View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Developmental Biology 297 (2006) 141–157 www.elsevier.com/locate/ydbio

FGF regulated gene-expression and neuronal differentiation in the developing –hindbrain region ⁎ Tomi Jukkola a,1, Laura Lahti a,1, Thorsten Naserke b,c, Wolfgang Wurst b,c, Juha Partanen a,

a Institute of Biotechnology, Viikki Biocenter, P.O. Box 56, 00014 University of Helsinki, Finland b GSF-National Research Center for Environment and Health, Technical University Munich, Institute of Developmental Genetics, Ingolstaedter Landstrasse 1, 85764 Munich/Neuherberg, Germany c Max-Planck-Institute of Psychiatry, Kraepelinstrasse 2, 80804 Munich, Germany Received for publication 1 July 2005; revised 3 April 2006; accepted 3 May 2006 Available online 13 May 2006

Abstract

The neuroectodermal tissue close to the midbrain–hindbrain boundary (MHB) is an important secondary organizer in the developing . This so-called isthmic organizer (IsO) secretes signaling molecules, such as fibroblast growth factors (FGFs), which regulate cellular survival, patterning and proliferation in the midbrain and rhombomere 1 (R1) of the hindbrain. We have previously shown that FGF-receptor 1 (FGFR1) is required for the normal development of this brain region in the mouse embryo. Here, we have compared the gene expression profiles of midbrain–R1 tissues from wild-type embryos and conditional Fgfr1 mutants, in which FGFR1 is inactivated in the midbrain and R1. Loss of Fgfr1 results in the downregulation of several genes expressed close to the midbrain–hindbrain boundary and in the disappearance of gene expression gradients in the midbrain and anterior hindbrain. Our screen identified 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. Interestingly, the neuronal progenitors in midbrain and R1 show different responses to the loss of signaling through FGFR1. © 2006 Elsevier Inc. All rights reserved.

Keywords: FGFR1; Isthmic organizer; Midbrain; Hindbrain; Neuronal progenitors; Proliferation; Midbrain dopaminergic ; Raphe nuclei; Locus coeruleus; Microarray

Introduction Katahira et al., 2000; Millet et al., 1999). MHB also coincides with the boundary of cell-lineage restriction (Zervas et al., Isthmic organizer (IsO) acts as an signaling center that 2004). Expression of IsO signal Wnt1 gets restricted to the regulates the development of the midbrain and rhombomere 1 caudal midbrain and Fgf8 to the rostral R1 (Li and Joyner, (R1) by secreted molecules, such as fibroblast growth factors 2001). The early development of midbrain–hindbrain region (FGFs) FGF8/17/18 and WNT1 (reviewed in Echevarria et al., requires both WNT1 and FGF8. In Wnt1 null mutant and Fgf8 2003; Liu and Joyner, 2001; Nakamura, 2001; Wurst and Bally- midbrain–R1 mutant embryos, cells in the midbrain and R1 die Cuif, 2001). These signals are thought to regulate cellular apoptotically around E8.5 (Chi et al., 2003; McMahon and survival, proliferation and differentiation in the developing mid- Bradley, 1990; Thomas and Capecchi, 1990). and hindbrain. In addition to cell survival, IsO also appears to regulate Where the expression of two transcription factors precursor cell proliferation in the MHB region. Studies in Otx2 and Gbx2 abuts defines the position of the midbrain– zebrafish and mouse have indicated that cells in the posterior hindbrain boundary (MHB) and IsO (Broccoli et al., 1999; midbrain and anterior R1 express basic helix–loop–helix transcription factors belonging to Hairy/E(spl) family and ⁎ Corresponding author. Fax: +358 9 191 59366. undergo neuronal differentiation later than the cells further E-mail address: [email protected] (J. Partanen). away from the IsO (Geling et al., 2003; Hirata et al., 2001; 1 Equal contribution. Ninkovic et al., 2005). Lineage tracing studies in zebrafish

0012-1606/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2006.05.002 142 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 suggest that these progenitor cells contribute to the structures the early MHB tissue. Instead, the loss of Fgfr1 leads to a of midbrain and anterior hindbrain (Tallafuss and Bally-Cuif, failure to establish a coherent midbrain–hindbrain boundary 2003). In the mouse, premature differentiation of neuronal and maintain IsO specific gene-expression after E9.5 precursors is prevented by transcription factors Hes1 and (Trokovic et al., 2003, 2005). Fgfr2 and Fgfr3 expression Hes3 (Hirata et al., 2001). These genes also maintain the diminishes towards the MHB. However, they are still expression of isthmus specific genes, such as Pax2/5, Fgf8 expressed in the midbrain–R1 region, especially in the and Wnt1. Later, FGF8 and FGF17 also redundantly ventral part where they may contribute to IsO signaling participate in keeping the cells of dorsal R1 in a proliferative (Blak et al., 2005; Trokovic et al., 2005). state thus inhibiting neuronal differentiation in the vermis In this study, we further define the molecular and cellular primordium (Xu et al., 2000). The control of cellular defects in Fgfr1 conditional mutant embryos. To identify proliferation is also involved in the development of a coherent genes, which are expressed at the midbrain–R1 region, we boundary cell population at the midbrain–hindbrain border. have compared gene-expression profiles of E10.5 wild-type We have previously characterized a slowly proliferating and conditional Fgfr1 mutant embryos. In our screen, we have narrow boundary cell population which is located on either identified several new genes that may be involved in the side of the MHB (Trokovic et al., 2005). These FGFR1- development of the midbrain and anterior hindbrain. Observed dependent cells express specific cell cycle regulators and cell- changes in gene expressions suggest alterations in the adhesion molecules and are needed for the development of the maintenance of the neuronal progenitors on both sides of the isthmic constriction. Compared to Hes3, Fgf8 or Wnt1 MHB. Surprisingly, midbrain DA and hindbrain SA positive cells, this boundary cell population is much narrower progenitor populations appear to have different requirements in the A–P dimension and includes only a subset of these for signaling through FGFR1. cells. IsO also regulates differentiation of brain structures and Material and methods nuclei in the midbrain and R1. In ventral midbrain, nuclei Mice and genotyping including dopaminergic (DA) neurons of the substantia nigra (SN) and ventral tegmental area (VTA) develop. By contrast, En1-Cre (Kimmel et al., 2000) and Fgfr1Flox (Trokovic et al., 2005) alleles serotonergic neurons (SA) develop in the ventral hindbrain, have been described previously. Mutant embryos were generated by crossing and noradrenergic (NA) neurons of locus coeruleus (LC) and En1-Cre/+; Fgfr1Flox/+ males with Fgfr1Flox/Flox females in outbred (129sv/ICR) nucleus subcoeruleus (NSC) will form in the dorsal R1. The background. Follistatin mutant mice and their genotyping has been described location and the size of the colliculi, as well as earlier (Matzuk et al., 1995). Embryonic age (E) was estimated by counting the somites or considering noon of the day of a vaginal plug as E0.5. For genotyping DA, SA and NA cell populations is determined by the of the mice and embryos, see Trokovic et al. (2003). All the experiments were position of the IsO (Brodski et al., 2003; Lam et al., 2003; approved by the committee of experimental animal research of the University of Prakash et al., 2006; Andersson et al., 2006). According to Helsinki. the current view, the development of DA and SA neurons is induced by both SHH from the floor plate and FGF8 from the Microarray analysis β isthmus (Ye et al., 1998). Other factors, such as TGF- s are Tissue containing posterior midbrain and anterior rhombomere 1 of E10.5 also needed for the induction and maintenance of midbrain wild-type and Fgfr1 mutant embryos was dissected (Supplementary Fig. 1 to DA neurons (Farkas et al., 2003). In addition to FGF8 show approximate area). Individual tissues were snap frozen in liquid nitrogen signals, FGF4 is needed for SA neuron induction in vitro and pooled after the genotype was confirmed (5–6 tissue samples/pool). Tissues cultured R1 explants, but its role in vivo has not been were lysed in pooled groups (2 wild-type pools, 2 mutant pools) and total RNA was extracted from the tissue lysate using Trizol reagent (Invitrogen) and demonstrated (Ye et al., 1998). Inductive signals in the RNeasy kit (Qiagen) clean-up procedure. Approximately 5 μg of total RNA specification of LC include BMPs from the roof plate and from each of the four independent samples was processed to produce FGF8 from the isthmus (Lam et al., 2003). biotinylated cRNA targets, which were hybridized to Affymetrix Genechip Four FGF receptors (FGFR1–4) have been identified in mouse U74A version 2 arrays following standard Affymetrix procedures (http:// vertebrates. Fgfr1–3 are expressed during the early brain www.affymetrix.com). 12 arrays were used and each demonstrated control parameters within development in the mouse (Blak et al., 2005; Liu et al., recommended limits (Raw Q < 30, background < 85, GAPDH 3′/5′ ratios below 2003; Trokovic et al., 2005; Walshe and Mason, 2000). Our 1.5). To allow comparison between U74A arrays, each was analyzed using earlier results have shown that FGFR1 is required for the global scaling with a target intensity of 300. The scaling factors used to development of anterior hindbrain and posterior midbrain normalize to the target value were within 4-fold of each other in all comparisons (Trokovic et al., 2003). Fgfr1 conditional knock out mice, (A chip: 0.683 to 1.226; B chip 1.597 to 3.782; C chip: 1.612 to 4.027). Affymetrix Microarray Suite version 5.0 (MAS 5.0) software was used to make where FGFR1 has been inactivated in the midbrain and R1 each pairwise comparison between the two wild-type and the two mutant arrays. at E8.5, lack vermis of the cerebellum and inferior colliculi. MAS 5.0 data were then exported to Lotus Notes database, in which the ‘Signal Ventrally, the major neuronal subtypes are present, but Log Ratios’ were converted to fold changes. Using the default MAS 5.0 settings, appear disorganized. The phenotype of the conditional Fgfr1 a probe set with a detection P value less than 0.05 were considered present. Only mutants is much less severe than the phenotype of transcripts called as present (detection P value < 0.05) in at least one wild-type sample, and showing a 0.74-fold or smaller and 1.41-fold or greater difference conditional Fgf8 mutants produced by a similar approach (change P value for up-regulated genes 0 to 0.0025; down-regulated genes 0.997 (Chi et al., 2003). In contrast to Fgf8 mutants, the phenotype to 1) between the wild-type and mutant MHB samples in at least 3 out of the 4 observed in Fgfr1 mutant mice is not due to the apoptosis in comparisons were included. We chose 0.75-fold change as our empirical cut-off T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 143

