A primary cilia-dependent etiology for midline facial disorders

A primary cilia-dependent etiology for midlinefacial disorders

Samantha A. Brugmann, Nancy C. Allen, Aaron W. James, Zesemayat Mekonnen,

Elena Madan and Jill A. Helms�

Department of Surgery, Division of Plastic and Reconstructive Surgery, Stanford University, Stanford, CA, USA

Received November 14, 2009; Revised and Accepted January 21, 2010

Human faces exhibit enormous variation. When pathological conditions are superimposed on normal vari-ation, a nearly unbroken series of facial morphologies is produced. When viewed in full, this spectrumranges from cyclopia and hypotelorism to hypertelorism and facial duplications. Decreased Hedgehog path-way activity causes holoprosencephaly and hypotelorism. Here, we show that excessive Hedgehog activity,caused by truncating the primary cilia on cranial neural crest cells, causes hypertelorism and frontonasaldysplasia (FND). Elimination of the intraflagellar transport protein Kif3a leads to excessive Hedgehog respon-siveness in facial mesenchyme, which is accompanied by broader expression domains of Gli1, Ptc and Shh,and reduced expression domains of Gli3. Furthermore, broader domains of Gli1 expression correspond toareas of enhanced neural crest cell proliferation in the facial prominences of Kif3a conditional knockouts.Avian Talpid embryos that lack primary cilia exhibit similar molecular changes and similar facial phenotypes.Collectively, these data support our hypothesis that a severe narrowing of the facial midline and excessiveexpansion of the facial midline are both attributable to disruptions in Hedgehog pathway activity. Thesedata also raise the possibility that genes encoding ciliary proteins are candidates for human conditions ofhypertelorism and FNDs.

INTRODUCTION

Craniofacial abnormalities comprise approximately one-thirdof all birth defects and of those developmental defects of theforebrain and midface, such as holoprosencephaly (HPE),are the most common (1,2). Although the phenotypic presen-tation of HPE is variable, HPE is associated with a distinctfacial ‘gestalt’, a reduced facial midline. The most extremecases of HPE are characterized by a complete collapse ofthe facial midline, such as cyclopia and the congenitalabsence of a mature nose. Less severe forms of HPE featureclose-set eyes (hypotelorism), defects of the upper lip andnose, and variable central nervous system (CNS) defects(reviewed in 3–6).

In contrast to defects associated with a reduced facialmidline, there is a second class of midline disorders that areassociated with an expanded facial midline. The mostextreme cases of midline expansion result in craniofacialduplication, or diprosopus. The phenotype comprises a widespectrum and ranges from partial duplication of a few facial

structures to complete dicephalus (7). Less severe forms ofmidline expansion are the hallmark of syndromes like fronto-nasal dysplasia (FND) in which a broad nasal root, medialclefting and extreme ocular hypertelorism are common (8).Although the phenotypic presentation of syndromes such asFND is variable, hypertelorism and an expanded facialmidline are the characteristic features.

The molecular basis for midline disorders has been thesubject of intense scrutiny. For syndromes associated withmidline collapse loss-of-function mutations in the Hedgehogsignaling pathway are common, but no direct correlationshave been uncovered between genotype and phenotype (9).Some evidence suggests a relationship between the timing ofHedgehog disruption, and the severity of the HPE phenotype(10,11), and other genetic studies indicate a dose-dependentrelationship between pathway activity and the extent towhich the facial midline is reduced (12,13).

Understanding the genetic basis for midline expansionshas not been as straightforward. Numerous syndromescharacterized by midline expansion including, frontorhiny,

�To whom correspondence should be addressed at: Department of Surgery, Division of Plastic and Reconstructive Surgery, Stanford University, 257Campus Drive Stanford, CA 94305, USA. Tel: þ1 6507360300; Fax: þ1 6507364374; Email: [email protected]

# The Author 2010. Published by Oxford University Press. All rights reserved.For Permissions, please email: [email protected]

Human Molecular Genetics, 2010, Vol. 19, No. 8 1577–1592doi:10.1093/hmg/ddq030Advance Access published on January 27, 2010

Downloaded from https://academic.oup.com/hmg/article-abstract/19/8/1577/572010by gueston 15 February 2018

craniofrontonasal syndrome and FND have been attributed toloss of Alx3 (14), mutations in Ephrin-B1 (15) and defectsin Six2 (16), respectively. Loss of the Hedgehog repressorGli3 has been linked to Greig cephalopolysyndactyly (a syn-drome characterized by hypertelorism) (17), and introductionof ectopic Hedgehog has been shown to increase the widthof the midline facial prominence (frontonasal prominence) inan avian model (18). Our own studies have demonstratedthat Wnt pathway activity is important for midline patterning(19); however, it was the loss of Wnt dependent proliferationin the maxillary prominence that permitted the expansion ofthe midline.

To gain insight into the mechanisms that produce midlineexpansion, we opted for an approach that, rather than disrupt-ing any member of a signaling pathway such as Wnt or Hh,disrupted the ability of the cell to respond to signals in theirenvironment. For this, we exploited an ubiquitous organelle,the primary cilia. Primary cilia have been reported to berequired for both Hedgehog and Wnt signal transduction(20–23). Loss of the intraflagellar transport protein (IFT)Kif3a results in non-functional primary cilia (24). We used aconditional knockout approach to eliminate Kif3a and thusdisrupt the function of primary cilia in the cells that giverise to the facial skeleton, the neural crest cells. Our datademonstrate that Kif3a-mediated truncation of primary ciliacauses a gain of Hedgehog function, and aberrant neuralcrest cell proliferation in the facial midline. The resultingembryos exhibit a form of FND, characterized by hypertelor-ism and diprosopus. Our data support a common molecularetiology linking loss and gain of Hedgehog function with aspectrum of midline facial anomalies.

RESULTS

Cranial neural crest cells extend primary cilia

We began our study by demonstrating that neural crest cellsextend primary cilia in vivo. We identified primary ciliausing immunostaining for acetylated tubulin (25) andArl13b, a small GTPase of the Arf/Arl family (26). Cranialneural crest cells in the frontonasal prominence showedrobust immunostaining for both proteins (Fig. 1A and C).We next tested the consequences of deleting Kif3a fromcranial neural crest cells. A complete inactivation of Kif3aresults in embryonic lethality at early stages of development(27); we circumvented this difficulty by crossing floxedKif3a mice (28) with the neural crest deleter Wnt1-Cre (29)to produce Wnt1-Cre::Kif3afl/fl embryos (referred to as Kif3aCKO).

Kif3a CKO embryos specifically lost primary cilia in neuralcrest derived facial mesenchyme. Acetylated tubulin immu-nostaining identified long, finger-like primary cilia in thee10.5 wild-type frontonasal prominence (Fig. 1A). The punc-tate pattern of immunostaining demonstrated that primary ciliawere truncated in the Kif3a CKO frontonasal prominence(Fig. 1B). Arl13b immunostaining confirmed the loss ofprimary cilia [compare wild-type (Fig. 1C) with Kif3a CKO(Fig. 1D)].

We next verified the tissue specificity of our Kif3a CKO.The neuroepithelium lies just dorsal to the neural crest-derived

facial mesenchyme, and is derived from ectoderm. Arl13bimmunostaining confirmed the extension of primary ciliafrom the neuroectoderm into the ventricle in both the hind-brain (Fig. 1E) and forebrain (Fig. 1F). Kif3a CKO mutanthindbrain and forebrain neuroectodermal cells were alsoimmunopositive for Arl13b (Fig. 1G and H). Facial ectoderm,which surrounds the neural crest mesenchyme, is derived fromsurface ectoderm. Although immunopositive cells were diffi-cult to detect at e10.5, we nonetheless saw an equivalent dis-tribution of Arl13b-positive cells in wild-type and Kif3a CKOfacial ectoderm (arrows, Fig. 1I and J). Collectively, these dataconfirmed that the loss of primary cilia was confined to neuralcrest cells in Kif3a CKO embryos. We next analyzed the facialphenotype of Kif3a CKO embryos.

Kif3a is required for normal craniofacial patterning

At e10.5, when neural crest migration into the facial promi-nences is complete (30), wild-type and Kif3a CKO embryoshave very similar appearances (n ¼ 17 wild-type, n ¼ 15mutant, data not shown), but shortly thereafter, subtle altera-tions in facial phenotype became evident. Compared to wild-type littermates, e11.5 Kif3a CKO faces were wider. Usinginfra-nasal measurements, we ascertained that e11.5 Kif3aCKO embryos had a 19% increase in frontonasal width (wild-type n ¼ 15, mutant n ¼ 7; Fig. 2A, B and I). After e11.5, theKif3a CKO midline defect was exacerbated. Compared towild-type littermates, the Kif3a CKO infra-nasal width wasincreased by 97% at e12.5 (wild-type n ¼ 9, mutant n ¼ 3;dotted yellow lines, Fig. 2C, D and I).

By e14.5, the Kif3a CKO frontonasal prominence wasalmost 100% wider than normal (wild-type n ¼ 12, mutantn ¼ 4; dotted yellow lines, Fig. 2E, F and I). At e17.5, theKif3a CKO infra-nasal distance was 120% wider than thewild-type (wild-type n ¼ 5, mutant n ¼ 3; dotted yellowlines, Fig. 2G, H and I). These measurements demonstratedthat the increased mid-facial width of Kif3a CKOs was signifi-cant and maintained relative to wild-type embryos throughoutcraniofacial development (Fig. 2I).

