Developmental Biology 347 (2010) 392–403

Contents lists available at ScienceDirect

Developmental Biology

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

Genomes & Developmental Control

Otx2 and Otx1 protect diencephalon and mesencephalon from caudalization intometencephalon during early brain regionalization

Yusuke Sakurai a,1, Daisuke Kurokawa a,c,1, Hiroshi Kiyonari b, Eriko Kajikawa a,Yoko Suda a, Shinichi Aizawa a,b,⁎a Laboratory for Vertebrate Body Plan, Center for Developmental Biology, RIKEN Kobe, 2-2-3 Minatojima Minamimachi, Chuo-ku, Kobe 650-0047, Japanb Laboratory for Animal Resources and Genetic Engineering, Center for Developmental Biology, RIKEN Kobe, 2-2-3 Minatojima Minamimachi, Chuo-ku, Kobe 650-0047, Japanc Misaki Marine Biological Station, Graduate School of Science, The University of Tokyo, 1024 Koajiro, Misaki, Miura, Kanagawa 238-0225, Japan

⁎ Corresponding author. Laboratory for Vertebrate Bmental Biology, RIKEN Kobe, 2-2-3 Minatojima Minam0047, Japan. Fax: +81 78 306 3148.

E-mail address: saizawa@cdb.riken.go.jp (S. Aizawa)1 The first two authors contributed equally to this wo

0012-1606/$ – see front matter © 2010 Elsevier Inc. Adoi:10.1016/j.ydbio.2010.08.028

a b s t r a c t

a r t i c l e i n f o

Article history:Received for publication 18 June 2010Revised 24 August 2010Accepted 25 August 2010Available online 16 September 2010

Keywords:Otx2Otx1EnhancerForebrainMidbrainIsthmus

Otx2 is expressed in each step and site of head development. To dissect each Otx2 function we have identifieda series of Otx2 enhancers. TheOtx2 expression in the anterior neuroectoderm is regulated by the AN enhancerand the subsequent expression in forebrain andmidbrain later than E8.5 by FM1 and FM2 enhancers; the Otx1expression takes place at E8.0. In telencephalon later than E9.5 Otx1 continues to be expressed in the entirepallium, while the Otx2 expression is confined to the most medial pallium. To determine the Otx functions inforebrain and midbrain development we have generated mouse mutants that lack both FM1 and FM2enhancers (DKO: Otx2ΔFM1ΔFM2/ΔFM1ΔFM2) and examined the TKO (Otx1−/−Otx2ΔFM1ΔFM2/ΔFM1ΔFM2) pheno-type. Themutants develop normally until E8.0, but subsequently by E9.5 the diencephalon, including thalamiceminence and prethalamus, and the mesencephalon are caudalized into metencephalon consisting of isthmusand rhombomere 1; the caudalization does not extend to rhombomere 2 and more caudal rhombomeres. Inrostral forebrain, neopallium, ganglionic eminences and hypothalamus in front of prethalamus develop; wepropose that they become insensitive to the caudalization with the switch from the Otx2 expression under theAN enhancer to that under FM1 and FM2 enhancers. In contrast, the medial pallium requires Otx1 and Otx2 forits development later than E9.5, and the Otx2 expression in diencepalon and mesencephalon later than E9.5 isalso directed by an enhancer other than FM1 and FM2 enhancers.

ody Plan, Center for Develop-imachi, Chuou-ku, Kobe 650-

.rk.

ll rights reserved.

© 2010 Elsevier Inc. All rights reserved.

Introduction

Otx2 is expressed and plays essential roles in each step and site ofhead development; first in epiblast and visceral endoderm, subse-quently in anterior mesendoderm, then in anterior neuroectodermand in cephalic neural crest cells at the migratory phase (Simeoneet al., 1992, 1993; Kimura et al., 1997). Otx2 homozygous mutantsexhibit the defects of the earliest Otx2 functions in visceral endoderm,and heterozygous mutants the defects of cephalic neural crest cells(Matsuo et al., 1995; Rhinn et al., 1998; Acampora et al., 1995, 2009;Kimura et al., 2000; Kimura-Yoshida et al., 2005, 2007). Otx2expression in the anterior neuroectoderm also changes dynamically.It is first found in the entire anterior neuroectoderm induced by

anterior visceral endoderm and anterior mesendoderm; its caudallimit is obscured in front of preotic sulcus. From E8.5 Otx2 expressionis gradually lost in the rostral forebrain, and at E10.5 the expression islost in telencephalon except for the most medial pallium. At this stagethe caudal limit of the Otx2 expression is sharply delineated at themidbrain/hindbrain boundary. Otx1 expression takes place at 1–3somite stage initially withinOtx2-positive domain, but continues to beexpressed in dorsal telencephalon. It is Otx1, but not Otx2, that isexpressed in developing cortex later than E12.5; later, Otx1 isexpressed in layer 5 (Frantz et al., 1994). These changes in theexpression pattern suggest that during brain development Otx2 andOtx1 play a series of roles differing by stage. This has, in fact, beendemonstrated with mouse mutants.

Anterior neuroectoderm initially develops in mouse or human Otx1knock-in mutants into Otx2 locus (Otx2 mOtx1/mOtx1 or Otx2 hOtx1/hOtx1)but is subsequently lost by E9.0 (Acampora et al., 1998; Suda et al.,1999). Thesemutantsmay primarily represent theOtx2 functions in theinitial anterior neuroectoderm. The Otx2 hOtx1/hOtx1 mutants, however,would include defects by the loss of Otx2 functions in anteriormesendoderm; the hOTX1 protein was concluded not to be produced

393Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

in epiblast-derived tissues (Acampora et al., 1998). The Otx2 mOtx1/mOtx1

phenotype was brought about by the reduction, but not the completeloss, of OTX2 expression in the anterior neuroectoderm (Suda et al.,1999); the development of the anterior neuroectoderm requires a OTX2dose higher than that required for anterior visceral endoderm andanterior mesendoderm development. Hypomorphic mutants by theinsertion of a λ phage DNA fragment into a 3'UTR or a frt-neo into anintron of Otx2 gene (Otx2λ/− or Otx2frt-neo/−), respectively, would alsorepresent the same condition (Boyl et al., 2001; Tian et al., 2002). Incontrast, Otx1+/−Otx2+/−, Otx1−/−Otx2+/− and Otx2λ/λ mutants maysuggest Otx1 and Otx2 functions in subsequent forebrain/midbraindevelopment (Acampora et al., 1997; Suda et al., 1997; Boyl et al., 2001).Prethalamus and telencephalon develop in the Otx1/Otx2 doublemutants, but it is not certain whether or not these regions requireOtx1/Otx2; there remains some Otx2 and/or Otx1 expression in theOtx1/Otx2 double mutants. In zebrafish it was proposed that Otx is notrequired for the prethalamus development (Scholpp et al., 2007). At thesame time, the doublemutant phenotypemight also include the defectsresulting from the partial loss of earlier Otx2 functions in the anteriorneuroectoderm. Of note is that the phenotypes in these mutants withthe partial loss of Otx2 and/or Otx1were highly variable.

