a RESEARCH ARTICLE

Dynamic Expression Pattern of SonicHedgehog in Developing Cochlear SpiralGanglion NeuronsZhiyong Liu,1,2 Thomas Owen,1,3 Lingli Zhang,1 and Jian Zuo1*

Sonic hedgehog (Shh) signaling plays important roles in the formation of the auditory epithelium.However, little is known about the detailed expression pattern of Shh and the cell sources from whichShh is secreted. By analyzing ShhCreEGFP/1 mice, we found that Shh was first expressed in all cochlear spi-ral ganglion neurons by embryonic day 13.5, after which its expression gradually decreased from base toapex. By postnatal day 0, it was not detected in any spiral ganglion neurons. Genetic cell fate mappingresults also confirmed that Shh was exclusively expressed in all spiral ganglion neurons and not in sur-rounding glia cells. The basal-to-apical wave of Shh decline strongly resembles that of hair cell differen-tiation, supporting the idea that Shh signaling inhibits hair cell differentiation. Furthermore, thisShhCreEGFP/1 mouse is a useful Cre line in which to delete floxed genes specifically in spiral ganglion neu-rons of the developing cochlea. Developmental Dynamics 239:1674–1683, 2010. VC 2010 Wiley-Liss, Inc.

Key words: Sonic Hedgehog (Shh); cochlear duct; hair cell; spiral ganglion neuron; differentiation

Accepted 21 March 2010

INTRODUCTION

The mouse auditory epithelium, theorgan of Corti, consists of mechano-sensory hair cells (HCs) and sur-rounding supporting cells (SCs). HCsand SCs extend along the coiled coch-lear duct that can be approximatelydivided into basal, middle, and apicalportions, with a total of �1.75 turns(Kelley, 2006). HCs and SCs arebelieved to share the same progenitorcells, which are located in the prosen-sory area of the cochlear duct duringembryonic development. These pro-genitor cells keep proliferating untilp27Kip1, a CIP/KIP family cell cycle

inhibitor, is turned on in an apical-to-

basal gradient (Chen and Segil, 1999;

Lowenheim et al., 1999; Kanzaki

et al., 2006; Lee et al., 2006). After

cell cycle exit, however, these progeni-

tors start differentiation into either

HCs or SCs in a basal-to-apical gradi-

ent. Previous studies have suggested

that Notch signaling (Lanford et al.,

1999; Zheng et al., 2000; Zine et al.,

2001; Kiernan et al., 2005a; Brooker

et al., 2006; Tang et al., 2006; Take-

bayashi et al., 2007; Li et al., 2008;

Doetzlhofer et al., 2009), fibroblast

growth factor receptor signaling (Col-

vin et al., 1996; Mueller et al., 2002;

Pirvola et al., 2002; Hayashi et al.,

2007, 2008; Puligilla et al., 2007), and

Sox2 protein (Kiernan et al., 2005b)

are necessary to promote the forma-

tion of the prosensory area.Cochlear spiral ganglion neurons

are derived from the neuroblasts

delaminating at embryonic day (E)

8.5 from the otic placode or otocyst, a

thickening ectoderm adjacent to the

hindbrain that gives rise to the

vestibular end organs and the cochlea

(Rubel and Fritzsch, 2002). In con-

trast, the glia cells are derived from

neural crest cells (Noden, 1986).

These neuroblasts become postmitotic

around E11.5 to E15.5 in a basal-to-

apical gradient (Ruben, 1967) and

gradually differentiate into either

Dev

elop

men

tal D

ynam

ics

1Department of Developmental Neurobiology, St. Jude Children’s Research Hospital, Memphis, Tennessee2Integrated Program In Biomedical Sciences, University of Tennessee Health Science Center, Memphis, Tennessee3University of Bath, United KingdomGrant sponsor: National Institutes of Health Grant numbers: DC06471; DC05168; DC008800; CA21765; Grant sponsor: Office of NavalResearch; Grant number: N000140911014; Grant sponsors: the American Lebanese Syrian Associated Charities (ALSAC) of St. JudeChildren’s Research Hospital; The Hartwell Individual Biomedical Research Award.*Correspondence to: Jian Zuo, Department of Developmental Neurobiology, St. Jude Children’s Research Hospital, 262Danny Thomas Place, Memphis, TN 38105. E-mail: jian.zuo@stjude.org

DOI 10.1002/dvdy.22302Published online 20 April 2010 in Wiley InterScience (www.interscience.wiley.com).

DEVELOPMENTAL DYNAMICS 239:1674–1683, 2010

VC 2010 Wiley-Liss, Inc.

type I or type II cochlear ganglion

neurons. The type I neurons inner-

vate the inner hair cells (IHCs), and

type II neurons innervate the outer

hair cells (OHCs) in the cochlear sen-

sory epithelium.Sonic hedgehog (Shh) signaling is a

highly conserved signaling pathwaythat plays critical roles in the devel-opment of a variety of organs acrossdifferent species (Wijgerde et al.,2002; Fuccillo et al., 2006). The impor-tance of Shh signaling in the develop-ment of the cochlea was highlightedby the complete absence of cochlea inthe Shh knockout mouse (Riccomagnoet al., 2002). Shh signaling also caninteract with Wnt signaling to pat-tern the developing inner ear (Boket al., 2005; Riccomagno et al., 2005).Basically, the dorsal part (the vestibu-lar end organ) of the inner earrequires less Shh signaling than theventral part (the cochlea). Such a dor-sal–ventral Shh signaling gradient iscontrolled by various Gli activatorsand repressors (Bok et al., 2007). Inaddition, Shh signaling is involved inregulating otic capsule chondrogene-sis (Liu et al., 2002). Because the com-plete absence of the cochlear duct inthe Shh knockout mouse precludesfurther analysis of the detailed rolesof Shh signaling in cochlear develop-ment, a mutant mouse model(Gli3D699/D699), in which Shh signalingis only partially lost, was analyzed(Bose et al., 2002; Driver et al., 2008).Although the cochlear ducts ofGli3D699/D699 mutant pups at post-natal day (P) 0 were approximatelyhalf the length of those of their wild-type littermates, the auditory epithe-lium was wider, with ectopic patchesof HCs in Kolliker’s organ, an areamedial to the auditory epitheliumthat normally does not have HCs.Gain-of-function and loss-of-functionanalyses revealed that Shh signalingrepresses the prosensory area forma-tion with the consequence of havingfewer HCs in the organ of Corti(Driver et al., 2008).

