PATTERNS & PHENOTYPES

Tracking Gene Expression During ZebrafishOsteoblast DifferentiationNan Li,1† Katharina Felber,1† Phil Elks,1,2† Peter Croucher,2 and Henry H. Roehl1,3*

The transcription factors RUNX2 and OSX have been shown to act sequentially to direct mammalianosteoblast differentiation. RUNX2 is required during the early stages of commitment and acts in part toactivate Osx transcription. OSX and RUNX2 then act to direct transcription of bone matrix proteins. Here,we investigate the expression of these genes and others during zebrafish osteoblastogenesis. Using whole-mount in situ hybridization, we find that, during the formation of a given bone, the zebrafish homologuesof mouse Runx2 (runx2a and runx2b) are typically expressed before the onset of osx. osx expression isusually followed by up-regulation of the bone matrix proteins, col1a2 and osteonectin. These results suggestthat the mammalian pathway is conserved during development of the head and shoulder skeleton ofzebrafish. We also analyze the expression of three atypical bone markers (tcf7, cvl2, and col10a1) in an effortto place them within this canonical hierarchy. Developmental Dynamics 238:459–466, 2009.© 2009 Wiley-Liss, Inc.

Key words: dermal bone; osteoprogenetor; osteoblast; osteoblastogenesis; neural crest; bone; cranial;runx2/cbfa1/osf1/pebp2�A; osterix; collagenIa2; crossveinless2; collagen10a1; tcf7; osteonectin; mesenchymal stem cell

Accepted 24 November 2008

INTRODUCTION

Osteoblasts are specialized cells thatare responsible for the deposition ofbone (Karsenty and Wagner, 2002).Mammalian osteoblasts originatefrom mesenchymal stem cells that arecompetent to give rise to several celltypes, including myocytes, chondro-cytes, and adipocytes (Prockop, 1997;Pittenger et al., 1999). Molecular andgenetic analysis performed in micehas identified two transcription fac-tors to be key regulators of osteoblastdifferentiation. In mice lacking Runx2

(runt-related transcription factor2/core binding factor a1 /osteoblast-specific transcription factor-2/polyomaenhancer-binding protein 2 alpha A)or Osx (osterix), osteoblasts do not dif-ferentiate but the initial formation ofother mesenchymal derivatives islargely normal (Komori et al., 1997;Otto et al., 1997; Nakashima et al.,2002). Although Osx expression is lostin mice lacking Runx2, Runx2 expres-sion is unaffected in Osx knockoutmice. This has led to the model thatRunx2 acts upstream of Osx. In sup-

port of this, functional RUNX2 bind-ing sites have been identified in theOsx promoter (Nishio et al., 2006).Furthermore, both RUNX2 and OSXhave been shown in turn to regulateexpression of bone matrix genes suchas CollagenIA1 (Ducy et al., 1997;Koga et al., 2005). As constitutive ex-pression of Runx2 leads to osteopeniaand blocks osteoblast maturation, it islikely that Runx2 expression must bedown-regulated for the completion ofdifferentiation (Kanatani et al., 2006).

Genetic and molecular analysis of

Additional Supporting Information may be found in the online version of this article.1MRC Centre for Developmental and Biomedical Genetics, University of Sheffield, Sheffield, United Kingdom2Academic Unit of Bone Biology, Section of Musculoskeletal Science, University of Sheffield Medical School, Beech Hill Road, Sheffield,United Kingdom3Department of Biomedical Science, University of Sheffield, Sheffield, United KingdomGrant sponsors: Wellcome Trust; Grant number: 072346/Z/03/Z; Grant sponsor: Cancer Research UK; Grant number: C11413/A4072; Grantsponsor: European Union; Grant number: Cells into Organs NoE; Grant sponsor: Medical Research Council; Grant number: G070091.†Drs. Li, Felber, and Elks contributed equally to this work.*Correspondence to: Henry H. Roehl, MRC Centre for Developmental and Biomedical Genetics, Firth Court, Western Bank,University of Sheffield, Sheffield S10 2TN, UK. E-mail: h.roehl@sheffield.ac.uk

DOI 10.1002/dvdy.21838Published online 19 January 2009 in Wiley InterScience (www.interscience.wiley.com).

