Fibroblast growth factor (FGF) gene expression ... - Wiley Online Library

Download PDF
3 downloads 0 Views 12MB Size Report
Comment
Jun 19, 2009 - ... of Craniofacial Development, King's College London, Guy's Campus, London, United Kingdom ..... al., 1993; Han and Martin, 1993). How-.
DEVELOPMENTAL DYNAMICS 238:2058 –2072, 2009

PATTERNS & PHENOTYPES

Fibroblast Growth Factor (FGF) Gene Expression in the Developing Cerebellum Suggests Multiple Roles for FGF Signaling During Cerebellar Morphogenesis and Development Yuichiro Yaguchi,1,2 Tian Yu,1 Mohi U. Ahmed,1† Mary Berry,3 Ivor Mason,3 and M. Albert Basson1,3*

The cerebellum is derived from the anterior-most segment of the embryonic hindbrain, rhombomere 1 (r1). Previous studies have shown that the early development and patterning of r1 requires fibroblast growth factor (FGF) signaling. However, many of the developmental processes that shape cerebellar morphogenesis take place later in embryonic development and during the first 2 weeks of postnatal life in the mouse. Here, we present a more comprehensive analysis of the expression patterns of genes encoding FGF receptors and secreted FGF ligands during these later stages of cerebellar development. We show that these genes are expressed in multiple cell types in the developing cerebellum, in an astonishing array of distinct patterns. These data suggest that FGF signaling functions throughout cerebellar development to regulate many processes that shape the formation of a functional cerebellum. Developmental Dynamics 238: 2058 –2072, 2009. © 2009 Wiley-Liss, Inc. Key words: cerebellum; development; morphogenesis; FGF; gene expression Accepted 15 May 2009

INTRODUCTION The cerebellum plays a pivotal role in the fine coordination of posture and locomotion. The mammalian cerebellum consists of a medial vermis, flanked laterally by a pair of hemispheres. Based on molecular and functional data, the cerebellum can be subdivided further into distinct parasagittal domains and specific zones in the anteroposterior plane. The somatotopic and molecular regionalization

of the cerebellum is complex, and the mechanisms that control the development of specific regions are not fully understood (reviewed in Manni and Petrosini, 2004; Larouche and Hawkes, 2006; Sillitoe and Joyner, 2007). At the cellular level, cerebellar neurons are organized in distinct layers that are folded in a highly stereotypical manner (Altman and Bayer, 1997). The characteristic foliated structure and laminated neuronal organization of the

cerebellum are the result of the coordinated expansion, differentiation, and migration of the neuronal precursors that eventually form the mature neurons of the cerebellum (reviewed by Hatten and Heintz, 1995). Three neuronal layers can be distinguished in the adult cerebellar cortex: the molecular, Purkinje cell, and granule cell layers (see Fig. 1E,J). The molecular layer (ML), which is on the outside closest to the pial surface, con-

Additional Supporting Information may be found in the online version of this article. 1 Department of Craniofacial Development, King’s College London, Guy’s Campus, London, United Kingdom 2 Department of Otorhinolaryngology, The Jikei University School of Medicine, Tokyo, Japan 3 MRC Centre for Developmental Neurobiology, King’s College London, Guy’s Campus, London, United Kingdom Grant sponsor: The Wellcome Trust; Grant number: WT080470; Grant sponsor: Medical Research Council. † Dr. Ahmed’s present address is Department of Genetics and Genomic Sciences, The Mount Sinai School of Medicine, New York, NY 10029. *Correspondence to: M. Albert Basson, Department of Craniofacial Development, King’s College London, Floor 27, Guy’s Tower, London SE1 9RT, UK. E-mail: [email protected] DOI 10.1002/dvdy.22013 Published online 19 June 2009 in Wiley InterScience (www.interscience.wiley.com).

© 2009 Wiley-Liss, Inc.

FGF GENE EXPRESSION IN THE CEREBELLUM 2059

tains parallel fibers and several subpopulations of interneurons, including stellate and basket cells. The Purkinje cell layer (PCL) is formed by a monolayer of large Purkinje cells (PCs) and Bergmann glia (BG), which are located just below and interspersed between the Purkinje cell bodies. Beneath the PCL is the granule cell layer (GCL) that contains the densely packed granule neurons, as well as some Golgi cells. White matter (WM) tracts are found underneath the GCL, while a set of deep cerebellar nuclei (DCN) reside beneath the cerebellar cortex (reviewed in Hatten and Heintz, 1995; Altman and Bayer, 1997; Sillitoe and Joyner, 2007). The cerebellum develops from the dorsal part of rhombomere 1 (r1), the most anterior segment of the hindbrain (recently reviewed by Zervas et al., 2005; Sillitoe and Joyner, 2007). At early developmental stages, molecules that control both mesencephalic (mes) and r1 development are produced by a signaling center, the isthmic organizer (IsO) located at the mes/r1 boundary (reviewed by Echevarria et al., 2003; Nakamura and Watanabe, 2005; Partanen, 2007). These molecules include members of the fibroblast growth factor (FGF) family of secreted factors (reviewed by Itoh and Ornitz, 2004; Mason, 2007). Neurogenesis is initiated after the initial stages of r1 growth and patterning. Most types of cerebellar neurons are born within the ventricular zone from approximately embryonic day (E) 10.25 in the mouse. PC precursors are born in sequential waves between E10.5 and E12.5 (Hashimoto and Mikoshiba, 2003) and migrate radially on glial fibers toward the pial surface of the cerebellar anlage to become organized as a diffuse layer underneath the pial surface by birth (postnatal day [P] 0; Miale and Sidman, 1961; Goldowitz and Hamre, 1998; Wang and Zoghbi, 2001; Sotelo, 2004, Fig. 1A–C). BG are unipolar astrocytes that are derived from radial glia in the ventricular zone. Their cell bodies are located just beneath the PC soma within the PCL during late embryogenesis, from where they extend radial fibers to the pial surface (Yuasa, 1996, Fig. 1B,C). By contrast, GC precursors (GCps) are born in the ante-

rior rhombic lip (RL; reviewed by Wingate, 2001) and migrate tangentially over the pial surface of the cerebellar anlage in an anterior direction from approximately E12.5 in the mouse embryo (Fig. 1A), to form a thin layer, the external granule cell layer (EGL) that covers the whole cerebellum by E16.5 (Miale and Sidman, 1961; Fig. 1B,C). They proliferate extensively during the first fortnight after birth to generate a vast number of GCps (Fig. 1C,D). Groups of GCps in the EGL exit the cell cycle at defined times (Espinosa and Luo, 2008) and migrate radially through the PCL on Bergmann glial fibers (Goldowitz and Hamre, 1998; Wang and Zoghbi, 2001; Komuro and Yacubova, 2003; Sotelo, 2004), to form the internal granule cell layer (IGL) within which they mature into GCs (Fig. 1D). During this period, PCs organize themselves into a monolayer and extend complex dendritic trees into the ML, apparently in close association with glial fibers (Lordkipanidze and Dunaevsky, 2005). These regulated processes of expansion, migration, and differentiation exert forces within the developing cerebellum, such that fissures and folia form in a highly stereotypical manner, possibly involving the establishment of “anchoring centers” (Sudarov and Joyner, 2007). Our understanding of the molecular mechanisms that regulate these complex processes remain incomplete. Previous studies have demonstrated that FGF8 and FGF17 have important functions during the early development of the mes and r1 (Meyers et al., 1998; Reifers et al., 1998; Martinez et al., 1999; Shamim et al., 1999; Irving and Mason, 2000; Xu et al., 2000; Chi et al., 2003; Liu et al., 2003; Basson et al., 2008). The FGF signaling system has at its disposal 22 genes encoding different ligands that can broadly be subdivided into 7 subfamilies based on sequence homology. With the exception of the four genes in the Fgf11 subfamily (Fgf11–14), which encode proteins that remain intracellular (Itoh and Ornitz, 2004), these genes encode secreted FGF ligands that function by activating cell surface FGF receptors. These signaling receptors are encoded by four FGF receptor genes, which can give rise to “IIIb” or “IIIc”

type isoforms by using one of two unique exons to produce two different versions of the third extracellular Ig-like loop (reviewed by Mason, 2007; Itoh and Ornitz, 2008). The secreted FGF ligands exhibit different affinities for specific receptor isoforms resulting in preferential ligand-receptor specificities (Zhang et al., 2006). Recently, a new FGF receptor gene, Fgfr5 (also known as Fgfrl1), has been identified which encodes a receptor that lacks the intracellular signaling domain (Sleeman et al., 2001; Trueb and Taeschler, 2006; Catela et al., 2009). Fgfr1, Fgfr2, and Fgfr3 genes are expressed in the mes/r1 region at E9.5 (Trokovic et al., 2003, 2005; Blak et al., 2005) and conditional deletion of all three of these FGF receptor genes in mes/r1 are required to phenocopy mutants in which Fgf8 is deleted in the same region (Chi et al., 2003; Saarimaki-Vire et al., 2007). These defects have been attributed to early functions in mes/r1 development while the role of these receptors during later cerebellar development has not been investigated until recently (Lin et al., 2009). Some studies have reported the expression of Fgf1, Fgf2, Fgf3, Fgf5, Fgf7, Fgf9, Fgf10, and Fgf22 in the developing postnatal rodent cerebellum, suggesting that these FGF ligands may have roles during postnatal cerebellar development (Wilkinson et al., 1989a; Wilcox and Unnerstall, 1991; Matsuda et al., 1994; Ozawa et al., 1996; Hattori et al., 1997; McAndrew et al., 1998; Nakamura et al., 1999; Gimeno et al., 2003; Umemori et al., 2004). A direct comparison of the expression of all FGF ligands and receptors at key stages of cerebellar morphogenesis is required as a starting point to understand the role of FGF signaling during cerebellar development. We performed a comprehensive analysis of Fgf gene expression at five key developmental stages in cerebellar development: E12.5, E16.5, P0, P7, and P21 (Fig. 1) and uncovered a complex picture of spatiotemporal gene expression changes, indicating that various aspects of cerebellar development may be regulated by FGF signaling.

