Differential Requirements for FGF3, FGF8 and FGF10 During Inner Ear Development

Developmental Biology 308 (2007) 379–391 www.elsevier.com/developmentalbiology

Differential requirements for FGF3, FGF8 and FGF10 during inner ear development

Laura Cecilia Zelarayan b, Victor Vendrell b,c, Yolanda Alvarez b, Elena Domínguez-Frutos a, ⁎ Thomas Theil d, Maria Teresa Alonso a,b, Mark Maconochie e, Thomas Schimmang a,b,

a Instituto de Biología y Genética Molecular, Universidad de Valladolid y Consejo Superior de Investigaciones Cientificas, C/Sanz y Forés s/n, E-47003 Valladolid, Spain b Center for Molecular Neurobiology, University of Hamburg, Falkenried 94, D-20251 Hamburg, Germany c Zoology Department, Zoology Building, Trinity College Dublin, Dublin 2, Ireland d University Tübingen, Anatomical Institute, Österbergstr. 3, 72074 Tübingen, Germany e School of Life Sciences, University of Sussex, Falmer, Brighton BN1 9QG, UK Received for publication 17 October 2006; revised 21 May 2007; accepted 24 May 2007 Available online 2 June 2007

Abstract

FGF signaling is required during multiple stages of inner ear development in many different vertebrates, where it is involved in induction of the otic placode, in formation and morphogenesis of the otic vesicle as well as for cellular differentiation within the sensory epithelia. In this study we have looked to define the redundant and conserved roles of FGF3, FGF8 and FGF10 during the development of the murine and avian inner ear. In the mouse, hindbrain-derived FGF10 ectopically induces FGF8 and rescues otic vesicle formation in Fgf3 and Fgf10 homozygous double mutants. Conditional inactivation of Fgf8 after induction of the placode does not interfere with otic vesicle formation and morphogenesis but affects cellular differentiation in the inner ear. In contrast, inactivation of Fgf8 during induction of the placode in a homozygous Fgf3 null background leads to a reduced size otic vesicle or the complete absence of otic tissue. This latter phenotype is more severe than the one observed in mutants carrying null mutations for both Fgf3 and Fgf10 that develop microvesicles. However, FGF3 and FGF10 are redundantly required for morphogenesis of the otic vesicle and the formation of semicircular ducts. In the chicken embryo, misexpression of Fgf3 in the hindbrain induces ectopic otic vesicles in vivo. On the other hand, Fgf3 expression in the hindbrain or pharyngeal endoderm is required for formation of the otic vesicle from the otic placode. Together these results provide important insights into how the spatial and temporal expression of various FGFs controls different steps of inner ear formation during vertebrate development. © 2007 Elsevier Inc. All rights reserved.

Keywords: Fibroblast growth factor; Otic vesicle; Otic placode; Mouse; Chicken

Introduction such as placode maintenance and its invagination to form the otic vesicle. The otic vesicle then undergoes a complex process Induction of the otic placode is controlled by signals from the of morphogenesis and differentiation leading to the formation of endoderm, mesoderm and neural ectoderm. Whereas the endo- the mature inner ear containing sensory epithelia innervated by derm and mesoderm contain the initial signals for placode the cochleovestibular ganglion. In the auditory sensory epithe- induction, the neural ectoderm (i.e. the adjacent hindbrain) is lium sound is transduced by inner and outer sensory hair cells thought to complement these signals by directing later stages which are embedded between supporting cells. Members of the fibroblast growth factor (FGF) family are key signals during multiple stages of inner ear development required for otic ⁎ Corresponding author. Instituto de Biología y Genética Molecular, Univer- sidad de Valladolid y Consejo Superior de Investigaciones Cientificas, C/Sanz y placode induction, otic vesicle formation and its subsequent Forés s/n, E-47003 Valladolid, Spain. Fax: +34983423588. morphogenesis and differentiation leading to formation of the E-mail address: [email protected] (T. Schimmang). mature inner ear. During the early stages of inner ear deve-

