliferation et al., 1995). In the other hand, p27 in the CNS is profoundly deregulated, deletion in mice causes a phenotype resulting in excess dividing cells localized opposite of that of the CyclinD1/D2 to normally post-mitotic regions (Clarke double knockout, which shows reduced et al., 1992). These mice lacking Rb die body size and hypoplastic cerebella. In around E15.5 due to a severe anemia. p27 KO animals, the gene inactivation More recently, Wu et al. (2003) found that causes an overproduction of neuronal the failure of the placental development cells (Kiyokawa et al., 1996; Nguyen et caused the early lethality. To circumvent al., 2006). p27 has recently been shown the placental abnormalities they produced to function beyond cell cycle regulation rescued Rb-defi cient mice with a wild-type and promote both neuronal migration placentas using tetraploid complementation and differentiation of newborn cortical technique. Interestingly, no developmental neurons, through distinct and separable defects in the CNS were observed in these mechanisms. A report by Nguyen et al. rescued mice (de Bruin et al., 2003; Wu et (2006) shows that cytoplasmic p27 protein al., 2003). promotes neuronal migration by direct blocking of the small GTPase RhoA and 4.2.3. Cdk inhibitors in neuronal its downstream pathway. Morover, p27 can development act as a Neurogenin 2 (NGN2) stabilizer and promotes neuronal differentiation. Critical negative regulators of the cyclin/ Little is known about the regulation CDKs p21 are p27 are expressed in the of cell cycle inhibitors by transcription developing and adult nervous system in the factors. In vitro overexpression of basic differentiating neurons shortly after they HLH genes such as NeuroD in the P19 have withdrawn from the mitotic cycle or HeLa cell lines can induce p27 or p21, (Parker et al., 1995). The development of respectively (Farah et al., 2000; Mutoh neurons in the CNS requires that progenitor et al., 1998). The p21 promoter sequence cells leave the cell cycle and activate contains multiple bHLH activator-binding specifi c programs of differentiation and sites (E-boxes) which have been shown migration. Genetic studies have identifi ed to be functional in the up regulation

27 Review of the Literature of p21 (Prabhu et al., 1997). Elevated premature expression of neural-specific p21 expression and ectopic mitosis are markers, such as TUJ1 (Estivill-Torrus observed also in the enteric epithelium et al., 2002). Reduction of the cell cycle of NeuroD-knockout mice (Mutoh et al., length by decreasing G1 phase increases 1998). In addition, the basic helix-loop- the likelihood of cell cycle re-entry helix genes Hes1 which is an essential and might expand the pool of neuronal effector for Notch signaling can repress progenitor cells. For example, the transition expression of the p21 in vitro (Castella et of neuroepithelial to radial glial cells and al., 2000; Kabos et al., 2002). their progression from proliferative to differentiating divisions during embryonic 4.3. Mechanisms controlling the cell development is associated with an increase polarity and asymmetric cleavage in the length of their cell cycle (Cappello et al., 2006). Remarkably, this increase in The mechanisms that regulate the cell-cycle length is predominantly due to a interkinetic nuclear migration and the lengthening of the G1 phase. transition from predominantly symmetrical In addition to the apico-basal to predominantly asymmetrical divisions polarity, a plane parallel to the plane of are poorly understood. Apico-basal neuroepithelium defi nes planar polarity, polarity is an essential characteristic for which plays a role in the development of interkinetic nuclear migration. Recently, pattern across a fi eld of cells. The planar the role of small GTPase CDC42 in the polarity and signals may also contribute maintenance of apico-basal polarity and to cell cycle progress in adjacent brain self-renewal was showed in the cerebral compartments (Guthrie et al., 1991). cortex (Cappello et al., 2006). Conditional inactivation of Cdc42 at different 4.4. Cellular proliferation and developmental stages demonstrated that regulation in the rhombomeric Cdc42 is needed for the apically directed boundaries interkinetic nuclear migration. In addition to polarity, other cellular mechanisms such It is suggested that the highest level of as cell cycle length, orientation of mitosis proliferation is detected at rhombomeres and basement membrane composition whereas the inter rhombomeric boundaries are thought to regulate the decisions show reduced rates of proliferation/ between progenitor cell proliferation cell cycle progression (Guthrie et al., and differentiation. Clonal analysis on 1991). Guthrie et al. (1991) showed oligodendrocyte progenitor cells suggested that at rhombomere boundaries DNA that an intrinsic clock operates within each synthesis (S-phase) occurs closer to the cell to assist control when the cell stops ventricular surface compared to cell dividing and differentiates (Temple and within rhombomeres where S-phase cells Raff, 1986). Thus, a similar “cellular” were detected more basally. The molecular clock, which can be infl uenced by extrinsic mechanisms and regulators responsible for signals, may act to control the balance these changes in nuclear migration patterns between the cell cycle progression and in the rhombomeric boundaries are differentiation also in neural progenitors. currently unknown. Later in development, For example, at the start of cortical the proliferative cells of ventricular zone neurogenesis Pax6 inactivation resulted are restricting the cell mixing between in the increased cell cycle length and the rhombomeres after evident hindbrain

28 Review of the Literature segmentation disappears (Wingate and signaling by knockdown of zebrafi sh Wnt1 Lumsden, 1996). or Tcf3b results to deregulation of boundary Cadherins mediate cell adhesion regions (Riley et al., 2004) and ectopic and play a fundamental role in normal expression of boundary cell markers, development. In the neuroepithelium, they Rfng (radical fringe) and Foxb1.2, in participate in the maintenance of proper non-boundary regions of the hindbrain cell-cell contacts. Cadherins are associated (Amoyel et al., 2005). In the latter, authors with the actin cytoskeleton through β- and showed that the Wnt1-positive cells α-catenins to maintain adherens junctions expressed in the rhombomeric boundary between cells (Nelson and Nusse, 2004). and roof plate contributed to upregulation The canonical WNT pathway acts also of proneural and Delta gene expression. at the level of transcription, wherein the As Delta cell-autonomously blocks Notch WNT signal stabilizes and increases β- activation, this pathway mediates lateral catenin in the cytoplasm and nuclei. inhibition that prevents spreading of β-catenin regulates the expression the Wnt1 boundary expression domain. of numerous target genes, including Thus, boundary cells play a key role in CyclinD1 that progresses through the G1- the regulation of cell differentiation in the to-S-phase cell-cycle transition (Shtutman zebrafi sh hindbrain (Amoyel et al., 2005). et al., 1999). Because cadherins sequester These data suggest that WNT and Notch- β-catenin into the adhesion complex on Delta signaling regulate neurogenesis the cell surface, it is thought to have dual in non-boundary regions, which in regulatory roles there, acting positively turn blocks ectopic boundary marker on cell adhesiveness and negatively on expression. Similar feedback mechanisms the WNT canonical pathway, which can operate in the Drosophila wing disc and reduce the progression of the cell cycle vertebrate limb bud, suggesting adaptation (Nelson and Nusse, 2004). of a conserved signaling module that In rhombomeres, CyclinD1 and spatially organizes cells in complex organ CyclinD2 show distinct segments specifi c systems. In the mouse, however, only the expression patterns, which are temporally midbrain and cerebellum is deleted in mice regulated along the AP axis (Wianny mutant for Wnt1 (McMahon and Bradley, et al., 1998). The rhombomere specific 1990; Thomas et al., 1991) suggesting expression of CyclinD1/2 could be the one redundancy between WNT signals. mechanism to regulate the length of the These results obtained from the G1 phase in rhombomeres. Interestingly, rhombomeric boundaries contribute to restricted expression of CyclinD1 and our understanding of how the cell cycle CyclinD2 are detected before the formation control can be linked to the patterning of inter-rhombomeric boundaries (3-5 programs to infl uence the balance between somite stage). proliferation and neuronal differentiation Several Wnt genes are expressed in in discrete progenitor domains. An developing hindbrain. For example, Wnt1 analogous situation to rhombomeric and Wnt3a are expressed in the zebrafi sh boundaries occurs in the IsO in which a hindbrain. In addition, the zebrafi sh Wnt signaling center expressing Fgf8 and Wnt1 genes (Wnt1, Wnt3a, Wnt8b, and Wnt10b) in the adjacent brain regions. Thus, the show elevated expression at rhombomere cellular mechanisms characteristic for the boundaries (Riley et al., 2004). rhombomeric cell cycle regulation might Consequently, partial inactivation of WNT be similar in the IsO.

29 Aims of the Study

AIMS OF THE STUDY

The aim of this study was to characterize the molecular and cellular mechanisms responsible for the midbrain and rhombomere 1 patterning defects observed in the conditional Fgfr1 mutant mice.

The aims were:

I. To identify FGFR1 and/or IsO target genes (I). II. To characterize expression of one of these, Drapc1 (II). III. To study the FGF-regulated neurogenesis in the midbrain-R1 region (I). IV. The aim was also to study the cell cycle regulation at the midbrain-hindbrain boundary (III, IV).

30 Materials and Methods

MATERIALS AND METHODS

Methods, mouse lines, probes and antibodies used in this study are listed in tables 1-4 and protocols and sources are referenced therein.

Table 1. Methods used in studies I-IV.

Method Reference/Source Study number BrdU incorporation and Cell proliferation kit (Amersham- III,IV immunohistochemical detection Pharmacia) - see table 4 Immunohistochemistry see table 4 I,III-IV Microarray analysis Affymetrix GeneChip I mouse U74A v2 arrays - Total-RNA extraction Trizol reagent (Invitrogen) I - Northern blot analysis of isolated (Frilander and Steitz, 1999) I total-RNA - cRNA labeling BioArray High Yield RNA I Transcript Labeling kit (Enzo) - Data analysis MAS 5.0 / GeneSpring software I Radioactive section in situ analysis Wilkinson and Green 1990 I-IV PCR for genotyping (Trokovic et al., 2003) I-IV TUNEL assay In situ cell death detection kit III (Roche) - see table 4 Whole-mount in situ hybridization (Henrique et al., 1995) I-IV analysis ß-galactosidase staining(Lobe et al., 1999) I V NADPH-diaphorase histochemistry IV IV

Table 2. Mouse lines used in this study. All animal procedures were performed in accordance with the national guidelines and instructions and the experiments were approved by the commit- tee of experimental animal research of the University of Helsinki.

Allele Reference Study number En1-Cre (Kimmel et al., 2000) I-III Fgfr1Flox (Trokovic et al., 2003) I-III En1+/Wnt1 (Panhuysen et al., 2004) III FST KO (Matzuk et al., 1995) I Trh KO (Yamada et al., 1997) not published MHB-Cre IV IV R26R (Soriano, 1999) IV

31 Materials and Methods

Table 3. Probes used for in situ hybridization. Probe Annotation / Reference Study number A) Mouse Aldh1 (Hermanson et al., 2003) I CyclinD1 IMAGE 3155470 III Fgfr1 (Trokovic et al., 2003) III Fgfr2 (Trokovic et al., 2003) III Fgfr3 (Trokovic et al., 2003) III Gata3 (Lillevali et al., 2004) I Hes3 clone UI-R-BO1-ajt-e-02-0-UI.r1 I Lmx1b a gift from H. Simon I Mash1 IMAGE 6415061 I Ngn2 IMAGE 2922473 I Nurr1 (Wallen et al., 1999) I Otx2 (Acampora et al., 1997) I,III,IV Pb-cadherin (Trokovic et al., 2003) III, IV Phox2a IMAGE 534970 I Pet1 clone UI-M-BH3-avj-b-02-0- I UI.s1 Pitx3 IMAGE 482871 I p21 A gift from Bert Vogelstein III,IV Sert (Slc6a4) A gift from Wolfgang Wurst I VaChT (Trokovic et al., 2003) I Wnt1 (Trokovic et al., 2003) I

Other probes and their references are listed in the results section (Tables 5 and 6).

B) Chicken ESTs* Drapc1 (or Pogo1) ChEST ID: C0000369 unpublished ChEST ID: C0003888 “ ChEST ID: C0001388 “

C) Zebrafish Cnpy1 IMAGE 7068815 I *Chicken cDNAs ordered from ARK-Genomics (Boardman et al. 2002).

32 Materials and Methods

Table 4. Antibodies used in immunohistochemical analyses.

Antibody Description, dilution & supplier Study number Primary antibodies anti-ß-galactosidase rabbit polyclonal (1:2000, Cappel) IV anti-BrdU mouse monoclonal (1:500, III, IV Amersham) anti-BrdU sheep monoclonal (1:800, Abcam) IV anti-KI67 rabbit polyclonal (1:500, IV Neomarkers) anti-p21 mouse monoclonal (1:100, BD IV Biosciences) anti-phosphohistone-H3 (PH3) rabbit monoclonal (1:800, Upstate) III anti-Tuj1 (or III ȕ-Tubulin) rabbit monoclonal (1:300, Covance) I anti-tyrosine hydroxylase (TH) rabbit monoclonal (1:500, I,IV Chemicon) anti-5-hydroxytryptamine (5-HT) rabbit monoclonal (1:5000, I, IV ImmunoStar) rabbit polyclonal (1:2000, MP I Biomedicals) Secondary antibodies anti-mouse-IgG Alexa 488/568 conjugated (1:300- I, III, IV 1:400, Molecular probes) HRP-conjugated (1:500, III Amersham) anti-rabbit-IgG Alexa 488/568 conjugated (1:300- III,IV 1:400, Molecular probes) HRP-conjugated (1:500, Chemicon) IV anti-sheep-IgG Alexa 488/568 conjugated (1:300- IV 1:400, Molecular probes) TUNEL analyses Fluorescein in situ detection kit used according to manufacturer’s III instructions (Roche) Counterstain 4´,6-diamino-2-phenylindole included in the mounting media I,III,IV (DAPI) Vectashield (Vector)

33 Results and Discussion

RESULTS AND DISCUSSION

1. Microarray analysis of Fgfr1 CKO Among the genes, which were found mutants (I-III) to be down-regulated in the mutants, there were many whose expression was To get better understanding of the Fgfr1 previously found to be down-regulated in regulated and/or isthmic organizer Fgfr1 CKO mutants using the candidate speciﬁ c genes, we used microarray-based gene approach (Table 5). Among the strategies for identiﬁ cation of genes whose down-regulated transcripts, there were also expression is affected in the conditional products of many genes, which previously Fgfr1 mutants. For these studies, total RNA were not known to be expressed at the from dissected midbrain-R1 tissues of midbrain-R1 region or which represent E10.5 wild-type and En1-Cre/+; Fgfr1Flox/ previously poorly characterized transcripts Flox (Trokovic et al., 2003, referred as Fgfr1 (e.g. expressed sequence tags, ESTs; see CKO mutants hereafter) embryos was Table 7). Many of the down-regulated isolated. The dissected region contained genes are components or regulators of the the posterior midbrain as well as the most signaling cascades downstream of receptor anterior part of the rhombomere 1 (R1) of tyrosine kinases (Table 7 and Fig. 13). the hindbrain (Fig. 12). The hybridizations were carried out on Affymetrix GeneChip arrays (mouse U74A version 2 arrays).

Figure 12. Experimental design for microarray analysis of midbrain-R1 region. Reproduced from Jukkola et al. (2006).

34 Results and Discussion

Table 5. Genes down regulated in Fgfr1 CKO mutants at E10.5.

Gene / EST Afymetrix Fold Validation by in situ Probe(s) for in situ probe set change hybridisation* hybridisation ID Fgf17 98730_at 0.06 Fig. 1B in I (Xu et al., 2000) Spry1 163956_at 0.24 Fig. 1C in I (Zhang et al., 2001) En1 96523_at 0.29 Fig. 2A in I (Davis and Joyner, 1988) Fgf8 97742_s_at 0.30 Fig. 1B in I (Crossley and Martin, 1995) Spry2 116938_at 0.30 Fig. 1C in I (Zhang et al., 2001) Trh 102665_at 0.31 Fig. 2E in I IMAGE 559851 IMAGE 559851 IMAGE 651225 Cnpy1 138472_at 0.35 Fig. 2C in I AI853839 (3’) IMAGE:6390940 (5') Fgf18 95316_at 0.35 Fig. 1B in I (Xu et al., 2000) Fgfr1 97509_f_at 0.38 Expresion (Trokovic et al., 2003) downregulated Tnfrsf19 (Trade) 160670_at 0.39 Fig. 2K in I (Pispa et al., 2003) Sef 133830_at 0.41 Fig. 1C in I IMAGE 1178421 Erm (Etv5) 163173_at 0.41 Fig. 1D in I IMAGE 3674281 Mkp3 (Dusp6) 93285_at 0.42 Fig. 1C in I IMAGE 874051 IMAGE 1138569 Nfia 130461_at 0.48 In WT embryos: Broad IMAGE 1053853 expression in the midbrain-R1 region and in the neural tube Mrp4 (Abcc4) 111137_at 0.50 Fig. 2F in I IMAGE 1153158 IMAGE 3325098 IMAGE 367701 Pax5 95890_r_at 0.50 Fig. 2D in I IMAGE 3333164 EST6 163120_at 0.51 In WT embryos: No IMAGE 534449 specific expression IMAGE 5363361 detected Pcx 171076_i_at 0.51 In WT embryos: No IMAGE 5044506 specific expression IMAGE 5053475 detected En2 98338_at 0.52 Fig. 2B in I (Liu and Joyner, 2001) Sfrp2 93503_at 0.52 Fig. 2L in I IMAGE 4487469 Fgf15 97721_at 0.54 Fig. 1B in I (McWhirter et al., 1997) Atp1a1 93797_g_at 0.54 In WT embryos: No UI-M-BH2.3-aqc-d-05- specific expression 0-UI (BMAP) detected 4921506J03Rik 99163_at 0.54 In WT embryos: No IMAGE 5721414 specific expression IMAGE 5290667 detected Igfbp5 100566_at 0.54 Fig. 2J in I IMAGE 318625 Nfib 160859_s_at 0.54 In WT embryos: IMAGE 617412 Expressed in the IMAGE 4038233 ventral midbrain-R1 IMAGE 4237079 region RhoA (Arha2) 101112_g_at 0.55 In WT embryos: Broad IMAGE 4190318 expression in the IMAGE 4234611 midbrain-R1 region and in the CNS Sox3 92264_at 0.56 Fig. 2I in I IMAGE 368804 Drapc1 96132_at 0.59 Fig. 2M in I IMAGE 1764103 IMAGE 374528 IMAGE 5038498 (Jukkola et al., 2004) Flrt3 110370_at 0.60 Fig. 1C in I IMAGE 5702881 Epcs3 (Ftsj3) 95756_at 0.60 - Ccnd2 97504_at 0.60 Fig. 2H in I IMAGE 367058 Fabp7 98967_at 0.62 In WT embryos: IMAGE 436894 Expressed in the IMAGE 334095 35 Results and Discussion

