Developmental Biology 336 (2009) 145–155

Contents lists available at ScienceDirect

Developmental Biology

j ourna l homepage: www.e lsev ie r.com/deve lopmenta lb io logy

Expression and requirement of T-box transcription factors Tbx2 and Tbx3 duringsecondary palate development in the mouse

Susann Zirzow a, Timo H.-W. Lüdtke a, Janynke F. Brons b, Marianne Petry a,Vincent M. Christoffels b, Andreas Kispert a,⁎a Institut für Molekularbiologie, OE5250, Medizinische Hochschule Hannover, Carl-Neuberg-Str. 1, D-30625 Hannover, Germanyb Department of Anatomy and Embryology, Academic Medical Center, University of Amsterdam, Meibergdreef 15 L2-108, 1105 AZ Amsterdam, The Netherlands

⁎ Corresponding author. Fax: +49 511 5324283.E-mail address: kispert.andreas@mh-hannover.de (A

0012-1606/$ – see front matter © 2009 Elsevier Inc. Adoi:10.1016/j.ydbio.2009.09.020

a b s t r a c t

a r t i c l e i n f o

Article history:Received for publication 14 May 2009Revised 27 August 2009Accepted 15 September 2009Available online 19 September 2009

Keywords:Secondary palateT-boxMousePhenotypic analysisCleft palateTbx2Tbx3BmpPhenotypeExpressionCyclinD1

Formation of the mammalian secondary palate is a highly regulated and complex process. Impairment of theunderlying cellular and molecular programs often results in cleft palate, a common birth defect in mammals.Here we report that Tbx2 and Tbx3, two closely related genes encoding T-box transcription factors, areexpressed in the mesenchyme of the mouse palatal structures during development. Mice homozygousmutant for Tbx2 and mice double heterozygous for Tbx2 and Tbx3 exhibit a cleft palate phenotype arguingfor an important contribution of Tbx2 and Tbx3 to palatogenesis. In Tbx2-deficient embryos, the bilateralprimordial palatal shelves form but are smaller and retarded in the outgrowth process. They do not makecontact but retain the potential to fuse. Development of other craniofacial structures appears normal,suggesting that impaired palate formation in Tbx2-mutant mice is caused by a primary defect in the palatalshelf mesenchyme. This is further supported by increased cell proliferation and apoptosis accompanied byincreased expression of Bmp4 and CyclinD1 in Tbx2-deficient palatal shelves. Hence, Tbx2 and Tbx3 functionoverlappingly to control growth of the palatal shelf mesenchyme.

© 2009 Elsevier Inc. All rights reserved.

Introduction

Cleft palate without or with association of a cleft lip (CP, CL/P) isthe most frequent human birth defect. This complex trait is caused bymultiple genetic disorders but also by environmental perturbations ofthe development of the secondary palate (Gritli-Linde, 2007; Murray,2002). The mammalian secondary palate arises from the medial sidesof the maxillary processes that flank the oral cavity. In the mouse, thebilateral palatal shelves (PS) initiate outgrowth at embryonic day (E)11.5 with a vertical movement down the side of the tongue. With theexpansion of the lower jaw, the tongue drops in the oral cavityproviding space for an elevation of the PS.With continued growth, thePS oppose in the midline and finally fuse. The multilayered epithelialseam initially present at the midline subsequently degenerates,establishing a continuous mesenchyme across the whole roof of theoral cavity at E15.5 (Ferguson, 1988).

PS morphogenesis depends on survival, continued proliferation,migration and differentiation of mesenchymal cells and their inter-

. Kispert).

ll rights reserved.

action with the surrounding epithelial cover. Impairment of any ofthese processes will delay or inhibit palatal outgrowth resulting in PSnon-fusion. The failure to separate the oral from the nasal cavity willlead to early postnatal lethality in mammals (Gritli-Linde, 2007; Rice,2005).

T-box (Tbx) genes encode a conserved family of transcriptionfactors that play important roles in numerous processes duringembryonic development including palatogenesis (Naiche et al.,2005). Phenotypic analyses of mice and human patients haveidentified three family members as critical mediators in this programto date. Tbx1-deficient mice display a wide range of developmentalanomalies including CP (Jerome and Papaioannou, 2001; Merscheret al., 2001). The phenotypic spectrum of Tbx1-deficient miceresembles that of heterozygous patients of DiGeorge/velocardiofacialsyndrome (DGS/VCFS), a common human disorder, usually associ-ated with deletions of chromosome 22q11 in which TBX1 resides(Shprintzen et al., 2005). Cleft palate in association with ankyloglos-sia, a rare disorder with X-linked inheritance (CPX) has recently beenshown to result from mutations in the T-box gene family memberTBX22 (Braybrook et al., 2001). Tbx22-deficient mice have a sub-mucous cleft palate (SMCP) and ankyloglossia, similar to the humanphenotype, with a small minority showing overt clefts (Pauws et al.,

146 S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

2009). Mice carrying the spontaneous Dancer (Dc) mutation exhibitCL/P in homozygotes and show significantly increased susceptibilityto CL/P in heterozygotes. Chromosomal analysis of the Dc locusrevealed an insertional mutation that causes ectopic expression of ap23-Tbx10 chimeric transcript indicating that gain of expression ofa T-box transcription factor gene can also cause CP pathogenesis(Bush et al., 2004).

Tbx2 and Tbx3 encode a closely related pair of T-box proteinswhose high sequence conservation, equivalent function as transcrip-tional repressors on common DNA binding sites and shared ex-pression domains suggest that the two genes arose by a recent geneduplication event during vertebrate evolution (Harrelson et al., 2004;Naiche et al., 2005). Phenotypic analyses of mutant mice have re-vealed unique functions of Tbx2 in heart and limb formation(Aanhaanen et al., 2009; Harrelson et al., 2004) and of Tbx3 in thedevelopment of limbs, yolk sac, mammary gland, heart and liver(Bakker et al., 2008; Davenport et al., 2003; Hoogaars et al., 2007b;Ludtke et al., 2009; Suzuki et al., 2008). Expression analysis and organculture experiments have implicated Tbx3 in control of cell prolifer-ation in the PS but an in vivo requirement for Tbx3 has not yet beenshown (Lee et al., 2007). Unfortunately, analysis of compoundmutantembryos to test the functional redundancy at sites of co-expression ofTbx2 and Tbx3 has been hampered by the difficulty of propagatingmice double heterozygous for both mutant alleles (Jerome-Majewskaet al., 2005).

Here, we extend the functional analysis of T-box genes in mousepalatogenesis and show that Tbx2 and Tbx3 are co-expressed duringpalate formation and are required for the formation of a secondarypalate. Mice homozygous mutant for Tbx2 and mice double hetero-zygous for Tbx2 and Tbx3 exhibit a CP phenotype arguing for animportant contribution of Tbx2 and Tbx3 to palatogenesis. We deter-mine the temporal onset of the phenotype and characterize theunderlying histochemical and molecular changes. We conclude thatformation of the mammalian secondary palate like formation of otherorgans requires the precise interplay of T-box transcription factoractivities and signaling pathways.