2060 YAGUCHI ET AL.

RESULTS FGF Signaling Persists in the Developing Cerebellum During Late Embryogenesis and Postnatal Life It is not known whether FGF signaling is required during cerebellar morphogenesis after E12.5. As a first step to address this possibility, we asked whether FGF signaling activity could be detected in the developing cerebellum after E12.5. To visualize FGF signaling in cerebellar tissue, we determined the expression of two transcriptional targets of the FGF signaling pathway, the ETS-related factor genes Etv4 (Pea3) and Etv5 (Erm; O’Hagan and Hassell, 1998; Roehl and Nusslein-Volhard, 2001; Klein et al., 2008). Both genes were expressed in an identical manner in the cerebellar anlage at E12.5, E16.5, P0, and P7 of development (Fig. 1K–O; Etv4 data not shown). At E12.5, the r1 territory had been established, GCps were being born in the rhombic lip and migrating over the pial surface toward the IsO (Fig. 1A,F). At this stage, Etv5 was expressed at high levels in the posterior mesencephalon (developing inferior colliculus, IC) and r1 (Fig. 1K). The most posterior region of the RL did not express Etv5, suggesting that these cells were not receiving FGF signals (Fig. 1K⬘). At E16.5, when PC and BG precursors were organizing underneath the pial surface, the EGL had formed and granule cell expansion and radial migration had been initiated (Fig. 1B,G), Etv5 was expressed in the PCL, in GCps in the EGL and in cells scattered through the anlage (Fig. 1L). Of interest, the level of Etv5 expression was highest in the EGL over the posterior region of the E16.5 cerebellum, suggesting that cells in this region were experiencing more FGF signaling than cells in other regions (Fig. 1L,L⬘). We also detected Etv5 expression in the ventricular zone (VZ) at this stage, indicating that FGF signaling might be involved in the generation of neurons from the VZ. At birth (P0), many GCps in the EGL were proliferating and some were beginning to differentiate and migrate on glial fibers to initiate fissure formation (Fig. 1C,H). Etv5 transcripts were present at high levels in

the PCL, although the strongest expression appeared to be in BG located below the PCs (Fig. 1M,M⬘). High FGF signaling was also evident in the EGL and migrating GCps in the posterior cerebellum (lobules VII–IX) at this stage (Fig. 1M,M⬘). At P7, PCs had organized into a monolayer, distinct folia had formed, and many GCs were migrating to form the IGL (Fig. 1D,I). Etv5 was detected in the PCL with highest expression in BG and in GCps in the posterior IGL (folia VII–IX; Fig. 1N,N⬘). At weaning (P21), cerebellar morphogenesis was complete and the characteristic layered, foliated structure of the cerebellum could be observed in histological sections (Fig. 1E,J). Etv5 expression was still evident in PCs, BG, and GCs at P21 (Fig. 1O,O⬘), suggesting that FGF signaling remained active in the cerebellum after the completion of morphogenesis. We confirmed that Etv5 expression was positively regulated by FGF signaling in the developing cerebellum by showing that its expression was reduced upon culturing cerebellar slices from newborn mice in the presence of a specific FGF receptor inhibitor, PD173074 (data not shown). Thus, we conclude that FGF signaling activity was not restricted to early stages of

mes/r1 development, but that it continued throughout cerebellar morphogenesis.

Multiple FGF Ligands Are Expressed During Cerebellar Development We determined the expression patterns of all 18 genes in the 6 subfamilies that encode secreted FGF ligands, together with the 5 genes that encode cell surface receptors in the E12.5, E16.5, P0, P7, and P21 mouse cerebellum.

The FGF1 Subfamily Members of the FGF1 subfamily (Fgf1 and Fgf2) were both expressed at extremely low levels during cerebellar development, and detecting the specific transcripts above any background staining by in situ hybridization was difficult. Fgf1 transcripts were detected in the area around the mid– hindbrain region at E12.5 (Fig. 2A) and in cells throughout the E16.5 cerebellar anlage, including cells in the EGL (Fig. 2B). At P0 and P7, both the EGL and developing IGL appeared to contain Fgf1 mRNA (Fig. 2C,C⬘,D,D⬘). Fgf1 mRNA was still

Fig 1. Fibroblast growth factor (FGF) signaling in the developing cerebellum from embryonic day (E) 12.5. A–E: Diagram depicting key stages during cerebellar morphogenesis as observed in sagittal sections, with anterior to the left. F–J: Cresyl violet (Nissl)- stained sagittal sections of mouse cerebella at the indicated stages of development to demonstrate structure and cellular organization. Developing cerebellar fissures at P0 are indicated in H (pcu, pr, ppy, se, po). The developing and final lobules are numbered in H–J using Roman numerals according to Inouye and Oda (1980). K–O: Etv5 gene expression as determined by in situ hybridization (blue) on sagittal sections counterstained with nuclear fast red (pink) are shown. The approximate position of the isthmic organizer is indicated by an arrow and the cerebellar anlage in dorsal r1 is outlined by broken black lines in K. BG, Bergmann glia; EGL, external granule cell layer; GCL, granule cell layer; IC, presumptive inferior colliculus (posterior mesencephalon); IGL, internal granule cell layer; IsO, isthmic organizer; ML, molecular layer; PCL, Purkinje cell layer; r1, rhombomere 1; RL, rhombic lip; vz, ventricular zone of the 4th ventricle; WM, white matter; pcu, preculminate; pr, primary; ppy, prepyramidal; se, secondary and po, posterior fissure. Scale bars ⫽ 250 ␮m in K–O; 50 ␮m in K⬘–O⬘. Fig 2. Fibroblast growth factor-1 (FGF1) subfamily expression during cerebellar morphogenesis. A–Fⴕ: Fgf1 (A–D⬘) and Fgf2 (E–F⬘) gene expression as determined by in situ hybridization (blue) on sagittal sections counterstained with nuclear fast red (pink). Anterior is to the left. The approximate position of the isthmic organizer is indicated by an arrow and the cerebellar anlage in dorsal r1 is outlined by broken black lines in A. CP, choroid plexus; other abbreviations as in legend to Figure 1. Fig 3. Fibroblast growth factor-4 (FGF4) subfamily expression during cerebellar morphogenesis. A–K: Fgf4 (A–C⬘), Fgf5 (D–G), and Fgf6 (H–K) gene expression as determined by in situ hybridization (blue) on sagittal sections counterstained with nuclear fast red (pink). The section in F was immunostained with an antibody to S100 to identify BG (brown). Anterior is to the left. Panels B⬘, C⬘, D⬘, I⬘, and J⬘ are higher magnification views of the regions indicated in B, C, D, I, and J, respectively. The approximate position of the isthmic organizer is indicated by arrows in the alar plate and the cerebellar anlage in dorsal r1 is outlined by broken black lines. mgs, meninges; X, lobule X; other abbreviations as in legend to Figure 1.

Fig. 1.

Fig. 2.

Fig. 3.

2062 YAGUCHI ET AL.

present in the PCL and GCL of the P21 cerebellum, although the PCL expression appeared not to localize to PC soma (Supp. Fig. 1A,a, which is available online). We could not detect any Fgf2 mRNA in the E12.5 and E16.5 cerebellum (not shown). At P0, Fgf2 was expressed in a few GCps in the inner EGL as well as some migrating GCps (Fig. 2E,E⬘). By P7, many cells in the IGL expressed Fgf2 (Fig. 2F,F⬘). Fgf2 was expressed in PCs in the P21 cerebellum (Supp. Fig. S1B,b).

The FGF4 Subfamily Neither Fgf4 nor Fgf5 were expressed in r1 at E12.5 (data not shown). Fgf4 mRNA was detected in many cells distributed throughout the cerebellar anlage at E16.5, including cells accumulating in the forming PCL (Fig. 3A). Fgf4 was expressed in the EGL and PCL at P0 (Fig. 3B,B⬘). Expression was also evident in small, migrating GCps and cells within the forming IGL (Fig. 3B⬘). This general pattern of expression was maintained at P7 with transcripts in PCs and strong expression in the IGL (Fig. 3C,C⬘). At P21, Fgf4 mRNA was present in some PCs and possibly also BG in the PCL and in some cells in the GCL (Supp. Fig. S1C,c). Fgf5 was first detected in a discrete population of cells in the posterior cerebellum at E16.5 (Fig. 3D,D⬘). Fgf5 was not detected in LHX1/5⫹ PC precursors but was expressed in some S100⫹ BG (not shown), suggesting that many of these cells were glia. Fgf5 was expressed in a similar posterior domain at P0 (Fig. 3E), around the secondary (se) fissure (prospective lobules VII–IX). At P0, BG cell bodies were located just underneath the PC somas. Staining these sections with an antibody to LHX1/5 to detect PC bodies, showed that cells expressing Fgf5 were located just underneath the PCL (data not shown), whereas staining with an S100 antibody to detect BG confirmed that many Fgf5-expressing cells were indeed glia (Fig. 3F). Many of the small GCps that were migrating from the EGL to the IGL also expressed Fgf5 (Fig. 3F). By P7, Fgf5 was expressed at high levels in almost all GCs in the IGL with the exception of those cells in folium X