Clefting of the secondary palate accompanied the widenedfrontonasal prominence in Kif3a CKO embryos (Fig. 2J andK). Examining the palate in frontal section revealed that thewild-type medial edge epithelium fused, and the underlyingpalatine bones nearly merged in the midline (Fig. 2L, dottedlines), whereas in the Kif3a CKO, the palatine bones were dys-morphic. The bones did not extend towards the midline(Fig. 2M). Some small ectopic ossifications were detectablein the midline (Fig. 2M, dotted black lines), but whetherthese represented fragments of the palatine bones was unclear.

Kif3a CKO embryos phenocopy human frontonasaldysplasia

Deletion of Kif3a from cranial neural crest cells resulted in anextreme and significant expansion of the facial midline. Inhumans, abnormal expansion of the facial midline can mani-fest as FND. Defining features of FNDs include ocular hyper-telorism; a flat broad nose; in more severe cases, the nose mayseparate vertically into two parts (bifid nasal septum), and cleftlip and/or palate. In addition, an abnormal skin-covered gap in

1578 Human Molecular Genetics, 2010, Vol. 19, No. 8


the front of the head (anterior cranium occultum) and CNSdefects are frequently present (31,32). We next examinedKif3a CKO embryos for defects analogous to human FND.

At e12.5, the wild-type nasal septum begins as a singlemesenchymal condensation in the frontonasal prominencewhich becomes continuous with condensations of the nasal

capsule (Fig. 3A). In Kif3a CKO littermates, the nasalseptum was evident as a bifid condensation (Fig. 3B). Bye16.5, the wild-type nasal septum condensation has maturedinto a midline cartilaginous rod (Fig. 3C). In Kif3a CKO litter-mates, the bifid condensation had matured into a duplicatednasal septum (Fig. 3D). Given previously published reports

Figure 1. Cranial neural crest cells do not extend primary cilia in Kif3a CKO embryos. (A) Acetylated tubulin staining (green) in WT frontonasal neural crestcells. (B) Punctate acetylated tubulin staining in Kif3a CKO frontonasal neural crest cells. (C) Arl13b (green) expression in WT frontonasal neural crest cells. (D)Arl13b expression is lost in Kif3a CKO frontonasal neural crest cells. (E–H) Primary cilia remain present in the neuroectoderm. (E and F) Arl13b expression inWT hindbrain (hb) and forebrain (fb) neuroectoderm. (G and H) Arl13b expression in Kif3a CKO hindbrain (hb) and forebrain (fb) neuroectoderm. Dottedyellow lines outline the ventricle (v). (I and J) Arl13b expression (white arrows) in WT and Kif3a CKO facial ectoderm (fe). Dotted yellow lines outline bound-ary between facial ectoderm and neural crest (nc). Note Arl13b expression in mesenchyme (green arrow) in WT but not Kif3a CKO. Cell nuclei appear blue withHoechst counterstain. Scale bars denote 10 mm.

Human Molecular Genetics, 2010, Vol. 19, No. 8 1579


(33), we were surprised to find that maturation of the cranio-facial skeleton was unaffected by loss of Kif3a. Forexample, Kif3a CKO chondrocytes were similar to wild-typechondrocytes in morphology (Fig. 3C and D; insets), distri-bution and intensity of alkaline phosphatase activity (datanot shown). The primary defect with the Kif3a CKO skeletonappeared not to be in the differentiation of chondrogenic orosteogenic tissues, but rather in their patterning: the wild-typenasal septum is located in the midline and is surrounded by thepre-maxillary bones and the nasal bones (dorsal view, Fig. 3E;frontal view, Fig. 3G), but in Kif3a CKO embryos, the nasalseptum was duplicated, which displaced both the nasal andpre-maxillary bones laterally (dorsal view, Fig. 3F; frontalview, 3H).

We continued our examination of the Kif3a CKO craniofa-cial skeleton by examining the palate, the skull and the mand-ible. In Kif3a CKO embryos, the bones of the palate and

ventral cranial midline, including the maxilla, trabecularbasal plate, palatine and basisphenoid were either laterally dis-placed, or if they were midline elements, reduced to bonynodules or absent (Fig. 3I and J). The Kif3a CKO craniumwas also severely dysmorphic. Compared to wild-type litter-mates (Fig. 3K), mutant embryos had laterally displaced,underdeveloped frontal bones, which resulted in an abnormalopening in the skull (i.e. cranium occultum; Fig. 3L) thatwas covered by an epithelial membrane (see whole mountsin Fig. 2). In addition, we found that the Kif3a CKO mandibleswere 30% shorter than their wild-type littermates, and lacked adiscernable ramus and condyle (Fig. 3M versus N).

Finally, we examined Kif3a CKO embryos for the CNSdefects associated with FND: namely agenesis of the corpuscallosum (32). Nerve fibers, which normally connect the leftand right hemispheres of the brain (Fig. 3O), failed to spanthe midline in Kif3a CKO mutants (Fig. 3P). Taken together,

Figure 2. Truncation of primary cilia in cranial neural crest cells causes facial dysmorphology. (A and B) Frontal view of e11.5 wild-type and Kif3a CKOembryos. White lines measure distance between nasal pits. (C and D) Frontal view of e12.5 wild-type and Kif3a CKO embryos. The dotted yellow linesoutline frontonasal prominence (F) and distance between the nasal pits (�). (E–H) Frontal view of e14.5 and 17.5 wild-type and Kif3a CKO embryos, respect-ively. (I) Quantification of the percent increase of infra-nasal width at various developmental stages in Kif3a CKO embryos relative to WT embryos, �P � 0.01.(J and K) Palatal view of the e14.5 WT and Kif3a CKO embryo. (^) denotes clefting. (L) Pentachrome stain of coronal sections of e17.5 WT palate. Palatinebones (pa) are outlined with dotted black lines. (M) Pentachrome stain of coronal sections of e17.5 Kif3 CKO palate. Palatine bones (pa) and midline bonynodules are outlined with dotted black lines. Maxillary prominence (mx), mandibular prominence (mn), lateral nasal prominence (l), median nasal prominence(m), nasal capsule (nc) and tooth (to). White scale bar ¼ 10 mm, black scale bar ¼ 100 mm.



these analyses revealed that Kif3a CKO mutants shared anumber of features found in human conditions that fallwithin the FND sequence, including bifid nasal septum, cleftpalate, cranium occultum and agenesis of the corpus callosum(31). Thus, the Kif3a CKO mutants displayed the murineequivalent of a human FND.

Truncation of primary cilia causes region-specificalterations in facial Wnt signaling

We sought to clarify the molecular basis for the Kif3a CKOmidline phenotype. Previous work from our lab has shownthat Wnt pathway activity plays an integral role in patterningthe vertebrate face (19). Although some studies indicate thatprimary cilia are not essential for Wnt signal transduction(34,35), other experiments suggest a key role for primarycilia in regulating b-catenin-dependent Wnt signaling (23).To address the possibility that loss of cilia disrupted thepattern of Wnt pathway activity in the face, we evaluatedthe expression of a transgenic reporter of canonical Wnt

signaling in Kif3a CKO embryos. We crossed TOPgalreporter mice (36) into the mutant background to produceWnt1-Cre::Kif3a f l/f l::TOPgal embryos (referred to as Kif3aCKO::TOPgal). At e10.5, prior to obvious morphologic differ-ences, the patterns of Xgal staining in wild-type and Kif3aCKO mutants appeared similar (n ¼ 3; Fig. 4A and B). Forexample, both embryos showed evidence of Xgal staining inthe telencephalon, and in the lateral nasal, maxillary and man-dibular prominences (Fig. 4A and B). Careful examinationfrom a lateral perspective showed a slight expansion of thelateral nasal and maxillary domains of Xgal staining(Fig. 4C and D).

We previously showed that domains of Wnt responsivenessin the face coincide with localized regions of proliferation, andthat these regions are ultimately responsible for shaping thevertebrate face (19). We reasoned that the expanded midlinein the Kif3a CKO mutant could be the consequence ofectopic Wnt responsiveness (and therefore increased prolifer-ation) in the frontonasal prominence. We were surprised tofind that despite the gross morphologic variation in the

Figure 3. Skeletal analyses of Kif3a CKO embryos. (A and B) Pentachrome stain on transverse sections of e12.5 WT and Kif3a CKO embryos. The nasal septumcondensation (dotted white line, ns) is a single skeletal element in wild-type embryos. The nasal septum of the Kif3a CKO mutant is a bridged, bifid skeletalelement. (C and D) Safranin-O staining on transverse sections of e16.5 WT and Kif3a CKO embryos. The nasal septum remains as a single skeletal element inWT embryos and is duplicated in Kif3a CKO embryos. Note the duplicated nasal cavities (black and yellow asterisks). High magnification of nasal septum(insets) shows similar patterning of the cartilages. (E and F) Dorsal view of whole mount alcian blue/alizarin red bone and cartilage staining of e17.5 WTand Kif3a CKO embryos. (G and H) Frontal view of whole mount alcian blue/alizarin red staining in e17.5 WT and Kif3a CKO embryos. (I and J) Ventralview of whole mount alcian blue/alizarin red staining in e17.5 WT and Kif3a CKO embryos. (K and L) Dorsal view of whole mount alcian blue/alizarinred staining in e17.5 WT and Kif3a CKO embryos. Distance between frontal bones (fb, white line) is significantly increased in Kif3a CKO embryo resultingin anterior cranium occultum. (M) The mandible (mn) of a WT embryo. (N) The mandible of a Kif3a CKO mandible lacks the ramus (r) and condyle (c). (O)H&E staining of corpus callosum (dotted yellow line) in frontal section of e17.5 WT embryos. (P) Failure of neural fibers to span the midline in Kif3a CKOmutant (dotted yellow lines). Nasal cavity (�), nasal capsule (ncp), whisker primordia (w), nasal bone (n), premaxilla (pmx), trabecular basal plate (tbp), palatine(pl), basisphenoid (bs), ectopic midline bone (white asterisk), ventricle (v). Black scale bars ¼ 100 mm, white scale bars ¼ 400 mm.