Materials and methods

Mice

Mice carrying a null allele of Tbx2 (Tbx2tm1.1(cre)Vmc, synonym:Tbx2cre) (Aanhaanen et al., 2009) or Tbx3 (Tbx3tm1.1(cre)Vmc, synonym:Tbx3cre) (Hoogaars et al., 2007b) were maintained on an NMRIoutbred and FvB/N inbred background, respectively. For the gene-ration of Tbx2 or Tbx3 mutant embryos, mice heterozygous for eithermutant allele were intercrossed. Mice double heterozygous for Tbx2and Tbx3 mutant alleles were mated to obtain compound mutantembryos. Embryos for Tbx2 and Tbx3 expression analysis werederived from matings of NMRI wildtype mice. For timed pregnancies,vaginal plugs were checked in the morning after mating, noon wastaken as embryonic day (E) 0.5. Embryos were harvested in PBS, fixedin 4% paraformaldehyde overnight and stored in 100% methanol at−20 °C before further use.Wildtype littermateswere used as controls.Genomic DNA prepared from yolk sacs or tail biopsies was used forgenotyping by PCR (Hoogaars et al., 2007b).

Organ culture

PS were isolated from E12.5 mouse maxillae and cultured inmedium with fetal bovine serum at 37 °C and 5% CO2 for 48 h using aslightmodification of the culturemethod reported by Trowell (Taya etal., 1999). The culture medium (DMEM/F12, Gibco) was supplemen-ted with 20 μg/ml ascorbic acid (Sigma) and 1% penicillin/streptomycin. Tissues were then fixed and processed for histologicalanalysis.

Histological analysis

Heads of the embryos and organ culture tissues were embedded inparaffin wax and sectioned to 5 μm. Sections were stained withhematoxylin and eosin.

Cellular assays

Cell proliferation in tissues of E12.5 and E14.0 embryos wasinvestigated by the detection of incorporated BrdU on 5-μm sectionsof paraffin wax-embedded specimens, similar to published protocols(Bussen et al., 2004). Four sections each of anterior and posteriorpalate were taken from three embryos of each genotype at both stagesfor quantification. The BrdU-labeling indexwas defined as the numberof BrdU-positive nuclei relative to the total number of nuclei asdetected by DAPI counterstain in the mesenchyme of PS tip region.Detection of apoptotic cells in 5-μm paraffin sections of E11.5-14.0embryos was based on the modification of genomic DNA usingterminal deoxynucleotidyl transferase (TUNEL assay) and indirectdetection of positive cells by fluorescein-conjugated anti-digoxigeninantibody. The procedure followed exactly the recommendation of themanufacturer (Serologicals Corp.) of the ApopTag kit used. The rate ofapoptosis was defined as the number of TUNEL positive nuclei relativeto the total number of nuclei as detected by DAPI counterstain in themesenchyme of PS tip region. Cell densities in the PSwere determinedby counting DAPI-positive nuclei in a defined square of 60×80 μmcentered in the PS tip region. PS size was measured using ImageJsoftware (NIH freeware). Statistical analysis was performed using thetwo-tailed Student's t-test. Data were expressed as mean +/−standard deviation. Differences were considered significant (⁎) whenthe P-value was below 0.05 and highly significant (⁎⁎) when theP-value was below 0.005.

In situ hybridization analysis

Whole mount in situ hybridization analysis was performed withdigoxigenin-labeled antisense riboprobes following a standard pro-cedure (Wilkinson, 1992). Stained specimens were transferred into80% glycerol prior to documentation. In situ hybridization analysis on10-μm frontal sections of embryo heads followed a previously des-cribed protocol (Moorman et al., 2001).

Semiquantitative reverse transcription PCR

Total RNA was extracted from microdissected E12.5 and E14.0 PSwith RNAPure reagent (Peqlab). RNA (100–500 ng) was reversetranscribed with RevertAid M-MuLV Reverse Transcriptase (Fermen-tas). For semiquantitative PCR, the number of cycles was adjusted tothe midlogarithmic phase for each PCR independently. Expressionlevels were normalised to Hprt expression. Quantification was per-formed with Quantity One software (Bio-Rad). Assays were per-formed in duplicates and with two different concentrations of cDNAon three independent pools of four PS each. Statistical analysis wasperformed using the two-tailed Student's t-test. Data were expressedas mean ± standard deviation. Differences were considered signifi-cant (⁎) when the P-value was below 0.05 and highly significant (⁎⁎)when the P-value was below 0.005.

Primers and PCR conditions are available on request.

Documentation

Whole mount specimens were photographed on a Leica M420microscope with a Fujix digital camera HC-300Z. Sections werephotographed using a Leica DM5000 microscope with a LeicaDFC300FX digital camera. All images were processed in AdobePhotoshop CS.

147S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

Results

Tbx2 and Tbx3 are co-expressed in the developing palate

Earlier studies have reported expression of Tbx2 and Tbx3 infacial primordia at E9.5 (Chapman et al., 1996) and of Tbx3 in PSmesenchyme at E13.5 and E14.5 of mouse development (Lee et al.,2007). However, a detailed analysis of expression of the two genesduring development of the maxilla and the secondary palate has notyet been documented. We therefore analyzed expression of the twogenes during all stages of palate development using in situhybridization analysis (Fig. 1). Analysis of whole heads revealedhigh expression of Tbx2 and Tbx3 in all prominences of thedeveloping face, including the fronto-nasal process, the maxillaryprocess and the cranial half of the mandible from E9.5 to E11.5 (Figs.1A–E, G). At E11.5, PS primordia became first visible as outgrowths ofthe maxillary processes on the oral side (Figs. 1E–H). Tbx2 and Tbx3were expressed on all following stages of outgrowth (E11.5–E14.0),elevation (E14.0–E14.5) and fusion (E14.5–E15.5) in the PS (Figs. 1I–X). Expression of Tbx2 and Tbx3 was confined to the mesenchymalcore of the PS (Figs. 1F, H, J, L, N, P, R, T) with Tbx3 expressionextending slightly more laterally in the posterior PS region than thatof Tbx2 (arrows in Figs. 1I, K, M, O, Q, S). After completion of thefusion process at E15.5, expression persisted in the rugae of the

Fig. 1. Tbx2 and Tbx3 are co-expressed during palate development. RNA in situ hybridizationQ, S, U, W) and on adjacent frontal sections (F, H, J, L, N, P, R, T, V, X). (E, G, I, K, M, O, Q, S, Uindicated. Black bars indicate section planes shown in the adjacent image; (F, H, J, L, N, P, R, T,indicated. Tbx2 and Tbx3 are co-expressed in the mesenchyme of the facial processes and inTbx3 (K, O, S) expression in the PS. Ea, ear vesicle; eye, eye vesicle; fnp, frontonasal process; mpsp, palatal shelf primordium; ps, palatal shelf; ru, rugue; to, tongue.

palate and in the mesenchyme around the midline epithelial seam(Figs. 1U–X).