DEVELOPMENTAL DYNAMICS 238:459–466, 2009

© 2009 Wiley-Liss, Inc.

bone development has predominantlyfocused on the long bones of tetrapodlimbs. These bones are both mesoder-mally derived and develop from atransient cartilage template which issubsequently ossified in a processcalled endochondral ossification. Incontrast, most of the bones in the headare derived from cranial neural crestand form both by the endochondralroute as well as in the absence of acartilage template (called dermal ossi-fication). These differences in embry-onic origins are reflected in differencesin the expression of osteoblast mark-ers and activities of signaling mole-cules during cranial osteoblastogen-esis (Eames and Helms, 2004;Abzhanov et al., 2007). Furthermore,these differences may also underliethe distinct effects that pharmaceuti-cals such as bisphosphonates have oncranial versus trunk skeleton (Sarinet al., 2008).

Although zebrafish has not beenused extensively for the analysis ofbone development and disease, sev-eral studies have highlighted its po-tential. As with tetrapods, the skullbones of zebrafish develop directly(called dermal bones) as well as usinga cartilage template (called cartilagebones; Cubbage and Mabee, 1999;Verreijdt et al., 2006). In addition, ze-brafish cranial neural crest plays asimilar role as in tetrapods (Schillingand Kimmel, 1994), suggesting thatmost of the skull is neural crest de-rived. Also, members of the hedgehog,runx, and dlx family of genes whichare known to regulate osteoblast dif-ferentiation in mammals are ex-pressed during zebrafish bone forma-tion (Flores et al., 2004, 2006; Pinto etal., 2005; van der Meulen et al., 2005;Avaron et al., 2006). Finally, a for-ward genetic screen for skeletal dys-plasia in adult zebrafish identified amutation in collagenIA1 that accu-rately models the human disease os-teogenesis imperfecta (Fisher et al.,2003).

Here, we track the expression ofbone markers in zebrafish larvae atsix time-points to establish an over-view early osteoblast differentiation.We find that, in general, bone devel-opment can be broken down into threeoverlapping stages based upon geneexpression: early differentiation withthe expression of runx2a and/or

runx2b, intermediate differentiationwith the expression of osx and a ma-ture state of differentiation distin-guished by the expression of the bonematrix proteins colIa2 (collagen type Ialpha 2; Fisher et al., 2003) and osn(osteonectin/sparc). These findingssuggest that the transcriptional hier-archy found in mammals is conservedduring zebrafish osteoblastogenesis.We also analyze the expression ofthree genes not typically associatedwith osteoblastogenesis (col10a1, col-lagen type X alpha 1; cvl2, crossvein-less 2; tcf7, transcription factor 7;Veien et al., 2005; Avaron et al., 2006;Rentzsch et al., 2006) and find thattheir expression suggests key newroles during osteoblast differentia-tion. Finally, we focus on one bone, theopercle to give a more detailed de-scription of the progress of differenti-ation within an individual skeletalprimordia.

RESULTS AND DISCUSSION

Early Stage Expression:runx2a and runx2b

As seen by others (Flores et al., 2004,2006; van der Meulen et al., 2005), wefound that runx2a and runx2b are ex-pressed in both cartilage and bone pri-mordia (Figs. 1A–J, 2A–L). Althoughtheir early expression is diffuse in thecranial region, after 36 hours postfer-tilization (hpf), their expression pro-gressively resolves to developing skel-etal elements starting with thecleithrum condensation at 36 hpf (Fig.2G). All of the bone anlage analyzedexpress both runx2a and runx2b (Ta-ble 1) and because they are probablyfunctionally interchangeable, they arelikely to be redundant in regionswhere they are coexpressed. Their ex-pression usually presages the expres-sion of osx and the bone matrix genes,and their expression around cranialbones is strongly diminished by 120hpf (Figs. 1E,J, 2F,L).