To dissect Otx2 functions at each stage and site of headdevelopment more clearly and to identify their gene cascades, wehave undertaken Otx2 enhancer analysis (Kimura et al., 1997;Kurokawa et al., 2004a,b, 2006, 2010; Kimura-Yoshida et al., 2007).Indeed, Otx2 enhancers for each expression could be dissected; thosefor the expression in visceral endoderm, anterior mesendoderm andcephalic neural crest cells exist near the transcription start site, andthose for the expression in epiblast and anterior neuroectoderm at90–95 kb 5' upstream. The enhancer that directs the expression inanterior neuroectoderm, AN, is active throughout the entire anteriorneuroectoderm, and the caudal limit of its activity is obscured in frontof the preotic sulcus as the endogenous Otx2 expression at presomitestage is. The AN enhancer, however, loses its activity, except formedial pallium, by E9.0 and the subsequent Otx2 expression inforebrain and midbrain is regulated by an enhancer located 75 kb 5'upstream (FM1) and another enhancer located 115 kb 3' downstream(FM2). The caudal limit of the FM1 and FM2 enhancers coincides withthe midbrain/hindbrain boundary, and rostrally they do not haveactivities in neopallium, ganglionic eminences or hypothalamus. AN,FM1 and FM2 enhancers continue to be active in medial pallium. TheFM2 enhancer is unique to rodents. AN and FM1 enhancers areconserved in tetrapods, coelacanth and skate Otx2 orthologues, whileteleost or bichir Otx2 orthologues do not conserve the AN enhancer,and the FM1 enhancer has the AN activity (Kurokawa et al., 2006; ourunpublished data).

The AN enhancer mutant (Otx2ΔAN/−) demonstrated that Otx2 isessential to protect the caudalization of the entire anterior neuroec-toderm to metencephalon (Kurokawa et al., 2004a). In this study, tospecify the later Otx1 and Otx2 functions upon brain regionalization,mutant mice were generated in which both FM1 and FM2 enhancersand Otx1 are lost (TKO: Otx1−/−Otx2ΔFM1ΔFM2/ΔFM1ΔFM2) and in whichthe Otx2 expression in forebrain/midbrain later than E8.5 isspecifically lost together with Otx1. The TKO mutants indeed developnormally until E8.0. Subsequently by E9.5, however, mesencephalonand diencephalon caudalized into metencephalon (isthmus andrhombomere 1). Anteriorly thalamic eminence and prethalamus arelost, but neopallium, ganglionic eminences and hypothalamusdevelop in the TKO mutants; we propose that the latter regionsbecome resistant to caudalization or OTX-independent by earlysomite stage when the AN enhancer becomes inactive and the FM1and FM2 enhancers become active. In contrast, Otx2 and Otx1 arerequired for the medial pallium development later than E9.5. The FM1and FM2 double enhancer mutants (DKO: Otx2ΔFM1ΔFM2/ΔFM1ΔFM2)also demonstrated that there is another enhancer for the Otx2expression in diencephalon and mesencephalon later than E9.5.

Materials and methods

Mutant mice

Transgenic lines that express βGal under AN, FM1 and FM2enhancers (Acc. No.: CDB0005T, CDB0027T and CDB0003T, respec-tively; http://www.cdb.riken.jp/arg/mutant%20mice%20list.html)were established previously (Kurokawa et al., 2004a,b). The ANenhancer is the 15 kb XhoI fragment at –92 kb to –77 kb 5' upstream,FM1 enhancer the 10 kb XhoI–ClaI fragment at the −77 kb to −68 kb5' upstream and FM2 enhancer the 10 kb MluI–NotI fragment at+106 kb to +116 kb 3' downstream. Each enhancer is conjugatedwith the 1.8 kb Otx2 promoter proximal to the translation start siteand lacZ gene. Otx1 mutant mice (Acc. No. CDB0017K) wereestablished previously (Suda et al., 1997). A mutant mouse line(Acc. No. CDB0050K) of which Otx2 FM1 enhancer (1314 bp ApaI–HindIII fragment, Fig. 1A) was replaced with a Neo cassette wasestablished previously; the cassette consists of Pgk1 promoter,neomycin resistance gene and SV40 polyadenylation signal and isflanked by loxP sequences (Kurokawa et al., 2004b). ES cells wereestablished from a homozygously FM1mutant blastocyst as described(Yagi et al., 1993a). To target the FM2 locus, the targeting vector wasconstructed by replacing the 995 bp BsmI–BamHI fragment encodingthe FM2 enhancer with a Puro cassette; the cassette consists of Pgk1promoter, puromycin resistance gene and SV40 polyadenylationsignal and is flanked by frt sequences. The construction of thetargeting vector, isolation of recombinant ES clones and production ofthe Otx2ΔFM1ΔFM2 enhancer mutant mice (Acc. No. CDB0064K) wereperformed as described (Yagi et al., 1993a,b), the details of which willbe provided upon request. Mice were housed and experiments wereperformed under the CDB guidelines for animal and recombinant DNAexperiments.

RNA in situ hybridization and βGal staining

Embryos were dissected in phosphate-buffered saline (PBS) andfixed overnight at 4 °C in 4% paraformaldehyde in PBS. Whole-mountand section in situ hybridization was performed using digoxigeninprobes as described (Wilkinson, 1993; Suda et al., 2001). The probesused are as described for the genes: Dlx1 (Bulfone et al., 1993), Dmbx1(Miyamoto et al., 2002), Emx2 (Yoshida et al., 1997), En1 (Davis andJoyner, 1988), Ephb1 (IMAGE clone AA058194), Ephrin A2 (Flennikenet al., 1996), Fezl (Hirata et al., 2004), Fgf8 (Crossley and Martin,1995), Gbx2 (Bulfone et al., 1993), Hoxb1 (Kimura et al., 2001), Lhx1(Fujii et al., 1994),Msx1 (Hill et al., 1989),Nkx2.1 (Kimura et al., 1996),Otx1 (Matsuo et al., 1995), Otx2 (Matsuo et al., 1995), Pax6 (Waltherand Gruss, 1991), Shh (Echelard et al., 1993), Six3 (Oliver et al., 1995),Tcf4 (Korinek et al., 1998), TTR (Wakasugi et al., 1985), Vgll1 (Murataet al., unpublished data),Wnt1 (McMahon and Bradley, 1990),Wnt3a(Roelink and Nusse, 1991), Wnt7b (Parr et al., 1993) and Wnt8b(IMAGE clone AA170920). βGal expression was determined asdescribed (Kimura et al., 1997, 2000).