(Fig. 3G), which are the last GCps to be generated from the RL (Machold and Fishell, 2005). No clear expression was evident in BG at P7 or P21 (not shown), although low levels of expression in a small number of cells could not be excluded. According to earlier reports, Fgf6 was expressed exclusively in myogenic cells during embryogenesis (deLapeyriere et al., 1993; Han and Martin, 1993). However, Ozawa et al. detected Fgf6 mRNA in the P5 cerebellum (Ozawa et al., 1996). We detected high levels of Fgf6 expression in the meninges and in cells of the more posterior regions of the cerebellar anlage at E12.5 (Fig. 3H). By E16.5, strong Fgf6 expression was observed in a small population of cells in the posterior cerebellum (Fig. 3I,I⬘). These cells appeared to be a small number of PCs that would subsequently expand to populate the developing lobules VI and VII between the primary (pr) and prepyramidal (ppy) fissures at P0 (Fig. 3J,J⬘). Note that these cells were localized more superficially (i.e., closer to the EGL) than the Fgf5-expressing glia and GCps (Fig. 3E,F). Fgf6 expression was still evident in the PCs in lobules VI–VII at P7 (Fig. 3K). There was also a population of PCs in the posterior cerebellum (lobules VIII–X) that expressed lower levels of Fgf6 from E16.5 onward (Fig. 3I,J). Some GCps in the EGL expressed Fgf6 at E16.5 and P0 (Fig. 3I,J), an expression pattern maintained during postnatal development, as a subpopulation of GCs in the IGL remained Fgf6-positive at P7 (Fig. 3K). A few other cells of unknown identity expressed Fgf6 at E16.5 and P0. Some of these cells were located adjacent to the VZ (Fig. 3I,J), suggesting that they were newly generated neuronal precursors. Fgf6 was expressed in cells in the ML, PCL (some PCs and possibly BG), and the GCL of the P21 cerebellum (Supp. Fig. S1D,d).

The FGF7 Subfamily We did not detect significant amounts of Fgf3 transcript in the E12.5 cerebellum (data not shown). By E16.5 of development, Fgf3 was expressed in cells adjacent to the VZ and cells in the PCL, with the highest expression in the posterior half of the cerebellar anlage (Fig. 4A). This observation is in agreement with a previous study dem-

onstrating Fgf3 expression in PC precursors leaving the VZ at E14.5 (Wilkinson et al., 1989a). Wholemount preparations revealed a complex expression pattern, with Fgf3 mRNA present in distinct regions along the mediolateral axis of the cerebellar anlage. In the prospective vermis, Fgf3 was expressed in two stripes that straddled an Fgf3-negative midline region (Fig. 4B). Where these stripes reached the posterior cerebellum, they extended further laterally to the approximate edge of the prospective vermis (Fig. 4B). Fgf3 was expressed strongly in the anterior halves of the prospective cerebellar hemispheres, and this expression extended all along the lateral edges of the hemispheres (Fig. 4B). Strong Fgf3 expression was still evident in the PCL at P0; however, considerable changes in its expression pattern were apparent at this stage. Most strikingly, PCs in the anterior cerebellum encompassing those that would contribute to lobes III– VII, expressed the highest levels of Fgf3 (Fig. 4C), in agreement with a previous study (Wilkinson et al., 1989). As Fgf3 expression was strong enough to allow combined in situ hybridization and immunohistochemistry, we could demonstrate that Fgf3 was indeed expressed in LHX1/5⫹ PC precursors (Fig. 4D). We confirmed this observation by showing that most Fgf3⫹ cells in the PCL were not S100⫹ BG (Fig. 4E). Curiously, some GCps migrating through the Fgf3-expressing PCL also expressed high levels of Fgf3 (Fig. 4D), suggesting that factors specific to this region and acting over a short range can induce Fgf3 expression in migrating GCps. We also determined the expression pattern of Fgf3 in whole brains and coronal sections at P0. Fgf3 was expressed in distinct parasagittal domains in the vermis (Fig. 4F,G) and hemispheres (Fig. 4F). At P7, the expression of Fgf3 was maintained in the PCL of lobules I–VII (Fig. 4H). Fewer cells in lobules I–III and VIb–VII expressed Fgf3 compared with lobules IV–VIa (Fig. 4H). Of interest, GCps in the EGL also expressed Fgf3 at this stage of development with highest levels of expression evident in lobules I–VIa and IXb–X (Fig. 4H). Fgf3 was no longer expressed at P21. Fgf7 mRNA was first detected in a small population of PCs in the lateral cerebellar anlage at P0 (data not shown). By P7, it was expressed in

FGF GENE EXPRESSION IN THE CEREBELLUM 2063

Fgf10 was expressed at low levels in the EGL of the E16.5 and P0 cerebellum (data not shown). At P7, many cells in the EGL and IGL expressed Fgf10 (Fig. 4J), in agreement with a previous study (Umemori et al., 2004). Of interest, the cells expressing Fgf10 mRNA in the anterior zone (lobules I–V) appeared to be primarily in the crowns of the folia, with few expressing cells in their basal regions. Low amounts of Fgf10 transcripts could be detected in the PCL of the P21 cerebellum (Supp. Fig. S1G,g). We detected Fgf22 transcripts in both the EGL and IGL of the P7 cerebellum (Fig. 4K). Fgf22 expression was not found at other developmental stages.

The FGF8 Subfamily

Fig 4. Fibroblast growth factor-7 (FGF7) subfamily expression during cerebellar morphogenesis. A,C–E,H–K: Fgf3 (A,C–E,H), Fgf7 (I), Fgf10 (J), and Fgf22 (K) gene expression as determined by in situ hybridization (blue) on sagittal sections counterstained with nuclear fast red (pink). Anterior is to the left. B,F: In situ hybridization for Fgf3 on brains in whole-mount are shown (anterior to the top). The approximate positions of the medial vermis and cerebellar hemispheres are indicated in B and the whole cerebellar anlage outlined with broken lines in F. The approximate position of the sagittal section shown in A, which shows strong expression in the posterior cerebellum is indicated by a broken red line in the whole-mount preparation in B. D,E: Combined in situ hybridization for Fgf3 (blue) with immunohistochemical identification of LHX1/5⫹ Purkinje cells or S100⫹ Bergmann glia (brown) in sagittal sections. PCs and migrating GCps are indicated. G: Fgf3 gene expression in coronal sections of a newborn cerebellum is shown (dorsal is to the top) with the vermis indicated. Abbreviations as in Figure 1.

PCs in the anterior zone up to and including the anterior edge of lobule VIa, as well as those in lobule IX and the posterior edge of lobule VIII. Most PCs in the central zone of the P7 cerebellum

(folia VIb–VII) did not express Fgf7. Of interest, Fgf7 transcripts were readily detectable in most PCs in the P21 cerebellum, with the exception of cells in lobule X (Supp. Fig. S1E,e).

We observed Fgf8 expression in anterior r1 with a sharp boundary just anterior to the morphological IsO at E12.5 (Fig. 5A), in agreement with previous studies (Hoshikawa et al., 1998; Xu et al., 2000). Note that the section in Figure 5A included the floor plate (basal plate midline), which did not express Fgf8 (Mahmood et al., 1995). Fgf17 (Fig. 5E) and Fgf18 (Fig. 5G) were expressed in a similar, overlapping pattern to Fgf8 with Fgf18 expressed at considerably lower levels than Fgf8 and Fgf17. A previous study reported that the expression of Fgf8-related genes was maintained in the IsO up to approximately E12.5 (Fgf8, Fgf18) and through E14.5 (Fgf17) of development (Xu et al., 2000). We also investigated later stages of development, and detected low levels of Fgf8 mRNA in GCps within the posterior EGL of the E16.5 cerebellum (Fig. 5B). In addition, some cells in the velum medullare (VM), the thin tissue derived from the IsO connecting the IC with the developing cerebellum, expressed Fgf8 (Fig. 5B). Likewise, Fgf17 was strongly expressed in the VM at E16.5, but unlike Fgf8, it was not expressed in the EGL (Fig. 5F). Fgf8 expression was maintained in the posterior EGL and VM at birth (Fig. 5C), and remained in the EGL of the IXth folium at P7 (Fig. 5D). We confirmed the apparent low Fgf8 transcript level by

Fig 5. Fibroblast growth factor-8 (FGF8) subfamily expression during cerebellar morphogenesis. A–H: Fgf8 (A–D), Fgf17 (E,F), and Fgf18 (G,H) gene expression as determined by in situ hybridization (blue) on sagittal sections counterstained with nuclear fast red (pink). I: Hemisected cerebellum from an Fgf8lacZ/⫹ newborn brain stained with X-gal in whole-mount, anterior is to the left in all cases. J: Vibratome section to confirm expression in VM and EGL. VM, velum medullare; other abbreviations as in Figure 1.

Fig 6. Expression of fibroblast growth factor-9 (FGF9) and FGF15 subfamilies during cerebellar morphogenesis. A–F,H: Fgf9 (A–C) and Fgf15 (D–F,H) gene expression as determined by in situ hybridization (blue) on sagittal sections counterstained with nuclear fast red (pink). Anterior is to the left. G: Fgf15 gene expression in coronal sections of a newborn cerebellum (dorsal is to the top) with parasagittal expression domains indicated by arrowheads. Abbreviations as in Figure 1.

staining whole brains from E16.5 embryos (data not shown) and newborn Fgf8lacZ/⫹ animals (Ilagan et al., 2006) for lacZ activity (Fig. 5I). This experiment confirmed the expression of Fgf8 in the posterior EGL, and also demonstrated an abundance of ␤-galactosidase activity in the VM (Fig. 5I,J). An explanation for the relative abundance of ␤-galactosidase activity compared with low amounts of Fgf8 mRNA detectable in the VM, may be that these cells proliferate much less than GCps, thus allowing accumulation of the stable ␤-galactosidase enzyme over time. In contrast to Fgf8, Fgf17 expression was not detected after E16.5 (data not shown). Fgf18 mRNA, which could not be detected in the E16.5 or P0 cerebellum (data not shown), was observed in some GCs in the EGL and IGL at P7, with slightly lower amounts in lobules VI–VII (Fig. 5H). Fgf18 expression was maintained in the P21 cerebellum when it could be detected in the GCL and some cells in the PCL (Supp. Fig. S1H,h).