Tbx2 and Tbx3 are required for palate development

Co-expression of Tbx2 and Tbx3 during palatogenesis argued for apossible functional overlap of the two genes in this process. Previousstudies reported embryonic lethality of Tbx2tm1Pa/tm1Pa mice at E14.5and of Tbx3tm1Pa/tm1Pa mice between E11.5 and E14.5 due tocardiovascular defects (Davenport et al., 2003; Harrelson et al.,2004), precluding analysis of a requirement of either gene inpalatogenesis. We wondered whether embryonic viability could beextended beyond completion of palatogenesis by maintaining nullalleles of Tbx2 (Tbx2cre) (Aanhaanen et al., 2009) and Tbx3 (Tbx3cre)(Hoogaars et al., 2007a) on an NMRI outbred rather than on a mixedinbred C57/129/ICR background as used in the previous studies(Davenport et al., 2003; Harrelson et al., 2004). Indeed, Tbx2cre/cre

embryos maintained on NMRI background exhibited less severecardiac defects at E14.5, were viable and represented in the expectedMendelian ratio at E18.5 (Aanhaanen et al., 2009). Tbx2cre/cre miceshowed a bilateral hindlimb-specific duplication of digit IV (datanot shown), consistent with previous observations in the Tbx2tm1Pa

strain (Harrelson et al., 2004). Inspection of whole heads and offrontal sections at E18.5 and E15.5 revealed a complete (isolated) cleft

analysis for Tbx2 and Tbx3 expression was carried out on whole heads (A–E, G, I, K, M, O,, W) show ventral views of the upper jaw region, the posterior (p)–anterior (a) axis isV, X) sections show higher magnification of the left PS, the dorsal (d)–ventral (v) axis isthe PS throughout development. Arrows indicate the spatial extent of Tbx2 (I, M, Q) andes, midline epithelial seam; mn, mandible; mx, maxilla; nc, nasal cavity; oc, oral cavity;

148 S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

of the secondary palate in 86% of embryos analyzed (18 of 21)(Figs. 2A–D) proving a requirement for Tbx2 in secondary palatedevelopment.

Mice homozygous for the Tbx3cre null allele maintained on NMRIoutbred background, similar to Tbx3tm1Pa/tm1Pa embryos on C57/129/ICR background, died around E14.5 most likely due to a combinationof cardiac, vascular and hepatic defects (Davenport et al., 2003; Ludtkeet al., 2009; Suzuki et al., 2008). In the one embryo that surviveduntil E15.5, the PS were fused in the middle region but not anteriorly(Figs. 2E–I) suggesting that PS growth and fusion occur normally butmay be delayed due to the general developmental retardation of theseembryos. Since a conditional approach is not available to date tostringently test the individual requirement of Tbx3 in palatogenesis,we analyzed embryos compoundmutant for Tbx2 and Tbx3 for defectsin this process.

Mice heterozygous for Tbx2cre or Tbx3cre alleles were pheno-typically normal on the NMRI outbred background they were main-tained on. In contrast, 38% (5 of 13) of Tbx2cre/+Tbx3cre/+ embryosharvested at E18.5 exhibited a complete CP suggesting a crucialbut less important contribution of Tbx3 to palatogenesis than Tbx2(Figs. 2J–M). The few embryos with three or four mutant alleles ofTbx2 and Tbx3 obtained from intercrosses of Tbx2cre/+Tbx3cre/+

mice died at E10.5 due to hemodynamic failure precluding furtheranalysis of the redundant role of the two genes in secondary palatedevelopment. Cardiovascular defects observed in these embryos arein agreement with a redundant function of the two genes in the

Fig. 2. Requirement of Tbx2 and Tbx3 in the development of the secondary palate. Analysis oregions at E18.5 (A, B, J, K, N, O) and at E15.5 (E), the posterior (p)–anterior (a) axis is indicaat E18.5 (C, D, L, M, P, Q) and at E15.5 (F–I), the dorsal (d)–ventral (v) axis is shown. Sectigenetic backgrounds (NMRI or FvB/N) are as indicated in the figure. 85% of Tbx2cre/cre embr86% of Tbx2cre/+Tbx3cre/+ on FvB/N genetic background (P, Q) but not Tbx3cre/cre embryos onseptum; oc, oral cavity; p, palate; ps, palatal shelf; to, tongue.

formation of the atrioventricular canal and will be presentedelsewhere.

Mice double heterozygous for null alleles of Tbx2 and Tbx3that were maintained on an FvB/N genetic background exhibitedcleft palate with a penetrance similar to Tbx2 single mutants (85%, 11of 13) (Figs. 2N–Q) arguing for a strong contribution of the geneticbackground to CP etiology in Tbx2- and Tbx3-deficient mice.

Tbx2-deficient mice exhibit retardation of PS elevation

To investigate which step of secondary palate formation isdisturbed in Tbx2cre/cre embryos, we carried out a detailed morpho-logical and histological analysis throughout palate development(Fig. 3). At E12.5, no obvious difference in morphology and histologyof the PS was observed betweenwildtype and Tbx2-deficient embryos(Figs. 3A–D). At E13.5, PS of Tbx2-deficient embryos appeared smallerand had moved less down the tongue compared to the wildtype(Figs. 3E–H). PS of wildtype embryos were elevated at E14.25 (Figs. 3I,K) and fused between E14.5 and E15.5 (Figs. 3M, O, Q, S). In contrast,in Tbx2-deficient embryos morphogenesis of the smaller PS wasseverely retarded and eventually came to a premature hold asevidenced by the appearance of elevated shelves only at E14.5 thatfailed to undergo the fusion process (Figs. 3J, L, N, P, R, T). The failureof PS fusion may be caused by a direct requirement for Tbx2 in thefusion process or by physical constrain due to the increased distancebetween the shelves. We tested these possibilities by juxtaposing PS

f defects in cleft palate formation in embryos by morphological inspection of upper jawted; and by histological staining with hematoxylin and eosin of frontal sections of headson planes are indicated by black bars in whole upper jaw preparations. Genotypes andyos (B, D) and 38% of Tbx2cre/+Tbx3cre/+ (K, M) embryos on NMRI genetic background,NMRI background (G, I) show a complete isolated cleft of the secondary palate. ns, nasal

Fig. 3. Retardation of PS growth in Tbx2-deficient embryos. Morphological (left two columns, the posterior (p)–anterior (a) axis is indicated) and histological (right two columns, thedorsal (d)–ventral (v) axis is shown) analysis of secondary palate formation in wildtype and Tbx2cre/cre mice. Developmental stages and genotypes are as indicated. In Tbx2-deficientembryos, PS are smaller, their outgrowth and elevation is retarded and fusion does not occur. md, mandible; mes, midline epithelial seam; ns, nasal septum; oc, oral cavity; ps, palatalshelf; to, tongue.

149S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

isolated from E13.5 Tbx2-deficient embryos in an organ culturesystem for 48 h. Under these conditions the PS of the Tbx2cre/cre

embryos fused as readily as the wildtype control (Fig. 4). Hence, lossof Tbx2 function primarily affects growth but not fusion of the PS.

Tbx2 plays a role in survival and proliferation of the PS mesenchyme

In agreement with the histological observations (Fig. 3) quantita-tive analysis of PS size revealed no change at E12.5 but a significantreduction in the anterior region at E14.0 in Tbx2-deficient embryos(Fig. 5A). Cell densities in both anterior and posterior aspects of the PSwas not significantly altered at both time points between wildtypeand Tbx2cre/cre embryos (Fig. 5B). To investigate the cellular mechan-isms that may underlie the decreased size and the retardation in PSmorphogenesis in Tbx2-deficient embryos, we analyzed cell prolifer-

Fig. 4. Tbx2-deficient PS retain fusion competence. Histological analysis by hematoxylin and eTbx2cre/cre PS undergo fusion as in the control. Arrows indicate position of former midline e

ation in the PS mesenchyme using the BrdU incorporation assay (Fig.5C). We detected a significant increase of the cell proliferation ratefrom 19.5% to 25.4% in the anterior but not in the posterior PS regionof Tbx2-deficient mice at E12.5. At E14.0, cell proliferation was notsignificantly changed throughout the PS mesenchyme compared tocontrol embryos (Fig. 5C).