Table 5 continuing Gene / EST Afymetrix Fold Validation by in situ Probe(s) for in situ probe set change hybridisation* hybridisation ID ventral midbrain Eef1a1 94766_at 0.63 - Pip92 (Ier2) 99109_at 0.63 Expressed strongly in IMAGE 3376243 the ventral neural tube and in the MHB boundary region which is lost in mutants Pea3 (Etv4) 92979_at 0.64 Fig. 1D in I (Lin et al., 1998) Kik1 (Hsd17b12) 94276_at 0.65 In WT embryos: no IMAGE 6394399 expression detected IMAGE 747788 Snx5 117186_at 0.65 In WT embryos: no IMAGE 3492262 expression detected IMAGE 6812794 Tcf7 97994_at 0.65 Fig. 2N in I A gift from Irma Thesleff Bnip3l 96255_at 0.66 In WT embryos: : no IMAGE 1513890 expression detected Gstm5 100629_at 0.67 In WT embryos: : no IMAGE 671545 expression detected Prkrir 99975_at 0.67 - Ccnb1 160159_at 0.69 Expressed broadly in IMAGE 3971364 the CNS. Absent from IMAGE 3973833 the MHB border region IMAGE 4014448 and downregulated in the mutants Spred2 161070_at 0.69 In WT embryos: no IMAGE 4988566 specific exprresion detected Nrp 95016_at 0.69 - Rraga 94257_at 0.70 In WT embryos: IMAGE 5401569 expressed broadly in the CNS and limb buds 1200007D18Rik 160184_at 0.71 In WT embryos: no IMAGE 3494155 specific exprresion detected Pea15 100548_at 0.71 Expressed weakly in IMAGE 406749 the ventral midbrain-R1 region which is lost in mutants CN04 162901_i_at 0.71 In WT embryos: no IMAGE 1245431 (6430527G18Rik) specific exprresion detected at MHB Sox21 167903_at 0.72 - D7Wsu128e 103861_s_at 0.74 - Jmj 94341_at 0.74 Fig. 2G in I IMAGE 6406875 * Published in I or if not stated refers to unpublished data. The microarray screen revealed the loss of isthmic signals can be expected also genes, whose expression was up- to result in premature differentiation. regulated in the Fgfr1 CKO mutants Therefore, it was hypothesized that the (Table 6). Among these were many factors up-regulated gene list may include novel that have been associated with neuronal components of the neuronal differentiation differentiation (see Table 8). In addition, pathway at the midbrain-R1 region.

36 Results and Discussion

Table 6. Genes upregulated in Fgfr1 CKO mutants at E10.5.

Gene / EST Afymetrix Fold Validation by in situ Probe for in situ probe set change hybridisation* hybridisation ID Fst 98817_at 4.00 Fig. 3M, N in I (Wang et al., 2004) EST 166028_s_at 2.22 - Tal-2 129118_at 2.00 In WT embryos: expressed in the tectum (Mori et al. 1999) Mab21l1 165815_r_at 1.89 Fig. 3F in I IMAGE 4526962 Ednrb 163124_s_at 1.87 Fig. 3I in I IMAGE 4971909 Math1 168404_at 1.84 Fig. 3D in I IMAGE 4218223 EST 113895_at 1.77 - Rgma 108717_at 1.66 Fig. 3C in I IMAGE 1001290 EST 139205_at 1.66 Fig. 3K in I IMAGE 2655939 Vtn 98549_at 1.65 Fig. 3A in I IMAGE 5366291 Uncx4.1 92499_at 1.60 Fig. 3H in I IMAGE 5716567 Ngfr (p75) 108762_at 1.57 Fig. 3G in I (Qun et al., 1999) EST 137358_at 1.55 Fig. 3L in I IMAGE 3329477 Maf 115390_at 1.52 - EST 167322_at 1.52 - Ckb 137242_f_at 1.49 Supp. Fig. 7J in I IMAGE 6395794 Pltp 100927_at 1.46 Broad expression in the IMAGE 4979759 neural tube Nhlh2 166814_f_at 1.44 - EST 166871_at 1.44 Fig. 3J in I IMAGE 3823589 Wfdc1 166414_at 1.43 Fig. 3B in I IMAGE 1497229 Tcpn1 108515_at 1.41 - IMAGE 653006 Dach1 114999_at 1.41 Fig. 3E in I IMAGE 6826750

* Published in I

Microarray data were categorized are listed in the tables 7 and 8. Fig. 13 using gene ontology (GO) annotation. GO shows that the representation of GO classes provides a controlled vocabulary organized between up- and downregulated genes was in a hierarchical way that can be applied quite different. The set of up-regulated to describe gene products from mouse genes was enriched with genes involved regarding: molecular function, biological in intercellular signalling (e.g. secreted process, and cellular component. The GO proteins) and neuronal transcription assignment relies on information from the factors but lacks genes involved in signal AMIGO server http://www.geneontology. transduction (intracellular), whereas genes org/index.shtml. The GO classifications involved in responses to external stimulus of downregulated and upregulated genes (growth factors; MAPK pathway), cell in the Fgfr1 CKO mutants at E10.5 are adhesion, and cell proliferation were shown in tables 7 and 8, respectively. A overrepresented among down-regulated summary of only the most specific GO genes. annotations withdrawn from the AMIGO

37 Results and Discussion

Table 7. Functional classiﬁ cation of downregulated genes. GO Cellular Gene GO Biological Process GO Molecular Function Fold Gene Title Component Symbol Description Description change Description Fgf17 fibroblast growth factor signal transduction cell- fibroblast growth factor extracellular space 0.06 17 cell signaling nervous receptor binding growth system developpment factor activity Spry1 sprouty homolog 1 ureteric bud development protein binding cytoplasm membrane 0.24 (Drosophila) induction of an organ regulation of signal transduction negative regulation of MAPK activity En1 engrailed 1 regulation of DNA binding transcription nucleus voltage-gated 0.29 transcription, DNA- factor activity voltage- potassium channel dependent potassium ion gated potassium channel complex transport dorsal/ventral activity sequence-specific pattern formation neuron DNA binding differentiation midbrain- hindbrain boundary development Fgf8 fibroblast growth factor regulation of progression fibroblast growth factor extracellular space 0.3 8 through cell cycle receptor binding growth MAPKKK cascade signal factor activity transduction positive regulation of cell proliferation negative regulation of apoptosis generation of neurons Spry2 sprouty homolog 2 cell-cell signaling protein binding microtubule membrane 0.3 (Drosophila) development regulation of signal transduction MAPK activity cell fate commitment Trh thyrotropin releasing signal transduction cell- hormone activity extracellular space 0.31 hormone cell signaling hormone- neuropeptide hormone soluble fraction mediated signaling activity protein binding thyrotropin-releasing hormone activity Cnpy1 Canopy 1 homolog fibroblast growth factor --- endoplasmic reticulum 0.35 (Zebrafish) receptor signalling (expressed sequence pathway AI853839) midbrain-hindbrain orgainizer development Fgf18 fibroblast growth factor signal transduction cell- fibroblast growth factor extracellular space 0.35 18 cell signaling positive receptor binding growth nucleolus regulation of cell factor activity proliferation Fgfr1 fibroblast growth factor MAPKKK cascade protein protein kinase activity extracellular space 0.38 receptor 1 amino acid protein serine/threonine membrane fraction phosphorylation signal kinase activity protein- integral to plasma transduction brain tyrosine kinase activity membrane development regulation receptor activity heparin of cell proliferation cell binding growth negative regulation of apoptosis Tnfrsf19 tumor necrosis factor induction of apoptosis receptor activity tumor extracellular space 0.39 receptor superfamily, JNK cascade necrosis factor receptor integral to membrane member 19 activity protein binding Sef Simlar expression to dorsal/ventral pattern transmembrane receptor integral to membrane 0.41 Fgfs specification activity negative regulation of FGF receptor signalling pathway Erm ets variant gene 5 ets regulation of DNA binding transcription nucleus 0.41 (or Etv5) variant gene 5 transcription, DNA- factor activity pseudogene dependent organ transcriptional activator morphogenesis positive activity regulation of transcription Mkp3 dual specificity protein amino acid phosphoprotein MAP --- 0.42 (or Dusp6) phosphatase 6 dephosphorylation kinase phosphatase protein amino acid activity,tyrosine/serine/thre dephosphorylation onine phosphatase activity protein binding hydrolase activity Nfia nuclear factor I/A DNA replication DNA binding transcription intracellular nucleus 0.48 regulation of factor activity iron ion transcription, DNA- binding electron carrier dependent electron activity transport Mrp4 ATP-binding cassette, ion transport multidrug nucleotide binding chloride integral to membrane 0.5 (or Abcc4) sub-family C transport channel activity ATP (CFTR/MRP), member 4 binding multidrug efflux pump activity ATPase activity 38 Results and Discussion

Table 7 continuing Pax5 Paired box gene 5 regulation of DNA binding transcription nucleus transcription 0.5 transcription, DNA- factor activity protein factor complex dependent transcription binding from RNA polymerase II promoter nervous system development EST6 163120_at ------0.51

Pcx 171076_i_at ------0.51

En2 engrailed 2 regulation of DNA binding transcription nucleus 0.52 transcription, DNA- factor activity sequence- dependent development neuron differentiation midbrain- hindbrain boundary development Sfrp2 secreted frizzled-related somitogenesis transmembrane receptor membrane 0.52 protein 2 anterior/posterior pattern activity formation Wnt receptor signaling pathway differentiation Fgf15 fibroblast growth factor neural crest cell migration fibroblast growth factor extracellular space 0.54 15 signal transduction receptor binding growth nervous system factor activity development Atp1a1 ATPase, Na+/K+ ion transport cation nucleotide binding membrane fraction 0.54 transporting, alpha 1 transport potassium ion magnesium ion binding integral to plasma polypeptide transport sodium ion catalytic activity membrane transport monovalent sodium:potassium- sodium:potassium- inorganic cation transport exchanging ATPase activity exchanging ATPase ATP hydrolysis coupled transmembrane movement complex proton transport of ions, phosphorylative metabolism mechanism 4921506J03R RIKEN cDNA 4921506J03 ------0.54 ik gene Igfbp5 insulin-like growth regulation of cell growth insulin-like growth factor extracellular space 0.54 factor binding protein 5 signal transduction binding insulin-like growth factor binding protein complex Nfib nuclear factor I/B DNA replication DNA binding transcription intracellular nucleus 0.54 regulation of factor activity transcription, DNA- dependent negative regulation of cell proliferation hindbrain development Rhoa ras homolog gene cell morphogenesis cell nucleotide binding GTPase intracellular nucleus 0.55 (or Arha2) family, member A adhesion cell-matrix activity signal transducer cytosol similar to aplysia ras- adhesion small GTPase activity cytoskeleton related homolog A2 mediated signal plasma membrane similar to aplysia ras- transduction actin related homolog A2 cytoskeleton organization and biogenesis cell differentiation positive regulation of NF-kappaB import into nucleus negative regulation of neuron apoptosis Sox3 SRY-box containing establishment and/or DNA binding transcription nucleus 0.56 gene 3 maintenance of factor activity transcription factor chromatin architecture complex regulation of transcription, DNA- dependent central nervous system development forebrain development Drapc1 adenomatosis polyposis ------integral to membrane 0.59 (or Apcdd1) coli down-regulated 1 Flrt3 fibronectin leucine rich --- protein binding integral to membrane 0.6 transmembrane protein 3 Epcs3 FtsJ homolog 3 (E. coli) rRNA processing methyltransferase activity nucleus 0.6 (or Ftsj3) transferase activity Ccnd2 cyclin D2 regulation of progression protein binding cyclin- nucleus 0.6 through cell cycle dependent protein kinase regulator activity Fabp7 fatty acid binding transport binding lipid binding cell projection cell soma 0.62 protein 7, brain Eef1a1 eukaryotic translation protein biosynthesis translation elongation cytoplasm 0.63 elongation factor 1 translational elongation factor activity nucleotide alpha 1 binding GTP binding Pip92 immediate early ------cytoplasm 0.63 (or Ier2) response 2 39 Results and Discussion

Table 7 continuing

Pea3 ets variant gene 4 (E1A regulation of DNA binding transcription nucleus 0.64 (or Etv4) enhancer binding transcription, DNA- factor activity protein, E1AF) dependent motor axon guidance positive regulation of transcription Kiki1 hydroxysteroid (17-beta) steroid biosynthesis estradiol 17-beta- endoplasmic reticulum 0.65 (or dehydrogenase 12 metabolism lipid dehydrogenase activity integral to membrane Hsd17b12) biosynthesis oxidoreductase activity Snx5 sorting nexin 5 transport cell protein binding --- 0.65 communication protein phosphoinositide binding transport Tcf7 transcription factor 7, T- regulation of DNA binding transcription nucleus transcription 0.65 cell specific transcription, DNA- factor activity factor complex dependent Wnt receptor signaling pathway regulation of cell proliferation regulation of apoptosis Bnip3l BCL2/adenovirus E1B induction of apoptosis protein binding nucleus 0.66 interacting protein 3- negative regulation of mitochondrionintegral like survival gene product tomembrane activity Gstm5 glutathione S- establishment of blood- glutathione transferase --- 0.67 transferase, mu 5 nerve barrier metabolism activity

Prkrir protein-kinase, regulation of translation DNA binding protein nucleus 0.67 interferon-inducible signal transduction binding zinc ion binding double stranded RNA negative regulation of metal ion binding protein dependent inhibitor, cell proliferation dimerization activity repressor of (P58 repressor) Ccnb1 cyclin B1, related regulation of progression cyclin-dependent protein nucleus 0.69 sequence 1 cyclin B1 through cell cycle mitosis kinase regulator activity cytokinesis protein protein binding kinase targeting activity Spred2 sprouty-related, EVH1 inactivation of MAPK stem cell factor receptor plasma membrane 0.69 domain containing 2 activity regulation of protein binding cytoplasmic vesicle signal transduction Nrp1 neuropilin 1 cell adhesion signal receptor activity membrane fraction 0.69 transduction cell-cell semaphorin receptor endoplasmic reticulum signaling nervous system activity integral to membrane development positive regulation of cell proliferation Rraga Ras-related GTP binding apoptosis regulation of nucleotide binding protein nucleus cytoplasm 0.7 A cytolysis binding GTP binding protein homodimerization activity protein heterodimerization activity 1200007D18 endoplasmic reticulum- transport ER to Golgi protein binding endoplasmic reticulum 0.71 Rik golgi intermediate vesicle-mediated ER-Golgi intermediate compartment (ERGIC) 1 transport transport compartment integral to membrane Pea15 phosphoprotein transport regulation of protein kinase C binding membrane fraction 0.71 enriched in astrocytes apoptosis protein binding cytosol microtubule 15 associated complex CN04 RIKEN cDNA --- protein binding zinc ion nucleus 0.71 6430527G18 6430527G18 gene binding metal ion binding Rik Sox21 SRY-box containing transcription regulation DNA binding transcription chromatin nucleus 0.72 gene 21 of transcription, DNA- factor activity dependent D7Wsu128e DNA segment, Chr 7, protein modification ------0.74 Wayne State University 128, expressed Jarid2 (or jumonji, AT rich negative regulation of DNA binding protein intracellular nucleus 0.74 Jmj) interactive domain 2 transcription from RNA binding transcriptional polymerase II promoter repressor activity regulation of transcription, negative regulation of cell proliferation

40 Results and Discussion

Table 8. Functional classiﬁ cation of upregulated genes.

GO Molecular GO Cellular GO Biological Process Fold Gene Symbol Gene Title Function Component Description change Description Description Fst follistatin BMP signaling pathway protein binding extracellular region 4 pattern specification activin inhibitor extracellular space negative regulation of activity cell differentiation 2610034M16 RIKEN cDNA protein targeting protein binding --- 2.22 Rik 2610034M16 gene Tal-2 T-cell acute regulation of DNA binding nucleus 2 lymphocytic transcription, DNA- transcription leukemia dependent regulator activity Mab21l1 mab-21-like 1 (C. positive regulation of protein kinase Golgi trans face cytosol 1.89 elegans) cell proliferation binding protein plasma membrane protein localization kinase A binding Ednrb endothelin neural crest cell rhodopsin-like extracellular space 1.87 receptor type B migration G-protein receptor activity G- plasma membrane coupled receptor protein coupled protein signaling receptor activity pathway nervous system development Math1 atonal homolog 1 regulation of DNA binding nucleus 1.84 (or Atoh1) (Drosophila) transcription, DNA- transcription factor dependent transcription activity protein from RNA polymerase II binding transcription promoter brain regulator activity development cell chromatin DNA differentiation neuron binding differentiation EST A730008L19 ------1.77 Rgma RGM domain neural tube closure GPI anchor binding cell surface membrane 1.66 family, member A C130071C03R RIKEN cDNA ------1.66 ik C130071C03 gene Vtn vitronectin cell adhesion heparin binding extracellular space 1.65 cell-matrix adhesion protein binding proteinaceous extracellular matrix Uncx4.1 Unc4.1 homeobox regulation of transcription factor nucleus transcription 1.6 (C. elegans) transcription, DNA- activity sequence- factor complex dependent specific DNA binding Ngfr nerve growth apoptosis induction of catalytic activity extracellular space 1.57 (or p75) factor receptor apoptosis signal receptor activity plasma membrane (TNFR superfamily, transduction axon protein binding integral to membrane member 16) guidance central nerve growth factor nervous system binding protein development cell binding differentiation D930028M14Ri RIKEN cDNA ------1.55 k D930028M14 gene Maf avian regulation of DNA binding chromatin nucleus 1.52 musculoaponeuroti progression through cell transcription factor cytoplasm c fibrosarcoma (v- cycle regulation of activity RNA maf) AS42 transcription, DNA- polymerase II oncogene homolog dependent transcription transcription factor from RNA polymerase II activity promoter Lcor ligand dependent regulation of DNA binding RNA nucleus 1.52 nuclear receptor transcription, DNA- polymerase II corepressor dependent transcription transcription factor from RNA polymerase II activity promoter Ckb creatine kinase, --- creatine kinase cytoplasm 1.49 brain activity kinase mitochondrion activity Pltp phospholipid lipid metabolism lipid lipid binding extracellular space 1.46 transfer protein transport

41 Results and Discussion

Table 8 continuing

GO Molecular GO Cellular GO Biological Process Fold Gene Symbol Gene Title Function Component Description change Description Description Nhlh2 nescient helix loop regulation of DNA binding nucleus 1.44 helix 2 transcription, DNA- transcription dependent central regulator activity nervous system development cell differentiation Fgfbp3 fibroblast growth ------1.44 factor binding protein 3 Wfdc1 WAP four-disulfide --- endopeptidase extracellular space 1.43 core domain 1 inhibitor activity Tpcn1 two pore channel 1 ion transport ion channel activity integral to membrane 1.41 sugar binding Dach1 dachshund 1 regulation of DNA binding nucleus transcription 1.41 (Drosophila) transcription, DNA- transcription factor factor complex dependent activity protein binding

Figure 13. A summary of molecular functions (GO) of differentally expressed genes. (A) downregulated genes; (B) upregulated genes.