FGF GENE EXPRESSION IN THE CEREBELLUM 2065

The FGF9 Subfamily Fgf9 expression first appeared in cells throughout the E16.5 cerebellum, albeit at low levels (Fig. 6A). At birth, Fgf9 was expressed at high levels in many GCps in the EGL and expression was maintained in migrating GCps (Fig. 6B). Fgf9 expression was still evident in most GCs, including those that had completed their migration to the IGL at P7 (Fig. 6C), and was occasionally detected in isolated cells that appeared to be BG (not shown). Of interest, Fgf9 transcripts were detected primarily in PCs at P21 (Supp. Fig. S1F,f). Neither Fgf16 nor Fgf20 mRNA were detected in the developing cerebellum at any of the stages investigated (data not shown).

The FGF15 Subfamily An earlier report demonstrated that Fgf15 was expressed within cells in the posterior mes and r1, but excluded from the IsO at E9.5–E11.5 (Gimeno et al., 2003). At E12.5, we found Fgf15 expression in the posterior mes (prospective IC) and posterior r1, but not in the IsO and anterior r1 (Fig. 6D). At E16.5, Fgf15 expression remained at high levels in the developing IC and was also expressed in PCs in the cerebellar anlage (Gimeno et al., 2003; Fig. 6E). Fgf15 could be detected at high levels in the PCL at P0 (Fig. 6F) and as previously reported (Gimeno et al., 2003), Fgf15 was expressed in parasagittal domains (Fig. 6G). Low levels of Fgf15 expression were detected in the EGL, PCL, and IGL of the P7 cerebellum in cells that may also include BG (Fig. 6H). Fgf15 was not expressed in the P21 cerebellum. We did not detect any Fgf21 or Fgf23 mRNA in the developing cerebellum (data not shown), although low levels of Fgf21 transcripts may be present at P21 (Supp. Fig. S1H,h).

FGF Receptor Gene Expression in the Developing Cerebellum The expression patterns of genes encoding secreted FGF ligands in the developing cerebellum were found to be complex. To identify the potential target cells for these FGF ligands, we determined the expression of genes

encoding FGF receptors. We detected all FGF receptor genes in r1 at E12.5. Fgfr1 was expressed at similar levels in both dorsal and alar plates throughout the mid– hindbrain region (Fig. 7A). Fgfr2 was expressed at high levels in the anterior mes, low levels in posterior r1 and excluded from the IsO (Fig. 7E). Expression of Fgfr3 was much stronger in the basal plate than the alar plate of mes/r1 but was detected at low levels in some cells within the cerebellar anlage (Fig. 7I). By contrast, Fgfr4 was expressed strongly in the cerebellar anlage (Fig. 7M). We also detected expression of Fgfr5 in both dorsal and alar plates, including the IsO (Fig. 7Q). At E16.5, Fgfr1 was detected in the PCL, GCps in the EGL, and in most cells of the VZ (Fig. 7B). Fgfr2 was expressed at high levels in the hindbrain choroid plexus (CP) and in a similar pattern to Fgfr1 in the cerebellar anlage, albeit at lower levels (Fig. 7F). By contrast, Fgfr3 expression was only detected in the VZ (Fig. 7J). Fgfr4 was expressed exclusively within the EGL, suggesting that this receptor was the primary receptor for transducing FGF signals in GCps (Fig. 7N). Fgfr5 was expressed in cells within the ventricular side of the IC and in the VZ of r1, while some cells in the EGL, PCL, and CP also expressed Fgfr5 (Fig. 7R). By P0, Fgfr1 was expressed at high levels in the PCL, in some migrating GCps and a subpopulation of cells within the posterior EGL (Fig. 7C,C⬘). Fgfr2 mRNA was detected in BG beneath and surrounding PCs, in a few GCps within the EGL and in GCps migrating to the IGL (Fig. 7G,G⬘). We could not exclude the possibility that Fgfr1 was also expressed in a few BG in close association with the PCs in the PCL (Fig. 7C⬘). Similarly, Fgfr2 might also be expressed in some PCs in the PCL (Fig. 7G⬘). Fgfr3 was expressed in some cells within the DCN (Fig. 7K,K⬘). Fgfr4 was expressed exclusively in the EGL (Fig. 7O,O⬘), whereas Fgfr5 expression was detected in several cell types in the newborn cerebellum, including BG, GCps, and cells in the VZ (Fig. 7S,S⬘). By P7, both Fgfr1 and Fgfr2 were still detected in the PCL and in some cells located within the IGL, particularly those closest to the PCL (Fig. 7D,

D⬘,H,H⬘). Within the PCL, Fgfr1 and Fgfr2 expression was found in both PCs and BG, but with Fgfr1 preferentially expressed in PCs and Fgfr2 in BG (Fig. 7D⬘,H⬘). Fgfr3 mRNA was detected in the DCN and was weakly expressed in the IGL and BG (Fig. 7L,L⬘). Fgfr4 expression was maintained in the EGL as previously reported (Miyake et al., 1995; Fig. 7P,P⬘). Fgfr5 mRNA was present in many cells in the cerebellum with highest levels of expression in BG (Fig. 7T,T⬘). At P21, Fgfr1, Fgfr2, Fgfr3, and Fgfr5 were all expressed in the PCL with Fgfr1 and Fgfr2 in PCs and BG, Fgfr3 mainly in BG, and Fgfr5 in PCs and BG. Fgfr1 was also present in cells in the GCL (Supp. Fig. S1J–M). The Fgfr1, -2, and -3 genes can generate two different types of receptor isoforms with different ligand specificity. These “IIIb” or “IIIc” isoforms are produced by alternative splicing of exon 3 (Johnson and Williams, 1993). Only some exon-specific probes have proved to be suitable for in situ hybridization (I. Mason, unpublished observations) and reverse transcriptasepolymerase chain reaction (RT-PCR) was therefore used to determine exactly which receptor isoforms were expressed (Hajihosseini and Dickson, 1999). We determined the receptor isoform expression profile of RNA isolated from cerebellar anlagen at the same stages analyzed by in situ hybridization. At all stages investigated, Fgfr1(IIIc), Fgfr2(IIIc), Fgfr3(IIIc), and Fgfr4 mRNA were detected (Fig. 7U–X). Comparatively low amounts of Fgfr3(IIIb) mRNA were present at all stages (Fig. 7U–X). Neither Fgfr1(IIIb) nor Fgfr2(IIIb) isoforms were detected. We confirmed that our PCR method was able to detect all receptor isoforms by performing the same analysis on RNA isolated from embryonic lung tissue, which expressed all these isoforms (Fig. 7Y).

DISCUSSION The role of FGF signaling during early (⬍E12.5) stages of cerebellar development has been the subject of a large number of studies. Here, we have taken the first step toward understanding the function of this signaling pathway during later stages of cere-

2066 YAGUCHI ET AL.

Fig 7. Fibroblast growth factor (FGF) receptor gene expression during cerebellar morphogenesis. A–T: Gene expression of Fgfr1 (A–D), Fgfr2 (E–H), Fgfr3 (I–L), Fgfr4 (M–P), and Fgfr5 (Q–T) expression as determined by in situ hybridization (blue) on sagittal sections counterstained with nuclear fast red (pink). Anterior is to the left. Arrows indicate the IsO and the cerebellar anlage in dorsal r1 is outlined by broken black lines in A, E, I, M, and Q. C⬘, D⬘, G⬘, H⬘, K⬘, L⬘, O⬘, P⬘, S⬘, T⬘, are high magnification views of the areas indicated in panels C, D, G, H, K, L, O, P, S, and T, respectively. U–X: Expression of FGF receptor isoforms as determined by reverse transcriptase-polymerase chain reaction (RT-PCR) from total RNA isolated from dissected cerebellar tissue at the stages indicated. Y: Expression of FGF receptor isoforms in embryonic lung tissue as positive controls for the PCR reactions in U–X. The sizes of the amplified PCR products are 306 bp (Fgfr1), 310 bp (Fgfr2), 348 bp (Fgfr3), and 195 bp (Fgfr4). White arrows indicate 348bp Fgfr3(IIIc) products. Abbreviations as in Figure 1 and 1b ⫽ Fgfr1(IIIb); 1c ⫽ Fgfr1(IIIc); 2b ⫽ Fgfr2(IIIb); 2c ⫽ Fgfr2(IIIc); 3b ⫽ Fgfr3(IIIb); 3c ⫽ Fgfr3(IIIc); 4 ⫽ Fgfr4.

bellar morphogenesis. We found that genes encoding several different FGF ligands were expressed in the cerebellar anlage during late embryogenesis and early postnatal development (Table 1). The expression of certain ligands is restricted to and maintained within specific cell types during development, whereas the expression patterns of others are dynamic, such that they are expressed in different cell types and/or regions at different times of development. These observations suggest that multiple FGF ligands are used during cerebellar development and that a single ligand may perform distinct functions at different developmental stages.

We were surprised to find so many FGF ligands expressed. The different expression patterns of FGF genes within the same subfamily, suggest that these genes have evolved gene regulatory elements that differ substantially among members of the same subfamily. Our findings suggest that FGF signaling performs important functions during cerebellar morphogenesis and that several ligands may cooperate, resulting in significant functional redundancy between them. The possibility of functional redundancy between these ligands is supported by the observation that cerebellar morphogenesis apparently proceeds normally in Fgf3⫺/⫺, Fgf5⫺/⫺, Fgf6⫺/⫺,

and Fgf18⫺/⫺ mice (Hebert et al., 1994; Fiore et al., 1997; Floss et al., 1997; Ohbayashi et al., 2002; Alvarez et al., 2003).