In situ hybridization analysisshowed that the cell sources whereShh is secreted are distributed in thecochlear spiral ganglion area (Driveret al., 2008). However, cochlear spiralganglion neurons and glia cells(Schwann cells) are mixed in thisarea. It is not clear whether either or

both of them express Shh. To date,the detailed expression pattern of Shhremains unknown. Here, we analyzedShhCreEGFP/þ knock-in mutant micein which the fusion protein Cre/EGFPis controlled by endogenous Shh pro-moter. There are two apparent advan-tages to using this ShhCreEGFP/þ

mouse line. First, with the high-qual-ity green fluorescent protein (GFP)antibody, enhanced GFP (EGFP) canbe used to recapitulate the endoge-nous Shh expression pattern. Second,by crossing the ShhCreEGFP/þ mousewith a Rosa26-EYFPloxp/þ reportermouse, we can perform in vivo geneticcell fate mapping, an approach widelyused for cell lineage analysis in nerv-ous systems (Joyner and Zervas,2006; Hatch et al., 2009). All cellsthat express Shh can be historicallyrecorded by the reporter gene EYFPin a Cre-mediated manner, irrespec-tive of their temporal and spatialexpression.

In this study, we found that Shhwas first expressed in all cochlear spi-ral ganglion neurons around E13.5,after which the expression graduallydeclined in a basal-to-apical wave andbecame undetectable at P0. Itstrongly resembles the basal-to-apicalgradient differentiation pattern of thecochlear prosensory progenitor cells,suggesting that Shh signaling con-trols the timing of when prosensoryprogenitor cells start their intrinsicdifferentiation program. Moreover,genetic cell fate mapping confirmedthat Shh expression was exclusive tospiral ganglion neurons, all of whichtransiently expressed Shh. Therefore,the ShhCreEGFP/þ mouse is an excel-lent genetic tool for studying func-tions of genes that are expressed incochlear spiral ganglion neurons.

RESULTS

Expression Pattern of Shh

Before E12.5

The mouse inner ear derives from athickened ectodermal region adjacentto the hindbrain, referred as oticplacode, at �E8.0. To show thedetailed Shh expression patternduring the development of the innerear, we analyzed the heterozygousShhCreEGFP/þ knock-in mutant mouse,in which the fusion protein Cre/EGFP

with a nuclear localization signal wasinserted to the endogenous Shh locus,resulting in a Shh-null allele (Harfeet al., 2004). Although homozygousShhCreEGFP/CreEGFP mice die asembryos, heterozygous mice are indis-tinguishable from their wild-type lit-termates and display no noticeablephenotypes (Harfe et al., 2004).Therefore, we used nuclear EGFPreporter to recapitulate the Shhexpression pattern. We found that theneurosensory progenitor cells, whichwere competent to develop into bothneurons and sensory cells, were Shh-negative at E8.75–E9 (Fig. 1A,B).Similarly, the NeuroD1-positivedelaminating neuroblasts at E10.5–E11 had not yet expressed Shh (Fig.1C). However, Shh was highlyexpressed in the notochord at bothages (Fig. 1A and inset in Fig. 1C),which served as internal positive con-trols. Additionally, we analyzed theE12.5 cochleae where the Sox2-posi-tive spiral ganglion neurons werefound to be Shh-negative (Fig. 2).

Expression Pattern of Shh at

E13.5

Although we did not find Shh-expressing cells by E12.5, Shh wasfound to be highly expressed in thespiral ganglion area at E13.5 (Fig. 3).Double immunostaining of EGFP andthe widely used neuronal markerTUJ1 confirmed the Shh-expressingcells were spiral ganglion neurons(Fig. 3). Furthermore, both trans-sec-tion (Fig. 3A) and whole-mount (Fig.3B) image analysis showed Shh inspiral ganglion neurons, irrespectiveof their locations. However, not all thespiral ganglion neurons were Shh-positive (Fig. 3D,F).

Dynamic Expression Pattern

of Shh at E15.5 and E16.5

Prosensory progenitor cells start dif-ferentiation into either HCs or SCsbetween E14.5 and E17.5 in a basal-to-apical gradient, with HCs appear-ing first in basal turns and last in api-cal turns. We analyzed cochleae atE15.5 and E16.5 to correlate the Shhexpression pattern to the differentia-tion pattern along the whole cochlearturns. Of interest, at E15.5, bothwhole-mount (Fig. 4) and trans-

Dev

elop

men

tal D

ynam

ics

SHH EXPRESSION IN COCHLEAR NEURONS 1675

section (Fig. 5) image analysis showedno Shh in the spiral ganglion neuronsin the entire basal turn and the mid-dle-basal turn, but it was expressed inthe apical–middle turn and the entire

apical turn. Under the same set-up ofthe Zeiss META 510 confocal micro-scope, the fluorescence signal washighest in neurons at the apex andlowest in the apical–middle turn.

