differential expression of sox2 and sox3 in neuronal and sensory progenitors of...

Differential Expression of Sox2 andSox3 in Neuronal and Sensory

Progenitors of the Developing Inner Earof the Chick

JOANA NEVES,1 ANDRES KAMAID,1,2 BERTA ALSINA,1

AND FERNANDO GIRALDEZ1*1Biologia del Desenvolupament, Departament de Ciencies Experimentals i de la Salut

(DCEXS), Universitat Pompeu Fabra, Parc de Recerca Biomedica de Barcelona (PRBB),08003-Barcelona, Spain

2Departamento de Histologıa y Embriologıa, Facultad de Medicina, Universidad de laRepublica, Montevideo, Uruguay 11200

ABSTRACTThe generation of the mechanosensory elements of the inner ear during development

proceeds in a precise temporal and spatial pattern. First, neurosensory precursors formsensory neurons. Then, prosensory patches emerge and give rise to hair and supporting cells.Hair cells are innervated by cochleovestibular neurons that convey sound and balanceinformation to the brain. SOX2 is an HMG transcription factor characteristic of the stem-cellgenetic network responsible for progenitor self-renewal and commitment, and its loss offunction generates defects in ear sensory epithelia. The present study shows that SOX2protein is expressed in a spatially and temporally restricted manner throughout developmentof the chick inner ear. SOX2 is first expressed in the neurogenic region that gives rise tosensory neurons. SOX2 is then restricted to the prosensory patches in E4 and E5 embryos, asrevealed by double and parallel labelling with SOX2 and Tuj1, MyoVIIa, or Islet1. Prolifer-ating cell nuclear antigen labelling showed that SOX2 is expressed in proliferating cellsduring those stages. By E5, SOX2 is also expressed in the Schwann cells of the cochleoves-tibular ganglion, but not in the otic neurons. At E8 and E17, beyond stages of sensory cellspecification, SOX2 is transiently expressed in hair cells, but its level remains high insupporting cells. SOX3 is concomitantly expressed with SOX2 in the neurogenic domain ofthe otic cup, but not in prosensory patches. Our data are consistent with a role for SOX2 inspecifying a population of otic progenitors committed to a neural fate, giving rise to neuronsand hair cells. J. Comp. Neurol. 503:487–500, 2007. © 2007 Wiley-Liss, Inc.

Indexing terms: Sox; otic; ear progenitors; proneural; sensory organ

The basic functional unit of the inner ear consists of fourelements of neural origin: the mechanotransducing haircells (HCs), the supporting cells (SC), the primary afferentneurons, and the ganglion Schwann cells. Those elementsdevelop in a stereotyped manner, with small variationsamong species (Adam et al., 1998; Torres and Giraldez,1998; Fritzsch et al., 2006). The specification of the celltypes of the chick inner ear proceeds sequentially and in aprecise spatial pattern. First, neuroblasts are specified inthe ventral epithelium of the otic cup, and they delami-nate to form the cochleovestibular ganglion (CVG, VIIIcranial nerve). Ganglionar neuroblasts still proliferate,and they expand generating the bipolar otic neurons thatconnect the hair cells to the brainstem (D’Amico-Martel,1982; Whitehead and Morest, 1985; Alsina et al., 2003).Delayed by one day in the chick, sensory patches emergein the otocyst at precise positions and develop into the

sensory organs that contain the mechanotransducing haircells (HCs) and the supporting cells (SCs). During the past

This article includes Supplementary Material available via the Internetat http://www.interscience.wiley.com/jpages/0021-9967/suppmat.

Grant sponsor: Ministerio de Educacion y Ciencia (MEC); Grant number:BMC2002-00355; Grant number: BFU2005-03045; Grant sponsor: Minis-terio de Sanidad y Consumo; Grant number: XT G03/203 ISCIII; Grantsponsor: Fundacao para a Ciencia e Tecnologia (FCT), Portugal; Grantnumber: SFRH/BD/25688/2005 (to J.N.).

*Correspondence to: Fernando Giraldez, DCEXS-Universitat PompeuFabra, c/Dr.Aiguader 88, 08003-Barcelona, Spain.E-mail: [email protected]

Received 12 June 2006; Revised 20 October 2006; Accepted 26 December2006

DOI 10.1002/cne.21299Published online in Wiley InterScience (www.interscience.wiley.com).

THE JOURNAL OF COMPARATIVE NEUROLOGY 503:487–500 (2007)

© 2007 WILEY-LISS, INC.

years, we have expanded the identification and knowledgeof the genes involved in the specification and differentia-tion of neurons and hair cells, particularly the expressionand function of proneural genes (Lewis, 1998; Berming-ham et al., 1999) and the role of cell communication sig-nals (Alsina et al., 2004; Barald and Kelley, 2004; Pujadeset al., 2006). However, we are still far from understandingthe mechanisms by which the different neural cell types ofsensory patches are generated from cell progenitors. Evi-dence has been obtained supporting the idea that neuronsand hair cells can originate from a common progenitor cell(Lang and Fekete, 2001; Matei et al., 2005; Satoh andFekete, 2005), and common genetic networks may under-lie the development of sensory cells of the eye and the ear(Fritzsch et al., 2006).

The SOX2 transcription factor is expressed early in em-bryonic and neural stem cells. SOX genes contain an HMG-box closely related to that of the mammalian testes-determining gene SRY and are highly conserved throughoutevolution. The C-terminal region of the SOX protein carriesa cryptic transactivating domain that is uncovered only afterspecific interaction with partner factors. To date, 24 verte-brate SOX genes have been identified and are classified intoseven subgroups (A–G) based on sequence identity, and atleast 12 members of the SOX gene family are expressed inthe nervous system (Wegner and Stolt, 2005; Pevny andPlaczek, 2005). Studies in the chick embryo have providedevidence that neural inducing signals directly regulateSOX2 expression in the neural tube and that SOX2 is re-sponsible for commitment of actively proliferating cells to aneural fate (Rex et al., 1997; Graham et al., 2003). Recentstudies also revealed that the SOX2 regulatory region con-tains a domain that responds directly to neural-inducingsignals, which is conserved across diverse animal species.Sox2-deficient mice such as light coat and circling (Lcc) andyellow submarine (Ysb) show hearing and balance impair-ment. Lcc/Lcc mutant mice fail to establish a prosensorydomain, and neither hair cells nor supporting cells differen-tiate (Kiernan et al., 2005). Ysb/Ysb mice show abnormaldevelopment, with disorganized and fewer hair cells. Thesephenotypes are a direct consequence of the absence or re-duced expression of the transcription factor SOX2 in thedeveloping inner ear (Kiernan et al., 2005). Moreover, mu-tations of SOX2 in humans cause anophthalmia, sensorineu-ral hearing loss, and global brain defects (Hagstrom et al.,2005) and also regulate retinal neural progenitor compe-tence (Taranova et al., 2006).

This background suggests that SOX2 plays an importantrole in the development of neural elements of the ear, par-ticularly in the regulation of the function of neural progeni-tors. Here, we report the first detailed analysis of SOX2expression in the developing inner ear, from the early oticvesicle until postdifferentiation stages. The results showthat SOX2 is expressed in proliferating cells of neurogenicand prosensory domains from the otic cup to early otocyst. Asdevelopment proceeds, SOX2 is not expressed in differenti-ated neurons and hair cells but remains intensely expressedin supporting cells. We propose that SOX2 may play criticalroles in ear development. First, it is associated with theself-renewal state and the proneural character of ear neuro-sensory progenitors, becoming an early marker for prosen-sory progenitor cells. Second, SOX2 is absent from hair cellsand neurons after terminal division but it is retained bysupporting cells along with their capacity to divide, suggest-

ing that SOX2 may be a key element in hair cell regenerationafter damage.