A Comprehensive Study of FGF Gene Expression in the Developing Cerebellum Although a comparison between the expression patterns of different FGF ligands during cerebellar morphogenesis has not been made until now, several groups have reported FGF gene expression at early postnatal stages of cerebellar development. Our observations largely agree with these studies with a few exceptions as outlined be-

FGF GENE EXPRESSION IN THE CEREBELLUM 2067

TABLE 1. Summary of FGF Gene Expression in the Developing Cerebelluma E12.5

Fgf1 Fgf2 Fgf3 Fgf4 Fgf5 Fgf6 Fgf7 Fgf8 Fgf9 Fgf10 Fgf15 Fgf16 Fgf17 Fgf18 Fgf20 Fgf21 Fgf22 Fgf23 Fgfr1 Fgfr2 Fgfr3 Fgfr4 Fgfr5

⫹ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫺ ⫹⫹ ⫺ ⫺ ⫹⫹ ⫺ ⫹⫹ ⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫹⫹ ⫹ ⫹⫹ ⫹⫹

E16.5

P0

P7

PCL

EGL

VZ

PCL

EGL

⫹/⫺ ⫺ ⫹⫹ ⫹⫹ ⫹ ⫹⫹ ⫺ ⫺ ⫹ ⫹/⫺ ⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫹ ⫺ ⫺ ⫹

⫹ ⫺ ⫺ ⫹⫹ ⫺ ⫹⫹ ⫺ ⫹⫹ ⫹⫹ ⫹/⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ ⫹⫹ ⫹

⫹/⫺ ⫺ ⫹ ⫹ ⫺ ⫹ ⫺ VM* ⫹ ⫹/⫺ ⫺ ⫺ VM* ⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫹ ⫹⫹ ⫺ ⫹

⫹/⫺ ⫺ ⫹⫹ ⫹ ⫹⫹ ⫹⫹ ⫹ ⫺ ⫹⫹ ⫺ ⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫹⫹ ⫺ ⫺ ⫹⫹

⫹ ⫹ ⫺ ⫹⫹ ⫹ ⫹⫹ ⫺ ⫹⫹ ⫹⫹ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ ⫹⫹ ⫹⫹

P21

VZ

PCL

EGL

IGL

PC

BG

GCL

⫹/⫺ ⫺ ⫺ ⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺** ⫺ ⫹⫹

⫺ ⫺ ⫹⫹ ⫹⫹ ⫹/⫺ ⫹⫹ ⫹⫹ ⫺ ⫹ ⫹ ⫹ ⫺ ⫺ ⫹/⫺ ⫺ ⫺ ⫹ ⫺ ⫹⫹ ⫹⫹ ⫹ ⫺ ⫹⫹

⫹ ⫺ ⫹⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ⫹⫹ ⫹⫹ ⫹ ⫺ ⫺ ⫹ ⫺ ⫺ ⫹⫹ ⫺ ⫹ ⫹ ⫺ ⫹⫹ ⫹

⫹ ⫹⫹ ⫺ ⫹⫹ ⫹⫹ ⫹⫹ ⫺ ⫺ ⫹⫹ ⫹⫹ ⫹ ⫺ ⫺ ⫹⫹ ⫺ ⫺ ⫹⫹ ⫺ ⫹ ⫹ ⫹⫹ ⫺ ⫹

⫺ ⫹ ⫺ ⫹ ⫺ ⫹⫹ ⫹⫹ ⫺ ⫹⫹ ⫹ ⫺ ⫺ ⫺ ⫹⫹ ⫺ ⫹ ⫺ ⫺ ⫹/⫺ ⫹/⫺ ⫹/⫺ ⫺ ⫹⫹

⫹/⫺ ⫹/⫺ ⫺ ⫹/⫺ ⫺ ⫹ ⫺ ⫺ ⫹/⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫹⫹ ⫹⫹ ⫺ ⫹

⫹/⫺ ⫺ ⫺ ⫹ ⫺ ⫹⫹ ⫺ ⫺ ⫹/⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫺ ⫹/⫺ ⫺ ⫺ ⫹⫹ ⫺ ⫺ ⫺ ⫺

a Relative strength of gene expression is indicated as “⫺” where no expression was detected, “⫹/⫺” where expression was difficult to confirm; “⫹” where low expression was observed and “⫹⫹” where clear expression of transcripts was seen. *Expressed in the velum medullare. **Expressed in the DCN. E, embryonic day; P, postnatal day; PCL, Purkinje cell layer; EGL, external granule cell layer; VZ, ventricular zone; PC, Purkinje cell; BG, Bergmann glia.

low. Others have reported Fgf1 expression in the EGL and IGL of the mouse cerebellum between P3 and P10 (Wilcox and Unnerstall, 1991; McAndrew et al., 1998) in agreement with our findings. However, we additionally observed low expression in the PCL at P21, although it was difficult to confirm expression in PC soma. Of interest, others have detected Fgf1 expression in PCs in the chick cerebellum (Schnurch and Risau, 1991). Taken together, these observations suggest that the expression of Fgf1 in GCps and PCs is conserved between chick and mouse. Matsuda et al. (1994) reported FGF2 protein in PCs in the postnatal rat cerebellum. We detected Fgf2 expression in a significant number of GCs at P0 and P7 but expression levels were so low that it was difficult to demonstrate unequivocal expression in PCs. Despite the weak in situ hybridization signal, we also managed to detect FGF2 protein immunohistochemically in GCs and

PCs in the early postnatal mouse cerebellum, in agreement with Matsuda et al. (T. Yu, data not shown). The intriguing expression pattern of Fgf3 in PCs in the anterior cerebellum around birth was originally reported by Wilkinson et al. (1989). Because several genes and proteins, expressed in PCs during the early postnatal stages of cerebellar development, are expressed in distinct parasagittal domains (Hawkes and Gravel, 1991; Hawkes and Eisenman, 1997; Herrup and Kuemerle, 1997; Oberdick et al., 1998), we also determined the expression pattern of Fgf3 in whole-mount. We observed complex three-dimensional expression patterns in wholemount preparations at E16.5 and P0. Fgf3 gene expression was further regionalized in the anterior–posterior axis at P7. In addition to Fgf3, we also provide a more detailed expression analysis for Fgf5 and Fgf6 than previously reported (Ozawa et al., 1996; Hattori et al., 1997), while the highly

regionalized expression pattern we report for Fgf7 has not been reported previously to our knowledge. Perhaps one of our most surprising observations is the detection of Fgf8 expression in the VM and a defined region of the EGL between E16.5 and P7 (see Fig. 5). The low levels of Fgf8 transcripts in these cells might have been missed in earlier studies, as was the case for Fgf8 expression in the cardiac outflow tract (Ilagan et al., 2006). Even though Fgf17 is also expressed in the VM at E16.5, Fgf17⫺/⫺;Fgf8⫹/⫺ animals still appear to have a VM, suggesting that this mild reduction of FGF signaling is not sufficient to prevent VM development (Basson et al., 2008). Proliferation and differentiation defects in neuronal precursors by E14.5 have been reported in these mutants to explain the failure of the most anterior cerebellum to develop properly (Xu et al., 2000). An obstacle to understanding the function of FGF ligands during post-

2068 YAGUCHI ET AL.

natal cerebellar development is that many of the FGF gene knockouts are embryonic or perinatal lethal. Fgf8 is a prime example— due to the important functions that FGF8 perform during embryogenesis, Fgf8 gene knockout, hypomorphic (Meyers et al., 1998) and even mes/r1-specific knockouts (Chi et al., 2003) die before the important early postnatal stages of cerebellar development. Thus, experiments in which Fgf8 function is removed at later stages of development or specifically in the IsO are required to determine whether the maintenance of Fgf8 (and Fgf17) expression beyond E14.5 is required for cerebellar development. Similarly, Fgf8 gene function would need to be removed specifically from cells in the EGL to determine its role in the EGL. In addition, Fgf18 would need to be knocked out specifically in the cerebellum to bypass perinatal lethality associated with Fgf18 loss of function (Ohbayashi et al., 2002). Fgf9 is expressed at low levels in GCps from E16.5 and this expression is maintained in both migratory and postmigratory GCs at P7 (see Fig. 6). Nakamura et al. reported FGF9 protein in GCs, PCs, and a few glia in the P14 rat cerebellum (Nakamura et al., 1999). Of interest, although we did not detect Fgf9 in PCs at P7, we found expression in PCs in the P21 cerebellum (Supp. Fig. S1F,f), suggesting that Fgf9 expression is initiated in PCs after P7. We were intrigued by the observation that several FGF genes are coexpressed in the posterior region of the cerebellar anlage at E16.5 and P0, notably Fgf3 (only at E16.5), Fgf5, Fgf6, and Fgf8. This nested expression pattern seems to have a significant impact on FGF signaling levels in this region as downstream target genes such as Etv5 are expressed at much higher levels in this region than in the rest of the cerebellum (see Fig. 1). These observations further raised the possibility that these gene expression domains are interdependent, such that FGF3 may be required to maintain the expression of Fgf5 and Fgf8. However, we found that Fgf5 and Fgf8 expression was unaffected in Fgf3⫺/⫺ cerebella at E16.5, arguing against the possibility that FGF3 function is required to maintain the expression of

the other FGF genes (data not shown). The formal possibility that FGF5 or FGF8 regulate the expression of the other FGF genes, remain.