Table 1 Table 1 (continued) Genes downregulated in Fgfr1 mutants Gene/EST Affymetrix Fold Validation Probe for Gene/EST Affymetrix Fold Validation Probe for probe set ID change by in situ in situ probe set ID change by in situ in situ hybridisation hybridisation hybridisation hybridisation CN04 162901_i_at 0.71 – Fgf17 98730_at 0.06 Fig. 1B Xu et al. (2000) (6430527G18Rik) Spry1 163956_at 0.24 Fig. 1C Zhang et al. Sox21 167903_at 0.72 Data not (2001) shown En1 96523_at 0.29 Fig. 2A Davis and D7Wsu128e 103861_s_at 0.74 – Joyner (1988) Jmj 94341_at 0.74 Fig. 2G IMAGE 6406875 Fgf8 97742_s_at 0.30 Fig. 1B Crossley and Martin (1995) Spry2 116938_at 0.30 Fig. 1C Zhang et al. (2001) for down-regulated genes according our validation analysis by whole-mount in Trh 102665_at 0.31 Fig. 2E IMAGE 559851 situ hybridization. Corresponding cut-off value for up-regulated genes was 1.41- Cnpy1 138472_at 0.35 Fig. 2C AI853839 (3′) fold. The data were sorted by average fold-change to produce the two lists of IMAGE:6390940 gene expression levels shown in Tables 1 and 2. This filtering of the data means (5′) that some genes with spatially and temporally restricted patterns may be Fgf18 95316_at 0.35 Fig. 1B Xu et al. (2000) excluded. Especially this is likely to be the case for genes with low expression Fgfr1 97509_f_at 0.38 Data not Trokovic et al. levels. shown (2003) Tnfrsf19 (Trade) 160670_at 0.39 Fig. 2K Pispa et al. In situ hybridization analyses (2003) Sef 133830_at 0.41 Fig. 1C IMAGE 1178421 Whole mount mRNA in situ hybridization analyses of E9.5–E11.5 day Erm (Etv5) 163173_at 0.41 Fig. 1D IMAGE 3674281 embryos (n≥3) were performed as described (Henrique et al., 1995). Mkp3 (Dusp6) 93285_at 0.42 Fig. 1C IMAGE 874051 Digoxigenin (DIG) -labeled antisense and sense RNA probes were synthesized Nfia 130461_at 0.48 – from linearized plasmid DNAs using DIG-labeling mix (Roche) and T3, T7 and Mrp4 (Abcc4) 111137_at 0.50 Fig. 2F IMAGE 1153158 SP6 RNA polymerases. After whole-mount staining, some of the embryos were Pax5 95890_r_at 0.50 Fig. 2D IMAGE 3333164 dehydrated, embedded in paraffin and sectioned coronally at 10 μm. Sections EST6 163120_at 0.51 – were counterstained with nuclear red (Vector). Radioactive in situ hybridization Pcx 171076_i_at 0.51 – on 5 μm sagittal sections was done according to Wilkinson and Green (1990) En2 98338_at 0.52 Fig. 2B Liu and Joyner using 35S labeled RNA probes. Plasmids used for validating microarray results (2001) are listed in Tables 1 and 2. Other probes used for in situ analysis were Ngn2 Sfrp2 93503_at 0.52 Fig. 2L IMAGE 4487469 (IMAGE 2922473), Phox2a (IMAGE 534970), Pitx3 (IMAGE 482871), Pet1 Fgf15 97721_at 0.54 Fig. 1B McWhirter et al. (clone UI-M-BH3-avj-b-02-0-UI.s1), Nurr1 (Wallen et al., 1999), GATA3 (1997) (Lillevali et al., 2004), Mash1 (IMAGE 6415061), Hes3 (clone UI-R-BO1-ajt-e- Atp1a1 93797_g_at 0.54 – 02-0-UI.r1), Aldh1 (Hermanson et al., 2003), and Otx2 (Acampora et al., 1997). 4921506J03Rik 99163_at 0.54 – Igfbp5 100566_at 0.54 Fig. 2J IMAGE 318625 Nfib 160859_s_at 0.54 – RhoA (Arha2) 101112_g_at 0.55 – Table 2 Sox3 92264_at 0.56 Fig. 2I IMAGE 368804 Genes upregulated in Fgfr1 mutants Drapc1 96132_at 0.59 Fig. 2M Jukkola et al. (2004) Gene/ Afymetrix Fold Validation by Probe for in situ Flrt3 110370_at 0.60 Fig. 1C IMAGE 5702881 EST probe set ID change in situ hybridisation hybridisation Epcs3 (Ftsj3) 95756_at 0.60 – Fst 98817_at 4.00 Fig. 3M, N Wang et al. (2004) Ccnd2 97504_at 0.60 Fig. 2H IMAGE 367058 EST 166028_s_at 2.22 – Fabp7 98967_at 0.62 – Tal-2 129118_at 2.00 – Eef1a1 94766_at 0.63 – Mab21l1 165815_r_at 1.89 Fig. 3F IMAGE 4526962 Pip92 (Ier2) 99109_at 0.63 Data not Ednrb 163124_s_at 1.87 Fig. 3I IMAGE 4971909 shown Math1 168404_at 1.84 Fig. 3D IMAGE 4218223 Pea3 (Etv4) 92979_at 0.64 Fig. 1D Lin et al. (1998) EST 113895_at 1.77 – Kik1 (Hsd17b12) 94276_at 0.65 – Rgma 108717_at 1.66 Fig. 3C IMAGE 1001290 Snx5 117186_at 0.65 – EST 139205_at 1.66 Fig. 3K IMAGE 2655939 Tcf7 97994_at 0.65 Fig. 2N A gift from Irma Vtn 98459_at 1.65 Fig. 3A IMAGE 5366291 Thesleff Uncx4.1 92499_at 1.60 Fig. 3H IMAGE 5716567 Bnip3l 96255_at 0.66 – Ngfr (p75) 108762_at 1.57 Fig. 3G Qun et al. (1999) Gstm5 100629_at 0.67 – EST 137358_at 1.55 Fig. 3L IMAGE 3329477 Prkrir 99975_at 0.67 – Maf 115390_at 1.52 – Ccnb1 160159_at 0.69 Data not EST 167322_at 1.52 – shown Ckb 137242_f_at 1.49 Supp. Fig. 7J IMAGE 6395794 Spred2 161070_at 0.69 – Pltp 100927_at 1.46 Data not shown IMAGE 4979759 Nrp 95016_at 0.69 – Nhlh2 166814_f_at 1.44 – Rraga 94257_at 0.70 – EST 166871_at 1.44 Fig. 3J IMAGE 3823589 1200007D18Rik 160184_at 0.71 – Wfdc1 166414_at 1.43 Fig. 3B IMAGE 1497229 Pea15 100548_at 0.71 Data not Tcpn1 10851 5_at 1.41 – IMAGE 653006 shown Dach1 114999_at 1.41 Fig. 3E IMAGE 6826750 144 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157