We further performed TUNEL assays to analyze the role ofapoptosis for the defect in shape and morphogenesis of PS in theTbx2-deficient embryos (Fig. 6). We compared the situation inhomozygous Tbx2-mutant embryos with that of wildtype as well asof heterozygous control embryos to exclude a contribution of Creexpression to changes in cell survival (Naiche and Papaioannou,2007). We detected increased apoptosis throughout the PS mesen-chyme at E11.5 and in the anterior PS region at E12.5 of Tbx2cre/cre

embryos. At E13.5 and E14.0, apoptosis was unchanged compared to

osin stainings of sections of 48 h cultures of juxtaposed PS isolated from E13.5 embryos.pithelial seam.

Fig. 5. Changes of PS size, cell density and proliferation in Tbx2cre/cre embryos. (A) Analysis of PS size, measured on frontal sections of wildtype (black column) and Tbx2-deficient embryos (white column) at E12.5 (left panel) and E14.0 (right panel). PS size at E12.5 did not show highly significant differences anteriorly or posteriorly (anterior:wt: 0.0157±0.002 mm2, mut: 0.0187±0.004 mm2, p=0.03; posterior: wt: 0.0434±0.007 mm2, mut: 0.0397±0.009 mm2, p=0.19). At E14.0, the PS size was significantlydecreased in the anterior and posterior region (anterior: wt: 0.0472±0.0087 mm2, mut: 0.0188±0.0023 mm2, pb0.0001; posterior: wt: 0.0718±0.0106 mm2, mut: 0.0626±0.0093 mm2, pb0.005). (B) Cell densities were not significantly changed at E12.5 (anterior, wt: 13791±1416 cells/mm2, mut: 12416±3024 cells/mm2, p=0.209; posterior,wt: 13541±1725 cells/mm2, mut: 12041±878 cells/mm2, p=0.023) and at E14.0 (anterior: wt: 9948±1107 cells/mm2, mut: 10833±1339 cells/mm2, p=0.294; posterior,wt: 11125±2524 cells/mm2, mut: 11718±1510 cells/mm2, p=0.547). (C) Analysis of cell proliferation in frontal sections of the palate region of wildtype (wt) and Tbx2-deficient (Tbx2cre/cre) embryos at E12.5 (left panel) and E14.0 (right panel) by the BrdU incorporation assay. The PS region used for quantification is shown in the histologicalsections. Quantification of cell proliferation was performed by the ratio of BrdU-positive cells to total cell number as determined by DAPI counterstain, the BrdU-labeling index,in the analyzed area at an anterior and posterior position in the PS. The BrdU labeling index is significantly increased in the anterior region of the Tbx2-mutant PS mesenchymeat E12.5 (wt: 19.5±0.4, mut: 25.4±0.9, pb0.0001), while in the posterior region at this stage (wt: 22.5±1.7, mut: 26.0±1.6, p=0.15), and at E14.0 in the anterior region(wt: 22.2±0.4, mut: 23.0±0.5, p=0.023) and in the posterior region (wt: 30,3±0.6, mut: 28.0±0.6, p=0.009) no significant changes were detected.

150 S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

the controls (data not shown). Thus, changes in cell proliferation andsurvival precede morphological and histological alteration of PSmorphology and growth.

Bmp4 and Ccnd1 expression are up-regulated in Tbx2-deficient PS

To analyze the molecular changes underlying disturbed PSformation in Tbx2cre/cre embryos, we analyzed expression of a largepanel of genes which mutational analyses in the mouse implicatedin the palatal growth process (Gritli-Linde, 2007) by in situ hybri-dization analysis of sections and whole heads between E12.5 andE14.0.

Bone morphogenetic protein 2 and 4 (Bmp2 and Bmp4) and theirtarget gene homeobox, msh-like 1 (Msx1) are weakly expressed in thePS mesenchyme where they regulate in an autoregulatory loopproliferation mainly in the anterior region (Zhang et al., 2000, 2002).Notably, Msx1 and Tbx2/Tbx3 are targets genes of Bmp signaling in avariety of tissues including the atrioventricular canal (Yamada et al.,2000) where the encoded proteins form complexes to represstranscription of genes (Boogerd et al., 2008). Expression of Bmp2and Msx1 as well as distribution of phospho-Smad1,5,8, theintracellular mediator of Bmp signaling, appeared unchanged in theTbx2-deficient palate (Supplemental Fig. 1). Bmp4 levels appearedslightly elevated in some of our experiments (Fig. 7A).

Mesenchymal fibroblast growth factor 10 (Fgf10) signalingregulates proliferation in the PS mesenchyme and epithelium.

Interestingly, Fgf10 activates sonic hedgehog (Shh) expression in theepithelium that in turn signals back into the mesenchyme (Rice et al.,2004). Expression of Shh in the PS epithelium, and of Fgf10, theintracellular Fgf inhibitor gene sprouty homolog 2 (Drosophila)(Spry2), the Fgf target ets variant gene 4 (Etv4 or Pea3) and of theShh target gene patched homolog 1 (Ptch1) in the PS mesenchyme wasunaffected in the mutant arguing against an involvement of Fgf- andShh-signalling pathways in the etiology of CP in Tbx2-deficient mice(Supplemental Figs. 2 and 3).

It was recently shown that loss of Eph receptor B1 (Ephb1) alone orcompound loss of Ephb2 and Ephb3 signaling also affects palategrowth and elevation (Orioli et al., 1996; Risley et al., 2009).Expression of these genes as well as other members of the familiesof Eph receptor A and B genes, and the ephrin A and ephrin B ligandgenes was unchanged in the mutant PS mesenchyme (SupplementalFigs. 4 and 5). Finally, expression ofwingless-related MMTV integrationsite 5A gene (Wnt5a) that was recently implicated in cellularproliferation and migration in the developing palate (He et al.,2008) was similarly unaffected (Supplemental Fig. 6).

PS are regionalized into an anterior and a posterior compartment.Short stature homeobox 2 (Shox2) is expressed in the anterior PSmesenchyme at E12.5 and E13.5 (Hilliard et al., 2005; Yu et al., 2005).In contrast, Tbx22 expression is found in the midposterior mesen-chyme at E12.5 and posteriorly at E13.5 (Braybrook et al., 2002; Bushet al., 2002; Herr et al., 2003). Paired box gene 9 (Pax9) is expressedalong the anterior-posterior extent of the PS mesenchyme with an

Fig. 6. Changes of apoptosis in Tbx2cre/cre PS. TUNEL assay on frontal sections through the head detects increased apoptosis in the entire PS mesenchyme of Tbx2-deficient embryos atE 11.5 (A, left panel) (anterior: wt: 0.18±0.18, mut: 3.68±0.79, p=0.00005; posterior: wt: 1.26±0.38, mut: 1.47±0.51, p=0.69) and in the anterior part at E12.5 (B, left panel)compared to wildtype and heterozygous controls (wt: 0.21±0.13, mut: 1.55±0.08, p=0.001). Apoptosis was unchanged in the posterior portion at E12.5 (A, right panel; B, rightpanel) (wt: 0.17±0.26), mut: 0.53±0.34, p=0.27) between controls and Tbx2-deficient embryos.

151S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

enhanced posterior domain (Peters et al., 1998). Expression of allthree markers was unchanged in the mutant arguing that anterior–posterior PS regionalization is not controlled by Tbx2 (SupplementalFig. 7).