Etv5 and FGF Receptor Gene Expression Suggest Roles for FGF Signaling in All Cerebellar Cell Types During Development The expression of Etv4 and Etv5 during cerebellar development suggests that FGF signaling may function during several different processes, including the generation of neuronal precursors in the ventricular zone; GCp proliferation, differentiation, and migration; and the differentiation and maturation of BG and PCs. Which FGF receptors are most likely to transduce the FGF signals in different cell types during cerebellar development? Fgfr1 and Fgfr2 are expressed at high levels in PCs and BG in the PCL and GCps during the first week after birth. Fgfr3 is expressed in some glia, while Fgfr4 mRNA is found exclusively within the EGL. Because we only detect the “IIIc” isoforms of Fgfr1 and Fgfr2 by RT-PCR, our data suggest that cells in the developing cerebellum are primarily responding to ligands in the FGF1, FGF4 and FGF8, FGF9, and/or FGF15 subfamilies, and not FGF7, FGF10, or FGF22. Of interest, these cells express many of these FGF ligands themselves, suggesting that signaling may occur in both autocrine and paracrine manner. Although Fgfr4 is the only gene detected at high levels in the EGL, Fgfr4⫺/⫺ mice have a normal cerebellum (Weinstein et al., 1998). This observation suggests either that the other FGF receptors expressed in GCps (Fgfr1(IIIc) and Fgfr2(IIIc)) are able to compensate for the loss of Fgfr4 during development, or that FGF signaling in the EGL is not required for GC development. Taken together, our observations suggest that Fgfr1 and Fgfr2 primarily relay FGF signals to PCs, Fgfr1, -r2, and -r4 to GCps and Fgfr1, -r2, and -r3 to BG during early postnatal development. We also observed high Fgfr3 expression in the DCN at P0 and P7 suggesting that FGFR3, in addition to FGFR1 (Marzban and

Hawkes, 2007), may be responsible for some of the reported effects of FGF on the differentiation of these cells. Conditional gene deletion studies in which various combinations of receptors are eliminated in specific cell types are required to determine their function(s). These studies are currently being undertaken in our laboratory.

FGF Signaling May Regulate GCp Proliferation and Differentiation Our results suggest that proliferating GCps in the EGL express Fgfr1(IIIc), Fgfr2(IIIc), and Fgfr4. These cells are, therefore, expected to respond to ligands of the FGF1, FGF4, FGF8, FGF9, and FGF15 subfamilies (Itoh and Ornitz, 2008). Many of these ligands are expressed in cells in the postnatal cerebellum, including in GCps themselves. Fogarty et al. applied exogenous FGF2 intracisternally next to the posterior cerebellum between P4 and P6 (Fogarty et al., 2007) and observed a reduction in the proliferation of GCps in the caudal EGL close to the injection site. Their data are consistent with the demonstration that FGF2 can inhibit Sonic hedgehog (SHH) -mediated GCp proliferation in vitro (Wechsler-Reya and Scott, 1999; Fogarty et al., 2007). The expression of Fgf2 in GCps in the EGL at P0 indicates that FGF2 is expressed in the appropriate cells at the correct time of development to perform such a role (Fig. 2). However, no cerebellar phenotype has been reported in mice lacking FGF2 function (Dono et al., 1998; Ortega et al., 1998), and further studies will be required to determine whether the normal function of FGF2 produced by GCps is to inhibit SHH-mediated GCp proliferation. Given the possibility that many FGF ligands function redundantly to control GCp proliferation and differentiation, a more suitable approach to elucidate the function of FGF signaling during this process may be to ablate various combinations of the Fgfr1, Fgfr2, and Fgfr4 genes specifically in GCps using a conditional gene deletion approach. Uncovering the role of the Fgfr5 gene in GCp development will also require a conditional gene knockout approach because Fgfr5⫺/⫺ animals

FGF GENE EXPRESSION IN THE CEREBELLUM 2069

die perinatally (Baertschi et al., 2007; Catela et al., 2009).

FGF Signaling May Control GCp Migration and Differentiation of Bergmann Glia Many FGF ligands are expressed in BG and migrating GCps, suggesting that these cell populations use FGF signaling to communicate with each other during the migration process. Indeed, a recent publication reports a role for FGF9 in the differentiation of BG and GCp migration (Lin et al., 2009). Our expression analyses suggest that FGFR1(IIIc) and FGFR2(IIIc) are the key receptors expressed on BG, although we cannot exclude a role for FGFR3 that has been shown previously to be expressed in astrocytes and glia (Pringle et al., 2003). Although in vitro binding assays have indicated that FGF9 has little affinity for soluble FGFR1 (Chellaiah et al., 1999), cell culture analyses suggested otherwise (Lin et al., 2009). This study by Lin et al. suggested that FGFR1 and FGFR2 acted redundantly in BG to receive FGF9 signals, which is consistent with our expression analysis. Deleting these receptors specifically in glia during development will be required to clarify the effect of FGF9 signaling on BG in vivo. The high expression of Fgfr5 in BG at P0 and P7 also suggests a role for this nonsignaling receptor in these cells.

FGF Signaling May Regulate PC Differentiation and Maturation Our gene expression analyses predict FGFR1(IIIc) and FGFR2(IIIc) to be the important FGF receptors during PC development. Whether FGF signaling is strictly required for normal PC development needs to be established by experiments in which both Fgfr1 and Fgfr2 are specifically deleted from PC precursors.

FGF7 Subfamily Genes May Function as Synaptic Organizing Molecules in the Cerebellum Several FGF genes are expressed in defined subpopulations of PCs, sug-

gesting that they may be part of a molecular code required for the development of distinct functional domains in the cerebellum (Sillitoe and Joyner, 2007; Sillitoe et al., 2008). One of the most intriguing expression patterns found was for Fgf3, which is expressed at high levels in defined domains of PC precursors during the late embryonic and early postnatal period. Because Fgf3⫺/⫺ animals survive to adulthood, it was possible to examine whether cerebellar morphology was abnormal at P21. We detected no difference between Fgf3⫺/⫺ cerebella and those from control animals (Y. Yaguchi, T. Yu, and M. A. Basson, unpublished observations). This observation suggests either that FGF3 functions redundantly with other FGF ligands or that FGF3 is not required for cerebellar morphogenesis. Fgf3 belongs to the Fgf7 subfamily, which preferentially signals through FGFR1(IIIb) and FGFR2(IIIb) (Yeh et al., 2003; Zhang et al., 2006), neither of which are expressed in cells in the developing cerebellum. Fgf7 is also expressed in PCs at P7, suggesting that this gene may control a similar process to Fgf3 and even functionally compensate for the loss of Fgf3. In addition to Fgf3 and Fgf7, two other FGF7 family ligands, Fgf10 and Fgf22, are expressed in the postnatal cerebellum, albeit in GCps (Umemori et al., 2004). The expression of Fgf7 family ligand genes during cerebellar morphogenesis seems curious in the light of the observation that no mRNA encoding for their cognate receptor genes could be detected in the cerebellum by RT-PCR. Of interest, Umemori et al. detected the presence of the FGFR2(IIIb) protein in the P8 cerebellum, presumably expressed on the ingrowing axons of extracerebellar pontine and vestibular neurons (Umemori et al., 2004). They demonstrated that FGF22 acts as a presynaptic organizer molecule in granule neurons. The spatially restricted expression patterns we describe here for Fgf3 and Fgf7, lead us to hypothesize that FGF3 and FGF7 may also be involved in organizing synapse formation on these FGF-expressing PCs. Because Fgf3 and Fgf7 are expressed in distinct subpopulations of PCs at P7, they may be important functional components of the genetic circuitry that control the subdivision of the cerebellum into functional compartments. Thus, the main

role of FGF7 family ligands in the developing cerebellum may not be the regulation of morphogenesis, but rather the connection of specific cerebellar neurons with the rest of the central nervous system.

Concluding Remark In conclusion, many FGF genes are expressed during cerebellar development and our analysis suggest functions in processes as diverse as morphogenesis, differentiation, proliferation, migration, and synaptogenesis. The characterization of multiple compound, conditional gene knockout combinations is expected to yield exciting insights into the function of FGF signaling during cerebellar development.

EXPERIMENTAL PROCEDURES Mice Outbred CD-1 mice were obtained from Charles River or bred in our animal facility. Brain tissue was obtained from timed-mated CD-1 embryos or postnatal mice. Noon of the day when a vaginal plug was detected was considered E0.5. Noon of the day when newborn pups were observed was considered P0. The Fgf8lacZ mouse line was kindly provided by Gail Martin (University of California, San Francisco). Tail DNA was genotyped by PCR as described previously (Ilagan et al., 2006). All experiments were approved by the UK Home Office.

Histology, Gene Expression Analysis by In Situ Hybridization, and Immunohistochemistry Dissected brains were fixed overnight in 4% paraformaldehyde at 4°C, washed in PBS, dehydrated through an ethanol series, and embedded in paraffin wax. Sections (10 ␮m) were cut and dried overnight at 42°C. Some sections were stained with Cresyl Violet (Nissl) for histological analysis, whereas others were processed for in situ hybridization using standard methods (Wilkinson et al., 1989). After completion of the color reaction, sections were counterstained with nu-

2070 YAGUCHI ET AL.

clear fast red (Sigma), depending on the intensity of the hybridization signal and dehydrated and mounted with DPX (BDH laboratory). To identify specific cell types in which genes were expressed, some sections were stained with cell type-specific antibodies after in situ hybridization. Primary antibodies used were raised either in mouse against LHX1/5 (DSHB, 1:100) to identify PCs (Morales and Hatten, 2006), or in rabbit against S100 (Dako, 1:100) to identify BG. Bound antibodies were detected using peroxidase-conjugated secondary antibodies (Dako, 1:200) and diaminobenzidine (Sigma) as colored substrate. Images were acquired with a 5M pixel Nikon DS color camera connected to a Nikon Eclipse 80i microscope, collated using Adobe Photoshop and annotated in Adobe Illustrator. Gene expression patterns were analyzed in sagittal sections from medial (vermis) and lateral (hemisphere) cerebellar tissue. Where differences between sections taken at different mediolateral positions were observed, gene expression patterns were also analyzed in whole-mount and coronal sections to investigate the possibility that some genes were not expressed homogeneously across the mediolateral extent of the cerebellar anlage. Details of the antisense probes used in this study can be provided upon request. The identities of all plasmids and PCR templates were confirmed by sequencing before use.

