Differential regulation of fibroblast growth factor receptors in the regenerating amphibian spinal cord in vivo

PII: S 0 3 0 6 - 4 5 2 2 ( 0 2 ) 0 0 3 2 1 - 4

Neuroscience Vol. 114, No. 4, pp. 837^848, 2002 < 2002 IBRO. Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain 0306-4522 / 02 $22.00+0.00

www.neuroscience-ibro.com

DIFFERENTIAL REGULATION OF FIBROBLAST GROWTH FACTOR RECEPTORS IN THE REGENERATING AMPHIBIAN SPINAL CORD IN VIVO F. ZHANG,a J. D. W. CLARKE,b L. SANTOS-RUIZa and P. FERRETTIa a

Developmental Biology Unit, Institute of Child Health, UCL, 30 Guilford Street, London WC1N 1EH, UK b

Department of Anatomy and Developmental Biology, UCL, London WC1E 6BT, UK

Abstract1Unlike mammals, adult urodele amphibians can regenerate their spinal cord and associated ganglia, but the molecular mechanisms controlling regeneration are not fully understood. We have recently shown that expression of FGF2, a member of the ¢broblast growth factor family, is induced in the progenitor cells of the regenerating spinal cord and appears to play a role in their proliferation and possibly in their di¡erentiation. In order to investigate which receptor(s) may mediate FGF2 signaling and their role in regeneration, we have studied expression of the four ¢broblast growth factor receptors, FGFR1, FGFR2, FGFR3 and FGFR4, and of the spliced variants, sFGFR and KGFR, in the regenerating spinal cord of the adult urodele, Pleurodeles waltl, following tail amputation. We show that all FGFRs are expressed in normal and regenerating spinal cord, with the exception of the spliced variants that are expressed only in non-neural tissues of the tail. FGFR1 and 4 show the more interesting spatio-temporal patterns of expression. They are not detectable in the ependymal cells of normal cords, from which neural progenitors for regeneration are believed to originate, though they are expressed in some mature neurons. During regeneration, signi¢cant up-regulation of FGFR1 precedes that of FGFR4 in the ependymal tube from which the new cord will form. FGFR4 is highly expressed in these cells at later stages of regeneration, when neuronal di¡erentiation is becoming apparent, and like FGFR1 is also expressed in some newborn neurons. In addition to the known form of FGFR1, the antibody against this receptor reacts also with a non-phosphorylated protein that appears to be present only during regeneration, and might represent a yet undescribed variant of the receptor. Altogether this study shows that ¢broblast growth factor signaling is ¢nely modulated during tail and spinal cord regeneration, and points to FGFR1 and FGFR4 as key players in this process, suggesting that FGFR1 is primarily associated with proliferation of progenitor cells and FGFR4 with early stages of neuronal di¡erentiation. < 2002 IBRO. Published by Elsevier Science Ltd. All rights reserved. Key words: ependymal cell, ¢broblast growth factor receptor, neural progenitor, Pleurodeles waltl, regeneration, tail.

sequent proliferation to give rise to an ependymal tube from which all the di¡erent neural cell types of the regenerating spinal cord are believed to originate (Holtzer, 1956; Butler and Ward, 1967; Egar and Singer, 1972; Arsanto et al., 1992; O’Hara et al., 1992). The ependymal response is not observed in species which cannot regenerate their spinal cord, and it is likely to be crucial for the occurrence of regeneration. The molecular mechanisms underlying this early response to cord injury in urodeles, however, are still poorly understood. Fibroblast growth factor (FGF) signaling has been shown to play multiple and complex roles in nervous system development (Mason, 1996; Deng et al., 1997; Kalyani et al., 1997; Pittack et al., 1997; Qian et al., 1997; Alvarez et al., 1998). Furthermore, it has been suggested that FGFs can act as neuroprotective and neurotrophic factor following injury of the nervous system (Laird et al., 1995; Fagan et al., 1997; Huber et al., 1997; Jacques et al., 1999; Klimaschewski et al., 1999; Lee et al., 1999). We have recently shown that FGF2, a member of the large FGF family (Goldfarb, 1996; Szebenyi and Fallon, 1999; Powers et al., 2000), is upregulated in the ependymal cells at an early stage following tail amputation (Zhang et al., 2000). As regeneration

In the mammalian CNS neurogenesis is largely complete by the early postnatal stages of development and is not re-established in response to injury or disease. Notwithstanding recent exciting work suggesting the presence of neural stem cells also in mature mammalian CNS, the way of reactivating these cells in vivo to repair a CNS injury has yet to be found. In contrast, urodele amphibians maintain the extraordinary ability to regenerate their spinal cord and other parts of their nervous system as adults (Clarke and Ferretti, 1998). In the case of the spinal cord, regeneration occurs both after spinal cord transection and tail amputation. The ¢rst response to spinal cord injury is migration of ependymal cells from the central canal to seal the cut surface, and their sub-

*Corresponding author. Tel. : +44-20-78298894 (direct line), +44-2079052715 (secretary) ; fax: +44-20-78314366. E-mail address: [email protected] (P. Ferretti). Abbreviations : FGF, ¢broblast growth factor; FGFR, ¢broblast growth factor receptor ; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Ig, immunoglobulin; PBS, phosphate-bu¡ered saline ; PCR, polymerase chain reaction ; RT, reverse transcription; SDS, sodium dodecyl sulfate. 837

NSC 5754 25-9-02

Cyaan Magenta Geel Zwart

838

F. Zhang et al.

proceeds, FGF2 is gradually down-regulated in the proliferative cells of the ependymal tube, and up-regulated in a subset of newborn neurons. The coincidence between FGF2 expression pattern and that of the cell proliferation marker PCNA, together with evidence that exogenous FGF2 increases proliferation in the ependymal tube in vivo, has suggested that FGF2 does play a role in proliferation of neural progenitors in the regenerating spinal cord (Zhang et al., 2000). FGF signaling is mediated by binding to high-a⁄nity tyrosine kinase membrane receptors (FGFRs). Receptor stimulation can evoke di¡erent cellular responses, such as cell proliferation, migration and di¡erentiation, and these diverse e¡ects seems to be mediated through activation of di¡erent signaling pathways, possibly dependent on level and duration of receptor activation, and on the cell type (Klein et al., 1996; Klint and Claesson-Welsh, 1999; Szebenyi and Fallon, 1999; Boilly et al., 2000; Powers et al., 2000). Four distinct FGFR genes (FGFR1, 2, 3, and 4) have been identi¢ed in most species including the urodele amphibian, Pleurodeles waltl (Shi et al., 1992; Shi et al., 1994a; Shi et al., 1994b). Pleurodeles’FGFRs have the same basic structure as those from higher vertebrates, including three extracellular immunoglobulin-like loops (IgI, IgII, IgIII), a transmembrane domain and an intracellular tyrosine kinase domain. Several receptor isoforms that display di¡erent ligand-binding speci¢city have been identi¢ed in all of the four mammalian FGFRs (Szebenyi and Fallon, 1999; Takaishi et al., 2000). In contrast, in Pleurodeles only FGFR2 appears to be spliced and ¢ve variants generated by alternative splicing of transcripts have been identi¢ed (Shi et al., 1994b). FGFR2IIIc and FGFR2IIIb are generated as the consequence of splicing in the second half of the IgIII domain. The excision of the IgI domain from FGFR2 generates the variants sFGFR2-IIIc and sFGFR2-IIIb, known also as bek and K-sam, respectively, in other species. The KGFR variant is generated by excision of both the IgI domain and of the acid box; only the IIIb exon is transcribed in this variant. Though in recent years the involvement of FGF sig-