Intermediate StageExpression: osx

Unlike the runx2 genes, osx is not ex-pressed broadly before 36 hpf and itsexpression at all times is specific tothe developing bony skeleton (Figs.1K–O, 2M–R). In addition, while the

runx2 genes tend to each label a re-gion of a bone primordium, osx labelsthe whole structure. This is perhapsdue to the lack of osx gene duplicationas is seen for the runx2 genes. Thisspecific and comprehensive labeling ofstructures by the in situ probe, makesosx a good marker for the intermedi-ate stage of osteoblast differentiation.Expression of osx is either coincidentwith that of runx2a and runx2b, orfollows shortly afterward (Table 1).Like the runx2 genes, expression ofosx diminishes by 120 hpf, suggestingthat it plays less of a role in osteo-blasts that are actively secreting bonematrix (Figs. 1O, 2R).

Mature Stage Expression:colIa2 and osn

The bone matrix genes, colIa2 and osnshow remarkably similar expressionpatterns (Figs. 1P–Y, 2S–DD). In ad-dition to being expressed in regionsthat undergo ossification, both are ex-pressed in epidermis, the fin fold andconnective tissue. osn is also stronglyexpressed in the developing ear. Thisrelatively nonspecific expressionmakes their interpretation difficult,particularly with deep bones such asthe parasphenoid. In general, their ex-pression in bone forming tissues is ei-ther coincident with that of osx or fol-lows shortly afterward (Table 1).Usually within 24 hr after the onset ofbone matrix gene expression, ossifica-tion of the bone is detectable by Aliza-rin Red staining. Expression of colIa2and osn perdures around skeletal ele-ments after osx, runx2a and runx2bare not detectable. This suggests thatosx, runx2a and runx2b are not highlyexpressed in mature osteoblasts.

Atypical Gene Expression:col10a1, tcf7, and cvl2

Although expression of col10a1 is typ-ically described as being associatedwith hypertrophic chondrocytes in tet-rapods, it has also been described asbeing expressed during formation ofcartilage and dermal bones in ze-brafish (Avaron et al., 2006). We havefound that its expression overlapswith that of osx as well as the bonematrix genes (Figs. 3A–E, 4A–F). Thissuggests that it plays a role duringboth the intermediate and mature

460 LI ET AL.

stages of osteoblast differentiation. Itis expressed around all of the develop-ing bones analyzed here as well asaround the retroarticular and quad-rate in some fish at 120 hpf (data notshown). Like osx, it is expressedthroughout an entire bone primor-dium, but unlike osx, expression ismaintained after bone has formed,presumably in mature osteoblasts. Inaddition to its expression in early os-sification centers, it also shows peri-chondral expression on the medialside of the Meckel’s cartilage, the ven-tral surface of the most distal part ofthe ceratohyal and on the medial sur-face of the hyosymplectic adjacent to

the interhyal at 120 hpf (data notshown). This probe shows a clear andcomprehensive expression pattern,making it a very useful marker forintermediate and mature stages of dif-ferentiation.

tcf7 is a known mediator of theWNT (wingless-type MMTV integra-tion site family) signaling pathway inzebrafish (Veien et al., 2005; Bonneret al., 2008; Nagayoshi et al., 2008).Although tcf7 is expressed in manytissues during development, we havefound that its expression is enhancedaround developing dermal bones(Figs. 3F–J, 4G–L). In general, thetiming of tcf7 expression matches that

of osx suggesting a role for WNT sig-naling during intermediate stages ofosteoblast differentiation. However,unlike osx and col10a1, its expressiondoes not always label an entire boneprimordium, and it is very weakly ex-pressed in the parasphenoid and ento-pterygoid. This may be explained byredundancy with other members ofthe lef/tcf family of genes.

Zebrafish cvl2 is a known mediatorof BMP (bone morphogenetic protein)signaling that has been shown to beexpressed in many cranial tissues in-cluding around forming cranial bones(Rentzsch et al., 2006). We have foundthat like col10a1, it is expressed in all of

Fig. 1. A–Y: Osteoblast gene expression: side views. Five stages of development are shown highlighting the development of the two related cranialbones, the third brachiostegal ray and the opercle, as well as the cleithrum, which is part of the shoulder. In addition, the parasphenoid and fifthceratobranchial are labeled in F and H. The double arrow in L marks the distal edge of the second arch for reference. All views are of the left side ofthe head, focusing on the plane of the opercle, with the eye and yolk visible to either side (labeled in A). bs, brachiostegal ray; cb5, ceratobranchialarch 5 (including teeth); ch, ceratohyal; cl, cleithrum; de, dentary; en, entopterygoid; hm, hyomandibular; mx, maxilla; op, opercle; ps, parasphenoid.A high resolution image of this figure is available online. Scale bar � 100 �M in Y.