Figure 4. Wnt signaling is altered in a site-specific manner in Kif3a CKO embryos. Xgal staining (blue) indicates sites of Wnt responsiveness. (A and B) Frontalview of e10.5 TOPgal and Kif3a CKO::TOPgal embryos. (C and D) Lateral view of e10.5 TOPgal and Kif3a CKO::TOPgal embryos. (E and F) Frontal view ofe13.0 control and mutant embryos. The frontonasal prominence is significantly wider in the mutant [compare white lines in (E) and (F)]. (G and H) Ventral viewof the palate showing differential patterns Wnt pathway activity in the dental lamina (dl). Wnt responsiveness is punctate (see white ^) in the control, whereas it iscontinuous (see dotted red line) in Kif3a CKO::TOPgal. (I) Dorsal view of the tongue (t), dental epithelium (dotted yellow lines) and mandible (mn) in e13.0control embryo. (J) Dorsal view of the tongue remnant (�), reduced dental epithelium (dotted yellow lines) and mandible in a Kif3a CKO::TOPgal mutantembryo. (K) Xgal staining in transverse sections of e13.0 WT and (L) Kif3a CKO::TOPgal mutant embryo. Wnt responsive domain in the nasal epithelium(ne), the frontonasal mesenchyme (compare white and black arrows) and lateral nasal mesenchyme (compare dotted black lines) is lost or reduced in theKif3a CKO::TOPgal mutant embryo. (M and N) Frontal view of e16.5 TOPgal and Kif3a CKO::TOPgal embryos. Dotted yellow lines outline the frontonasalprominence. (O and P) Transverse sections through the nasal septum of e16.5 TOPgal and Kif3a CKO::TOPgal embryos. Wnt activity is maintained in thecartilaginous bifid nasal septum (dotted red line) and ectopic cartilage nodules (�), and expanded in the surface ectoderm (dotted black lines) of Kif3a CKO::-TOPgal embryos. (Q) High magnification of area in dotted white box in (P) shows Wnt responsive ectopic cartilage nodule. (R) Pentachrome staining of ectopicskeletal elements in Kif3a CKO embryos. Nasal pit (np), telencephalon (te), frontonasal prominence (f), lateral nasal prominence (l), maxillary prominence (mx),mandibular prominence (mn), eye (dotted black circle), facial ectoderm (fe), bone nodule (bn). Black scale bars ¼ 100 mm, white scale bars ¼ 300 mm.



width of the mutant frontonasal prominence, the boundariesand domains of Wnt responsiveness in wild-type TOPgaland Kif3a CKO::TOPgal embryos were largely maintained(Fig. 4E and F). For example, the lateral nasal and maxillarydomains of Xgal staining were remarkably well conserved inboth wild-type and mutants, and the frontonasal prominence,which in wild-type embryos is normally free of Xgal staining,maintained this Xgal free status in the mutant (Fig. 4E–H).

We examined other domains of Wnt responsiveness in theface. Focal domains of Wnt and Shh signaling define thesites where teeth will develop (37). In wild-type TOPgalembryos, discrete Xgal-positive domains were evident in thedental lamina (Fig. 4G), but in Kif3a CKO::TOPgalembryos, the domain was an unbroken line (Fig. 4H). Thedental lamina staining pattern differed slightly between theupper and lower jaws: in the mutant mandible, Xgaldomains were smaller in comparison to wild-type littermates(Fig. 4I and J), and the mutant tongue was reduced to aremnant.

Analyses of Xgal-stained tissue sections from Kif3a CKO::-TOPgal embryos revealed other changes in the pattern of Wntpathway activity. Initially, ectoderm is the first Wnt responsivetissue in the face, followed by Wnt responsiveness in under-lying neural crest-derived mesenchyme (Fig. 4K and see(19)). In Kif3a CKO::TOPgal embryos, the mesenchymalXgal domains were smaller or in some cases, absent(Fig. 4L). By e16.5, the Xgal staining pattern shifted again:whole-mount analyses showed ectopic patches of Xgal posi-tive tissue in the mutant frontonasal prominence (n ¼ 3;Fig. 4M wild-type compared to Fig. 4N). In tissue sections,these ectopic Xgal patches demarcated small cell aggregates(compare control, Fig. 4O with P and Q). Morphologic andhistologic analyses demonstrated that these condensationswere undergoing chondrogenesis and osteogenesis (Fig. 4R),and ultimately formed the ectopic islands of bone and cartilagewe observed in the e17.5 mutant facial midline (Figs 2Mand 3L).

Taken together, the analyses of Kif3a CKO::TOPgalembryos clearly demonstrated that b-catenin-dependent Wntsignaling was altered in certain domains that have lost func-tional primary cilia. An expansion in Wnt responsiveness,however, did not appear to be the direct cause the expandedKif3a CKO frontonasal prominence as originally hypoth-esized. We next employed another genetic strategy to deter-mine whether a disruption in Hedgehog signaling was theroot cause for this facial malformation.

Loss of Kif3a results in an expansion of Hedgehogresponsiveness in the face

Most studies suggest that a mutation in Kif3a reduces Hedge-hog signal transduction (38–40), and it is known that the lossof Hedgehog signaling causes hypotelorism and cyclopia(10,41,42). Since our mutant embryos have hypertelorism,we hypothesized that neural crest-specific deletion of Kif3amight cause a gain of Hedgehog pathway activity within theface. This hypothesis is not without precedent, as other inves-tigators have reported that Hedgehog target genes areexpressed in broader domains after the loss or truncationof primary cilia (33,43). The mechanism(s) behind this

contradictory effect have not been elucidated, but our con-ditional genetic knock-out offered a unique opportunity toexplore this problem in more detail.

To directly assess the effect of Kif3a deletion on Hedgehogsignal transduction, we crossed Ptc-LacZ reporter mice (44)into the mutant background to produce Wnt1-Cre::Kif3afl/fl::Ptc-LacZ embryos (referred to as Kif3a CKO::Ptc-LacZ).We began analyses at e10.5, a time-point prior to anyobvious differences between wild-type and Kif3a CKOfacial morphology. In Ptc-LacZ controls, Xgal staining wasdetectable in all known sites of Hedgehog pathway activity(n ¼ 4; Fig. 5A). By whole-mount analysis, the stainingpattern appeared similar in Kif3a CKO::Ptc-LacZ embryos(n ¼ 6; Fig. 5B), but sagittal tissue sections revealed subtlechanges in the pattern of Xgal staining. For example, relativeto e10.5 Ptc-LacZ controls, Kif3a CKO::Ptc-LacZ embryosshowed broader domains of Xgal staining in the midlinefacial mesenchyme and facial ectoderm (Fig. 5C and D,white and black arrows). Most notable was an expansion ofHedgehog responsiveness within the facial ectoderm itself.Although in our conditional knockout facial ectodermal cellsmaintain their primary cilia (Fig. 1I and J), the ectodermalXgal domain was clearly expanded (Fig. 5D). Broaderdomains of Hedgehog activity were also observed in e14.5Kif3a CKO::Ptc-LacZ embryos relative to Ptc-LacZ litter-mates (Fig. 5E and F). In particular, the regions of Xgal stain-ing in the Kif3a CKO::Ptc-LacZ facial mesenchyme weremarkedly expanded (Fig. 5G and H). The most dramaticincrease in Xgal staining was observed in the mesenchymeof the primary palate (Fig. 5I and J). The molecular basisand cellular consequences for this effect on Hedgehogpathway activity within the facial prominences became thefocus of our next series of experiments.

Shh and Gli1 expression domains are expanded in Kif3aCKO facial prominences

Hedgehog signals from forebrain neuroectoderm, facial ecto-derm and pharyngeal endoderm are required for proper devel-opment of the craniofacial complex (10,45,46). We firstexamined tissue sections to determine whether the pattern ofShh expression in these epithelial domains was altered. Ate11.0, the domains of Shh in the telencephalon, diencephalon,ventral hindbrain and oral ectoderm are equivalent betweenwild-type and mutant tissues (Fig. 6A–D), indicating thatKif3a deletion in cranial neural crest cells does not affectthe initial Shh domains in epithelia established during neuro-genesis (18). The targets of Shh signaling, however, are ecto-pically expressed. The Gli transcription factors are directtargets of Hedgehog signaling (47). In the presence of aHedgehog ligand, Gli1 is transcriptionally activated and func-tions to extend the duration and strength of a Hedgehog signal(47,48). At e11.0, Gli1 expression domains were dramaticallyaltered in the Kif3a CKO facial prominences (Fig. 6E–H). Inthe proximal frontonasal prominence, the mutant Gli1 domainwas smaller, but in the distal frontonasal prominence andlateral nasal prominences the Gli1 domains were larger(Fig. 6E and F). Examination of Gli1 expression in sagittalsection reveals an expansion of the expression domain



posteriorly into the palatal mesenchyme and within the nasalepithelium (Fig. 6G and H, dotted yellow line).

Although early domains of Shh were normal (Fig. 6A–D),between e11.0 and e12.5 the Hedgehog signaling was dis-rupted in Kif3a CKO embryos. By e12.5, the neuroectodermaldomains of Shh were laterally displaced from their normalmidline position, and the oral ectodermal domain of Shh wasexpanded in Kif3a CKO embryos (Fig. 6I and J). Althoughthe tongue, a midline structure, was absent in Kif3a CKOembryos, the Shh domain that marks the lingual ectoderm per-sists, and was expanded (Fig. 6K and L). The Shh-positiveincisor dental lamina was also expanded in Kif3a CKO oralectoderm (Fig. 6M and N).