0012-1606/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2007.05.033 380 L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 lopment, placode induction and vesicle formation, FGF3 ap- the surface ectoderm of chicken embryos, leading to the form- pears to play a highly conserved role across different verte- ation of ectopic placodes and vesicles expressing otic markers brates, including the mouse, chicken and zebrafish (Baker and (Vendrell et al., 2000). However, these experiments were Bronner-Fraser, 2001; Noramly and Grainger, 2002, Whitfield, performed at a stage when Fgf3 expression is not observed in 2002, Brown et al., 2003, Riley and Phillips, 2003, Barald and the surface ectoderm, but rather is found in the hindbrain Kelley, 2004, Groves, 2005; Schlosser, 2006 and summarized adjacent to the area where the otic placode develops (Mahmood below). et al., 1995; Kil et al., 2005). Thus, the exact role of hindbrain In mouse, the restricted expression of FGF3 in the deve- FGF3 expression in inner ear development has not been loping hindbrain during and prior to otic placode induction addressed. (Wilkinson et al., 1988) implicated FGF3 as the inducer of the In the present study we further define the sufficiency and inner ear. This notion was reinforced by a similar expression unique or redundant requirements of several FGF family pattern in chick and by the demonstration that antisense oligo- members for different steps of vertebrate inner ear development. nucleotides and antibodies, directed against human FGF3, Thus, hindbrain-derived FGF10 rescues otic vesicle formation blocked the formation of the otic vesicle in chick embryo in FGF mouse mutants deficient for this process and is able to explants (Represa et al., 1991). Later studies did question if the induce ectopic Fgf8 expression. On the other hand, misexpres- oligonucleotides used in this study were able to specifically sion of Fgf3 leads to ectopic expression of otic markers in the target Fgf3 expression since they contained several mismatches hindbrain. Inactivation of Fgf8 during different timepoints of to the chicken Fgf3 cDNA sequence (Mahmood et al., 1995) mouse development does not affect otic vesicle formation but and indicated that further experiments were required to resolve affects differentiation of PCs in the cochlear sensory epithelium. the inductive role of Fgf3 in the chick. There appeared to be However, FGF8 is redundantly required with FGF3 for otic some considerable doubt to this inductive role in mouse as well vesicle formation during, but not after inner ear placode induc- since analysis of the two different Fgf3 mutant alleles generated tion. The combined loss of FGF3 and FGF8 during inner ear to date revealed that otic vesicles are formed during develop- induction is more detrimental for otic vesicle formation than ment (Mansour et al., 1993; Alvarez et al., 2003). Recently, it noted in the previously reported combined inactivation of FGF3 has been shown that FGF3 is redundantly required together with and FGF10 (Alvarez et al., 2003; Wright and Mansour, 2003). FGF10 to direct the expression of otic marker genes in the However, FGF3 and FGF10 are crucially involved in the developing otic placode of Fgf3−/−/Fgf10−/− double mutants, morphogenesis of the vestibular system. In the chicken, Fgf3 and such mutants fail to form the otic vesicle or only develop expression in the hindbrain is able to induce ectopic otic microvesicles (Alvarez et al., 2003; Wright and Mansour, vesicles, whereas its inactivation impairs invagination of the 2003). otic placode to form the otic vesicle. A similar phenotype is also Interestingly, FGF10, and to a much lesser degree FGF3, observed upon blocking Fgf3 expression in the pharyngeal induces the formation of ectopic otic vesicles when misex- endoderm. The central role of FGF3 acting alone or together pressed in the developing mouse hindbrain during the period of with other FGF family members in different tissues and cell inner ear induction (Alvarez et al., 2003). In addition, loss-of- types of the developing embryo to control inner ear formation function mouse mutants for FGF10 develop smaller sized otic and differentiation is discussed. vesicles and fail to develop the semicircular ducts of the vestibular system (Ohuchi et al., 2000, 2005; Pauley et al., 2003). Materials and methods Recently, mouse mutant embryos carrying a hypomorphic and a null allele for FGF8 on an FGF3 homozygous mutant back- Transgenic mice ground (Fgf3−/−/Fgf8H/−) have been shown to develop a very similar phenotype to Fgf3−/−/Fgf10−/− double mutants (Ladher The following mouse lines used in this study have been described −/− −/− et al., 2005). Unfortunately, the use of a hypomorphic allele for previously: Fgf3 and Fgf10 knockout mutants and transgenic mice expressing Fgf3 or Fgf10 under the control of the EphA4 enhancer (Alvarez FGF8 does not allow the demonstration of exactly when FGF8 et al., 2003), mutants carrying a conditional (Fgf8flox) or a null allele is redundantly required together with FGF3 for otic vesicle (Fgf8d2,3) for Fgf8 (Meyers et al., 1998), mouse lines in which cre has either formation. During inner ear induction, Fgf8 is expressed in the been targeted to the Foxg1 (BF-1; Hebert and McConnell, 2000), or Mox2 endoderm and mesoderm underlying the future otic placode and (Tallquist and Soriano, 2000) locus, transgenic mice which express lacZ in the preplacodal ectoderm where the placode will develop under the control of Fgf3 regulatory sequences (Powles et al., 2004) and the ROSA26 Cre reporter strain (Soriano, 1999). (Crossley and Martin, 1995; Ladher et al., 2005). During later development, Fgf8 expression is also observed in the inner hair Histology, RNA in situ hybridization, β-galactosidase staining and cells (IHCs) of the cochlear sensory epithelium (Pirvola et al., paint-fillings of inner ears 2002; Shim et al., 2005) where it is postulated to control the formation of neighboring supporting cells called pillar cells Preparation of histological sections stained with Toluidine Blue O, (PCs; Mueller et al., 2002; Shim et al., 2005). β-galactosidase staining, RNA whole-mount in situ hybridization and the A conserved role in inner ear induction in chick embryogen- sectioning of stained embryos have been described previously (Alvarez et al., 2003). Sections from embryos stained for β-galactosidase activity were esis was suggested by the antibody/oligonucleotide blocking counterstained with hematoxylin and eosin. The riboprobes corresponding to experiments of Represa et al. mentioned above (Represa et al., chicken Fgf3 (Aragon et al., 2005) and murine Fgf8 have been described 1991). This was further supported by overexpression of Fgf3 in (Crossley and Martin, 1995). All other riboprobes used in this study have been L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 381 referred to previously (Vendrell et al., 2000; Alvarez et al., 2003). Paint-filling of heterozygous Fgf3+/−/Fgf10+/− mice carrying the Fgf10 trans- inner ears was performed as described previously (Morsli et al., 1998). For gene. Otic vesicle formation was restored together with Dlx5 immunohistochemistry cryostat sections were prepared and processed as outlined earlier (Alvarez et al., 2003; Carnicero et al., 2004). Calretinin marker expression when the transgene was present in the full (Swant) and p75 (Chemicon) antibodies were used at a dilution of 1:200. mutant background (Fig. 1B; n=2). This demonstrates that hindbrain-derived Fgf10 expression is sufficient to rescue the Electroporation early inner ear defects observed in mutants lacking both FGF3 and FGF10 and leads to otic vesicle formation. For in ovo electroporation, embryos were incubated until stage HH8–9 (after To define the molecular mechanism(s) involved in the rescue Hamburger and Hamilton, 1992). The solution containing the different plasmids of inner ear formation in Fgf3−/−/Fgf10−/− double mutants, we encoding either a murine Fgf3 cDNA (cloned into pCS2) or siRNAs (cloned into pSuppressor; Biocarta) directed against different FGFs (2 μg/μl) plus a GFP next examined the expression of genes previously implicated in reporter plasmid pLP-EGFP-C1 (Clontech) at 0.8 μg/μl in PBS was injected into early inner ear formation, (including Pax2, Dlx5, Lnfg, Sox9, the lumen of the neural tube or the pharynx with a glass microcapillary. The Fgf8, NeuroD, Notch1, EphA4, HoxB1 and kreisler/MafB)in pSuppressor vectors contained DNA duplexes encoding siRNAs directed embryos misexpressing Fgf10 in the developing hindbrain (see against different FGFs which were formed from the following oligonucleotides: above; Alvarez et al., 2003) between E8.5 and E9.5. In parallel, for FGF3 (3 different siRNAs), 5′TCGAAAGCCAGTGCGGAGAGACTCTT- TCAAGAGAAGAGTCTCTCCGCACTGGCTTTTTTT-3′ and 5′CTAGAAA- we also examined embryos misexpressing Fgf3 under the AAAAGCCAGTGCGGAGAGACTCTTCTCTTGAAAGAGTCTCT- control of the EphA4 enhancer that usually do not form ectopic CCGCACTGGCTT-3′,5′TCGAAAGGGCTTGTTCTCTGGCAGATTCAA- otic vesicles (Alvarez et al., 2003). Interestingly, differential GAGATCTGCCAGAGAACAAGCCCTTTTTTT-3′ and 5′CTAGAAAAAA- induction of various genes was observed between FGF3 and AGGGCTTGTTCTCTGGCAGATCTCTTGAATCTGCCAGAGAA- FGF10 transgenic embryos. Ectopic expression of Pax2 (n=4), CAAGCCCTT-3′,5′TCGAAAACACGCAGGACACAGAAATTTCAAGA- GAATTTCTGTGTCCTGCGTGTTTTTTTT-3′ and 5′CTAGAAAAAAAAC- Dlx5 (n=5) and kreisler/MafB (n=4) in r3 and r5 was observed ACGCAGGACACAGAAATTCTCTTGAAATTTCTGTGTCCTGCGTGTTT- following hindbrain misexpression of Fgf3, but not in FGF10 3′, for FGF4 5′-TCGACCGATACAGTCTGCTGGAATTCAAGAGATTC- transgenic embryos (Figs. 1C–F and Supplementary Fig. 1). In CAGCAGACTGTATCGGTTTTT-3′ and 5′CTAGAAAAACCGATA- contrast, the latter embryos showed ectopic induction of Fgf8 ′ CAGTCTGCTGGAATCTCTTGAATTCCAGCAGACTGTATCGG-3 and for expression in r3 and r5, which was not detected in FGF3 FGF8, 5′-TCGAAAGCCCAGGTAACTGTTCAGTTTCAAGAGAACTGAA- CAGTTACCTGGGCTTTTTTT-3′ and 5′CTAGAAAAAAAGCCCAGG- transgenic embryos (Figs. 1G, H). The expression of Lnfg, TAACTGTTCAGTTCTCTTGAAACTGAACAGTTACCTGGGCTT-3′.Two Sox9, NeuroD, Notch1, HoxB1 and EphA4 was found to be parallel platinum electrodes (0.5 mm width and 4 mm length) with a distance unaffected by the ectopic expression of these FGFs between of 5 mm between them were positioned on both sides of the embryo. Sub- E8.5 and E9.5 (Supplementary Fig. 1 and data not shown). sequently 4 pulses of 30 V of 50 ms duration each and an interval of 1 ms were Therefore, FGF3 and FGF10 differentially induce otic markers. applied using a BTX electroporator. For electroporation of the endoderm, embryos were cultured ventral-side up on a filter paper carrier (Chapman et al., 2001). Three 17-V pulses of 50 ms each with an interval of 1 ms were applied Foxg1-Cre-mediated inactivation of FGF8 does not interfere using Tungsten electrodes. The negative electrode was inserted below the with otic vesicle formation but affects differentiation of pillar embryo parallel to its anterior–posterior axis, and a positive electrode was held cells parallel to and above the embryo. Functionality of siRNAs directed against Fgf8 was confirmed by electroporation into the pharyngeal endoderm which caused downregulation of Fgf8 expression. Ectopic expression of Fgf10 in the developing hindbrain frequently results in the formation of ectopic otic vesicles (Alvarez et al., 2003). The specific induction of Fgf8 expression Results in FGF10 transgenic mice (see above) suggested the participa- tion of this FGF family member during otic vesicle formation. Hindbrain-derived Fgf10 rescues otic vesicle formation in We thus decided to examine the consequences of a loss of Fgf8 Fgf3−/−/Fgf10−/− double mutants and induces expression of expression during inner ear development. During otic induction Fgf8 between E7 and E8.5, Fgf8 expression is observed in the preplacodal surface ectoderm, the underlying mesoderm and We and others have recently shown that double homozygous pharyngeal endoderm (Fig. 2A; Crossley and Martin, 1995; mutant embryos for FGF3 and FGF10 (Fgf3−/−/Fgf10−/−) show Ladher et al., 2005). Null mutants for Fgf8 show severe defects a severe loss or absence of otic tissue and otic marker gene during gastrulation (Meyers et al., 1998; Sun et al., 1999) expression, including Dlx5 (Fig. 1A; Alvarez et al., 2003; thereby precluding analysis of inner ear formation. Therefore, a Wright and Mansour, 2003). Since Fgf3 and Fgf10 are conditional approach using Cre-LoxP-mediated disruption of coexpressed in the developing hindbrain during otic vesicle Fgf8 was used, where Cre recombinase expression is controlled formation (Alvarez et al., 2003), we were interested to examine by the Foxg1 promoter, which has previously been shown to if ectopic misexpression of FGFs in the hindbrain would rescue efficiently inactivate floxed alleles in the developing otic vesicle the defects observed in Fgf3−/−/Fgf10−/− mutant embryos. We (Pirvola et al., 2002; Brooker et al., 2006). To assay precisely took advantage of a transgenic mouse line, which expresses when Cre activity is first observed in Foxg1-Cre mice during Fgf10 under the control of an EphA4 enhancer in rhombomeres early inner ear development, we crossed Foxg1-Cre transgenic (r) 3 and 5 of the developing hindbrain, coincident with the mice with ROSA26 reporter mice (Soriano, 1999). Cre activity formation of the otic placode and vesicle in the neighboring was first detected by X-gal staining at E8.25 as weak staining in surface ectoderm (Alvarez et al., 2003). We thus crossed double the surface ectoderm and the underlying mesoderm and pha- 382 L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391