GENE EXPRESSION IN ZEBRAFISH OSTEOBLAST 461

Fig. 2. A–Z: Osteoblast gene expression: dorsal and ventral views. Six stages of development are shown in dorsal views at 36 and 48 hourspostfertilization (hpf; A,B,G,H,M,N,S,T,Y,Z), and in ventral views at 63, 72, 96, and 120 hpf (C–F,I–L,O–R,U–X,AA–DD). A high resolution image of thisfigure is available online. See Figure 1 for abbreviations. Scale bar � 200 �M in DD.

Fig. 3. Atypical osteoblast gene expression: side views. Five stages shown as described in Figure 1. See Figure 1 for abbreviations and scale bar.A high resolution image of this figure is available online.

462 LI ET AL.

the bone primordia analyzed here,throughout a given bone primordium,and during intermediate and maturestages of osteoblast differentiation(Figs. 3K–O, 4M–R). This synexpres-sion suggests that these genes may betranscriptionally regulated by the samemechanism. Intriguingly, col10a1 hasbeen shown to be a transcriptional tar-get of BMP signaling during chondro-cyte maturation (Drissi et al., 2003) andcvl2 has been shown to be a target ofBMP signaling in zebrafish embryos(Rentzsch et al., 2006). Thus, our anal-ysis raises the possibility that BMP sig-naling also induces col10a1 expressionduring osteoblast differentiation.

Expression Profile of aDeveloping Dermal Bone:The Opercle

At 120 hpf, the opercle is fan-shapedand made up of a strut proximallythat articulates with the hyosymplec-tic cartilage by means of a ball andsocket joint, and a thin plate distallythat serves to give support to the gillcover as well as provide muscle at-tachment sites for four cranial mus-cles (Fig. 5A,B). The opercle first be-gins to ossify in the strut region before96 hpf and proceeds to ossify moredistally to give rise to the fan by 120hpf (data not shown; Verreijdt et al.,2006). This ossification sequence ispresaged by the expression of osteo-blast markers in the second arch mes-enchyme. Both runx2 genes show dif-fuse expression in the second arch at36 hpf, which subsequently resolvesinto the opercle primordium as well assecond arch precartilage condensa-tions (Figs. 1A,F, 2A,B,G,H; Flores etal., 2004, 2006; van der Meulen et al.,2005). By 63 hpf, all of the genes an-alyzed show expression in the opercleprimordium, with tcf7, runx2a andrunx2b expressed in a large domain(Figs. 1B,G, 3G), and the other mark-ers expressed in a smaller, strut-shaped strip of cells (Figs. 1L,Q,V,3B,L). One interpretation of this oper-cular expression pattern is that thereis a large patch of early osteoblast pre-cursors marked by tcf7 and runx2genes surrounding a smaller subset ofcells that have begun to express inter-mediate and mature stage genes (Fig.5C). By 72 hpf, tcf7 and runx2 geneshave begun to show diminished ex-

pression, especially in the centre ofthe primordium (Figs. 1C,H, 3H),while osx expression has expanded toencompass the entire anlage (Fig.1M). The other markers are still ex-pressed in a smaller, strut-shapedcluster cells (Figs. 1R,W, 3C,M). Thedown-regulation of early markers mayindicate that cells in the centre of thecondensation are maturing and begin-ning to form the strut of the opercle,

while cells on the periphery are notmature (Fig. 5D). By 96 hpf, the pri-mordium has taken on a fan shape,and osx, tcf7, and runx2 genes are alldiminished in the middle region (Figs.1D,I,N, 3I). The other genes are ex-pressed throughout the anlage (Figs.1S,X, 3D,N). This indicates that thereis now a fan-shaped cluster of matureosteoblasts actively secreting bonematrix, surrounded by osteoblast pre-

TABLE 1. Summary of In Situ Data for Both Dermal (D)and Cartilage (C) Bones

GENE EXPRESSION IN ZEBRAFISH OSTEOBLAST 463

cursors in early and/or intermediatestages of differentiation (Fig. 5E).Taken together, these data suggestthat differentiation begins in the cen-tre of the opercle condensation andmoves in a wave to the periphery. Thisanalysis tentatively places tcf7 as anearly differentiation marker whilecol10a1 and cvl2 are intermediate andlate stage markers.