nalling in development of the vertebrate nervous system has been the focus of several studies, the speci¢c roles of di¡erent FGFRs in mediating FGF signaling in the nervous system, particularly in vivo, have yet to be fully elucidated. In order to gain further insight into the role of FGF signaling during spinal cord regeneration in urodeles we have studied the expression of FGFR transcripts and proteins during spinal cord regeneration in P. waltl. We have found that FGFR1, but not other FGFRs, is up-regulated in ependymal cells at an early stage of regeneration, whereas FGFR4 is greatly elevated in two-week regenerating spinal cord. This suggests that whereas activation of FGFR1 may mediate FGF2induced proliferation of neural progenitors, FGFR4 is likely to play a role in the early stages of neural di¡erentiation and that this action may be mediated by a di¡erent ligand.

EXPERIMENTAL PROCEDURES

All chemicals were supplied by Sigma (Pool, Dorset, UK) unless stated otherwise. Animals and surgery All experiments were carried out on juvenile urodele amphibian P. waltl (supplied by Biopharm Technology, France) in accordance with procedures approved under the Animals Scienti¢c Procedures Act 1986. Adult Pleurodeles were maintained in the laboratory at 19^20‡C and fed bloodworms daily. Tail amputation was performed on Pleurodeles anesthetized in 0.1% tricaine (3-aminobenzoic acid ethylester methanesulfonate salt, Sigma) by transversally cutting the distal third of the tail. Operated animals were maintained at 25‡C and tail blastemas harvested at di¡erent times after amputation. Blastemas were either processed for in situ hybridization and immunocytochemistry (see below), or rapidly frozen in liquid nitrogen and stored at 370‡C until used for RNA or protein extraction. In order to analyze FGFRs mRNA expression in di¡erent tissues, dissection was carefully performed to separate the normal spinal cord from amputated tails, and regenerating ependymal tube from twoweek regenerated tail blastemas. In order to obtain enough total RNA for the quanti¢cation assay, at least 15 of such spinal cords and 20 regenerated ependymal tubes were pooled. The tissues were stored as above.

Table 1. Oligonucleotide primers for the semi-quantitative RT-PCR assay

FGFR-1 FGFR-2 sPFR-2 PKGFR PFR-2 IIIc PFR-2 IIIb FGFR-3 FGFR-4 GAPDH a

Nucleotide sequence 5PC3P

Position from translation start site

Size (bp)

For: Rev: For: Rev: For: Rev: For: Rev: For: Rev: For: Rev: For: Rev: For: Rev:

102^123 618^598 41^60 372^351 13^31 473^455 963^983 1116^1094 62^82a 203^185 117^135 432^413 164^183 384^366 33^51a 181^162

517

AGCCAAGGTTGAAGTAGAGTCG AACCTTATATCCACCGATCCG CCACGGCAACTCTTTCTCTC GATGTTCACAATGAAGTAGCGG AGTTATCTTATGGGCTTGG CTCTTGTGGTTGTTGTCG GATCGAAGTCCTCTATGTGCG CAACCATGCAGAGTGATAGGAG AGCGCTCGAATTAATAGCTCC ATGCAGGCAGTACCGTGAG CTCCCTGGTGGAAGAGCTC CGTGAATTTTGGTGCATTTG ACACTTCCTGTTGGACCCTG AGATGGTGAAGTTGCGCAG CGGAATCAACGGATTTGG GCGTCCATGGGTAGAGTCAT

Primers were designed from unpublished partial sequence provided by Dr. D.L. Shi.

NSC 5754 25-9-02

Cyaan Magenta Geel Zwart

332 206 116 105 142 316 221 148

FGF signaling in spinal cord regeneration

Analysis of FGFRs mRNA expression by reverse transcription-polymerase chain reaction (RT-PCR) Total RNA was isolated using TRI REAGENT1 according to the manufacturer’s instruction. RNA concentrations were measured spectrophotometrically at 260 nm. To ensure that the RNA samples were not contaminated with genomic DNA, PCR was carried out in all samples with the same primers used to amplify the cDNA obtained by RT (see below). No ampli¢ed band was ever observed. RT was carried out in the presence of 1 Wg of RNA, 1U¢rst strand bu¡er (Gibco/BRL, Paisley, UK), 0.5 mM of each dNTP, 1 mM dithiothreitol (DTT), 10 U Moloney Murine Leukemia Virus Transcriptase (Gibco/BRL Paisley, UK), 1 U RNasin (Promega, Southampton, UK), 40 pmol random hexamers (Pharmacia, Herts, UK) in a total reaction volume of 30 Wl. The reaction was carried out as follows: 42‡C for 60 min, 95‡C for 10 min, and then cooled at 4‡C before storing at 320‡C. An equal amount of cDNA from di¡erent RT reactions was subjected to PCR using primers for FGFRs and glyceraldehyde3-phosphate dehydrogenase (GAPDH). The PCR reaction contains a ¢nal concentration of 1UPCR bu¡er (Gibco/BRL, 20 mM Tris^HCl pH 8.4, 50 mM KCl), 1.5 mM MgCl2 , 0.2 mM each of dNTP, 0.05% detergent W-1 (Gibco/BRL), 25 pmol of each primer, 2 U of Taq DNA polymerase (Gibco/ BRL). In order to rule out any template contamination, PCR reactions in the absence of primers were run as negative controls. To carry out a semi-quantitative analysis of the amount of each FGFR transcripts present in di¡erent tissues, equal amount of PCR products were size fractionated in 1.8% agarose gel. Di¡erences in FGFR mRNA levels were evaluated by densitometric analysis of the gel bands using Phoretix software, version 4.01 (Phoretix International, Newcastle, UK) and normalized using GAPDH expression. Each experiment was repeated three times. All primers used in this study were designed using PRIMER software (version 0.5, written by Whitehead Institute for Biomedical Research, 1991). The nucleotide sequence, the position of primers and the expected length of PCR products for FGFRs and GAPDH are shown in Table 1. Data were analyzed by Student’s t-test. In situ hybridization All tissues were ¢xed in 4% paraformaldehyde in phosphatebu¡ered saline (PBS, 100 mM, pH 7.4) overnight. After ¢xation, normal tails were decalci¢ed by treatment with 0.5 EDTA pH 7.4, at 4‡C for three to four days before para⁄n embedding. In situ hybridization to detect FGFR1 and FGFR4 transcripts was performed using digoxigenin-labeled riboprobes synthesized from pBluescript SK vector containing the Pleurodeles FGFR1