MATERIALS AND METHODS

Embryos

Fertilized hens’ eggs (Granja Gibert, Tarragona, Spain)were incubated at 38°C for the designated times, and em-bryos were staged according to Hamburger and Hamilton(1992). Embryos were dissected from the yolk and fixed byimmersion in 4% paraformaldehyde in phospate-bufferedsaline (pH 7.5; 4% PFA/PBS) overnight at 4°C; dehydratedwith consecutive washings in 25%, 50%, 75%, and 100%methanol in PBST (PBS 0.1% Tween); and stored at –20°C.

Cryostat sectionning

For cryostat sectioning, embryos fixed in 4% parafor-maldehyde were immersed in 15% sucrose and embeddedin 7.5% gelatin/15% sucrose. Blocks were frozen in isopen-tane to improve tissue preservation and then sectioned at20 �m thickness onto Superfrost Plus Slides (Fisher,Pittsburg, PA) and stored at –20°C. Sections were usedeither for immunohistochemistry or in situ hybridization.

Immunostaining

For immunodetection, the sections were thawed andpermeabilized for 20 minutes with PBST. For antigenretrieval, sections were treated with citrate buffer (pH 6.1)at 95°C for 30 minutes. Nonspecific binding was blockedfor 1 hour at room temperature in 0.1% Tween, 10% heat-inactivated goat serum in PBS (blocking solution). Slideswere incubated overnight at 4°C with diluted antibodiesand covered with Parafilm.

The antibody against SOX2 protein (Abcam ab15830, lot103889) was a rabbit polyclonal antibody raised againstthe synthetic peptide of residues 9–20 of human SOX2(GNQKNSPDRVKR) conjugated to KLH. The immunogenhas 100% identity with the chicken protein. The antibodystained the nucleus of human ES cells (H9 line) and de-tected bands of apparent mw 34–42 kDa (predicted size34 kDa) in Western blots of different human and mousestem cells, where it did not cross-react with SOX1 (Abcamhttp://www.abcam.com/index.html?datasheet�15830).Antigen amino acid identity with other SOX proteins was42% with SOX3 and SOX1 and less than 25% for other SOXproteins (SOX4, -8, -9, -10, -11, -12, -14, -21). We performedfour controls for immunostaining with SOX2 antibody. First,specificity of antibody binding was tested by preadsorptionof the antiserum against the synthetic peptide (Suppl. Fig.1). All staining was abolished from neural tube and otocystwhen 1 ml of rabbit polyclonal anti-SOX2 antibody wasincubated with 2.5 ml of SOX2 antigen peptide used to raisethe antibody (Abcam ab15831, synthetic peptide derivedfrom residues 9–20 of human SOX2). Positive controls in theabsence of peptide were run in alternate parallel sections.Second, we carried out a Western blot analysis of chickneuroectoderm extracts, which were compared with mouseES-CGR8 cells and extraembryonic membrane extracts aspositive and negative controls, respectively (Suppl. Fig. 2). Aband of an apparent mw 38 kDa was detected in nuclearextracts of chick neuroectoderm and in mouse stem cells.This band was absent from extraembryonic tissues and fromcytoplasmic extracts of neural ectoderm. Chick embryonicneuroectoderm also showed another band of 35 kDa. This is

The Journal of Comparative Neurology. DOI 10.1002/cne

488 J. NEVES ET AL.

the expected mw of the SOX2 protein. SOXB proteins areknown to undergo posttranslational modifications, includingSUMOylation (Savare et al., 2005), and the chicken SOX2protein has five possible SUMOylation sites (SUMOPLOT;Abgent). The two bands observed in chick extracts maytherefore correspond to posttranslational modifications ofthe SOX2 protein and are within the range of molecularweights observed in mouse and human extracts (see above).Third, to ensure that the SOX2 antibody did not react withSOX3, which was also expressed at early developmentalstages, we checked for cross-reactivity of the SOX2 antibodywith the chSOX3 protein. For this purpose, the chickenSox3-GFP fusion gene was transfected in HeLa CCL-2 cells,which were then probed by immunofluorescence with SOX2or SOX3 antibodies (Suppl. Fig. 3). These experimentsshowed that anti-SOX2 antibody was unable to recognisethe chicken SOX3 protein when expressed in HeLa cells. Theexpression of SOX3 was assessed by the concomitant detec-tion of green fluorescent protein (GFP) in the same cells andalso by immunofluorescence with the SOX3 antibody in par-allel plates. Finally, the SOX2 immunostaining showed aclose overlapping and colocalized with Sox2 in situ hybrid-ization for the Sox2 mRNA (see Fig. 7). The anti-SOX2antibody was used at a dilution of 1:400.

The anti-SOX3 antibody was a rabbit polyclonal anti-serum, kindly provided by Thomas Edlund and JonasMuhr. The immunogen was a fusion protein expressed inE. coli that contained the 76 C-terminal amino acids ofchicken SOX3 protein (Thomas Edlund, unpublisheddata). This antibody has been used to detect chicken SOX3protein (Wilson et al., 2001; Bylund et al., 2003). Theauthors used this antibody to compare the expression ofSOX1, -2, and -3 (Fig. 1 in Bylund et al., 2003), and theyshowed that transcriptional activity of the Sox3 gene iscorrelated by corresponding levels of SOX3 protein asdetected by the SOX3 antibody (Figs. 3, 4 in Bylund et al.,2003) and that immunoreactivity is down-regulated byproneural genes and neural-inducing signals, both in vitroan in vivo (Wilson et al., 2001; Bylund et al., 2003). In ourhands, the SOX3 antibody corresponded well with Sox3mRNA expression in the otic cup and neural tube (see Fig.7). Whereas SOX3 expression in the neural tube persistedalong with that of SOX2 throughout development, SOX3staining was negative in the otocyst beyond E3 (see Fig.7). Because SOX2 was distinctly expressed in sensorypatches at these stages, we concluded that the SOX3 an-tibody did not cross-react with SOX2 in the developingear. The SOX3 antibody was used at a dilution of 1:300.

Other antibodies used and dilutions were as follows. 1)Mouse monoclonal anti-HNK1 (CD57, clone HNK1, super-natant; Becton Dickinson; gift from Lopez-Botet, Univer-sitat Pompeu Fabra, Barcelona, Spain). This antibody wasderived from hybridization of mouse P3-X63-Ag8.653 my-eloma cells with lymph node cells from BALB/c mice im-munized with membrane extracts of the HSB-2T-lymphoblastoid cell line. It recognizes a carbohydrate-structure antigen from myelin-associated glycoproteinsexpressed in a subset of natural killer (NK) and T lym-phocytes (Abo and Balch, 1981). Dilution was 1:400. 2)Mouse monoclonal antineuronal class III �-tubulin (Tuj1,MMS-435P; Berkeley Antibody Company, Berkeley, CA).This antibody was raised against microtubules derivedfrom rat brain and recognizes the epitope CEAQGPK,corresponding to the N-terminal, isotype-defining domainof class III �-tubulin. This antibody is well characterized