Fig. 1. Rescue of otic vesicle formation in Fgf3−/−/Fgf10−/− mutants and ectopic induction of genes upon misexpression of Fgf3 or Fgf10 in the hindbrain. Embryos have been hybridized with the indicated probes and carry transgenes (tg) misexpressing Fgf3 or Fgf10 in rhombomeres 3 and 5 of the developing hindbrain. (A, B) Otic vesicle formation in Fgf3−/−/Fgf10−/− mutant embryos lacking (A) or carrying a transgene misexpressing Fgf10 at E9.5 (B). Note the loss of otic tissue (circumference indicated by the punctuated circle) in the Fgf3−/−/Fgf10−/− mutant (A) whereas a normally developed otic vesicle (ov) has been formed in the presence of the transgene. (C–H) Differential induction of Pax2, Dlx5 and Fgf8 expression in rhombomeres 3 and 5 upon ectopic expression of Fgf3 or Fgf10. Note the ectopic expression of Pax2 (C) and Dlx5 (E) in FGF3 transgenic embryos whereas Fgf8 (H) is induced by FGF10. ryngeal endoderm (Fig. 2B; Hebert and McConnell, 2000), mediated deletion of Fgf8 addresses the functions of FGF8 after overlapping partially with the expression of Fgf8 in this part of inner ear placode induction has been completed. Fgf8flox/flox the embryo (Fig. 2A). At this stage the placodal epithelium in mice were crossed with Fgf8d2,3/+;Foxg1Cre/+ animals to obtain mouse embryos has already started to thicken and is most likely Fgf8flox/d2,3;Foxg1Cre/+ mutants. As described previously fully committed to form the otic vesicle. Therefore, Foxg1- Fgf8flox/d2,3;Foxg1Cre/+ mutant embryos exhibited defects of the forebrain and frontonasal structures and die shortly after birth (Figs. 3A, B; Kawauchi et al., 2005; Storm et al., 2006). The inner ears of the conditional mutants were examined between E11 until birth (n=26). No obvious morphological defects using paint-fill analysis could be detected in mutant animals (Figs. 3C, D). FGF8 has previously been suggested to be involved in the formation of inner ear sensory neurons and the generation of PCs in the developing cochlea (Pirvola et al., 2002; Mueller et al., 2002; Shim et al., 2005). By breeding an Fgf3/lacZ reporter transgene (Powles et al., 2004) into the mutant background we could confirm the presence of Fgf3 expression that characterizes the ventrolateral sensory domain from which inner ear sensory neurons are derived (Figs. 3E, F) and the formation of the cochleovestibular ganglion in Fgf8flox/d2,3;Foxg1Cre/+ mutants (Fig. 3N). Using neurofilament and tubulin antibodies, we observed a normal innervation pattern of vestibular and cochlear sensory epithelia (data not Fig. 2. Expression of Fgf8, Fgf3, Foxg1 and Mox2 in tissues relevant for inner shown). Staining against calretinin antibodies confirmed the ear formation. Embryos have been hybridized with a Fgf8 riboprobe (A) or presence of vestibular hair cells (data not shown) and cochlear assayed for β-galactosidase activity (B–D) and sectioned transversally at the level where inner formation takes place. (A) At E8 Fgf8 is expressed in the IHCs, the latter of which usually express Fgf8 (Figs. 3I, J; pharyngeal endoderm (e), mesoderm (m) and surface ectoderm (se). (B) As Pirvola et al., 2002; Shim et al., 2005). On histological sections revealed by lacZ staining of a R26RlacZ/+;Foxg1Cre/+ embryo at E8.25 Foxg1 is and by using antibodies directed against p75 we were also able also expressed in these tissues. (C) As revealed by lacZ staining of a R26RlacZ/+; to detect a pair of PCs in both wild-type and Fgf8flox/d2,3; Cre/+ Mox2 embryo at E8.25, Mox2 is broadly expressed throughout the embryo. Foxg1Cre/+ mutants at E18 and at birth (Figs. 3G, H, K, L; (D) An embryo which carries a lacZ gene under the control of Fgf3 regulatory sequences shows lacZ staining in the endoderm, surface ectoderm and neural n=5). However, as revealed by staining with hematoxylin and tube (nt) at E8.5. Scale bars: in A, 100 μm for A and 110 μm for B, C; in D, eosin, the PCs of mutant animals showed less eosinophilic 100 μm for D. staining indicating a loss of microtubular structures typically L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 383

Fig. 3. Inner ear phenotype of Fgf8flox/d2,3;Foxg1Cre/+ and Fgf3−/−/Fgf8flox/d2,3;Foxg1Cre/+ mutants. (A, B) Wild-type (wt) and Fgf8flox/d2,3;Foxg1Cre/+ mutant heads at E16. Note the severe loss of frontonasal tissue in the mutants. (C, D) Paint-filled inner ears of newborn wild-type and Fgf8flox/d2,3;Foxg1Cre/+ mutant embryos. The cochlea (co) and the posterior (pc), anterior (ac) and lateral (lc) semicircular canals are indicated. To allow better visualization of the semicircular canals, the endolymphatic duct has not been paint-filled. (E, F) Transverse sections through the inner ear of control and Fgf8flox/d2,3;Foxg1Cre/+ mutant embryos at E11 carrying an Fgf3/lacZ reporter transgene. A normally developing otic vesicle with lacZ staining marking the ventrolateral sensory domain is observed in both control and mutant embryos. (G–L) Cross-sections through the organ of Corti of newborn wild-type (G, I, K) and Fgf8flox/d2,3;Foxg1Cre/+ mutants (H, J, L) stained with hematoxylin– eosin (G, H) or antibodies against calretinin (I, J) and p75 (K, L). Inner (i) and outer (o) hair cells and the circumference of pillar cells (p) are indicated. Note the lack of eosinophilc structures in the cytoplasm of the pillar cells (H) and the reduction of the pillar cell head (L, arrow) in the mutant. (M, N) Transverse sections through the inner ear of Fgf3−/− and Fgf3−/−/Fgf8flox/d2,3;Foxg1Cre/+ mutant embryos at E12 carrying an Fgf3/lacZ reporter transgene. In both mutant embryos otic vesicles show normal development and lacZ staining is observed in the ventrolateral sensory domain. The developing cochlea (co), cochleovestibular ganglion (g) and lateral semicircular canal (lc) are indicated. Orientation of tissue sections along the dorsal (d)–lateral (l) axis is indicated in F. Scale bars: in E, 200 μm for E, F, M and N; in I, 30 μm for I–L and 10 μm for G, H. found in PCs (Figs. 3G, H). Moreover p75 staining showed that Mox2-Cre-mediated inactivation of FGF8 the size of the PC heads was reduced in the mutants (Figs. 3K, L). Therefore, Foxg1-mediated deletion of FGF8 after inner ear The absence of early inner ear defects due to conditional induction does not affect its formation but affects later cellular inactivation of FGF8 prompted us to inactivate Fgf8 expression differentiation in the cochlea. more extensively and earlier in development. The Mox2Cre