(1.1 kb) and 4 (1.5 kb) sequence spanning the extracellular domains (provided by Dr. D.L. Shi). The synthesis of the digoxigenin-labeled riboprobes were carried out according to the manufacturer’s instruction (Boehringer Mannheim, Sussex, UK). Seven to eight Wm sections were processed for in situ hybridization as previously described (Zhang et al., 2000). The FGFR1 riboprobe was hybridized at 60‡C and the FGFR4 at 64‡C overnight. After high-stringency washings and RNaseA digestion, the sections were incubated with an alkaline phosphatase-conjugated anti-digoxigenin antibody (Boehringer Mannheim, Sussex, UK). Bound anti-digoxigenin antibody was detected using the alkaline phosphatase substrate nitroblue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate (Promega, Southampton, UK). Color development was allowed to proceed in a humid chamber in the dark for 2^20 h (usually overnight). Immunohistochemistry All procedures were performed at room temperature unless stated otherwise. Tissues for immunohistochemistry were ¢xed, embedded and cut as described for in situ hybridization. All antibodies used were diluted in PBS containing 10% goat serum. Controls where either no primary antibody was added or a mouse IgG was used were always negative. The procedures were performed as previously described (Zhang et al., 2000). After rehydration, sections were microwaved in 0.01 M citric acid bu¡er, pH 6.0, and treated with 1% hydrogen peroxide for 10 min to block endogenous peroxidase. After blocking in 10% goat serum for 1 h sections were incubated sequentially with avidin D and biotin solutions (Vector, Peterborough, UK) for 15 min to block any endogenous avidin/biotin activity. After the blocking step sections were incubated overnight at 4‡C with polyclonal antibodies against either FGFR1 or FGFR4 (1:100). After washing with PBS, sections were incubated with a goat anti-rabbit biotinylated antibody (Dako, High Wycombe, UK; 1:800) for 1 h, washed with PBS and incubated with streptavidin/biotinylated horseradish peroxidase (Dako, High Wycombe, UK) for 30 min. Peroxidase activity was detected by using 3,3P-diaminobenzidine as a substrate. Sections processed for in situ hybridization and immunocytochemistry were viewed under Nomarski, bright-¢eld and epi£uorescence optics using a Zeiss Axiophot microscope. Images were either recorded photographically or captured electronically using a Kontron ProgRes 3012 digital camera. Figure montages were compiled using Adobe Photoshop. Western blotting analysis All procedures were performed at room temperature unless stated otherwise.

Fig. 1. Schematic structure of FGF receptor and the primer spanning regions. (A) Schematic structure of FGF receptor. (B) Primer spanning regions for FGFRs and their isoforms. SP: signal peptide. Ig-I, Ig-II and Ig-III: immunoglobulin-like loops. A: acid box. TM: transmembrane region. TK I and TK II: tyrosine kinase domains. C: carboxyl region.

NSC 5754 25-9-02

839

Cyaan Magenta Geel Zwart

840

F. Zhang et al.

Fig. 2. mRNA expression of FGFR1, FGFR2, FGFR3, FGFR4 and of the sFGFR2 and KGFR2 receptor variants in normal and regenerating tail by RT-PCR. (A) Size fractionation in agarose gel of PCR products of FGFRs and GAPDH from the same RT reaction from normal tail, tail blastemas at di¡erent times after tail amputation (w = week) and spinal cord (S.C.). (B) Bar chart showing the relative levels of expression of FGFR transcripts in the regenerating tail blastema expressed as a percentage of transcript levels in normal tail (100%). Bars represent the means U S.E.M. of three separate experiments. *P 6 0.05, **P 6 0.01 from control. (C) Size fractionation of PCR products of FGFR2IIIb, FGFR2IIIc and GAPDH from the same RT reaction as shown in A.

Normal tails and tail blastemas were homogenized in lysis bu¡er [50 mM Tris^HCl pH 8.0, 150 mM NaCl, 0.1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS)] containing protease inhibitors (1 WM sodium orthovanadate, 100 Wg/ml phenylmethylsulfonyl £uoride, 30 Wl/ml aprotinin). The homogenates were spun at 12 000 r.p.m. for 20 min at

NSC 5754 25-9-02

4‡C. Supernatants were collected and stored at 370‡C in small aliquots. Protein concentration in the supernatants was measured using the BCA protein assay kit (Pierce, Chester, UK). Proteins were separated by SDS gel electrophoresis using a 5% stacking gel and a 6% separating gel. 70 Wg/lane of blastema supernatant proteins, and 5 Wl/lane of rainbow-colored molec-

Cyaan Magenta Geel Zwart

FGF signaling in spinal cord regeneration

841

ular weight marker RPN 800 (Amersham, Buckinghamshire, UK) were used. The proteins were then transferred to a Hybond-C membrane (Amersham, Buckinghamshire, UK) by semi-dry electroblotting. After electroblotting ¢lters were stained with Ponceau red to assess the extent of transfer and as a further control that comparable amounts of protein were going to be assayed in extracts from di¡erent samples. The ¢lters were placed overnight in blocking solution (5% Marvel, 0.1% Tween 20 in PBS) and then incubated with polyclonal antibodies (Santa Cruz UK) either to FGFR1 (1:200) or FGFR4 (1:150) diluted in blocking solution. The speci¢city of the antibodies was checked by pre-incubating them with the appropriate antigenic peptide. Incubations with normal or pre-absorbed antibodies were carried out at 4‡C, for 1 h in the case of FGFR1 and for 2 h in the case of FGFR4. Membranes were washed three times in blocking solution, and then incubated for 30 min with a peroxidase-conjugated goat anti-rabbit antibody (Dako, High Wycombe UK) diluted 1:1500 in blocking solution. After three washings in PBS, the positive signal was detected using an enhanced chemiluminescence reagent kit (ECL, Amersham, Buckinghamshire, UK).

RESULTS

FGFR mRNA regulation in regenerating tail blastema The initial analysis of the expression of the four FGFRs and of their variants in normal and regenerating tails of Pleurodeles was performed by RT-PCR. All primers were designed against the ¢rst Ig-loop domain, as this is the most divergent regions within the receptor gene sequence (Basilico and Moscatelli, 1992) and allows speci¢c detection of the sFGFR and KGFR variants. The primers initially designed could not distinguish the IIIb and IIIc isoforms as the sequence for Pleurodeles IIIb was not available at that time. The primer spanning regions for FGFRs and variants are shown in Fig. 1. We have studied expression of FGFR1, FGFR2, and its isoforms sFGFR2 and KGFR2, FGFR3 and FGFR4 mRNA in normal tail, normal spinal cord and tail blastemas at di¡erent stages of regeneration (Fig. 2). In the normal tail, that includes the spinal cord, all four receptors and the isoforms sFGFR2 and KGFR2 are clearly present, with FGFR2, sFGFR2 and FGFR3 being the predominant forms expressed (Fig. 2A). Comparison of FGFR mRNA levels from regenerating tail blastemas and normal tails shows that only FGFR1 is up-regulated in one-week regenerates, though a small increase in the expression of KGFR2 is also observed. In contrast, the expression of FGFR2, FGFR3 and FGFR4 is down-regulated (Fig. 2A). In two-week blastemas, when neuronal di¡erentiation is becoming apparent, the expression of FGFR1 is higher than in one-week blastemas, and so is that of FGFR4, as it becomes the predominant FGFR expressed at this stage. The levels of FGFR2 and 3 are higher at two weeks than at one week, but are not signi¢cantly di¡erent from those in normal tail. With time (three and four weeks of regeneration) the levels of both FGFR1 and 4 decline, whereas the level of expression of the other FGFRs remains constant and is comparable to that of unamputated controls. Semi-quantitative analysis of mRNA levels for FGFR1, 2, 3, and 4 shows that FGFR1 is indeed signi¢cantly increased in one-, two-,