and highly reactive to neuron-specific class III �-tubulin(�III). Tuj1 does not identify �-tubulin in glial cells (Lee etal., 1990), and it has been extensively characterized in thedeveloping ear by Molea et al. (1999). Dilution was 1:400.3) Mouse monoclonal anti-Islet1 (clone 39.4D5, cell culturesupernantant, Developmental Studies Hybridoma Bank,University of Iowa). This antibody was developed byThomas M. Jessel (Columbia University) and was raisedagainst a bacterial expression generated peptide corre-sponding to the fragment extending from the beginning ofthe homeodomain to the C-terminal end of the protein.This antibody has been characterized in the developingear by Li et al. (2004). Dilution was 1:400. 4). Rabbitpolyclonal antimyosin VIIa (MyoVIIa; Tama Hasson, Uni-versity of California, San Diego). This antibody was raisedagainst human myosinVIIa tail, and it is specific to theportion comprising amino acids 880–1,070. The antibodywas developed and characterized by Tama Hasson (Has-son et al., 1995). Dilution was 1:5,000. 5) Mouse monoclo-nal antiproliferating cell nuclear antigen (PCNA; ab29,clone pc-10; Abcam, Cambridge, United Kingdon). Thisantibody was raised against the protein A-rat PCNA fu-sion obtained from pC2T. It is specific for PCNA p36protein expressed at high levels in proliferating cells(Waseem and Lane, 1990). It reacts with human, mouse,rat, chicken, and fruit fly (Drosophila melanogaster) anddetects a 30-kDa band in WB of chicken DT40 cell totalextract (see abcam Abreviews at http://www.abcam.com/index.html?pageconfig�reviews&intAbID�29). Dilutionwas 1:200.

All antibodies were diluted in blocking solution. Unboundantibody was removed by successive washes with PBST for 6hours and incubated for 2 hours with Alexa-488- and -594-conjugated goat anti-mouse and anti-rabbit, respectively,secondary antibodies (Molecular Probes, Eugene, OR), di-luted 1:400 in blocking solution. Slides were washed over-night in PBST and mounted with Vectashield mountingmedium with DAPI (Vector Laboratories, Burlingame, CA),which provided DNA counterstaining. Slides were analyzedby conventional fluorescence microscopy (Leica DMRB Flu-orescence Microscope with Leica CCD camera DC300F). Im-ages were acquired with Leica IM50 image manager, andAdobe Photoshop 7.0.1 was used to obtain the merged im-ages and to adjust contrast and brightness. Figures 2b,c,h,iand 7k–n were corrected for uneven illumination with thePhotoshop dodging toolbar.

In situ hybridization

In situ hybridization in sections was performed essen-tially as described by Nieto et al. (1996), using digoxigenin(DIG)-labelled riboprobes. DIG was detected with anti-DIG-AP and NBT/BCIP (Roche), which gives a purplestaining. The riboprobes were chicken Sox2 and Sox3 (Rexet al., 1997). After the in situ hybridization protocol, sec-tions were immunostained with Tuj1 antibody, followingthe method described above.

RESULTS

SOX2 is expressed in the neurogenicdomain of the the otic vesicle (E2 and E3)

The experiments that follow show the expression of theSOX2 protein throughout the development of the inner


489SOX2 IN THE CHICK INNER EAR

ear, between early otic vesicle and prehatching stages.The proneural domain of the inner ear is specified early indevelopment, during initial stages of the formation of theotic primordium. It is located in the anterior aspect of theotic cup and otic vesicle, and it is complementary to theposterior nonneural region of the otic cup. The latter ischaracterized by the expression of patterning genes andthe carbohydrate epitope HNK1 (Alsina et al., 2004). Theexpression of SOX2 in an E2.5 (HH16) otic vesicle isshown in Figure 1. Parasagittal sections of otic vesicleswere doubly stained for SOX2 and HNK1, red and green,respectively (SOX2 is always red throughout the paper).Figure 1b shows that SOX2 was abundantly expressed incell nuclei but only in cells residing in the anterior domainof the otic vesicle, a domain corresponding to the proneu-ral domain (Alsina et al., 2004). The expression of SOX2was complementary to that of HNK1, as shown by com-paring Figure 1b,c and in the merged images (Fig. 1a,d).The relation between SOX2 expression and neuroblastgeneration was also explored by doubly labelling otic ves-icles with SOX2 and either Islet1 or Tuj1 (Fig 1e–l). Islet1is a Lim-homeodomain protein expressed in young neuro-blasts during delamination and in auditory and vestibular

neurons (Adam et al., 1998; Li et al., 2004). Figure 1e–hshows that SOX2 expression overlapped with the domainof Islet1 expression and the region of generation of thecochleovestibular neurons. Only few scattered cells in theotic epithelium showed coexpression of both SOX2 andIslet1 proteins, indicating that initiation of Islet1 expres-sion can occur before SOX2 protein is down-regulated (Fig.1h, orange, and see below). SOX2, however, was neverexpressed in delaminated neuroblasts in the cochleoves-tibular ganglion. This is also illustrated by double label-ling otic vesicles for SOX2 and Tuj1, which showed thatearly neuroblasts were generated from the SOX2-positivedomain, and that SOX2 was not expressed by Tuj1-positive neuroblasts (Fig. 1i–l).

SOX2 is expressed in the prosensorypatches of the otocyst (E4 and E5)

By E3.5–E4 (HH22–24), ear prosensory domains be-come specified, first the vestibular and then the auditorysensory organs, and, by E5 (HH26), the different sensoryorgans can be clearly identified (Wu et al., 1998; Cole etal., 2000; Sanchez-Calderon et al., 2005). Figure 2 showsserial coronal sections of an E4 and E5 embryo that were

Fig. 1. SOX2 is expressed in the neurogenic domain of chick E2.5and E3 otic vesicles. Double immunostaining of parasagittal sectionsof HH16 otic vesicles for SOX2 protein (red), and the HNK1 epitope(green; a–d). SOX2 expression was detected in the nuclei of the cellslocated in the anterior domain (b) and HNK1 expression in a comple-mentary fashion in the posterior domain (c). Double immunostainingof parasagittal sections of HH18 otic vesicles against SOX2 protein(red) and Islet1 (green; e–h). Islet1 expression in the cochlear-vestibular ganglion (CVG, arrow in e) The scarce yellow nuclei corre-

sponded to cells positive for Islet1 and SOX2 (h, arrowheads). Expres-sion of SOX2 and Tuj1 (i–l). Tuj1 expression was restricted to thecochlear-vestibular ganglion (i,k,l). No overlapping between Tuj1 andSOX2 positive cells was detected (l). DAPI is shown in a, e, and i.Dashed frames indicate the proneural domain enlarged at the right.D, dorsal; A, anterior; CVG, cochleovestubular ganglion; OV, oticvesicle. Scale bars � 50 �m in a (applies to a–d); 100 �m in e (appliesto e,i); 25 �m in f (applies to f–h,j–l).