Additionally, neurons with highEGFP expression were found in thecentral cochlear nerve area (which ispart of cranial nerve VIII; Fig. 5A–C).The central cochlear nerve area iswhere axons of the spiral ganglionneurons aggregate, project to thebrainstem, and synapse with cochlearnuclei, but that was not the focus ofthis study.At E16.5, differentiating HCs that

were myosin-VIIa-positive were iden-tified exclusively in the basal turns(Fig. 6A–D). Similar to what was seenat E15.5, Shh expression again wasfound exclusively in the spiral gan-glion neurons distributed in the coch-lear apical turns (Fig. 6E–H) and cellsin the central cochlear nerve (Fig.6A,E). Again, at both E15.5 andE16.5, not all the spiral ganglion neu-rons were Shh-positive (Figs. 5F, 6H).

Expression Pattern of Shh

at E17.5

Differentiating HCs and SCs with dis-tinct morphology and molecularmarkers were easily identified incochleae at E17.5. In contrast to whatwas seen at E16.5, myosin-VIIa-posi-tive HCs were found in all cochlearturns at E17.5 (Fig. 7A–D). The audi-tory epithelium and the greater epi-thelial ridge were marked with Sox2,a transcription factor indispensable tocochlear development (Kiernan et al.,2005b; Dabdoub et al., 2008). Whole-mount analysis showed Shh only inthe spiral ganglion neurons at themost apical cochlear ducts (Fig. 7F–H,L–O). In addition, Shh-expressingcells still were found in the centralcochlear nerve area (Fig. 7E,I–K).Taken together, these findingsshowed that Shh expression declinedin the late embryonic ages and wasnot detected in spiral ganglion neu-rons except those in the most apicalcochlear ducts.

Genetic Fate Mapping of

Shh-Expressing Cells

As discussed, during our analysis ofthe Shh expression pattern at eachembryonic age, we found that not allthe spiral ganglion neurons wereShh-positive. There are two possibil-ities we could speculate. One is thatthere indeed are some portions of

Fig. 1. Sonic hedgehog (Shh) expression pattern at embryonic day (E) 8.75–E9 and E10.5–E11.A–B: Whole embryo trans-section image at E8.75–E9. Shh was highly expressed in the notochord(arrows) but not in the otic pit. C: Whole-embryo trans-section image at E10.5–E11. Shh was notexpressed in the NeuroD1-positive neuroblasts delaminating from the otic vesicle. Insets wereShh-positive notochord cells. Green: enhanced green fluorescent protein (EGFP); Red, Sox2 (B)and NeuroD1 (C); blue, Hoechst 33342. hb, hindbrain; op, otic pit; ov, otic vesicle; nc, notochord;dn, delaminating neuroblast. Scale bars ¼ 100 mm in A,B, 20 mm in C, 10 mm inset of C.

Fig. 2. Sonic hedgehog (Shh) expression pattern at embryonic day (E) 12.5. A–C: Low-magnifi-cation image of the inner ear and surrounding central nervous tissues. Shh was expressed inthe diencephalon cells but not in cochlear spiral ganglion neurons. D,E: High-magnificationimages of the dotted rectangular area in C. Green, enhanced green fluorescent protein (EGFP);red, Sox2; blue, Hoechst 33342. cd, cochlear duct; sg, spiral ganglion; sa, sacculus. Scale bars¼ 500 mm in A–C, 50 mm D,E.

Dev

elop

men

tal D

ynam

ics

1676 LIU ET AL.

spiral ganglion neurons that neverexpress Shh during embryonic devel-opment. The other is that rapiddynamic changes in Shh expression

occur in all neurons such that differ-ent fractions were detected at differ-ent times. To clarify this, we tookadvantage of the fusion protein Cre/

EGFP in the heterozygous ShhCreEGFP/þ

mouse. Cre-mediated permanentreporter gene expression has beenwidely used for genetic fate mappinganalysis (Joyner and Zervas, 2006).We generated ShhCreEGFP/þ;Rosa26-EYFPloxp/þ mice for analysis andShhCreEGFP/þ;Rosa26-EYFPþ/þ litter-mates as controls. In the absence ofCre, EYFP expression is blocked bythe floxed stop sequences precedingthe EYFP coding frame. When thefloxed stop sequences are deleted byCre, EYFP is expressed permanentlyas long as the Rosa26 promoter is con-tinuously active (Fig. 8A). Followingthis scheme, all Shh-expressing cellsand their offspring that also expressCre can be traced by EYFP, irrespec-tive of their temporal and spatialexpression differences.We first analyzed the control

ShhCreEGFP/þ;Rosa26-EYFPþ/þ litter-mates at P0, and no fluorescent signalwas seen in the whole-mount analysis(Fig. 8B,C). This showed that Shhexpression was completely undetect-able at P0 in spiral ganglion neuronsin all cochlear turns, including themost apical. Then we analyzed theShhCreEGFP/þ;Rosa26-EYFPloxp/þ miceat P0. As expected, EYFP was highlyexpressed in almost 100% of the coch-lear spiral ganglion neurons (Fig. 8D–F), and lower expression was detectedin the peripheral cochlear nerve fiberssurrounding the HCs and SCs (Fig.8F). This clarified that all spiral gan-glion neurons had expressed Shhtransiently during embryonic devel-opment. Although GFP antibody canrecognize both EGFP and EYFP, weconcluded that all the fluorescencesignals were from EYFP (induced bythe Rosa26 promoter) not EGFP(driven by the Shh promoter) becauseno fluorescent signal was found inShhCreEGFP/þ;Rosa26-EYFPþ/þ controlcochleae. Moreover, an identicalEYFP expression pattern was foundin ShhCreEGFP/þ;Rosa26-EYFPloxp/þ

mice at P6 (Fig. 9A,B), demonstratingthat no additional cell typesexpressed Shh between P0 and P6.Furthermore, double immunostainingof EYFP and TUJ1 confirmed thatEYFP was expressed exclusively inthe TUJ1-positive spiral ganglionneurons (large round nuclei), whilenone of the surrounding glia cells(small nuclei) were EYFP-positive

Fig. 3. Shh expression pattern at embryonic day (E) 13.5. A,B: Trans-section (A) image of thewhole inner ear and whole-mount (B) image of the cochlear duct at E13.5. C–F: High-magnifica-tion images of the square area in A. D: The white dotted line circle is an example of an areawhere Shh was very low or undetectable. Red, TUJ1; green, enhanced green fluorescent protein(EGFP); blue, Hoechst 33342. cd, cochlear duct; sg, spiral ganglion; ut, utricle; sa, sacculus.Scale bars ¼ 100 mm in A, 200 mm in B, 50 mm in C–F.