42 Results and Discussion

To validate our microarray results, this of various signal transduction cascades analysis was followed by a whole-mount (Table 7 and Fig. 13). mRNA in situ hybridization analysis. First, we concentrated on expression analysis of 2.1. Down-regulated genes (I, III and a down-regulated gene list showed in Table unpublished) 5. Initially this list produced by the microarray analysis consists of sets of DNA 2.1.1. Down-regulation of FGF-FGFR probes, each chosen to record expression signaling components at E10.5 (I,III) of specifi c genes. In Affymetrix GeneChip methodology, the set of probes relating to The down-regulated genes identified a gene is referred to as a probeset, and the by microarray screen included several sequence which is best associated with the known members of the FGF signal transcribed region being interrogated by transduction pathway (Table 5). These the probeset is referred to as its represent- included FGF ligands, signal regulators ative sequence (Liu et al., 2003a). The se- and nuclear effectors. At E9.5 the quence annotations referring to the infor- expression of Fgf8/17/18 was not down- mation about the representative sequence regulated in the Fgfr1 mutant embryos. for a probeset are presented in the Affyme- However, comparison of expression trix NetAffx service (http://www.affyme- levels by microarray analysis showed that trix.com). This tool was utilized to locate Fgf8/15/17/18 were down-regulated in the probes within the mouse genome and E10.5 Fgfr1 mutant mice (Table 5). This the corresponding cDNAs were ordered was confirmed by in situ hybridization from the IMAGE consortium. These plas- analysis. Expression of Fgf17 was mids were then used to produce the in situ completely abolished from the midbrain-R1 hybridization probes for consequent vali- region whereas Fgf8 and Fgf18 expression dation experiments. was maintained in the most ventral region. This suggests, that other FGF receptors (FGFR2 and 3) are mediating isthmic FGF 2. Identifi cation of FGF regulated genes signals in the Fgfr1 mutants (see below). at the midbrain-R1 region (I, III and Consistent with this, the most evident unpublished) downregulation of Fgf15 was observed in the lateral midbrain, whereas the ventral In addition to the components of the Fgf15 expression domain was unchanged FGF-FGFR signal transduction pathway or even upregulated. (see section 2.1.1.) our microarray screen Sef (similar expression to Fgfs) is a identified several other genes being receptor molecule that is associated with differentially expressed in the Fgfr1 mutant the FGF/Ras-MAPK signaling. At E10.5 embryos (see sections 2.1.2. and 2.1.3.). wild-type embryos, Sef expression was Comparison of the midbrain-R1 tissues detected in the midbrain-R1 region and from E10.5 wild-type and Fgfr1 mutant extended posterior to the isthmus. In the embryos identifi ed 51 down-regulated and mutant E10.5 embryos, a down-regulation 20 up-regulated genes, respectively (see of Sef was seen in the most dorsal part of Tables 5 and 6). The downregulated list the midbrain-R1 region. Similarly to Pea3 included genes involved in the patterning and Erm at E9.5, Sef expression was also of midbrain and R1, MHB-specifi c genes, downregulated close to the border between cell-cycle regulators, and components midbrain and R1. At E10.5, expression of

43 Results and Discussion

Erm and Pea3 was clearly down-regulated maintenance of local organizing centers. at the midbrain-R1 region in the Fgfr1 In Xenopus and zebrafish, expression mutant embryos compared to the wild-type of Mkp3 has been shown to induce by embryos. However, in the Fgfr1 mutants, RA and β-catenin pathways (Tsang and some expression of Erm and Pea3 was Dawid, 2004; Tsang et al., 2004). In Fgfr1 still detected in the ventral midbrain-R1 mutants, Mkp3 expression was lost from region. the dorsal midbrain-R1 region already at Sprouty (or SPRY) and MKP3 E9.5. Interestingly, at this stage, Wnt1 and proteins act as negative regulators of Drapc1 expression is still detected in the the FGF/Ras-MAPK signaling pathway. dorsal region. Thus, it can be speculated Both Spry1 and Spry2 are expressed at that FGF8/17/18 are the key molecules the midbrain-R1 region at E10.5. In the which are regulating the MAPK-RAS Fgfr1 mutant embryos, Spry1 transcripts pathway, which in turn is required for the were abolished from the dorsal midbrain- dorsal patterning of R1. R1 domain, but remained in the ventral Recent studies have shown that rat midbrain-R1 region. By contrast, and Xenopus Flrt3 genes are coexpressed expression of Spry2 and Mkp3 was with Fgfs (Bottcher et al., 2004; Lacy down-regulated throughout the midbrain- et al., 1999; Robinson et al., 2004). We R1 region in the Fgfr1 mutant mice at detected Flrt3 expression in wild-type E10.5 compared to wild-type embryos. embryos from E10.5 onwards at the Although, Mkp3 expression was clearly narrow midbrain-R1 region. In the Fgfr1 down-regulated at E10.5 in Fgfr1 mutants, mutant embryos, Flrt3 expression domain some expression was detected in the most at the midbrain-R1 region was abolished. ventral midbrain-R1 domain when whole- The lack of Flrt3 expression in the mount embryos were sectioned after in situ boundary cells between the midbrain and hybridizations. A positive feedback loops R1 may be the mechanism to the observed provides a basis for the formation and tissue segregation in Fgfr1 mutants. Thus,

Figure 14. Cnpy1 is speciﬁ cally expressed in the midbrain-R1 region. Expression pattern of Cnpy1 is conserved between mouse (A) and zebraﬁ sh (B). Arrowheads point to the IsO and an arrow marks the more posterior expression of Cnpy1. hpf, hours post fertilization; ss, somite stage.

44 Results and Discussion similarly than WNT signaling modulates the FGF signal pathway were identiﬁ ed cadherin-mediated cell adhesion in as being down-regulated in the Fgfr1 Drosophila imaginal disc cells (Wodarz et mutants at E10.5. Taken together, most of al., 2006), the FGFR signaling is needed these FGF-FGFR1 signaling components for the homophilic FLRT-mediated cell- described here retained expression in the cell adhesion. In the Fgfr1 mutants the ventral midbrain-R1 region in the Fgfr1 dorsal patterning was more affected. Thus, mutants indicating the presence of residual it is possible that residual FGFR signaling FGF signaling. may be required for maintaining the FLRT- mediated adhesion in the ventral region. 2.1.2. Cnpy1 as a novel regulator of In addition to Flrt3, Flrt1 and Flrt2 are FGF signaling at the midbrain-R1 expressed in the ventral midbrain-R1 region (I and unpublished) region as showed by (Haines et al., 2006). In Fgfr1 mutants, expression of Flrt1 or Expression of IsO regulated transcription Flrt2 was not studied. factors En1/2 and Pax5 was abolished Thus, validating our microarray at the dorsal part of the midbrain-R1 approach, known members of the FGF region of mutant embryos. Interestingly, synexpression group and components of previously uncharacterized gene (NM_

Figure 15. Correlation of Cnpy1 expression to the ventral midbrain DA neurons at E12.5. Radioactive in situ hybridization with Cnpy1 (A,B). A bracket marks the strongest Cnpy1 expression domain in the ventral midbrain. TH-positive domain is marked with the broken line in A-C. Insert in D shows the pseudocolored staining of Cnpy1 in situ signal (red) overlaid with the TH (green) and counterstained with the DAPI (blue). Lines indicate the most posterior expression boundary of Otx2 expression on parallel section (not shown). Broken lines in A-C mark the TH- positive region shown in D. cp, choroid plexus; DA, dopaminergic neurons, mb, midbrain; r1, rhombomere 1.

45 Results and Discussion

175651; Affymetrix probe ID: 138472_at) The role of Cnpy1 in the maturation of showed similar expression pattern to the ventral DA neurons might be less obvious. En2 gene specifi cally at the midbrain-R1 However, further functional analyses of (Fig. 14A). The zebrafish ortholog was mouse Cnpy1 are needed to clarify its role recently named as Canopy1 (Cnpy1, NM_ in the FGF-FGFR regulated patterning and 001039497; Fig. 14B) and its inactivation differentiation in the midbrain-R1 region. by antisense morpholinos resulted in MHB defects (Hirate and Okamoto, 2006). 2.1.3. Down-regulated genes expressed Given the importance of En1 and as a narrow stripe near the MHB (I and En2 in the development of midbrain DA unpublished) neurons (Alberi et al., 2004; Simon et al., 2001; Simon et al., 2004; Simon et Genes that were identifi ed to be expressed al., 2005) we also studied mouse Cnpy1 in a narrow stripe near the MHB included expression in relation to the known DA Trh, Mrp4, Igfbp5 and Jumonji (Jmj). markers (TH and En1). At E12.5, Cnpy1 The role of Jmj in the midbrain-R1 was expressed strongly in the dorsal region is discussed in section 5.4. Trh midbrain-R1 region whereas its expression and Mrp4 showed specifi c expression in in the TH-positive DA neurons was weaker. the midbrain-R1 region whereas Igfbp5 In addition, the co-localization of Cnpy1 expression was detected in various brain and TH was restricted to the most posterior DA neurons adjacent to the isthmal region (Fig. 15). Later at E15.5 and E16.5, Cnpy1 expression in TH-positive cell populations was even weaker (Fig. 16). Thus, our analysis indicates that Cnpy1 might be needed for the development of the most posterior midbrain DA neurons and in the early patterning of midbrain-R1 region. The proposed model for this patterning activity is the direct interaction of CNPY1 with the extracellular domain of FGFR1. Hirate and Okamoto (2006), suggest that together with other FGF-FGFR signaling components CNPY1 promotes the FGF signaling in the midbrain-R1 region.

Figure 16. Correlation of Cnpy1 expression to the ventral midbrain DA neurons at E15.5 and E16.5. Comparison of TH (A,C) and En1 with Cnpy1 (B,D) on parallel sagittal sections. Broken circles (B) or lines (D) indicate the strongest TH immunostaings at E15.5 and E16.5 in parallel sections. Square in A shows the depicted TH staining (green) in A. cb, cerebellar primordium; cp, choroid plexus; mb, midbrain.

46 Results and Discussion regions between E9.5 and E11.5. The 2002; Yamada et al., 1997). We studied the importance of these genes during the expression of early patterning genes in Trh midbrain-R1 development has not been KO embryos. At E10.5, expression of the reported. Mouse knockouts of Trh, Mrp4, genes studied was unchanged in the Trh Igfbp5 and Jumonji has been generated, KO embryos compared with the wild-type but only Jmj knockout mice show obvious (+/+) litter mates (Fig. 17). At the later brain phenotype (Assem et al., 2004; Lai stage (E15.5), as well, no obvious changes and Tan, 2002; Motoyama et al., 1997; in gene expression patterns of Otx2, En1 Ning et al., 2007; Takeuchi et al., 1995; or Drapc1 were detected (Fig. 18). Takeuchi et al., 1999; Urayama et al.,

Figure 17. Expression of patterning genes in the Trh KO at E10.5. Radioactive in situ hybridization analysis with the probes indicated (A-F). Lines mark the most posterior Otx2 expression boundary. mb, midbrain; r1, rhombomere 1.

47 Results and Discussion

2.1.4. Other down-regulated genes in the 2.2. Up-regulated genes show loss of Fgfr1 CKO mutants (I) expression gradients in Fgfr1 CKO mutant embryos (I) Among the other down-regulated genes (Table 5) we observed Cyclins (CyclinD1/ Taken together, the majority of upregulated D2/B1) and Sox3 and Sox21, which were genes (Table 6) showed mainly two main broadly expressed in the dorsal midbrain types of expression in wild-type embryos and hindbrain. Interestingly, a narrow - dorsal gradients or combined dorsal domain of Cyclins- and Sox3-negative and ventral gradients. Genes such as cells was detected in the border between Fst, Vtn, Wfdc1, RgmA, Math1, Dach1 mid- and hindbrain. Expression of these and Mab21l1 were expressed as dorsal genes was downregulated in the dorsal gradients decreasing towards the MHB midbrain-R1 region of Fgfr1 mutant in wild-type embryos. In Fgfr1 mutants, embryos. It has been suggested that the midbrain and R1 gradients were lost expression of Sox genes is dependent of and the expression continued uniformly FGF signaling (Takemoto et al., 2006). strong towards the boundary or across it. In addition, Sox3 expression is inhibited This effect was the most prominent for in the neuroepithelial boundary regions to Fst, which is a known BMP signaling prevent the early exit from the cell cycle component. (Rizzoti et al., 2004). Thus, these cellular mechanisms linking FGF signaling to proliferation and differentiation prompted us to study the genes important for the neuronal differentiation in the vicinity of the MHB (see section 3 below).

Figure 18. Expression of patterning genes in the Trh KO at E15.5. Radioactive in situ hybridization analysis with the probes indicated (A-C). Lines mark the most posterior Otx2 expression boundary. cb, cerebellum; mb, midbrain.

48 Results and Discussion

In summary, several upregulated neuronal progenitors at certain stages of genes showed marked similarity in their development. expression patterns in E10.5 embryos. In the Fgfr1 mutant embryos, the Most of them were expressed as gradients Ngn2 expressing cell populations of the decreasing towards MHB, both in midbrain ventral midbrain and hindbrain were less and rhombomere 1. In Fgfr1 mutant compact and spread towards the boundary embryos, these gradients disappeared both (E10.5 and E11.5). Especially, a strongly in dorsal and in ventral regions. expressing cell population near the BMPs are thought to regulate MHB in ventral R1 showed an extensive the development of the LC. Thus, to scattering effect. Also the LC domain in address the role of follistatin (FST) on E10.5 mutants seemed less compact as neurogenesis, the expression of Phox2a compared to the wild-type. The III and IV and Ngn2 were analyzed on E10.5 Fst cranial ganglia were the same size in the knock-out embryos (Matzuk et al., 1995). Fgfr1 mutants as in the wild-type embryos The expression of these both genes in as marked by the Phox2a expression. mutants was, however, identical to the However, Phox2a expression domains wild-type controls and no changes in the appeared to be fused together, presumably LC were seen at this stage. In addition, as a result of the loss of the boundary cell there were no observable changes in the population between mid- and hindbrain. appearance of locus coeruleus, SA neurons As the development of III and IV or midbrain DA neurons in Fst mutants at cranial ganglia as well as LC is initiated E14.5. This suggests that the upregulation earlier compared to other neuronal of Fst alone would not lead to notable populations, such as SA or DA neurons, changes in neurogenesis. In addition, they can be less dependent on the spreading of Fst expression anteriorly in isthmic signaling. Thus, the major events Fgfr1 mutant embryos does not prevent concerning their formation takes place the development of LC. before the need for FGFR1 function. This would explain their mild phenotype in Fgfr1 conditional mutant adults - III 3. Neuronal progenitor cell populations and IV ganglia appear normal, and LC is are affected in Fgfr1 CKO mutant somewhat disorganized (Trokovic et al., embryos (I) 2003). To study the fate of the neuronal 3.1. Neuronal progenitor cell precursor cells in Fgfr1 mutants, we populations are affected in Fgfr1 CKO analyzed the expression of Hes3, a bHLH mutant embryos (I) transcriptional factor and a known Notch effector, which is inhibiting neuronal Since our results suggested loss of differentiation in the MHB (Hirata et al., neuronal differentiation gradients in Fgfr1 2001). In E10.5 Fgfr1 mutant embryos, mutants, we wanted to know what effects the expression of Hes3 was downregulated FGFR1 conditional inactivation would but was still weakly detectable in the have on different subsets of neuronal caudal midbrain near the IsO. The residual progenitor cells in the midbrain-R1 expression in the R1 near the IsO was region. We analyzed several genes which confined to the neuroectoderm. Thus are known to be expressed in different the loss of Hes3 expression supports the