Quantitative RT-PCR

RNA isolation and quantitative RT-PCRwere carried out as described(Shibata et al., 2008). For all primer sets tested, correlation (R2) washigher than0.98, and the slopewas−3.1 to−3.6 in eachstandard curve.Primers to detect the expression of each gene were designed in a singleexon encoding 3'UTR: forward (5'-CAATGTCCCAGGCTCATTCA-3') andreverse (5'-TCAGTGCCAACTACCTGTTGGT-3') for Otx2; forward (5'-GGAGACGGACTGCCTTACATC-3') and reverse (5'-TGCGGAGAG-TACCTGTGTACC-3') for Otx1; forward (5'-GTGATGTGAAGTTCCCCA-TAAGG-3') and reverse (5'-CTACTGAACTGCTGGTGGGTCA-3') for Tbp(Svingen et al., 2009).

Fig. 1. DKO (Otx2ΔFM1ΔFM2/ΔFM1ΔFM2) mutants. (A) Schematic representation of the wild type and DKO mutant Otx2 alleles. Open boxes indicate enhancers and filled boxes Otx2 codingexons. Filled triangles indicate loxP sequences and open triangles frt sequences. (B) Double enhancer mutant phenotype at E8.5 (a, d, k, n), E9.5 (b, e, g, i, l, o), E10.5 (c, f, h, j, m, p), E12.5(q, t), E15.5 (r, u) and E18.5 (s, v); (a–c, g, h, k–m, q–s) give the wild type and (d–f, i, j, n–p, t–v) double enhancer mutant embryos. (a-p) give lateral views (anterior is at left) of wholemount embryos stained for theOtx1 (a–f),Dmbx1 (g–j) andGbx2 (k–p) expression; (q, r, t, u) parasagittal sections and (s, v) frontal sections at the telencephalic/diencephalic level stainedwith hematoxylin-eosin (q, t) and cresyl violet (r, s, u, v). In (a–f) dotted lines indicate the planes of the dissections for RT-PCR analysis in Fig. 9a–f, and solid lines the planes of frontalsections for the Otx2 in situ hybridization in Fig. 9g–p. Asterisks in (c, f) indicate the dorsal mesencephalon where the Otx1 expression is enhanced in the DKO mutants.

394 Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

BrdU labeling

In order to investigate cell proliferation, single pulse of BrdU(Roche) (50 mg/kg body weight) was delivered to pregnant mice30 min before sacrifice. Embryos were fixed in 4% paraformaldehydeand embedded in paraplast. Sections were subsequently prepared(thickness of 8 μm). Immunohistochemical staining for BrdU wasperformed using a monoclonal antibody ( Purified Mouse Anti-BrdU;BD pharmingen ), a M.O.M kit (Vector) and a NovaRed substrate(Vector) according to the manufacturer's instruction.

TUNEL assay

Serial paraffin sections (8 μm) derived from embryos fixed in 4%paraformaldehyde were subjected to TUNEL assay according to themanufacturer's protocol (Apop Tag Peroxidase In Situ ApoptosisDetection Kit; CHEMICON international, Inc).

Histology

Mouse embryos were fixed overnight with Bouin's or Carnoy'sfixative solution at room temperature. Specimens were subsequentlydehydrated and embedded in paraplast. Serial sections (10 μm) wereprepared and stained with hematoxylin and eosin or with cresyl violet.

Results

FM1 and FM2 double enhancer mutants

To dissect the Otx2 functions under the FM1 and FM2 enhancers,mutant mice were generated which lack both of these enhancers(Otx2ΔFM1ΔFM2 ; DKO; Fig. 1A). To make this mutation ES cells werefirst established from an Otx2ΔFM1/ΔFM1 mutant blastocyst in whichthe FM1 enhancer was homozygously replaced with a neo cassetteflanked by loxP (Kurokawa et al., 2004b). In the ES cells the FM2enhancer was replaced with a Puro cassette flanked by frt.

395Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

Otx1-positive anterior and Gbx2-positive posterior neuroectodermwas developing normally in homozygously DKOmutants at E8.5whenthe FM1 and FM2 enhancers became active (Fig. 1Ba, d, k, n). At E9.5,however, Otx1-positive rostral brain was greatly reduced (Fig. 1Bb, e),the Dmbx1 expression in caudal forebrain/midbrain was faint(Fig. 1Bg, i) and Gbx2-positive anterior hindbrain was expanded(Fig. 1Bl, o). The telencephalic region was apparently normal, and thereduction was pronounced in the mesencephalic and diencephalicregions; they were not completely lost. Noticeably in E10.5 DKOmutants Otx1-positive rostral brain was largely restored (Fig. 1Bc, f),though Gbx2-positive isthmus was somewhat expanded (Fig. 1Bm, p)and the caudal outline of the Otx1-positive forebrain was slightlydepressed. Dmbx1 expression in caudal forebrain/midbrain was lowand not fully restored (Fig. 1Bh, j). At E12.5 histologically telenceph-alon was normal, while diencephalon and mesencephalon weremoderately reduced (Fig. 1Bq, t). Concomitantly, cerebellum primor-dium and isthmus were anteriorly expanded. In E15.5 and E18.5 DKOmutants, no defects were apparent, and the brain was developingalmost normally (Fig. 1Br, s, u, v). Homozygous DKO mutants werelive born apparently normally in a Mendelian ratio after the crossbetween heterozygotes; they were fertile. These phenotypes wereuniform among DKO mutants.