Immunohistochemistry (n = 6) were pooled after the genotype was confirmed. The gene expression levels of wild-type (n = 2) and Fgfr1 mutant The antibodies used for immunohistochemistry on paraffin sections were (n = 2) tissue sample pools were analyzed by Affymetrix anti-5-HT antibody (1:5000; Immunostar) for E15.5 serotonergic neurons, anti- 5-HT (1:2000, MPBiomedicals Cappel) for adult serotonergic neurons, anti-TH microarrays (see Materials and methods and Supplementary antibody (1:500; Chemicon international) for dopaminergic and noradrenergic Fig. 1 for detailed description). Data analysis of gene neurons and anti-Tuj1 antibody (1:300; Covance). Secondary antibodies used expression levels between duplicate wild-type and mutant were anti-rabbit-IgG (1:300; Alexa-488, Molecular Probes) and anti-mouse-IgG tissue pools revealed that 75% of sequences on the array were (1:300; Alexa-488, Molecular Probes). identified as present. Comparison of the four data sets (2 wild- For quantification of the dopaminergic and serotonergic neurons, embryos (3 mutant and 3 wild-type embryos) were fixed for 1 week in 4% PFA at +4°C, type control vs. 2 Fgfr1 mutant samples) with <0.75-fold or dehydrated and embedded in paraffin. Embryos were sectioned sagittally at >1.41-fold difference thresholds in at least 3 out of 4 5 μm. After deparaffination, the antigen retrieval was done by heating the independent comparisons, identified 51 down-regulated and sections in 10 mM sodium-citrate buffer, pH 6.0, in a microwave oven. The 20 up-regulated genes, respectively. The lists of differentially – sections were washed in PBS and blocked for 30 60 min in PBT (PBS + 1% expressed genes are shown in Tables 1 and 2. BSA + 10% goat serum + 0.3% Triton-X). The samples were incubated in the primary antibody in PBT over 3 nights at +4°C, rinsed in PBS and placed in the Microarray analysis was followed by whole-mount mRNA secondary antibody in PBT for 3 h at RT. Sections were rinsed twice in PBS and in situ hybridization to confirm the pattern of expression of a set once in sterile water before mounting them in Vectashield with DAPI (Vector). of differentially expressed genes. Expression of 25 out of 51 Immunostaining was visualized with an Olympus AX70 microscope with down-regulated and 15 out of 20 up-regulated genes was Olympus DP70 camera. analyzed by in situ hybridization. In addition, real-time relative RT-PCR assay showed similar abundance values for four genes Quantification that were differentially expressed in microarray analysis (Fgf17, For E15.5 serotonergic neurons, only cell populations in rostral hindbrain Igfbp5, Follistatin and Wfdc1, data not shown). were counted. To simplify the comparison between the samples, the results were combined into two populations: dorsal and ventral serotonergic neurons. Since Down-regulated genes the distance between two analyzed sections was always 50 μm and the diameter – μ of one positive cell was estimated to be 12 15 m, the number of cell layers Components of the FGF signal transduction pathway between two sections was approximated to be 4. Total number of positive cells detected in each population was then multiplied by the number of cell layers to The down-regulated genes included several known members get the total amount of positive cells in the whole region. of the FGF signal transduction pathway (Fig. 1A). These For dopaminergic neurons, the positive cells in ventral tegmental area and included FGF ligands, signal regulators and nuclear effectors. substantia nigra were counted and the results divided into two populations: Fgf family members Fgf8, Fgf15, Fgf17,andFgf18 are caudal (c) and rostral (r) dopaminergic neurons. The location of the most caudal expressed in the midbrain–R1 region. Comparison of gene DA neurons in the wild-type embryos determined the border between r- and c - sectors. The caudal boundary of the DA neurons in mutant embryos was used to expression levels by microarray analysis showed that all these draw the caudal boundary of the c-sector. The rostral sector was drawn around Fgfs were down-regulated in E10.5 Fgfr1 mutant mice (Table all ventral DA neurons located rostrally to r–c boundary. 1). This was confirmed by in situ hybridization analysis of Student's t test was used for the comparison of the mean values, and sample E10.5 embryos (Fig. 1B). In Fgfr1 mutants, Fgf17 mRNA was variance was compared using Levene's test. For quantification of adult 5-HT completely abolished from the midbrain–R1 region. In contrast, neurons, brains were sectioned horizontally and 5-HT immunohistochemistry was performed according to Brodski et al. (2003). The number of 5-HT positive Fgf8 was abolished from the dorsal part of the anterior R1, but cells was determined using Stereoinvestigator software (MBF Bioscience). remained in a small ventral patch (Fig. 1B, red arrow; Supplementary Fig. 2). Similarly to Fgf17, down-regulation of Fgf18 expression was observed in the midbrain–R1 region in Results and discussion the Fgfr1 mutant embryos. Expression of Fgf15 was decreased especially in the alar plate of the midbrain–R1 region of the To address the role of FGF signaling and its gene Fgfr1 mutants (Fig. 1B, Supplementary Fig. 2). expression networks in the mouse midbrain–R1 region we Members of the Sprouty, SEF (similar expression to Fgfs), compared the gene expression profiles of wild-type embryos and mitogen-activated protein kinase phosphatase (MKP) and embryos carrying a tissue-specific inactivation of the families are negative modulators of FGF signaling, whereas Fgfr1 gene in the neuroectoderm of midbrain and R1 (En1- nuclear targets of FGF signaling include the ETS transcription Cre/+; Fgfr1Flox/Flox; Trokovic et al., 2003). For this, we factors Erm and Pea3. These molecules affect the FGF dissected the midbrain–R1 tissues from E10.5 wild-type and signaling cascade at different levels to regulate the final output conditional Fgfr1 mutant mice. Stage E10.5 was selected for of the FGFR mediated signal transduction (Fig. 1A). the analysis because a large set of genes was expected to SEF, Sprouty proteins (SPRY1-5), and MKP3 act as negative change their expression in the mutants by this stage. These regulators of the FGF/Ras–MAPK–ERK signaling pathway genes would include both direct and indirect targets of FGF (Furthauer et al., 2002; Mason et al., 2004; Minowada et al., 1999; signaling. All the direct targets, presumably affected already at Tsang et al., 2002; Yusoff et al., 2002). In the Fgfr1 mutant an earlier time point, would be expected to be found among the embryos, Sef and Spry1 transcripts were abolished from the dorsal up- or downregulated genes also at E10.5. In addition, larger midbrain–R1 domain, but remained in the ventral midbrain–R1 amount of tissue available at E10.5 allowed probe preparation region (Fig. 1C, dotted line; Supplementary Fig. 2). By contrast, without an amplification step. Individual midbrain–R1 tissues expression of Spry2 and Mkp3 was down-regulated throughout T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 145

Fig. 1. Expression patterns of FGFR1 signaling cascade components at E10.5. Scheme of the FGF signal transduction pathway (A). Whole mount in situ hybridization of wild-type and En1-Cre/+;Fgfr1Flox/Flox mutant embryos show altered gene expression of known FGF signaling components at E10.5 (B–D). Lateral views of the whole-mount-stained embryos, anterior rightwards. The MHB is marked with arrowheads and a dotted line. Ventral midbrain–R1 region is marked with a red broken line. Down-regulation of different FGF genes (B). At E10.5, expression of Fgf17/8/18/15 in the dorsal midbrain and R1 was clearly down-regulated, but some expression was maintained in the ventral domain (arrows in B, see also Supplementary Fig. 2). Molecules that are known to be associated with FGF signal transduction (C) and proximal downstream targets of FGF signaling (D) were also down-regulated in Fgfr1 mutant embryos. 146 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 the midbrain–R1 region in the Fgfr1 mutant mice at E10.5 (Supplementary Fig. 4; Fig. 2E). Section in situ analysis compared to wild-type embryos (Fig. 1C). Earlier expression showed that Trh expression was localized to the Otx2 positive studies have shown that Mkp3 expression is co-localized with cells near the MHB (Supplementary Fig. 5). In the Fgfr1 Fgf8 also near the MHB (Dickinson et al., 2002; Echevarria et al., mutant embryos, Trh expression was strongly down-regulated 2005; Klock and Herrmann, 2002). Mkp3 expression was (Fig. 2E′). Interestingly, TRH knockout mice showed de- gradually down-regulated in the Fgfr1 mutant embryos starting creased expression of cell cycle regulator PF-TAIRE in the from E9.0 (20 somite stage) (Supplementary Fig. 2). cerebellum (Hashida et al., 2002). Although no obvious CNS The FLRT family of proteins structurally resembles small defects were detected in adult mice, detailed analysis of the leucine-rich proteoglycans found in the extracellular matrix and role of Trh in the early brain development has not been are thought to promote FGF signaling. Recent studies showed performed. MRP4 (or ABCC4) belongs to the ABC transporter that rat and Xenopus Flrt3 genes are coexpressed with Fgfs and molecules and is further grouped to the multidrug resistance are regulated by FGF signaling (Bottcher et al., 2004; Lacy et related protein (MRP) subfamily. We found that Mrp4 al., 1999; Robinson et al., 2004). At E10.5, expression of Flrt3 expression was restricted to a region near the MHB at E9.5– in wild-type embryos was detected at the MHB region, both in E10.5 (Fig. 2F; Supplementary Fig. 4). At E10.5, Mrp4 midbrain and R1 (Fig. 1C, Supplementary Fig. 5). In the Fgfr1 expression was detected in the most anterior R1 (Supplemen- mutant embryos, Flrt3 expression domain at the MHB region tary Fig. 5). Weaker expression was also detected in the was abolished. olfactory bulb and eye primordium (Supplementary Fig. 4). In Pea3 and Erm are members of ETS family transcription the Fgfr1 mutant embryos, Mrp4 expression was abolished factors and thought to be proximal transcriptional targets of from the midbrain–R1 region (Fig. 2F′). At E10.5, expression FGF signaling (Chotteau-Lelievre et al., 2001; Monte et al., of Igfbp5 was localized in distinct cell populations on both 1994). At E10.5, expression of Erm and Pea3 was clearly sides of the MHB (Fig. 2J, Supplementary Fig. 5). In Fgfr1 down-regulated in the midbrain–R1 region in the Fgfr1 mutants, Igfbp5-positive cells were not detected at the border mutant embryos compared to the wild-type (Fig. 1D). between mid- and hindbrain (Fig. 2J′). By contrast, in the However, in Fgfr1 mutants, some expression of Erm and dorsal R1 Igfbp5 expression domain was expanded rostrally Pea3 was still detected in the ventral region (Supplementary (red arrowheads in Fig. 2J′). Jumonji can function as a Fig. 2), indicating the presence of residual FGF signaling. transcriptional repressor of CyclinD1, D2, and cdc2 (Jung et Thus, validating our microarray approach, many known al., 2005; Takahashi et al., 2004; Toyoda et al., 2003). At members of the FGF synexpression group and components of E10.5, we observed Jmj expression as a narrow stripe close to the FGF signal pathway were identified as being down- the MHB in the wild-type embryos (Fig. 2G), but Fgfr1 mutant regulated in the conditional Fgfr1 mutants. embryos lack most of the Jmj expression (Fig. 2G′). Together, these results demonstrate both similarities and differences in Other down-regulated genes the gene-expression in the slowly proliferating boundary cells We also found several other genes being downregulated in in the posterior midbrain and the most anterior R1 (Trokovic et Fgfr1 mutants. These include genes involved in the patterning al., 2005). of midbrain and R1, MHB-specific genes, cell-cycle regulators, D-type cyclins promote cell cycle progression under the and components of various signal transduction cascades. control of extracellular signals such as FGFs or WNTs. At Expression of IsO regulated transcription factors En1/2 and E10.5, CyclinD2 was broadly expressed in the dorsal midbrain Pax5 was abolished at the dorsal part of the midbrain–R1 region and hindbrain (Fig. 2H), but a narrow domain of CyclinD2- of mutant embryos (Figs. 2A′,B′,D′). Interestingly, previously negative cells were detected in the border between mid- and uncharacterized gene (NM_175651; Affymetrix probe ID: hindbrain (Trokovic et al., 2005). In the Fgfr1 mutant embryos, 138472_at) showed similar expression pattern to the En2 gene the overall CyclinD2 expression level was reduced both in (Figs. 2C, C′). The zebrafish ortholog was recently named as the midbrain and R1 and the negative CyclinD2 domain Canopy1 (Cnpy1, NM_001039497) and its inactivation by between midbrain and R1 was abolished (Fig. 2H′). Jmj antisense morpholinos resulted in midbrain–R1 defects (Hirate expression (see above) might repress CyclinD2 levels near the and Okamoto, 2006). Physically, mouse Cnpy1 gene is located MHB leading to decreased cell cycle progression of the 32 kb downstream of En2 locus in chromosome 5, but it is boundary cell population. transcribed from the opposite strand (see detailed description in Sox3 is broadly expressed throughout the developing neural Supplementary Fig. 3). tube. In the wild-type E10.5 controls, Sox3 expression was Genes expressed as a narrow stripe near the MHB include detected broadly in the CNS (Fig. 2I). Similarly to CyclinD2, Trh, Mrp4, Igfbp5 and Jumonji (Jmj). Thyrotropin-releasing Sox3-positive cells were not seen in the narrow population of hormone (TRH), originally isolated as a hypothalamic cells between mid- and hindbrain. In Fgfr1 mutants, Sox3- neuropeptide hormone, most likely acts also as a neuromodu- negative domain near the MHB was abolished (Fig. 2I′), but lator and/or neurotransmitter in the the overall signal was weaker in the midbrain–R1 region (Fig. (Urayama et al., 2002; Yamada et al., 1997). We detected 2I′, red bracket). Recent study by Rizzoti et al. (2004) has strong expression of Trh in the neuroepithelium of prospective shown that cells lacking SOX3 are not able to maintain brain already at E8.0 and at E10.5, expression of Trh was endogenous cell proliferation in the ventral diencephalon detected in a narrow band near the MHB in wild-type embryos leading to abnormal development of Rathke's pouch. Thus, T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 147