Odd-skipped related 1 and 2 (Osr1 and Osr2) encode transcriptionfactors that are expressed and required for proliferation in the PSmesenchyme (Lan et al., 2001, 2004). Both genes showed unalteredexpression in the Tbx2-deficient palate at E12.5-E14.0 (SupplementalFig. 8).

We additionally tested expression of genes directly regulating cellcycle progression and proliferation since Tbx2 has been implicated incell cycle control in other contexts (Bilican and Goding, 2006; Jerome-Majewska et al., 2005). Expression of CyclinD2 (CycD2, Ccnd2), Cy-clinD3 (CycD3, Ccnd3), of cyclin-dependent kinase inhibitor 2A (Cdkn2aor p16Ink), 2C (Cdkn2c or p18), 2D (Cdkn2d or p19Arf) and of cyclin-dependent kinase inhibitor 1A (Cdkn1a or p21Waf), 1B (Cdkn1b orp27Kip1), 1C (Cdkn1c or p57Kip2) was not detectably altered by lossof Tbx2 in the PS mesenchyme in this assay (Supplemental Fig. 9).Expression of CyclinD1 (CycD1, Ccnd1), however, appeared up-regulated (Fig. 7B).

Analysis of chondrogenesis by in situ hybridization of the chon-drogenic regulator SRY-box containing gene 9 (Sox9) and its targetgene collagen type II, alpha 1 (Col2a1) and histochemical analysisof ECM composition by Alcian blue staining revealed no distur-bance of chondrogenesis in the Tbx2-deficient PS (SupplementalFigs. 10 and 11).

In order to confirm and expand this molecular phenotype analysis,we additionally employed a semiquantitative RT-PCR analysis onmRNA of microdissected whole PS from E12.5 and E14.0 embryos(Figs. 7C, D). In the later case, we used independent pools fromanterior and posterior PS regions. We successfully established PCRsfor most of the genes that have been implicated or are likely to beinvolved in growth control of the PS mesenchyme including genes ofthe Bmp (Bmp2, Bmp4, Msx1), and Fgf pathways (Fgf10, Spry2, Pea3),

Osr genes (Osr1, Osr2), and cell cycle control genes (CycD1, CycD2,CycD3, p16Ink, p19Arf, p21Waf, p27Kip1, p27Kip2). We omitted genesthe spatially restricted expression of which cannot be properlyreflected in this assay (Pax9, Tbx22, Shox2, Shh, Ptch1).

RT-PCR analysis reliably identified significant up-regulation ofBmp4 (1.7 fold) and CyclinD1 (3.2 fold) expression in Tbx2-deficientPS at E12.5 (Fig. 7C) confirming the results of our in situ hybridizationassays (Figs. 7A, B). All other genes tested were unchanged at bothtime points (Figs. 7C, D) again in agreement with our in situhybridization data (Supplemental Figs. 1, 2, 8, and 9). Thus, ouranalyses excludes a number of signaling pathways including Fgf, Shh,Wnt and Ephrins as molecular mediators of the defects in secondarypalate development in Tbx2-deficient mice.

Since CyclinD1 has been characterized as a factor positivelyinfluencing cellular proliferation, its elevated expression can explainthe observed proliferation increase in the E12.5 anterior PS mesen-chyme of Tbx2-deficient embryos. Bmp-signaling affects a largenumber of cellular processes in diverse embryonic context. Notably,elevated levels of Bmp have been shown to induce apoptosis in anumber of tissues but may also affect cellular proliferation (Kiyonoand Shibuya, 2006; Koide et al., 2009; Lagna et al., 2006; Mukho-padhyay et al., 2006). Thus, dysregulation of cell proliferation andapoptosis in Tbx2-deficient PS mesenchyme possibly caused byaltered levels of CyclinD1 and Bmp4 expression in the early outgrowthphasemay translate to alteredmorphology andmorphogenesis of thistissue, finally resulting in a CP phenotype.

Discussion

CP is the most frequent human birth defect caused by genetic andenvironmental perturbations of the development of the secondarypalate. Recovery of more than 60 gene loci in the mouse the mutationof which causes this phenotype has allowed the assignment of specificmolecular and cellular pathways that control palate development

Fig. 7. Molecular changes in palate development in Tbx2-deficient embryos. (A, B) RNA in situ hybridization analysis for Bmp4 and CyclinD1 expression in PS at E12.5. Images showexpression on sections through the anterior, middle and posterior region of PS with dorsal oriented upwards. Bmp4 and CyclinD1 expression appears increased in the anterior region ofPS. (C, D) Quantitative RT-PCR analysis detects molecular changes in Tbx2-deficient PS at E12.5 (C) but not at E14.5 (D). QRT-PCR analysis of marker genes was done on mRNA fromE12.5 and E14.0 PS as indicated in the figure. Expression levels are relative to the wildtype/heterozygous control pool (100%). Expression of Bmp4 and CyclinD1 is significantly up-regulated with a mean±SD for Bmp4 of 1.798±0.2802, p=0.0154 and 3.210±0.7005, p=0.0069 for CyclinD1. All other tested genes remained unchanged in the Tbx2cre/cre embryo.

152 S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

(reviewed in Gritli-Linde, 2007). Here, we have shown that Tbx2 andTbx3, two closely related T-box transcription factor genes, arespecifically required for PS growth during secondary palate develop-ment. Increased apoptosis and altered proliferation may be cellular

mediators of this phenotype, while increased expression of CyclinD1and Bmp4 may be a molecular cause. Our identification adds to thecomplexity of the T-box transcription factor network controllingmammalian palate development.

153S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

Genetic control of PS growth

Palatogenesis is a multistep process that comprises an initialbilateral expansion of the mandibulary processes followed by acomplex growth program of the PS before a fusion process takesplace. Developmental analysis of our Tbx2-deficient embryosshowed that PS vertical outgrowth and elevation occurred butwere severely delayed while the fusion potential of the shelves wasunaffected. Disturbed proliferation and apoptosis in the Tbx2-deficient palate at E11.5 to 12.5 indicates a requirement for Tbx2in tissue growth by maintaining the cell number during the initialoutgrowth phase. Even slight changes of this cellular parameter arelikely to accumulate and translate into a morphogenetic halt asevidenced by the large number of CP phenotypes associated withchanges of cellular survival and proliferation (Chou et al., 2004;Gritli-Linde, 2007). Palate development is tightly coupled to thegrowth and morphogenesis of the whole craniofacial complex anddisturbances in the development of the later may perturb PSelevation and result in CP (Chai and Maxson, 2006). Tbx2 and Tbx3are strongly expressed in all facial primordia, compatible with awider requirement for these genes in craniofacial development. Yet,our histological analysis has not revealed any gross defects outsidethe secondary palate pointing to primary role for Tbx2 and Tbx3 inpalatogenesis.