490 J. NEVES ET AL.

stained for SOX2 and Tuj1. Tuj1 recognizes a neuron-specific �III tubulin epitope that is also expressed in nas-cent hair cells (Stone et al., 1996; Stone and Rubel, 2000).At E4, the prosensory patches can be detected by geneexpression domains that anticipate future sensory organs

(see diagram in Fig. 2a; see also Cole et al., 2000; Sanchez-Calderon et al., 2005). The experiment shown in Figure2b–f, illustrates the expression of SOX2 in the prospectivecristae (Fig. 2b), macula utriculi (Fig. 2c,d), macula sacculi(Fig. 2e), and basilar papilla (Fig. 2f). The correspondence

Fig. 2. SOX2 expression is restricted to the prosensory patches ofotocysts in the E4 and E5 embryo. Diagrams represent otocysts fromE4 (a) and E5 (g) stage embryos. Sensory patches are represented inred, as defined by gene expression, and were modified from Sanchez-Calder´on et al. (2005). Planes represent the extreme levels of thesections shown for each stage. All sections were immunostained forSOX2 (red), Tuj1 (green), and DAPI (blue). Serial coronal sections ofan E4 otocyst, in a dorsal-ventral sequence (b–f). SOX2 expression investibular (b–e) and auditory (f) prosensory patches. Serial coronalsections of an E5 otocyst in a dorsal to ventral sequence (h–i). SOX2

expression in vestibular (h–k) and auditory (l) sensory patches. Tuj1labeled neurons of the vestibular (vg) and cochlear ganglion (cg) andthe fibers innervating the prosensory domains (d–f,i–l). SOX2 inposteior crista (b,i), superior crista (b,h), lateral crista (i), maculautriculi (c,d,j), macula sacculi (e,k), and basilar papilla (f,l). pc, Pos-terior crista; sc, superior crista; lc, lateral crista; mu, macula utriculi;ms, macula sacculi; bp, basilar papilla; ml, macula lagena; vg, vestib-ular ganglion; cg, cochlear ganglion. Orientation: D, dorsal; M, me-dial; A, anterior. Scale bars � 100 �m in b (applies to b–f); 100 �m inh (applies to h–l).



between SOX2 expression and the sensory patches contin-ued in E5, as illustrated in Figure 2h–l, which shows anexperiment processed as described above for SOX2 andTuj1. Again, the expression of SOX2 was restricted to thesensory patches. From dorsal to ventral, SOX2 expressionis shown in cristae (Fig. 2h,i), maculae (Fig. 2j,k), andbasilar papilla (Fig. 2l). The diagrams in Figure 2a,g showthe location of sensory patches as drawn from gene ex-pression patterns and were modified from Sanchez-Calderon et al. (2005).

To analyze better the correspondence between SOX2expression and cell differentiation in sensory patches, weperformed a comparison between SOX2 expression andMyoVIIa, which is a well-established cellular marker forthe incipient differentiation of hair cells (Sahly et al.,1997; Wolfrum et al., 1998). Serial alternate sections of E5embryos were immunostained in parallel for SOX2/Tuj1(Fig. 3a–c) and MyoVIIa/Tuj1 (Fig. 3d–f). The experimentillustrates that the expression of SOX2 and MyoVIIa over-lapped in sensory patches. To analyze this correspondencefurther at the cellular level, selected patches were exam-ined at higher resolution and are shown in Figure 3g,h. Asshown, SOX2 was concentrated in the basal cell layer,whereas MyoVIIa was restricted to the luminal layer oc-cupied by hair cells. At these early stages of hair cellproduction, SOX2 was sometimes detected in the hair celllayer, but, later in development, SOX2 was not expressedin hair cells (see below).

Interestingly, and in contrast to what was observed inprevious stages, SOX2 was now expressed in the co-chleovestibular ganglion. In E5 embryos, SOX2 wasdistinctly expressed in a spotted pattern that intermin-gled with neurons (Fig. 3i). The staining was nuclear, asconfirmed at high magnification with DAPI counter-staining (Fig. 3j). These cells corresponded well toSchwann glial cells that populate the ganglion and thatsurround the neurons and nerve fibers. Note that thelarge nuclei of Tuj1-positive neurons (Fig. 3j) are nega-tive for SOX2, suggesting that SOX2 was not expressedin postmitotic otic neurons.

Islet1 is known to be expressed in nascent hair andsupporting cells, and it persists until later stages ofdevelopment in differentiating sensory cells (Li et al.,2004). The correspondence between SOX2 and Islet1expression in E5 otocysts is shown in Figure 4a– c (su-perior and posterior cristae, Fig. 4a; basilar papilla, Fig.4b). Both in crista and basilar papilla there was a re-stricted and overlapping expression of SOX2 and Islet1.At the cellular level, sensory patches showed cells thatwere positive for SOX2, for Islet1, and for both (Fig. 4c).Generally, the SOX2 domain was contained within theIslet1 patch (Fig. 4b). Taking Islet1 as an early markerfor sensory cell commitment, this suggests that SOX2expression persisted transiently during early stages ofcell fate acquisition (see below). Double-labelling exper-iments with SOX2 and the proliferation marker PCNAshowed that most SOX2 expression in sensory patchesof the E5 otocyst was related to proliferating cells (Fig.4d–f). Double labelling with SOX2 and PCNA showed avery intense proliferative activity in the otocyst at thisstage and a great extent of overlap (Fig. 4d–f). At thecellular level, Figure 4f illustrates that most SOX2-positive cells also expressed PCNA. This would indicatethat there is a transition state in which sensory progen-

itors are committed but still proliferative (Doetzlhoferet al., 2006).

SOX2 is expressed in differentiation stages(E8 and E17)

Expression of SOX2 was followed further in develop-ment and studied at stages E8 and E17, well beyond theperiod of cell specification and when hair and supportingcells are, respectively, in the process of early and latedifferentiation. The results in Figure 5 show the expres-sion of SOX2 and Tuj1 in a stage E8 otocyst. SOX2 was, asbefore, restricted to the sensory organs, which were nowclearly identified by morphological criteria and by theexpression of Tuj1 that labelled both hair cells and theinnervating fibers. SOX2 expression was intense in allsensory organs, some of which are shown as examples inFigure 5: posterior crista (Fig. 5b), macula utriculi (Fig.5c), macula sacculi (Fig. 5d), and basilar papilla (Fig. 5e).A detailed analysis with cellular resolution is shown inFigure 5f for the macula sacculi and in Figure 5g for thebasilar papilla. It can be seen that SOX2 expression waslow or absent in hair cells but remained intense in sup-porting cells (Fig. 5f,g).

The results obtained from stage E17 otocysts are shownin Figure 6. Again, SOX2 was expressed in all sensoryorgans. Examples are shown of the posterior cristae (Fig.6a,b), macula utriculi (Fig. 6c), macula sacculi (Fig. 6d),and basilar papilla (Fig. 6e,f, inset shows the maculalagena). High magnifications in Figure 6b,e show thatSOX2 was not expressed in hair cells of vestibular andauditory organs, but it persisted at high levels in the layerof supporting cells. This was clearly distinguishable by theordered position of nuclei and the negative staining forTuj1. Further confirmation of this pattern of expressionwas obtained by double labelling E17 otocysts with SOX2and Islet1. The latter is typically expressed in supportingcells at this stage (Fig. 6g,h; Li et al., 2004). This is seen athigh resolution in Figure 6i–k. Islet1 (green) was coex-pressed with SOX2 in the supporting cells of macula utric-uli (Fig. 6i) and in the basilar papilla (Fig. 6j,k). Note alsothat hair cells in the macula (Fig. 6i) still expressed Islet1,whereas it was down-regulated in the papilla (Fig. 6j,k;see also Li et al., 2004). SOX2 was absent from hair cellsin both sensory organs.

Examination of E17 otocysts with SOX2 and PCNArevealed that some, but not all, of the SOX2-positive sup-porting cells were actively proliferating. As shown in Fig-ure 6l,m, SOX2 labelling concentrated in the ordered sup-porting cell layer, which was spotted with PCNA-positivecells (yellow-green, arrows). At higher magnification,double-labelled cells were clearly visible in the supportingcell layer. Some of them are indicated with arrows in theexamples shown in Figure 6o–q.