Expression domains of the Gli transcription factors werealso altered at this later time point. The Gli1 domain wasexpanded in the oral ectoderm (Fig. 6O and P). Expressionof the transcriptional repressor, Gli3, on the other hand, wasreduced in Kif3a CKO oral ectoderm (Fig. 6Q and R). Thus,molecular data support a model whereby loss of primarycilia in neural crest cells results in ectopic Hedgehog signalingin the face.

Loss of primary cilia results in a Hedgehog-dependentincrease in neural crest cell proliferation

We next sought to understand the cellular consequence of theloss of primary cilia. Neural crest cells undergo four distinctcellular behaviors; epithelial to mesenchymal transformation,migration, proliferation and differentiation. Previous studiesin zebrafish proposed that a loss of cilia adversely affectedthe migration of neural crest cells (49). To rule out a migrationdefect and to confirm that neural crest cells were able to popu-late the facial prominences in the Kif3a CKO mutants, wecrossed the Kif3a CKO mutant into the Wnt1-Cre::R26Rreporter line (Supplementary Material, Fig. S1). At e10.5,Xgal staining in Wnt1-Cre::R26R and Wnt1-Cre::R26R::Kif3aCKO was similarly localized to areas populated by neural crestcells (Supplementary Material, Fig. S1A–D). To verify Xgalstaining was marking neural crest cells, we sectioned theseembryos and found that the mesenchyme of the frontonasal,maxillary and mandibular prominence of Wnt1-Cre::R26Rand Wnt1-Cre::R26R::Kif3a CKO were both Xgal positive,whereas the forebrain and surface ectoderm were Xgal nega-tive (Supplementary Material, Fig. S1E–H). Furthermore,we confirmed that it was the neural crest mesenchyme thatlacked primary cilia in the Kif3a CKO by performing doubleimmunostaining for the neural crest marker, AP2a, andprimary cilia marker, Arl13b. We found double positivecells in the wild-type mandibular prominence, but noco-localization of AP2a and Arl13b in the Kif3a CKO man-dibular prominence (Supplementary Material, Fig. S1I andJ). These results allowed us to confirm that although Kif3aCKO neural crest cells lacked primary cilia, they were stillable to migrate out of the neural tube and populate the facialprominences.

Since neural crest cells were able to exit the neural tube,arrived in the facial prominences and underwent skeletaldifferentiation, we assayed neural crest cell proliferation.Using BrdU immunostaining (Fig. 7A–F), we foundexuberant proliferation throughout the Kif3a CKO facial pro-

Figure 5. Hedgehog pathway activity is expanded in Kif3a CKO embryos. (Aand B) Frontal view of whole mount Xgal stained e10.5 Ptc-LacZ and Kif3aCKO::Ptc-LacZ embryos. (C and D) Sagittal sections through e10.5 Xgalstained Ptc-LacZ and Kif3a CKO::Ptc-LacZ embryos. Xgal staining in thesurface ectoderm of the mutant is expanded anteriorly (dashed white line,arrows). (E and F) Frontal view of whole mount Xgal stained e14.5Ptc-LacZ and Kif3a CKO::Ptc-LacZ embryos. Dashed white and yellowlines indicate plane of section shown in (G and H) and (I and J), respectively.(G and H) Transverse sections through nasal capsule of (E) and (F) showingincreased Hedgehog pathway activity in facial mesenchyme and nasal epi-thelium (ne) of Kif3a CKO::Ptc-LacZ mutant (compare black and whitearrows and asterisks). (I and J) Transverse sections through the oral cavity(oc) of (E) and (F) showing increased Hedgehog pathway activity in themedial facial mesenchyme of the Kif3a CKO::Ptc-LacZ mutant. Frontonasalprominence (f), nasal pit (np), mandibular prominence (mn), eye (e), maxillaryprominence (mx), whisker primordia (w), primary palate (pp). Black scalebars ¼ 300 mm, white scale bar ¼ 150 mm.



http://hmg.oxfordjournals.org/cgi/content/full/ddq030/DC1









minences, in comparison to the localized areas of proliferationseen in wild-type facial prominences. The enhanced cell pro-liferation was most notable in the Kif3a CKO facial midline.Analysis of BrdU incorporation was performed throughvarious section planes including the nasal capsule (Fig. 7Aand B) and palatal mesenchyme (Fig. 7C–F). The mostrobust increase in proliferation was observed in the ventralmidline of the frontonasal prominence (future primarypalate) in the Kif3a CKO (Fig. 7C–F). To confirm thisobserved increase in proliferation, we quantified BrdUincorporation in various planes of section via ImageJ software.

Quantification of BrdU incorporation showed an over 5-foldincrease in neural crest proliferation in Kif3a CKO facial pro-minences compared with wild-type control (Fig. 7G, P ¼0.002). We next explored the molecular basis for thisincreased proliferation.

The most notable molecular difference in Kif3a CKOembryos was the expansion of the expression domain of theHedgehog target Gli1. We hypothesized that the increasedproliferation in the Kif3a CKO mutant was a Hedgehogdependent process. To test this hypothesis, we first sought todetermine whether the increased proliferation and increased

Figure 6. Expression of Hedgehog ligand and targets are perturbed in Kif3a CKO embryos. (A–D) In situ hybridization for Shh in e11.0 WT and Kif3a CKO insagittal and transverse sections. (E–H) In situ hybridization for Gli1 in e11.0 WT and Kif3a CKO embryos in transverse and sagittal sections. Gli1 expression isup-regulated in the median nasal (m) and lateral nasal (l) prominences, nasal epithelium (ne) and posteriorly into the medial mesenchyme [see dotted yellow linein (G) and (H)]. (I and J) Shh expression in frontal sections of e12.5 WT and Kif3a CKO embryos. (K and L) High magnification of the Shh expression domain inthe oral ectoderm (oe) shown in (I) and (J), respectively. (M and N) Shh expression domain around the oral cavity (oc) in WT and Kif3a CKO e12.5 embryos. (Oand P) Gli1 expression in e12.5 WT and Kif3a CKO embryos. (Q and R) Gli3 expression in e12.5 WT and Kif3a CKO embryos. Frontonasal prominence (f),mandibular prominence (mn), telencephalon (te), diencephalon (di), rathke’s pouch (rp), floor plate (fp), ventral neuroectoderm (vn), tongue (t), oral ectoderm(oe), incisor dental lamina (idl), Scale bars ¼ 100 mm.



Gli1 expression were co-localized. Using equivalent planes ofsection, we found that the expanded Gli1 domains (Fig. 7Hand I) in the Kif3a CKO coincided with regions of increasedBrdU incorporation (Fig. 7J and K) in the medial mesench-yme. Co-localization of Gli1 expression and BrdU incorpor-ation was also found in lateral maxillary mesenchyme withinthe Kif3a CKO mutants (Fig. 7L–O).

To support a causal relationship between Hedgehog signal-ing and proliferation in the facial mesenchyme, frontonasalneural crest cells were isolated and exposed in vitro to exogen-ous N-terminal Sonic Hedgehog (Shh-N). First, we confirmedthat the concentration of Shh-N was sufficient to induce thedownstream target, Gli1 (Fig. 7P). Next, we assayed theamount of proliferation by measuring BrdU incorporation inwild-type and Kif3a CKO neural crest cells in the presenceof increasing doses of Shh-N. Relative to wild-type cells,

which showed no significant increase in proliferation inresponse to varying doses of Shh-N, the Kif3a CKO facialneural crest cells showed a significant, dose-dependentincrease in BrdU incorporation (Fig. 7Q, P ¼ 0.0028). There-fore, we hypothesize that Kif3a deletion altered the responseof neural crest cells to a Hedgehog signal and led to theirexuberant proliferation in the frontonasal prominence.

Primary cilia function in the cranial neural crest isconserved among species

We next examined the facial phenotype of another animal witha genetic defect that disrupts primary cilia function. Talpidchicken embryos exhibit polydactyly (50) as well asfacial malformations (51). Talpid3 phenotypes arise as aconsequence of mutation in a basal body protein that eliminates

Figure 7. Loss of Kif3a results in a Hedgehog-dependent increase in neural crest cell proliferation. (A–F) BrdU incorporation in e12.5 WT and Kif3a CKOembryos. (A and B) BrdU incorporation in the dorsal nasal capsule of e12.5 WT and Kif3a CKO embryos. (C and D) BrdU incorporation around the oralcavity of e12.5 WT and Kif3a CKO embryos. (E and F) BrdU incorporation in the ventral nasal capsule of e12.5 WT and Kif3a CKO embryos. (G) Quanti-fication of BrdU incorporation in WT and Kif3a CKO e12.5 embryos, �P ¼ 0.002. (H–K) Gli1 expression and adjacent BrdU incorporation in oral ectoderm(oe) and facial mesenchyme in e12.5 WT and Kif3a CKO embryos. (L–O) Gli1 expression and adjacent BrdU incorporation in lateral facial mesenchyme in WTand Kif3a CKO embryos. (P) Quantitative RT–PCR analysis of Gli1 expression in micro-dissected WT and Kif3a CKO cranial neural crest cells treated withShh-N, �P � 0.01. (Q) BrdU incorporation in WT and Kif3a CKO cranial neural crest cells treated with various concentrations of Shh-N,�P ¼ 0.0028. Fronto-nasal prominence (f), nasal pit (�), primary palate (pp), oral cavity (oc), maxillary prominence (mx), oral ectoderm (oe) and tongue (t). White scale bar ¼ 10 mm,black scale bar ¼ 100 mm.



primary cilia (52). We found that cranial neural crest cells ofTalpid2 embryos also lacked primary cilia (Fig. 8A and B)and basal bodies (Fig. 8C and D). As a consequence of theloss of functional primary cilia, Talpid2 embryos exhibitednumerous facial malformations (Fig. 8E and F). For example,by St. 19, the maxillary prominences and the stomodeumwere dysmorphic. Later in embryonic development, dispropor-tionate growth of the frontonasal prominence relative to themaxillary prominences resulted in facial clefting (Fig. 8G andH). Furthermore, like the Kif3a CKO, Talpid2 embryos exhib-

ited midline defects including the absence of a tongue (Fig. 8Gand H, asterisk), duplication of the midline nasal septum (Fig. 8Iand J) and expanded expression domains of Shh and Hedgehogtargets in the face (Fig. 8K–N, and see what follows).