PS growth and morphogenesis are controlled by a complexinterplay of several epithelial and mesenchymal signaling pathwaysand their downstream transcriptional mediators (Gritli-Linde,2007). Tbx2 and Tbx3 are specifically expressed in the mesenchymeof the developing secondary palate pointing to regulation byepithelial and/or mesenchymal signals and a function in themesenchyme itself. Tbx2 and Tbx3 are targets of Bmp-signaling in anumber of embryonic contexts including the limb and the atrioven-tricular canal (Behesti et al., 2006; Ma et al., 2005; Yamada et al.,2000). Notably, Lee et al. (2007) have recently shown that Bmp4induces Tbx3 expression and that ectopic Tbx3 in turn down-regulates Bmp4 in a palate culture system. Our molecular charac-terization of Tbx2-deficient PS revealed a significant up-regulation ofBmp4 in this tissue supporting the notion that Tbx3 and Tbx2 areinvolved in a negative feedback loop with Bmp4. Zhang et al. (2002)recently showed that Bmp4 and its target gene Msx1 engage in apositive feedback loop arguing for a complex network to tightlycontrol expression of this potent signaling molecule. Since we didnot detect changes of Msx1 expression in our assays, Tbx2/Tbx3 andMsx1 may represent independent and differently sensitive targets ofBmp-signaling in the PS mesenchyme. Unchanged levels of P-Smad1,5,8 do not exclude that Bmp-signaling is augmented sincealternative pathways (including p38 MAPK) are known to operatedownstream of the ligand activated receptors (Xu et al., 2008).Increased proliferation in the anterior PS mesenchyme of E12.5Tbx2cre/cre embryos may be caused by increased Bmp4 levels sinceBmp4 is known to be a powerful mitogen in this tissue (Hilliard etal., 2005; Zhang et al., 2002). However, elevated levels of Bmp4 havealso been associated with increased apoptosis, which we detected inearly PS outgrowth in Tbx2-deficient embryos (Kiyono and Shibuya,2006; Koide et al., 2009; Lagna et al., 2006; Mukhopadhyay et al.,2006). A mitogenic effect of Bmp4 may be mediated or furtherenhanced by increased expression of CyclinD1, a regulatory subunitof CDK4/6 kinases that controls G1-S progression in the cell cycle.Increased expression of CyclinD1 and increased proliferation inTbx2-deficient PS is unexpected since Tbx2 (and Tbx3) havefrequently been associated with the repression of cell cycle inhibitorgenes including p19Arf and p21Waf, thus, cell proliferation andsenescence bypass (Jacobs et al., 2000; Prince et al., 2004). Lee et al.(2007) have shown that ectopic expression of Tbx3 leads toincreased cellular proliferation in cultures of E13.5 and E14.5 PS.The finding that gain of Tbx3 in vitro and loss of Tbx2 in vivo both

lead to increased proliferation is surprising and cannot be explainedat this moment.

It remains unclear how exactly altered proliferation and apoptosisin early PS development translates in reduced PS size and retardedoutgrowth. At this point, we cannot exclude that the delay in PSelevation in Tbx2-deficient mice may also be caused by changes ofECM composition, cytoskeletal architecture or cell movements.However, histological analysis of the ECM, of Wnt5a expression (Heet al., 2008) and Ephrin signaling (Risley et al., 2009) provided noindication of an involvement of these factors and parameters in CPetiology in Tbx2-deficient embryos.

It is noteworthy that the phenotypic penetrance of CP in Tbx2/Tbx3 compound heterozygous mice was strongly dependent on thegenetic background, the defect being much more frequent on FvB/Nthan on NMRI background. We assume that relatively little changes ofTbx2 and Tbx3 activity levels result in changes in cellular parameters.However, since PS growth and morphogenesis is exquisitely sensitiveto proliferation changes, even minor changes of cell cycle regulatorymolecules may accumulate over time and translate in a delay ofoutgrowth and eventually cleft palate.

T-box genes and palate development

T-box genes regulate tissue and organ development in a complexmanner that may include hierarchal activities, but also independent,opposing and/or redundant regional functions of individual familymembers (Hoogaars et al., 2007a; King et al., 2006). Previous studieshave identified three T-box genes in human (and mouse) palatedevelopment. Tbx1 seems to act early in patterning and outgrowthof the pharyngeal pouches preceding secondary palate development.Tbx22 is expressed in the posterior PS arguing for a role in region-alization or compartmentalization of the palate in anterior andposterior halves (Arnold et al., 2006; Braybrook et al., 2002; Bushet al., 2002; Herr et al., 2003). Expression of Tbx22 depends onthe transcription factor Mn1 that is preferentially required forposterior palate growth (Liu et al., 2008). Recent analysis of Tbx22-deficient mice revealed submucous cleft palate (SMCP) and ankylo-glossia. Only a small minority of the animals showed overt cleftsindicating a less important role in palatogenesis in mice comparedto humans (Pauws et al., 2009). Tbx2 and Tbx3, as shown in thisstudy, show no regional restricted expression in the palatal mesen-chyme but may synergize with Tbx22 in the control of cellularprograms important for the elevation process including proliferationand apoptosis.

Interestingly, ubiquitous overexpression of Tbx10 in the spon-taneous Dancer mutant or in transgenic mice causes CP (Bushet al., 2004) arguing that not only loss but also ectopic expressionof T-box genes can result in the same phenotypic consequences. Itis unclear whether Tbx10 acts as a transcriptional activator orrepressor but it is possible that ectopic Tbx10 competes withendogenous Tbx function in the palate including Tbx1, Tbx2/Tbx3and Tbx22. A paradigm for this has been established in the heartwhere the transcriptional repressor Tbx2 acts by competing withthe transcriptional activation activity of Tbx5 and Tbx20, explain-ing the dramatic cardiac arrest phenotype in in vivo misexpres-sion experiments of Tbx2 (Christoffels et al., 2004; Habets et al.,2002). The effect of ectopic Tbx10 in cleft palate induction isstrongly dose dependent, being much more frequent in homo-zygous than in heterozygous conditions (Bush et al., 2004). Ouranalysis has shown that the function of Tbx2 and Tbx3 is alsostrongly dose- and background-dependent arguing that even slightinterference (e.g., by Tbx10) of endogenous T-box transcriptionfactor function may lead to perturbed palate development. In sum-mary, it seems clear that the precise temporal and spatial control ofT-box activity levels is mandatory for normal palatogenesis in miceand man.

154 S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

Acknowledgment

Work in the laboratory of A.K. is supported by grants from theGerman Research Foundation (DFG).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ydbio.2009.09.020.

References

Aanhaanen, W.T., Brons, J.F., Dominguez, J.N., Rana, M.S., Norden, J., Airik, R., Wakker, V.,de Gier-de Vries, C., Brown, N.A., Kispert, A., Moorman, A.F., Christoffels, V.M., 2009.The Tbx2+ primary myocardium of the atrioventricular canal forms theatrioventricular node and the base of the left ventricle. Circ Res.

Arnold, J.S., Werling, U., Braunstein, E.M., Liao, J., Nowotschin, S., Edelmann, W., Hebert,J.M., Morrow, B.E., 2006. Inactivation of Tbx1 in the pharyngeal endoderm results in22q11DS malformations. Development 133, 977–987.

Bakker, M.L., Boukens, B.J., Mommersteeg, M.T., Brons, J.F., Wakker, V., Moorman, A.F.,Christoffels, V.M., 2008. Transcription factor Tbx3 is required for the specification ofthe atrioventricular conduction system. Circ. Res. 102, 1340–1349.

Behesti, H., Holt, J.K., Sowden, J.C., 2006. The level of BMP4 signaling is critical for theregulation of distinct T-box gene expression domains and growth along the dorso-ventral axis of the optic cup. BMC Dev. Biol. 6, 62.

Bilican, B., Goding, C.R., 2006. Cell cycle regulation of the T-box transcription factortbx2. Exp. Cell Res. 312, 2358–2366.