SOX2 but not SOX3 is expressed inprosensory patches in E5

Sox3 is known to be expressed in the neurogenic pla-codes (Abu-Elmagd et al., 2001), and we had some evi-dence of its early role in the specification of the proneuraldomain of the otic placode (Abello et al., 2007). We wereinterested in exploring the possibility of SOX2 and SOX3playing specific roles in proneural commitment by com-paring their expression in the proneural domain of theearly otic cup and in the sensory patches of E5 otocysts.Figure 7a–d shows the expression of SOX3 in the otic cup


492 J. NEVES ET AL.

Fig. 3. Expression of SOX2 and MyoVIIa in E5 embryos and SOX2expression in Schwann cells. Alternate coronal sections of an E5otocyst were immunostained for SOX2 (red) and Tuj1 (green; a–c) orMyoVIIa (red) and Tuj1 (green; d–f), respectively. SOX2 and MyoVIIastaining in vestibular prosensory patches, in a dorsal-ventral se-quence: posterior crista and superior crista (a,d), lateral crista (b,e),macula utriculi and macula sacculi (c,f). High magnifications of supe-rior crista show MyoVIIa (g) and SOX2 (h) in alternate sections. Notethat SOX2 was expressed in some cells corresponding to the hair cell

(HC) layer but not in the most apical ones (arrowheads). SOX2 ex-pression in Schwann cells of the cochleovestibular ganglion (i,j). SOX2was expressed in small nuclei typical of Schwann cells and not in largenuclei of Tuj1-positive neurons. pc, Posterior crista; sc, superior cris-ta; lc, lateral crista; mu, macula utriculi; ms, macula sacculi; vg,vestibular ganglion. Brackets in g,h indicate the position of the row ofhair cells. A, anterior; M, medial. Scale bars � 100 �m in a (applies toa–f); 20 �m in g (applies to g–i); 10 �m in j.



of an E2 (HH14) embryo. Similarly to SOX2, SOX3 wasexpressed in the proneural domain of the otic cup (Fig.7a,b), in a manner complementary to HNK1 (a–c). Figure7e–j shows the expression of SOX3 in an E5 embryo.Parallel, equivalent sections were analyzed for SOX2 andSOX3 expression in vestibular (Fig. 7e,h) and auditory(Fig. 7f,i) domains. No SOX3 immunoreactivity was de-tected in the otocyst, either in the vestibular or in theauditory domains. The positive control of the SOX3 was itsstrong positive reaction in the ventricular zone of theneural tube (Fig. 7j), which was very similar to that ofSOX2 (Fig. 7g). A similar difference in the expressionpattern between Sox2 and Sox3 was observed when com-paring the mRNA expression of both genes at E5. This isshown in Figure 7k–n. In situ hybridization is shown as ablue precipitate and immmunostaining Tuj1 in brown.Sox2 was expressed in the basilar papilla, macula utriculi,and lateral crista (Fig. 7k), whereas Sox3 expression wasabsent from the otocyst (Fig. 7l). A detail of the differentialexpression in the papilla is shown in Figure 7m,n. Sox2and Sox3 were strongly coexpressed in the ventricularzone of the hindbrain (Fig. 7k,l). These results suggest thepossibility that the neurogenic potential of otic progeni-tors may require the expression of both SOX2 and SOX3,whereas the prosensory potential would be specificallyassociated with that of SOX2 (Uchikawa et al., 1999).

DISCUSSION

SOX2 and the state of commitment ofsensory progenitors

SOX2 is a HMG box transcription factor that is ex-pressed in multipotent neural stem cells at all stagesduring mouse ontogeny (Wegner and Stolt, 2005). Sox2expression in the early embryonic CNS is pieced togetherby separate enhancers with distinct spatiotemporal spec-ificities, and enhancers for Sox2 expression in the lens andnasal/otic placodes have been identified (Uchikawa et al.,2003). SOX2 belongs to the stem-cell cassette that main-tains the self-renewal state and pluripotency of progeni-tors (Takahashi and Yamanaka, 2006). Sox1–3 interactwith various partner transcription factors and participatein defining distinct cell states that depend on the partnerfactors, Pax6 for lens differentiation, Oct3/4 for establish-ing the epiblast/ES cell state, and Brn2 for the neuralprimordia. All three factors are coexpressed in proliferat-ing neural progenitors of the embryonic and adult CNS.The SoxB2 subgroup of Sox factors, including Sox14 andSox21, are very similar to SoxB1 in their HMG-DNA bind-ing domain but act as transrepression domains. A keycommon feature of SoxB1, SoxB2, and SoxE, however, istheir ability to maintain neural progenitor or stem cellidentity (for review see Wegner and Stolt, 2005).

Fig. 4. SOX2, Islet1 and PCNA expression in developing inner earof the E5 embryo. Coronal sections a–c were immunostained forSOX2 (red) and Islet1 (green), and sagittal sections (d,e) were immu-nostained for SOX2 (red) and PCNA (green). DAPI staining (blue) isshown in a and d. Vestibular (a) and auditory (b) sensory patches of anE5 embryo otocyst labelled with SOX2 and Islet1. SOX2-positivedomains overlapped with Islet1. In c, a high-magnification detail ofthe basilar papilla shown in b, with nuclei labelled only with Islet1(green), with both Islet1 and SOX2 (yellow), and only with SOX2 (red).Vestibular (d) and auditory (e) sensory patches of an E5 embryo

otocyst labelled with SOX2 and PCNA. PCNA expression was widelyspread through the otocyst, and SOX2 labelling was mostly overlap-ping PCNA-positive cells at their restricted domains. f Shows a high-magnification detail of the basilar papilla represented in e, where thehigh number of yellow nuclei can be observed, showing that mostSOX2-positive cells were also PCNA positive. pc, Posterior crista; mu,macula utriculi; ms, macula sacculi; bp, basilar papilla. Orientation:D, dorsal; M, medial; A, anterior. Scale bars � 100 �m in a (applies toa,b,d,e); 50 �m in c (applies to c,f).


494 J. NEVES ET AL.

As discussed by Fritzsch et al. (2006), the vertebratesensory organ requires a mechanism for rapidly expand-ing the basic sensory unit, so that placodal epithelial cellsbear characteristics of stem cells. Therefore, it is expectedthat they express typical genes of the stem cell cassette(Takahashi and Yamanaka, 2006). Our experiments showthat the expression of SOX2 in the ear is reminiscent ofthis general stem-cell function but restricted to neuralcommitted progenitors. Early in development, during oticvesicle stages, SOX2 is found in proliferating cells but onlywithin the proneural domain of the early otocyst. Theproneural domain of the otic vesicle is associated withneuron generation and is characterized by the expressionof proneural genes such as Ngn1, NeuroD, and NeuroM,and Notch signalling elements such as Dl1, Hes5, andLfng (Adam et al., 1998; Cole et al., 2000; Alsina et al.,2004). Null mutations of Ngn1 and NeuroD in mice showthat these genes are necessary for neuronal determinationand differentiation (Ma et al., 2000; Kim et al., 2001;Matei et al., 2005). Cell communication signals are also

expressed in this domain, including several FGFs, whichhave been shown to be required for neuronal production(Alsina et al., 2004; Sanchez-Calderon et al., 2004, 2005).Other genes are expressed in the complementary posteriorand dorsal domains of the otic cup and vesicle (Alsina etal., 2004; Abello et al., 2006). Our results show that SOX2expression is temporally and spatially restricted to thedomain expected for the activity of proliferating cells thatare committed to generate neurons and sensory cells.