CONCLUSIONS

We have found that runx2 genes areexpressed before the onset of osx in 7/10bones analyzed and osx is expressedearlier than bone matrix genes in 5/10

bones analyzed (see Table 1 for over-view). For the bone primordia wherethe initiation of expression matches (forexample osx and bone matrix genes inthe maxilla at 72 hpf), analysis at moretime points may tease apart the onsettimes. The only bone that does not showa correlation is the parasphenoid wherethe onset of matrix genes precedes thatof osx. However, on the whole, this se-quential expression suggests that thetranscriptional hierarchy required formammalian osteoblast differentiationis conserved in zebrafish dermal andcartilage bone formation. In addition,our analysis of tcf7, col10a1 and cvl2

expression provides evidence that BMPand WNT signaling could play a roleduring differentiation of osteoblasts asis seen in mammals (Wan and Cao,2005; Hartmann, 2007). Our descrip-tion of the dynamic gene expressionduring the formation of the opercle sug-gests a model for how differentiationtakes place in an individual dermalbone anlage (Fig. 5). However, thesedata also raise important questions,such as how recruitment of osteoblastprecursors is balanced with the forma-tion of other cell types so as not to de-plete the embryonic mesenchymal stemcell population; are markers for differ-ent stages of differentiation expressedconcurrently in individual cells; how aremorphogenesis and differentiation or-chestrated during bone development?The ease with which transgenic linescan be made to manipulate signalingpathways and fluorescently label indi-vidual cell types, coupled with the abil-ity to image cell behavior in live em-bryos, makes zebrafish uniquely suitedto answer these and other questions re-lating to vertebrate bone formation.

EXPERIMENTALPROCEDURES

RNA in situ hybridization was done asdescribed by Thisse and Thisse (2008).Several individual fish were analyzedand in some cases dissection was nec-

Fig. 4. A–R: Atypical osteoblast gene expression: dorsal and ventral views. Six stages of development are shown highlighting the development thecleithrum in dorsal views at 36 and 48 hpf (A,B,G,H,M,N), and the development of other bones in ventral views at 63, 72, 96, and 120 hpf (C–F,I–L,O–R).See Figure 1 for abbreviations and Figure 2 for scale bar sizing. A high resolution image of this figure is available online.

Fig. 5. Model for osteoblast differentiation during formation of the opercle. A: Alizarin Redstaining of the fan-shaped opercle (op) at 144 hours postfertilization (hpf) showing its connectionto the hyosymplectic (hs) cartilage proximally. B: Camera lucida drawing of A showing the joint,strut, and plate of the opercle. C–E: See text for explanation.

464 LI ET AL.

essary to confirm expression. Thebones that were analyzed are the fol-lowing: brachiostegal ray 3, cerato-branchial arch 5 (including teeth),ceratohyal, cleithrum, dentary, ento-pterygoid, hyomandibular, maxilla,opercle, parasphenoid (Cubbage andMabee, 1999; Verreijdt et al., 2006).For Table 1, hatching indicates thatexpression was weak or not present insome fish.

Please note that the expression timepoints marked in Table 1 as well asshown in the Figures indicates theminimum period of expression. Allow-ing the in situ color reaction to con-tinue for longer periods than used forthis study may reveal weak stainingoutside of the indicated period.