Fig. 4. Whole-mount analysis of Sonic hedgehog (Shh) expression pattern at embryonic day,(E) 15.5. A–C0: Whole-mount image of cochlear basal (A,A0), middle (B,B0), and apical turns(C,C0). Shh was absent in spiral ganglion neurons in the basal and middle-basal turns but waspresent in those in the middle-apical (green part in B0) and apical turns. A–C: White dottedcurved lines label the approximate medial boundary of the prosensory area; yellow stars markthe end (toward the basal orientation) of each turn. D–F: High-magnification images of the apicalspiral ganglion neurons in the dotted rectangular area in C. Red, TUJ1; green, enhanced greenfluorescent protein (EGFP); blue, Hoechst 33342. Scale bars ¼ 500 mm in A–C, 25 mm in D–F.

Dev

elop

men

tal D

ynam

ics

SHH EXPRESSION IN COCHLEAR NEURONS 1677

(Fig. 9C–F). Additionally, we per-formed double staining of EYFP andSox10, a glia cell marker (Davies,2007; Puligilla et al., 2010), and wedid not find any cells that were dou-ble-positive for EYFP and Sox10 (Fig.10A–H), which confirmed that Shhwas expressed in neurons but not inglia cells.As discussed, neurons distributed

in the cochlear central nerve area alsoexpressed Shh. With cell lineage anal-ysis, we also found EYFP-positiveneurons in the cochlear central nervearea that were Sox10-negative (Fig.10I–L). In addition, we determinedthe onset of Cre activity in spiral gan-glion neurons in ShhCreEGFP/þ;Rosa26-LacZloxp/þ mice, which canhelp us to distinguish EGFP and thereporter gene beta-galactosidase. Wefound that �E15.5 was the earliesttime expression of beta-galactosidasecould be detected in TUJ1-positiveneurons (Fig. 11).

Fig. 5.

Fig. 6.

Fig. 6. Shh expression pattern at embryonicday (E) 16.5. A–H: Trans-section images ofthe cochlear part at E16.5. B–D: High-magnifi-cation images of each cochlear turn shown inA. Myosin-VIIa–positive hair cells (HCs)appeared at the basal (B) but not middle (C)or apical (D) turns. F–H: High-magnificationimages of the spiral ganglion area of eachcochlear turn shown in E. Shh was present inspiral ganglion neurons at the apical (H) butnot basal (F) or middle (G) turns of the coch-leae. White arrows in H are examples of spiralganglion neurons that were Shh-negative. A,E:Note that Shh was expressed in the centralcochlear nerve area as well (white dottedcircle area). Red, myosin-VIIa (A-D) and TUJ1(E-H); green, enhanced green fluorescent pro-tein (EGFP); blue, Hoechst 33342. cn, centralcochlear nerve (cranial nerve VIII); sg, spiralganglion. Scale bars ¼ 200 mm in A,E, 50 mmin B–D,F–H.

Fig. 5. Trans-section analysis of Sonichedgehog (Shh) expression pattern at embry-onic day (E) 15.5. A–C: Low-magnificationimage of the cochlea. C: Shh was expressedin the apical spiral ganglion neurons and neu-rons distributed in the central cochlear nervearea (white curved lines). D–F: High-magnifi-cation image of white rectangular area in C. F:Note that some apical spiral ganglion neuronswere Shh-negative (white curved lines). Red,TUJ1; green, enhanced green fluorescent pro-tein (EGFP); blue: Hoechst 33342. cn, centralcochlear nerve (cranial nerve VIII). Scale bars¼ 200 mm in A–C, 25 mm D–F.

Dev

elop

men

tal D

ynam

ics

1678 LIU ET AL.

Fig. 7.

Fig. 8.

Fig. 8. Sonic hedgehog (Shh) expression pat-tern at postnatal day (P) 0 and genetic cell fatemapping analysis. A: Illustration of genetic cellfate mapping analysis. B–E: Whole-mountimmunofluorescence of EYFP and myosin- VIIain control ShhCreEGFP/þ mice (B,C) andShhCreEGFP/þ; Rosa26-EYFPloxp/þ mice (D,E);yellow stars mark the end (toward the basal ori-entation) of each turn. F: High-magnificationimage (confocal projection of all layers in Zpanel) of the rectangular area in D. Red, myosin-VIIa; green, EYFP; blue, Hoechst 33342. oc,organ of Corti; sg, spiral ganglion. Scale bars ¼200 mm in B–E, 50 mm in F.