Fig. 2. Expression patterns of other down-regulated genes detected by the microarray screen. Whole mount in situ hybridization of the indicated genes at E10.5. Analysis confirmed the down-regulation of genes in the midbrain–hindbrain region of the En1-Cre/+;Fgfr1Flox/Flox embryos. Lateral views of the whole-mount-stained wild-type (A–N) and Fgfr1 mutant (A′–N′) embryos (anterior rightwards). The MHB is marked with a white arrowhead. Red arrows in K′,L,L′, and M′ indicate altered gene expression. di, diencephalon; te, telencephalon. 148 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157

SOX3 activity might be necessary to maintain normal rates expression was restricted to the Otx2 positive cells in Fgfr1 of proliferation in the neuroepithelial precursors also at the mutants (Supplementary Fig. 6) although some Wfdc1 positive midbrain–R1 region. cells in R1 was observed in the lateral boundary area (a black Several down-regulated genes were associated with other arrowhead in Fig. 3B′). signal transduction cascades, including IGF (Igfbp5, see Other genes, such as low-affinity NGF receptor, NGFR1 or above), TNF (Tnfrsf19), and WNT (Sfrp2, Drapc1, Tcf7) p75 (Figs. 3G, G′, dotted line indicates the MHB) and pathways. At E10.5, strong Tnfrsf19 expression was detected homeobox gene Uncx4.1 (Figs. 3H, H′, Supplementary Fig. both in the midbrain–R1 region and in the diencephalon (Fig. 7D), were detected both as dorsal and ventral gradients in R1 2K; Supplementary Fig.4). In Fgfr1 mutant embryos, and midbrain, being absent near the boundary. Expression Tnfrsf19 expression was down-regulated in the lateral changes showed again the loss of gradients in Fgfr1 mutants, midbrain–R1 region (Fig. 2K′, red arrows). Unexpectedly, although the upregulation effect was less obvious in p75 in the domain of Tnfrsf19 -positive cells expanded in the dorsal ventral midbrain. The C. elegans homologue of the Uncx4.1 R1. In E10.5 wild-type mice, Sfrp2 expression was detected participates in VA motor neuron specification (Miller and in the midbrain and R1 (Fig. 2L). Strong Sfrp2 expression Niemeyer, 1995; Miller et al., 1992; White et al., 1992). was detected in the dorsal R1 whereas in the dorsal midbrain, However, the role of Uncx4.1 in neuronal differentiation in the expression domain was graded diminishing towards the mammals and other vertebrates is unclear, since the null mutants MHB. In the Fgfr1 mutant embryos, the amount of Sfrp2- do not show a CNS phenotype (Leitges et al., 2000; Mansouri positive cells was decreased in the R1 (Fig. 2L′) and the et al., 2000). gradient in the midbrain was not clearly seen in the mutant The expression changes in ventral R1 were not as uniform embryos. Drapc1 (or Apcdd1) is a novel in vivo target gene as in the dorsal structures and ventral midbrain. For example, of Wnt/β-catenin signaling pathway (Takahashi et al., 2002). the expression of endothelin receptor type B (Ednrb) in the Its expression pattern shows high similarity to β-catenin ventral R1 domain, near the boundary, was surprisingly activity pattern during (Jukkola et downregulated in the mutant embryos (Figs. 3I, I′; asterisks al., 2004; Maretto et al., 2003). At E10.5, expression of point to the downregulation in R1, see Supplementary Fig. Drapc1 together with Tcf7 was found in the Wnt1/Otx2 7F,). In the ventral midbrain, the Ednrb positive cells showed positive cells close to the MHB (Figs. 2M, N; Supplementary an opposite effect spreading both laterally (the region Fig. 5). At the same stage in the Fgfr1 mutant embryos, between the black arrowheads in Fig. 3I′) and towards the Drapc1 expression was dispersed into adjacent dorsal MHB. rhombomere 1 (Fig. 2M′, red arrow). Overall, Drapc1 and The expression of three yet uncharacterized genes or ESTs Tcf7 expression was clearly down-regulated in the Fgfr1 (Expressed Sequence Tags), found in the microarray screen, mutants (Figs. 2M′,N′). were also validated using in situ hybridization. Sequence analysis of NM_028263 (Affymetrix probe ID 166871_at; Upregulated genes show loss of expression gradients in Fgfr1 UniProt Q7TNS6) found an FGF-binding protein (FGF-bp) mutant embryos domain, suggesting that the gene belonged to the FGF-bp family. FGF-bps are secreted molecules that mobilize FGFs We performed whole mount in situ hybridization of the from the extracellular matrix and enhance their activity, upregulated genes (Table 2) on E9.5 and E10.5 wild-type and probably by presenting them to FGF-receptors (Tassi et al., Fgfr1 conditional mutant embryos. No observable changes in 2001). In the mutant embryos, the ventral midbrain expression were detected in E9.5 mutants, except for follistatin expression of NM_028263 spread towards the boundary (Fst). This was due to very weak overall expression of the genes (Figs. 3J, J′, see Supplementary Fig. 7M), and a globular in the midbrain and hindbrain at E9.5 (See Supplemetary Fig. 7). strong expression patch appeared in the R1 side near the In wild-type E10.5 embryos, the upregulated genes showed MHB (black arrowheads). mainly two main types of expression—dorsal gradients or Two other uncharacterized genes, NM_177100 (Affymetrix combined dorsal and ventral gradients. probe ID 139205_at; UniProt Q8BH22) and an EST corresponding Genes such as Vtn (Figs. 3A, A′), Wfdc1 (Figs. 3B, B′), to GeneBank accession BM932503 (Affymetrix probe ID RgmA (Figs. 3C, C′), Math1 (Figs. 3D, D′), Dach1 (Figs. 3E, 137358_at) were expressed widely in the developing CNS. In E′) and Mab21l1 (Figs. 3F, F′) were expressed as dorsal Fgfr1 mutants, the expression gradients in both tectum and ventral gradients decreasing towards the MHB in wild-type embryos midbrain were lost in NM_177100 (Figs. 3K, K′, see also (see also Supplementary Fig. 7). In mutants, the gradients were Supplementary Fig. 7L). Similar changes were seen in the ESTwith lost and the expression continued uniformly strong towards the ID 137358_at, although it showed less obvious changes in ventral boundary. Interestingly, WAP four disulfide core domain midbrain (Figs. 3L, L′, see also Supplementary Fig. 7K). protein 1 (Wfdc1) codes for a growth inhibitor ps20 (Larsen Follistatin (Figs. 3M, M′,N,N′, see also Supplementary Fig. et al., 1998). Its function would suggest a role in controlling 7A) binds and inactivates several members of the TGF-β cell cycle, proliferation and transition to postmitotic phase superfamily of proteins, such as activin and BMP-4. In the which is associated with neuronal differentiation. Therefore, Fgfr1 mutants at E9.5, R1 expression domain spread almost to the upregulation of Wfdc1 in mutants might imply that there is the MHB and its rostral part was diffuse (compare to the gap in excessive premature differentiation in the tectum. Wfdc1 expression in Fig. 3M, marked by brackets, see also T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 149

Fig. 3. Expression patterns of the upregulated genes detected by the microarray screen. Whole mount in situ hybridizations of E9.5 (M) and E10.5 (A–L, N) wild-type (A–N) and En1-Cre/+;FgfrFlox/Flox embryos (A′–N′) with the probes indicated. Lateral views of the whole-mount-stained embryos, anterior rightwards. The MHB is indicated with a white arrowhead and a dotted line. Brackets, asterisks and black arrowheads point to alterations in gene expression. Embryos in panels I and N have been sagittally bisected after staining. 150 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157

Supplementary Fig. 8). This effect was even more prominent in E10.5 mutant mice, where the whole R1 expressed Fst (Fig. 3N′). Near the boundary, the expression also extended ventrally in the mutants. In summary, several upregulated genes showed marked similarity in their expression patterns in E10.5 embryos. Most of them were expressed as gradients decreasing towards MHB, both in midbrain and rhombomere 1. In Fgfr1 mutant embryos, these gradients disappeared both in dorsal and in ventral regions.