Fig. 4. Inner ear phenotype of Fgf8flox/d2,3;Mox2Cre/+ and Fgf3−/−/Fgf8flox/d2,3;Mox2Cre/+ mutants. (A–C) Wild-type (wt), Fgf8flox/d2,3;Mox2Cre/+ and Fgf3−/−/ Fgf8flox/d2,3;Mox2Cre/+ mutant embryos at E9.5 hybridized with a Pax2 riboprobe. (A) Pax2 expression is detected in the otic vesicle (ov) and midbrain–hindbrain (mhb) boundary of wild-type embryos. (B) Pax2 expression is also observed in the otic vesicle of the Fgf8flox/d2,3;Mox2Cre/+ mutant embryo. (C) In the Fgf3−/−/ Fgf8flox/d2,3; Mox2Cre/+ mutant no Pax2 expression is observed in the area where the otic vesicle usually forms. Note also the loss of the first branchial arch (ba, asterisks) and the absence of Pax2 staining in the area of the midbrain–hindbrain boundary (arrows) in the mutant embryos. (D, F) Sections through wild-type (wt) embryos reveal the presence of a normal otic vesicle expressing Pax2 or Dlx5 in their corresponding domains at E9.5 and E10.5, respectively. (E, G) Fgf3−/−/ Fgf8flox/d2,3;Mox2Cre/+ embryos only develop microvesicles lacking Pax2 expression or showing Dlx5 expression in an abnormal ventral domain. The circumference of the microvesicles and the neural tube (nt) is indicated. (H–J) Histological sections through the inner ear of wild-type (wt) and Fgf3−/−/Fgf8flox/d2,3;Mox2Cre/+ mutant embryos at E12. (H) In the wild-type embryo the developing cochlea (co) and the posterior (pc) and lateral semicircular canals (lc) are indicated. Sections of the mutant embryo reveal the presence of an undifferentiated otic vesicle (I) or a complete lack of otic tissue (J). Orientation of tissue sections along the dorsal (d)–lateral (l) axis is indicated in F. Abbreviation: neural tube (nt). Scale bars: in D, 120 μm for D, E and G and 150 μm for F; in H, 200 μm for H–J. 384 L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 mouse line expresses Cre recombinase from the Mox2 locus and To examine the redundant requirements for FGF3 and FGF8 drives expression throughout the epiblast (Tallquist and during and after otic placode induction, we next analyzed the Soriano, 2000). Thus X-gal staining of embryos derived from effects of conditional inactivation of Fgf8 induced by Cre mating ROSA26 reporter mice with the Mox2Cre mice leads to driven either by the Foxg1 or Mox2 locus in an Fgf3 null widespread β-galactosidase activity in the embryo. We wanted mutant background at the otic vesicle stage. We therefore to ensure that tissues relevant to early inner ear development crossed either Fgf3+/−/Fgf8flox/d2,3;Foxg1Cre/+ or Fgf3+/−/ would be within the Mox2Cre expression domain. Compared to Fgf8flox/d2,3;Mox2Cre/+ animals with Fgf3−/−/Fgf8flox/flox mice. R26RlacZ/+;Foxg1Cre/+ embryos, R26RlacZ/+;Mox2Cre/+ embryos Histological analysis of Fgf3−/−/Fgf8flox/d2,3; Foxg1Cre/+ mu- showed strong and widespread expression of β-galactosidase in tants at E11 and E12 (n=5) revealed that otic vesicle formation the preplacodal ectoderm, the underlying mesoderm and was unaffected (Figs. 3M, N). Similar to Fgf3−/− control em- pharyngeal endoderm during otic induction between E7 and bryos (Fig. 3M) which show normal inner ear development E8.5 (Fig. 2C and data not shown), and thus in tissues that have (Alvarez et al., 2003), formation of the cochleovestibular been shown to be critical for early stages in inner ear ganglion, the cochlea and the vestibule including the ventrolate- development and at the appropriate developmental time points. ral sensory domain characterized by Fgf3 expression was found Next, to obtain Fgf8flox/d2,3;CreMox2Cre/+ mutants, Fgf8flox/flox to be unaltered in Fgf3−/−/Fgf8flox/d2,3;Foxg1Cre/+ mutants (Fig. mice were crossed with Fgf8d2,3/+;Mox2Cre/+ animals. Similar 3N). The presence of hair cells and pillar cells was also to Fgf8flox/d2,3;Foxg1Cre/+ mutants, Fgf8flox/d2,3;Mox2Cre/+ confirmed in Fgf3−/−/Fgf8flox/d2,3;Foxg1Cre/+ mutants at E17 embryos showed defective forebrain and craniofacial develop- (Supplementary Fig. 2). ment and died after birth. As judged by external appearance, the However in contrast, Fgf3−/−/Fgf8flox/d2,3; Mox2Cre/+ condi- phenotype of the latter mutants was more variable compared to tional mutants showed defects during otic vesicle formation. At Fgf8flox/d2,3;Foxg1Cre/+ mutants, most likely reflecting mosaic E9.5 and E10.5 these mutants showed a severe loss of otic tissue expression of Cre from the Mox2 locus. Whole-mount in situ (n=6). Whole-mount in situ hybridization revealed the absence hybridization of embryos with Pax2 revealed the loss of the of Pax2 expression in the area where the otic vesicle is usually midbrain–hindbrain boundary, a previously described defect formed (Fig. 4C). Sections through this region revealed the attributed to the loss of Fgf8 expression (Chi et al., 2003)in presence of microvesicles that lacked Pax2 expression (Figs. Fgf8flox/d2,3;Mox2Cre/+ mutant embryos at E10 (Figs. 4A, B). 4D, E). Dlx5 expression was detected, although not in its nor- This illustrates the efficacy of the conditional deletion of Fgf8 mal dorsal domain but rather concentrated in the ventral part of by expression of Cre from the Mox2 locus. However, the otic the microvesicle (Figs. 4F, G). To examine the developmental vesicle in Fgf8flox/d2,3;Mox2Cre/+ mutants was normally formed capacity of these microvesicles to undergo further develop- and showed no obvious morphological or histological defects ment and differentiation, we examined Fgf3−/−/Fgf8flox/d2,3; between E9 and E12 (n=7). Therefore, Mox2-Cre-mediated Mox2Cre/+ mutants at E12 (n=4) and found a variable inner ear inactivation of FGF8 does not interfere with otic vesicle phenotype. Some microvesicles at E10 had increased in size to formation and its initial morphogenesis. form an elongated otic vesicle (Fig. 4I). However, these vesicles lacked any signs of normal morphogenesis, such as the Requirements for FGF3 and FGF8 during inner ear formation formation of the endolymphatic duct or the cochleovestibular ganglion (n=2/8). In the majority of cases (n=6/8) we were The lack of an inner ear phenotype during early inner ear unable to detect any otic tissue upon examination of serial formation in mouse mutants for FGF8 suggests the involvement sections at E12 through the cranial region (Fig. 4J). Therefore, of other FGF family members acting in a redundant manner. In the combined loss of FGF3 and FGF8 at early stages of inner ear the zebrafish, several studies have demonstrated a joint induction results in a severe reduction of otic tissue which requirement for FGF8 and FGF3 for inner ear induction eventually disappears or is only able to form highly aberrant otic (Phillips et al., 2001; Leger and Brand, 2002; Maroon et al., vesicles that fail to undergo proper morphogenesis. 2002). Similarly, mouse embryos carrying a null allele for Fgf3 and a compound hypomorphic and null allele for Fgf8 (Fgf3−/−/ Loss of FGF3 and FGF10 affects the formation of semicircular Fgf8H/−) fail to form otic vesicles and show reduced or absent canals expression of otic placode markers (Ladher et al., 2005). Fgf3 expression has been widely cited for its evolutionary conserved The above phenotype is reminiscent of the failure to form presence in the posterior hindbrain, but in contrast, Fgf8 is otic vesicles or microvesicle formation observed in Fgf3−/−/ absent in the embryonic mouse hindbrain (Wilkinson et al., Fgf10−/− mutants (Alvarez et al., 2003; Wright and Mansour, 1988; Crossley and Martin, 1995). Relevant to early inner ear 2003). To further compare the redundant requirements during induction, Fgf8 expression is found in the surface ectoderm and inner ear development between FGF3 and FGF8, or FGF3 and the underlying pharyngeal endoderm during otic placode and FGF10, we examined the fate of the microvesicles observed in vesicle formation (Mahmood et al., 1996; McKay et al., 1996). Fgf3−/−/Fgf10−/− mutants after E10. After the isolation of 96 Similar to Fgf8, Fgf3 expression is also found in these same embryos derived from matings of mice that were heterozygous sites, as demonstrated by staining embryos containing the for null alleles of both FGF3 and FGF10, we were able to Fgf3lacZ reporter gene (Powles et al., 2004) for β-galactosidase recover only two Fgf3−/−/Fgf10−/− embryos at E11 and E12.5. activity at E8.5 (Fig. 2D). Therefore the number of embryos after E10 is 2.1% rather than L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 385 the expected 6.25%. These embryos are likely to correspond to mally detected in the developing hindbrain (Mahmood et al., the subset of Fgf3−/−/Fgf10−/− mutants which form less affected 1995; Kil et al., 2005). Successful transfection was confirmed otic vesicles (Alvarez et al., 2003). Histological analysis of an by strong GFP expression in the electroporated side of the Fgf3−/−/Fgf10−/− embryo at E12 revealed a smaller but overtly neural tube as early as 2 h after electroporation (Fig. 6A), and normal developing cochlea and cochleovestibular ganglion ectopic expression of Fgf3 was demonstrated by RNA whole- (Figs. 5A, B). In contrast, the formation of the semicircular mount in situ hybridization (data not shown). Analysis of em- canals was found to be defective in the vestibule. Instead of the bryos 36 h after electroporation (n=35) revealed the generation anterior, lateral and posterior canal detected in wild-type of ectopic structures close to the normal otic vesicle in 71% of controls (Fig. 5C and data not shown), only a single, dorsally electroporated embryos. Histological analysis of the ectopic protruding canal was identified (Fig. 5D). To further character- structures revealed a very organized vesicular structure with an ize this phenotype after E12, we generated compound mutants epithelium similar to the otic vesicle (data not shown), and their consisting of homozygous null alleles at Fgf3 and a single otic character was confirmed by in situ hybridization with a Fgf10 null mutant allele. These compound mutants (Fgf3−/−/ RNA probe against Lmx1 (Giraldez, 1998), which is expressed Fgf10+/−) survived until adulthood and showed a variable inner dorso-laterally within the otic vesicle (Fig. 6B). Together, these ear phenotype. Histological analysis between E10 and E12.5 data further support the role of FGF3 as a conserved hindbrain- revealed that about half of the mutant embryos (n=8/15) derived signal sufficient to direct the formation of the otic showed no gross abnormalities during inner ear development. vesicle in chicken embryos in vivo. The remaining mutants examined formed otic vesicles which were smaller and positioned in a more ventral position com- Electroporation of siRNA directed against FGF3 blocks pared to wild-type controls (Figs. 5E, F). As adults, several of formation of the chicken otocyst the Fgf3−/−/Fgf10+/− mutants (n=3/7) showed abnormal behav- ior consistent with defects in the vestibular system, including We next determined whether FGF3 is not only sufficient but head tilting, circling and hyperactivity. To further characterize also required for otic vesicle formation in chicken embryos. In these defects we performed paint-filling of adult inner ears. In order to do so, loss-of-function experiments using siRNA were Fgf3−/−/Fgf10+/− mutants we were only able to detect a single carried out. canal (Fig. 5H) and not the three canals observed in wild-type Three different sequences encoding siRNAs directed against specimens (Fig. 5G). This canal appeared to be comprised of the the chicken Fgf3 cDNA were cloned into the pSuppressor plas- anterior semicircular canal connected with the posterior semi- mid (see Materials and methods). Plasmids encoding Fgf3- circular canal that had fused with the common cross, with the siRNAs and a GFP reporter plasmid were coelectroporated into lateral semicircular canal completely absent. In summary, these the neural tube of chicken embryos at stage HH8. Six hours data reveal that FGF3 and FGF10 are redundantly required for after electroporation, in situ hybridization revealed a reduction the formation of semicircular canals. of Fgf3 mRNA on the electroporated side of the neural tube at the level of rhombomeres 4 and 5 (Figs. 6C, D; n=5/6). Fgf3 Ectopic expression of FGF3 in the neural tube induces the expression in the neural tube of embryos electroporated with formation of ectopic otic vesicles in chicken embryos plasmids producing siRNAs directed against Fgf8 (see Mate- rials and methods) was unchanged (Fig. 6E). Next, embryos Recent studies in the chicken embryo have shown that electroporated with Fgf3 siRNA plasmids were incubated until endoderm-derived Fgf8 initiates otic induction (Ladher et al., the otic vesicle stage and analyzed histologically. 33% of the 2005). The role of FGF3 during chicken inner ear formation is embryos (n=25/76) presented with either an open otocyst, more controversial (Represa et al., 1991; Mahmood et al., where invagination of the placode is incomplete (Fig. 6F) or 1995). However, we have previously demonstrated the capacity formed an otic placode only just initiating invagination on the of FGF3 to induce otic vesicles when misexpressed in the sur- electroporated side (Fig. 6G). Otic vesicles formed normally on face ectoderm of chicken embryos via viral infection (Vendrell the non-electroporated side and in embryos electroporated with et al., 2000). Such evidence, together with the highly conserved the GFP reporter plasmid alone or plasmids encoding siRNAs expression pattern of Fgf3 in several species, lends support to directed against Fgf8 or Fgf4 (Figs. 6F, G and data not shown). the postulated role of FGF3 as a key hindbrain-derived signal Therefore, knockdown of Fgf3 expression in the neural tube of that directs early inner ear formation (Baker and Bronner- chicken embryos using siRNA demonstrates the requirement of Fraser, 2001; Noramly and Grainger, 2002; Alvarez et al., 2003; this factor for otic vesicle formation. Brown et al., 2003; Riley and Phillips, 2003). Since our studies in the mouse revealed exquisite timing requirements for early Blocking FGF3 in the pharyngeal endoderm also inhibits otic inner ear formation controlled by FGF signaling, we looked to vesicle formation further define the role of FGF3 in the avian embryo by mani- pulating its expression in the developing neural tube. To over- Prominent Fgf3 expression is also observed in pharyngeal express Fgf3, we electroporated a cDNA for Fgf3 together with endoderm next to the developing inner ear from HH9 (Mah- a GFP reporter plasmid (see Materials and methods) into the mood et al., 1995; Karabagli et al., 2002). To determine the neural tube of chicken embryos at stages HH8 (after Hamburger function of FGF3 in this domain, we blocked Fgf3 expression and Hamilton, 1992), when prominent Fgf3 expression is nor- using siRNA. Explant cultures of chicken embryos at stage 386 L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391