In Talpid2 embryos, we found alterations in gene expressionextending to all craniofacial tissues. For example, Shh wasreduced in the Talpid2 midbrain (Fig. 8O and P) eventhough Shh and Gli1 expression in the limb bud wereunchanged (51). Since Talpid2 phenotypes have previouslybeen attributed to a constitutive activation of Hedgehog sig-

Figure 8. The avian primary cilia mutant Talpid2 has craniofacial abnormalites. (A and B) Acetylated tubulin staining (green) in control and Talpid2 mutantneural crest cells. (C and D) Gamma tubulin staining (green) in control and Talpid2 mutant neural crest cells. (E and F) Frontal view of control and Talpid2mutants at St. 19. Dotted white line outlines maxillary prominences (mx). Dotted black lines outline stomodeum and nasal pits (np). (G and H) Lateral view ofcontrol and Talpid2 mutants at St. 36. Dotted white line indicates cleft in Talpid2 mutant. Asterisk indicates lack of tongue (t) in Talpid2 mutant. (I and J) Alcianblue staining in control and Talpid2 mutants. Black arrow points to single nasal septum in control and duplicated nasal septum in Talpid2 mutant. (K and L) Shhexpression domains are expanded in the face of Talpid2 mutants relative to stage matched controls. (M and N) Ptc expression domains are expanded in the faceof Talpid2 mutants relative to stage-matched controls. (O and P) Shh expression is down-regulated in the diencephalon of Talpid2 mutants relative to stage-matched controls. (Q and R) Gli1 expression is up-regulated in the frontonasal and maxillary prominence of Talpid2 embryos. (S and T) Gli3 expression islost in the facial mesenchyme but up-regulated in facial ectoderm (white arrow) of Talpid2 embryos as seen in sagittal sections. Frontonasal prominence (f),maxillary prominence (mx), mandibular prominence (mn), tongue (t), eye (e), nasal pits (np), diencephalon (di), telecephalon (tel). White scale bar ¼10 mm. Black scale bars ¼ 200 mm.



naling (53), we extended our evaluation to include the Gligenes. As expected, the Gli1 domain was expanded (n ¼ 3;Fig. 8Q and R) and the Gli3 domain was reduced in mesench-yme (Fig. 8S and T).

DISCUSSION

The face as a barometer of Hedgehog pathway activity

In 1908, the anatomist Harris Hawthorn Wilder postulated aunifying theory to explain the relationship between extremeforms of facial dysmorphologies. On either side of a normalface, Wilder speculated, symmetrical anomalies that consti-tuted a single malformation spectrum existed. At oneextreme was cyclopia, characterized by a single median eyeand a proboscis replacing the nose, and at the other extremewas a duplication of facial features (Fig. 9). In betweenwere the conditions of hypo- and hypertelorism and theirassociated craniofacial and neurologic defects. In effect,Wilder’s continuum ranged from a complete collapse to aduplication of the face. Wilder hypothesized that thesedeformities were due to ‘some modification in the germitself, leading the organisms to develop in accordance withlaws as definite and natural, though not as usual, as thosegoverning normal development.’ Wilder sought experimentalevidence to support his theory on the ‘morphology of Cosmo-bia’, but was unable to produce embryos exhibiting this rangeof phenotypes (54).

Accumulating evidence indicates that Wilder’s theory ofCosmobia has merit. Wilder’s collapse of the face is nowknown as HPE. The most severe form of HPE is cyclopia,with milder forms appearing as hypotelorism (Fig. 9). Cyclo-pia is caused by disruptions in the Hedgehog pathway thatresult in the elimination (41,55) or reduction (10,12,13) ofHedgehog signaling. At the other end of the Cosmobic spec-trum are hypertelorism and facial duplications. Herein weprovide the first evidence that these malformations are alsocaused by disruptions in the Hedgehog pathway, but of thekind that produce expanded or ectopic activity (Fig. 9).

Primary cilia are an essential component for Hedgehogsignal transduction (21). Most studies indicate that mutationsin IFT proteins, including Kif3a, result in a down-regulationof Hedgehog signaling (21,43,56). Our data show that Kif3a-mediated disruption in primary cilia function results inexpanded Hedgehog pathway activity in the cranial neuralcrest. Other investigators have noted that Kif3a and otherIFT mutations can have similar paradoxical effects (43,57),and have attributed this inconsistency to which Gli protein,Gli1 or Gli3 is acting predominantly in the tissue(40,43,58,59). What determines this Gli predominance inany given tissue is not clear. A Gli12/2 mutation does notaffect craniofacial morphogenesis (60), suggesting that it isnot the predominant Gli in facial tissues. Humans and micewith inactivating Gli3 mutations, however, exhibit dramati-cally wider faces (17,61,62) implicating Gli3 as an importantregulator of craniofacial patterning. Kif3a CKO embryos showreduced Gli3 expression (Fig. 6) and they also exhibit widerfaces (Fig. 2), supporting the Gli3 predominance theory.

Primary cilia may integrate signals from multiplepathways in the developing face

Hedgehog signaling directly regulates Gli3 expression, but insome tissues Gli3 expression is also directly regulated by Wntpathway activity (63). In the facial prominences, we foundboth the Gli3 expression and Wnt responsiveness in facialmesenchyme. Gli3 expression levels were also reduced inthe Kif3a CKO mutant, coincident with a reduction in Wntresponsiveness (Fig. 4) and an expansion of Hedgehog respon-siveness (Fig. 5). Do primary cilia integrate Wnt and Hedge-hog signals, thereby regulating Gli expression? Or doesKif3a have another function besides its role as an IFT motorprotein? In some tissues, Kif3a regulates CKI-mediated phos-phorylation of Disheveled, and that loss of Kif3a leads toincreased Wnt pathway activation (23). However, we observedthat the loss of Kif3a leads to a localized reduction in Wntresponsiveness (Fig. 4). While our genetic approach providesa demonstration that b-catenin-dependent Wnt signaling isaffected by truncating the primary cilia on cranial neuralcrest cells, there is no proof that it is direct. In light of thesedata, more questions are raised than answered about whetherprimary cilia are absolutely required for Wnt pathway activity.

Mechanistic insights into craniofacial deformities

We found that a loss of primary cilia in the facial mesenchymeresulted in expanded Hedgehog activity, including the upregulation of Gli1 and simultaneous down regulation of Gli3(Figs 6 and 8), resulting in inappropriate cell proliferation inthe facial mesenchyme. There are a number of human con-ditions where excessive Hedgehog signaling leads to uncon-trolled cell proliferation. For example, Gorlin syndromepatients carry mutations in Ptc1 or Smo that lead to ectopicHedgehog pathway activity (64–66). These mutations resultin the characteristic, hyper-proliferative nevi that predomi-nantly mark the patients’ faces and trunk, and which fre-quently transform into basal cell carcinomas (31,67). Inaddition to these skin lesions, Gorlin syndrome patientsexhibit moderate to severe hypertelorism, a broad nasal root,large heads (macrocephaly) and an increased incidence offacial clefting (31). Our data suggest that mutations resultingin excessive Hedgehog signaling lead directly to increasedproliferation of neural crest cells and that this can manifestas hypertelorism and in extreme cases, FND.

Syndromes such as FND (OMIM 136760; which includefrontonasal malformation; median facial cleft syndrome andfrontorhiny), acrofrontofacionasal dysostosis 1 (OMIM201180), frontofacionasal dysplasia (OMIM 229400), ocu-loauriculofrontonasal syndrome (OMIM 601452) and cranio-frontonasal syndrome (OMIM 304110) all present with amidline expansion accompanied by hypertelorism, a flatbroad nose, and cleft lip and/or cleft palate. Although FNDshave been attributed to neural tube defects (7), abnormal celldeath or aberrant patterns of vascularization in the face (31),our data and one previous report (68) argue for a differentetiology. We propose that some forms of FND are due to again-of-Hedgehog function and Hedgehog-dependent hyper-proliferation of the cranial neural crest. Finally, in view ofthe fact that the basis of the Kif3a CKO phenotype was



ciliopathic, we put forward the possibility that genes encodingciliary proteins could now represent a novel class of candi-dates for genetic analysis of disorders characterized by ahyperteloric phenotype.

MATERIALS AND METHODS

Embryo collection and preparation of tissues

Embryos were collected in 48C PBS then fixed in 4% parafor-maldehyde (PFA) overnight at 48C, dehydrated through anethanol series and stored in 100% ethanol. Most tissueswere embedded in paraffin and cut at 8 mm using a standardmicrotome.