The transition between the proneural domain, which isclearly defined at the otic vesicle stage (E2–3), and theprosensory patches that are identifiable at E4 has not yetbeen resolved unambiguously. This occurs between E3and E4, and it is difficult to assess whether it is the resultof the development of a common domain or the result ofthe emergence of different, perhaps overlapping, indepen-dent prosensory patches. Some genes expressed in theneurogenic domain, such as Lfng and FGF10, persist inthe prospective sensory domains during the stages of sen-sory organ development (Cole et al., 2000; Pujades et al.,

Fig. 5. SOX2 expression in the sensory patches of the E8 embryo.Diagram representing an otocyst from E8 chicken embryo (a). Sensorypatches are represented in red (adapted from Sanchez-Calderon et al.2005). Sections (b–e) were immunostained for SOX2 (red), Tuj1(green), and DAPI (blue). b–e: Serial coronal sections of an E8 otocystare shown in a dorsal-ventral sequence. SOX2 expression was de-tected in vestibular (b–d) and auditory (e) sensory organs (pink fordouble staining with SOX2 and DAPI). SOX2-labelled cells in poste-rior crista (b), macula utriculi (c), macula sacculi (d), and basilar

papilla (e). High magnification of macula saculi (f). In this sensorypatch, hair cells, indicated with arrows, did not express SOX2, whichwas detected in supporting cells (pink nuclei). High magnification ofbasilar papilla shown in e, illustrating SOX2 staining in supportingcells (g). pc, Posterior crista; mu, macula utriculi; ms, macula sacculi;bp, basilar papilla; cg, cochlear ganglion. Orientation in a: D, dorsal;M, medial; A, anterior. Orientation in b applies to b–e: M, medial; A,anterior. Scale bars � 100 �m in b (applies to b–e); 50 �m in f (appliesto f,g).



Fig. 6. SOX2 expression is detected at postdifferentiation stages insupporting cells of E17 sensory organs. Sections (a–f) were immuno-stained for SOX2 (red) and Tuj1 (green). Sections (g–k) were immu-nostained for SOX2 (red) and Islet1 (green) and sections (l–q) wereimmunostained for SOX2 (red) and PCNA (green). DAPI staining isshown in (b,e,g,h,j,l,m) and (p). Transverse sections of E17 otocystsshowing vestibular (a,c,d) and auditory (f) sensory organs. SOX2-labelled cells in posterior crista (a), macula utriculi (c), macula sacculi(d), and basilar papilla (f). High-magnification details of posteriorcrista and basilar papilla (b,e). SOX2 was restricted to the supportingcells (red nuclei) and it was excluded from the hair cell layer (bluenuclei), that was also labelled with Tuj1. Vestibular (g, macula utric-uli) and auditory (h, basilar papilla) sensory patches of an E17 embryootocyst doubly labelled for SOX2 and Islet1. High magnification ofmacula utricculi (i) and basilar papilla (j,k). SOX2 expression oc-

curred in the supporting cell row. Islet1 expression in supporting cellsin vestibular (i) and auditory (j,k) sensory organs and in hair cells ofthe macula utriculi (i). Macula utricculi (l) and cristae (m) organs ofan E17 embryo otocyst labelled with SOX2 and PCNA. o–q: High-magnification detail of macula utriculi (o) and crista (p–q). SOX2expression was restricted to the supporting cell layer. Proliferatingcells were detected with PCNA (yellow nuclei with arrows). No PCNA-positive cells were detected in the hair cell layer, with the exception ofone cell in the macula utriculi (l,o, green nucleus). pc, Posterior crista;mu, macula utriculi; ms, macula sacculi; bp, basilar papilla; ml, mac-ula lagina; HC, hair cells, SC, supporting cells. Brackets in b,e,i,k,o,qindicate rows of hair and supporting cells. Scale bars � 100 �m in a(applies to a,c,d,f); 20 �m in b (applies to b,e,i–k,o–q); 50 �m in g(applies to g,h,l,m).


496 J. NEVES ET AL.

Fig. 7. SOX2 and SOX3 expression during ear development. SOX3expression in the E2 otic cup. Double immunostaining of parasagittalsections of E2, HH14, otic cup for SOX3 protein, red, and the HNK1epitope (green; a). SOX3 expression was detected in the nuclei of thecells located in the anterior domain (b,d) and HNK1 expression in acomplementary fashion in the posterior domain (c). Sections wereoriented with anterior to the right and dorsal to the top. SOX2 andSOX3 expression in sensory organs of E5 embryos (e–j). Coronalsections of an E5 otocyst, processed for SOX2 expression (red) and

Tuj1 (green; e–g). SOX2 expression in the macula sacculi and maculautriculi (e), basilar papilla (f), and ventricular zone of the neural tube(g). Corresponding alternate sections, but processed for SOX3 expres-sion (h–j). Macula sacculi and macula utriculi (h), basilar papilla (i),and neural tube (j). In situ hybridization for mRNA of Sox2 and Sox3in sensory organs of E5 embryos (k–n); sections were counterstainedfor Tuj1. mu, Macula utriculi; ms, macula sacculi; bp, basilar papilla;nt, neural tube. D, dorsal; M, medial; A, anterior. Scale bars � 50 �min a (applies to a–d); 100 �m in e (applies to e–n).



2006). Other genes, such as BMP4, and neurotrophin fac-tors foreshadow the sensory domains (Wu et al., 1998;Farinas et al., 2001; Pirvola and Ylikoski, 2003), but it isnot until E4 when Atho1 (Cath1)-positive cells are clearlyidentifiable at the sensory patches (Pujades et al., 2006).Serrate1 is probably accompanying the prospective sen-sory domain from very initial steps of specification on-ward, and it has been recently shown to be required forsensory development (Cole et al., 2000; Daudet and Lewis,2005). Because SOX2 is expressed in both early neuro-genic and prosensory domains, it appears that SOX2 maybe used by all proneural-neurogenic and prosensory pro-genitors as a common requirement for neural commit-ment. Otic neurons and hair cells are neural cell types ina strict sense and both are specified upon activation ofproneural basic helix-loop-helix (bHLH) genes into neu-rons by NeuroD or hair cells by Atho1 (Alsina et al., 2004).In the neural tube, SOX2 is responsible for proneuralcompetence, but it maintains repression of the activity ofproneural genes until the expression of the SoxB2 genegroup, which counteracts this effect (Bylund et al., 2003).SOX2 may play a similar function in the ear and maintainthe state of proneural commitment and the capacity forself-renewal, its activity being down-regulated upon ter-minal division. This would be consistent with the obser-vation that SOX2 is expressed not in neurons and haircells but in supporting cells.

Identification of otic proneural progenitors

Otic progenitors undergo several cycles of cell divisionand progressive specialization to generate committedcells, first neurons and then sensory cells. It has beenshown that neurons and sensory cells can derive from acommon progenitor cell that resides in the otic placode(Fekete et al., 1998; Satoh and Fekete, 2005). Cell progen-itors may change their properties because of intrinsicmechanisms related to the rounds of cell divisions, be-cause they are exposed to different signals, or because ofboth. On the other hand, it may be that an initially com-mon proneural progenitor located within the proneuraldomain of the otic placode becomes further specified by itsposition within the proneural domain. Studies in the ver-tebrate retina indicate that progenitors undergo a seriesof competence states, progressively changing their respon-siveness to instructive extracellular signals that influencethe type of cells that they will become (Livesey and Cepko,2001; Cayouette et al., 2003). Little is known about howthese mechanisms operate in the determination of oticlineages. It has been suggested that coevolutionary his-tory of eyes and ears shows conserved cellular develop-mental programs (Fritzsch et al., 2006).