Fig. 5. Inner ear phenotypes of Fgf3−/−/Fgf10−/− and Fgf3−/−/Fgf10+/− mutants. (A–D) Transversal histological sections through the inner ears of a wild-type (wt) and an Fgf3−/−/Fgf10−/− mutant embryo at E12.5. (A, B) In both wild-type and mutant inner ears the cochlea (co) and the cochleovestibular ganglion (g) are present. (C, D) In the wild-type inner ear (C) the posterior (pc) and lateral (lc) semicircular canals are observed, whereas in the mutant (D) only a single rudimentary canal (c) is formed. (E, F) Histological sections through the otic vesicle of a wild-type (E) and an Fgf3−/−/Fgf10+/− mutant embryo (F) at E10. The otic vesicle in the mutant is smaller, in a more ventralized position and lacks the endolymphatic duct (arrow in E) compared to the wild-type. (G, H) Paint-filled inner ears of a wild-type (G) and an Fgf3−/−/Fgf10+/− mutant (H) adult inner ear. In the wild-type inner ear the cochlea (co), common cross (cc), lateral (lc), posterior (pc) and anterior (ac) semicircular canal are labeled. In the mutant adult inner ear only the cochlea and a single semicircular canal (c) can be observed. To allow better visualization of the semicircular canals, the endolymphatic duct has not been paint-filled. Scale bar: 200 μm for A–F.

HH8–9 were coelectroporated with plasmids encoding siRNAs the vectors to the pharyngeal endoderm was confirmed by directed against Fgf3 and a GFP reporter vector into the monitoring GFP expression (Fig. 7A). Specific downregulation endoderm (see Materials and methods). Successful targeting of of Fgf3 expression in the pharyngeal endoderm on the electroporated side of the explants was observed, but not in the neighboring neural tube, nor contralateral unelectroporated endoderm (Fig. 7B, n=8/13). In contrast, Fgf3 expression was unaffected by electroporation of plasmids encoding siRNAs directed against Fgf4 (Fig. 7C, n=5/5). We next electroporated explants with Fgf3 siRNA plasmids at HH8–9, which were subsequently cultured for 24 h and then examined for the expression of a variety of otic markers, including Pax2, Lmx1 and EphA4. As revealed by the downregulation or absence of staining for these markers, formation of the otic vesicle had been blocked in about half of the embryos on the electroporated side (Figs. 7D–F, n=4/9). Therefore, Fgf3 expression in the pharyngeal endoderm also appears to be required for formation of the otic vesicle.

Fig. 6. Effects of misexpression and inhibition of Fgf3 expression in the neural tube on chicken otic vesicle formation. The embryos shown were electroporated Discussion into the left side of the neural tube. (A) Expression of GFP in the neural tube 2 h after electroporation at stage HH8 with a vector containing Gfp. (B) Staining of Hindbrain-derived FGFs direct otic vesicle formation and a chicken embryo (HH18) with a Lmx1 riboprobe after electroporation at HH8 differentially induce expression of genes involved in inner ear with a plasmid containing Fgf3 reveals the presence of an ectopic otic vesicle development (arrow) next to the endogenous otic vesicle (ov). (C–E) Embryos at stage HH10 which have been electroporated at HH8 (D, E) and stained with a riboprobe against chicken Fgf3. Fgf3 expression is downregulated on the left side of the Loss of FGF3 and FGF10 during murine inner ear deve- embryo upon electroporation with a vector encoding siRNAs directed against lopment leads to a loss of otic tissue, frequently resulting in the Fgf3 (D), but not in unelectroporated controls (C; ctrl) or embryos electro- formation of microvesicles (Alvarez et al., 2003; Wright and porated with a vector encoding siRNAs directed against Fgf8 (E). (F, G) Trans- Mansour, 2003). Both FGFs are expressed in the hindbrain and versal sections through embryos 24 h after electroporation with a vector encoding siRNAs directed against Fgf3 at HH8. Note that on the electroporated show some overlapping expression during otic placode and otic side otic vesicle formation has not been completed (F) or the invagination of the vesicle formation. In the mouse, FGF10 has been shown to act otic placode has only been initiated (G). Scale bar: 150 μm in F and G. as a potent inducer of ectopic vesicles expressing otic markers L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 387