Immunohistochemistry

For acetylated and gamma tubulin antibody staining,de-paraffinized tissue sections were immersed in coldacetone and treated with 0.1% TritonX-100. Sections wereincubated overnight at 48C with a 1:1000 dilution of incu-bation in monoclonal anti-acetylated tubulin and anti-gammatubulin (Sigma, St Louis, MO, USA) 1% donkey IgG over-night, and incubated for 1 h at room temperature in a 1:1000dilution of FITC conjugated anti-mouse secondary antibody(Jackson Immunoresearch) in 1% donkey IgG. Slides werewashed in 1:10 000 dilution of Hoechst in PBS. Arl13b (giftfrom T. Caspary) staining was performed under similarconditions with a 1:500 dilution for the primary antibody

incubation and a 1:200 dilution for secondary antibody incu-bation. For AP2a (Developmental Studies HybridomaBank), similar conditions were used with a 1:25 dilution forprimary antibody incubation and 1:200 dilution for secondaryantibody incubation.

For BrdU antibody staining, pregnant dams were givenintraperitoneal injections of BrdU labeling reagent (Zymed,Carlsbad, CA, USA) and euthanized 2 h post injection. BrdUdetection was carried out as per manufacturer’s instructions(Zymed, Carlsbad, CA, USA).

Whole mount and histological staining

Whole-mount bone and cartilage staining of embryos was per-formed as described (69). Movat’s Pentachrome stain wascarried out as previously described (70) and Safranin-O stain-ing was used to detect cartilage. Briefly slides were de-waxed,washed in Weigert’s hematoxylin, acid alcohol, aqueous fastgreen, 1% acetic acid and 0.1% aqueous Safranin-O.

Generating Kif3a CKO, Kif3a CKO::TOPgal, Kif3aCKO::Ptc-LacZ mice and Wnt1-Cre::R26R::Kif3aCKO

All mouse experiments were done in accordance with theStanford University institutional guidelines. The Cre-lox Psystem was used to generate mice in which Kif3a is con-ditionally inactivated in the neural crest. Female mice, homo-zygous for the floxed Kif3a allele (Kif3afl/fl) (71), were crossed

Figure 9. A molecular model of cosmobia. (A) Schematic diagram indicating Hedgehog levels in embryonic tissue sections corresponding to various intervals ofthe cosmobic spectrum. (A) ‘Normal state’ has moderate Hedgehog activity in the midline (light blue). Various degrees of midline expansion including hyperte-lorism, facial duplications and diprosopus have excessive Hedgehog activity (dark blue). Various degrees of midline collapse including cyclopia and hypotelor-ism are caused by insufficient levels of Hedgehog activity (white). (B) Facial phenotypes of Wilder’s cosmobic spectrum. (C) Previously identified craniofacialdisorders that have been attributed to a loss or gain in Hedgehog activity. We propose that craniofacial ciliopathies can now be added to this list. Greig cepha-lopolysyndactyly syndrome (GCPS), Smith-Lemli-Opitz Syndrome (SLOS).



with males, heterozygous for the floxed Kif3a allele and eitherheterozygous or homozygous for Wnt1-Cre (Wnt1-Cre Kif3afl/þ)mice, to produce Kif3a CKO mutant embryos. Genotypewas confirmed by PCR for the Kif3a allele using wild-type(50-TCTGTGAGTTTGTGACCAGCC-30), common (50-AGGGCAGACGGAAGGGTGG-30) and deletion (50-TGGCAGGTCAATGGACGCAG-30) primers. The floxed Kif3a allelegenerated a 490 bp product, the wild-type allele a 360 bpproduct and the null allele a 200 bp product. Wnt1-Cre:R26R:Kif3afl/þ males were crossed with Kif3afl/fl females to createWnt1-Cre::R26R::Kif3aCKO embryos.

Kif3a CKO::TOPgal mice were generated by crossingTOPgal mice, which carry a b-galactosidase transgene down-stream of a c-fos minimal promoter and three consensusTcf-binding motifs (36), to mice homozygous for the floxedKif3a. To produce Kif3a CKO::Ptc-LacZ mice, Ptc-LacZmice (72) were bred into Kif3afl/fl mice and the Kif3afl/fl

Ptc-LacZ offspring were then crossed to Wnt1-Cre Kif3afl/þ

males to generate Kif3a CKO::Ptc-LacZ.

LacZ detection

b-Galactosidase activity was visualized by Xgal staining. Inbrief, freshly collected tissues were fixed with 0.2% glutaralde-hyde for 15 min, and stained with Xgal (5-bromo-4-chloro-3-indolyl-beta-D-galactopyranoside; Invitrogen, Carlsbad,CA, USA) overnight at 378C (73). For Xgal staining oftissue sections, freshly collected embryos were fixed in 0.4%paraformaldehyde for 2 h, infused with 30% sucrose for24 h, embedded in OCT media, frozen with dry ice in isopen-tane and cryosectioned. Xgal staining for cryosections fol-lowed the same protocol.

In situ hybridization

Templates for the relevant mRNAs for in situ hybridizationwere amplified from embryonic mouse cDNA by PCR usingsequence-specific primers which included the promoter sitesfor T3 or T7 RNA polymerase. Antisense riboprobe for eachgene was transcribed with either T3 or T7 RNA polymerasein the presence of Dig-11-UTP (Roche; Indianapolis, IN,USA). Whole-mount and section in situ hybridizations wereperformed by incubating tissue sections in hybridizationbuffer (Ambion Corporation, Austin, TX, USA) at 708C for12 h, followed by the addition of riboprobe (approximate con-centration of 0.2–0.3 mg/ml probe per kilobase of probe com-plexity). Non-specifically bound probe was removed with highstringency washes (0.1� SSC, 658C). For color detection,embryos or slides were blocked with 10% sheep serum, 1%Boehringer-Mannheim Blocking Reagent (Roche, Indianapo-lis, IN, USA) and levamisole, and developed using nitroblue tetrazolium chloride (NBT; Roche, Indianapolis, IN,USA) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP;Roche, Indianapolis, IN, USA). After developing, the slideswere cover-slipped with aqueous mounting medium.

Tissue harvest and primary cell culture

Frontonasal neural crest cells were isolated from e11.5 wild-type and Kif3a CKO embryos. Embryos were collected,

dissected in cold, sterile PBS and digested in 1.26 U/mlDispase (BD Biosciences, San Jose, CA, USA). Facial ecto-derm and neuroectoderm were removed from the frontonasalprominences with tungsten needles. Isolated frontonasalneural crest cells were centrifuged, resuspended and platedin standard growth medium, containing DMEM, 10% FBSand 100 IU/ml penicillin/streptomycin. Cells were expandedfor a period of 2 days at 378C, 21% O2, 5% CO2. Cellswere passaged by trypsinization; only passage 1 cells were uti-lized for analyses.

In vitro BrdU incorporation assay

Cellular proliferation was assessed by bromodeoxyuridine(BrdU) incorporation assays (74). Briefly, e11.5 wild-typeand Kif3a CKO neural crest cells were seeded in 96-wellplates (1000 cells/well, n ¼ 6), treated with 0, 100 or250 ng/ml of N-terminal Sonic Hedgehog (Shh-N) (R&D,Minneapolis, MN, USA) or vehicle as a control (0.1% PBS).After 48 h, BrdU assays were performed according to the man-ufacturer’s instructions (Roche Applied Science, Indianapolis,IN, USA). Means and standard deviations were calculated.

RNA isolation and quantitative real-time polymerase chainreaction

Cranial neural crest cells from the frontonasal prominencewere isolated from e11.5 wild-type and Kif3a CKOembryos. Tissues were snap-frozen and homogenized by soni-cation, and RNA isolation was performed with the RNeasyMini Kit (Qiagen Sciences, MD, USA). After DNase treat-ment, reverse transcription was performed with TaqmanReverse Transcription Reagents (Applied Biosystems, FosterCity, CA, USA). Quantitative real-time polymerase chainreaction was carried out using the Applied Biosystems Prism7900HT Sequence Detection System and Power Sybr GreenMaster Mix (Applied Biosystems, Foster City, CA, USA).Specific primers for the genes examined were designedbased on their PrimerBank (http://pga.mgh.harvard.edu/primerbank) sequence. Primers were first tested to determineoptimal concentrations, and products were run on a 2%agarose gel to confirm the appropriate size and RNA integrity.The levels of gene expression were determined by normalizingto the values of GAPDH. All reactions were performed intriplicate.

Talpid2 embryo collection

Fertilized eggs were obtained from Talpid2 flocks maintainedat the University of California, Davis. All eggs were incubatedat 388C in a humidified forced draft chamber and were stagedas previously described (75).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.




ACKNOWLEDGEMENTS

We would like to thank Sue McConnell (Stanford University)for kindly sharing Kif3afl/fl and Kif3afl/þ::Wnt1-Cre mice,Tamara Caspary (Emory University) for kindly providing uswith Arl13b antibody and Jackie Pisenti at UC Davis for sup-plying Talpid2 embryos.

Conflict of Interest statement. None declared.

FUNDING

This work was supported by the March of Dimes (FY06-335to J.A.H.); the National Institutes of Health(NRSA-F32DE017499-01 to S.A.B., RO1-DE012462-06A1to J.A.H.); and the Lucile Packard Foundation for Children’sHealth (1K99DE019853-01 to S.A.B.).

REFERENCES

1. Ming, J.E. and Muenke, M. (1998) Holoprosencephaly: from homer tohedgehog. Clin. Genet., 53, 155–163.

2. Muenke, M. and Beachy, P.A. (2000) Genetics of ventral forebraindevelopment and holoprosencephaly. Curr. Opin. Genet. Dev, 10,262–269.