Fig. 7. Sonic hedgehog (Shh) expression pat-tern at embryonic day (E) 17.5. A: Immunofluo-rescence of Sox2 and myosin-VIIa on cochleartrans-sections. B–D: High-magnification imagesof the sensory epithelium in each cochlear turnshown in A. E: Immunofluorescence of TUJ1and enhanced green fluorescent protein (EGFP)on cochlear trans-sections. Closed white dottedlines mark each cochlear duct. F–H: Whole-mount TUJ1 and EGFP immunofluorescence ofcochlear basal (F), middle (G), and apical (H)turns; yellow stars mark the end (toward thebasal orientation) of each turn. I–K: High-magni-fication images of the white rectangular area(central cochlear nerve region) in E. Arrows inI–K mark a neuron (the same one in all images)that was Shh-negative. L–O: High-magnificationimages of the squared area in H. Arrows in L-Olabel a spiral ganglion neuron (the same one inall images) that was Shh-negative. Red, myosin-VIIa (A–D) and TUJ1 (E–O); green, Sox2 (A–D)and enhanced green fluorescent protein (EGFP)(E–O); blue, Hoechst 33342. sg, spiral ganglion;GER, greater epithelial ridge. Scale bars ¼ 200mm in A,E,F–H, 20 mm in B–D,I–O.

DISCUSSION

Shh Signaling and Cell Cycle

Exit of the Prosensory

Progenitors

Previous studies have shown thatcochlear prosensory progenitor cellsin apical turns exit the cell cycle firstat �E12.5, but those in the middleand basal turns continue proliferatinguntil �E14.5. This apical-to-basalgradient cell cycle exit is correlatedwith a similar apical-to-basal wave ofp27Kip1 expression between E12.5 andE14.5 (Lee et al., 2006). The factorsregulating p27Kip1 transcription arenot clear, but the transcription isthought to be repressed by the Notch-Hes1 pathway (Murata et al., 2009).Here, we wondered whether therewas a genetic interaction betweenShh signaling and p27Kip1. We foundthat Shh is present in spiral ganglionneurons at E13.5 along the entirecochlear duct (which has only 0.75�1.0 turn). Because only basal turnprogenitor cells are proliferating atE13.5, it seems to support the ideathat Shh signaling might not promoteproliferation of progenitor cells. How-ever, it is also possible that Shh sig-naling does promote their prolifera-tion in the basal turn cochlear duct,while that signaling is neutralized byother unknown signals present exclu-sively in the apical and middle coch-lear duct at E13.5. This possibility isindirectly supported by the fact thatShh signaling could promote prolifer-ation, in an Atoh1-dependent manner,of external granule cells in the cere-bellum in vivo and otic progenitorcells (isolated from E9 mouse otic ves-icle) in vitro (Zhao et al., 2006; Floraet al., 2009). The exact roles of Shhsignaling can be delineated by tar-geted deletion of Shh in cochlear spi-ral ganglion neurons (Doris Wu,NIDCD, personal communication).

Shh Signaling and

Differentiation of the

Cochlear Prosensory

Progenitors

While cochlear prosensory progenitorcells exit the cell cycle at an apical-to-basal gradient, they start differentia-tion at an opposite basal-to-apicalgradient. How this differentiation

Fig. 10. No Sonic hedgehog (Shh) expression in cochlear glia cells. A–D: Double staining ofEYFP and glia cell marker Sox10 in cochleae (basal turn) dissected from ShhCreEGFP/þ; Rosa26-EYFP loxp/þ mice at postnatal day (P) 0. E–H: High-magnification image of the white square areain B (#1). None of the Sox10-positive glia cells in the spiral ganglion area was EYFP-positive. I–L: High-magnification image of the white rectangular area in B (#2). None of the Sox10-positiveglia cells in the central cochlear nerve area was EYFP-positive. Red, Sox10; green, EYFP; blue,Hoechst33342. sg, spiral ganglion; oc, organ of Corti; cn, central cochlear nerve. Scale bars ¼500 mm in A–D, 20 mm E–L.

Fig. 9. Genetic cell fate mapping analysis at postnatal day (P) 6. A,B: Whole-mount immunofluo-rescence of EYFP and myosin-VIIa in ShhCreEGFP/þ; Rosa26-EYFP loxp/þ mice. C–F: Whole-mountdouble staining of EYFP and TUJ1. Arrows label a TUJ1-positive spiral ganglion neuron (big roundnucleus) that was traced by EYFP. White dotted circle in E marks a glia cell (small and irregular nu-cleus) that was TUJ1-negative and not traced by EYFP. Red, myosin-VIIa (A,B) and TUJ1 (C–F);green, EYFP; blue, Hoechst 33342. Scale bars ¼ 200 mm in A,B, 20 mm in C–F.

Dev

elop

men

tal D

ynam

ics

1680 LIU ET AL.

pattern occurs is still an enigma. Inthis study, we comprehensively ana-lyzed the dynamic Shh expression

pattern between E8.75 and P0. Itsdown-regulation in a basal-to-apicalwave matches the basal-to-apical gra-

dient differentiation pattern of thecochlear prosensory progenitor cells,highlighting the reverse correlationbetween Shh signaling and the intrin-sic differentiation program insideprosensory progenitor cells (see a pro-posed working model in Fig. 12).Moreover, this opposite correlation isfurther supported by the fact thatShh signaling blocked HC formationwhen E13 cochlear explants were cul-tured and treated 6 days in vitro withexogenous Shh protein (Driver et al.,2008). Furthermore, the earlier down-regulation of Shh in basal turn spiralganglion neurons is also supported bythe fact that basal turn cochlearexplants were less sensitive to Shhtreatment than middle and apicalturn explants (Driver et al., 2008).In addition, taking middle turn coch-

lear ducts at E16.5 as an example, Shhwas undetectable in spiral ganglionneurons. However, myosin-VIIa-posi-tive HCs were absent in the prosen-sory area, indicating the differentia-tion program had not been initiatedyet in the progenitor cells. This showsthat Shh signaling should be turnedoff before progenitors start to differen-tiate. Nonetheless, there is no directevidence that Shh signaling doesprevent premature differentiation of

Fig. 12.

Fig. 11.