Role of Fst in neuronal development in the midbrain–hindbrain region

To analyze the role of follistatin on neurogenesis, the expression of Phox2a (Figs. 4A, A′) and Ngn2 (Figs. 4B, B′) were analyzed on E10.5 follistatin knock-out embryos (Matzuk et al., 1995). If Fst normally regulates the BMP signals from the roof plate and this way affected the positioning of LC, we would expect to see a shift in the position of LC. However, the expression of both genes in mutants was identical to the wild-type controls and no changes in LC was seen at this stage. There were no observable changes in the appearance of locus coeruleus (Figs. 4C, C′), midbrain dopaminergic neurons (Figs. 4D, D′) or serotonergic neurons (Figs. 4E, E′) in E14.5 follistatin mutant embryos either, as shown in immunohistochemistry experiments. This suggests that the upregulation of Fst would not lead to notable changes in neurogenesis, either.

Neuronal progenitor cell populations are affected in Fgfr1 mutant embryos

Neural precursors near the MHB differentiate later in development, and their premature differentiation is prevented by Hes3. The expression gradients diminishing towards the MHB in the wild-type embryos would, in this light, indicate that these genes are expressed in more fate-restricted and/or differentiating cells. The loss of the expression gradients suggests that the region of less differentiated neurons became narrower in Fgfr1 mutants and that differentiation proceeded prematurely towards the boundary on both sides. In Ednrb, however, an opposite effect was seen—downregulation of expression in ventral R1 near the boundary in mutants. The Ednrb-positive cells in ventral R1 seem to respond differently to the loss of FGF signaling at the IsO. This may correlate with impaired differentiation of the serotonergic neurons in Fgfr1 mutants (see below). Since our results suggested loss of neuronal differentiation gradients in Fgfr1 mutants, we wanted to know what effects would FGFR1 inactivation have on different subsets of neuronal progenitor cells in the mid- and hindbrain. We analyzed several Fig. 4. Analysis of neurogenesis in follistatin null mutant embryos. Whole genes which are known to be expressed in different neuronal mount in situ hybridizations of E10.5 (A, B) embryos with the probes indicated. progenitors at certain stages of development. Embryos in panel A have been sagittally cut after staining. Immunohistochem- istry on E14.5 sagittal paraffin sections, anterior rightwards (C–E). TH-staining Hes3 is a bHLH transcriptional factor and a known Notch in panels C and D, 5-HT staining in panel E. Wild-type embryos in A–E; effector. In E10.5 Fgfr1 mutant embryos, the expression of follistatin knock-out embryos in A′–E′. lc, locus coeruleus; III, third cranial Hes3 is downregulated but is still weakly detectable in the ganglion; IV, fourth cranial ganglion. T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 151 caudal midbrain and anterior R1 near the MHB (Figs. 5A, A′, In wild type embryos, a gap of mature neurons was evident B, B′). The loss of Hes3 expression supports the results near the MHB region, as seen in Tuj1 staining (Figs. 5C, C′). In obtained from the upregulated genes, which suggested that Fgfr1 mutants the gap disappeared, although fewer Tuj1 neuronal differentiation was shifted towards the MHB region positive cells were still detected close to the MHB compared in Fgfr1 mutant embryos. to more distal regions. In tectum, no differences were seen (data not shown). Neurogenin2 (Figs. 5D, D′,E,E′) is a basic-helix–loop– helix (bHLH) transcriptional regulator, which functions as a neuronal determination factor (Sommer et al., 1996). In the Fgfr1 conditional mutant embryos, the Ngn2 expressing cell populations of the ventral midbrain and hindbrain are less compact and have spread towards the boundary. Especially a strongly expressing cell population near the MHB in ventral R1 (marked with an asterisk in Figs. 5E, E′) shows an extensive scattering effect. Also locus coeruleus (LC) in E10.5 mutants seems less compact as compared to the wild-type. At E11.5, the expression of Ngn2 in the dorsal R1 cells was shifted rostrally in the mutant embryos (compare to the area indicated with a bracket in Fig. 5E). These results are supported by the expression analysis of a homeobox Phox2a, which is expressed in the developing LC and in the oculomotoric (III) and trochlear (IV) nuclei (Figs. 5F, F′). The III and IV cranial ganglia are the same size in the Fgfr1 mutants as in the wild-type embryos. However, they appear to be closer together, presumably as a result of the loss of the boundary cell population between mid- and hindbrain. III and IV cranial ganglia as well as LC develop earlier compared to other neuronal populations, such as SA or DA neurons, in the area. They can be detected already in E9 embryos, which suggest that their development is less dependent on the isthmic signaling and that the major events concerning their formation takes place before the need for FGFR1 function. This would explain their mild phenotype in Fgfr1 conditional mutant adults—III and IV ganglia appear normal, and LC is somewhat disorganized (Trokovic et al., 2003).

Midbrain dopaminergic neuron progenitor pool spreads in Fgfr1 mutants

Since several genes were upregulated in the ventral midbrain, we examined what happens to the dopaminergic neuron progenitors in the area. For this, we analyzed the expression of several genes known to be involved in the early

Fig. 5. Expression patterns of genes regulating neuronal development. Whole mount in situ hybridization analysis (A, B, D–F) with the probes and embryonic stages indicated. Immunohistochemistry on sagittal paraffin sections with Tuj1- antibody in C, close-up view of the ventral midbrain and R1. Wild-type embryos in A–E and En1-Cre/+;Fgfr1Flox/Flox embryos in A′–E′. Embryos in panel A have been cut after staining and sections are depicted in panel B. White lines indicate cutting planes. Lateral views of the whole-mount-stained embryos, anterior rightwards. Awhite open arrowhead indicates the MHB in panels A, D– F and a black arrowhead in panel B. Asterisks and brackets point to altered gene expression. E11.5 embryos have been sagittally bisected at the midline after staining. Lc, locus coeruleus; III, oculomotoric nucleus; III, trochlear nucleus; mb, midbrain; r1, rhombomere 1. Scale bars, 100 μm. 152 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157

Fig. 6. Midbrain dopaminergic neurons in Fgfr1 mutant embryos. Whole mount (A–D) and radioactive (E, F) in situ hybridization analysis with the probes and embryonic stages indicated. Lateral views of the embryos, anterior rightwards (A) white open arrowhead indicates the midbrain–hindbrain boundary in panels A–D. Black arrowheads point to altered gene expression. Ventral brain morphology has been visualized with a red broken line in panels B and D. All embryos have been sagittally bisected at the midline after staining. Tyrosine hydroxylase immunohistochemistry on paraffin sections in panel G. Sagittal views, anterior rightwards. Orange broken line indicates the quantification area and white arrowheads point to altered TH expression. Wild-type embryos in A–E and G, En1-Cre/+;Fgfr1Flox/Flox embryos in A′–E′, F and F′ and in G′. Quantification results in H. DA-all, total number of DA neurons; DA-C, number of DA neurons in caudal sector; DA-R, number of DA neurons in rostral sector; c, caudal sector; r, rostral sector. development of DA neurons and quantified the neurons in (Fig. 6G, G′). The comparison between the total number of E15.5 embryos. dopaminergic neurons showed no statistically significant In E11.5 Fgfr1 mutant embryos, the expression of Nurr1 difference between wild-type and mutant embryos (Fig. (Figs. 6A, A′), Pitx3 (Figs. 6B, B′) and Aldh1 (also known as 6H). The neurons were divided into caudal and rostral Aldh1a1 and Raldh1) (Figs. 6C, C′,D,D′) spread posteriorly sectors (marked with an orange broken line in Figs. 6G, G′) (black arrowheads). The spreading of Aldh1 expression was using the mesencephalic flexure as a landmark, and the detected already at E10.5. The spreading of Aldh1 follows the sectors were compared individually. The analysis revealed an caudal shift in the expression of Otx2 (Figs. 6F, F′). increase in the caudal sector and conversely a decrease in the We next analyzed the tyrosine hydroxylase (TH) positive rostral sector of dopaminergic neurons in the Fgfr1 mutant dopaminergic neurons of the ventral tegmental area and embryos. Thus, consistent with changes in gene expression substantia nigra E15.5 wild-type and Fgfr1 mutant embryos patterns, TH immunostaining at E15.5 stage revealed a caudal T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 153