Gene Expression Analysis by RT-PCR Cerebellar tissue from timed mated CD-1 embryos or newborn animals were dissected in phosphate buffered saline (PBS) and total RNA extracted using PureYield RNA Midiprep System (Promega). Before the RT reaction, potentially contaminating residual genomic DNA was eliminated by DNAseI treatment (Promega). Access RT-PCR System (Promega) was used to carry out both reverse transcription and PCR amplification of the RNA in a single reaction tube using avian myeloblastosis virus reverse transcriptase (AMV RT) and thermus flavus (Tfl) DNA polymerase. One microliter of RNA was amplified in a 25-␮l reaction containing 25 pmol of each

primer, AMV RT and Tfl DNA polymerase. Primers were the same as described by Hajihosseini and Dickson (1999). The thermal cycling profile for RT-PCR was as follows: 1 cycle of 45°C for 45 min and 94°C for 2 min; 30 cycles of 94°C for 30 sec, 55°C for 1 min, 68°C for 2 min, and a final extension of 68°C for 7 min. PCR products were analyzed by electrophoresis on 1.5% agarose gel (containing ethidium bromide) in 1⫻TAE buffer. The bands were visualized under ultraviolet light. For the molecular weight markers, 1 kb plus DNA ladder (Invitrogen) was used.

ACKNOWLEDGMENTS We thank Gail Martin, Hisashi Umemori, and Clive Dickson who generously provided in situ probes; Gail Martin for the Fgf8lacZ mouse line; Hagen Schmidt, Karun Sagar, and Samantha Martin for excellent technical assistance; Brian Emmenegger, Rob Wechsler-Reya, and Fen Wang for discussions and for communicating data before publication; and Leigh Wilson, David Chambers, and our laboratory colleagues for comments on the manuscript. The LHX1/5-specific antibody, developed by Thomas Jessell were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences (Iowa City, IA 52242).

REFERENCES Altman J, Bayer SA. 1997. Development of the cerebellar system: in relation to its evolution, structure and functions. Boca Raton, FL: CRC Press. Alvarez Y, Alonso MT, Vendrell V, Zelarayan LC, Chamero P, Theil T, Bosl MR, Kato S, Maconochie M, Riethmacher D, Schimmang T. 2003. Requirements for FGF3 and FGF10 during inner ear formation. Development 130:6329 – 6338. Baertschi S, Zhuang L, Trueb B. 2007. Mice with a targeted disruption of the Fgfrl1 gene die at birth due to alterations in the diaphragm. FEBS J 274: 6241– 6253. Basson MA, Echevarria D, Ahn CP, Sudarov A, Joyner AL, Mason IJ, Martinez S, Martin GR. 2008. Specific regions within the embryonic midbrain and cerebellum require different levels of FGF signaling during development. Development 135: 889 – 898.

Blak AA, Naserke T, Weisenhorn DM, Prakash N, Partanen J, Wurst W. 2005. Expression of Fgf receptors 1, 2, and 3 in the developing mid- and hindbrain of the mouse. Dev Dyn 233:1023–1030. Catela C, Bilbao-Cortes D, Slonimsky E, Kratsios P, Rosenthal N, Te Welscher P. 2009. Multiple congenital malformations of Wolf-Hirschhorn syndrome are recapitulated in Fgfrl1 null mice. Dis Model Mech 2:283–294. Chellaiah A, Yuan W, Chellaiah M, Ornitz DM. 1999. Mapping ligand binding domains in chimeric fibroblast growth factor receptor molecules. Multiple regions determine ligand binding specificity. J Biol Chem 274:34785–34794. Chi CL, Martinez S, Wurst W, Martin GR. 2003. The isthmic organizer signal FGF8 is required for cell survival in the prospective midbrain and cerebellum. Development 130:2633–2644. deLapeyriere O, Ollendorff V, Planche J, Ott MO, Pizette S, Coulier F, Birnbaum D. 1993. Expression of the Fgf6 gene is restricted to developing skeletal muscle in the mouse embryo. Development 118: 601– 611. Dono R, Texido G, Dussel R, Ehmke H, Zeller R. 1998. Impaired cerebral cortex development and blood pressure regulation in FGF-2-deficient mice. EMBO J 17:4213– 4225. Echevarria D, Vieira C, Gimeno L, Martinez S. 2003. Neuroepithelial secondary organizers and cell fate specification in the developing brain. Brain Res Brain Res Rev 43:179 –191. Espinosa JS, Luo L. 2008. Timing neurogenesis and differentiation: insights from quantitative clonal analyses of cerebellar granule cells. J Neurosci 28:2301–2312. Fiore F, Planche J, Gibier P, Sebille A, deLapeyriere O, Birnbaum D. 1997. Apparent normal phenotype of Fgf6-/- mice. Int J Dev Biol 41:639 – 642. Floss T, Arnold HH, Braun T. 1997. A role for FGF-6 in skeletal muscle regeneration. Genes Dev 11:2040 –2051. Fogarty MP, Emmenegger BA, Grasfeder LL, Oliver TG, Wechsler-Reya RJ. 2007. Fibroblast growth factor blocks Sonic hedgehog signaling in neuronal precursors and tumor cells. Proc Natl Acad Sci U S A 104:2973–2978. Gimeno L, Brulet P, Martinez S. 2003. Study of Fgf15 gene expression in developing mouse brain. Gene Expr Patterns 3:473– 481. Goldowitz D, Hamre K. 1998. The cells and molecules that make a cerebellum. Trends Neurosci 21:375–382. Hajihosseini MK, Dickson C. 1999. A subset of fibroblast growth factors (Fgfs) promote survival, but Fgf-8b specifically promotes astroglial differentiation of rat cortical precursor cells. Mol Cell Neurosci 14:468 – 485. Han JK, Martin GR. 1993. Embryonic expression of Fgf-6 is restricted to the skeletal muscle lineage. Dev Biol 158:549 – 554. Hashimoto M, Mikoshiba K. 2003. Mediolateral compartmentalization of the cer-

FGF GENE EXPRESSION IN THE CEREBELLUM 2071

ebellum is determined on the “birth date” of Purkinje cells. J Neurosci 23: 11342–11351. Hatten ME, Heintz N. 1995. Mechanisms of neural patterning and specification in the developing cerebellum. Annu Rev Neurosci 18:385– 408. Hattori Y, Miyake A, Mikami T, Ohta M, Itoh N. 1997. Transient expression of FGF-5 mRNA in the rat cerebellar cortex during post-natal development. Brain Res Mol Brain Res 47:262–266. Hawkes R, Eisenman LM. 1997. Stripes and zones: the origins of regionalization of the adult cerebellum. Perspect Dev Neurobiol 5:95–105. Hawkes R, Gravel C. 1991. The modular cerebellum. Prog Neurobiol 36:309 –327. Hebert JM, Rosenquist T, Gotz J, Martin GR. 1994. FGF5 as a regulator of the hair growth cycle: evidence from targeted and spontaneous mutations. Cell 78:1017–1025. Herrup K, Kuemerle B. 1997. The compartmentalization of the cerebellum. Annu Rev Neurosci 20:61–90. Hoshikawa M, Ohbayashi N, Yonamine A, Konishi M, Ozaki K, Fukui S, Itoh N. 1998. Structure and expression of a novel fibroblast growth factor, FGF-17, preferentially expressed in the embryonic brain. Biochem Biophys Res Commun 244:187–191. Ilagan R, Abu-Issa R, Brown D, Yang YP, Jiao K, Schwartz RJ, Klingensmith J, Meyers EN. 2006. Fgf8 is required for anterior heart field development. Development 133:2435–2445. Inouye M, Oda SI. 1980. Strain-specific variations in the folial pattern of the mouse cerebellum. J Comp Neurol 190: 357–362. Irving C, Mason I. 2000. Signalling by FGF8 from the isthmus patterns anterior hindbrain and establishes the anterior limit of Hox gene expression. Development 127:177–186. Itoh N, Ornitz DM. 2004. Evolution of the Fgf and Fgfr gene families. Trends Genet 20:563–569. Itoh N, Ornitz DM. 2008. Functional evolutionary history of the mouse Fgf gene family. Dev Dyn 237:18 –27. Johnson DE, Williams LT. 1993. Structural and functional diversity in the FGF receptor multigene family. Adv Cancer Res 60:1– 41. Klein OD, Lyons DB, Balooch G, Marshall GW, Basson MA, Peterka M, Boran T, Peterkova R, Martin GR. 2008. An FGF signaling loop sustains the generation of differentiated progeny from stem cells in mouse incisors. Development 135:377– 385. Komuro H, Yacubova E. 2003. Recent advances in cerebellar granule cell migration. Cell Mol Life Sci 60:1084 –1098. Larouche M, Hawkes R. 2006. From clusters to stripes: the developmental origins of adult cerebellar compartmentation. Cerebellum 5:77– 88. Lin Y, Chen L, Lin C, Luo Y, Tsai RY, Wang F. 2009. Neuron-derived FGF9 is essential for scaffold formation of Berg-