Fig. 12. A model to summarize the correla-tion between Sonic hedgehog (Shh) decliningand hair cell (HC) differentiation. At embryonicday (E) 13.5, no HCs are present and Shh isexpressed in all spiral ganglion neurons. AtE16.5, HCs are present in the basal turn,where spiral ganglion neurons lose Shhexpression. At postnatal day (P) 0, HCs arepresent in all cochlear turns, and Shh is notdetected in any spiral ganglion neurons. Thiscorrelation supports the concept that Shhinhibits HC differentiation in cochleardevelopment.

Fig. 11. Efficient Cre activity in ShhCreEGFP/þ

mice at approximately embryonic day (E) 15.5.A–D: Low-magnification image of doublelabeling TUJ1 (A, D) and beta-galactosidase(B,D) in cochleae dissected from ShhCreEGFP/þ;Rosa26-LacZloxp/þ mice at �E15.5. D: The redline inside the cochlear duct (yellow arrows) isnon-specific staining. E–H: High-magnificationimages of the white squared area in D; arrowsmark a TUJ1-positive spiral ganglion neuronthat was labeled by reporter b-galactosidase.Red, b-galactosidase; green, TUJ1; blue,Hoechst33342. sg, spiral ganglion; cn, centralcochlear nerve (cranial nerve VIII). Scale bars¼ 200 mm in A–D, 20 mm in E–H.

Dev

elop

men

tal D

ynam

ics

SHH EXPRESSION IN COCHLEAR NEURONS 1681

progenitor cells. It would be strongsupportive evidence if Shh signalingcould be specifically blocked (such asby deleting the smoothened gene) inthe sensory epithelium of the cochlearapical turns between E14.5 and E16.5and progenitor cells were able to startpremature differentiation. Alterna-tively, the deletion of Shh in spiralganglion neurons can also test the cor-relation between Shh signaling andhair cell differentiation.

Potential Applications of

ShhCreEGFP/1 Cre Mouse Line

in Spiral Ganglion Neurons

Since its first use in studying limb pat-terning and development (Harfe et al.,2004), the ShhCreEGFP/þ Cre mouse linehas been widely used in a variety of bi-ological processes (http://jaxmice.jax.org/strain/005622.html). Our geneticmapping analysis has shown therobust Cre activity of the fusion proteinCre/EGFP, as almost 100% of thecochlear spiral ganglion neurons, butnot surrounding glia cells, were tracedby EYFP at P0 and P6 in theShhCreEGFP/þ; Rosa26-EYFPloxp/þ mice,and beta-galactosidase at �E15.5 inthe ShhCreEGFP/þ; Rosa26-LacZloxp/þ

mice. Given the availability of manymutant mouse lines in which the inter-ested genes are floxed, the efficient Creactivity allows ShhCreEGFP/þ to bewidely used to study functions of a va-riety of floxed genes in spiral ganglionneurons, especially for those whosecomplete loss results in embryonic fa-tality that precludes further analysis oftheir functions at later ages.

EXPERIMENTAL

PROCEDURES

Mouse Strain Maintenance

ShhCreEGFP/þ (stock number: 005622),Rosa26-EYFPloxp/þ (stock number:006148) and Rosa26-LacZloxp/þ mouse(stock number: 003310) strains werepurchased from The Jackson Labora-tory (Bar Harbor, ME). For genetic fatemapping analysis, male ShhCreEGFP/þ

were crossed with female Rosa26-EYFPloxp/þ to get the ShhCreEGFP/þ;Rosa26-EYFPloxp/þ mice. This breedingstrategy was recommended by TheJackson Laboratory. All animal workconducted during the course of this

study was approved by the Institu-tional Animal Care and Use Commit-tee at St. Jude Children’s ResearchHospital and was performed accordingto NIH guidelines.

Embryonic Age

Determination

Heterozygous male ShhCreEGFP/þ micewere crossed with female Shhþ/þ

mice, or male ShhCreEGFP/þ mice werecrossed with female Rosa26-LacZloxp/þ

mice, at 5 pm, and the vaginal plugwas checked at 7 AM the next day.Given the presence of plugs, thatmorning was designated as E0.5 forthe embryonic age. To guarantee theaccuracy of the embryonic age deter-mination and to avoid false-negativeoutcomes from plug checking, onlythe females that were plugged after 1night of breeding were used in ourcurrent study.

Histology and

Immunofluorescence

The whole embryos at various ageswere immersed in 4% paraformalde-hyde overnight at 4�C, after which ei-ther the whole embryos (before E12.5)were analyzed together or the innerear was carefully dissected out first(after E13.5). The inner ear was thenimmersed in 4% paraformaldehydefor 3 hr at 4�C. For trans-section anal-ysis, the whole embryo or inner earwas immersed in 30% sucrose over-night at 4�C and then embedded inOptimum Cutting Temperature com-pound, frozen in dry ice, and cut intosections 12 mm thick. For whole-mount analysis, the whole cochlearduct and corresponding medial spiralganglion tissues were divided intothree parts, the basal, middle, and ap-ical turns, with the exception that theE13.5 cochlear duct was maintainedas a whole.