49 Results and Discussion results obtained from the upregulated comparison between the total numbers genes (see above), which suggested of DA neurons showed not statistically that neuronal differentiation was shifted signifi cant difference between the wild- towards the MHB region in Fgfr1 mutant type and mutant embryos. Next, we embryos. looked at the differences between the In wild type embryos, a gap of mature caudal and rostral DA populations as neurons was evident near the MHB region, these were compared individually. The as seen with anti-TUJ1 staining. In Fgfr1 analysis revealed an increase in the caudal mutants the gap disappeared. The number population and conversely a decrease of radial glia cells and mitotic cells outside in the rostral population of DA neurons the boundary region, as determined by in the Fgfr1 mutant embryos. Thus, anti-RC2 and anti-PH3 immunostainings, consistent with observed changes in gene respectively, was not signifi cantly altered expression patterns of DA markers TH in Fgfr1 mutants (data not shown). Earlier immunostaining at E15.5 stage revealed a quantification studies did not show any caudal shift in the position of DA neurons changes in cell proliferation and apoptosis in Fgfr1 mutants. Thus, it seems that in the ventral MHB region (Trokovic et al., although the area where the DA neurons 2003). However, the changes in the ratio are located expands in mutants, the total of progenitor neurons to mature neurons number of these neurons does not change. might be too subtle to be easily detectable The caudal shift is probably due to the in these early embryonic stages. caudal spreading of Otx2 expression. In ventral midbrain, Otx2 expression is 3.2. Midbrain dopaminergic neuron scattered in Fgfr1 mutants at E15.5, and progenitor pool spreads in Fgfr1 CKO the boundary between Otx2 positive and mutants (I) negative areas is less obvious than in the wild-type. This would suggest that Since FGF is thought to control the the mixing of ventral midbrain and R1 development of DA neurons in the ventral cells forms a tissue mosaic in the Fgfr1 midbrain, we examined what happens to mutants. A similar phenomenon has bee the DA neuron progenitors in the midbrain. observed in zebrafi sh Ace mutants where For this, we analyzed the expression of R1 is transformed into midbrain in the several genes known to be involved in the absence of Fgf8 from the IsO (Jaszai et al., early development of midbrain DA neurons 2003). However, in Fgfr1 mutants, instead and quantifi ed the neurons in E15.5 Fgfr1 of complete fate transformation of R1 an mutant and wild-type embryos. extensive mixing of ventral midbrain cells In E11.5 Fgfr1 mutant embryos, the into R1 is observed. expression of Nurr1, Pitx3 and Aldh1 (also known as Aldh1a1 and Raldh1) 3.3. The most rostral serotonergic was expanded posteriorly towards the neuron precursors are lost in Fgfr1 border between midbrain and R1. Then CKO mutant embryos (I) we analyzed the tyrosine hydroxylase (TH) positive dopaminergic neurons FGF8 from the IsO is thought to be one of of the ventral tegmental area (VTA) the major signals guiding the development and substantia nigra at E15.5 wild- of serotonergic (SA) neurons in the raphe type and Fgfr1 mutant embryos. The nuclei (Ye et al., 1998). To fi nd out whether

50 Results and Discussion precursor of SA neurons are affected in Taken together, it is interesting that Fgfr1 mutant embryos, we analyzed the the downregulation of transcription factors expression patterns of several genes that Mash1, Gata3 and Pet1 could be seen are known to be involved in the SA neuron only in the most anterior R1. The early specifi cation and quantifi ed the rostral SA downregulation of the proneural gene neurons in Fgfr1 mutant and wild-type Mash1 suggests that the SA progenitors embryos at E15.5. near the MHB require a high dose of In Fgfr1 mutants, several genes FGF-signals from the IsO in order to which are involved in the specification develop normally. In the explant culture of SA neurons, such as Gata3, Mash1/ experiments, FGF4 in combination with Ascl1 and Pet1 were downregulated FGF8 and SHH could induce SA neuron near the midbrain-R1 region in the development (Ye et al., 1998). However, ventral R1. Next we performed anti-5HT FGF4 itself is not expressed in the vicinity immunohistochemistry on E11.5 and E15.5 of developing rostral SA neurons at E10.5- sagittal sections to characterize whether 11.5. It has been suggested that FGF4 downregulation of the transcription factors expressed in the primitive streak primes has an effect on SA neuron development. neuroectoderm for the development of At E11.5 in wild-type embryos, 5-HT was SA precursors (Ye et al., 1998). However, detected in the same area as Pet1. In Fgfr1 since FGFR1 is inactivated by 10 somite mutants, 5-HT expression was abolished stage in our conditional mutants (Trokovic near the boundary region, similarly to et al., 2003), the early FGF4 signaling Pet1. At 15.5, when all SA neurons have functions normally in Fgfr1 CKO mutants. developed in mouse, 5-HT could be seen This suggests that other FGFs from the IsO in a large area extending to the top of the are the major molecular regulators in SA mesencephalic fl exure in the wild-type. In neuron development. Thus, it is possible Fgfr1 mutants, the rostral boundary of SA that in the explant culture experiments neuron population was shifted caudally FGF4 mimicked the high FGF signaling and almost no 5-HT positive neurons were activity near the MHB, which is normally seen near the top of the mesencephalic achieved by the combination FGF8/ flexure. The SA neurons located more FGF17/FGF18 molecules expressed by caudally seemed unaffected in mutants. the IsO. In Fgfr1 mutants, the amount of When the total number of neurons activating isthmic signals is reduced at was counted and compared, their number E10.5 and the SA neuron developmental was significantly decreased in mutants. pathway is impaired. Serotonergic neurons The neurons were divided into dorsal further away from the boundary express and ventral populations which were normally all the genes in the SA pathway, compared individually (similarly to the as well as 5-HT, and seem less dependent DA neurons). The loss of the most rostral on the isthmic signaling. It is likely that SA neurons explains the decrease in the the FGFs from the IsO can only diffuse dorsal population. The ventral population and activate the genes needed for SA seems to be unaffected in Fgfr1 mutants. development in the anterior R1. The quantification results reflect the The caudal spreading of Otx2 changes in early gene expression, where expression is not enough to explain the downregulation was seen in the rostral the observed changes in DA and SA area only. populations. Firstly, its expression does

51 Results and Discussion not extend to the whole region where region which disrupts the patterning of the downregulation of the SA marker midbrain and R1. Additionally, the slowly genes was observed. Secondly, although proliferating boundary cell population is Otx2 shows a larger caudal shift in its needed for the positioning and stabilization expression in ventral midbrain, the region of the IsO that function by informing the where Otx2 spreads is not completely DA and SA progenitor cells in the adjacent void of SA neurons, although their brain regions. number is significantly reduced. At the same time, Otx2 expression is scattered and somewhat downregulated in ventral 4. Identiﬁ cation of a putative target midbrain-R1 region. Based on these gene of WNT-β-catenin pathway (II) observations, it seems that the mixing of MHB region cells, already visible at The first transcript among the down- E10.5, continues during development. At regulated probesets showing an interesting E15.5, the observed cell mixing results to expression pattern was annotated to be a a mosaic midbrain-R1 region where cells gene called EIG180 (ethanol induced of the ventral midbrain, including DA gene product). Further sequence analysis, neurons have extensively scattered into however, revealed that this transcript is ventral R1, but R1 has not changed its fate a part of the novel mouse gene located completely into midbrain. Consequently, upstream of EIG180 locus on the mouse there are a mixture of ectopic DA neurons chromosome 18 (Fig. 19B). This novel and some SA neurons in the same area. mouse gene was found to be orthologous This scattering effect enhances the already to a human gene called Drapc1 (or impaired development of rostral SA Apcdd1), which was earlier showed to neurons by disrupting the patterning of be down-regulated in cancer cells by midbrain-R1 region. exogenous APC (Takahashi et al., 2002). It is interesting that various neuronal In addition, they showed that human populations differ in their requirement DRAPC1 is capable to promote growth for FGF signaling through FGFR1 (see of cancer cells in vitro and it is a putative sections 3.1.-3.3). Whereas the midbrain target of WNT/β-catenin pathway. The DA neurons shift posteriorly but their latter notion was further supported by numbers are largely unaffected, the rostral our in situ hybridization analysis. Mouse SA neurons are greatly reduced in Fgfr1 Drapc1 expression pattern showed mutants. Thus, isthmic FGF signaling striking similarities to active WNT/β- is required for the proper development catenin signaling (Maretto et al., 2003) of the rostral SA neurons in R1. Our in the developing embryos (Fig. 19A). To model suggests that high FGF-signaling support our sequence analysis of the mouse levels from the IsO prevent neuronal Drapc1 locus and predicted exon structure, differentiation in the midbrain side we detected similar expression patterns near the MHB. When FGF-signaling with three individual probes transcribed from IsO is abolished in Fgfr1 mutants, from different cDNA clones (Fig. 19B). neurogenesis can proceed closer to MHB At the midbrain-R1 region, Drapc1 in the midbrain but is impaired in the expression was located in the Wnt1- and ventral R1. This effect is enhanced by the Otx2-positive cells in the midbrain (Fig. mixing of the cells in ventral midbrain-R1 19D-F). In the Fgfr1 mutants, Drapc1 was

52 Results and Discussion down-regulated at E10.5 and abolished at Genbank) expression was not restricted E11.5 (Fig. 19C). to the midbrain at the MHB region. This A blast search against the GenBank is probably due to the species difference EST database (dbest) revealed that Drapc1 in the regulation of WNT signaling. At was conserved in many vertebrate species this stage (HH 15), however, cell lineage as well as in the chordates. To further restriction along the anterior and posterior support our sequence comparisons, we border of the midbrain-R1 boundary studied chicken Drapc1 expression in might be leaky (Jungbluth et al., 2001). early embryos. A marked similarity to the Conservation of the Drapc1 expression mouse Drapc1 expression pattern was pattern between mouse and chicken observed (Fig. 20). However, chicken further suggests DRAPC1 as a putative Drapc1 (or Pgo1, NP_001012959 in component of the WNT pathway.

Figure 19. Genomic location of Drapc1 (or Apcdd1) and its gene expression pattern. For the Drapc1 mRNA detection by in situ hybridization we used three EST clones derived from I.M.A.G.E consortium (IMAGE: 5038498, 374528 and 1764103; marked with blue boxes in B). Expression of Drapc1 at E9.0 (A). In Fgfr1 mutants, Drapc1 expression was decreased at E10.5 and E11.5 (C). Localization of Drapc1 in the midbrain overlapping with the Otx2 expression (D- F). A predicted model of Drapc1 transmembrane helices according to TMpred software (G).

53 Results and Discussion

5. A slowly proliferating narrow 5.1.1. Initial gene expression changes population of cells exists at the of FGF-FGFR signaling components at boundary between the midbrain and E9.5 (III) R1 (III,IV). A downregulation of Fgf15, Spry1, 5.1. Initial changes in gene expression Spry2, Pea3 and Erm was detected in occur close to the midbrain-hindbrain Fgfr1 mutants at E10.5 by microarray boundary in Fgfr1 mutants (III) analysis (Table 5). Thus these genes were analyzed at E9.5 with whole-mount in situ hybridization. Interestingly,

Figure 20. Drapc1 expression pattern in chick embryos at 15 and 17 HH stage (Hamburger– Hamilton stage). Whole mount in situ hybridization with Drapc1 antisense (A,C,D) and sense (B) probes. ba1 & 2, branchial arch 1 & 2; IsO, isthmic organizer; mb, midbrain; ov, otic vesicle; psm. presomitic mesoderm; r1, rhombomere 1; sc, spinal cord; te, telencephalon; ZLI, zona limitans intrathalamica.

54 Results and Discussion expression of Spry1, Spry2, Pea3 and Erm 5.3. Isthmic constriction fails to was downregulated only at the narrow develop in the Fgfr1 CKO mutant boundary region between midbrain and embryos (III) R1 in Fgfr1 mutants compared to wild- type embryos. Conversely, in Fgfr1 The most obvious morphological landmark mutants Fgf15 expression was expanded at the IsO is a constriction which forms at to the narrow boundary region where it is the boundary between midbrain and R1. not normally detected. These analyses of In the mouse embryos, the constriction gene expressions at E9.5 thus suggested starts to form around E9.25 and it is that initial changes in the isthmic gene clearly visible at E10.5. In Fgfr1 mutants, expression in Fgfr1 mutant embryos occur however, the isthmic constriction was close to the midbrain-hindbrain boundary. not evidently recognized at E10.5 and was clearly absent at E11.5. These 5.2. Other FGFRs may mediate morphological defects further suggested FGF signaling at a distance from the that the cell population at the boundary midbrain-hindbrain boundary (III) failed to develop normally and might be required for a stabilization of the IsO. As there are other FGF receptors which can mediate the isthmic FGF signals we 5.4. Expression of cell cycle regulators compared the expressions of Fgfr1, Fgfr2 at the midbrain-R1 region (I, III, IV) and Fgfr3 in E7.5-E9.5 wild type embryos. In addition, expression of Fgfr2 and Fgfr3 Several genes responsible for cell cycle was analyzed in E9.5 Fgfr1 mutants. regulation were differentially expressed Based on these analyses Fgfr1 seems to in Fgfr1 mutants compared with wild-type be the only FGF receptor expressed at the embryos at E10.5 (Tables 5 and 7). These boundary region whereas the expression were validated by in situ hybridization. of Fgfr2 and Fgfr3 in the midbrain- We found out that CyclinD1 and CyclinD2 hindbrain region was graded gradually (or Ccnd1 and 2) and CyclinB2 (or Ccnb1) diminishing towards the boundary and were broadly expressed in the dorsal could not be detected in the cells near the midbrain and hindbrain, but a narrow midbrain-hindbrain border by our in situ domain of CyclinD1, CyclinD2 and hybridization analyses. Recently, Blak et CyclinB2 negative cells were detected al. (2005) reported that Fgfr2 expression in the border between midbrain and is spanning the midbrain-R1 boundary R1. Microarray screen identified also ventrally at E9.5, which is inconsistent with a negative regulator of the cell cycle, our data. This, however, would explain Jumonji (Jmj or Jarid), which expression why Spry1, Sef, Pea3, Erm and possibly was detected as a stripe at the midbrain- Mkp3 were not completely downregulated hindbrain boundary in wild-type embryos in the most ventral midbrain-R1 region in at E9.5 and E10.5. In the Fgfr1 mutant Fgfr1 mutants. Fgfr2 and Fgfr3 expression embryos, Jmj was clearly down-regulated was not upregulated in Fgfr1 mutants at at these both stages. Subsequently, we E9.5. At E10.5, however, only few Erm looked at the expression of other cell and Pea3 positive cells were detected cycle inhibitors such as p21, p27 and in the midbrain of Fgfr1 mutants where p57. Among these only p21 showed a Fgfr2 is thought to be expressed at E10.5 specific expression at the midbrain-R1 (Blak et al., 2005). region between E9.0-E11.5. In the wild-

55 Results and Discussion type embryos p21 was expressed as a were located on both sides of the border narrow stripe at the midbrain-hindbrain between midbrain and R1. The location of boundary. In contrast, p21 positive cells p21 expression was further analyzed using were dispersed both in midbrain and R1 immunostaining with anti-p21 antibody in the Fgfr1 embryos at E9.5. Later, p21 and Otx2 in situ hybridization on parallel expression was completely abolished from sections. Consistent with our in situ Fgfr1 embryos. hybridization analysis, p21 proteins were located on both sides of the midbrain-R1 5.5. Cell proliferation at the border boundary (Fig. 21). In addition, they were between midbrain and R1 (III,IV) profoundly in the pial side of the neural tube. 5.5.1. The location of p21 expressing cells at the border between midbrain and R1 (III,IV)

Our in situ hybridization analysis revealed the p21-positive cell population within a midbrain-R1 region. Interestingly, p21 expression domain appeared to correspond to the region where CyclinD1/2 and CyclinB1 negative cells were observed. Comparative analysis with Otx2, Wnt1, and Fgf8 determined that these cells

Figure 21. Location of the p21 transcripts and proteins in the boundary cells between midbrain and anterior hindbrain (r1) at E11.5. Whole-mount in situ hybridization analysis of p21 expression in wild-type embryos. Parallel coronal sections of E11.5 wild-type embryos were analyzed using anti- p21 antibody (D) and Otx2 radioactive in situ hybridization (E,F). Parasagittal frozen sections immunostained with anti-p21 and anti- PH3 (G-J). The rectangles in C and G show the areas presented in the close-up micrographs D- F and H-J, respectively. Arrowheads show the nuclear p21 expression in the boundary cells (insets in H, I). Arrow in J shows weakly p21- positive cells in the r1, at a distance from the boundary cell population. mb, midbrain; r1, rhombomere 1. Scale bars: 50 μm.

56 Results and Discussion

5.5.2. Slowly proliferating cell Moreover, we could not detect any p21- population at the midbrain-R1 boundary positive cells which were in S phase of the (III,IV) cell cycle (BrdU analysis) between E10.5 and E11.5 (Fig. 22A-C). Thus, these data As several genes regulating the cell cycle suggest that p21-expressing cells are a showed specific expression patterns at slow-proliferating population which is the midbrain-hindbrain boundary we needed to maintain the coherent boundary analyzed the cellular proliferation in the between midbrain and R1. To further midbrain-R1 region. In wild-type embryos characterize the role of p21-positive cells (at E10.5), a narrow cell population was in the midbrain-R1 boundary region we found to show a reduced proliferation analyzed the expression of p21 protein index assayed by the BrdU incorporation together with KI67 that is expressed at all and mitotic cell analyses. Subsequently, phases of the cell cycle but is absent from this cell population was showed to quiescent cells (G0) (Schluter et al., 1993). correlate with the p21 expression at We could not detect any p21 proteins the border between midbrain and R1. overlapping with the KI67-positive cells

Figure 22. Cell proliferation and p21 expression in the boundary cells. Micrographs of coronal sections of the midbrain-R1 region at E11.5 immunostained with anti-p21 and anti-BrdU (A-C). Coronal sections were also immunostained with anti-p21 and anti-KI67 to detect all cell cycle phases (D-J). DAPI was used to stain the nuclei. BrdU injection was given 1 hour before the mice were sacriﬁ ced. Dorsal (A-C,D,E, H-J) and ventral (F,G) regions are shown. Orange arrowheads indicate cells expressing both p21 and KI67 in the midbrain (E). White arrowheads point to the boundary between midbrain and R1. mb, midbrain; r1, rhombomere 1. Scale bars: 50 μm.