NSC 5754 25-9-02

Fig. 3. Expression of FGFR mRNA in neural and non-neural tissues of normal tail and two-week tail blastema. (A) Size fractionation in agarose gel of PCR products of FGFRs and GAPDH from the same RT-PCR reaction in (1) normal tail without spinal cord, (2) two-week blastema without ependymal tube, (3) normal spinal cord, (4) two-week ependymal tube. (B) Bar chart showing the relative levels of expression of the FGFR transcripts in twoweek blastema without ependymal tube and ependymal tube expressed as a percentage of transcript levels in normal tail without spinal cord and normal spinal cord, respectively. Bars represent the means U S.E.M. of three separate experiments. *P 6 0.05, **P 6 0.01 from control.

and three-week regenerating tail blastema and that expression of FGFR4 is signi¢cantly up-regulated at two, three and four weeks (Fig. 2B). Since alternative splicing in the third Ig-loop domain of FGF receptors can confer di¡erent ligand-binding speci¢city and can be tissue speci¢c, we analyzed also the expression of the IIIb and IIIc isoforms of FGFR2 by RT-PCR (Fig. 2C). Both isoforms are present in the normal tail, with IIIb being expressed at higher levels. The two isoforms are transcribed also in the normal spinal cord, but the IIIc isoform is expressed at higher levels than IIIb in this tissue. The levels of IIIb and IIIc expression do not show any signi¢cant change during tail regeneration, but ¢ne modulation of their expression

Cyaan Magenta Geel Zwart

842

F. Zhang et al.

Fig. 4. Localization of FGFR1 (A, B) and FGFR4 (C, D) mRNA in cross sections of normal and regenerating tail blastema using digoxigenin-labeled probes; positive cells are dark blue; pigments are brown; dorsal is up. (A) FGFR1 mRNA expression in normal tail; high levels of expression are observed in a subset of neurons in the spinal cord whereas the transcript is barely detectable in the ependymal cells lining the central canal. (B) FGFR1 mRNA expression one-week regenerating tail blastema ; FGFR1 mRNA is expressed at high levels in ependymal cells of one-week regenerating ependymal tube (arrow) ; relatively low expression is observed in some of the surrounding blastema mesenchyme. (C) FGFR4 mRNA expression in normal tail ; expression is observed only in a subset of neurons in the spinal cord and in the dorsal root ganglia (arrows). (D) FGFR4 mRNA expression in two-week regenerating tail blastema; FGFR4 mRNA is expressed at high levels in ependymal cells and new born neurons, signal is also observed in the regenerating dorsal root ganglia (arrows). Scale bars = 50 Wm.

would be di⁄cult to detect by RT-PCR. In fact, in the normal tail and tail blastema, the transcripts of IIIb may re£ect expression of both FGFR2 and of the KGFR2 variant, and the transcripts of IIIc that of FGFR2 and sFGFR2, but not of KGFR2, as this variant transcribes only the IIIb exon (Shi et al., 1994b). In contrast, since sFGFR2 and KGFR2 appear to be hardly expressed in normal spinal cord, the transcripts of IIIb and IIIc in normal cord must derive from splicing of FGFR2 only. Altogether, these data show a dynamic regulation of FGFRs during tail regeneration and point to particular key roles for FGFR1 and FGFR4 in tail regeneration. Tissue-speci¢c regulation of FGFRs in normal tail and regenerating tail blastema Since the tail blastema consists of di¡erent cell populations and includes the regenerating spinal cord, it is unclear whether the changes in FGFRs expression observed are mainly associated with regeneration of the spinal cord or of the other tissues of the tail. To address this question, we have studied FGFR expression in twoweek regenerates by RT-PCR (Fig. 3) using RNA isolated from the regenerating ependymal tube and the tail blastema from which the ependymal tube had been dissected (‘blastema minus tube’). RNA extracts from normal spinal cord and normal tail without the spinal cord were used as controls. Expression of FGFRs in the normal tail without spinal

NSC 5754 25-9-02

cord (Fig. 3A) parallels that observed in the whole tail (Fig. 2A), but di¡erences in FGFR expression pattern are observed in the dissected regenerating tissues. Expression of FGFR1 is up-regulated both in the ‘blastema minus tube’ and in the ependymal tube, though at a lower level (Fig. 3B). In contrast, expression of FGFR4 is not detectable in the ‘blastema minus tube’, but dramatic up-regulation of this transcript is observed in the regenerating ependymal tube. No signi¢cant changes in FGFR2 and FGFR3 expression are observed either in regenerating neural or non-neural tissues. In keeping with our initial RT-PCR results (Fig. 2A), no signi¢cant expression of sFGFR2 and KGFR2 is observed either in normal or regenerating spinal cord, and their expression in the ‘blastema minus tube’ is not di¡erent from that in controls (Fig. 3A). These tissue-speci¢c changes in FGFR levels of expression suggest that whereas FGFR1 is important for regeneration of neural and non-neural tissues, FGFR4 seems to play a role only in neural regeneration. Cellular localization of FGFR1 and FGFR4 mRNA To gain further insight into which speci¢c cell populations express FGFR1 and FGFR4 during tail regeneration, we have studied the distribution of both transcripts and proteins by in situ hybridization and immunocytochemistry in cross sections of normal tail and tail blastemas at di¡erent times after amputation. Both FGFR1

Cyaan Magenta Geel Zwart

FGF signaling in spinal cord regeneration

and FGFR4 transcripts are expressed in the normal tail, consistent with the RT-PCR results. Their patterns of expression, however, show a number of di¡erences. As shown in Fig. 4A, in normal tail FGFR1 transcript is present in some neurons, but is barely detectable in the ependymal cells lining the central canal. High levels of FGFR1 expression are observed also in dorsal root ganglia and muscle (not shown). Also FGFR4 is expressed in a subset of neurons and in the dorsal root ganglia in the normal tail, but is not detected in non-neural tissues with the exception of some expression in blood vessel walls (data not shown). This pattern of expression is consistent with detection of FGFR4 mRNA in normal tail minus spinal cord by RT-PCR (Fig. 3A). In one-week tail blastemas (Fig. 4B), intense expression of FGFR1 mRNA is observed in the ependymal tube. FGFR1 is also expressed in a sub-population of blastemal cells, though the staining is generally less intense than in the ependymal tube. In contrast, after one week of regeneration, blastemal cells do not express FGFR4 and it is only weakly expressed in a few ependymal cells (not shown). In two-week regenerates, however, the transcript is highly expressed in the ependymal cells and in putative newborn neurons in the spinal cord and in the dorsal root ganglia (Fig. 4D).