Fig. 7. Inhibition of Fgf3 expression in the pharyngeal endoderm blocks chicken otic development. Embryos were electroporated into endoderm on the right side of the embryo. (A) Transverse section through an embryo at the level of the otic placode 12 h after electroporation at stage HH8 with a vector containing Gfp. Expression of GFP is observed in the endoderm (e) and surface ectoderm (se) which are indicated by stippled lines. (B–D) Embryos were electroporated with the indicated vectors encoding siRNAs directed against different FGFs at stage HH8–9, incubated until the desired stage and stained with the indicated riboprobes. (B, C) Embryos at HH10 which have been electroporated with a vector encoding siRNAs against Fgf3 (B) or Fgf4 (C) and hybridized with a ripobrobe for Fgf3. Fgf3 expression is unaffected in the endoderm (arrows) on both sides of the embryo in the presence of siRNAs against Fgf4 (C), whereas siRNAs against Fgf3 downregulate Fgf3 expression in the pharyngeal endoderm on the electroporated side (B). (D–F) Embryos 24 h after electroporation with a vector encoding siRNAs directed against Fgf3 and stained with the indicated riboprobes which identify the developing otic placode and vesicle. Note the absence of Lmx1 expression in the otic cup (punctuated) and the reduced amount of otic tissue and EphA4 expression on the electroporated side of the embryos. Scale bar: in A, 100 μm. when misexpressed to anterior regions of the developing hind- and FGF10, appears to be required for otic vesicle formation brain (Alvarez et al., 2003). Ectopic expression of Fgf10 in the during normal development (Ladher et al., 2005) and in ectopic hindbrain of Fgf3−/−/Fgf10−/− embryos in this study rescued the locations in transgenic embryos. A similar cascade of FGF sig- formation of the otic vesicle. This demonstrates that hindbrain- nals, involving FGF8 and FGF10 as well as FGF4, participate derived FGF10 is able to reinstruct the mutants to proceed with during limb development (Martin and Groves, 2006; Capdevila inner ear development at least until the otic vesicle stage. This and Izpisua Belmonte, 2001). However, the exact temporal and result is consistent with previously postulated functions of the spatial sequence of FGF signals required for otic vesicle form- hindbrain for inner ear development, such as directing the ation remains to be defined. completion of otic placode induction and its invagination to Transgenic embryos misexpressing Fgf3 fail to induce Fgf8 form the otic vesicle (Baker and Bronner-Fraser, 2001; Noramly expression but, in contrast to FGF10 transgenic embryos, ecto- and Grainger, 2002; Whitfield, 2002; Brown et al., 2003; Riley pically induce a number of genes previously implicated in otic and Phillips, 2003; Barald and Kelley, 2004; Groves, 2005; vesicle formation, including Pax2, Dlx5 and kreisler/MafB in Schlosser, 2006). Nevertheless, despite the larger size of the their hindbrains. Pax2 and Dlx5 are usually expressed in the vesicle observed in the presence of the EphA4Fgf10 transgene otic placode and vesicle but not in hindbrain rhombomeres in Fgf3−/−/Fgf10−/− embryos, it is unclear if the rescued vesicle during normal development (Rinkwitz-Brandt et al., 1996; has recapitulated the entire otic program characteristic for this Acampora et al., 1999; Depew et al., 1999). Therefore, ectopic stage. expression of Pax2 and Dlx5 in r3 and r5 induced by mis- Most interestingly, misexpression of Fgf10 induced the expression of Fgf3 may be interpreted as part of the program ectopic expression of Fgf8 in the hindbrain, a location where it normally initiated in the placodal ectoderm via endogenous is not normally detected during normal mouse development hindbrain Fgf3 expression. However, these genes are all also (Crossley and Martin, 1995). Since FGF8 has recently been normally expressed at other locations in the developing brain, implicated as a crucial signal for otic vesicle formation in mice and specifically in regions influenced by FGF signaling (the (Ladher et al., 2005), the upregulation of FGF8 by FGF10 may midbrain–hindbrain border and diencephalon). Finally, we have be a key event in rescue experiments and in the generation of also addressed the question whether the EphA4Fgf3 transgene ectopic vesicles. Thus, in the context of transgenic embryos is able to rescue otic vesicle development in Fgf3−/−/Fgf10−/− which misexpress Fgf10 in the hindbrain, the ectopic induction embryos but have so far been unable to isolate double mutants of FGF8 and presence of endogenous FGF3 may be sufficient to carrying the transgene at E10 (n=62). initiate formation of ectopic vesicles. In contrast, in transgenic Induction of kreisler/MafB expression by FGF3 signaling embryos misexpressing Fgf3, exogenous Fgf3 and endogenous has been observed in the chicken and zebrafish hindbrain Fgf10 expression overlap only very transiently (Alvarez et al., (Maves et al., 2002; Aragon et al., 2005). The reverse has also 2003) and thus apparently fail to induce FGF8 and ectopic otic been demonstrated, where ectopic kreisler/MafB expression in- vesicles. Therefore, a triad of FGFs, consisting of FGF3, FGF8 duces Fgf3 expression in mice (Theil et al., 2002), whereas 388 L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 kreisler mouse mutants lose Fgf3 expression in r5 and r6 of the are likely to be required in a redundant fashion for PC hindbrain (Frohman et al., 1993; McKay et al., 1994). On the differentiation. other hand valentino (val), the zebrafish ortholog of kreisler/ MafB, represses Fgf3 expression in the hindbrain (Kwak et Redundant requirements for FGF3 and FGF8 or FGF3 and al., 2002). Therefore, the expression of Fgf3 and kreisler/MafB FGF10 during otic vesicle formation in the hindbrain is mutually regulated in several vertebrate species in either positive or negative feedback loops. This study has addressed issues of timing and redundancy by Loss of both Fgf3 and Fgf8 expression in the zebrafish generating and analyzing transgenic mice of different FGF hindbrain results in the absence of val expression and other mutant allele combinations. Combined loss of FGF3 and FGF8 hindbrain markers, paralleled by a failure to induce the otic after otic placode induction in Fgf3−/−/Fgf8flox/d2,3;Foxg1Cre/+ placode (Phillips et al., 2001; Leger and Brand, 2002; Maroon mutant embryos did not affect formation of the otic vesicle, nor et al., 2002; Maves et al., 2002; Walshe et al., 2002). In contrast, later inner ear morphogenesis and innervation. Earlier loss of in mouse which lacks Fgf8 expression in the hindbrain during FGF8 during placode induction in the FGF3 mutant background normal development, loss of FGF3 and FGF10 does not affect (Fgf3−/−/Fgf8flox/d2,3;Mox2Cre/+ mutant embryos) revealed a the hindbrain, but otic vesicle formation is impaired (Alvarez et redundant requirement for both FGFs during otic vesicle al., 2003; Wright and Mansour, 2003). Likewise, Fgf3−/−/ formation. Variation in phenotypes from small microvesicles Fgf8H/− mouse embryos show normal kreisler/MafB expression to simply smaller sized otic vesicles that failed to undergo proper but lack otic vesicles (Ladher et al., 2005). Therefore, in con- morphogenesis and differentiation was probably due to mosaic trast to the zebrafish, evidence for effects of loss of FGF signal- expression of the Mox2Cre transgene. The remaining otic ing on hindbrain patterning which may indirectly influence otic tissue observed in more severely affected Fgf3−/−/Fgf8flox/d2,3; vesicle formation has so far not been obtained in the mouse. Mox2Cre/+ mutant embryos showed a complete loss of Pax2 expression. Loss of Pax2 was also observed in the otic placode Requirements for FGF8 during murine inner ear development of Fgf3−/−/Fgf8H/− and in Fgf3−/−/Fgf10−/− mutant mouse embryos (Alvarez et al., 2003; Wright and Mansour, 2003; Ladher Both Foxg1-Cre-and Mox2-Cre-mediated inactivation of et al., 2005). Therefore, FGF signaling appears to be crucial for Fgf8 did not affect the formation of the otic vesicle. Thus FGF8 initiating the expression of Pax2 in the otic placode of mice, an appears not to be uniquely required for otic vesicle formation in observation which has also been made in zebrafish (Phillips et mice. Similar observations have been recently made in hy- al., 2001; Leger and Brand, 2002; Maroon et al., 2002). Inter- pomorphic mutant mouse embryos for Fgf8 (Ladher et al., estingly, recent results in zebrafish suggest that Pax2 may be 2005). These results differ from observations made in the necessary for maintaining otic precursor cells responsive to chicken where siRNA-mediated downregulation of Fgf8 led to FGF signaling (Hans et al., 2004; Mackereth et al., 2005). a reduced or absent placode (Ladher et al., 2005). Species- Important phenotypic differences are apparent between specific FGF8 requirements during otic vesicle formation may Fgf3−/−/Fgf8flox/d2,3;Mox2Cre/+ and Fgf3−/−/Fgf10−/− mutant explain the different phenotypes observed. Alternatively, differ- embryos. In severely affected Fgf3−/−/Fgf8flox/d2,3;Mox2Cre/+ ences between the type and timing of experimental manipula- mutant embryos we observed the absence or only minor rem- tions directed to reduce or abolish Fgf8 expression may also nants of otic tissue. In contrast during our analysis of Fgf3−/−/ influence the different phenotypes obtained. Fgf10 −/− double homozygous mouse embryos, we always FGF8 by itself does however play a later role in inner ear observed the formation of microvesicles (Alvarez et al., 2003). development. FGF8 expression is observed in the first cell type Moreover, in some cases these microvesicles appear to maintain to start to differentiate within the auditory sensory epithelium, the potential to initiate the development of the cochlear portion the IHC, and this expression persists throughout life (Pirvola et of the future inner ear but vestibular morphogenesis is arrested. al., 2002; Shim et al., 2005). Fgf8 expression in the IHCs has Defective development of the dorsal (vestibular) portion of mi- been postulated to control the differentiation of neighboring crovesicles was also noted in embryos carrying one mutant PCs which express the high-affinity receptor for FGF8, FGFR3 Fgf10 allele in an FGF3 mutant background (Fgf3−/−/Fgf10+/−), (Mueller et al., 2002). Importantly, in mouse mutants for and when these are less severely affected, mutants proceed with FGFR3, PCs fail to differentiate (Colvin et al., 1996; Hayashi et development to present with vestibular defects in the mature al., 2007). In contrast, mouse mutants lacking Sprouty2, which inner ear. Instead of three semicircular canals, only a single encodes a negative regulator of FGF receptor signaling, develop canal is formed. Whereas the anterior canal is still observed in an additional PC (Shim et al., 2005). In this case a direct role of Fgf3−/−/Fgf10+/− mutants, the lateral canal has been lost and the FGF8 is indicated by the fact that reducing Fgf8 gene dosage posterior canal most likely fused with the common cross. Loss restores the normal number of PCs. In the present study we of the posterior canal has also been observed in one of the Fgf3 observed the presence of IHCs and PCs using specific markers mutant strains (Mansour et al., 1993). This compares with loss in Fgf8flox/d2,3;Foxg1Cre/+ mutants. However, PCs in these of all canals in Fgf10 mouse mutants (Pauley et al., 2003; mutants appeared less differentiated compared to controls. Ohuchi et al., 2005). A role for FGFs in semicircular canal Since FGFR3 mutants show a more severe phenotype during formation has also been shown in the chicken embryo (Chang et PC differentiation compared to Fgf8flox/d2,3;Foxg1Cre/+ mutants al., 2004). Using an inhibitor of FGF receptor signaling, Chang additional members of the FGF gene family binding to FGFR3 et al. demonstrated a graded dose–response of blocking semi- L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 389 circular canal development; medium doses of inhibitor blocked ryngeal endoderm underlying the developing otic placode at posterior canal development whereas higher doses blocked the stages HH9–10 (Mahmood et al., 1995; Karabagli et al., 2002). development of all canals. Our analysis of mouse mutants Interestingly, strong pharyngeal endoderm expression of Fgf3 is doubly mutant for FGF3 and FGF10 now further defines the conserved in mouse embryos (Mahmood et al., 1996; McKay et differential requirements for both FGFs during the formation of al., 1996; Powles et al., 2004). Blocking this domain of Fgf3 the different semicircular canals. FGF3 and FGF10 are redund- expression inhibited the formation of the otic vesicle and the antly required for development of the lateral and posterior expression of otic marker genes in chicken embryos. Defects in semicircular canal, whereas the induction of the anterior semi- the pharyngeal endoderm affecting otic vesicle formation have circular canal only depends on FGF10. been observed in zebrafish van gogh mutants for Tbx1 (Pio- trowski and Nusslein-Volhard, 2000; Piotrowski et al., 2003). FGF3 is required for the transition from the otic placode to the Therefore, signals expressed in the pharyngeal endoderm may otic vesicle stage in chicken be transmitted via the intervening mesoderm to the developing otic vesicle to influence its formation and morphogenesis In this study we provide further evidence that FGF3 is suffi- (Mansour and Schoenwolf, 2005). cient and required for at least one important step of chicken inner ear development, the invagination of the otic placode to Acknowledgments form the otic vesicle. We previously had shown that ectopic expression of Fgf3 in the surface ectoderm of chicken embryos We thank Gail Martin, Susan McConnell, Jean Hebert, at HH8–9 results in the formation of ectopic otic placodes Dieter Riethmacher and Jose Luis de la Pompa for transgenic (Vendrell et al., 2000). However, Fgf3 expression is not ob- mouse lines, Merce Alonso, Judith Acebes and Sybille served in the surface ectoderm at this stage, but rather in the Koehnke for technical support and Cristina Pujades and hindbrain next to the developing otic placode (Mahmood et al., Andreas Trumpp for cDNA clones and comments on the 1995; Kil et al., 2005). We now show that ectopic expression of manuscript. Supported by the Spanish MEC (BFU2004-860) Fgf3 in the hindbrain at this stage, and thus in a more deve- and the DFG (SFB 444). lopmentally normal context, also leads to the formation of ectopic otic placodes. We previously had observed no induction of Appendix A. Supplementary data ectopic placodes upon viral misexpression of Fgf3 in the neural tube (Vendrell et al., 2000). Most likely, ectopic Fgf3 expression Supplementary data associated with this article can be found, levels obtained by electroporation and the amounts of trans- in the online version, at doi:10.1016/j.ydbio.2007.05.033. fected cells in the neural tube are thus higher than those observed in infected embryos. Although the otic placode is already specified at HH8 References (Groves and Bronner-Fraser, 2000; Kil et al., 2005) the fact that FGF3 is also required for placode invagination in vivo argues Acampora, D., Merlo, G.R., Paleari, L., Zerega, B., Postiglione, M.P., Mantero, S., Bober, E., Barbieri, O., Simeone, A., Levi, G., 1999. Craniofacial, strongly in favor that the intense expression of Fgf3 in the vestibular and bone defects in mice lacking the Distal-related gene Dlx. hindbrain from stage HH8 onwards reflects its function to Development 126, 3795–3809. promote the formation of the otic vesicle from the otic placode. Alvarez, Y., Alonso, M.T., Vendrell, V., Zelarayan, L.C., Chamero, P., Theil, T., Using siRNAs directed against Fgf3, we observed that this step Bosl, M.R., Kato, S., Maconochie, M., Riethmacher, D., Schimmang, T., is blocked leading to partially invaginated placodes or otic pits 2003. Requirements for FGF3 and FGF10 during inner ear formation. Development 130, 6329–6338. which fail to close. Since the otic placode is already specified at Aragon, F., Vazquez-Echeverria, C., Ulloa, E., Reber, M., Cereghini, S., Alsina, HH8–9(Groves and Bronner-Fraser, 2000; Kil et al., 2005) and B., Giraldez, F., Pujades, C., 2005. vHnf1 regulates specification of caudal the block of Fgf3 expression obtained in our experiments is rhombomere identity in the chick hindbrain. Dev. Dyn. 234, 567–576. likely to be incomplete, an earlier arrest of inner ear deve- Baker, C.V., Bronner-Fraser, M., 2001. Vertebrate cranial placodes I. Embryonic – lopment might be expected if Fgf3 expression is inhibited induction. Dev. Biol. 232, 1 61. Barald, K.F., Kelley, M.W., 2004. From placode to polarization: new tunes in before these stages in the neural tube. Additionally, Fgf3 inner ear development. Development 131, 4119–4130. expression is also observed in the mesoderm at HH7 before the Brooker, R., Hozumi, K., Lewis, J., 2006. Notch ligands with contrasting otic placode is specified (Karabagli et al., 2002; Kil et al., 2005). functions: Jagged1 and Delta1 in the mouse inner ear. Development 133, At this stage Fgf3 is coexpressed with Fgf19 (Kil et al., 2005), 1277–1286 (electronic publication 2006 Feb 22). which also induces Fgf3 in presumptive neural tissue (Ladher et Brown, S.T., Martin, K., Groves, A.K., 2003. Molecular basis of inner ear induction. Curr. Top. Dev. Biol. 57, 115–149. al., 2005). Thus the timing and location of Fgf3 expression, Capdevila, J., Izpisua Belmonte, J.C., 2001. Patterning mechanisms controlling either alone or in combination with Fgf19, may account for the vertebrate limb development. Annu. Rev. Cell. Dev. Biol. 17, 87–132. induction of the otic placode in chicken embryos (Brown et al., Carnicero, E., Zelarayan, L.C., Ruttiger, L., Knipper, M., Alvarez, Y., Alonso, 2003; Phillips et al., 2004; Kil et al., 2005). M.T., Schimmang, T., 2004. Differential roles of fibroblast growth factor-2 Our experiments have also revealed a novel site of Fgf3 during development and maintenance of auditory sensory epithelia. J. Neurosci. Res. 77, 787–797. expression necessary for otic vesicle formation, the pharyngeal Chang, W., Brigande, J.V., Fekete, D.M., Wu, D.K., 2004. The development of endoderm. Shortly after the onset of strong Fgf3 expression in semicircular canals in the inner ear: role of FGFs in sensory cristae. the hindbrain, Fgf3 transcripts are also observed in the pha- Development 131, 4201–4211. 390 L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391