Boogerd, K.J., Wong, L.Y., Christoffels, V.M., Klarenbeek, M., Ruijter, J.M., Moorman, A.F.,Barnett, P., 2008. Msx1 and Msx2 are functional interacting partners of T-boxfactors in the regulation of Connexin43. Cardiovasc. Res. 78, 485–493.

Braybrook, C., Doudney, K., Marcano, A.C., Arnason, A., Bjornsson, A., Patton, M.A.,Goodfellow, P.J., Moore, G.E., Stanier, P., 2001. The T-box transcription factor geneTBX22 is mutated in X-linked cleft palate and ankyloglossia. Nat. Genet. 29,179–183.

Braybrook, C., Lisgo, S., Doudney, K., Henderson, D., Marcano, A.C., Strachan, T., Patton,M.A., Villard, L., Moore, G.E., Stanier, P., Lindsay, S., 2002. Craniofacial expression ofhuman and murine TBX22 correlates with the cleft palate and ankyloglossiaphenotype observed in CPX patients. Hum. Mol. Genet. 11, 2793–2804.

Bush, J.O., Lan, Y., Maltby, K.M., Jiang, R., 2002. Isolation and developmental expressionanalysis of Tbx22, themouse homolog of the human X-linked cleft palate gene. Dev.Dyn. 225, 322–326.

Bush, J.O., Lan, Y., Jiang, R., 2004. The cleft lip and palate defects in Dancer mutant miceresult from gain of function of the Tbx10 gene. Proc. Natl. Acad. Sci. U. S. A. 101,7022–7027.

Bussen, M., Petry, M., Schuster-Gossler, K., Leitges, M., Gossler, A., Kispert, A., 2004. TheT-box transcription factor Tbx18 maintains the separation of anterior and posteriorsomite compartments. Genes Dev. 18, 1209–1221.

Chai, Y., Maxson Jr, R.E., 2006. Recent advances in craniofacial morphogenesis. Dev. Dyn.235, 2353–2375.

Chapman, D.L., Garvey, N., Hancock, S., Alexiou, M., Agulnik, S.I., Gibson-Brown, J.J.,Cebra-Thomas, J., Bollag, R.J., Silver, L.M., Papaioannou, V.E., 1996. Expression of theT-box family genes, Tbx1-Tbx5, during early mouse development. Dev. Dyn. 206,379–390.

Chou, M.J., Kosazuma, T., Takigawa, T., Yamada, S., Takahara, S., Shiota, K., 2004. Palatalshelf movement during palatogenesis: a fate map of the fetal mouse palate culturedin vitro. Anat. Embryol. (Berl). 208, 19–25.

Christoffels, V.M., Hoogaars, W.M., Tessari, A., Clout, D.E., Moorman, A.F., Campione, M.,2004. T-box transcription factor Tbx2 represses differentiation and formation of thecardiac chambers. Dev. Dyn. 229, 763–770.

Davenport, T.G., Jerome-Majewska, L.A., Papaioannou, V.E., 2003. Mammary gland, limband yolk sac defects in mice lacking Tbx3, the gene mutated in human ulnarmammary syndrome. Development 130, 2263–2273.

Ferguson, M.W., 1988. Palate development. Development 103 (Suppl), 41–60.Gritli-Linde, A., 2007. Molecular control of secondary palate development. Dev. Biol.

301, 309–326.Habets, P.E., Moorman, A.F., Clout, D.E., van Roon, M.A., Lingbeek, M., van Lohuizen, M.,

Campione, M., Christoffels, V.M., 2002. Cooperative action of Tbx2 and Nkx2.5inhibits ANF expression in the atrioventricular canal: implications for cardiacchamber formation. Genes Dev. 16, 1234–1246.

Harrelson, Z., Kelly, R.G., Goldin, S.N., Gibson-Brown, J.J., Bollag, R.J., Silver, L.M.,Papaioannou, V.E., 2004. Tbx2 is essential for patterning the atrioventricularcanal and for morphogenesis of the outflow tract during heart development.Development 131, 5041–5052.

He, F., Xiong, W., Yu, X., Espinoza-Lewis, R., Liu, C., Gu, S., Nishita, M., Suzuki, K., Yamada,G., Minami, Y., Chen, Y., 2008. Wnt5a regulates directional cell migration and cellproliferation via Ror2-mediated noncanonical pathway in mammalian palatedevelopment. Development 135, 3871–3879.

Herr, A., Meunier, D., Muller, I., Rump, A., Fundele, R., Ropers, H.H., Nuber, U.A., 2003.Expression of mouse Tbx22 supports its role in palatogenesis and glossogenesis.Dev. Dyn. 226, 579–586.

Hilliard, S.A., Yu, L., Gu, S., Zhang, Z., Chen, Y.P., 2005. Regional regulation of palatal growthand patterning along the anterior–posterior axis in mice. J. Anat. 207, 655–667.

Hoogaars, W.M., Barnett, P., Moorman, A.F., Christoffels, V.M., 2007a. T-box factorsdetermine cardiac design. Cell. Mol. Life Sci. 64, 646–660.

Hoogaars, W.M., Engel, A., Brons, J.F., Verkerk, A.O., de Lange, F.J., Wong, L.Y., Bakker,M.L., Clout, D.E., Wakker, V., Barnett, P., Ravesloot, J.H., Moorman, A.F., Verheijck,E.E., Christoffels, V.M., 2007b. Tbx3 controls the sinoatrial node gene program andimposes pacemaker function on the atria. Genes Dev. 21, 1098–1112.

Jacobs, J.J., Keblusek, P., Robanus-Maandag, E., Kristel, P., Lingbeek, M., Nederlof, P.M., van Welsem, T., van de Vijver, M.J., Koh, E.Y., Daley, G.Q., van Lohuizen, M.,2000. Senescence bypass screen identifies TBX2, which represses Cdkn2a (p19(ARF)) and is amplified in a subset of human breast cancers. Nat. Genet. 26,291–299.

Jerome, L.A., Papaioannou, V.E., 2001. DiGeorge syndrome phenotype in mice mutantfor the T-box gene, Tbx1. Nat. Genet. 27, 286–291.

Jerome-Majewska, L.A., Jenkins, G.P., Ernstoff, E., Zindy, F., Sherr, C.J., Papaioannou, V.E.,2005. Tbx3, the ulnar-mammary syndrome gene, and Tbx2 interact in mammarygland development through a p19Arf/p53-independent pathway. Dev. Dyn. 234,922–933.

King, M., Arnold, J.S., Shanske, A., Morrow, B.E., 2006. T-genes and limb buddevelopment. Am. J. Med. Genet. A. 140, 1407–1413.

Kiyono, M., Shibuya, M., 2006. Inhibitory Smad transcription factors protect arterialendothelial cells from apoptosis induced by BMP4. Oncogene 25, 7131–7137.

Koide, Y., Kiyota, T., Tonganunt, M., Pinkaew, D., Liu, Z., Kato, Y., Hutadilok-Towatana, N.,Phongdara, A., Fujise, K., 2009. Embryonic lethality of fortilin-null mutant mice byBMP-pathway overactivation. Biochim. Biophys. Acta 1790, 326–338.

Lagna, G., Nguyen, P.H., Ni, W., Hata, A., 2006. BMP-dependent activation of caspase-9and caspase-8 mediates apoptosis in pulmonary artery smooth muscle cells. Am. J.Physiol., Lung Cell. Mol. Physiol. 291, L1059–1067.