843

Fig. 5. Detection of FGFR1 and FGFR4 proteins in normal tail (nt) and two-week tail blastema (bl) by Western blot. Total protein extracts from normal tail and two-week regenerated tail blastema (70 Wg/lane for FGFR1, 90 Wg/lane for FGFR4) were used. FGFR1 and FGFR4 were detected by using anti-human FGFR1 and FGFR4 polyclonal antibodies. The anti-FGFR1 antibody detects a 125-kDa band both in normal tail and two-week blastemas, but the band intensity is higher in the blastema. In addition, in the regenerating tail the anti-FGFR1 antibody reacts with a 70kDa protein. The anti-FGFR4 antibody detects a band of 100kDa band both in normal tail and two-week tail blastemas, but the intensity of this band is higher in the blastema. The position of the molecular weight standards is indicated.

Expression and activity of FGFR1 and FGFR4 proteins In order to assess whether expression of FGFR proteins parallels that of the mRNA, we have studied expression of FGFR1 and FGFR4 proteins by Western blot and immunocytochemistry. As shown in Fig. 5, the anti-FGFR1 antibody used detects a main band of 125 kDa in both normal tail and in two-week tail blastema. The intensity of the band is stronger in the regenerating blastema as compared to normal tail, suggesting that the FGFR1 protein is up-regulated as is its transcript. The apparent molecular weight of Pleurodeles FGFR1 is higher than predicted from its amino acid sequence (91.5 kDa) (Shi et al., 1992). This is very likely due to glycosylation of the receptor as reported for human FGFR1 (Xu et al., 1992; Stachowiak et al., 1997). In the regenerating blastema, but not in the normal tail, the anti-FGFR1 antibody detects also a smaller protein of 70 kDa. Reactivity of the anti-FGFR1 antibody with this protein is highly speci¢c, as the 70-kDa band is not detected by any of the other antibodies used in this study. The anti-FGFR4 antibody detects a single band of approximately 100 kDa that is more intense in the two-week tail blastema than in the normal tail. The size of this band is very similar to the size predicted from the FGFR4 amino acid sequence (92 kDa) (Shi et al., 1992). The spatio-temporal distribution of FGFR1 and FGFR4 proteins parallels that of the mRNA. In normal tail FGFR1 is mainly expressed in a subset of spinal cord neurons, in the muscle (Fig. 6A) and in the dorsal root ganglia (not shown). FGFR4 expression is restricted to the spinal cord and dorsal root ganglia (Fig. 6D), and to some blood vessels (not shown). In one-week blastemas (Fig. 6B, E) FGFR1 expression in the ependymal tube is

NSC 5754 25-9-02

higher than that of FGFR4, which increases by two weeks (Fig. 6F). At this stage of regeneration both receptors are detected in the ependymal cells, in the putative newborn neurons and in the regenerating dorsal root ganglia (Fig. 6C, F). FGFR1 is also expressed in some blastemal cells and in the regenerating muscle (Fig. 6B, C). Overall these data show that the levels of expression of the FGFR1 and FGFR4 proteins and their spatio-temporal distribution are very similar to that of their transcripts.

DISCUSSION

FGF signaling in the regenerating spinal cord The present study shows that in addition to the previously demonstrated up-regulation of FGF2 (Zhang et al., 2000) there is also speci¢c regulation of some of the receptors that mediate FGF signaling in response to spinal cord injury in adult Pleurodeles, further supporting the initial proposition that these signaling pathways play an important role in the regenerative process. Analysis of FGFR expression pattern indicates that in normal Pleurodeles tail and spinal cord, the four FGFRs and two spliced isoforms of FGFR2, sFGFR2 and KGFR2, also known as the two-loop variants, have distinctive patterns of expression and that they are di¡erently regulated during regeneration of these tissues. With the exception of the two FGFR2 isoforms, we ¢nd that all receptors are expressed in the normal spinal cord. Only FGFR1 and FGFR4 expression, however, are clearly regulated dur-

Cyaan Magenta Geel Zwart

844

NSC 5754 25-9-02 F. Zhang et al.

Cyaan Magenta Geel Zwart Fig. 6. Localization of FGFR1 (A, B, C) and FGFR4 (D, E, F) proteins in cross sections of normal and regenerating tails by immunohistochemistry. Positive staining is brown; dorsal is up. (A) FGFR1 protein expression in normal tail ; note expression in a subset of neurons in the spinal cord and in muscle (insert). (B) FGFR1 expression one week after tail amputation: the protein is strongly expressed in ependymal cells lining the central canal (arrow); some positive cells are also observed in the surrounding mesenchyme. (C) FGFR1 expression in two-week tail blastema ; expression is lower than at one week in the ependymal cells, but strong signal is detected in newborn neurons, regenerating dorsal root ganglia (DRG) and muscle ¢bers (Mu); cartilage (Ca) is negative. (D) FGFR4 protein expression in normal tail; note expression in a subset of neurons in the spinal cord. (E) FGFR4 expression one week after tail amputation ; some positive staining is observed in some of the ependymal cells lining the central canal (arrow). (F) FGFR4 expression in two-week blastema ; intense expression is observed in the ependymal cells and in newborn neurons (arrows) in the regenerating spinal cord; cartilage (Ca) is negative. Scale bars = 50 Wm.

FGF signaling in spinal cord regeneration

ing spinal cord regeneration, with signi¢cant up-regulation of FGFR1 preceding that of FGFR4. The spatiotemporal regulation of FGFR1 and FGFR4 we have reported here suggest a role for these receptors in spinal cord regeneration. FGFR1 and FGFR4 play a role in spinal cord regeneration Expression of FGFR1 and FGFR4 appears to be regulated primarily at the mRNA level, as an identical tissue localization of their mRNA and protein, as detected by in situ hybridization and immunocytochemistry, is observed both in normal and regenerating tissues. During regeneration these receptors are clearly up-regulated both at the mRNA and protein level. FGFR1 and FGFR4 display some signi¢cant di¡erences in their pattern of expression, with FGFR4 being expressed almost exclusively in neural tissue. The proposition that these receptors play important role in spinal cord regeneration is consistent with studies in Xenopus embryos showing that FGFR1 and FGFR4 are important for CNS development. Overexpression of a dominant-negative FGFR1 in Xenopus resulted in inhibition of neural induction by noggin and chordin, suggesting that FGFR1 activation plays a permissive role in neural induction (Launay et al., 1996; Sasai et al., 1996). In contrast, overexpression of a dominant-negative FGFR4 led to complete disruption of anterior neural tube development, providing evidence of a direct role of FGF signaling in neural ectoderm development (Hongo et al., 1999). An increase in FGFR1 mRNA, as well as FGF2, has been reported following brain injury in mammalian systems, and it has been suggested that FGF2 and FGFR1 may play a role in the cascade of events underlying CNS wound repair, particularly in the recruitment and activation of astrocytes and possibly in glial scar formation (Logan et al., 1992; GomezPinilla et al., 1995; Smith et al., 2001). FGFR1 expression in neural progenitors, rather than in astrocytes, during regeneration of Pleurodeles spinal cord, suggests that its role in this regenerating system may be more akin to that played during neural development than to that played following traumatic injury of the mature mammalian brain. It is likely that the ependymal cells of the urodele spinal cord are closely related to the radial glial cells of the mammalian CNS. They have a typical radial morphology and express the astrocytic marker GFAP (Holder et al., 1990) and unpublished observations). Mammalian radial glia also express molecular markers typical of astrocytes (Shibata et al., 1997; Hartfuss et al., 2001) and recently these cells have been shown to be mitotically active neuronal progenitors (Miyata et al., 2001; Noctor et al., 2001; Noctor et al., 2002). Even in the adult mammal, astrocytic cells in the subventricular zone have been identi¢ed as adult neural stem cells both in vitro and in vivo (Doetsch et al., 1999). Taken together with the remarkable regenerative capacity of urodele ependymal cells in vivo, these observations point to the possibility that manipulation of cells of an astrocytic lineage may hold some hope for regenerative repair in mammalian systems.