Chi, C.L., Martinez, S., Wurst, W., Martin, G.R., 2003. The isthmic organizer Mansour, S.L., Schoenwolf, G.C., 2005. Morphogenesis of the inner ear. In: signal FGF8 is required for cell survival in the prospective midbrain and Kelley, M.W., Wu, D.K., Popper, A.N., Fay, R.R. (Eds.), Development of cerebellum. Development 130, 2633–2644. the Inner Ear. Springer, New York, pp. 43–84. Chapman, S.C., Collignon, J., Schoenwolf, G.C., Lumsden, A., 2001. Improved Mansour, S.L., Goddard, J.M., Cappechi, M.R., 1993. Mice homozygous for a method for chick whole-embryo culture using a filter paper carrier. Dev. targeted disruption of the proto-oncogene int-2 have developmental defects Dyn. 220, 284–289. in the tail and the inner ear. Development 117, 13–28. Colvin, J.S., Bohne, B.A., Harding, G.W., McEwen, D.G., Ornitz, D.M., 1996. Maroon, H., Walshe, J., Mahmood, R., Kiefer, P., Dickson, C., Mason, I., 2002. Skeletal overgrowth and deafness in mice lacking fibroblast growth factor Fgf3 and Fgf8 are required together for formation of the otic placode and receptor 3. Nat. Genet. 12, 390–397. vesicle. Development 129, 2099–2108. Crossley, P.H., Martin, G.R., 1995. The mouse Fgf8 gene encodes a family of Martin, K., Groves, A.K., 2006. Competence of cranial ectoderm to respond to polypeptides and is expressed in regions that direct outgrowth and patterning Fgf signaling suggests a two-step model of otic placode induction. in the developing embryo. Development 121, 439–451. Development 133, 877–887 (electronic publication 2006 Feb 1). Depew, M.J., Liu, J.K., Long, J.E., Presley, R., Meneses, J.J., Pedersen, R.A., Maves, L., Jackman, W., Kimmel, C.B., 2002. FGF3 and FGF8 mediate a Rubenstein, J.L., 1999. Dlx5 regulates regional development of the rhombomere 4 signaling activity in the zebrafish hindbrain. Development branchial arches and sensory capsules. Development 126, 3831–3846. 129, 3825–3837. Frohman, M.A., Martin, G.R., Cordes, S.P., Halamek, L.P., Barsh, G.S., 1993. McKay, I.J., Muchamore, I., Krumlauf, R., Maden, M., Lumsden, A., Lewis, J., Altered rhombomere-specific gene expression and hyoid bone differentia- 1994. The kreisler mouse: a hindbrain segmentation mutant that lacks two tion in the mouse segmentation mutant, kreisler (kr). Development 117, rhombomeres. Development 120, 2199–2211. 925–936. McKay, I.J., Lewis, J., Lumsden, A., 1996. The role of FGF-3 in early inner ear Giraldez, F., 1998. Regionalized organizing activity of the neural tube revealed development: an analysis in normal and kreisler mutant mice. Dev. Biol. by the regulation of lmx1 in the otic vesicle. Dev. Biol. 203, 189–200. 174, 370–378. Groves, A.K., 2005. The induction of the otic placode. In: Kelley, M.W., Wu, Meyers, E.N., Lewandoski, M., Martin, G.R., 1998. An Fgf8 mutant allelic D.K., Popper, A.N., Fay, R.R. (Eds.), Development of the Inner Ear. series generated by Cre- and Flp-mediated recombination. Nat. Genet. 18, Springer, New York, pp. 10–42. 136–141. Groves, A.K., Bronner-Fraser, M., 2000. Competence, specification and Morsli, H., Choo, D., Ryan, A., Johnson, R., Wu, D.K., 1998. Development of commitment in otic placode induction. Development 127, 3489–3499. the mouse inner ear and origin of its sensory organs. J. Neurosci. 18, Hamburger, V., Hamilton, H.L., 1992. A series of normal stages in the 3327–3335. development of the chick embryo. 1951. Dev. Dyn. 195, 231–272. Mueller, K.L., Jacques, B.E., Kelley, M.W., 2002. Fibroblast growth factor Hans, S., Liu, D., Westerfield, M., 2004. Pax8 and Pax2a function syner- signaling regulates pillar cell development in the organ of Corti. J. Neurosci. gistically in otic specification, downstream of the Foxi1 and Dlx3b trans- 22, 9368–9377. cription factors. Development 131, 5091–5102. Noramly, S., Grainger, R.M., 2002. Determination of the embryonic inner ear. Hayashi, T., Cunningham, D., Bermingham-McDonogh, O., 2007. Loss of J. Neurobiol. 53, 100–128. FGFR3 leads to excess hair cell development in the mouse organ of Corti. Ohuchi, H., Hori, Y., Yamasaki, M., Harada, H., Sekine, K., Kato, S., Itoh, N., Dev. Dyn. 236, 525–533. 2000. FGF10 acts as a major ligand for FGF receptor 2 IIIb in mouse Hebert, J.M., McConnell, S.K., 2000. Targeting of cre to the Foxg1 (BF-1) locus multi-organ development. Biochem. Biophys. Res. Commun. 277, mediates loxP recombination in the telencephalon and other developing 643–649. head structures. Dev. Biol. 222, 296–306. Ohuchi, H., Yasue, A., Ono, K., Sasaoka, S., Tomonari, S., Takagi, A., Itakura, Karabagli, H., Karabagli, P., Ladher, R.K., Schoenwolf, G.C., 2002. Com- M., Moriyama, K., Noji, S., Nohno, T., 2005. Identification of cis-element parison of the expression patterns of several fibroblast growth factors during regulating expression of the mouse Fgf10 gene during inner ear deve- chick gastrulation and neurulation. Anat. Embryol. (Berl.) 205, 365–370. lopment. Dev. Dyn. 233, 177–187. Kawauchi, S., Shou, J., Santos, R., Hebert, J.M., McConnell, S.K., Mason, I., Pauley, S., Wright, T.J., Pirvola, U., Ornitz, D., Beisel, K., Fritzsch, B., 2003. Calof, A.L., 2005. Fgf8 expression defines a morphogenetic center required Expression and function of FGF10 in mammalian inner ear development. for olfactory neurogenesis and nasal cavity development in the mouse. Dev. Dyn. 227, 203–215. Development 132, 5211–5223 (electronic publication 2005 Nov 2). Phillips, B.T., Bolding, K., Riley, B.B., 2001. Zebrafish fgf3 and fgf8 encode Kil, S.H., Streit, A., Brown, S.T., Agrawal, N., Collazo, A., Zile, M.H., redundant functions required for otic placode induction. Dev. Biol. 235, Groves, A.K., 2005. Distinct roles for hindbrain and paraxial mesoderm in 351–365. the induction and patterning of the inner ear revealed by a study of Phillips, B.T., Storch, E.M., Lekven, A.C., Riley, B.B., 2004. A direct role for vitamin-A-deficient quail. Dev. Biol. 285, 252–271. Fgf but not Wnt in otic placode induction. Development 131, 923–931. Kwak, S.J., Phillips, B.T., Heck, R., Riley, B.B., 2002. An expanded domain of Piotrowski, T., Nusslein-Volhard, C., 2000. The endoderm plays an important fgf3 expression in the hindbrain of zebrafish valentino mutants results in role in patterning the segmented pharyngeal region in zebrafish (Danio mis-patterning of the otic vesicle. Development 129, 5279–5287. rerio). Dev. Biol. 225, 339–356. Ladher, R.K., Wright, T.J., Moon, A.M., Mansour, S.L., Schoenwolf, G.C., Piotrowski, T., Ahn, D.G., Schilling, T.F., Nair, S., Ruvinsky, I., Geisler, R., 2005. FGF8 initiates inner ear induction in chick and mouse. Genes Dev. 19, Rauch, G.J., Haffter, P., Zon, L.I., Zhou, Y., Foott, H., Dawid, I.B., Ho, 603–613. R.K., 2003. The zebrafish van gogh mutation disrupts tbx1, which is Leger, S., Brand, M., 2002. Fgf8 and Fgf3 are required for zebrafish ear involved in the DiGeorge deletion syndrome in humans. Development 130, placode induction, maintenance and inner ear patterning. Mech. Dev. 119, 5043–5452. 91–108. Pirvola, U., Ylikoski, J., Trocovik, R., Hérbert, J.M., McConnell, S., Partanen, Mackereth, M.D., Kwak, S.J., Fritz, A., Riley, B.B., 2005. Zebrafish pax8 is J., 2002. FGFR1 is required for the development of the auditory sensory required for otic placode induction and plays a redundant role with Pax2 ephitelium. Neuron 35, 671–680. genes in the maintenance of the otic placode. Development 132, 371–382 Powles, N., Marshall, H., Economou, A., Chiang, C., Murakami, A., Dickson, (electronic publication 2004 Dec 16). C., Krumlauf, R., Maconochie, M., 2004. Regulatory analysis of the mouse Mahmood, R., Kiefer, P., Guthrie, S., Dickson, C., Mason, I., 1995. Multiple Fgf3 gene: control of embryonic expression patterns and dependence upon roles for FGF-3 during cranial neural development in the chicken. Deve- sonic hedgehog (Shh) signalling. Dev. Dyn. 230, 44–56. lopment 121, 1399–1410. Represa, J., Leon, Y., Miner, C., Giraldez, F., 1991. The int-2 proto-oncogene is Mahmood, R., Mason, I.J., Morriss-Kay, G.M., 1996. Expression of Fgf-3 in responsible for induction of the inner ear. Nature 353, 561–563. relation to hindbrain segmentation, otic pit position and pharyngeal arch Riley, B.B., Phillips, B.T., 2003. Ringing in the new ear: resolution of cell morphology in normal and retinoic acid-exposed mouse embryos. Anat. interactions in otic development. Dev. Biol. 261, 289–312. Embryol. (Berl.) 194, 13–22. Rinkwitz-Brandt, S., Arnold, H.H., Bober, E., 1996. Regionalized expression of L.C. Zelarayan et al. / Developmental Biology 308 (2007) 379–391 391