Fig. 7. Rostral serotonergic neurons in Fgfr1 mutant embryos. Whole mount in situ hybridization analysis (A–E) with the probes and embryonic stages indicated. A white open arrowhead indicates the midbrain–hindbrain boundary. Lateral views of the whole-mount-stained embryos, anterior rightwards. Black arrowheads point to altered gene expression. Ventral brain morphology has been visualized with a red broken line in D and D′. All embryos have been sagittally bisected at the midline after staining. Anti-5-HT immunohistochemistry on paraffin E15.5 sections (F, F′,F″). Sagittal sections, anterior rightwards. Orange broken line indicates the quantification area, white arrows mark the caudal Otx2 expression boundary. Asterisks point to altered 5-HT expression. Quantification results in G. Radioactive in situ hybridisation on P0 stage coronal sections with the probes indicated. Bright-field images in panel H, dark-field images in panels I and J. Wild-type embryos in A–F and H–J, En1-Cre/ +;Fgfr1Flox/Flox embryos in A′–F′ and H′–J′. SA-all, total number of SA neurons; SA-D; number of SA neurons in the dorsal population; SA-V, number of SA neurons in the caudal population; d, dorsal population; v, ventral population. shift in the position of DA neurons in Fgfr1 mutants (Figs. suggest that the mixing of ventral midbrain and R1 cells 6G, G′, compare the area between white arrows). It seems forms a tissue mosaic in the Fgfr1 mutants (see below). that although the area where the DA neurons are located expands in mutants, the total number of these neurons does The most rostral serotonergic neuron precursors are lost in not change. The caudal shift is probably due to the caudal Fgfr1 mutant embryos spreading of Otx2 expression (Figs. 6G and G′). In ventral midbrain, Otx2 expression is scattered in Fgfr1 mutants at FGF8 from the IsO is thought to be one of the major signals E15.5, and the boundary between Otx2 positive and negative guiding the development of serotonergic neurons in the raphe areas is less obvious than in the wild-type. This would nuclei (Ye et al., 1998). To find out whether serotonergic 154 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 neurons are affected in Fgfr1 mutant embryos, we analyzed the develop normally. In the explant culture experiments, FGF4 in expression patterns of several genes that are known to be combination with FGF8 and SHH could induce SA neuron involved in the SA neuron specification and quantified the development (Ye et al., 1998). However, FGF4 itself is not rostral serotonergic neurons in E15.5 embryos. expressed in the vicinity of developing rostral SA neurons at In Fgfr1 mutants, several genes which are involved in the E10.5–11.5. Since FGFR1 is inactivated by 10 somite stage specification of serotonergic neurons, such as zinc-finger (Trokovic et al., 2003), the early FGF4 signaling functions transcription factor Gata3 (Figs. 7A, A′) a bHLH transcription normally in Fgfr1 mutants. This suggests that other FGFs from factor Mash1/Ascl1 (Figs. 7B, B′) and Pet1 (Figs. 7D, D′,F,F′), the IsO are the major regulators in SA neuron development. It is were downregulated near the MHB region in ventral R1 (black possible that in the explant culture experiments FGF4 mimicked arrowheads). The expression of Mash1 and Gata3 in the ventral the high FGF signaling activity near the MHB, which is midbrain showed, however, an opposite effect, spreading normally achieved by the combination of FGFs expressed by strongly towards the boundary. This is consistent with posterior the IsO. In Fgfr1 mutants, the amount of activating IsO signals expansion in the expression of DA neuron specification genes. The rostral Pet1 expression boundary is narrower and disorganized compared to the wild-type controls. This result combined with the downregulation of upstream effectors Mash1 and Gata3 would directly imply that the rostralmost population of raphe nuclei does not develop in conditional Fgfr1 mutants. Double in situ hybridization with midbrain specific Otx2 and Mash1 shows that at E10.5, Mash1 downregulation in R1 is not only due to Otx2 spreading into R1 area (Figs. 7C, C′). However, at E11.5 Otx2 extensively spreads into ventral R1 and masks Pet1 downregulation (Figs. 7E, E′). We performed anti-5HT immunohistochemistry on E15.5 sagittal sections to characterize whether downregulation of the transcription factors has an effect on SA neuron development. 5-HT could be seen in a large area extending to the top of the mesencephalic flexure in the wild-type embryos (Fig. 7F). In Fgfr1 mutants, the rostral boundary of SA neuron population was shifted caudally and less 5-HT positive neurons were seen near the top of the flexure (compare the area marked with asterisks in Figs. 7F, F′,F″). The caudal shift of Otx2 expression in mutants (caudal expression boundary of Otx2 is indicated with a white arrowhead in panel F) did not completely prevent rostral SA neuron development. The SA neurons located more caudally seemed unaffected in mutants. We quantified the serotonergic neurons in rostral R1 at E15.5 wild-type and Fgfr1 mutant embryos (the quantified area is depicted with an orange broken line in Figs. 7F, F′,F″). When the total number of neurons was compared, their number was significantly decreased in mutants (Fig. 7G). The neurons were divided into dorsal and ventral populations which were compared individually. The loss of the most rostral SA neurons explains the decrease in the dorsal population. The ventral Fig. 8. Schematic illustration of midbrain and anterior rhombomere 1 neuronal cell populations and summarizing gene expression patterns of FGF regulated population seems to be unaffected in Fgfr1 mutants. The genes at the MHB region. Lateral views of E10.5 wild-type (A and B) and Fgfr1 quantification results reflect the changes in early gene mutant (A′ and B′) embryos. Gene expression changes in Fgfr1 mutant embryos expression, where the downregulation was seen in the rostral (A′) compared with wild-type (A). A narrow slowly proliferating cell population area only. Results from P0 stage support these observations. Sert was lost in Fgfr1 mutants (yellow line in A). In addition, the loss of Fgfr1 is downregulated and diffuse in mutants (Figs. 7H, H′,I,I′). activity lead to downregulation of several genes expressed at the posterior midbrain and anterior R1 (A and A′). The expression patterns of these genes are Interestingly, consistent with the expression of Phox2a at E10.5, marked with a green hatching (A and A′). Red dotted areas represents a pattern the development of the trochlear nucleus appears normal in of genes that showed gradient-like expression both in the midbrain and R1 and mutants (Figs. 7J, J′). were found to be downregulated in the midbrain–R1 region in the Fgfr1 mutants It is interesting that the downregulation of transcription (A and A′). By contrast, orange gradients show a pattern for upregulated genes ′ – factors Mash1, Gata3 and Pet1 could be seen only in the (A and A ). Neuronal progenitor populations at the midbrain R1 region are marked as indicated with boxes (B and B′). Black arrows represent the direction vicinity of the boundary. The early downregulation of the of a shift in the neuron-specific gene expressions (B). Arrowheads and broken proneural gene Mash1 suggests that the SA progenitors near the lines mark the MHB. mb, midbrain; r1, rhombomere 1; DA, dopaminergic MHB require a high dose of FGF-signals from IsO in order to neurons; LC, locus coeruleus; SA, serotonergic neurons. T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 155 is reduced (Supplementary Fig. 2) and the SA neuron Acknowledgments developmental pathway is impaired. Serotonergic neurons further away from the boundary express normally all the We would like to thank Eija Koivunen, Päivi Hannuksela genes in the SA pathway, as well as 5-HT, and seem less and Marjo Virtanen for the excellent technical assistance and M. dependent on the isthmic signaling. Sc Kaia Kala for the help with immunohistochemistry. Henri Koivula in Dr. Pertti Panula's laboratory helped with the Conclusions zebrafish experiments and Marika Suomalainen from Dr. Irma Thesleff's laboratory with follistatin knock-out mice. We A proposed model on the role of FGF-signaling from the acknowledge Dr. Martyn Matzuk for the Fst mutant mice and isthmic organizer in regulating neurogenesis in the midbrain and Dr. Martyn James for providing the Tcf1/7. For microarray R1 of the hindbrain is shown in Fig. 8. Our screen identified analysis, we want to acknowledge the Turku Centre for several genes responsible for cell cycle regulation, which were Biotechnology's Affymetrix service. This work was supported downregulated in Fgfr1 mutant embryos. They can be grouped by the Academy of Finland, Biocentrum Helsinki, Sigfrid into three classes based on their response to FGF signaling Juselius Foundation and Helsinki Graduate School in Biotech- through FGFR1 in the midbrain–R1. First, Flrt3, Trh, Mrp4, nology and Molecular Biology (T.J. and L.L.) and the Igfbp5 and Jmj are expressed in narrow bands near the MHB. Bundesministerium für Bildung und Forschung (NGFN- They are completely abolished in the Fgfr1 mutants (Figs. 8A, A 2,01GS0476 to W.W.) and Deutsche Forschungsgemeinschaft ′). Second, a group of genes expressed in a broader domain such (DFG to W.W.). as Hes3 which maintain the expression of IsO-specific genes (e.g., Fgf8 and Pax5) and inhibit neuronal differentiation (Fig. Appendix A. Supplementary data 8A). Expression of also these genes requires FGFR1, except for the most ventral region where a small domain of ventral Supplementary data associated with this article can be found, expression remains possibly due to the expression of Fgfr2 and in the online version, at doi:10.1016/j.ydbio.2006.05.002. Fgfr3 ventrally (Fig. 8A′). Third, cell cycle regulators CyclinD2 and Sox3 are expressed throughout the midbrain–R1 region being absent close to the MHB (Fig. 8A). In the Fgfr1 mutants, CyclinD2- and Sox3-negative domains were abolished, but the References overall expression levels were decreased at the midbrain–R1 region (Fig. 8A′). The downregulation of cell cycle activators in Acampora, D., Avantaggiato, V., Tuorto, F., Simeone, A., 1997. Genetic control of brain morphogenesis through Otx gene dosage requirement. Development the midbrain and R1 of the Fgfr1 mutants would indicate that 124, 3639–3650. cell proliferation in the region was decreased. The disappearance Andersson, E., Tryggvason, U., Deng, Q., Friling, S., Alekseenko, Z., Robert, B., of Hes3-domain might lead to premature neuronal differentia- Perlmann, T., Ericson, J., 2006. Identification of intrinsic determinants tion in the midbrain–R1 region. of midbrain dopamine neurons. Cell 2, 393–405. The data from the upregulated genes supported the Blak, A.A., Naserke, T., Weisenhorn, D.M., Prakash, N., Partanen, J., Wurst, W., 2005. Expression of Fgf receptors 1, 2, and 3 in the developing mid- and deregulation of cell differentiation in the area, as expression hindbrain of the mouse. Dev. Dyn. 233, 1023–1030. analysis showed disappearance of gene expression gradients in Bottcher, R.T., Pollet, N., Delius, H., Niehrs, C., 2004. The transmembrane tectum as well as in the ventral midbrain/R1 (Figs. 8A, A′). protein XFLRT3 forms a complex with FGF receptors and promotes FGF There are genes (e.g., Vtn, Wfdc1 and Fst) expressed in the signalling. Nat. Cell Biol. 6, 38–44. anterior midbrain, expression of which expanded posteriorly in Broccoli, V., Boncinelli, E., Wurst, W., 1999. The caudal limit of Otx2 expression positions the isthmic organizer. Nature 401, 164–168. the Fgfr1 mutants. Other genes (e.g., Math1, Uncx4.1, Fst and Brodski, C., Weisenhorn, D.M., Signore, M., Sillaber, I., Oesterheld, M., Ednrb) were expressed in the posterior R1 and their expression Broccoli, V., Acampora, D., Simeone, A., Wurst, W., 2003. Location and shifted anteriorly in the midbrain mutants. However, various size of dopaminergic and serotonergic cell populations are controlled by the neuronal populations differ in their requirement for FGF position of the midbrain–hindbrain organizer. J. Neurosci. 23, 4199–4207. signaling through FGFR1. Whereas the midbrain dopaminergic Chi, C.L., Martinez, S., Wurst, W., Martin, G.R., 2003. The isthmic organizer signal FGF8 is required for cell survival in the prospective midbrain and neurons shift posteriorly but their numbers are largely cerebellum. Development 130, 2633–2644. unaffected, the rostral serotonergic neurons are greatly reduced Chotteau-Lelievre, A., Dolle, P., Peronne, V., Coutte, L., de Launoit, Y., in Fgfr1 mutants. Thus, high signaling levels are required for Desbiens, X., 2001. Expression patterns of the Ets transcription factors from the proper development of the rostral SA neurons in R1. Our the PEA3 group during early stages of mouse development. Mech. Dev. 108, – model suggests that high FGF-signaling levels from the IsO 191 195. Crossley, P.H., Martin, G.R., 1995. The mouse Fgf8 gene encodes a family of prevent neuronal differentiation in the midbrain side near the polypeptides and is expressed in regions that direct outgrowth and patterning MHB. When FGF-signaling from IsO is abolished in Fgfr1 in the developing embryo. Development 2, 439–451. mutants, neurogenesis can proceed closer to MHB in the Davis, C.A., Joyner, A.L., 1988. Expression patterns of the homeo box- midbrain but is impaired in the ventral R1. In zebrafish Ace containing genes En-1 and En-2 and the proto-oncogene int-1 diverge during – mutants, R1 is rostralized into midbrain in the absence of mouse development. Genes Dev. 12B, 1736 1744. Dickinson, R.J., Eblaghie, M.C., Keyse, S.M., Morriss-Kay, G.M., 2002. isthmic Fgf8 signaling (Jaszai et al., 2003). However, in Fgfr1 Expression of the ERK-specific MAP kinase phosphatase PYST1/MKP3 in mutants, instead of complete fate transformation extensive mouse embryos during morphogenesis and early organogenesis. Mech. Dev. mixing of ventral midbrain cells into R1 is detected. 113, 193–196. 156 T. Jukkola et al. / Developmental Biology 297 (2006) 141–157