Even at E9.5 the DKOmutation did not result in the loss of the entireFM1andFM2-positivebrain. This could beexplainedby complementaryfunctions of an Otx2 paralogue, Otx1. Consequently, the phenotype ofthe additional Otx1 mutation (Otx1−/−Otx2ΔFM1ΔFM2/ΔFM1ΔFM2; TKO)was examined. The phenotypes described below were uniform amongTKO mutants. Histologically the TKO mutants indeed appeared to losethe diencephalon and mesencephalon at E12.5 (Fig. 2a and k), and thephenotypewas not restored at E15.5 and E18.5 (Fig. 2b–e, h–j, k–o, r–t).Rostrally at E12.5 and E15.5, neopalliumwasmostly normal, butmedialpallium was residual (Fig. 2a–d, k–n). Residual medial pallium wascaudally deformed (Fig. 2l); choroid plexus developed (Fig. 2l and n).Hypothalamus developed almost normally, but ganglionic eminenceswere rather hyperplastic. At E18.5 cortical layer formation wasapparently abnormal (Fig. 2h, j, r, t). Hyperplasia of ganglioniceminences was pronounced; this may be explained by a defect in thetangential migration of interneurons into the cortex (Pleasure et al.,2000; Shinozaki et al., 2002). Hippocampus, dentate gyrus or fimbriawas not histologically apparent, and medial pallium was residual.Caudally, a cerebellum-like structure was anteriorly expanded wherenormallymesencephalon anddiencephalondevelopandwas connectedto the residual medial pallium through choroid plexus (Fig. 2b, h, l, r). Itis indeed the cerebellum as confirmed by the presence of the Pax6-positive external granule layer that is also uniquely dark-stained byNissle at E18.5 (Fig. 2g, i, q, s). Isthmus and pons also expandedanteriorly, but medulla did not (Fig. 2b, e, h, i, l, o, r, s); the isthmic cellswere abnormally organized with the feature of the overgrowth in theventricular zone cells that faces the fourth ventricle (Fig. 2f and p).

Diencephalic/mesencephalic structures are lost in the TKO mutants

The loss of mesencephalon and diencephalon was characterized indetail by marker analyses at E12.5. Tcf4 is expressed intensively inpretectum and thalamus (Fig. 3Aa), Gbx2 in thalamus (Fig. 3Ba), and

Fig. 2. Histological views of TKO (Otx1−/−Otx2ΔFM1ΔFM2/ΔFM1ΔFM2) mutant phenotype.(a–j) Give wild type and (k–t) TKO mutant embryos. (a, b, h, k, l, r) Show parasagittalsections (anterior is at left) at E12.5 (a, k), E15.5 (b, l) and E18.5 (h, r) stained withhematoxylin–eosin (a, b, k, l) and cresyl violet (h, r); (c–e, m–o) frontal sections stainedwith hematoxylin-eosin, at the level indicated in (b, l) at E15.5; (f, p) enlarged views ofcerebellar/isthmic area boxed in (b, l); (g, q) parasagittal sections showing Pax6expression in external granule layer of cerebellum at E15.5; (i, s) enlarged views ofcerebellar/isthmic area boxed in (h, r) showing external granule layer of cerebellumintensively stained with cresyl violet; (j, t) enlarged views of medial pallium regionboxed in (h, r). Arrowheads in (l, p) indicate cells darkly stained with hematoxylin–eosin in the expanded isthmus, and arrows in (b–d, l, n ) the choroid plexus.

Fig. 3. Marker analyses of diencephalic/mesencephalic defects at E12.5. Genes examinedby RNA in situ hybridization are: (A) Tcf4, (B)Gbx2, (C)Dlx1, (D) Pax6, (E) Lhx1, (F)Wnt8b,(G) Emx2, (H) Ephrin A2, (I)Dmbx1 and (J)Nkx2.1. Embryos in (a) panels arewild type, (b)DKOmutant and (c) TKOmutant embryos; all are parasagittal sections (anterior is at left).Insets in (B) give the Gbx2 expression in ganglionic eminences. Pax6, Lhx1 and Wnt8bexpressions in thalamic eminence (arrows in (D–F)) are lost in TKO mutants, and weconsider the Emx2 expression indicated by an open arrowhead in (Gc) does not representthe expression in thalamic eminence; its origin remains to be determined. Openarrowheads in (Cc) indicate ectopic Dlx1 expression in the expanded isthmic region (cf.Fig. 2p). Filled arrowheads indicate: in (A) the Tcf4 expression in themost anterior tectum;in (B, E, I) the Gbx2, Lhx1 and Dmbx1 expression, respectively, in cerebellum primordium;in (C, J) Dlx1 and Nkx2.1 expression, respectively, in hypothalamus; in (F) the Wnt8bexpression in medial pallium; in (G) the Emx2 expression in tegmentum. Open arrows in(E, J) indicate Lhx1 and Nkx2.1 expression, respectively, in the mammillary region.

396 Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

Dlx1 and Pax6 in prethalamus (Fig. 3Ca. Da). Lhx1 is expressed in zonalimitans, prethalamus, pretectum, tegmentum, isthmus, pons andmedulla (Fig. 3Ea; Fujii et al., 1994). The loss of intensive Tcf4, Gbx2,Dlx1 and Pax6 expressions indicates the lack of pretectum, thalamusand prethalamus in the TKO mutants (Fig. 3Ac, Bc, Cc, Dc).Furthermore,Wnt8b, Emx2, Pax6 and Lhx1 positive thalamic eminencewas lost in the TKO mutants (arrows in Fig. 3Fa, Ga, Da, Ea).Consequently, we assume that Lhx1-positive zona limitans, prethala-mus, pretectum and tegmentum are also lost, and Lhx1-positiveisthmus and pons are placed where normally diencephalic structuresexist (Fig. 3Ec). Emx2 is also expressed in tegmentum (Fig. 3Ga), andthe tegmentumwas also lost in TKOmutants (Fig. 3Gc). In addition topretectum and thalamus, Tcf4 is also expressed in the most anteriortectum which is also lost in the TKO mutants (Fig. 3Aa, c). EphrinA2-positive superior colliculus and Dmbx1-positive inferior colliculus areabsent in the TKO mutants (Fig. 3H and I). Gbx2, Lhx1 and Dmbx1 areexpressed in anterior cerebellum (arrowheads in Fig. 3Ba, Ea, Ia), andtheir expression is found in the anterior region of expandedcerebellum of the TKO mutants (arrowheads in Fig. 3Bc, Ec, Ic).

Ventrally Lhx1 is expressed in the mammillary region, Dlx1 inposterior entopeduncular area and hypothalamus, and Nkx2.1 in themammillary region and hypothalamus (Fig. 3C, E, J). The hypothal-amus develops, and the mammillary region also appears to havedeveloped in E12.5 TKO mutants. However, the details of ventralstructures that are lost or present in the TKO mutants remain forfuture studies. DKO mutants expressed these forebrain and hindbrainmarkers almost normally at E12.5 (Fig. 3A–Jb).

TKO defects in medial pallium

In telencephalon at E12.5 Pax6 and Emx2-positive cortex and Gbx2,Dlx1 and Nkx2.1-positive ganglionic eminences (Fig. 3Da, Ga, Ba, Ca,Ja) developed normally in the TKO mutants (Fig. 3Dc, Gc, Bc, Cc, Jc).However, Emx2-positive medial pallium was greatly reduced

Fig. 4.Marker analyses of medial pallium defects at E12.5. Genes examined are: (A) TTR,(B)Msx1, (C)Wnt3a, (D)Wnt8b and (E) Ephb1. Embryos in (a) panels are wild type, (b)DKO mutant and (c) TKO mutant embryos. All show frontal sections.

397Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

(Fig. 3Ga, c). Transthyretin (TTR)- and Msx1-positive choroid plexus,and TTR-negative andMsx1-positive choroidal roof were also reducedin the TKO mutants (Fig. 4A and B). In the medial pallium a series ofWnt genes is expressed in a nested pattern. Wnt3a marks the corticalhem adjacent to the choroid plexus (Fig. 4Ca; Yoshida et al., 2006);Wnt3a-positive cortical hem was present but reduced in the TKOmutants (Fig. 4Cc). Normally Wnt8b is widely expressed in medialpallium (Fig. 4Da; Fig. 3Fa). In the TKO mutants, Wnt8b-positivemedial pallium was greatly reduced (Fig. 4Dc, Fig. 3Fc). In the medialpallium, Ephb1 expression marks the hippocampus field; it wasresidually present in the TKO mutants (Fig. 4Ea, c). At E12.5 thereduction of medial pallium, being assessed with these markers, wasmoderate in DKO mutants (Fig. 4A–Eb).

Medial pallium development starts with the neural tube closure atthe telencephalic level around E9.0. Otx1 continues to be expressed indorsal telencephalon except for choroid plexus and choroidal roofduring its development (Fig. 5A). Otx2 expression is initially found inthe entire pallial region, but is restricted to the most medial region atE10.5, and thereafter continues to be expressed in cortical hem,choroid plexus and choroidal roof (Fig. 5B). The activities of FM1, FM2

Fig. 5. Onset of medial pallium defects. (A) shows Otx1 mRNA expression, (B) Otx2 mRNA eE10.5 and (b) E12.5 wild type embryos. (F) gives Otx2 expression, (G, H, K, L) Wnt8b expresembryos at E9.5 (G, H) and E10.5 (F, I–L). All show frontal sections at telencephalic level; (G,the1.8 kb Otx2 promoter and lacZ gene, and βGal expression is shown in each typical trans

and AN enhancer cover this Otx2 expression (Fig. 5C–E). The Otx2expression also exists in the most ventral telencephalon (Fig. 5Ba, b);the enhancer that directs this expression has not yet been identified(Kurokawa et al., 2004b).

In TKO mutants the Otx2 expression was greatly reduced butpresent in the most medial pallium (Fig. 5F); this may represent theOtx2 expression under the AN enhancer. The Otx2 expression alsoremains in the most ventral telencephalon of TKO mutants (Fig. 5Band F). At E9.5 Wnt8b-positive medial pallium developed almostnormally in TKO mutants (Fig. 5G and H). However, at E10.5dorsomedial telencephalon poorly invaginated, and Wnt3a-positivecortical hem and Wnt8b-positive medial pallium were reduced in theTKO mutants (Fig. 5I–L). Changes in cell death or cell growth mightaccount for the poor invagination; cell death is prominent in choroidalroof (Shinozaki et al., 2004). No change in TUNEL-positive cells was,however, apparent in the E10.0 TKO roof or medial pallium (data notshown). No difference in BrdU uptake was also apparent betweenwild type and TKO mutant medial pallium (Supplementary Fig. 2).DKOmutants exhibitedmilder defects inmedial pallium developmentat E10.5 (Fig. 7F–Lb).

xpression and (C-E) βGal expression by FM1 (C), FM2 (D) and AN (E) enhancers in (a)sion and (I, J) Wnt3a expression in wild type (a), DKO mutant (b) and TKO mutant (c)I, K) show rostral and (H, J, L) caudal sections. In (C-E) each enhancer is conjugated withgenic line (Kurokawa et al., 2004a,b).

398 Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

Onset of mesencephalic/diencephalic defects

In E9.5 wild type forebrain Emx2 and Fezl expression coverstelencephalon (Fig. 6Aa, Ba; Hirata et al., 2004; Kimura et al., 2005),and Pax6 expression telencephalon and diencephalon (Fig. 6Ca;Kimura et al., 2005). In the TKO mutants Emx2- and Fezl-positivetelencephalon was normally present, while Pax6-positive forebrainwas greatly reduced, indicating the loss of diencephalon (Fig. 6Ac, Bc,Cc). Ventrally Shh-positive diencephalon from which zona limitansdevelops was also lost (arrows in Fig. 6D; Shimamura et al., 1995);however, Shh-positive most rostral ventral forebrain developednormally in the TKO mutants (arrowheads in Fig. 6D; Echelard et al.,1993). Dmbx1 expression covers posterior diencephalon and theentire mesencephalon (Fig. 6Ea; Ohtoshi et al., 2002). Dmbx1-positiveregion is completely lost in the TKO mutants (Fig. 6Ec). Wnt1expression is normally present in the dorsomedial region ofdiencephalon and mesencephalon and the mesencephalic side ofthe midbrain/hindbrain boundary (Fig. 6Fa; Rowitch and McMahon,1995; Bally-Cuif et al., 1995). TheseWnt1-positive domains are lost inthe TKO mutants (Fig. 6Fc). Fgf8 expression is normally confined tothe metencephalic side of the midbrain/hindbrain boundary oristhmus (Fig. 6Ga). At E9.5 the isthmus also expresses Gbx2(Fig. 6Ha). Fgf8- and Gbx2-positive isthmus is greatly expandedanteriorly in the TKO mutants (Fig. 6Gc, Hc), whereas the Fgf8-positive anterior neural ridge has developed normally (arrowheads inFig. 6G).

We had a question about the caudal end of the expanded area inhindbrain. Locations of Vgll1-positive rhombomere 2 and Hoxb1-positive rhombomere 4 in respect to the otic vesicle were normal inthe TKO mutants (Fig. 6I and J). Consistently, the anterior end of Pax6,Wnt1 and Gbx2 expression in the TKOmutant hindbrain was normallyfound at rhombomere 2 (Fig. 6Cc, Ec, Hc). No marker specific torhombomere 1 is available, but apparently Fgf8- and Gbx2- negativerhombomere 1 (Fig. 6Ga, Ha) is expanded in the TKO mutants(Fig. 6Gc, Hc).

In E9.5 DKO mutants, Otx1-positive forebrain/midbrain andDmbx1-positive caudal forebrain/midbrain was reduced (Fig. 1Be, i;Fig. 6Eb), and Wnt1-, Fgf8- and Gbx2-positive isthmus was expanded(Fig. 1Bo; Fig. 6F–Hb). Emx2- and Fezl-positive telencephalon, Pax6-positive rostral forebrain, Vgll1-positive rhombomere 2, and Hoxb1-positive rhombomere 4 were apparently normal (Fig. 6Ab–Cb, Ib, Jb).