NSC 5754 25-9-02

845

The tissue distribution of FGFR1 and FGFR4 during spinal cord regeneration, and the fact that signi¢cant up-regulation of FGFR1 precedes that of FGFR4, is suggestive of distinct roles for these receptors in the regenerative process. FGFR4 expression increases dramatically two weeks post-amputation at the onset of neuronal di¡erentiation and is detected both in the ependymal cells and in the newborn neurons. This suggests that signaling through FGFR4 is important at the early stages of neural di¡erentiation. In contrast, the increased expression of FGFR1 at earlier stages of regeneration suggests that signaling via this receptor may be associated with cell proliferation in the growing ependymal tube. This is consistent with the early up-regulation of FGF2 in the ependymal cells, and the fact that proliferation of neural precursors in higher vertebrates is stimulated by FGF2 and appears to be mediated mainly by FGFR1 both in mouse (Tropepe et al., 1999) and humans (Mackay and Ferretti, in preparation). So far our attempts to block the activity of FGFR1 in vivo using speci¢c inhibitors of its tyrosine kinase activity have been unsuccessful, most likely due to the technical problem of applying and maintaining su⁄ciently high concentrations of the drug. An alternative approach will be to use dominant-negative receptor forms of the FGFR1 and FGFR4 receptors to speci¢cally to block their activities. Unlike FGFR4, FGFR1 expression is detected also in mesenchymal tissues of the normal tail, particularly in muscle, and during regeneration in some blastemal cells and regenerating muscle. Expression of FGFR1 in normal muscle is of interest and may be relevant to the high ability of urodele muscle to re-enter the cell cycle (Brockes, 1998), as in species with limited regenerative capability FGFR1 is down-regulated with the development of muscle ¢bers (Itoh et al., 1996). Its expression in the tail blastema resembles that observed in the regenerating limb, though in the limb the majority of blastemal cells seems to express FGFR1 (Poulin et al., 1993). Therefore, it appears that FGFR1, like FGFR2, is expressed both in tail and limb blastemas. FGFR2 isoforms are di¡erently expressed in neural and non-neural tissues FGFR2 is the only receptor that has been shown to be di¡erentially spliced in Pleurodeles. We have shown here that in the adult tail both IIIb and IIIc are transcribed in neural and non-neural tissues, although IIIc transcripts appear to be more abundant than IIIb in neural tissue. In contrast, in Pleurodeles embryos, the tissue distribution of IIIb- and IIIc-containing FGFRs has been shown to be mutually exclusive (Shi et al., 1994b). IIIb transcripts are expressed mainly in the epidermis, while IIIc transcripts are activated in many neural structures during development, supporting the view that IIIc-containing FGFRs may play a role in neural development (Shi et al., 1994b). Therefore it appears that these receptors may have distinct roles in developing and adult animals and di¡erent sensitivity to various FGFs (Goldfarb, 1996; Szebenyi and Fallon, 1999; Powers et al., 2000). In con-

Cyaan Magenta Geel Zwart

846

F. Zhang et al.

trast to the ubiquitous expression of IIIb- and IIIc-containing FGFR2, the two-loop variants of this receptor, sFGFR2 and KGFR2, are selectively expressed in nonneural tissue of the tail. Although there is no signi¢cant change in their total mRNA levels in the tail blastema during regeneration, analysis of their pattern of expression, yet to be carried out, might show dynamic changes, as observed in regenerating newt limbs where these isoforms have been proposed to play distinct roles in limb regeneration (Poulin and Chiu, 1995). Di¡erent ligands may mediate FGFR1 and FGFR4 signaling in regenerating spinal cord As mentioned above, the spatio-temporal expression of FGFR1 is very similar to that of FGF2 previously described. Furthermore our previous work has indicated that FGF2 plays a role in cell proliferation in vivo during spinal cord regeneration (Zhang et al., 2000). It is therefore possible that growth of the ependymal tube is mainly mediated through FGF2^FGFR1 signaling. It has been shown that FGF2 can also bind to FGFR4 using an in vitro assay (Vainikka et al., 1992; Goldfarb, 1996), but it is not clear at present whether FGF2 plays a major role in signaling via FGFR4 in the regenerating cord. Signi¢cant up-regulation of FGF2 occurs before that of FGFR4, and in three-week regenerates FGF2 expression has dramatically decreased in the regenerating spinal cord as compared to one week. In contrast FGFR4 is still expressed at high levels in three-week regenerating spinal cords. Given the distinct temporal pattern of expression of these two genes, it is possible that the activation of FGFR4 in the regenerating spinal cord is not mediated by FGF2 but by another member(s) of the FGF family. FGFR4 can indeed bind several FGFs that have been shown to be developmentally regulated in the nervous system (Goldfarb, 1996; Kannan and Givol, 2000). It will be important to identify which FGFs, in addition to FGF2, are regulated during spinal cord regeneration to further elucidate the di¡erent roles played by FGF signaling in this system.

blastema, but not in normal tail. Our preliminary studies suggest that this protein is not phosphorylated as indicated by the lack of reactivity of an anti-phosphotyrosine antibody with this band (not shown). The size of this band is the same as the soluble form of FGFR1 described in the adult retina (Guillonneau et al., 1998). This type of soluble receptor is generated by the splicing of the extracellular domain from IgIIIa. The mRNA transcript of this splice variant encodes a protein that lacks the hydrophobic membrane-spanning domain and the intracellular domain, and may therefore be a secreted form of the receptor. Such soluble forms of FGFR1 are often observed in di¡erent tissues and cell types, and are thought to be involved in inhibiting the biological e¡ects of their ligands (Root and Shipley, 2000; Wang et al., 2000). However, since the commercial antibody we have used has been raised against a peptide at the carboxyl terminus of the FGFR1 protein, it should not recognize this soluble form. Finally, it is possible that a yet undiscovered two-loop variant of FGFR1 exists also in Pleurodeles as it does in other vertebrates. However, upon FGF binding, such a receptor would dimerize and become phosphorylated. The anti-FGFR1-positive 70-kDa protein does not seem to be phosphorylated (not shown), and this is more consistent with the hypothesis that it may be a soluble variant of the receptor. Further work will have to be carried out to establish the identity of this 70-kDa protein induced during regeneration.