Echevarria, D., Vieira, C., Gimeno, L., Martinez, S., 2003. Neuroepithelial Lillevali, K., Matilainen, T., Karis, A., Salminen, M., 2004. Partially secondary organizers and cell fate specification in the developing brain. overlapping expression of Gata2 and Gata3 during inner ear development. Brain Res. Brain Res. Rev. 43, 179–191. Dev. Dyn. 231, 775–781. Echevarria, D., Martinez, S., Marques, S., Lucas-Teixeira, V., Belo, J.A., 2005. Lin, J.H., Saito, T., Anderson, D.J., Lance-Jones, C., Jessell, T.M., Arber, S., Mkp3 is a negative feedback modulator of Fgf8 signaling in the mammalian 1998. Functionally related motor neuron pool and muscle sensory isthmic organizer. Dev. Biol. 277, 114–128. afferent subtypes defined by coordinate ETS gene expression. Cell 3, Farkas, L.M., Dunker, N., Roussa, E., Unsicker, K., Krieglstein, K., 2003. 393–407. Transforming growth factor-beta(s) are essential for the development of Liu, A., Joyner, A.L., 2001. Early anterior/posterior patterning of the midbrain midbrain dopaminergic neurons in vitro and in vivo. J. Neurosci. 23, and cerebellum. Annu. Rev. Neurosci. 24, 869–896. 5178–5186. Liu, A., Li, J.Y., Bromleigh, C., Lao, Z., Niswander, L.A., Joyner, A.L., 2003. Furthauer, M., Lin, W., Ang, S.L., Thisse, B., Thisse, C., 2002. Sef is a FGF17b and FGF18 have different midbrain regulatory properties from feedback-induced antagonist of Ras/MAPK-mediated FGF signalling. Nat. FGF8b or activated FGF receptors. Development 130, 6175–6185. Cell Biol. 4, 170–174. Mansouri, A., Voss, A.K., Thomas, T., Yokota, Y., Gruss, P., 2000. Uncx4. 1 is Geling, A., Itoh, M., Tallafuss, A., Chapouton, P., Tannhauser, B., Kuwada, J.Y., required for the formation of the pedicles and proximal ribs and acts Chitnis, A.B., Bally-Cuif, L., 2003. bHLH transcription factor Her5 links upstream of Pax9. Development 127, 2251–2258. patterning to regional inhibition of neurogenesis at the midbrain–hindbrain Maretto, S., Cordenonsi, M., Dupont, S., Braghetta, P., Broccoli, V., Hassan, boundary. Development 130, 1591–1604. A.B., Volpin, D., Bressan, G.M., Piccolo, S., 2003. Mapping Wnt/beta- Hashida, T., Yamada, M., Hashimoto, K., Shibusawa, N., Monden, T., Satoh, T., catenin signaling during mouse development and in colorectal tumors. Proc. Mori, M., 2002. A novel TRH-PFTAIRE protein kinase 1 pathway in the Natl. Acad. Sci. U. S. A. 100, 3299–3304. cerebellum: subtractive hybridization analysis of TRH-deficient mice. Mason, J.M., Morrison, D.J., Bassit, B., Dimri, M., Band, H., Licht, J.D., Endocrinology 143, 2808–2811. Gross, I., 2004. Tyrosine phosphorylation of Sprouty proteins regulates Henrique, D., Adam, J., Myat, A., Chitnis, A., Lewis, J., Ish-Horowicz, D., their ability to inhibit growth factor signaling: a dual feedback loop. Mol. 1995. Expression of a Delta homologue in prospective neurons in the chick. Biol. Cell 15, 2176–2188. Nature 375, 787–790. Matzuk, M.M., Lu, N., Vogel, H., Sellheyer, K., Roop, D.R., Bradley, A., 1995. Hermanson, E., Joseph, B., Castro, D., Lindqvist, E., Aarnisalo, P., Wallen, A., Multiple defects and perinatal death in mice deficient in follistatin. Nature Benoit, G., Hengerer, B., Olson, L., Perlmann, T., 2003. Nurr1 regulates 374, 360–363. dopamine synthesis and storage in MN9D dopamine cells. Exp. Cell Res. 2, McMahon, A.P., Bradley, A., 1990. The Wnt-1 (int-1) proto-oncogene is 324–334. required for development of a large region of the mouse brain. Cell 62, Hirata, H., Tomita, K., Bessho, Y., Kageyama, R., 2001. Hes1 and Hes3 regulate 1073–1085. maintenance of the isthmic organizer and development of the mid/hindbrain. McWhirter, J.R., Goulding, M., Weiner, J.A., Chun, J., Murre, C., 1997. A novel EMBO J. 20, 4454–4466. fibroblast growth factor gene expressed in the developing nervous system is Hirate, Y., Okamoto, H., 2006. Canopy1, a novel regulator of FGF signaling a downstream target of the chimeric homeodomain oncoprotein E2A-Pbx1. around the midbrain–hindbrain boundary in zebrafish. Curr. Biol. 16, Development 17, 3221–3232. 421–427. Miller III, D.M., Niemeyer, C.J., 1995. Expression of the unc-4 homeoprotein in Jaszai, J., Reifers, F., Picker, A., Langenberg, T., Brand, M., 2003. Isthmus-to- Caenorhabditis elegans motor neurons specifies presynaptic input. midbrain transformation in the absence of midbrain–hindbrain organizer Development 121, 2877–2886. activity. Development 130, 6611–6623. Miller, D.M., Shen, M.M., Shamu, C.E., Burglin, T.R., Ruvkun, G., Dubois, Jukkola, T., Sinjushina, N., Partanen, J., 2004. Drapc1 expression during mouse M.L., Ghee, M., Wilson, L., 1992. C. elegans unc-4 gene encodes a embryonic development. Gene Expression Patterns 4, 755–762. homeodomain protein that determines the pattern of synaptic input to Jung, J., Kim, T.G., Lyons, G., Kim, H.R., Lee, Y., 2005. Jumonji regulates specific motor neurons. Nature 355, 841–845. cardiomyocyte proliferation via interaction with retinoblastoma protein. J. Millet, S., Campbell, K., Epstein, D.J., Losos, K., Harris, E., Joyner, A.L., 1999. Biol. Chem. 280, 30916–30923. A role for Gbx2 in repression of Otx2 and positioning the mid/hindbrain Katahira, T., Sato, T., Sugiyama, S., Okafuji, T., Araki, I., Funahashi, J., organizer. Nature 401, 161–164. Nakamura, H., 2000. Interaction between Otx2 and Gbx2 defines the Minowada, G., Jarvis, L.A., Chi, C.L., Neubuser, A., Sun, X., Hacohen, N., organizing center for the optic tectum. Mech. Dev. 91, 43–52. Krasnow, M.A., Martin, G.R., 1999. Vertebrate Sprouty genes are induced Kimmel, R.A., Turnbull, D.H., Blanquet, V., Wurst, W., Loomis, C.A., Joyner, by FGF signaling and can cause chondrodysplasia when overexpressed. A.L., 2000. Two lineage boundaries coordinate vertebrate apical ectodermal Development 126, 4465–4475. ridge formation. Genes Dev. 14, 1377–1389. Monte, D., Baert, J.L., Defossez, P.A., de Launoit, Y., Stehelin, D., 1994. Klock, A., Herrmann, B.G., 2002. Cloning and expression of the mouse dual- Molecular cloning and characterization of human ERM, a new member of specificity mitogen activated protein (MAP) kinase phosphatase Mkp3 the Ets family closely related to mouse PEA3 and ER81 transcription during mouse embryogenesis. Mech. Dev. 116, 243–247. factors. Oncogene 9, 1397–1406. Lacy, S.E., Bonnemann, C.G., Buzney, E.A., Kunkel, L.M., 1999. Identification Nakamura, H., 2001. Regionalisation of the optic tectum: combinations of gene of FLRT1, FLRT2, and FLRT3: a novel family of transmembrane leucine expression that define the tectum. Trends Neurosci. 24, 32–39. rich repeat proteins. Genomics 62, 417–426. Ninkovic, J., Tallafuss, A., Leucht, C., Topczewski, J., Tannhauser, B., Lam, C.S., Sleptsova Friedrich, I., Munro, A.D., Korzh, V., 2003. SHH Solnica-Krezel, L., Bally-Cuif, L., 2005. Inhibition of neurogenesis at the and FGF8 play distinct roles during development of noradrenergic zebrafish midbrain–hindbrain boundary by the combined and dose- neurons in the locus coeruleus of the zebrafish. Mol. Cell. Neurosci. dependent activity of a new hairy/E(spl) gene pair. Development 132, 22, 501–515. 75–88. Larsen, M., Ressler, S.J., Lu, B., Gerdes, M.J., McBride, L., Dang, T.D., Pispa, J., Mikkola, M.L., Mustonen, T., Thesleff, I., 2003. Ectodysplasin, Rowley, D.R., 1998. Molecular cloning and expression of ps20 growth Edar and TNFRSF19 are expressed in complementary and overlapping inhibitor. A novel WAP type “four disulfide core” domain protein expressed patterns during mouse embryogenesis. Gene Expression Patterns 5, in smooth muscle. J. Biol. Chem. 273, 4574–4584. 675–679. Leitges, M., Neidhardt, L., Haenig, B., Herrmann, B.G., Kispert, A., 2000. Prakash, N., Brodski, C., Naserke, T., Puelles, E., Gogoi, R., Hall, A., The paired homeobox gene Uncx4. 1 specifies pedicles, transverse Panhuysen, M., Echevarria, D., Sussel, L., Weisenhorn, D.M., et al., 2006. A processes and proximal ribs of the vertebral column. Development 127, Wnt1-regulated genetic network controls the identity and fate of midbrain- 2259–2267. dopaminergic progenitors in vivo. Development 1, 89–98. Li, J.Y., Joyner, A.L., 2001. Otx2 and Gbx2 are required for refinement and not Qun, L.X., Pirvola, U., Saarma, M., Ylikoski, J., 1999. Neurotrophic factors in induction of mid–hindbrain gene expression. Development 128, 4979–4991. the auditory periphery. Ann. N. Y. Acad. Sci. 884, 292–304. T. Jukkola et al. / Developmental Biology 297 (2006) 141–157 157