3. Cohen, M.M.J. and Sulik, K.K. (1992) Perspectives onholoprosencephaly: Part II. Central nervous system, craniofacial anatomy,syndrome commentary, diagnostic approach, and experimental studies.J. Craniofac. Genet. Dev. Biol., 12, 196–244.

4. Ming, J.E. and Muenke, M. (2002) Multiple hits during early embryonicdevelopment: digenic diseases and holoprosencephaly. Am. J. Hum.Genet., 71, 1017–1032.

5. McKusick, V.A. (2000) Online Mendelian Inheritance in Man, OMIM(TM). McKusick-Nathans Institute for Genetic Medicine, Johns HopkinsUniversity, Baltimore, MD and National Center for BiotechnologyInformation, National Library of Medicine, Bethesda, MD, Vol. 2002.

6. Schachter, K.A. and Krauss, R.S. (2008) Murine models ofholoprosencephaly. Curr. Top. Dev. Biol., 84, 139–170.

7. Wu, J., Staffenberg, D.A., Mulliken, J.B. and Shanske, A.L. (2002)Diprosopus: a unique case and review of the literature. Teratology, 66,282–287.

8. Guion-Almeida, M.L., Richieri-Costa, A., Saavedra, D. and Cohen,M.M. Jr (1996) Frontonasal dysplasia: analysis of 21 cases and literaturereview. Int. J. Oral Maxillofac. Surg., 25, 91–97.

9. Nanni, L., Ming, J.E., Bocian, M., Steinhaus, K., Bianchi, D.W.,Die-Smulders, C., Giannotti, A., Imaizumi, K., Jones, K.L., Campo, M.D.et al. (1999) The mutational spectrum of the sonic hedgehog gene inholoprosencephaly: SHH mutations cause a significant proportion ofautosomal dominant holoprosencephaly. Hum. Mol. Genet., 8,2479–2488.

10. Cordero, D., Marcucio, R., Hu, D., Gaffield, W., Tapadia, M. and Helms,J.A. (2004) Temporal perturbations in sonic hedgehog signaling elicitthe spectrum of holoprosencephaly phenotypes. J. Clin. Invest., 114,485–494.

11. Marcucio, R., Cordero, D.R., Hu, D. and Helms, J.A. (2005) Molecularinteractions coordinating development of the forebrain and face. Dev.Biol., 284, 48–61.

12. Zhang, W., Kang, J.S., Cole, F., Yi, M.J. and Krauss, R.S. (2006) Cdofunctions at multiple points in the Sonic Hedgehog pathway, andCdo-deficient mice accurately model human holoprosencephaly. Dev.Cell., 10, 657–665.

13. Allen, B.L., Tenzen, T. and McMahon, A.P. (2007) TheHedgehog-binding proteins Gas1 and Cdo cooperate to positively regulateShh signaling during mouse development. Genes Dev., 21, 1244–1257.

14. Twigg, S.R., Versnel, S.L., Nurnberg, G., Lees, M.M., Bhat, M.,Hammond, P., Hennekam, R.C., Hoogeboom, A.J., Hurst, J.A., Johnson,D. et al. (2009) Frontorhiny, a distinctive presentation of frontonasaldysplasia caused by recessive mutations in the ALX3 homeobox gene.Am. J. Hum. Genet., 84, 698–705.

15. Twigg, S.R., Kan, R., Babbs, C., Bochukova, E.G., Robertson, S.P., Wall,S.A., Morriss-Kay, G.M. and Wilkie, A.O. (2004) Mutations of ephrin-B1(EFNB1), a marker of tissue boundary formation, cause craniofrontonasalsyndrome. Proc. Natl Acad. Sci. USA, 101, 8652–8657.

16. Fogelgren, B., Kuroyama, M.C., McBratney-Owen, B., Spence, A.A.,Malahn, L.E., Anawati, M.K., Cabatbat, C., Alarcon, V.B., Marikawa, Y.and Lozanoff, S. (2008) Misexpression of Six2 is associated with heritablefrontonasal dysplasia and renal hypoplasia in 3H1 Br mice. Dev. Dyn.,237, 1767–1779.

17. Vortkamp, A., Gessler, M. and Grzeschik, K.H. (1991) GLI3 zinc-fingergene interrupted by translocations in Greig syndrome families. Nature,352, 539–540.

18. Hu, D. and Helms, J.A. (1999) The role of sonic hedgehog in normal andabnormal craniofacial morphogenesis. Development, 126, 4873–4884.

19. Brugmann, S.A., Goodnough, L.H., Gregorieff, A., Leucht, P., Ten Berge,D., Fuerer, C., Clevers, H., Nusse, R. and Helms, J.A. (2007) Wntsignaling mediates regional specification in the vertebrate face.Development, 134, 3283–3295.

20. Corbit, K.C., Aanstad, P., Singla, V., Norman, A.R., Stainier, D.Y. andReiter, J.F. (2005) Vertebrate smoothened functions at the primary cilium.Nature, 437, 1018–1021.

21. Rohatgi, R., Milenkovic, L. and Scott, M.P. (2007) Patched1 regulateshedgehog signaling at the primary cilium. Science, 317, 372–376.

22. Haycraft, C.J., Zhang, Q., Song, B., Jackson, W.S., Detloff, P.J., Serra, R.and Yoder, B.K. (2007) Intraflagellar transport is essential forendochondral bone formation. Development, 134, 307–316.

23. Corbit, K.C., Shyer, A.E., Dowdle, W.E., Gaulden, J., Singla, V., Chen,M.H., Chuang, P.T. and Reiter, J.F. (2008) Kif3a constrainsbeta-catenin-dependent Wnt signalling through dual ciliary andnon-ciliary mechanisms. Nat. Cell Biol., 10, 70–76.

24. Kondo, S., Sato-Yoshitake, R., Noda, Y., Aizawa, H., Nakata, T.,Matsuura, Y. and Hirokawa, N. (1994) KIF3A is a new microtubule-basedanterograde motor in the nerve axon. J. Cell Biol., 125, 1095–1107.

25. Alieva, I.B., Gorgidze, L.A., Komarova, Y.A., Chernobelskaya, O.A. andVorobjev, I.A. (1999) Experimental model for studying the primary ciliain tissue culture cells. Membr. Cell Biol., 12, 895–905.

26. Caspary, T., Larkins, C.E. and Anderson, K.V. (2007) The gradedresponse to Sonic Hedgehog depends on cilia architecture. Dev. Cell, 12,767–778.

27. Takeda, S., Yonekawa, Y., Tanaka, Y., Okada, Y., Nonaka, S. andHirokawa, N. (1999) Left-right asymmetry and kinesin superfamilyprotein KIF3A: new insights in determination of laterality and mesoderminduction by kif3A2/2 mice analysis. J. Cell Biol., 145, 825–836.

28. Marszalek, J.R., Liu, X., Roberts, E.A., Chui, D., Marth, J.D., Williams,D.S. and Goldstein, L.S. (2000) Genetic evidence for selective transport ofopsin and arrestin by kinesin-II in mammalian photoreceptors. Cell, 102,175–187.

29. Chai, Y., Jiang, X., Ito, Y., Bringas, P. Jr, Han, J., Rowitch, D.H., Soriano,P., McMahon, A.P. and Sucov, H.M. (2000) Fate of the mammaliancranial neural crest during tooth and mandibular morphogenesis.Development, 127, 1671–1679.

30. Osumi-Yamashita, N., Ninomiya, Y., Doi, H. and Eto, K. (1994) Thecontribution of both forebrain and midbrain crest cells to the mesenchymein the frontonasal mass of mouse embryos. Dev. Biol., 164, 409–419.

31. Gorlin, R.J., Cohen, M.M. and Levin, L.S. (1990) Syndromes of the Head

and Neck., 3rd edn. Oxford University Press, New York.

32. Guion-Almeida, M.L. and Richieri-Costa, A. (2001) Frontonasaldysplasia, macroblepharon, eyelid colobomas, ear anomalies,macrostomia, mental retardation and CNS structural anomalies: definingthe phenotype. Clin. Dysmorphol., 10, 81–86.

33. Koyama, E., Young, B., Nagayama, M., Shibukawa, Y.,Enomoto-Iwamoto, M., Iwamoto, M., Maeda, Y., Lanske, B., Song, B.,Serra, R. et al. (2007) Conditional Kif3a ablation causes abnormalhedgehog signaling topography, growth plate dysfunction, and excessivebone and cartilage formation during mouse skeletogenesis. Development,134, 2159–2169.

34. Eggenschwiler, J.T. and Anderson, K.V. (2007) Cilia and developmentalsignaling. Annu. Rev. Cell Dev. Biol., 23, 345–373.

35. Ocbina, P.J., Tuson, M. and Anderson, K.V. (2009) Primary cilia are notrequired for normal canonical Wnt signaling in the mouse embryo. PLoS

One, 4, e6839.



36. DasGupta, R. and Fuchs, E. (1999) Multiple roles for activated LEF/TCFtranscription complexes during hair follicle development anddifferentiation. Development, 126, 4557–4568.

37. Sarkar, L., Cobourne, M., Naylor, S., Smalley, M., Dale, T. and Sharpe,P.T. (2000) Wnt/Shh interactions regulate ectodermal boundary formationduring mammalian tooth development. Proc. Natl Acad. Sci. USA, 97,4520–4524.

38. Han, Y.G., Spassky, N., Romaguera-Ros, M., Garcia-Verdugo, J.M.,Aguilar, A., Schneider-Maunoury, S. and Alvarez-Buylla, A. (2008)Hedgehog signaling and primary cilia are required for the formation ofadult neural stem cells. Nat. Neurosci., 11, 277–284.