CONCLUSION

The results presented here show that FGF signaling is ¢nely modulated during tail and spinal cord regeneration, and point to FGFR1 and FGFR4 as key players in the process, suggesting that FGFR1 is primarily associated with proliferation of progenitor cells and FGFR4 with early stages of neuronal di¡erentiation. In addition we have shown that expression of a 70-kDa protein speci¢cally recognized by an antibody to FGFR1 is induced in the blastema.

Is there a spliced isoform of FGFR1 in the regenerating tail blastema? The anti-human FGFR1 antibody we have used in this study, in addition to reacting with a 125-kDa band which is consistent with a glycosylated form of FGFR1, detects a strong band of 70 kDa in the tail

Acknowledgements9We wish to thank D.L. Shi and J.-C. Bocaut for sharing with us the unpublished DNA sequence of Pleurodeles FGFR2-IIIb and GAPDH, and for providing the FGFR1 and FGFR4 plasmids. This study was supported by The Wellcome Trust. L.S.R. was the recipient of a short-term FPDI fellowship from the Junta Andalucia.

REFERENCES

Alvarez, I.S., Araujo, M., Nieto, M.A., 1998. Neural induction in whole chick embryo cultures by FGF. Dev. Biol. 199, 42^54. Arsanto, J.P., Komorowski, T.E., Dupin, F., Caubit, X., Diano, M., Geraudie, J., Carlson, B.M., Thouveny, Y., 1992. Formation of the peripheral nervous system during tail regeneration in urodele amphibians: ultrastructural and immunohistochemical studies of the origin of the cells. J. Exp. Zool. 264, 273^292. Basilico, C., Moscatelli, D., 1992. The FGF family of growth factors and oncogenes. Adv. Cancer Res. 59, 115^165. Boilly, B., Vercoutter-Edouart, A.S., Hondermarck, H., Nurcombe, V., Le Bourhis, X., 2000. FGF signals for cell proliferation and migration through di¡erent pathways. Cytokine Growth Factor Rev. 11, 295^302. Brockes, J.P., 1998. Progenitor cells for regeneration : origin by reversal of di¡erentiation. In: Ferretti, P., Ge¤raudie, J. (Eds.), Cellular and Molecular Basis of Regeneration: From Invertebrates to Humans. Wiley, Chichester. Butler, E.G., Ward, M., 1967. Reconstitution of the spinal cord after ablation in adult Triturus. Dev. Biol. 15, 464^486.