Both whole mounts and trans-sec-tions were permeabilized and blockedat room temperature for 1 hr in solu-tions containing 1% bovine serum al-bumin and 1% Triton X-100 in 10 mMphosphate-buffered saline (PBS, pH7.4). Tissues were then incubated withprimary antibodies in blocking solu-tion (1% bovine serum albumin and0.1% Triton X-100 in 10 mM PBS)overnight at 4�C, followed by 3 washes

for 10 min each in 10 mM PBS. Then,tissues were incubated with secondaryantibodies in the same blocking solu-tion overnight at 4�C, followed by 3washes for 10 min each in 10 mM PBS.Tissues were then incubated for 30min at room temperature in Hoechst33342 in 10 mM PBS (Invitrogen,H3570, 1:1,000), followed by 3 washesfor 10 min each in 10 mM PBS, andfinally were mounted in ProLong Goldantifade reagent (Invitrogen, P36934).Samples were dried at room tempera-ture for at least 24 hr. All whole-mountsamples were analyzed with a ZeissMETA 510 confocal microscope. Trans-section samples at E17.5 were ana-lyzed with the Zeiss META 510 confo-cal microscope, and those at other em-bryonic ages were analyzed with ourOlympus BX60 fluorescence.The following primary antibodies

were used: anti-myosin-VIIa (rabbit,1:200, Proteus Bioscience, 25-6790),anti-GFP (chicken, 1:1,000, Abcam,ab13970), anti-Sox10 (goat, 1:250,Santa Cruz Biotechnology, sc-17342),anti-NeuroD1 (goat, 1:100, Santa CruzBiotechnology, sc-1084), anti-Sox2 (goat,1:1,000, Santa Cruz Biotechnology, sc-17320), anti–b-galactosidase (rabbit,1:500, ICN, 55976), and anti-TUJ1(mouse, 1:1,000, MMS-435P, Covance).The following secondary antibodiesfrom Invitrogen Company were used:donkey anti-rabbit Alexa Fluor 647(1:1,000, A-31573), donkey anti-goatAlexa Fluor 568 (1:1,000, A11057), goatanti-chicken Alexa Fluor 488 (1:1,000,A11039), goat anti-mouse Alexa Fluor568 (1:1,000, A11031), and goat anti-rabbit Alexa Fluor 568 (1:1,000,A11036).

ACKNOWLEDGMENTSWe thank J. Woods and D. Wash fordetermining the embryonic ages; S.Connell, L. Zhang, J. Peters, and Y.Ouyang for expertise in confocalimaging; and D. Wu at NIDCD for dis-cussion. J.Z. is a recipient of TheHartwell Individual BiomedicalResearch Award.

REFERENCES

Bok J, Bronner-Fraser M, Wu DK. 2005.Role of the hindbrain in dorsoventralbut not anteroposterior axial specifica-tion of the inner ear. Development 132:2115–2124.

Dev

elop

men

tal D

ynam

ics

1682 LIU ET AL.

Bok J, Dolson DK, Hill P, Ruther U,Epstein DJ, Wu DK. 2007. Opposinggradients of Gli repressor and activa-tors mediate Shh signaling alongthe dorsoventral axis of the inner ear.Development 134:1713–1722.

Bose J, Grotewold L, Ruther U. 2002.Pallister-Hall syndrome phenotype inmice mutant for Gli3. Hum Mol Genet11:1129–1135.

Brooker R, Hozumi K, Lewis J. 2006.Notch ligands with contrasting func-tions: Jagged1 and Delta1 in the mouseinner ear. Development 133:1277–1286.

Chen P, Segil N. 1999. p27(Kip1) linkscell proliferation to morphogenesis inthe developing organ of Corti. Develop-ment 126:1581–1590.

Colvin JS, Bohne BA, Harding GW, McE-wen DG, Ornitz DM. 1996. Skeletalovergrowth and deafness in mice lack-ing fibroblast growth factor receptor 3.Nat Genet 12:390–397.

Dabdoub A, Puligilla C, Jones JM, FritzschB, Cheah KS, Pevny LH, Kelley MW.2008. Sox2 signaling in prosensory do-main specification and subsequent haircell differentiation in the developing coch-lea. Proc Natl Acad Sci U S A 105:18396–18401.

Davies D. 2007. Temporal and spatial reg-ulation of alpha6 integrin expressionduring the development of the cochlear-vestibular ganglion. J Comp Neurol502:673–682.

Doetzlhofer A, Basch ML, Ohyama T,Gessler M, Groves AK, Segil N. 2009.Hey2 regulation by FGF provides aNotch-independent mechanism formaintaining pillar cell fate in the organof Corti. Dev Cell 16:58–69.

Driver EC, Pryor SP, Hill P, Turner J,Ruther U, Biesecker LG, Griffith AJ,Kelley MW. 2008. Hedgehog signalingregulates sensory cell formation and au-ditory function in mice and humans. JNeurosci 28:7350–7358.

Flora A, Klisch TJ, Schuster G, ZoghbiHY. 2009. Deletion of Atoh1 disruptsSonic Hedgehog signaling in the devel-oping cerebellum and prevents medullo-blastoma. Science 326:1424–1427.

Fuccillo M, Joyner AL, Fishell G. 2006.Morphogen to mitogen: the multipleroles of hedgehog signalling in verte-brate neural development. Nat RevNeurosci 7:772–783.

Harfe BD, Scherz PJ, Nissim S, Tian H,McMahon AP, Tabin CJ. 2004. Evidencefor an expansion-based temporal Shhgradient in specifying vertebrate digitidentities. Cell 118:517–528.

Hatch EP, Urness LD, Mansour SL. 2009.Fgf16(IRESCre) mice: a tool to inacti-vate genes expressed in inner ear cris-tae and spiral prominence epithelium.Dev Dyn 238:358–366.

Hayashi T, Cunningham D, Bermingham-McDonogh O. 2007. Loss of Fgfr3 leadsto excess hair cell development in the

mouse organ of Corti. Dev Dyn 236:525–533.

Hayashi T, Ray CA, Bermingham-McDonoghO. 2008. Fgf20 is required for sensory epi-thelial specification in the developing coch-lea. J Neurosci 28:5991–5999.

Joyner AL, Zervas M. 2006. Genetic in-ducible fate mapping in mouse: estab-lishing genetic lineages and defininggenetic neuroanatomy in the nervoussystem. Dev Dyn 235:2376–2385.