At E7.75 when the Otx1 expression is absent, FM1 and FM2activities are inactive and the Otx2 expression is directed by the ANenhancer (Kurokawa et al., 2004a,b), Otx2-positive anterior neuroec-todermwas developing normally in the TKOmutants (Fig. 7A). At E8.5when the AN enhancer was still active, several abnormalities werealready apparent by marker analyses. The Six3-positive most anteriorneuroectoderm, Pax6-positive forebrain, and Emx2-positive caudalforebrain primordium were mostly normal (Fig. 7B–D). However,Wnt1-expression in midbrain primordium was reduced (Fig. 7E), andGbx2-positive anterior hindbrain expanded rostrally with the reduc-tion of its expression intensity (Fig. 7F and G). The Fgf8 expressionwasalso reduced and expanded anteriorly (Fig. 7H). En1 expressionnormally covers posterior midbrain and anterior hindbrain. Itsexpression intensity was also reduced (Fig. 7I); it is probable thatEn1-positive midbrain primordium was reduced and En1-positiveanterior hindbrain primordium was concurrently expanded in theTKO mutants. Therefore, the caudalization of caudal forebrain and

Fig. 6.Marker analyses of E9.5 phenotype. Genes examined by RNA in situ hybridizationare: (A) Emx2, (B) Fezl, (C) Pax6, (D) Shh, (E) Dmbx1, (F) Wnt1, (G) Fgf8, (H) Gbx2, (I)Vgll1 and (J) Hoxb1. Embryos in (a) panels are wild type, (b) DKO mutant and (c) TKOmutant embryos; all are whole mount lateral views (anterior is at left). Arrows in(C, F,H, I) indicate the rhombomere 2 and in (D) ventral diencephalon where zona limitansdevelop; arrowheads in (D) most rostral ventral forebrain and in (G) anterior neuralridge; brackets in (G, H) rhombomere 1. Solid lines numbered (a-l) in (Ja) and (Jc)indicate the sites of BrdU uptake (Fig. 8) and TUNEL assays.

Fig. 7. Onset of diencephalic/mesencephalic defects. (A) shows the Otx2 expression atE7.75, and all the others the gene expression at E8.5. Genes examined are: (B) Six3,(C) Pax6, (D) Emx2, (E) Wnt1, (F, G) Gbx2, (H) Fgf8 and (I) En1. Embryos in (a) panelsare wild type, (b) DKO mutant and (c) TKO mutant embryos. (A-F, H, I) show lateral(anterior is at left) and (G) dorsal (anterior is at bottom) whole mount views. In (E–I)filled arrowheads indicate the anterior ends and open ones the posterior ends of eachexpression.

399Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

midbrain starts at E8.5, when the FM1 and FM2 enhancers arebecoming active and the AN enhancer is going to lose its activity, andis completed by E9.5. In E8.5 DKO mutants no defects were apparentin the expression of these markers (Fig. 7A–Ib).

A decrease in cell growth in diencephalic/mesencephalic primor-dium and its increase inmetencephalic primordiummight explain thecaudalization in TKO mutants. However, no change was observed inBrdU uptake in these regions at E9.5 by the TKO mutation (Fig. 6J;Fig. 8). An increase in cell death in diencephalic/mesencephalicprimordium and its decrease in metencephalic primordium mightalso explain the caudalization in TKO mutants. The number of TUNELpositive cells was counted at the sites same with the BrdU assay(Fig. 6J). In wild type embryos one TUNEL-positive cell was foundamong 182±65 cells throughout diencephalic/mesencephalic andmetencephalic regions. The TKO mutation caused no change in thisfrequency (150±37 cells).

Recovery of the DKO defect: the third enhancer

Mesencephalic region was caudalized into metencephalon in E9.5DKO mutants, but subsequently restored by E10.5. No regionalincrease in cell growth or cell death was prominent in E9.5 or E10.5DKO mutants by the BrdU uptake or TUNEL assay (data not shown);the recovery was most likely the retransformation of expandedmetencephalon into mesencephalon. It is most easily explained by theincrease in the Otx1 expression and/or the recovery of the Otx2expression later than E9.5 in the DKO mutants. Quantitative RT-PCRassay was conducted to determine the Otx level at each stage;embryos were dissected as illustrated in Fig. 1Ba–f. At E8.0 and E9.5the level of the Otx1 expression in DKO mutants was almost the sameas that in wild type, while it was somewhat increased in DKOdiencephalon/mesencephalon at E10.5 (Fig. 9Aa–c). In situ hybrid-ization suggested an increase in the dorsal mesencephalic region(Fig. 1Bc, f; asterisk).

At E8.0 when the AN enhancer is active and FM1 and FM2enhancers are inactive, the level of the Otx2 expression in the DKOmutants was nearly the same as that in wild type (Fig. 9Ad). At E9.5the level of the Otx2 expression in the DKO was about one fourth ofthat in wild type embryos (Fig. 9Ae). The AN enhancer has the activityin medial pallium even at this stage as described above (Fig. 5Ea;Fig. 9Be, f; Kurokawa et al., 2004a); Otx2 is also expressed inventral telencephalon under an unidentified enhancer (Fig. 5Ba, Fc;Fig. 9Be, f). Furthermore, the AME and CM enhancers for the Otx2expression in cephalic mesenchyme, which locate in the promoterregion, are active at this stage (Kimura et al., 1997; Kurokawa et al.,2010). We ascribe the residual Otx2 expression to these Otx2expressions. Indeed, by in situ hybridization the Otx2 expressionwas not apparent in diencephalic ormesencephalic region (Fig. 9Ba–d).However, at E10.5 the level of the Otx2 expression in DKO dien/mesencephalic region was about 70% of that in wild type (Fig. 9Af). Insitu hybridization demonstrated the significant Otx2 expression indiencephalic and mesencephalic regions (Fig. 9Be–j); the expression intelencephalon was confined to medial pallium and the most ventraltelencephalon as the endogeneous Otx2 expression (Fig. 9Be, f).

The AN enhancer, which is normally inactive later than E9.0, mighthave been compensatorily activated in the absence of the FM1 andFM2 enhancers. To assess this possibility we introduced a lacZ genecombined with the 15 kb XhoI fragment that covers the AN enhancerinto the DKO mutant background (Kurokawa et al., 2004a). However,the fragment did not exhibit any βGal expression in E10.5 DKOdiencephalic/mesencephalic region (data not shown). Therefore, it ismost probable that there is another enhancer for the Otx2 expressionin diencephalon and mesencephalon later than E9.5.