Details for the individual probes arelisted below: runx2a and runx2b:cDNA clones in the pGEM-T vector (agift from T. van der Meulen) were sub-cloned into the pBSK vector using Sa-cII and PstI. These were then used togenerate a polymerase chain reaction(PCR) template using M13 forwardand reverse primers to and T7 poly-merase for probe synthesis. tcf7: acDNA fragment encoding 916 basepa-irs (bp) was amplified using PCR (tcat-cactggtcagcgagac and accagtccgtctgtt-ggttc) and cloned into the PCRII-TOPO vector. This was then processedas for the runx2 genes. osx: a cDNAfragment encoding 920 bp was ampli-fied using PCR (ttagacatgacgcatcct-tacg and ggttaaatctccagcagtccac) andcloned into the PCRII-TOPO vector.This was then processed as for therunx2 genes. col10a1: a cDNA clone inthe PCRII-TOPO vector (a gift fromM. Akimenko) was used to generate atemplate using PCR (ggatccattaaccct-cactaaagggaaccgagctcggatccactagta-acg and M13for), which was then tran-scribed using T3 polymerase. cvl: acDNA clone in the PCS2� vector (agift from M. Hammerschmidt) wasused to generate a template usingPCR (cttgatttaggtgacactatagaa andM13rev), which was then transcribedusing T7 polymerase. osn: IMAGEclone 6899780 (1,688 bp) was sub-cloned into the pBSK vector usingEcoRI and HindIII. This was then pro-cessed as for the runx2 genes. col1a2:IMAGE clone 6963594 (5,182 bp) wasused to generate a template (M13for,

cagtccagaacatcacctacca) which wasthen transcribed using T7 polymer-ase.

ACKNOWLEDGMENTSH.H.R. was funded by grants fromCancer Research UK, the WellcomeTrust, and the European Commission.

REFERENCES

Abzhanov A, Rodda SJ, McMahon AP,Tabin CJ. 2007. Regulation of skeleto-genic differentiation in cranial dermalbone. Development 134:3133–3144.

Avaron F, Hoffman L, Guay D, AkimenkoMA. 2006. Characterization of two newzebrafish members of the hedgehog fam-ily: atypical expression of a zebrafish In-dian hedgehog gene in skeletal elementsof both endochondral and dermal origins.Dev Dyn 235:478–489.

Bonner J, Gribble SL, Veien ES, NikolausOB, Weidinger G, Dorsky RI. 2008. Pro-liferation and patterning are mediatedindependently in the dorsal spinal corddownstream of canonical Wnt signaling.Dev Biol 313:398–407.

Cubbage CC, Mabee PM. 1999. Develop-ment of the cranium and paired fins inthe zebrafish Danio rerio (Ostariophysi,Cyprinidae). J Morphol 229:121–160.

Drissi MH, Li X, Sheu TJ, Zuscik MJ,Schwarz EM, Puzas JE, Rosier RN,O’Keefe RJ. 2003. Runx2/Cbfa1 stimula-tion by retinoic acid is potentiated byBMP2 signaling through interactionwith Smad1 on the collagen X promoterin chondrocytes. J Cell Biochem90:1287–1298.

Ducy P, Zhang R, Geoffroy V, Ridall AL,Karsenty G. 1997. Osf2/Cbfa1: a tran-scriptional activator of osteoblast differ-entiation. Cell 89:747–754.

Eames BF, Helms JA. 2004. Conserved mo-lecular program regulating cranial andappendicular skeletogenesis. Dev Dyn231:4–13.

Fisher S, Jagadeeswaran P, Halpern ME.2003. Radiographic analysis of zebrafishskeletal defects. Dev Biol 264:64–76.

Flores MV, Tsang VW, Hu W, Kalev-Zylin-ska M, Postlethwait J, Crosier P, CrosierK, Fisher S. 2004. Duplicate zebrafishrunx2 orthologues are expressed in de-veloping skeletal elements. Gene ExprPatterns 4:573–581.

Flores MV, Lam EY, Crosier P, Crosier K.2006. A hierarchy of Runx transcriptionfactors modulate the onset of chondro-genesis in craniofacial endochondralbones in zebrafish. Dev Dyn 235:3166–3176.

Hartmann C. 2007. Skeletal develop-ment—Wnts are in control. Mol Cells 24:177–184.

Kanatani N, Fujita T, Fukuyama R, Liu W,Yoshida CA, Moriishi T, Yamana K,Miyazaki T, Toyosawa S, Komori T.2006. Cbf beta regulates Runx2 function

isoform-dependently in postnatal bonedevelopment. Dev Biol 296:48–61.

Karsenty G, Wagner EF. 2002. Reaching agenetic and molecular understanding ofskeletal development. Dev Cell 2:389–406.

Koga T, Matsui Y, Asagiri M, Kodama T,de Crombrugghe B, Nakashima K,Takayanagi H. 2005. NFAT and Osterixcooperatively regulate bone formation.Nat Med 11:880–885.