NSC 5754 25-9-02

Cyaan Magenta Geel Zwart

FGF signaling in spinal cord regeneration

847

Clarke, J.D.W., Ferretti, P., 1998. CNS regeneration in lower vertebrates. In: Ferretti, P., Ge¤raudie, J. (Eds.), Cellular and Molecular Basis of Regeneration : From Invertebrates to Humans. Wiley, Chichester, pp. 255^269. Deng, C., Bedford, M., Li, C., Xu, X., Yang, X., Dunmore, J., Leder, P., 1997. Fibroblast growth factor receptor-1 (FGFR-1) is essential for normal neural tube and limb development. Dev. Biol. 185, 42^54. Doetsch, F., Caille, I., Lim, D.A., Garcia-Verdugo, J.M., Alvarez-Buylla, A., 1999. Subventricular zone astrocytes are neural stem cells in the adult mammalian brain. Cell 97, 703^716. Egar, M., Singer, M., 1972. The role of ependyma in spinal cord regrowth. Exp. Neurol. 37, 422^430. Fagan, A.M., Suhr, S.T., LucidiPhillipi, C.A., Peterson, D.A., Holtzman, D.M., Gage, F.H., 1997. Endogenous FGF-2 is important for cholinergic sprouting in the denervated hippocampus. J. Neurosci. 17, 2499^2511. Goldfarb, M., 1996. Functions of ¢broblast growth factors in vertebrate development. Cytokine Growth Factor Rev. 7, 311^325. GomezPinilla, F., Vu, L., Cotman, C.W., 1995. Regulation of astrocyte proliferation by fgf-2 and heparan sulfate in vivo. J. Neurosci. 15, 2021^ 2029. Guillonneau, X., Regnier-Ricard, F., Laplace, O., Jonet, L., Bryckaert, M., Courtois, Y., Mascarelli, F., 1998. Fibroblast growth factor (FGF) soluble receptor 1 acts as a natural inhibitor of FGF2 neurotrophic activity during retinal degeneration. Mol. Biol. Cell 9, 2785^2802. Hartfuss, E., Galli, R., Heins, N., Gotz, M., 2001. Characterization of CNS precursor subtypes and radial glia. Dev. Biol. 229, 15^30. Holder, N., Clarke, J.D., Kamalati, T., Lane, E.B., 1990. Heterogeneity in spinal radial glia demonstrated by intermediate ¢lament expression and HRP labelling. J. Neurocytol. 19, 915^928. Holtzer, S., 1956. The inductive ability of the spinal cord in urodele tail regeneration. J. Morphol. 99, 1^39. Hongo, I., Kengaku, M., Okamoto, H., 1999. FGF signaling and the anterior neural induction in Xenopus. Dev. Biol. 216, 561^581. Huber, K., Meisinger, C., Grothe, C., 1997. Expression of ¢broblast growth factor-2 in hypoglossal motoneurons is stimulated by peripheral nerve injury. J. Comp. Neurol. 382, 189^198. Itoh, N., Mima, T., Mikawa, T., 1996. Loss of ¢broblast growth factor receptors is necessary for terminal di¡erentiation of embryonic limb muscle. Development 122, 291^300. Jacques, T.S., Skepper, J.N., Navaratnam, V., 1999. Fibroblast growth factor-1 improves the survival and regeneration of rat vagal preganglionic neurones following axon injury. Neurosci. Lett. 276, 197^200. Kalyani, A., Hobson, K., Rao, M.S., 1997. Neuroepithelial stem cells from the embryonic spinal cord: Isolation, characterization, and clonal analysis. Dev. Biol. 186, 202^223. Kannan, K., Givol, D., 2000. FGF receptor mutations : dimerization syndromes, cell growth suppression, and animal models. IUBMB Life 49, 197^205. Klein, S., Bikfalvi, A., Birkenmeier, T.M., Giancotti, F.G., Rifkin, D.B., 1996. Integrin regulation by endogenous expression of 18-kDa ¢broblast growth factor-2. J. Biol. Chem. 271, 22583^22590. Klimaschewski, L., Meisinger, C., Grothe, C., 1999. Localization and regulation of basic ¢broblast growth factor (FGF-2) and FGF receptor-1 in rat superior cervical ganglion after axotomy. J. Neurobiol. 38, 499^506. Klint, P., Claesson-Welsh, L., 1999. Signal transduction by ¢broblast growth factor receptors. Front Biosci. 4, D165^177. Laird, J.M., Mason, G.S., Thomas, K.A., Hargreaves, R.J., Hill, R.G., 1995. Acidic ¢broblast growth factor stimulates motor and sensory axon regeneration after sciatic nerve crush in the rat. Neuroscience 65, 209^216. Launay, C., Fromentoux, V., Shi, D.L., Boucaut, J.C., 1996. A truncated FGF receptor blocks neural induction by endogenous Xenopus inducers. Development 122, 869^880. Lee, T.T., Green, B.A., Dietrich, W.D., Yezierski, R.P., 1999. Neuroprotective e¡ects of basic ¢broblast growth factor following spinal cord contusion injury in the rat. J. Neurotrauma 16, 347^356. Logan, A., Frautschy, S.A., Gonzalez, A.M., Baird, A., 1992. A time course for the focal elevation of synthesis of basic ¢broblast growth-factor and one of its high-a⁄nity receptors (£g) following a localized cortical brain injury. J. Neurosci. 12, 3828^3837. Mason, I., 1996. Neural induction : do ¢broblast growth factors strike a cord? Curr. Biol. 6, 672^675. Miyata, T., Kawaguchi, A., Okano, H., Ogawa, M., 2001. Asymmetric inheritance of radial glial ¢bers by cortical neurons. Neuron 31, 727^741. Noctor, S.C., Flint, A.C., Weissman, T.A., Dammerman, R.S., Kriegstein, A.R., 2001. Neurons derived from radial glial cells establish radial units in neocortex. Nature 409, 714^720. Noctor, S.C., Flint, A.C., Weissman, T.A., Wong, W.S., Clinton, B.K., Kriegstein, A.R., 2002. Dividing precursor cells of the embryonic cortical ventricular zone have morphological and molecular characteristics of radial glia. J. Neurosci. 22, 3161^3173. O’Hara, C.M., Egar, M.W., Cherno¡, E.A., 1992. Reorganization of the ependyma during axolotl spinal cord regeneration: changes in intermediate ¢lament and ¢bronectin expression. Dev. Dyn. 193, 103^115. Pittack, C., Grunwald, G.B., Reh, T.A., 1997. Fibroblast growth factors are necessary for neural retina but not pigmented epithelium di¡erentiation in chick embryos. Development 124, 805^816. Poulin, M.L., Patrie, K.M., Botelho, M.J., Tassava, R.A., Chiu, I.M., 1993. Heterogeneity in the expression of ¢broblast growth factor receptors during limb regeneration in newts (Notophthalmus viridescens). Development 119, 353^361. Poulin, M.L., Chiu, I.M., 1995. Re-programming of expression of the KGFR and bek variants of ¢broblast growth factor receptor 2 during limb regeneration in newts (Notophthalmus viridescens). Dev. Dyn. 202, 378^387. Powers, C.J., McLeskey, S.W., Wellstein, A., 2000. Fibroblast growth factors, their receptors and signaling. Endocr.-Relat. Cancer 7, 165^197. Qian, X., Davis, A.A., Goderie, S.K., Temple, S., 1997. FGF2 concentration regulates the generation of neurons and glia from multipotent cortical stem cells. Neuron 18, 81^93. Root, L.L., Shipley, G.D., 2000. Normal human ¢broblasts produce membrane-bound and soluble isoforms of FGFR-1. Mol. Cell Biol. Res. Commun. 3, 87^97. Sasai, Y., Lu, B., Piccolo, S., De Robertis, E.M., 1996. Endoderm induction by the organizer-secreted factors chordin and noggin in Xenopus animal caps. EMBO J. 15, 4547^4555. Shi, D.L., Feige, J.J., Riou, J.F., DeSimone, D.W., Boucaut, J.C., 1992. Di¡erential expression and regulation of two distinct ¢broblast growth factor receptors during early development of the urodele amphibian Pleurodeles waltl. Development 116, 261^273. Shi, D.L., Fromentoux, V., Launay, C., Umbhauer, M., Boucaut, J.C., 1994a. Isolation and developmental expression of the amphibian homolog of the ¢broblast growth factor receptor 3. J. Cell Sci. 107, 417^425. Shi, D.L., Launay, C., Fromentoux, V., Feige, J.J., Boucaut, J.C., 1994b. Expression of ¢broblast growth factor receptor-2 splice variants is developmentally and tissue-speci¢cally regulated in the amphibian embryo. Dev. Biol. 164, 173^182. Shibata, T., Yamada, K., Watanabe, M., Ikenaka, K., Wada, K., Tanaka, K., Inoue, Y., 1997. Glutamate transporter GLAST is expressed in the radial glia-astrocyte lineage of developing mouse spinal cord. J. Neurosci. 17, 9212^9219. Smith, C., Berry, M., Clarke, W.E., Logan, A., 2001. Di¡erential expression of ¢broblast growth factor-2 and ¢broblast growth factor receptor 1 in a scarring and nonscarring model of CNS injury in the rat. Eur. J. Neurosci. 13, 443^456.

NSC 5754 25-9-02

Cyaan Magenta Geel Zwart

848

F. Zhang et al.

Stachowiak, M.K., Mo¡ett, J., Maher, P., Tucholski, J., Stachowiak, E.K., 1997. Growth factor regulation of cell growth and proliferation in the nervous system. A new intracrine nuclear mechanism. Mol. Neurobiol. 15, 257^283. Szebenyi, G., Fallon, J.F., 1999. Fibroblast growth factors as multifunctional signaling factors. Int. Rev. Cytol. 185, 45^106. Takaishi, S., Sawada, M., Morita, Y., Seno, H., Fukuzawa, H., Chiba, T., 2000. Identi¢cation of a novel alternative splicing of human FGF receptor 4: soluble-form splice variant expressed in human gastrointestinal epithelial cells. Biochem. Biophys. Res. Commun. 267, 658^662. Tropepe, V., Sibilia, M., Ciruna, B.G., Rossant, J., Wagner, E.F., van der Kooy, D., 1999. Distinct neural stem cells proliferate in response to EGF and FGF in the developing mouse telencephalon. Dev. Biol. 208, 166^188. Vainikka, S., Partanen, J., Bellosta, P., Coulier, F., Birnbaum, D., Basilico, C., Jaye, M., Alitalo, K., 1992. Fibroblast growth factor receptor-4 shows novel features in genomic structure, ligand binding and signal transduction. EMBO J. 11, 4273^4280. Wang, J.F., Shen, M., Fong, G.H., Hill, D.J., 2000. A soluble ¢broblast growth factor receptor is released from HL-60 promyelocytic leukemia cells: implications for paracrine growth control. Growth Factors 17, 203^214. Xu, J., Nakahara, M., Crabb, J.W., Shi, E., Matuo, Y., Fraser, M., Kan, M., Hou, J., McKeehan, W.L., 1992. Expression and immunochemical analysis of rat and human ¢broblast growth factor receptor (£g) isoforms. J. Biol. Chem. 267, 17792^17803. Zhang, F., Clarke, J.D.W., Ferretti, P., 2000. FGF-2 up-regulation and proliferation of neural progenitors in the regenerating amphibian spinal cord. Dev. Biol. 225, 381^391. (Accepted 6 June 2002)

NSC 5754 25-9-02

Cyaan Magenta Geel Zwart