Two allelic mouse mutants, light coat and circling (Lcc)and yellow submarine (Ysb), show hearing and balanceimpairment. These phenotypes are due to the absence (inLcc mutants) or reduced expression (in Ysb mutants) ofthe transcription factor SOX2, pointing toward a require-ment for SOX2 in sensory progenitor development (Kier-nan et al., 2005). Our results are consistent with thisnotion, because they show that SOX2 may be traced backto early stages of sensory progenitor specification. How-ever, both SOX2 and SOX3 are expressed during the gen-eration of neurons, but only SOX2 remains during sensoryorgan formation. This suggests the possibility that, at agiven stage of development, SOX3 expression would beswitched off, and the persistent SOX2 expression would

result in sensory cell generation. The possibility of a phe-notypic switch of cycling neural progenitors from neuronto hair cell fate has been suggested to occur in the Ngn1null mice (Matei et al., 2005).

The results also show the expression of SOX2 in theSchwann glial cells of the cochleovestibular ganglion. Ex-pression of Sox2 is down-regulated in the neural platewhen neural crest segregates from dorsal neural tube andremains low during crest cell migration. Sox2 expressionis subsequently up-regulated in some crest-derived cells inthe developing peripheral nervous system and is laterrestricted to glial sublineages. This is interesting in con-nection to the neural crest origin of the glial cells of thecochleovestibular ganglion as described by (D’Amico-Martel and Noden, 1983; Matei et al., 2005). SOX2 expres-sion in Schwann cells of the auditory and vestibular gan-glion may then reflect its neural crest origin.

Significance of SOX2 expression indifferentiation and postdifferentiation

stages

In addition to the mouse mutants defective for SOX2,which show hearing and balance impairment, a nonsensemutation of the Sox2 gene has been recently reported inone patient suffering from bilateral clinical anophthalmia,absence of optic pathways, and sensorineural deficit(Matei et al., 2005; Hagstrom et al., 2005). This illustratesthe potential importance of Sox2 in human developmentand disease. The persistent expression of SOX2 in thesupporting cells of the sensory organs during postdiffer-entiation stages may be important in connection to theregenerative properties of the avian inner ear. This is awell known ability in birds and other vertebrates thatseems to rely on the capacity of supporting cells to reacti-vate cell division. Damaged cells reactivate cell divisionand a dormant developmental program that results in thegeneration of a cluster of Atho1-positive cells and thesingling out of hair and supporting cells via the Delta-Notch pathway (Cotanche et al., 1994; Stone and Rubel,2000; Bermingham-McDonogh and Rubel, 2003). Our datashow that SOX2 is maintained in differentiated support-ing cells, which may suggest a mechanism to enable themto reenter the cell cycle in case of damage.

In conclusion, the present study illustrates the patternof expression of SOX2 protein throughout development ofthe inner ear. We show that SOX2 is initially expressedalong with SOX3 in the neurogenic domain. As otic devel-opment proceeds, SOX3 is lost, and SOX2 is maintained inprogenitor cells of the prosensory region. At postdifferen-tiation stages, SOX2 withdraws from hair cells and re-mains expressed in supporting cells from the sensorypatches. Our data suggest that SOX2 may have a dualcritical function in ear development, first, conferring theproneural potential to otic progenitors and, second, retain-ing the regenerative capacity of sensory supporting cells.

ACKNOWLEDGMENTS

We are grateful to Thomas Schimmang, MatıasHidalgo-Sanchez, Cristina Pujades, and Isabel Varela-Nieto for comments on the manuscript. We thank MartaLinares for excellent technical assistance. Islet-1 monoclo-nal antibody was obtained from the developmental Stud-ies Hybridoma Bank under the auspices of the National


498 J. NEVES ET AL.

Institute of Child Health and Human Development andmaintained by the University of Iowa Department of Bio-logical Sciences (Iowa City, IA). The anti-SOX3 antibodywas a kind gift of Thomas Edlund and Jonas Muhr. Probesfor the chick Sox2 and Sox3 genes were kindly provided byPaul Scotting, University of Nottingham.

LITERATURE CITED

Abello G, Khatri S, Giraldez F, Alsina B. 2007. Early regionalisation of theotic placode and its regulation by the notch signaling pathway. Mech-anisms of Development doi: 10.1016/j.mod.2007.04.002 (in press).

Abo T, Balch CM. 1981. A differentiation antigen of human NK and K cellsidentified by a monoclonal antibody (HNK-1). J Immunol 127:1024–1029.

Abu-Elmagd M, Ishii Y, Cheung M, Rex M, Le Rouedec D, Scotting PJ.2001. cSox3 expression and neurogenesis in the epibranchial placodes.Dev Biol 237:258–269.

Adam J, Myat A, Le Roux I, Eddison M, Henrique D, Ish-Horowicz D,Lewis J. 1998. Cell fate choices and the expression of Notch, Delta andSerrate homologues in the chick inner ear: parallels with Drosophilasense-organ development. Development 125:4645–4654.

Alsina B, Giraldez F, Varela-Nieto I. 2003. Growth factors and earlydevelopment of otic neurons: interactions between intrinsic and extrin-sic signals. Curr Top Dev Biol 57:177–206.

Alsina B, Abello G, Ulloa E, Henrique D, Pujades C, Giraldez F. 2004. FGFsignaling is required for determination of otic neuroblasts in the chickembryo. Dev Biol 267:119–134.

Barald KF, Kelley MW. 2004. From placode to polarization: new tunes ininner ear development. Development 131:4119–4130.

Bermingham NA, Hassan BA, Price SD, Vollrath MA, Ben Arie N, EatockRA, Bellen HJ, Lysakowski A, Zoghbi HY. 1999. Math1: an essentialgene for the generation of inner ear hair cells. Science 284:1837–1841.

Bermingham-McDonogh O, Rubel EW. 2003. Hair cell regeneration: wing-ing our way towards a sound future. Curr Opin Neurobiol 13:119–126.

Bylund M, Andersson E, Novitch BG, Muhr J. 2003. Vertebrate neurogen-esis is counteracted by Sox1–3 activity. Nat Neurosci 6:1162–1168.

Cayouette M, Barres BA, Raff M. 2003. Importance of intrinsic mecha-nisms in cell fate decisions in the developing rat retina. Neuron 40:897–904.

Cole LK, Le Roux I, Nunes F, Laufer E, Lewis J, Wu DK. 2000. Sensoryorgan generation in the chicken inner ear: contributions of bone mor-phogenetic protein 4, serrate1, and lunatic fringe. J Comp Neurol424:509–520.

Cotanche DA, Lee KH, Stone JS, Picard DA. 1994. Hair cell regenerationin the bird cochlea following noise damage or ototoxic drug damage.Anat Embryol 189:1–18.

D’Amico-Martel A. 1982. Temporal patterns of neurogenesis in avian cra-nial sensory and autonomic ganglia. Am J Anat 163:351–372.

D’Amico-Martel A, Noden DM. 1983. Contributions of placodal and neuralcrest cells to avian cranial peripheral ganglia. Am J Anat 166:445–468.

Daudet N, Lewis J. 2005. Two contrasting roles for Notch activity in chickinner ear development: specification of prosensory patches and lateralinhibition of hair-cell differentiation. Development 132:541–551.

Doetzlhofer A, White P, Lee YS, Groves A, Segil N. 2006. Prospectiveidentification and purification of hair cell and supporting cell progeni-tors from the embryonic cochlea. Brain Res (in press).

Farinas I, Jones KR, Tessarollo L, Vigers AJ, Huang E, Kirstein M, deCaprona DC, Coppola V, Backus C, Reichardt LF, Fritzsch B. 2001.Spatial shaping of cochlear innervation by temporally regulated neu-rotrophin expression. J Neurosci 21:6170–6180.

Fekete DM, Muthukumar S, Karagogeos D. 1998. Hair cells and support-ing cells share a common progenitor in the avian inner ear. J Neurosci18:7811–7821.

Fritzsch B, Pauley S, Beisel KW. 2006. Cells, molecules and morphogene-sis: the making of the vertebrate ear. Brain Res (in press).