57 Results and Discussion at the vicinity of the boundary cells (Fig. More detailed analysis showed that a 22E,G,I-J). However, p21 and KI67 were subset of recombined cells (MHB-Cre/+: co-expressed in the same cells further R26R/+ embryos) overlapped with the p21 away from the boundary cells (IV, Fig. expression in the most posterior midbrain 22E, orange arrowheads). and the most anterior R1. As we could The narrow boundary cell population not detect any p21-expressing cells that is dependent on Fgfr1 signaling and may were in S phase, this observation suggests prevent cell mixing between midbrain that boundary cells may represent a and R1. In Fgfr1 mutants, loss of p21 slowly proliferating stem cell population. expression might interrupt the normal cell Additional hypothesis suggests that cycle kinetics leading into cell sorting the p21-expressing cells might be and/or cell fate changes at the midbrain- differentiating within the Cre-expressing R1 boundary region (see Fig. 24 below). domain between E10.5-E11.5. Consistent with this, both Wnt1 and Otx2 Fate mapping data of MHB-Cre/+: were ectopically expressed in the R1 R26R/+ embryos revealed that the region in Fgfr1 mutants at E10.5. Because midbrain-R1 boundary cells marked by we also detected downregulation of cell MHB-Cre contributed to the midbrain and adhesion molecules (e.g. Pb-cadherin and R1 derivatives such as inferior colliculi Ob-cadherin, Fig. 23), a decrease in the and cerebellar vermis, respectively. p21 expression level might be due to the Interestingly, the majority of recombined loss of these cells. In addition, Fgfr1 is not cells remained near the border between absolutely required for the development midbrain and R1, further suggesting tightly of the boundary cells as scattered p21 controlled lineage restriction between expressing cells were observed at E9.5 in these brain compartments. Fgfr1 mutants. Our results suggest that p21- expressing boundary cells posses a 5.5.3. The contribution of the boundary decreased proliferation potential and cell population to the midbrain-R1 would require re-activation of self- development (IV)

The cell lineage properties of the boundary cell population were studied with a mouse line expressing the Cre-recombinase in a boundary cell –specific fashion (designated MHB-Cre mouse line). We detected a marked similarity between the MHB-Cre activity and the p21-expressing boundary cells between the E10.5-E11.5.

Figure 23. The loss of cell adhesion properties in the Fgfr1 mutant. In addition to the Pb-cadherin (Trokovic et al., 2003), we also detected the downregulation of Ob- cadherin at the midbrain-R1 region in Fgfr1 mutant embryos at E10.5 (B).

58 Results and Discussion renewal pathways. Decrease of Cyclin and Bayer 1997) and in E12.5 mouse (Li D1/2, by limiting the proliferation of early et al., 2002). In these studies the region of boundary cells, would reduce the number slowly proliferating cells was co-localized of cells arising from the early border with the broad Fgf8 (100 μm) expression between midbrain and R1. Although p21 domain in the R1, whereas in our analyses expressing nuclei are located at the basal the reduced cell proliferation rates and side there is fewer cells which are entering p21 (50 μm) expression were detected the cell cycle. In the midbrain-R1 region only in the most anterior part of the Fgf8 a domain of reduced cell proliferation has expression domain (app. 25% of it) and in been reported earlier in E14 rat (Altman the most posterior midbrain.

Figure 24. Maintenance of compartments between midbrain and R1 is dependent on Fgfr1 signaling. In a cell sorting-based model, cells cannot move to the other side of the boundary (Aa, a cell sorting-based mechanism) whereas in a cell fate-based mechanism cells regulate their gene expression to match that of their neighbors (Ab). Different cell adhesion properties of midbrain and R1 cells (e.g. Pb-cadherin) and our data of existence on boundary cell population favor the presence of cell sorting at the IsO. In wild-type embryos this sorting mechanism leads to a sharp Wnt1/Fgf8 expression boundary (A; E10.5). Conversely, In Fgfr1 mutants (Mut) disruption of a coherent signaling from the IsO and subsequent loss of the cell sorting mechanism leads to a cell mixing at the border between midbrain and R1 (B; E10.5). White arrowheads point to the cell mixing in the R1.

59 Conclusions

CONCLUSIONS

The Iso provides as one of the best models expression at the MHB is striking, the to study the cellular mechanism of the actual role of DRAPC1 in the midbrain regionalization and differentiation of the or hindbrain development needs to be brain along the AP axis. Two secreted resolved. factors, FGF8 and WNT1, are required for Prior to this study the role FGFR1 the IsO activity and the development of in the early development of neuronal the midbrain and cerebellum. populations such as DA and SA In this study we have studied the neurons was not addressed. The data role of FGFR1 in the development of from microarray screen showed that a midbrain-R1 region. We compared the number of genes regulating the neuronal gene-expression proﬁ les of wild-type and differentiation were up-regulated in the Fgfr1 CKO mutant embryos in order to Fgfr1 CKO mutants. Further studies discover new molecules regulating the with conditional Fgfr1 mutants provided IsO activity and development of this brain in vivo evidence for the involvement of region. FGF signaling in the development of DA In this work, we have characterized and SA cell-types characteristic for the mouse genes that are involved in the ventral midbrain and R1, respectively. FGF-FGFR1 signal transduction in the In the developing midbrain DA neuron midbrain-R1 region. Our gene-expression population, we suggest that FGF signaling analyses showed that a number of these is required for active proliferation of molecules are regulated by FGFR1 signal the early precursor cells rather than the transduction pathway and they show neural differentiation process itself. FGF specific expression at the IsO. These signaling likely regulates the development molecules are modulating the FGFR of other neuronal precursor pools in the signaling either positively (CNPY1, midbrain-R1 region in a similar fashion. FLRT3) or negatively (MKP3, SEF). However, due to redundancy of different The present study suggests that the FGFRs further analyses of FGFR1-3 mouse CNPY1 is a novel FGFR1 pathway compound mutants are needed to clarify component which is essential for the FGF the role of FGFR signaling in the neuronal signal transduction at the IsO. It has been development. hypothesized that CNPY1 activity is About ﬁ fteen years ago the existence needed to attenuate the FGFR signaling and importance of compartment boundaries through physical interaction with the lipid was discovered in the embryonic rafts on the plasma membrane (Hirate and rhombomeres (Fraser et al. 1990). It was Okamoto 2006). This assumption provides postulated that compartment boundaries in an interesting mechanism how the activity the developing CNS have a dual function of broadly expressed FGFR1 is modulated during development and serve a similar in the vicinity of the IsO. function with the Drosophila embryonic In addition to components of FGFR structures such as imaginal discs. Firstly, signal transduction, our gene-expression neuroepithelial compartment boundaries profiling revealed a transcript, Drapc1, are preventing the intermingling of cells which was found to be a novel component that are fated to contribute to different of the WNT signaling. Although, its parts of the embryo (e.g. midbrain vs.

60 Conclusions

R1). Secondly, compartment boundaries To support our hypothesis of are providing positional information to compartmentalization at the midbrain-R1 adjacent cell populations. In addition to region Brand and co-workers (Langenberg the compartmentalization of rhombomeres, and Brand, 2005) have reported recently a it was also shown that the positional state-of-the-art time-lapse study, in which identity of each rhombomere is defined they mapped the fates of hundreds of cells by the combinatorial expression of in the developing zebrafi sh midbrain-R1 homeotic selector genes such as the Hox region. They proposed that cells do not genes. Thus, after the initial studies of cross the border between midbrain and R1 rhombomeric boundaries there has been during the zebrafi sh development. In mice, continuous search for compartments and the existence of cell lineage restriction at these so-called selector genes in other the midbrain-R1 boundary was supported regions of the developing brain as well. by a study that used genetic fate mapping Before this study was initiated, it with an inducible transgene (Zervas et al., was not clear whether the midbrain-R1 2004). Both these analyses clearly showed compartment boundary really exists. For the presence of cell lineage restriction example, Jungbluth et al. (2001) proposed boundary between the midbrain and R1, that midbrain-R1 boundary does not but it was proposed that in the mouse the represent a compartment boundary in the cell lineage restriction is set up at different chick embryos (HH 11 stage). The loss of phases in ventral and dorsal midbrain-R1 isthmic constriction in the Fgfr1 mutants domains (Zervas et al., 2004). Importantly, by E10.5 together with the deregulation our analysis with the cells marked by of the cell cycle regulators prompted MHB-Cre further suggests limited cell us to study the cellular characteristics mixing within these brain compartments. behind this morphological landmark (see In addition, the role of slowly Fig. 24). Subsequent characterization proliferating boundary cells between of the cell proliferation at the boundary midbrain and R1 has been recently between midbrain and R1 identified a confi rmed by others. Baek et al. (2006) slowly proliferating cell population (p21- showed that a high expression levels of expressing cells) that may contribute to transcription factor Hes1 in the border compartmentalization of this brain region. between midbrain and R1 is needed for Furthermore, in the present study, we the coherent boundary formation. This present that one of the main aspect of the narrow expression domain of high Hes1 boundary between midbrain and R1 is to expression level correlates with the p21- position and stabilize the local signaling expressing cells. centre, the IsO.

61 Acknowledgements

ACKNOWLEDGEMENTS

I am very grateful for the opportunity to work in the lab managed by Docent Juha Partanen in the Institute of Biotechnology (BI), University of Helsinki. Also, I like to thank Juha sincerely for the extensive supervision and his useful advice and tremendous help for my PhD thesis. The working environment in the BI and Developmental Biology Program was excellent; actually I think it has to be one of the best in the world. For that, I want to thank BI’s director, professor Mart Saarma. I sincerely want to thank all the people who have been involved for this thesis: My co-authors and our laboratory’s present and past group members Ras Trokovic, Nina Trokovic, Natalia Sinjushina, Laura Lahti, Jonna Saarimäki-Vire, Paula Peltopuro, Kaia Kala, Emilia Carlsson and Dmitri Chilov. I am very grateful for the help, advice, discussions that I have had with my colleagues during my PhD work in the lab. Kindest appreciation to our lab technicians and research assistants Eija Koivunen, Marjo Virtanen, Päivi Hannuksela and Mona Augustin for their excellent technical expertise. Special recognition for Eija Koivunen whose histological expertise and help for my thesis was outstanding. The members of Prof. Irma Thesleff’s, Prof. Hannu Sariola’s, Doc. Tapio Heino’s, Doc. Marjo Salminen’s and Mikko Frilander’s research groups as well as the whole Developmental Biology Program both in Viikki and in Biomedicum. Especially, I want to thank for the great discussions and chats I have had with Janne Hakanen, Tanja Matilainen, Katja Piltti, Nina Perälä, Martyn James, Mari Palgi, Johanna Pispa and many others. I am also grateful for the D&C publication team and its editors Satu Kuure (editor-in-chief), Mark Tummers (editor-in-chief), Elina Järvinen and other editors in Turku and Oulu as well as other people who have contributed to this students’ journal. I want to acknowledge the Helsinki Biotechnology and Molecular Biology Graduate School (GSBM) and its director Dr. Pekka Lappalainen, co-ordinator Dr. Erkki Raulo and Ms. Anita Tienhaara for their excellent courses, seminars, steering days, Christmas parties and travel grants. Special thanks to Dr. Erkki Raulo for fruitful conversations and ideas. I thank reviewers professor Eero Castren and professor Seppo Vainio for their excellent comments and critical reading of this thesis book. Turku Affymetrix Core Facility and is acknowledged for the help with the microarray analysis. Also, Dr. Mikko Frilander and Dr. Martyn James who gave me valuable information and assistance that I needed for RNA isolation and analysis. For help with the data analysis I want to acknowledge Kai Lähteenmäki and Jarno Tuimala (CSC). To my parents Maarit and Tuomo, brother Tuomas and sisters Mari and Maija for all kind of support, encouragement and discussions. I also want to express grateful thanks for my parents-in-law Raija and Veikko Marttila for their continuous support. I also want to thank my friends Hemmo Jauhiainen, Perttu Salmenperä and Topi Törrönen for their support and friendship. Finally, to my lovely wife Minttu Marttila and to my little daughters Emma and Miusa, for being supportive and loving as well as helping me with many things.

Tomi Jukkola, Helsinki 1.5.2007

62 References

REFERENCES

Acampora,D., Avantaggiato,V., Tuorto,F. and Arber,S., Ladle,D.R., Lin,J.H., Frank,E. and Simeone,A. (1997) Genetic control of brain Jessell,T.M. (2000) ETS gene Er81 controls the morphogenesis through Otx gene dosage formation of functional connections between requirement. Development, 124, 3639-3650. group Ia sensory afferents and motor neurons. Alberi,L., Sgado,P. and Simon,H.H. (2004) Cell, 101, 485-498. Engrailed genes are cell-autonomously Arman,E., Haffner-Krausz,R., Chen,Y., required to prevent apoptosis in mesencephalic Heath,J.K. and Lonai,P. (1998) Targeted dopaminergic neurons. Development, 131, disruption of fi broblast growth factor (FGF) 3229-3236. receptor 2 suggests a role for FGF signaling Altman and Bayer, 1997. J. Altman and S.A. in pregastrulation mammalian development. Bayer, The Development of the Cerebellar Proc.Natl.Acad.Sci.U.S.A., 95, 5082-5087. System: In Relation to Its Evolution, Structure, Assem,M., Schuetz,E.G., Leggas,M., Sun,D., and Functions, CRC Press, Boca Raton, FL Yasuda,K., Reid,G., Zelcer,N., Adachi,M., (1997). Strom,S., Evans,R.M., Moore,D.D., Borst,P. Amoyel,M., Cheng,Y.C., Jiang,Y.J. and and Schuetz,J.D. (2004) Interactions between Wilkinson,D.G. (2005) Wnt1 regulates hepatic Mrp4 and Sult2a as revealed by the neurogenesis and mediates lateral inhibition constitutive androstane receptor and Mrp4 of boundary cell specifi cation in the zebrafi sh knockout mice. J.Biol.Chem., 279, 22250- hindbrain. Development, 132, 775-785. 22257. Andersson,E., Jensen,J.B., Parmar,M., Baek,J.H., Hatakeyama,J., Sakamoto,S., Guillemot,F. and Bjorklund,A. (2006a) Ohtsuka,T. and Kageyama,R. (2006) Persistent Development of the mesencephalic and high levels of Hes1 expression regulate dopaminergic neuron system is compromised boundary formation in the developing central in the absence of neurogenin 2. Development, nervous system. Development, 133, 2467- 133, 507-516. 2476. Andersson,E., Tryggvason,U., Deng,Q., Bailey,A.P., Bhattacharyya,S., Bronner- Friling,S., Alekseenko,Z., Robert,B., Fraser,M. and Streit,A. (2006) Lens Perlmann,T. and Ericson,J. (2006b) specifi cation is the ground state of all sensory Identification of intrinsic determinants of placodes, from which FGF promotes olfactory midbrain dopamine neurons. Cell, 124, 393- identity. Dev.Cell., 11, 505-517. 405. Bally-Cuif,L., Cholley,B. and Wassef,M. Ang,S.L. (2006) Transcriptional control of (1995) Involvement of Wnt-1 in the formation midbrain dopaminergic neuron development. of the mes/metencephalic boundary. Mech. Development, 133, 3499-3506. Dev., 53, 23-34. Ang,S.L., Conlon,R.A., Jin,O. and Rossant,J. Bayer,S.A., Altman,J., Russo,R.J., Dai,X.F. (1994) Positive and negative signals from and Simmons,J.A. (1991) Cell migration in mesoderm regulate the expression of mouse the rat embryonic neocortex. J.Comp.Neurol., Otx2 in ectoderm explants. Development, 120, 307, 499-516. 2979-2989. Ben-Arie,N., Bellen,H.J., Armstrong,D. Ang,S.L. and Rossant,J. (1993) Anterior L., McCall,A.E., Gordadze,P.R., Guo,Q., mesendoderm induces mouse Engrailed genes Matzuk,M.M. and Zoghbi,H.Y. (1997) Math1 in explant cultures. Development, 118, 139- is essential for genesis of cerebellar granule 149. neurons. Nature, 390, 169-172.