Discussion

We propose that at the stage when the rostral brain is regionalizedinto telencephalon, diencephalon, mesencephalon and metencepha-lon at E8.5–E9.5, Otx2 and Otx1 protect the caudalization of thediencephalon and mesencephalon into metencephalon. The rostral

Fig. 8. Cell proliferation by BrdU uptake assay in E9.5 diencephalic, mesencephalic and metencephalic regions. The number of BrdU-positive cells among total cells (labeling index)was counted on areas of 100 μm width in the E9.5 neuroepithelial layers given by horizontal sections. The sites of the counting are indicated in Fig. 6J.

400 Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

forebrain that generates neopallium, ganglionic eminences andhypothalamus becomes insensitive to the caudalization when theAN enhancer becomes inactive and FM1 and FM2 enhancers becomeactive around E8.5. In the telencephalon, however, Otx2 and Otx1 areessential to the medial pallium development later than E9.5.

Previous studies on neuroectoderm development with several Otxmutants have observed the phenotypes brought about by cumulativeeffects of Otx2 partial dysfunctions at different stages and sites; thephenotypes were inherently highly variable (Suda et al., 1997, 1999;Boyl et al., 2001; Acampora et al., 1995, 1997, 1998; Matsuo et al.,1995). Enhancer analysis, however, has demonstrated that Otx2functions can be differentiated into each stage and site (Kimura et al.,1997; Kimura-Yoshida et al., 2004; Kurokawa et al., 2004a,b). In thisreport we have focused on the Otx2 functions in forebrain andmidbrain at E8.5–E9.5 under conditions at which Otx2 functions atearlier stages and other sites (epiblast, visceral endoderm, anterior

Fig. 9. Otx1 and Otx2 expression in DKO mutants. (A) gives quantitative RT-PCR analyses of Oand DKOmutant brains at each stage were dissected as shown in Fig. 1Ba–f, respectively. Theet al., 2009). (B) shows Otx2 in situ hybridization at E9.5 (a–d) and E10.5 (e–j) at the plan

mesendoderm, anterior neuroectoderm at presomite stage andcephalic mesenchymal cells) are normally preserved. The phenotypeswere uniform among TKO or DKO mutants. We consider that the TKOphenotype is the caudalization but not the loss of diencephalon/mesencephalon. This is based on the following observations: (1) themorphological expansion of cerebellum, isthmus and pons, (2) theexpansion of Fgf8 and Gbx2-positive isthmus and -negative rhombo-mere 1, and (3) no apparent increase of cell death in the diencephalic/mesencephalic region or of cell growth in themetencephalic region bythe TKOmutation. This, however, must be confirmed in future studiesby tracing the fate of the initially Otx2-positive (under the ANenhancer) anterior neuroectoderm cells into expandedmetencephalicstructures in the TKO mutants.

We previously reported that Otx1 null and Otx2 FM1 or FM2enhancer double mutants lose mesencephalon and diencephalon(Kurokawa et al., 2004b). However, thalamic eminence and

tx1 (a–c) and Otx2 (d–f) expression at E8.0 (a, d), E9.5 (b, e) and E10.5 (c, f). Wild typeOtx1 or Otx2 expression is given as the relative quantity to the Tbp expression (Svingen

es indicated in Fig. 1Bb, c, e, f, respectively.

401Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

prethalamus developed in the most severe Otx1−/−Otx2 ΔFM1/ΔFM1

mutants. In Otx1−/−Otx2+/− mutants by Acampora et al. (1997) andin Otx1−/−Otx2+/− mutants by us (Suda et al., 1997) thalamiceminence and prethalamus also developed; the entire territory underor the anterior limit of Otx control for forebrain/midbrain develop-ment at E8.5–E9.5 has remained uncertain. In zebrafish it wasproposed that Otx is not required for the prethalamus development(Scholpp et al., 2007). With the TKO mutants we propose that theterritory comprises the mesencephalon and diencephalon thatincludes prethalamus and thalamic eminence anterior to zonalimitans. Hypothalamus may or may not be a part of diencephalon;it is a ventral and more anterior structure than prethalamus at thetelencephalic level (Puelles and Rubenstein, 2003). Hypothalamusdeveloped in the TKO mutants. Apparently during head development,each step and each site require a differentOtx dosage (Kurokawa et al.,2004b). Otx2 positive visceral endoderm, and probably anteriormesendoderm, develops at the minimum Otx dosage; their defectsoccur only with Otx2 homozygous null mutation (Acampora et al.,1995; Matsuo et al., 1995). Anterior neuroectoderm requires a higherOtx2 dosage for its development as seen in Otx2ΔAN/− mutants(Kurokawa et al., 2004a), forebrain/ midbrain even more (this study),and cephalic mesenchyme the highest dosage of Otx2 at theheterozygous level (Matsuo et al., 1995). Furthermore, in forebrain/midbrain development caudal mesencephalon requires a higher Otxdosage and prethalamus a lower dosage (Puelles et al., 2003;Kurokawa et al., 2004b; this study). It is an intriguing question howthis different dosage requirement has evolved and been establishedfor Otx usage at each step of head development.

The medial pallium develops after the neural tube closure at thetelencephalic level around E9.0. However, it is possible that theterritory is predetermined before the closure or that the TKO defect inmedial pallium forms a part of the caudalization of caudal forebrainandmidbrain at E8.5. In the prosomericmodel themedial palliumwasonce thought to constitute the p4 prosomere, together with thalamiceminence, in front of the p3 prethalamus (Bulfone et al., 1993;Rubenstein et al., 1998). In the TKOmutants, however, medial palliumdeveloped apparently normally at E9.5 just after the tube closure. Weconsider that the medial pallium defect in the TKO mutants is not apart of the caudalization of caudal forebrain and midbrain at E8.5–E9.0. Later than E9.5, Otx1 continues to be expressed throughout thepallium including the medial one. Otx2 also continues to be expressedin medial pallium under the AN, FM1 and FM2 enhancers. Hypomor-phic phenotypes in a series of mutants in previous and presentstudies (Otx1−/−, Otx2ΔAN/ΔAN, Otx2ΔFM1/ΔFM, Otx2ΔFM2/ΔFM ,Otx2ΔFM1ΔFM2/ΔFM1ΔFM2 and Otx1−/−Otx2ΔFM1 ΔFM2/ΔFM1ΔFM2) indicatethat the Otx1 expression and the Otx2 expression under AN, FM1 andFM2enhancers function complementarily in the development ofmedialpallium. It is of great interest why and how in telencephalon only themedial pallium continues to require Otx for its development; it alsoremains for a further study whether the medial pallium defect in theTKOmutants is due to its loss or a transformation into another structure.