Lan, Y., Kingsley, P.D., Cho, E.S., Jiang, R., 2001. Osr2, a new mouse gene related toDrosophila odd-skipped, exhibits dynamic expression patterns during craniofacial,limb, and kidney development. Mech. Dev. 107, 175–179.

Lan, Y., Ovitt, C.E., Cho, E.S., Maltby, K.M., Wang, Q., Jiang, R., 2004. Odd-skipped related2 (Osr2) encodes a key intrinsic regulator of secondary palate growth andmorphogenesis. Development 131, 3207–3216.

Lee, J.M., Kim, J.Y., Cho, K.W., Lee, M.J., Cho, S.W., Zhang, Y., Byun, S.K., Yi, C.K., Jung, H.S.,2007. Modulation of cell proliferation during palatogenesis by the interplaybetween Tbx3 and Bmp4. Cell Tissue Res. 327, 285–292.

Liu, W., Lan, Y., Pauws, E., Meester-Smoor, M.A., Stanier, P., Zwarthoff, E.C., Jiang, R.,2008. The Mn1 transcription factor acts upstream of Tbx22 and preferentiallyregulates posterior palate growth in mice. Development 135, 3959–3968.

Ludtke, T.H., Christoffels, V.M., Petry, M., Kispert, A., 2009. Tbx3 promotes liver budexpansion during mouse development by suppression of cholangiocyte differen-tiation. Hepatology 49, 969–978.

Ma, L., Lu, M.F., Schwartz, R.J., Martin, J.F., 2005. Bmp2 is essential for cardiac cushionepithelial-mesenchymal transition and myocardial patterning. Development 132,5601–5611.

Merscher, S., Funke, B., Epstein, J.A., Heyer, J., Puech, A., Lu, M.M., Xavier, R.J., Demay,M.B., Russell, R.G., Factor, S., Tokooya, K., Jore, B.S., Lopez, M., Pandita, R.K., Lia,M., Carrion, D., Xu, H., Schorle, H., Kobler, J.B., Scambler, P., Wynshaw-Boris, A.,Skoultchi, A.I., Morrow, B.E., Kucherlapati, R., 2001. TBX1 is responsible forcardiovascular defects in velo-cardio-facial/DiGeorge syndrome. Cell 104,619–629.

Moorman, A.F., Houweling, A.C., de Boer, P.A., Christoffels, V.M., 2001. Sensitivenonradioactive detection of mRNA in tissue sections: novel application of thewhole-mount in situ hybridization protocol. J. Histochem. Cytochem. 49, 1–8.

Mukhopadhyay, P., Singh, S., Greene, R.M., Pisano, M.M., 2006. Molecular fingerprintingof BMP2- and BMP4-treated embryonic maxillary mesenchymal cells. Orthod.Craniofac. Res. 9, 93–110.

Murray, J.C., 2002. Gene/environment causes of cleft lip and/or palate. Clin. Genet. 61,248–256.

Naiche, L.A., Harrelson, Z., Kelly, R.G., Papaioannou, V.E., 2005. T-box genes in vertebratedevelopment. Annu. Rev. Genet. 39, 219–239.

Naiche, L.A., Papaioannou, V.E., 2007. Cre activity causes widespread apoptosis andlethal anemia during embryonic development. Genesis 45, 768–775.

Orioli, D., Henkemeyer, M., Lemke, G., Klein, R., Pawson, T., 1996. Sek4 and Nukreceptors cooperate in guidance of commissural axons and in palate formation.EMBO J. 15, 6035–6049.

Pauws, E., Hoshino, A., Bentley, L., Prajapati, S., Keller, C., Hammond, P., Martinez-Barbera, J.P., Moore, G.E., Stanier, P., 2009. Tbx22null mice have a submucous cleftpalate due to reduced palatal bone formation and also display ankyloglossia andchoanal atresia phenotypes. Hum. Mol. Genet. (Advance Access published on July31, 2009). doi:10.1093/hmg/ddp368.

Peters, H., Neubuser, A., Kratochwil, K., Balling, R., 1998. Pax9-deficient mice lackpharyngeal pouch derivatives and teeth and exhibit craniofacial and limbabnormalities. Genes Dev. 12, 2735–2747.

Prince, S., Carreira, S., Vance, K.W., Abrahams, A., Goding, C.R., 2004. Tbx2 directlyrepresses the expression of the p21(WAF1) cyclin-dependent kinase inhibitor.Cancer Res. 64, 1669–1674.

Rice, D.P., 2005. Craniofacial anomalies: from development to molecular pathogenesis.Curr. Mol. Med. 5, 699–722.

Rice, R., Spencer-Dene, B., Connor, E.C., Gritli-Linde, A., McMahon, A.P., Dickson, C.,Thesleff, I., Rice, D.P., 2004. Disruption of Fgf10/Fgfr2b-coordinated epithelial-mesenchymal interactions causes cleft palate. J. Clin. Invest. 113, 1692–1700.

Risley, M., Garrod, D., Henkemeyer, M., McLean, W., 2009. EphB2 and EphB3 forwardsignalling are required for palate development. Mech. Dev. 126, 230–239.

Shprintzen, R.J., Higgins, A.M., Antshel, K., Fremont, W., Roizen, N., Kates, W., 2005.Velo-cardio-facial syndrome. Curr. Opin. Pediatr. 17, 725–730.

155S. Zirzow et al. / Developmental Biology 336 (2009) 145–155

Suzuki, A., Sekiya, S., Buscher, D., Izpisua Belmonte, J.C., Taniguchi, H., 2008. Tbx3controls the fate of hepatic progenitor cells in liver development by suppressingp19ARF expression. Development 135, 1589–1595.

Taya, Y., O'Kane, S., Ferguson, M.W., 1999. Pathogenesis of cleft palate in TGF-beta3knockout mice. Development 126, 3869–3879.

Wilkinson, D.G., 1992. Whole mount in situ hybridization of vertebrate embryos. In:Wilkinson, D.G. (Ed.), In Situ Hybridization: A Practical Approach. OxfordUniversity Press, Oxford, pp. 75–84.

Xu, X., Han, J., Ito, Y., Bringas Jr., P., Deng, C., Chai, Y., 2008. Ectodermal Smad4 and p38MAPK are functionally redundant in mediating TGF-beta/BMP signaling duringtooth and palate development. Dev. Cell 15, 322–329.

Yamada, M., Revelli, J.P., Eichele, G., Barron, M., Schwartz, R.J., 2000. Expression of chick

Tbx-2, Tbx-3, and Tbx-5 genes during early heart development: evidence for BMP2induction of Tbx2. Dev. Biol. 228, 95–105.

Yu, L., Gu, S., Alappat, S., Song, Y., Yan, M., Zhang, X., Zhang, G., Jiang, Y., Zhang, Z., Zhang,Y., Chen, Y., 2005. Shox2-deficient mice exhibit a rare type of incomplete clefting ofthe secondary palate. Development 132, 4397–4406.

Zhang, Y., Zhang, Z., Zhao, X., Yu, X., Hu, Y., Geronimo, B., Fromm, S.H., Chen, Y.P., 2000. Anew function of BMP4: dual role for BMP4 in regulation of Sonic hedgehogexpression in the mouse tooth germ. Development 127, 1431–1443.

Zhang, Z., Song, Y., Zhao, X., Zhang, X., Fermin, C., Chen, Y., 2002. Rescue of cleft palate inMsx1-deficient mice by transgenic Bmp4 reveals a network of BMP and Shhsignaling in the regulation of mammalian palatogenesis. Development 129,4135–4146.