63 References

Blaess,S., Corrales,J.D. and Joyner,A.L. Brodski,C., Weisenhorn,D.M., Signore,M., (2006) Sonic hedgehog regulates Gli activator Sillaber,I., Oesterheld,M., Broccoli,V., and repressor functions with spatial and Acampora,D., Simeone,A. and Wurst,W. temporal precision in the mid/hindbrain region. (2003) Location and size of dopaminergic and Development, 133, 1799-1809. serotonergic cell populations are controlled Blak,A.A., Naserke,T., Saarimaki-Vire,J., by the position of the midbrain-hindbrain Peltopuro,P., Giraldo-Velasquez,M., Vogt organizer. J.Neurosci., 23, 4199-4207. Weisenhorn,D.M., Prakash,N., Sendtner,M., Buckles,G.R., Thorpe,C.J., Ramel,M.C. and Partanen,J. and Wurst,W. (2007) Fgfr2 and Lekven,A.C. (2004) Combinatorial Wnt Fgfr3 are not required for patterning and control of zebrafish midbrain-hindbrain maintenance of the midbrain and anterior boundary formation. Mech.Dev., 121, 437- hindbrain. Dev.Biol., 303, 231-243. 447. Blak,A.A., Naserke,T., Weisenhorn,D.M., Cappello,S., Attardo,A., Wu,X., Iwasato,T., Prakash,N., Partanen,J. and Wurst,W. (2005) Itohara,S., Wilsch-Brauninger,M., Eilken,H. Expression of Fgf receptors 1, 2, and 3 in the M., Rieger,M.A., Schroeder,T.T., Huttner,W. developing mid- and hindbrain of the mouse. B., Brakebusch,C. and Gotz,M. (2006) The Dev.Dyn., 233, 1023-1030. Rho-GTPase cdc42 regulates neural progenitor Blum,M. (1998) A null mutation in TGF-alpha fate at the apical surface. Nat.Neurosci., 9, leads to a reduction in midbrain dopaminergic 1099-1107. neurons in the substantia nigra. Nat.Neurosci., Castella,P., Sawai,S., Nakao,K., Wagner,J. 1, 374-377. A. and Caudy,M. (2000) HES-1 repression Bottcher,R.T., Pollet,N., Delius,H. and of differentiation and proliferation in PC12 Niehrs,C. (2004) The transmembrane protein cells: role for the helix 3-helix 4 domain in XFLRT3 forms a complex with FGF receptors transcription repression. Mol.Cell.Biol., 20, and promotes FGF signalling. Nat.Cell Biol., 6170-6183. 6, 38-44. Chao,J. and Nestler,E.J. (2004) Molecular Bouwmeester,T., Kim,S., Sasai,Y., Lu,B. and neurobiology of drug addiction. Annu.Rev. De Robertis,E.M. (1996) Cerberus is a head- Med., 55, 113-132. inducing secreted factor expressed in the Chellaiah,A., Yuan,W., Chellaiah,M. and anterior endoderm of Spemann’s organizer. Ornitz,D.M. (1999) Mapping ligand binding Nature, 382, 595-601. domains in chimeric ﬁ broblast growth factor Brandeis,M., Rosewell,I., Carrington,M., receptor molecules. Multiple regions determine Crompton,T., Jacobs,M.A., Kirk,J., Gannon,J. ligand binding speciﬁ city. J.Biol.Chem., 274, and Hunt,T. (1998) Cyclin B2-null mice 34785-34794. develop normally and are fertile whereas Chi,C.L., Martinez,S., Wurst,W. and Martin,G. cyclin B1-null mice die in utero. Proc.Natl. R. (2003) The isthmic organizer signal FGF8 Acad.Sci.U.S.A., 95, 4344-4349. is required for cell survival in the prospective Brault,V., Moore,R., Kutsch,S., Ishibashi,M., midbrain and cerebellum. Development, 130, Rowitch,D.H., McMahon,A.P., Sommer,L., 2633-2644. Boussadia,O. and Kemler,R. (2001) Chizhikov,V.V., Lindgren,A.G., Currle,D. Inactivation of the beta-catenin gene by Wnt1- S., Rose,M.F., Monuki,E.S. and Millen,K. Cre-mediated deletion results in dramatic J. (2006) The roof plate regulates cerebellar brain malformation and failure of craniofacial cell-type specification and proliferation. development. Development, 128, 1253-1264. Development, 133, 2793-2804. Broccoli,V., Boncinelli,E. and Wurst,W. (1999) Chotteau-Lelievre,A., Dolle,P., Peronne,V., The caudal limit of Otx2 expression positions Coutte,L., de Launoit,Y. and Desbiens,X. the isthmic organizer. Nature, 401, 164-168. (2001) Expression patterns of the Ets

64 References transcription factors from the PEA3 group Crossley,P.H., Martinez,S. and Martin,G. during early stages of mouse development. R. (1996) Midbrain development induced by Mech.Dev., 108, 191-195. FGF8 in the chick embryo. Nature, 380, 66- Christen,B. and Slack,J.M. (1999) Spatial 68. response to ﬁ broblast growth factor signalling Darnell,D.K. and Schoenwolf,G.C. (1997) in Xenopus embryos. Development, 126, 119- Vertical induction of engrailed-2 and other 125. region-specific markers in the early chick Ciani,L. and Salincookeas,P.C. (2005) WNTs embryo. Dev.Dyn., 209, 45-58. in the vertebrate nervous system: from Davidson,D., Graham,E., Sime,C. and Hill,R. patterning to neuronal connectivity. Nat.Rev. (1988) A gene with sequence similarity to Neurosci., 6, 351-362. Drosophila engrailed is expressed during the Ciemerych,M.A., Kenney,A.M., Sicinska,E., development of the neural tube and vertebrae Kalaszczynska,I., Bronson,R.T., Rowitch,D. in the mouse. Development, 104, 305-316. H., Gardner,H. and Sicinski,P. (2002) Davis,C.A. and Joyner,A.L. (1988) Expression Development of mice expressing a single D- patterns of the homeo box-containing genes type cyclin. Genes Dev., 16, 3277-3289. En-1 and En-2 and the proto-oncogene int- Ciemerych,M.A. and Sicinski,P. (2005) Cell 1 diverge during mouse development. Genes cycle in mouse development. Oncogene, 24, Dev., 2, 1736-1744. 2877-2898. de Bruin,A., Wu,L., Saavedra,H.I., Clarke,A.R., Maandag,E.R., van Roon,M., van Wilson,P., Yang,Y., Rosol,T.J., Weinstein,M., der Lugt,N.M., van der Valk,M., Hooper,M.L., Robinson,M.L. and Leone,G. (2003) Rb Berns,A. and te Riele,H. (1992) Requirement function in extraembryonic lineages suppresses for a functional Rb-1 gene in murine apoptosis in the CNS of Rb-deficient mice. development. Nature, 359, 328-330. Proc.Natl.Acad.Sci.U.S.A., 100, 6546-6551. Colvin,J.S., Bohne,B.A., Harding,G.W., Deng,C., Wynshaw-Boris,A., Zhou,F., Kuo,A. McEwen,D.G. and Ornitz,D.M. (1996) and Leder,P. (1996) Fibroblast growth factor Skeletal overgrowth and deafness in mice receptor 3 is a negative regulator of bone lacking fibroblast growth factor receptor 3. growth. Cell, 84, 911-921. Nat.Genet., 12, 390-397. Deng,C., Zhang,P., Harper,J.W., Elledge,S.J. Cooke,J.E., Kemp,H.A. and Moens,C.B. and Leder,P. (1995) Mice lacking p21CIP1/ (2005) EphA4 Is Required for Cell Adhesion WAF1 undergo normal development, but are and Rhombomere-Boundary Formation in the defective in G1 checkpoint control. Cell, 82, Zebraﬁ sh. Current Biology, 15, 536-542. 675-684. Corson,L.B., Yamanaka,Y., Lai,K.M. and Deng,C.X., Wynshaw-Boris,A., Shen,M. Rossant,J. (2003) Spatial and temporal M., Daugherty,C., Ornitz,D.M. and Leder,P. patterns of ERK signaling during mouse (1994) Murine FGFR-1 is required for embryogenesis. Development, 130, 4527- early postimplantation growth and axial 4537. organization. Genes Dev., 8, 3045-3057. Crossley,P.H. and Martin,G.R. (1995) Ding,Y.Q., Marklund,U., Yuan,W., Yin,J., The mouse Fgf8 gene encodes a family of Wegman,L., Ericson,J., Deneris,E., Johnson,R. polypeptides and is expressed in regions L. and Chen,Z.F. (2003) Lmx1b is essential for that direct outgrowth and patterning in the the development of serotonergic neurons. Nat. developing embryo. Development, 121, 439- Neurosci., 6, 933-938. 451.

65 References

Eblaghie,M.C., Lunn,J.S., Dickinson,R. Fraser,S., Keynes,R. and Lumsden,A. (1990) J., Munsterberg,A.E., Sanz-Ezquerro,J.J., Segmentation in the chick embryo hindbrain Farrell,E.R., Mathers,J., Keyse,S.M., Storey,K. is defi ned by cell lineage restrictions. Nature, and Tickle,C. (2003) Negative feedback 344, 431-435. regulation of FGF signaling levels by Pyst1/ Friedman,G.C. and O’Leary,D.D. (1996) MKP3 in chick embryos. Curr.Biol., 13, 1009- Retroviral misexpression of engrailed genes in 1018. the chick optic tectum perturbs the topographic Echevarria,D., Martinez,S., Marques,S., targeting of retinal axons. J.Neurosci., 16, Lucas-Teixeira,V. and Belo,J.A. (2005) Mkp3 5498-5509. is a negative feedback modulator of Fgf8 Frilander,M.J. and Steitz,J.A. (1999) Initial signaling in the mammalian isthmic organizer. recognition of U12-dependent introns requires Dev.Biol., 277, 114-128. both U11/5’ splice-site and U12/branchpoint Echevarria,D., Vieira,C., Gimeno,L. and interactions. Genes Dev., 13, 851-863. Martinez,S. (2003) Neuroepithelial secondary Furthauer,M., Reifers,F., Brand,M., Thisse,B. organizers and cell fate specification in the and Thisse,C. (2001) sprouty4 acts in vivo as a developing brain. Brain Res.Brain Res.Rev., feedback-induced antagonist of FGF signaling 43, 179-191. in zebrafi sh. Development, 128, 2175-2186. Ekker,S.C., Ungar,A.R., Greenstein,P., von Galaktionov,K. and Beach,D. (1991) Specifi c Kessler,D.P., Porter,J.A., Moon,R.T. and activation of cdc25 tyrosine phosphatases by Beachy,P.A. (1995) Patterning activities of B-type cyclins: evidence for multiple roles of vertebrate hedgehog proteins in the developing mitotic cyclins. Cell, 67, 1181-1194. eye and brain. Curr.Biol., 5, 944-955. Gavalas,A. and Krumlauf,R. (2000) Retinoid Estivill-Torrus,G., Pearson,H., van signalling and hindbrain patterning. Curr.Opin. Heyningen,V., Price,D.J. and Rashbass,P. Genet.Dev., 10, 380-386. (2002) Pax6 is required to regulate the cell cycle and the rate of progression from German,D.C. and Manaye,K.F. (1993) symmetrical to asymmetrical division in Midbrain dopaminergic neurons (nuclei A8, mammalian cortical progenitors. Development, A9, and A10): three-dimensional reconstruction 129, 455-466. in the rat. J.Comp.Neurol., 331, 297-309. Fantl,V., Stamp,G., Andrews,A., Rosewell,I. Gimeno,L., Brulet,P. and Martinez,S. (2003) and Dickson,C. (1995) Mice lacking cyclin Study of Fgf15 gene expression in developing D1 are small and show defects in eye and mouse brain. Gene Expr.Patterns, 3, 473-481. mammary gland development. Genes Dev., 9, Gimeno,L., Hashemi,R., Brulet,P. and 2364-2372. Martinez,S. (2002) Analysis of Fgf15 Farah,M.H., Olson,J.M., Sucic,H.B., Hume,R. expression pattern in the mouse neural tube. I., Tapscott,S.J. and Turner,D.L. (2000) Brain Res.Bull., 57, 297-299. Generation of neurons by transient expression Glover,J.C., Renaud,J.S. and Rijli,F.M. (2006) of neural bHLH proteins in mammalian cells. Retinoic acid and hindbrain patterning. Development, 127, 693-702. J.Neurobiol., 66, 705-725. Farkas,L.M., Dunker,N., Roussa,E., Gotz,M. and Huttner,W.B. (2005) The cell Unsicker,K. and Krieglstein,K. (2003) biology of neurogenesis. Nat.Rev.Mol.Cell Transforming growth factor-beta(s) are Biol., 6, 777-788. essential for the development of midbrain Guo,S., Brush,J., Teraoka,H., Goddard,A., dopaminergic neurons in vitro and in vivo. Wilson,S.W., Mullins,M.C. and Rosenthal,A. J.Neurosci., 23, 5178-5186. (1999) Development of noradrenergic neurons in the zebrafi sh hindbrain requires BMP, FGF8,

66 References and the homeodomain protein soulless/Phox2a. prospective neurons in the chick. Nature, 375, Neuron, 24, 555-566. 787-790. Guo, C., Qiu, H.Y, Huang, Y., Chen, H., Yang, Hermanson,E., Joseph,B., Castro,D., R.Q., Chen, S.D., Johnson, R.L., Chen, Z.F., Lindqvist,E., Aarnisalo,P., Wallen,A., Ding, Y.Q. (2007) Lmx1b is essential for Fgf8 Benoit,G., Hengerer,B., Olson,L. and and Wnt1 expression in the isthmic organizer Perlmann,T. (2003) Nurr1 regulates dopamine during tectum and cerebellum development in synthesis and storage in MN9D dopamine mice. Development, 134, 317-325. cells. Exp.Cell Res., 288, 324-334. Guthrie,S., Butcher,M. and Lumsden,A. Hirata,H., Tomita,K., Bessho,Y. and (1991) Patterns of cell division and interkinetic Kageyama,R. (2001) Hes1 and Hes3 regulate nuclear migration in the chick embryo maintenance of the isthmic organizer and hindbrain. J.Neurobiol., 22, 742-754. development of the mid/hindbrain. EMBO J., Guthrie,S., Muchamore,I., Kuroiwa,A., 20, 4454-4466. Marshall,H., Krumlauf,R. and Lumsden,A. Hirate,Y. and Okamoto,H. (2006) Canopy1, a (1992) Neuroectodermal autonomy of Hox- novel regulator of FGF signaling around the 2.9 expression revealed by rhombomere midbrain-hindbrain boundary in zebrafish. transpositions. Nature, 356, 157-159. Curr.Biol., 16, 421-427. Haines,B.P., Wheldon,L.M., Summerbell,D., Hirsch,E.C., Faucheux,B., Damier,P., Mouatt- Heath,J.K. and Rigby,P.W. (2006) Regulated Prigent,A. and Agid,Y. (1997) Neuronal expression of FLRT genes implies a functional vulnerability in Parkinson’s disease. J.Neural role in the regulation of FGF signalling during Transm.Suppl., 50, 79-88. mouse development. Dev.Biol., 297, 14-25. Hirsch,M.R., Tiveron,M.C., Guillemot,F., Hamburger,V. (1988) Ontogeny of Brunet,J.F. and Goridis,C. (1998) Control neuroembryology. J.Neurosci., 8, 3535-3540. of noradrenergic differentiation and Phox2a Hanafusa,H., Torii,S., Yasunaga,T. and expression by MASH1 in the central and Nishida,E. (2002) Sprouty1 and Sprouty2 peripheral nervous system. Development, 125, provide a control mechanism for the Ras/ 599-608. MAPK signalling pathway. Nat.Cell Biol., 4, Holder,N. and Klein,R. (1999) Eph receptors 850-858. and ephrins: effectors of morphogenesis. Harland,R. and Gerhart,J. (1997) Formation Development, 126, 2033-2044. and function of Spemann’s organizer. Annu. Houart,C., Caneparo,L., Heisenberg,C., Rev.Cell Dev.Biol., 13, 611-667. Barth,K., Take-Uchi,M. and Wilson,S. (2002) Hemmati-Brivanlou,A., Stewart,R.M. and Establishment of the telencephalon during Harland,R.M. (1990) Region-speciﬁ c neural gastrulation by local antagonism of Wnt induction of an engrailed protein by anterior signaling. Neuron, 35, 255-265. notochord in Xenopus. Science, 250, 800-802. Huard,J.M., Forster,C.C., Carter,M.L., Hendricks,T.J., Fyodorov,D.V., Wegman,L. Sicinski,P. and Ross,M.E. (1999) Cerebellar J., Lelutiu,N.B., Pehek,E.A., Yamamoto,B., histogenesis is disturbed in mice lacking cyclin Silver,J., Weeber,E.J., Sweatt,J.D. and D2. Development, 126, 1927-1935. Deneris,E.S. (2003) Pet-1 ETS gene plays Hynes,M. and Rosenthal,A. (1999) a critical role in 5-HT neuron development Speciﬁ cation of dopaminergic and serotonergic and is required for normal anxiety-like and neurons in the vertebrate CNS. Curr.Opin. aggressive behavior. Neuron, 37, 233-247. Neurobiol., 9, 26-36. Henrique,D., Adam,J., Myat,A., Chitnis,A., Inglis-Broadgate,S.L., Thomson,R.E., Lewis,J. and Ish-Horowicz,D. (1995) Pellicano,F., Tartaglia,M.A., Pontikis,C. Expression of a Delta homologue in C., Cooper,J.D. and Iwata,T. (2005) FGFR3

67 References regulates brain size by controlling progenitor Hoekstra,M.F. and Evans,R.M. (1992) A cell proliferation and apoptosis during mouse cdc25 homolog is differentially and embryonic development. Dev.Biol., 279, 73- developmentally expressed. Genes Dev., 6, 85. 578-590. Irving,C. and Mason,I. (2000) Signalling Karaulanov,E.E., Bottcher,R.T. and Niehrs,C. by FGF8 from the isthmus patterns anterior (2006) A role for fibronectin-leucine-rich hindbrain and establishes the anterior limit of transmembrane cell-surface proteins in Hox gene expression. Development, 127, 177- homotypic cell adhesion. EMBO Rep., 7, 283- 186. 290. Isaacs,H.V., Pownall,M.E. and Slack,J.M. Katahira,T., Sato,T., Sugiyama,S., Okafuji,T., (1998) Regulation of Hox gene expression and Araki,I., Funahashi,J. and Nakamura,H. (2000) posterior development by the Xenopus caudal Interaction between Otx2 and Gbx2 deﬁ nes the homologue Xcad3. EMBO J., 17, 3413-3427. organizing center for the optic tectum. Mech. Ishibashi,M. and McMahon,A.P. (2002) Dev., 91, 43-52. A sonic hedgehog-dependent signaling Kawakami,Y., Rodriguez-Leon,J., Koth,C. relay regulates growth of diencephalic and M., Buscher,D., Itoh,T., Raya,A., Ng,J.K., mesencephalic primordia in the early mouse Esteban,C.R., Takahashi,S., Henrique,D., embryo. Development, 129, 4807-4819. Schwarz,M.F., Asahara,H. and Izpisua Itasaki,N. and Nakamura,H. (1996) A role Belmonte,J.C. (2003) MKP3 mediates the for gradient en expression in positional cellular response to FGF8 signalling in the speciﬁ cation on the optic tectum. Neuron, 16, vertebrate limb. Nat.Cell Biol., 5, 513-519. 55-62. Kele,J., Simplicio,N., Ferri,A.L., Mira,H., Jacobson and Rao, 2005. Rao, Mahendra S.; Guillemot,F., Arenas,E. and Ang,S.L. (2006) Jacobson, Marcus (Eds.), Developmental Neurogenin 2 is required for the development Neurobiology, Springer, New York. of ventral midbrain dopaminergic neurons. Development, 133, 495-505. Jaszai,J., Reifers,F., Picker,A., Langenberg,T. and Brand,M. (2003) Isthmus-to-midbrain Khokha,M.K., Yeh,J., Grammer,T.C. and transformation in the absence of midbrain- Harland,R.M. (2005) Depletion of three BMP hindbrain organizer activity. Development, antagonists from Spemann’s organizer leads to 130, 6611-6623. a catastrophic loss of dorsal structures. Dev. Cell., 8, 401-411. Jho,E.H., Zhang,T., Domon,C., Joo,C.K., Freund,J.N. and Costantini,F. (2002) Wnt/beta- Kiecker,C. and Lumsden,A. (2005) catenin/Tcf signaling induces the transcription Compartments and their boundaries in of Axin2, a negative regulator of the signaling vertebrate brain development. Nat.Rev. pathway. Mol.Cell.Biol., 22, 1172-1183. Neurosci., 6, 553-564. Jungbluth,S., Larsen,C., Wizenmann,A. and Kimmel,R.A., Turnbull,D.H., Blanquet,V., Lumsden,A. (2001) Cell mixing between the Wurst,W., Loomis,C.A. and Joyner,A.L. embryonic midbrain and hindbrain. Curr.Biol., (2000) Two lineage boundaries coordinate 11, 204-207. vertebrate apical ectodermal ridge formation. Genes Dev., 14, 1377-1389. Kabos,P., Kabosova,A. and Neuman,T. (2002) Blocking HES1 expression initiates Kiyokawa,H., Kineman,R.D., Manova- GABAergic differentiation and induces the Todorova,K.O., Soares,V.C., Hoffman,E. expression of p21(CIP1/WAF1) in human S., Ono,M., Khanam,D., Hayday,A.C., neural stem cells. J.Biol.Chem., 277, 8763- Frohman,L.A. and Koff,A. (1996) Enhanced 8766. growth of mice lacking the cyclin-dependent kinase inhibitor function of p27(Kip1). Cell, Kakizuka,A., Sebastian,B., Borgmeyer,U., 85, 721-732. Hermans-Borgmeyer,I., Bolado,J., Hunter,T.,