Rizzoti, K., Brunelli, S., Carmignac, D., Thomas, P.Q., Robinson, I.C., Lovell- Tsang, M., Friesel, R., Kudoh, T., Dawid, I.B., 2002. Identification of Sef, a Badge, R., 2004. SOX3 is required during the formation of the novel modulator of FGF signalling. Nat. Cell Biol. 4, 165–169. hypothalamo–pituitary axis. Nat. Genet. 36, 247–255. Urayama, A., Yamada, S., Kimura, R., Zhang, J., Watanabe, Y., 2002. Robinson, M., Parsons Perez, M.C., Tebar, L., Palmer, J., Patel, A., Marks, D., Neuroprotective effect and brain receptor binding of taltirelin, a novel Sheasby, A., De Felipe, C., Coffin, R., Livesey, F.J., Hunt, S.P., 2004. thyrotropin-releasing hormone (TRH) analogue, in transient forebrain FLRT3 is expressed in sensory neurons after peripheral nerve injury and ischemia of C57BL/6J mice. Life Sci. 72, 601–607. regulates neurite outgrowth. Mol. Cell. Neurosci. 27, 202–214. Wallen, A., Zetterstrom, R.H., Solomin, L., Arvidsson, M., Olson, L., Sommer, L., Ma, Q., Anderson, D.J., 1996. Neurogenins, a novel family of Perlmann, T., 1999. Fate of mesencephalic AHD2-expressing dopamine atonal-related bHLH transcription factors, are putative mammalian neuronal progenitor cells in NURR1 mutant mice. Exp. Cell Res. 253, 737–746. determination genes that reveal progenitor cell heterogeneity in the Walshe, J., Mason, I., 2000. Expression of FGFR1, FGFR2 and FGFR3 during developing CNS and PNS. Mol. Cell. Neurosci. 8, 221–241. early neural development in the chick embryo. Mech. Dev. 90, 103–110. Takahashi, M., Tsunoda, T., Seiki, M., Nakamura, Y., Furukawa, Y., 2002. Wang, X.P., Suomalainen, M., Jorgez, C.J., Matzuk, M.M., Werner, S., Identification of membrane-type matrix metalloproteinase-1 as a target of Thesleff, I., 2004. Follistatin regulates enamel patterning in mouse incisors the beta-catenin/Tcf4 complex in human colorectal cancers. Oncogene 21, by asymmetrically inhibiting BMP signaling and ameloblast differentiation. 5861–5867. Dev. Cell 7, 719–730. Takahashi, M., Kojima, M., Nakajima, K., Suzuki-Migishima, R., Motegi, Y., White, J.G., Southgate, E., Thomson, J.N., 1992. Mutations in the Caenorhab- Yokoyama, M., Takeuchi, T., 2004. Cardiac abnormalities cause early ditis elegans unc-4 gene alter the synaptic input to ventral cord motor lethality of jumonji mutant mice. Biochem. Biophys. Res. Commun. 324, neurons. Nature 355, 838–841. 1319–1323. Wilkinson, D.G., Green, J., 1990. In situ hybridization and the three- Tallafuss, A., Bally-Cuif, L., 2003. Tracing of her5 progeny in zebrafish dimensional construction of serial sections. In: Copp, A.J., Cockroft, D. transgenics reveals the dynamics of midbrain–hindbrain neurogenesis and L. (Eds.), Postimplatation Mammalian Embryos. Oxford Univ. Press, maintenance. Development 130, 4307–4323. pp. 155–171. Tassi, E., Al-Attar, A., Aigner, A., Swift, M.R., McDonnell, K., Karavanov, A., Wurst, W., Bally-Cuif, L., 2001. Neural plate patterning: upstream and Wellstein, A., 2001. Enhancement of fibroblast growth factor (FGF) activity downstream of the isthmic organizer. Nat. Rev., Neurosci. 2, 99–108. by an FGF-binding protein. J. Biol. Chem. 276, 40247–40253. Xu, J., Liu, Z., Ornitz, D.M., 2000. Temporal and spatial gradients of Fgf8 and Thomas, K.R., Capecchi, M.R., 1990. Targeted disruption of the murine int-1 Fgf17 regulate proliferation and differentiation of midline cerebellar proto-oncogene resulting in severe abnormalities in midbrain and cerebellar structures. Development 127, 1833–1843. development. Nature 346, 847–850. Yamada, M., Hashimoto, K., Satoh, T., Shibusawa, N., Kohga, H., Ozawa, Y., Toyoda, M., Shirato, H., Nakajima, K., Kojima, M., Takahashi, M., Kubota, M., Yamada, S., Mori, M., 1997. A novel transcript for the thyrotropin-releasing Suzuki-Migishima, R., Motegi, Y., Yokoyama, M., Takeuchi, T., 2003. hormone receptor in human pituitary and pituitary tumors. J. Clin. jumonji downregulates cardiac cell proliferation by repressing cyclin D1 Endocrinol. Metab. 82, 4224–4228. expression. Dev. Cell 5, 85–97. Ye, W., Shimamura, K., Rubenstein, J.L., Hynes, M.A., Rosenthal, A., 1998. Trokovic, R., Trokovic, N., Hernesniemi, S., Pirvola, U., Vogt Weisen- FGF and Shh signals control dopaminergic and serotonergic cell fate in the horn, D.M., Rossant, J., McMahon, A.P., Wurst, W., Partanen, J., anterior neural plate. Cell 93, 755–766. 2003. FGFR1 is independently required in both developing mid- and Yusoff, P., Lao, D.H., Ong, S.H., Wong, E.S., Lim, J., Lo, T.L., Leong, H.F., hindbrain for sustained response to isthmic signals. EMBO J. 22, Fong, C.W., Guy, G.R., 2002. Sprouty2 inhibits the Ras/MAP kinase 1811–1823. pathway by inhibiting the activation of Raf. J. Biol. Chem. 277, 3195–3201. Trokovic, R., Jukkola, T., Saarimaki, J., Peltopuro, P., Naserke, T., Weisenhorn, Zervas, M., Millet, S., Ahn, S., Joyner, A.L., 2004. Cell behaviors and genetic D.M., Trokovic, N., Wurst, W., Partanen, J., 2005. Fgfr1-dependent lineages of the mesencephalon and rhombomere 1. Neuron 3, 345–357. boundary cells between developing mid- and hindbrain. Dev. Biol. 278, Zhang, S., Lin, Y., Itaranta, P., Yagi, A., Vainio, S., 2001. Expression of Sprouty 428–439. genes 1, 2 and 4 during mouse organogenesis. Mech. Dev. 2, 367–370.