Nkx5-1, Nkx5-2, Pax2 and sek genes during mouse inner ear development. mice: a tool to distinguish embryonic vs. extra-embryonic gene function. Hear. Res. 99, 129–138. Genesis 26, 113–115. Schlosser, G., 2006. Induction and specification of cranial placodes. Dev. Biol. Theil, T., Ariza-McNaughton, L., Manzanares, M., Brodie, J., Krumlauf, R., 294, 303–351 (electronic publication 2006 May 3). Wilkinson, D.G., 2002. Requirement for downregulation of kreisler Shim, K., Minowada, G., Coling, D.E., Martin, G.R., 2005. Sprouty2, a mouse during late patterning of the hindbrain. Development 129, 1477–1485. deafness gene, regulates cell fate decisions in the auditory sensory epithe- Vendrell, V., Carnicero, E., Giraldez, F., Alonso, M.T., Schimmang, T., 2000. lium by antagonizing FGF signaling. Dev. Cell 8, 553–564. Induction of inner ear fate by FGF3. Development 127, 2011–2019. Soriano, P., 1999. Generalized lacZ expression with the ROSA26 Cre reporter Walshe, J., Maroon, H., McGonnell, I.M., Dickson, C., Mason, I., 2002. strain. Nat. Genet. 21, 70–71. Establishment of hindbrain segmental identity requires signaling by FGF3 Storm, E.E., Garel, S., Borello, U., Hebert, J.M., Martinez, S., McConnell, and FGF8. Curr. Biol. 12, 1117–1123. S.K., Martin, G.R., Rubenstein, J.L., 2006. Dose-dependent functions of Whitfield, T.T., 2002. Zebrafish as a model for hearing and deafness. Fgf8 in regulating telencephalic patterning centers. Development 133, J. Neurobiol. 53, 157–171. 1831–1844. Wilkinson, D.G., Peters, G., Dickson, C., McMahon, A.P., 1988. Expression of Sun, X., Meyers, E.N., Lewandoski, M., Martin, G.R., 1999. Targeted disruption the FGF-related proto-oncogene int-2 during gastrulation and neurulation in of FGF8 causes failure of cell migration in the gastrulating mouse embryos. the mouse. EMBO J. 7, 691–695. Genes Dev. 13, 1834–1846. Wright, T.J., Mansour, S.L., 2003. Fgf3 and Fgf10 are required for mouse otic Tallquist, M.D., Soriano, P., 2000. Epiblast-restricted Cre expression in MORE placode induction. Development 130, 3379–3390.