68 References

Klingensmith,J., Ang,S.L., Bachiller,D. Langenberg,T. and Brand,M. (2005) Lineage and Rossant,J. (1999) Neural induction and restriction maintains a stable organizer cell patterning in the mouse in the absence of the population at the zebrafi sh midbrain-hindbrain node and its derivatives. Dev.Biol., 216, 535- boundary. Development, 132, 3209-3216. 549. Lee,K.J. and Jessell,T.M. (1999) The Kovalenko,D., Yang,X., Nadeau,R.J., specification of dorsal cell fates in the Harkins,L.K. and Friesel,R. (2003) Sef vertebrate central nervous system. Annu.Rev. inhibits fi broblast growth factor signaling by Neurosci., 22, 261-294. inhibiting FGFR1 tyrosine phosphorylation Lekven,A.C., Buckles,G.R., Kostakis,N. and and subsequent ERK activation. J.Biol.Chem., Moon,R.T. (2003) Wnt1 and wnt10b function 278, 14087-14091. redundantly at the zebrafish midbrain- Kowalczyk,A., Filipkowski,R.K., Rylski,M., hindbrain boundary. Dev.Biol., 254, 172-187. Wilczynski,G.M., Konopacki,F.A., Jaworski,J., Leyns,L., Bouwmeester,T., Kim,S.H., Ciemerych,M.A., Sicinski,P. and Kaczmarek,L. Piccolo,S. and De Robertis,E.M. (1997) Frzb- (2004) The critical role of cyclin D2 in adult 1 is a secreted antagonist of Wnt signaling neurogenesis. J.Cell Biol., 167, 209-213. expressed in the Spemann organizer. Cell, 88, Kozar,K., Ciemerych,M.A., Rebel,V.I., 747-756. Shigematsu,H., Zagozdzon,A., Sicinska,E., Li,J.Y. and Joyner,A.L. (2001) Otx2 and Gbx2 Geng,Y., Yu,Q., Bhattacharya,S., Bronson,R. are required for refi nement and not induction of T., Akashi,K. and Sicinski,P. (2004) Mouse mid-hindbrain gene expression. Development, development and cell proliferation in the 128, 4979-4991. absence of D-cyclins. Cell, 118, 477-491. Li,J.Y., Lao,Z. and Joyner,A.L. (2002) Kudoh,T., Wilson,S.W. and Dawid,I.B. (2002) Changing requirements for Gbx2 in Distinct roles for Fgf, Wnt and retinoic development of the cerebellum and acid in posteriorizing the neural ectoderm. maintenance of the mid/hindbrain organizer. Development, 129, 4335-4346. Neuron, 36, 31-43. Kukoski,R., Blonigen,B., Macri,E., Renshaw,A. Lillevali,K., Matilainen,T., Karis,A. and A., Hoffman,M., Loda,M. and Datta,M.W. Salminen,M. (2004) Partially overlapping (2003) p27 and cyclin E/D2 associations in expression of Gata2 and Gata3 during inner testicular germ cell tumors: implications for ear development. Dev.Dyn., 231, 775-781. tumorigenesis. Appl.Immunohistochem.Mol. Morphol., 11, 138-143. Lin,J.C., Cai,L. and Cepko,C.L. (2001) The external granule layer of the developing chick Lacy,S.E., Bonnemann,C.G., Buzney,E.A. cerebellum generates granule cells and cells of and Kunkel,L.M. (1999) Identification of the isthmus and rostral hindbrain. J.Neurosci., FLRT1, FLRT2, and FLRT3: a novel family 21, 159-168. of transmembrane leucine-rich repeat proteins. Genomics, 62, 417-426. Lin,W., Furthauer,M., Thisse,B., Thisse,C., Jing,N. and Ang,S.L. (2002) Cloning of the Lai,L. and Tan,T.M. (2002) Role of glutathione mouse Sef gene and comparative analysis of in the multidrug resistance protein 4 (MRP4/ its expression with Fgf8 and Spry2 during ABCC4)-mediated efflux of cAMP and embryogenesis. Mech.Dev., 113, 163-168. resistance to purine analogues. Biochem.J., 361, 497-503. Liu,A. and Joyner,A.L. (2001a) Early anterior/ posterior patterning of the midbrain and Laing,M.A., Coonrod,S., Hinton,B.T., cerebellum. Annu.Rev.Neurosci., 24, 869-896. Downie,J.W., Tozer,R., Rudnicki,M.A. and Hassell,J.A. (2000) Male sexual dysfunction Liu,A. and Joyner,A.L. (2001b) EN and GBX2 in mice bearing targeted mutant alleles of play essential roles downstream of FGF8 in the PEA3 ets gene. Mol.Cell.Biol., 20, 9337- patterning the mouse mid/hindbrain region. 9345. Development, 128, 181-191. 69 References

Liu,D., Matzuk,M.M., Sung,W.K., Guo,Q., expression of the homeobox gene en. Neuron, Wang,P. and Wolgemuth,D.J. (1998) Cyclin 6, 971-981. A1 is required for meiosis in the male mouse. Martinez-Barbera,J.P., Signore,M., Boyl,P. Nat.Genet., 20, 377-380. P., Puelles,E., Acampora,D., Gogoi,R., Liu,G., Loraine,A.E., Shigeta,R., Cline,M., Schubert,F., Lumsden,A. and Simeone,A. Cheng,J., Valmeekam,V., Sun,S., Kulp,D. and (2001) Regionalisation of anterior Siani-Rose,M.A. (2003a) NetAffx: Affymetrix neuroectoderm and its competence in probesets and annotations. Nucleic Acids Res., responding to forebrain and midbrain inducing 31, 82-86. activities depend on mutual antagonism Liu,Y., Jiang,H., Crawford,H.C. and Hogan,B. between OTX2 and GBX2. Development, 128, L. (2003b) Role for ETS domain transcription 4789-4800. factors Pea3/Erm in mouse lung development. Matzuk,M.M., Lu,N., Vogel,H., Sellheyer,K., Dev.Biol., 261, 10-24. Roop,D.R. and Bradley,A. (1995) Multiple Lobe,C.G., Koop,K.E., Kreppner,W., defects and perinatal death in mice deﬁ cient in Lomeli,H., Gertsenstein,M. and Nagy,A. follistatin. Nature, 374, 360-363. (1999) Z/AP, a double reporter for cre- McGrew,L.L., Takemaru,K., Bates,R. and mediated recombination. Dev.Biol., 208, 281- Moon,R.T. (1999) Direct regulation of the 292. Xenopus engrailed-2 promoter by the Wnt Logan,C., Wizenmann,A., Drescher,U., signaling pathway, and a molecular screen for Monschau,B., Bonhoeffer,F. and Lumsden,A. Wnt-responsive genes, conﬁ rm a role for Wnt (1996) Rostral optic tectum acquires caudal signaling during neural patterning in Xenopus. characteristics following ectopic engrailed Mech.Dev., 87, 21-32. expression. Curr.Biol., 6, 1006-1014. McMahon,A.P. and Bradley,A. (1990) The Lumsden,A. and Keynes,R. (1989) Segmental Wnt-1 (int-1) proto-oncogene is required for patterns of neuronal development in the chick development of a large region of the mouse hindbrain. Nature, 337, 424-428. brain. Cell, 62, 1073-1085. Lumsden,A. and Krumlauf,R. (1996) McMahon,A.P., Joyner,A.L., Bradley,A. and Patterning the vertebrate neuraxis. Science, McMahon,J.A. (1992) The midbrain-hindbrain 274, 1109-1115. phenotype of Wnt-1-/Wnt-1- mice results from stepwise deletion of engrailed-expressing cells Maden,M. (2006) Retinoids and spinal cord by 9.5 days postcoitum. Cell, 69, 581-595. development. J.Neurobiol., 66, 726-738. McMahon,J.A., Takada,S., Zimmerman,L.B., Mailleux,A.A., Tefft,D., Ndiaye,D., Itoh,N., Fan,C.M., Harland,R.M. and McMahon,A.P. Thiery,J.P., Warburton,D. and Bellusci,S. (1998) Noggin-mediated antagonism of BMP (2001) Evidence that SPROUTY2 functions as signaling is required for growth and patterning an inhibitor of mouse embryonic lung growth of the neural tube and somite. Genes Dev., 12, and morphogenesis. Mech.Dev., 102, 81-94. 1438-1452. Maretto,S., Cordenonsi,M., Dupont,S., Mellitzer,G., Xu,Q. and Wilkinson,D.G. Braghetta,P., Broccoli,V., Hassan,A.B., (1999) Eph receptors and ephrins restrict cell Volpin,D., Bressan,G.M. and Piccolo,S. (2003) intermingling and communication. Nature, Mapping Wnt/beta-catenin signaling during 400, 77-81. mouse development and in colorectal tumors. Proc.Natl.Acad.Sci.U.S.A., 100, 3299-3304. Millen,K.J., Wurst,W., Herrup,K. and Joyner,A.L. (1994) Abnormal embryonic Martinez,S., Wassef,M. and Alvarado- cerebellar development and patterning of Mallart,R.M. (1991) Induction of a postnatal foliation in two mouse Engrailed-2 mesencephalic phenotype in the 2-day-old mutants. Development, 120, 695-706. chick prosencephalon is preceded by the early

70 References

Millet,S., Campbell,K., Epstein,D.J., Losos,K., Mutoh,H., Naya,F.J., Tsai,M.J. and Leiter,A. Harris,E. and Joyner,A.L. (1999) A role for B. (1998) The basic helix-loop-helix protein Gbx2 in repression of Otx2 and positioning BETA2 interacts with p300 to coordinate the mid/hindbrain organizer. Nature, 401, 161- differentiation of secretin-expressing 164. enteroendocrine cells. Genes Dev., 12, 820- Minowada,G., Jarvis,L.A., Chi,C.L., 830. Neubuser,A., Sun,X., Hacohen,N., Krasnow,M. Nelson,W.J. and Nusse,R. (2004) Convergence A. and Martin,G.R. (1999) Vertebrate Sprouty of Wnt, beta-catenin, and cadherin pathways. genes are induced by FGF signaling and can Science, 303, 1483-1487. cause chondrodysplasia when overexpressed. Nguyen,L., Besson,A., Heng,J.I., Development, 126, 4465-4475. Schuurmans,C., Teboul,L., Parras,C., Mohammadi,M., Dikic,I., Sorokin,A., Philpott,A., Roberts,J.M. and Guillemot,F. Burgess,W.H., Jaye,M. and Schlessinger,J. (2006) p27kip1 independently promotes (1996) Identification of six novel neuronal differentiation and migration in the autophosphorylation sites on fi broblast growth cerebral cortex. Genes Dev., 20, 1511-1524. factor receptor 1 and elucidation of their Ning,Y., Hoang,B., Schuller,A.G., Cominski,T. importance in receptor activation and signal P., Hsu,M.S., Wood,T.L. and Pintar,J.E. (2007) transduction. Mol.Cell.Biol., 16, 977-989. Delayed mammary gland involution in mice Mohammadi,M., Honegger,A.M., Rotin,D., with mutation of the insulin-like growth factor Fischer,R., Bellot,F., Li,W., Dionne,C.A., binding protein 5 gene. Endocrinology, . Jaye,M., Rubinstein,M. and Schlessinger,J. Ohkubo,Y., Chiang,C. and Rubenstein,J.L. (1991) A tyrosine-phosphorylated carboxy- (2002) Coordinate regulation and synergistic terminal peptide of the fi broblast growth factor actions of BMP4, SHH and FGF8 in the rostral receptor (Flg) is a binding site for the SH2 prosencephalon regulate morphogenesis of the domain of phospholipase C-gamma 1. Mol. telencephalic and optic vesicles. Neuroscience, Cell.Biol., 11, 5068-5078. 111, 1-17. Mori,S., Sugawara, S., Kikuchi,T., Tanji,M., Olander,S., Nordstrom,U., Patthey,C. and Narumi,O., Stoykova,A., Nishikawa,S.I. and Edlund,T. (2006) Convergent Wnt and FGF Yokota,Y. (1999) The leukemic oncogene tal- signaling at the gastrula stage induce the 2 is expressed in the developing mouse brain. formation of the isthmic organizer. Mech.Dev., Brain Res.Mol.Brain.Res., 642, 199-210. 123, 166-176. Morin,X., Cremer,H., Hirsch,M.R., Kapur,R. Olsen,S.K., Li,J.Y., Bromleigh,C., P., Goridis,C. and Brunet,J.F. (1997) Defects Eliseenkova,A.V., Ibrahimi,O.A., Lao,Z., in sensory and autonomic ganglia and absence Zhang,F., Linhardt,R.J., Joyner,A.L. and of locus coeruleus in mice deficient for the Mohammadi,M. (2006) Structural basis by homeobox gene Phox2a. Neuron, 18, 411- which alternative splicing modulates the 423. organizer activity of FGF8 in the brain. Genes Motoyama,J., Kitajima,K., Kojima,M., Dev., 20, 185-198. Kondo,S. and Takeuchi,T. (1997) Ong,S.H., Hadari,Y.R., Gotoh,N., Guy,G.R., Organogenesis of the liver, thymus and spleen Schlessinger,J. and Lax,I. (2001) Stimulation is affected in jumonji mutant mice. Mech.Dev., of phosphatidylinositol 3-kinase by fi broblast 66, 27-37. growth factor receptors is mediated by Murphy,M., Stinnakre,M.G., Senamaud- coordinated recruitment of multiple docking Beaufort,C., Winston,N.J., Sweeney,C., proteins. Proc.Natl.Acad.Sci.U.S.A., 98, 6074- Kubelka,M., Carrington,M., Brechot,C. and 6079. Sobczak-Thepot,J. (1997) Delayed early Ornitz,D.M. (2000) FGFs, heparan sulfate and embryonic lethality following disruption of the FGFRs: complex interactions essential for murine cyclin A2 gene. Nat.Genet., 15, 83-86. development. Bioessays, 22, 108-112. 71 References