39. Spassky, N., Han, Y.G., Aguilar, A., Strehl, L., Besse, L., Laclef, C., Ros,M.R., Garcia-Verdugo, J.M. and Alvarez-Buylla, A. (2008) Primary ciliaare required for cerebellar development and Shh-dependent expansion ofprogenitor pool. Dev. Biol., 317, 246–259.

40. Huangfu, D., Liu, A., Rakeman, A.S., Murcia, N.S., Niswander, L. andAnderson, K.V. (2003) Hedgehog signalling in the mouse requiresintraflagellar transport proteins. Nature, 426, 83–87.

41. Chiang, C., Litingtung, Y., Lee, E., Young, K.E., Corden, J.L., Westphal,H. and Beachy, P.A. (1996) Cyclopia and defective axial patterning inmice lacking Sonic hedgehog gene function. Nature, 383, 407–413.

42. Jeong, J., Mao, J., Tenzen, T., Kottmann, A.H. and McMahon, A.P. (2004)Hedgehog signaling in the neural crest cells regulates the patterning andgrowth of facial primordia. Genes Dev., 18, 937–951.

43. Huangfu, D. and Anderson, K.V. (2005) Cilia and Hedgehogresponsiveness in the mouse. Proc. Natl Acad. Sci. USA, 102, 11325–11330.

44. Goodrich, L.V., Johnson, R.L., Milenkovic, L., McMahon, J.A. and Scott,M.P. (1996) Conservation of the hedgehog/patched signaling pathwayfrom flies to mice: induction of a mouse patched gene by Hedgehog.Genes Dev., 10, 301–312.

45. Benouaiche, L., Gitton, Y., Vincent, C., Couly, G. and Levi, G. (2008)Sonic hedgehog signalling from foregut endoderm patterns the avian nasalcapsule. Development, 135, 2221–2225.

46. Etchevers, H.C., Couly, G., Vincent, C. and Le Douarin, N.M. (1999)Anterior cephalic neural crest is required for forebrain viability.Development, 126, 3533–3543.

47. Lee, J., Platt, K.A., Censullo, P., Ruiz, I. and Altaba, A. (1997) Gli1 is atarget of Sonic hedgehog that induces ventral neural tube development.Development, 124, 2537–2552.

48. Hynes, M., Stone, D.M., Dowd, M., Pitts-Meek, S., Goddard, A., Gurney,A. and Rosenthal, A. (1997) Control of cell pattern in the neural tubeby the zinc finger transcription factor and oncogene Gli-1. Neuron, 19,15–26.

49. Tobin, J.L., Di Franco, M., Eichers, E., May-Simera, H., Garcia, M., Yan,J., Quinlan, R., Justice, M.J., Hennekam, R.C., Briscoe, J. et al. (2008)Inhibition of neural crest migration underlies craniofacial dysmorphologyand Hirschsprung’s disease in Bardet-Biedl syndrome. Proc. Natl Acad.Sci. USA, 105, 6714–6719.

50. Cole, R.K. (1942) The ‘talpid’ lethal in the domestic fowl. J. Hered., 33,82–86.

51. Schneider, R.A., Hu, D. and Helms, J.A. (1999) From head to toe:conservation of molecular signals regulating limb and craniofacialmorphogenesis. Cell Tissue Res., 296, 103–109.

52. Yin, Y., Bangs, F., Paton, I.R., Prescott, A., James, J., Davey, M.G.,Whitley, P., Genikhovich, G., Technau, U., Burt, D.W. et al. (2009) TheTalpid3 gene (KIAA0586) encodes a centrosomal protein that is essentialfor primary cilia formation. Development, 136, 655–664.

53. Caruccio, N.C., Martinez-Lopez, A., Harris, M., Dvorak, L., Bitgood, J.,Simandl, B.K. and Fallon, J.F. (1999) Constitutive activation of sonichedgehog signaling in the chicken mutant talpid(2): Shh-independentoutgrowth and polarizing activity. Dev. Biol., 212, 137–149.

54. Wilder, H. (1908) The morphology of Cosmobia; speculations concerningthe significance of certain types of monsters. Am. J. Anat., 8, 355–440.

55. Belloni, E., Muenke, M., Roessler, E., Traverso, G., Siegel-Bartelt, J.,Frumkin, A., Mitchell, H.F., Donis-Keller, H., Helms, C., Hing, A.V. et al.

(1996) Identification of Sonic hedgehog as a candidate gene responsiblefor holoprosencephaly. Nat. Genet., 14, 353–356.

56. Kolpakova-Hart, E., Jinnin, M., Hou, B., Fukai, N. and Olsen, B.R. (2007)Kinesin-2 controls development and patterning of the vertebrateskeleton by Hedgehog- and Gli3-dependent mechanisms. Dev. Biol., 309,273–284.

57. Bisgrove, B.W. and Yost, H.J. (2006) The roles of cilia in developmentaldisorders and disease. Development, 133, 4131–4143.

58. Haycraft, C.J., Banizs, B., Aydin-Son, Y., Zhang, Q., Michaud, E.J. andYoder, B.K. (2005) Gli2 and Gli3 localize to cilia and require theintraflagellar transport protein polaris for processing and function. PLoSGenet., 1, e53.

59. May, S.R., Ashique, A.M., Karlen, M., Wang, B., Shen, Y., Zarbalis, K.,Reiter, J., Ericson, J. and Peterson, A.S. (2005) Loss of the retrogrademotor for IFT disrupts localization of Smo to cilia and prevents theexpression of both activator and repressor functions of Gli. Dev. Biol.,287, 378–389.

60. Bai, C.B., Auerbach, W., Lee, J.S., Stephen, D. and Joyner, A.L. (2002)Gli2, but not Gli1, is required for initial Shh signaling and ectopicactivation of the Shh pathway. Development, 129, 4753–4761.

61. Johnson, D.R. (1967) Extra-toes: a new mutant gene causing multipleabnormalities in the mouse. J. Embryol. Exp. Morph., 17, 543–581.

62. Hui, C.C. and Joyner, A.L. (1993) A mouse model of greigcephalopolysyndactyly syndrome: the extra-toesJ mutation contains anintragenic deletion of the Gli3 gene. Nat. Genet., 3, 241–246.

63. Alvarez-Medina, R., Cayuso, J., Okubo, T., Takada, S. and Marti, E.(2008) Wnt canonical pathway restricts graded Shh/Gli patterning activitythrough the regulation of Gli3 expression. Development, 135, 237–247.

64. Aszterbaum, M., Rothman, A., Johnson, R.L., Fisher, M., Xie, J., Bonifas,J.M., Zhang, X., Scott, M.P. and Epstein, E.H. Jr (1998) Identification ofmutations in the human PATCHED gene in sporadic basal cell carcinomasand in patients with the basal cell nevus syndrome. J. Invest. Dermatol.,110, 885–888.

65. Villavicencio, E.H., Walterhouse, D.O. and Iannaccone, P.M. (2000) Thesonic hedgehog-patched-gli pathway in human development and disease.Am. J. Hum. Genet., 67, 1047–1054.

66. Xie, J., Murone, M., Luoh, S.M., Ryan, A., Gu, Q., Zhang, C., Bonifas,J.M., Lam, C.W., Hynes, M., Goddard, A. et al. (1998) ActivatingSmoothened mutations in sporadic basal-cell carcinoma. Nature, 391,90–92.

67. Gorlin, R.J. (2004) Nevoid basal cell carcinoma (Gorlin) syndrome.Genet. Med., 6, 530–539.

68. Stevens, C.A. and Qumsiyeh, M.B. (1995) Syndromal frontonasaldysostosis in a child with a complex translocation involving chromosomes3, 7 and 11. Am. J. Med. Genet., 55, 494–497.

69. McLeod, M.J. (1980) Differential staining of cartilage and bone in wholemouse fetuses by alcian blue and alizarin red S. Teratology, 22, 299–301.

70. Sheehan, D.C. and Hrapchak, B.B. (1980) Theory and practice ofHistotechnology, 2nd edn. Batelle Press, Columbus, Ohio.

71. Marszalek, J.R., Ruiz-Lozano, P., Roberts, E., Chien, K.R. and Goldstein,L.S. (1999) Situs inversus and embryonic ciliary morphogenesis defects inmouse mutants lacking the KIF3A subunit of kinesin-II. Proc. Natl Acad.Sci. USA, 96, 5043–5048.

72. Goodrich, L.V., Jung, D., Higgins, K.M. and Scott, M.P. (1999)Overexpression of ptc1 inhibits induction of Shh target genes and preventsnormal patterning in the neural tube. Dev. Biol., 211, 323–334.

73. Maretto, S., Cordenonsi, M., Dupont, S., Braghetta, P., Broccoli, V.,Hassan, A.B., Volpin, D., Bressan, G.M. and Piccolo, S. (2003) MappingWnt/beta-catenin signaling during mouse development and in colorectaltumors. Proc. Natl Acad. Sci. USA, 100, 3299–3304.

74. James, A.W., Xu, Y., Wang, R. and Longaker, M.T. (2008) Proliferation,osteogenic differentiation, and fgf-2 modulation of posterofrontal/sagittalsuture-derived mesenchymal cells in vitro. Plast. Reconstr. Surg., 122,53–63.

75. Hamburger, V. and Hamilton, H.L. (1951) A series of normal stages in thedevelopment of the chick embryo. J. Morphol., 88, 49–92.



A primary cilia-dependent etiology for midline facial disorders

Documents

Transcript of A primary cilia-dependent etiology for midline facial disorders