Kanzaki S, Beyer LA, Swiderski DL, Izu-mikawa M, Stover T, Kawamoto K,Raphael Y. 2006. p27(Kip1) deficiencycauses organ of Corti pathology andhearing loss. Hear Res 214:28–36.

Kelley MW. 2006. Regulation of cell fatein the sensory epithelia of the innerear. Nat Rev Neurosci 7:837–849.

Kiernan AE, Cordes R, Kopan R, GosslerA, Gridley T. 2005a. The Notch ligandsDLL1 and JAG2 act synergistically toregulate hair cell development in themammalian inner ear. Development132:4353–4362.

Kiernan AE, Pelling AL, Leung KK, TangAS, Bell DM, Tease C, Lovell-Badge R,Steel KP, Cheah KS. 2005b. Sox2 isrequired for sensory organ developmentin the mammalian inner ear. Nature434:1031–1035.

Lanford PJ, Lan Y, Jiang R, Lindsell C,Weinmaster G, Gridley T, Kelley MW.1999. Notch signalling pathway medi-ates hair cell development in mamma-lian cochlea. Nat Genet 21:289–292.

Lee YS, Liu F, Segil N. 2006. A morphoge-netic wave of p27Kip1 transcriptiondirects cell cycle exit during organ ofCorti development. Development 133:2817–2826.

Li S, Mark S, Radde-Gallwitz K, Schlis-ner R, Chin MT, Chen P. 2008. Hey2functions in parallel with Hes1 andHes5 for mammalian auditory sensoryorgan development. BMC Dev Biol 8:20.

Liu W, Li G, Chien JS, Raft S, Zhang H,Chiang C, Frenz DA. 2002. Sonic hedge-hog regulates otic capsule chondrogene-sis and inner ear development in themouse embryo. Dev Biol 248:240–250.

Lowenheim H, Furness DN, Kil J, ZinnC, Gultig K, Fero ML, Frost D,Gummer AW, Roberts JM, Rubel EW,Hackney CM, Zenner HP. 1999. Genedisruption of p27(Kip1) allows cell pro-liferation in the postnatal and adultorgan of corti. Proc Natl Acad Sci U SA 96:4084–4088.

Mueller KL, Jacques BE, Kelley MW. 2002.Fibroblast growth factor signaling regu-lates pillar cell development in the organof corti. J Neurosci 22:9368–9377.

Murata J, Ohtsuka T, Tokunaga A, NishiikeS, Inohara H, Okano H, Kageyama R.2009. Notch-Hes1 pathway contributes tothe cochlear prosensory formation poten-tially through the transcriptional down-regulation of p27Kip1. J Neurosci Res 87:3521–3534.

Noden DM vdT. 1986. The developing ear:tissue origins and interactions. In:Ruben RJ, van de Water TR, Rubel EW,editors. The biology of change in otolar-yngology. New York: Excerpta Medica.p 15–46.

Pirvola U, Ylikoski J, Trokovic R, HebertJM, McConnell SK, Partanen J. 2002.FGFR1 is required for the developmentof the auditory sensory epithelium.Neuron 35:671–680.

Puligilla C, Feng F, Ishikawa K, BertuzziS, Dabdoub A, Griffith AJ, Fritzsch B,Kelley MW. 2007. Disruption of fibro-blast growth factor receptor 3 signalingresults in defects in cellular differentia-tion, neuronal patterning, and hearingimpairment. Dev Dyn 236:1905–1917.

Puligilla C, Dabdoub A, Brenowitz SD,Kelley MW. 2010. Sox2 induces neuro-nal formation in the developing mam-malian cochlea. J Neurosci 30:714–722.

Riccomagno MM, Martinu L, MulheisenM, Wu DK, Epstein DJ. 2002. Specifica-tion of the mammalian cochlea isdependent on Sonic hedgehog. GenesDev 16:2365–2378.

Riccomagno MM, Takada S, Epstein DJ.2005. Wnt-dependent regulation ofinner ear morphogenesis is balanced bythe opposing and supporting roles ofShh. Genes Dev 19:1612–1623.

Rubel EW, Fritzsch B. 2002. Auditory sys-tem development: primary auditoryneurons and their targets. Annu RevNeurosci 25:51–101.

Ruben RJ. 1967. Development of theinner ear of the mouse: a radioauto-graphic study of terminal mitoses. ActaOtolaryngol Suppl 220:221–244.

Takebayashi S, Yamamoto N, Yabe D,Fukuda H, Kojima K, Ito J, Honjo T.2007. Multiple roles of Notch signalingin cochlear development. Dev Biol 307:165–178.

Tang LS, Alger HM, Pereira FA. 2006.COUP-TFI controls Notch regulation ofhair cell and support cell differentia-tion. Development 133:3683–3693.

Wijgerde M, McMahon JA, Rule M, McMa-hon AP. 2002. A direct requirement forHedgehog signaling for normal specifica-tion of all ventral progenitor domains inthe presumptive mammalian spinalcord. Genes Dev 16:2849–2864.

Zhao Y, Wang Y, Wang Z, Liu H, Shen Y, LiW, Heller S, Li H. 2006. Sonic hedgehogpromotes mouse inner ear progenitorcell proliferation and hair cell generationin vitro. Neuroreport 17:121–124.

Zheng JL, Shou J, Guillemot F, KageyamaR, Gao WQ. 2000. Hes1 is a negativeregulator of inner ear hair cell differen-tiation. Development 127:4551–4560.

Zine A, Aubert A, Qiu J, Therianos S, Guil-lemot F, Kageyama R, de Ribaupierre F.2001. Hes1 and Hes5 activities arerequired for the normal development ofthe hair cells in the mammalian innerear. J Neurosci 21:4712–4720.

Dev

elop

men

tal D

ynam

ics

SHH EXPRESSION IN COCHLEAR NEURONS 1683