mann radial fibers and migration of granule neurons in the cerebellum. Dev Biol 329:44 –54. Liu A, Li JY, Bromleigh C, Lao Z, Niswander LA, Joyner AL. 2003. FGF17b and FGF18 have different midbrain regulatory properties from FGF8b or activated FGF receptors. Development 130:6175– 6185. Lordkipanidze T, Dunaevsky A. 2005. Purkinje cell dendrites grow in alignment with Bergmann glia. Glia 51:229 –234. Machold R, Fishell G. 2005. Math1 is expressed in temporally discrete pools of cerebellar rhombic-lip neural progenitors. Neuron 48:17–24. Mahmood R, Bresnick J, Hornbruch A, Mahony C, Morton N, Colquhoun K, Martin P, Lumsden A, Dickson C, Mason I. 1995. A role for FGF-8 in the initiation and maintenance of vertebrate limb bud outgrowth. Curr Biol 5:797– 806. Manni E, Petrosini L. 2004. A century of cerebellar somatotopy: a debated representation. Nat Rev Neurosci 5:241–249. Martinez S, Crossley PH, Cobos I, Rubenstein JL, Martin GR. 1999. FGF8 induces formation of an ectopic isthmic organizer and isthmocerebellar development via a repressive effect on Otx2 expression. Development 126:1189 –1200. Marzban H, Hawkes R. 2007. Fibroblast growth factor promotes the development of deep cerebellar nuclear neurons in dissociated mouse cerebellar cultures. Brain Res 1141:25–36. Mason I. 2007. Initiation to end point: the multiple roles of fibroblast growth factors in neural development. Nat Rev Neurosci 8:583–596. Matsuda S, Ii Y, Desaki J, Yoshimura H, Okumura N, Sakanaka M. 1994. Development of Purkinje cell bodies and processes with basic fibroblast growth factor-like immunoreactivity in the rat cerebellum. Neuroscience 59:651– 662. McAndrew PE, Frostholm A, Evans JE, Zdilar D, Goldowitz D, Chiu IM, Burghes AH, Rotter A. 1998. Novel receptor protein tyrosine phosphatase (RPTPrho) and acidic fibroblast growth factor (FGF-1) transcripts delineate a rostrocaudal boundary in the granule cell layer of the murine cerebellar cortex. J Comp Neurol 391:444 – 455. Meyers EN, Lewandoski M, Martin GR. 1998. An Fgf8 mutant allelic series generated by Cre- and Flp-mediated recombination. Nat Genet 18:136 –141. Miale IL, Sidman RL. 1961. An autoradiographic analysis of histogenesis in the mouse cerebellum. Exp Neurol 4:277–296. Miyake A, Minami M, Satoh M, Ohta M, Itoh N. 1995. Transient expression of FGF receptor-4 mRNA in the rat cerebellum during postnatal development. Brain Res Mol Brain Res 31:95–100. Morales D, Hatten ME. 2006. Molecular markers of neuronal progenitors in the embryonic cerebellar anlage. J Neurosci 26:12226 –12236. Nakamura S, Todo T, Motoi Y, Haga S, Aizawa T, Ueki A, Ikeda K. 1999. Glial expression of fibroblast growth factor-9

in rat central nervous system. Glia 28: 53– 65. Nakamura H, Watanabe Y. 2005. Isthmus organizer and regionalization of the mesencephalon and metencephalon. Int J Dev Biol 49:231–235. O’Hagan RC, Hassell JA. 1998. The PEA3 Ets transcription factor is a downstream target of the HER2/Neu receptor tyrosine kinase. Oncogene 16:301–310. Oberdick J, Baader SL, Schilling K. 1998. From zebra stripes to postal zones: deciphering patterns of gene expression in the cerebellum. Trends Neurosci 21:383– 390. Ohbayashi N, Shibayama M, Kurotaki Y, Imanishi M, Fujimori T, Itoh N, Takada S. 2002. FGF18 is required for normal cell proliferation and differentiation during osteogenesis and chondrogenesis. Genes Dev 16:870 – 879. Ortega S, Ittmann M, Tsang SH, Ehrlich M, Basilico C. 1998. Neuronal defects and delayed wound healing in mice lacking fibroblast growth factor 2. Proc Natl Acad Sci U S A 95:5672–5677. Ozawa K, Uruno T, Miyakawa K, Seo M, Imamura T. 1996. Expression of the fibroblast growth factor family and their receptor family genes during mouse brain development. Brain Res Mol Brain Res 41:279 –288. Partanen J. 2007. FGF signalling pathways in development of the midbrain and anterior hindbrain. J Neurochem 101:1185–1193. Pringle NP, Yu WP, Howell M, Colvin JS, Ornitz DM, Richardson WD. 2003. Fgfr3 expression by astrocytes and their precursors: evidence that astrocytes and oligodendrocytes originate in distinct neuroepithelial domains. Development 130: 93–102. Reifers F, Bohli H, Walsh EC, Crossley PH, Stainier DY, Brand M. 1998. Fgf8 is mutated in zebrafish acerebellar (ace) mutants and is required for maintenance of midbrain-hindbrain boundary development and somitogenesis. Development 125:2381–2395. Roehl H, Nusslein-Volhard C. 2001. Zebrafish pea3 and erm are general targets of FGF8 signaling. Curr Biol 11:503–507. Saarimaki-Vire J, Peltopuro P, Lahti L, Naserke T, Blak AA, Vogt Weisenhorn DM, Yu K, Ornitz DM, Wurst W, Partanen J. 2007. Fibroblast growth factor receptors cooperate to regulate neural progenitor properties in the developing midbrain and hindbrain. J Neurosci 27: 8581– 8592. Schnurch H, Risau W. 1991. Differentiating and mature neurons express the acidic fibroblast growth factor gene during chick neural development. Development 111:1143–1154. Shamim H, Mahmood R, Logan C, Doherty P, Lumsden A, Mason I. 1999. Sequential roles for Fgf4, En1 and Fgf8 in specification and regionalisation of the midbrain. Development 126:945–959. Sillitoe RV, Joyner AL. 2007. Morphology, molecular codes, and circuitry produce the three-dimensional complexity of the

2072 YAGUCHI ET AL.

cerebellum. Annu Rev Cell Dev Biol 23: 549 –577. Sillitoe RV, Gopal N, Joyner AL. 2008. Embryonic origins of ZebrinII parasagittal stripes and establishment of topographic Purkinje cell projections. Neuroscience [Epub ahead of print]. Sleeman M, Fraser J, McDonald M, Yuan S, White D, Grandison P, Kumble K, Watson JD, Murison JG. 2001. Identification of a new fibroblast growth factor receptor, FGFR5. Gene 271:171–182. Sotelo C. 2004. Cellular and genetic regulation of the development of the cerebellar system. Prog Neurobiol 72:295–339. Sudarov A, Joyner AL. 2007. Cerebellum morphogenesis: the foliation pattern is orchestrated by multi-cellular anchoring centers. Neural Dev 2:26. Trokovic R, Trokovic N, Hernesniemi S, Pirvola U, Vogt Weisenhorn DM, Rossant J, McMahon AP, Wurst W, Partanen J. 2003. FGFR1 is independently required in both developing mid- and hindbrain for sustained response to isthmic signals. EMBO J 22:1811–1823. Trokovic R, Jukkola T, Saarimaki J, Peltopuro P, Naserke T, Weisenhorn DM, Trokovic N, Wurst W, Partanen J. 2005. Fgfr1-dependent boundary cells between

developing mid- and hindbrain. Dev Biol 278:428 – 439. Trueb B, Taeschler S. 2006. Expression of FGFRL1, a novel fibroblast growth factor receptor, during embryonic development. Int J Mol Med 17:617– 620. Umemori H, Linhoff MW, Ornitz DM, Sanes JR. 2004. FGF22 and its close relatives are presynaptic organizing molecules in the mammalian brain. Cell 118: 257–270. Wang VY, Zoghbi HY. 2001. Genetic regulation of cerebellar development. Nat Rev Neurosci 2:484 – 491. Wechsler-Reya RJ, Scott MP. 1999. Control of neuronal precursor proliferation in the cerebellum by Sonic Hedgehog. Neuron 22:103–114. Weinstein M, Xu X, Ohyama K, Deng CX. 1998. FGFR-3 and FGFR-4 function cooperatively to direct alveogenesis in the murine lung. Development 125:3615–3623. Wilcox BJ, Unnerstall JR. 1991. Expression of acidic fibroblast growth factor mRNA in the developing and adult rat brain. Neuron 6:397– 409. Wilkinson DG, Bhatt S, McMahon AP. 1989. Expression pattern of the FGF-related proto-oncogene int-2 suggests multiple roles in fetal development. Development 105:131–136.

Wingate RJ. 2001. The rhombic lip and early cerebellar development. Curr Opin Neurobiol 11:82– 88. Xu J, Liu Z, Ornitz DM. 2000. Temporal and spatial gradients of Fgf8 and Fgf17 regulate proliferation and differentiation of midline cerebellar structures. Development 127:1833–1843. Yeh BK, Igarashi M, Eliseenkova AV, Plotnikov AN, Sher I, Ron D, Aaronson SA, Mohammadi M. 2003. Structural basis by which alternative splicing confers specificity in fibroblast growth factor receptors. Proc Natl Acad Sci U S A 100: 2266 –2271. Yuasa S. 1996. Bergmann glial development in the mouse cerebellum as revealed by tenascin expression. Anat Embryol (Berl) 194:223–234. Zervas M, Blaess S, Joyner AL. 2005. Classical embryological studies and modern genetic analysis of midbrain and cerebellum development. Curr Top Dev Biol 69: 101–138. Zhang X, Ibrahimi OA, Olsen SK, Umemori H, Mohammadi M, Ornitz DM. 2006. Receptor specificity of the fibroblast growth factor family. The complete mammalian FGF family. J Biol Chem 281:15694 –15700.