Komori T, Yagi H, Nomura S, YamaguchiA, Sasaki K, Deguchi K, Shimizu Y,Bronson RT, Gao YH, Inada M, Sato M,Okamoto R, Kitamura Y, Yoshiki S,Kishimoto T. 1997. Targeted disrup-tion of Cbfa1 results in a complete lackof bone formation owing to matura-tional arrest of osteoblasts. Cell 89:755–764.

Nagayoshi S, Hayashi E, Abe G, Osato N,Asakawa K, Urasaki A, Horikawa K,Ikeo K, Takeda H, Kawakami K. 2008.Insertional mutagenesis by the Tol2transposon-mediated enhancer trap ap-proach generated mutations in two de-velopmental genes: tcf7 and synembryn-like. Development 135:159–169.

Nakashima K, Zhou X, Kunkel G, Zhang Z,Deng JM, Behringer RR, de Crombrug-ghe B. 2002. The novel zinc finger-con-taining transcription factor osterix is re-quired for osteoblast differentiation andbone formation. Cell 108:17–29.

Nishio Y, Dong Y, Paris M, O’Keefe RJ,Schwarz EM, Drissi H. 2006. Runx2-me-diated regulation of the zinc finger Os-terix/Sp7 gene. Gene 372:62–70.

Otto F, Thornell AP, Crompton T, DenzelA, Gilmour KC, Rosewell IR, Stamp GW,Beddington RS, Mundlos S, Olsen BR,Selby PB, Owen MJ. 1997. Cbfa1, a can-didate gene for cleidocranial dysplasiasyndrome, is essential for osteoblast dif-ferentiation and bone development. Cell89:765–771.

Pinto JP, Conceicao NM, Viegas CS, Le-ite RB, Hurst LD, Kelsh RN, CancelaML. 2005. Identification of a newpebp2alphaA2 isoform from zebrafishrunx2 capable of inducing osteocalcingene expression in vitro. J Bone MinerRes 20:1440 –1453.

Pittenger MF, Mackay AM, Beck SC,Jaiswal RK, Douglas R, Mosca JD, Moor-man MA, Simonetti DW, Craig S, Mar-shak DR. 1999. Multilineage potential ofadult human mesenchymal stem cells.Science 284:143–147.

Prockop DJ. 1997. Marrow stromal cells asstem cells for nonhematopoietic tissues.Science 276:71–74.

Rentzsch F, Zhang J, Kramer C, Sebald W,Hammerschmidt M. 2006. Crossveinless2 is an essential positive feedback regu-lator of Bmp signaling during zebrafishgastrulation. Development 133:801–811.

Sarin J, DeRossi SS, Akintoye SO. 2008.Updates on bisphosphonates and poten-tial pathobiology of bisphosphonate-in-duced jaw osteonecrosis. Oral Dis14:277–285.

Schilling TF, Kimmel CB. 1994. Segment andcell type lineage restrictions during pharyn-

GENE EXPRESSION IN ZEBRAFISH OSTEOBLAST 465

geal arch development in the zebrafish em-bryo. Development 120:483–494.

Thisse C, Thisse B. 2008. High-resolutionin situ hybridization to whole-mount ze-brafish embryos. Nat Protoc 3:59–69.

van der Meulen T, Kranenbarg S, SchipperH, Samallo J, van Leeuwen JL, FranssenH. 2005. Identification and characterisa-tion of two runx2 homologues in zebrafish

with different expression patterns. Bio-chim Biophys Acta 1729:105–117.

Veien ES, Grierson MJ, Saund RS, DorskyRI. 2005. Expression pattern of zebrafishtcf7 suggests unexplored domains ofWnt/beta-catenin activity. Dev Dyn 233:233–239.

Verreijdt L, Debiais-Thibaud M, Borday-Birraux V, Van der Heyden C, Sire JY,

Huysseune A. 2006. Expression of thedlx gene family during formation of thecranial bones in the zebrafish (Danio re-rio): differential involvement in the vis-ceral skeleton and braincase. Dev Dyn235:1371–1389.

Wan M, Cao X. 2005. BMP signaling inskeletal development. Biochem BiophysRes Commun 328:651–657.

466 LI ET AL.