Ornitz,D.M. and Itoh,N. (2001) Fibroblast Prakash,N., Brodski,C., Naserke,T., growth factors. Genome Biol., 2, Puelles,E., Gogoi,R., Hall,A., Panhuysen,M., REVIEWS3005. Echevarria,D., Sussel,L., Weisenhorn,D. Panhuysen,M., Vogt Weisenhorn,D.M., M., Martinez,S., Arenas,E., Simeone,A. and Blanquet,V., Brodski,C., Heinzmann,U., Wurst,W. (2006) A Wnt1-regulated genetic Beisker,W. and Wurst,W. (2004) Effects network controls the identity and fate of of Wnt1 signaling on proliferation in the midbrain-dopaminergic progenitors in vivo. developing mid-/hindbrain region. Mol.Cell. Development, 133, 89-98. Neurosci., 26, 101-111. Prakash,N. and Wurst,W. (2006) Development Parker,S.B., Eichele,G., Zhang,P., Rawls,A., of dopaminergic neurons in the mammalian Sands,A.T., Bradley,A., Olson,E.N., Harper,J. brain. Cell Mol.Life Sci., 63, 187-206. W. and Elledge,S.J. (1995) p53-independent Preger,E., Ziv,I., Shabtay,A., Sher,I., Tsang,M., expression of p21Cip1 in muscle and other Dawid,I.B., Altuvia,Y. and Ron,D. (2004) terminally differentiating cells. Science, 267, Alternative splicing generates an isoform of 1024-1027. the human Sef gene with altered subcellular Partanen,J., Schwartz,L. and Rossant,J. (1998) localization and specifi city. Proc.Natl.Acad. Opposite phenotypes of hypomorphic and Sci.U.S.A., 101, 1229-1234. Y766 phosphorylation site mutations reveal a Puelles,E., Acampora,D., Lacroix,E., function for Fgfr1 in anteroposterior patterning Signore,M., Annino,A., Tuorto,F., Filosa,S., of mouse embryos. Genes Dev., 12, 2332- Corte,G., Wurst,W., Ang,S.L. and Simeone,A. 2344. (2003) Otx dose-dependent integrated control Pattyn,A., Goridis,C. and Brunet,J.F. (2000) of antero-posterior and dorso-ventral patterning Specification of the central noradrenergic of midbrain. Nat.Neurosci., 6, 453-460. phenotype by the homeobox gene Phox2b. Puelles,E., Annino,A., Tuorto,F., Usiello,A., Mol.Cell.Neurosci., 15, 235-243. Acampora,D., Czerny,T., Brodski,C., Ang,S. Pattyn,A., Simplicio,N., van Doorninck,J.H., L., Wurst,W. and Simeone,A. (2004) Otx2 Goridis,C., Guillemot,F. and Brunet,J.F. (2004) regulates the extent, identity and fate of Ascl1/Mash1 is required for the development neuronal progenitor domains in the ventral of central serotonergic neurons. Nat.Neurosci., midbrain. Development, 131, 2037-2048. 7, 589-595. Raible,F. and Brand,M. (2001) Tight Peters,K.G., Marie,J., Wilson,E., Ives,H.E., transcriptional control of the ETS domain Escobedo,J., Del Rosario,M., Mirda,D. and factors Erm and Pea3 by Fgf signaling during Williams,L.T. (1992) Point mutation of an early zebrafi sh development. Mech.Dev., 107, FGF receptor abolishes phosphatidylinositol 105-117. turnover and Ca2+ fl ux but not mitogenesis. Raible,F. and Brand,M. (2004) Divide et Nature, 358, 678-681. Impera – The Midbrain-Hindbrain Boundary Pinson,K.I., Brennan,J., Monkley,S., Avery,B. and Its Organizer. Trends in Neurosciences, J. and Skarnes,W.C. (2000) An LDL-receptor- 27, 727-734. related protein mediates Wnt signalling in Rhinn,M., Lun,K., Luz,M., Werner,M. and mice. Nature, 407, 535-538. Brand,M. (2005) Positioning of the midbrain- Placzek,M. (1995) The role of the notochord hindbrain boundary organizer through global and fl oor plate in inductive interactions. Curr. posteriorization of the neuroectoderm mediated Opin.Genet.Dev., 5, 499-506. by Wnt8 signaling. Development, 132, 1261- 1272. Prabhu,S., Ignatova,A., Park,S.T. and Sun,X.H. (1997) Regulation of the expression of cyclin- Riley,B.B., Chiang,M.Y., Storch,E.M., dependent kinase inhibitor p21 by E2A and Id Heck,R., Buckles,G.R. and Lekven,A.C. proteins. Mol.Cell.Biol., 17, 5888-5896. (2004) Rhombomere boundaries are Wnt signaling centers that regulate metameric 72 References patterning in the zebrafish hindbrain. Dev. Scholpp,S. and Brand,M. (2004) Endocytosis Dyn., 231, 278-291. controls spreading and effective signaling Rizzoti,K., Brunelli,S., Carmignac,D., range of Fgf8 protein. Curr.Biol., 14, 1834- Thomas,P.Q., Robinson,I.C. and Lovell- 1841. Badge,R. (2004) SOX3 is required during the Sesack,S.R. and Carr,D.B. (2002) Selective formation of the hypothalamo-pituitary axis. prefrontal cortex inputs to dopamine cells: Nat.Genet., 36, 247-255. implications for schizophrenia. Physiol.Behav., Roberts,J.M. (1999) Evolving ideas about 77, 513-517. cyclins. Cell, 98, 129-132. Shamim,H., Mahmood,R., Logan,C., Roberts,J.M. and Sherr,C.J. (2003) Bared Doherty,P., Lumsden,A. and Mason,I. (1999) essentials of CDK2 and cyclin E. Nat.Genet., Sequential roles for Fgf4, En1 and Fgf8 35, 9-10. in specification and regionalisation of the midbrain. Development, 126, 945-959. Robinson,M., Parsons Perez,M.C., Tebar,L., Palmer,J., Patel,A., Marks,D., Sheasby,A., De Sharrocks,A.D. (2001) The ETS-domain Felipe,C., Coffi n,R., Livesey,F.J. and Hunt,S.P. transcription factor family. Nat.Rev.Mol.Cell (2004) FLRT3 is expressed in sensory neurons Biol., 2, 827-837. after peripheral nerve injury and regulates Sherr,C.J. (1994) G1 phase progression: neurite outgrowth. Mol.Cell.Neurosci., 27, cycling on cue. Cell, 79, 551-555. 202-214. Sherr,C.J. and Roberts,J.M. (1999) CDK Roehl,H. and Nusslein-Volhard,C. (2001) inhibitors: positive and negative regulators of Zebrafi sh pea3 and erm are general targets of G1-phase progression. Genes Dev., 13, 1501- FGF8 signaling. Curr.Biol., 11, 503-507. 1512. Ruiz i Altaba,A. and Jessell,T.M. (1993) Shih,J. and Fraser,S.E. (1996) Characterizing Midline cells and the organization of the the zebrafi sh organizer: microsurgical analysis vertebrate neuraxis. Curr.Opin.Genet.Dev., 3, at the early-shield stage. Development, 122, 633-640. 1313-1322. Ryan,P.J., Paterno,G.D. and Gillespie,L. Shimizu,T., Bae,Y.K. and Hibi,M. (2006) Cdx- L. (1998) Identification of phosphorylated Hox code controls competence for responding proteins associated with the fi broblast growth to Fgfs and retinoic acid in zebrafi sh neural factor receptor type I during early Xenopus tissue. Development, 133, 4709-4719. development. Biochem.Biophys.Res. Shtutman,M., Zhurinsky,J., Simcha,I., Commun., 244, 763-767. Albanese,C., D’Amico,M., Pestell,R. and Ben- Saucedo-Cardenas,O., Quintana-Hau,J.D., Ze’ev,A. (1999) The cyclin D1 gene is a target Le,W.D., Smidt,M.P., Cox,J.J., De Mayo,F., of the beta-catenin/LEF-1 pathway. Proc.Natl. Burbach,J.P. and Conneely,O.M. (1998) Acad.Sci.U.S.A., 96, 5522-5527. Nurr1 is essential for the induction of the Sicinska,E., Aifantis,I., Le Cam,L., Swat,W., dopaminergic phenotype and the survival Borowski,C., Yu,Q., Ferrando,A.A., Levin,S. of ventral mesencephalic late dopaminergic D., Geng,Y., von Boehmer,H. and Sicinski,P. precursor neurons. Proc.Natl.Acad.Sci.U.S.A., (2003) Requirement for cyclin D3 in 95, 4013-4018. lymphocyte development and T cell leukemias. Schluter,C., Duchrow,M., Wohlenberg,C., Cancer.Cell., 4, 451-461. Becker,M.H., Key,G., Flad,H.D. and Gerdes,J. Sicinski,P., Donaher,J.L., Parker,S.B., (1993) The cell proliferation-associated Li,T., Fazeli,A., Gardner,H., Haslam,S.Z., antigen of antibody Ki-67: a very large, Bronson,R.T., Elledge,S.J. and Weinberg,R. ubiquitous nuclear protein with numerous A. (1995) Cyclin D1 provides a link between repeated elements, representing a new kind of development and oncogenesis in the retina and cell cycle-maintaining proteins. J.Cell Biol., breast. Cell, 82, 621-630. 123, 513-522. 73 References

Simon,H.H., Bhatt,L., Gherbassi,D., Sgado,P. Suda,Y., Matsuo,I. and Aizawa,S. (1997) and Alberi,L. (2003) Midbrain dopaminergic Cooperation between Otx1 and Otx2 genes neurons: determination of their developmental in developmental patterning of rostral brain. fate by transcription factors. Ann.N.Y.Acad. Mech.Dev., 69, 125-141. Sci., 991, 36-47. Takahashi,M., Fujita,M., Furukawa,Y., Simon,H.H., Saueressig,H., Wurst,W., Hamamoto,R., Shimokawa,T., Miwa,N., Goulding,M.D. and O’Leary,D.D. (2001) Fate Ogawa,M. and Nakamura,Y. (2002) Isolation of midbrain dopaminergic neurons controlled of a novel human gene, APCDD1, as a by the engrailed genes. J.Neurosci., 21, 3126- direct target of the beta-Catenin/T-cell factor 3134. 4 complex with probable involvement in Simon,H.H., Scholz,C. and O’Leary,D.D. colorectal carcinogenesis. Cancer Res., 62, (2005) Engrailed genes control developmental 5651-5656. fate of serotonergic and noradrenergic neurons Takemoto,T., Uchikawa,M., Kamachi,Y. and in mid- and hindbrain in a gene dose-dependent Kondoh,H. (2006) Convergence of Wnt and manner. Mol.Cell.Neurosci., 28, 96-105. FGF signals in the genesis of posterior neural Simon,H.H., Thuret,S. and Alberi,L. (2004) plate through activation of the Sox2 enhancer Midbrain dopaminergic neurons: control of N-1. Development, 133, 297-306. their cell fate by the engrailed transcription Takeuchi,T., Kojima,M., Nakajima,K. and factors. Cell Tissue Res., 318, 53-61. Kondo,S. (1999) jumonji gene is essential for Smidt,M.P., Smits,S.M., Bouwmeester,H., the neurulation and cardiac development of Hamers,F.P., van der Linden,A.J., Hellemons,A. mouse embryos with a C3H/He background. J., Graw,J. and Burbach,J.P. (2004) Early Mech.Dev., 86, 29-38. developmental failure of substantia nigra Takeuchi,T., Yamazaki,Y., Katoh-Fukui,Y., dopamine neurons in mice lacking the Tsuchiya,R., Kondo,S., Motoyama,J. and homeodomain gene Pitx3. Development, 131, Higashinakagawa,T. (1995) Gene trap capture 1145-1155. of a novel mouse gene, jumonji, required for Smits,S.M., Ponnio,T., Conneely,O. neural tube formation. Genes Dev., 9, 1211- M., Burbach,J.P. and Smidt,M.P. (2003) 1222. Involvement of Nurr1 in specifying the Temple,S. and Raff,M.C. (1986) Clonal neurotransmitter identity of ventral midbrain analysis of oligodendrocyte development in dopaminergic neurons. Eur.J.Neurosci., 18, culture: evidence for a developmental clock 1731-1738. that counts cell divisions. Cell, 44, 773-779. Soriano,P. (1999) Generalized lacZ expression Thomas,K.R. and Capecchi,M.R. (1990) with the ROSA26 Cre reporter strain. Nat. Targeted disruption of the murine int-1 proto- Genet., 21, 70-71. oncogene resulting in severe abnormalities in Stern,C.D. (2001) Initial patterning of the midbrain and cerebellar development. Nature, central nervous system: how many organizers? 346, 847-850. Nat.Rev.Neurosci., 2, 92-98. Thomas,K.R., Musci,T.S., Neumann,P.E. and Storey,K.G., Goriely,A., Sargent,C.M., Capecchi,M.R. (1991) Swaying is a mutant Brown,J.M., Burns,H.D., Abud,H.M. and allele of the proto-oncogene Wnt-1. Cell, 67, Heath,J.K. (1998) Early posterior neural 969-976. tissue is induced by FGF in the chick embryo. Trokovic,R., Trokovic,N., Hernesniemi,S., Development, 125, 473-484. Pirvola,U., Vogt Weisenhorn,D.M., Rossant,J., Streit,A., Berliner,A.J., Papanayotou,C., McMahon,A.P., Wurst,W. and Partanen,J. Sirulnik,A. and Stern,C.D. (2000) Initiation (2003) FGFR1 is independently required of neural induction by FGF signalling before in both developing mid- and hindbrain for gastrulation. Nature, 406, 74-78. sustained response to isthmic signals. EMBO J., 22, 1811-1823. 74 References

Tsang,M. and Dawid,I.B. (2004) Promotion Weidinger,G., Thorpe,C.J., Wuennenberg- and attenuation of FGF signaling through the Stapleton,K., Ngai,J. and Moon,R.T. (2005) Ras-MAPK pathway. Sci.STKE, 2004, pe17. The Sp1-related transcription factors sp5 and Tsang,M., Friesel,R., Kudoh,T. and Dawid,I.B. sp5-like act downstream of Wnt/beta-catenin (2002) Identifi cation of Sef, a novel modulator signaling in mesoderm and neuroectoderm of FGF signalling. Nat.Cell Biol., 4, 165-169. patterning. Curr.Biol., 15, 489-500. Tsang,M., Maegawa,S., Kiang,A., Habas,R., Weinstein,D.C., Ruiz i Altaba,A., Chen,W. Weinberg,E. and Dawid,I.B. (2004) A role S., Hoodless,P., Prezioso,V.R., Jessell,T.M. for MKP3 in axial patterning of the zebrafi sh and Darnell,J.E.,Jr. (1994) The winged-helix embryo. Development, 131, 2769-2779. transcription factor HNF-3 beta is required for notochord development in the mouse embryo. Urayama,A., Yamada,S., Kimura,R., Zhang,J. Cell, 78, 575-588. and Watanabe,Y. (2002) Neuroprotective effect and brain receptor binding of taltirelin, Wianny,F., Real,F.X., Mummery,C.L., a novel thyrotropin-releasing hormone (TRH) Van Rooijen,M., Lahti,J., Samarut,J. and analogue, in transient forebrain ischemia of Savatier,P. (1998) G1-phase regulators, cyclin C57BL/6J mice. Life Sci., 72, 601-607. D1, cyclin D2, and cyclin D3: up-regulation at gastrulation and dynamic expression during Vieira,C., Garda,A.L., Shimamura,K. and neurulation. Dev.Dyn., 212, 49-62. Martinez,S. (2005) Thalamic development induced by Shh in the chick embryo. Dev. Wilkinson DG, Green J (1990) In situ Biol., 284, 351-363. hybridization and the three-dimensional reconstruction of serial sections. In Copp Wallen,A. and Perlmann,T. (2003) AJ, Cockroft DL, eds. Postimplantation Transcriptional control of dopamine neuron Mammalian Embryos. New York, Oxford development. Ann.N.Y.Acad.Sci., 991, 48-60. University Press, 155-171. Wallen,A., Zetterstrom,R.H., Solomin,L., Wilson,S.I., Graziano,E., Harland,R., Jessell,T. Arvidsson,M., Olson,L. and Perlmann,T. M. and Edlund,T. (2000) An early requirement (1999) Fate of mesencephalic AHD2- for FGF signalling in the acquisition of neural expressing dopamine progenitor cells in cell fate in the chick embryo. Curr.Biol., 10, NURR1 mutant mice. Exp.Cell Res., 253, 421-429. 737-746. Wingate,R.J. and Lumsden,A. (1996) Wang,F., Kan,M., Yan,G., Xu,J. and Persistence of rhombomeric organisation in McKeehan,W.L. (1995) Alternately spliced the postsegmental hindbrain. Development, NH2-terminal immunoglobulin-like Loop I in 122, 2143-2152. the ectodomain of the fi broblast growth factor (FGF) receptor 1 lowers affinity for both Winnier,G., Blessing,M., Labosky,P.A. and heparin and FGF-1. J.Biol.Chem., 270, 10231- Hogan,B.L. (1995) Bone morphogenetic 10235. protein-4 is required for mesoderm formation and patterning in the mouse. Genes Dev., 9, Wassarman,K.M., Lewandoski,M., 2105-2116. Campbell,K., Joyner,A.L., Rubenstein,J. L., Martinez,S. and Martin,G.R. (1997) Wizenmann,A. and Lumsden,A. (1997) Specification of the anterior hindbrain and Segregation of rhombomeres by differential establishment of a normal mid/hindbrain chemoaffinity. Mol.Cell.Neurosci., 9, 448- organizer is dependent on Gbx2 gene function. 459. Development, 124, 2923-2934. Wodarz,A., Stewart,D.B., Nelson,W.J. and Waters,S.T. and Lewandoski,M. (2006) A Nusse,R. (2006) Wingless signaling modulates threshold requirement for Gbx2 levels in cadherin-mediated cell adhesion in Drosophila hindbrain development. Development, 133, imaginal disc cells. J.Cell.Sci., 119, 2425- 1991-2000. 2434.

75 References

Wu,L., de Bruin,A., Saavedra,H.I., Starovic,M., Yang,R.B., Ng,C.K., Wasserman,S.M., Trimboli,A., Yang,Y., Opavska,J., Wilson,P., Komuves,L.G., Gerritsen,M.E. and Topper,J. Thompson,J.C., Ostrowski,M.C., Rosol,T. N. (2003) A novel interleukin-17 receptor-like J., Woollett,L.A., Weinstein,M., Cross,J.C., protein identified in human umbilical vein Robinson,M.L. and Leone,G. (2003) Extra- endothelial cells antagonizes basic fi broblast embryonic function of Rb is essential for growth factor-induced signaling. J.Biol.Chem., embryonic development and viability. Nature, 278, 33232-33238. 421, 942-947. Ye,W., Shimamura,K., Rubenstein,J.L., Wurst,W., Auerbach,A.B. and Joyner,A. Hynes,M.A. and Rosenthal,A. (1998) FGF L. (1994) Multiple developmental defects and Shh signals control dopaminergic and in Engrailed-1 mutant mice: an early mid- serotonergic cell fate in the anterior neural hindbrain deletion and patterning defects in plate. Cell, 93, 755-766. forelimbs and sternum. Development, 120, Yusoff,P., Lao,D.H., Ong,S.H., Wong,E. 2065-2075. S., Lim,J., Lo,T.L., Leong,H.F., Fong,C.W. Wurst,W. and Bally-Cuif,L. (2001) Neural and Guy,G.R. (2002) Sprouty2 inhibits the plate patterning: upstream and downstream of Ras/MAP kinase pathway by inhibiting the the isthmic organizer. Nat.Rev.Neurosci., 2, activation of Raf. J.Biol.Chem., 277, 3195- 99-108. 3201. Xu,J., Liu,Z. and Ornitz,D.M. (2000) Temporal Zervas,M., Millet,S., Ahn,S. and Joyner,A. and spatial gradients of Fgf8 and Fgf17 L. (2004) Cell behaviors and genetic lineages regulate proliferation and differentiation of of the mesencephalon and rhombomere 1. midline cerebellar structures. Development, Neuron, 43, 345-357. 127, 1833-1843. Zetterstrom,R.H., Solomin,L., Jansson,L., Yamada,M., Saga,Y., Shibusawa,N., Hirato,J., Hoffer,B.J., Olson,L. and Perlmann,T. (1997) Murakami,M., Iwasaki,T., Hashimoto,K., Dopamine neuron agenesis in Nurr1-defi cient Satoh,T., Wakabayashi,K., Taketo,M.M. and mice. Science, 276, 248-250. Mori,M. (1997) Tertiary hypothyroidism and Zhang,X., Ibrahimi,O.A., Olsen,S.K., hyperglycemia in mice with targeted disruption Umemori,H., Mohammadi,M. and Ornitz,D. of the thyrotropin-releasing hormone gene. M. (2006) Receptor specificity of the Proc.Natl.Acad.Sci.U.S.A., 94, 10862-10867. fi broblast growth factor family. The complete mammalian FGF family. J.Biol.Chem., 281, 15694-15700.