Graham V, Khudyakov J, Ellis P, Pevny L. 2003. SOX2 functions tomaintain neural progenitor identity. Neuron 39:749–765.

Hagstrom SA, Pauer GJ, Reid J, Simpson E, Crowe S, Maumenee IH,Traboulsi EI. 2005. SOX2 mutation causes anophthalmia, hearing loss,and brain anomalies. Am J Med Genet A138:95–98.

Hamburger V, Hamilton HL. 1992. A series of normal stages in the devel-opment of the chick embryo. 1951. Dev Dyn 195:231–272.

Hasson T, Heintzelman MB, Santos-Sacchi J, Corey DP, Mooseker MS.1995. Expression in cochlea and retina of myosin VIIa, the gene prod-uct defective in Usher syndrome type 1B. Proc Natl Acad Sci U S A92:9815–9819.

Kiernan AE, Pelling AL, Leung KK, Tang AS, Bell DM, Tease C, Lovell-Badge R, Steel KP, Cheah KS. 2005. Sox2 is required for sensory organdevelopment in the mammalian inner ear. Nature 434:1031–1035.

Kim WY, Fritzsch B, Serls A, Bakel LA, Huang EJ, Reichardt LF, BarthDS, Lee JE. 2001. NeuroD-null mice are deaf due to a severe loss of theinner ear sensory neurons during development. Development 128:417–426.

Lang H, Fekete DM. 2001. Lineage analysis in the chicken inner ear showsdifferences in clonal dispersion for epithelial, neuronal, and mesenchy-mal cells. Dev Biol 234:120–137.

Lee MK, Tuttle JB, Rebhun LI, Cleveland DW, Frankfurter A. 1990. Theexpression and posttranslational modification of a neuron-specific beta-tubulin isotype during chick embryogenesis. Cell Motil Cytoskeleton17:118–132.

Lewis J. 1998. Notch signalling and the control of cell fate choices invertebrates. Semin Cell Dev Biol 9:583–589.

Li H, Liu H, Sage C, Huang M, Chen ZY, Heller S. 2004. Islet-1 expressionin the developing chicken inner ear. J Comp Neurol 477:1–10.

Livesey FJ, Cepko CL. 2001. Vertebrate neural cell-fate determination:lessons from the retina. Nat Rev Neurosci 2:109–118.

Ma Q, Anderson DJ, Fritzsch B. 2000. Neurogenin 1 null mutant earsdevelop fewer, morphologically normal hair cells in smaller sensoryepithelia devoid of innervation. J Assoc Res Otolaryngol 1:129–143.

Matei V, Pauley S, Kaing S, Rowitch D, Beisel KW, Morris K, Feng F,Jones K, Lee J, Fritzsch B. 2005. Smaller inner ear sensory epithelia inNeurog 1 null mice are related to earlier hair cell cycle exit. Dev Dyn234:633–650.

Molea D, Stone JS, Rubel EW. 1999. Class III beta-tubulin expression insensory and nonsensory regions of the developing avian inner ear.J Comp Neurol 406:183–198.

Nieto MA, Patel K, Wilkinson DG. 1996. In situ hybridization analysis ofchick embryos in whole mount and tissue sections. Methods Cell Biol51:219–235.

Pevny L, Placzek M. 2005. SOX genes and neural progenitor identity. CurrOpin Neurobiol 15:7–13.

Pirvola U, Ylikoski J. 2003. Neurotrophic factors during inner ear devel-opment. Curr Top Dev Biol 57:207–223.

Pujades C, Kamaid A, Alsina B, Giraldez F. 2006. BMP-signaling regulatesthe generation of hair-cells. Dev Biol 292:55–67.

Rex M, Orme A, Uwanogho D, Tointon K, Wigmore PM, Sharpe PT,Scotting PJ. 1997. Dynamic expression of chicken Sox2 and Sox3 genesin ectoderm induced to form neural tissue. Dev Dyn 209:323–332.

Sahly I, El Amraoui A, Abitbol M, Petit C, Dufier JL. 1997. Expression ofmyosin VIIA during mouse embryogenesis. Anat Embryol 196:159–170.

Sanchez-Calderon H, Martin-Partido G, Hidalgo-Sanchez M. 2004. Otx2,Gbx2, and Fgf8 expression patterns in the chick developing inner earand their possible roles in otic specification and early innervation. GeneExpress Patterns 4:659–669.

Sanchez-Calderon H, Martin-Partido G, Hidalgo-Sanchez M. 2005. Pax2expression patterns in the developing chick inner ear. Gene ExpressPatterns 5:763–773.

Satoh T, Fekete DM. 2005. Clonal analysis of the relationships betweenmechanosensory cells and the neurons that innervate them in thechicken ear. Development 132:1687–1697.

Savare J, Bonneaud N, Girard F. 2005. SUMO represses transcriptionalactivity of the Drosophila SoxNeuro and human Sox3 central nervoussystem-specific transcription factors. Mol Biol Cell 16:2660–2669.

Stone JS, Rubel EW. 2000. Cellular studies of auditory hair cell regener-ation in birds. Proc Natl Acad Sci U S A 97:11714–11721.

Stone JS, Leano SG, Baker LP, Rubel EW. 1996. Hair cell differentiationin chick cochlear epithelium after aminoglycoside toxicity: in vivo andin vitro observations. J Neurosci 16:6157–6174.

Takahashi K, Yamanaka S. 2006. Induction of pluripotent stem cells frommouse embryonic and adult fibroblast cultures by defined factors. Cell126:663–676.

Taranova OV, Magness ST, Fagan BM, Wu Y, Surzenko N, Hutton SR,Pevny LH. 2006. SOX2 is a dose-dependent regulator of retinal neuralprogenitor competence. Genes Dev 20:1187–1202.



Torres M, Giraldez F. 1998. The development of the vertebrate inner ear.Mech Dev 71:5–21.

Uchikawa M, Kamachi Y, Kondoh H. 1999. Two distinct subgroups ofGroup B Sox genes for transcriptional activators and repressors: theirexpression during embryonic organogenesis of the chicken. Mech Dev84:103–120.

Uchikawa M, Ishida Y, Takemoto T, Kamachi Y, Kondoh H. 2003. Func-tional analysis of chicken Sox2 enhancers highlights an array of di-verse regulatory elements that are conserved in mammals. Develop-mental Cell vol. 4, 509–519.

Waseem NH, Lane DP. 1990. Monoclonal antibody analysis of the prolif-erating cell nuclear antigen (PCNA). Structural conservation and thedetection of a nucleolar form. J Cell Sci 96:121–129.

Wegner M, Stolt CC. 2005. From stem cells to neurons and glia: a Soxist’sview of neural development. Trends Neurosci 28:583–588.

Whitehead MC, Morest DK. 1985. The development of innervation pat-terns in the avian cochlea. Neuroscience 14:255–276.

Wilson SI, Rydstrom A, Trimborn T, Willert K, Nusse R, Jessell TM,Edlund T. 2001. The status of Wnt signalling regulates neural andepidermal fates in the chick embryo. Nature 411:325–330.

Wolfrum U, Liu X, Schmitt A, Udovichenko IP, Williams DS. 1998. MyosinVIIa as a common component of cilia and microvilli. Cell Motil Cy-toskeleton 40:261–271.

Wu DK, Nunes FD, Choo D. 1998. Axial specification for sensory organs vs.non-sensory structures of the chicken inner ear. Development 125:11–20.


500 J. NEVES ET AL.

differential expression of sox2 and sox3 in neuronal and sensory progenitors of...

Documents