This study proposes it is the metencephalon (isthmus andrhombomere 1), and does not include myelencephalon (rhombomere2 and more posterior rhombomeres), into which the rostral braintransforms as a result of the Otx deficiency in TKO mutants. In the Otx2mutant that lacks the AN enhancer, Otx2ΔAN/−, the entire anteriorneuroectoderm is also transformed into themetencephalon (Kurokawaet al., 2004a; our unpublished results). Moreover, in Emx2−/−Otx2+/−

double mutants that lose diencephalon, the metencephalon is alsocompensatorily expanded (Suda et al., 2001; our unpublished data). Inaddition, the ectoderm that develops inOtx2/Cripto doublemutants hasthe nature of isthmus and rhombomere 1 (Kimura et al., 2001); Otx2mutant exhibits a headless phenotype andCriptomutant a trunkless one(Matsuo et al., 1995; Ding et al., 1998). Isthmus and rhombomere 1might have a special position in the grand design of the mammalianbody plan.

We consider that initially the entire anterior neuroectoderm issusceptible to the isthmus and rhombomere 1 fate, and that Otx2under the AN enhancer protects this (Kurokawa et al., 2004a;unpublished data). Subsequently the rostral forebrain that generatesneopallium, ganglionic eminences and hypothalamus becomes insen-sitive to the caudalization and does not require Otx to suppress itaround E8.5 when the AN enhancer is turned off and FM1 and FM2enhancers are turned on. However, caudal forebrain and midbrainremain susceptible to the isthmus and rhombomere 1 fate by E10.5(see below), and Otx2 under FM1/FM2 enhancers and Otx1 protectthe caudalization. The Otx targets to suppress the caudalization,however, have not yet been identified. Many studies have suggestedthat Otx2 and Gbx2 counteract each other to determine the midbrain/hindbrain boundary (Millet et al., 1999; Li and Joyner, 2001;Martinez-Barbera et al., 2001), but the precise mechanism remainsuncertain.

At E9.5 mesencephalon and diencephalon were greatly reduced inthe DKO mutants, but the reduction was largely recovered at E10.5.The recovery suggests that at E9.5–E10.5 the region is still susceptibleto both caudal forebrain/midbrain and anterior hindbrain fates; somedefects were recovered even after E12.5 by E15.5. The question washow the retransformation became possible. The initial transformationinto metencephalon was caused by the decrease in Otx dosage beyonda threshold at E8.5–9.5 in the absence of FM1 and FM2 enhancers.Retransformation is most easily explained by an Otx dosage increaselater than E9.5. Quantitative RT-PCR and in situ hybridizationsuggested that the Otx2 expression is absent in the E9.5 DKOdiencephalic/mesencephalic region, but recovers at E10.5. Thepossibility cannot be ruled out from the present analysis that theAN enhancer was compensatorily activated in the absence of FM1 andFM2 enhancers, directed by sequences located in other regions, butthe recovery is most probably caused by the third enhancer; itsactivity, if any, is minimal at E9.5 but increasesmarkedly by E10.5. Ourprevious enhancer survey did not identify such an enhancer in a 290-kb genomic region surrounding Otx2 gene (Kurokawa et al., 2004a).The enhancer may exist far away and could be unique to mammal(Kurokawa et al., 2004b). The level of the Otx2 expression by thisenhancer would be insufficient to restore the caudalization in TKOmutants.

Supplementarymaterials related to this article can be found onlineat doi:10.1016/j.ydbio.2010.08.028.

Acknowledgments

We are indebted to the Laboratory for Animal Resources andGenetic Engineering of RIKEN CDB for animal housing and mutantmouse production. We thank Drs. A.P. McMahon, J.L.R. Rubenstein, P.Gruss, H. Clevers, A. Joyner, D.G.Wilkinson, G. Martin and T. Miyamotofor their kind gift of in situ hybridization probes. This work wassupported by a Grant-in-Aid for Creative Scientific Research from theJapan Society for the Promotion of Science.

References

Acampora, D., Mazan, S., Lallemand, Y., Avantaggiato, V., Maury, M., Simeone, A., Brûlet,P., 1995. Forebrain and midbrain regions are deleted in Otx2-/- mutants due to adefective anterior neuroectoderm specification during gastrulation. Development121, 3279–3290.

Acampora, D., Avantaggiato, V., Tuorto, F., Simeone, A., 1997. Genetic control of brainmorphogenesis throughOtx genedosage requirement. Development 124, 3639–3650.

Acampora, D., Avantaggiato, V., Tuorto, F., Briata, P., Corte, G., Simeone, A., 1998. Visceralendoderm-restricted translation of Otx1 mediates recovery of Otx2 requirementsfor specification of anterior neural plate and normal gastrulation. Development125, 5091–5104.

Acampora, D., Di Giovannantonio, L.G., Di Salvio, M., Mancuso, P., Simeone, A., 2009.Selective inactivation of Otx2 mRNA isoforms reveals isoform-specific requirementfor visceral endoderm anteriorization and head morphogenesis and highlights celldiversity in the visceral endoderm. Mech. Dev. 126, 882–897.

Bally-Cuif, L., Cholley, B., Wassef, M., 1995. Involvement ofWnt-1 in the formation of themes/metencephalic boundary. Mech. Dev. 53, 23–34.

402 Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

Boyl, P.P., Signore, M., Acampora, D., Martinez-Barbera, J.P., Ilengo, C., Annino, A., Corte,G., Simeone, A., 2001. Forebrain and midbrain development requires epiblast-restricted Otx2 translational control mediated by its 3' UTR. Development 128,2989–3000.

Bulfone, A., Puelles, L., Porteus, M.H., Frohman, M.A., Martin, G.R., Rubenstein, J.L., 1993.Spatially restricted expression of Dlx-1, Dlx-2 (Tes-1), Gbx-2, and Wnt-3 in theembryonic day 12.5 mouse forebrain defines potential transverse and longitudinalsegmental boundaries. J. Neurosci. 13, 3155–3172.

Crossley, P.H., Martin, G.R., 1995. The mouse Fgf8 gene encodes a family of polypeptidesand is expressed in regions that direct outgrowth and patterning in the developingembryo. Development 121, 439–451.

Davis, C.A., Joyner, A.L., 1988. Expression patterns of the homeo box-containing genesEn-1 and En-2 and the proto-oncogene int-1 diverge during mouse development.Genes Dev. 2, 1736–1744.

Ding, J., Yang, L., Yan, Y.T., Chen, A., Desai, N., Wynshaw-Boris, A., Shen, M.M., 1998.Cripto is required for correct orientation of the anterior–posterior axis in the mouseembryo. Nature 395, 702–707.

Echelard, Y., Epstein, D.J., St-Jacques, B., Shen, L., Mohler, J., McMahon, J.A., McMahon, A.P.,1993. Sonic hedgehog, a member of a family of putative signaling molecules, isimplicated in the regulation of CNS polarity. Cell 75, 1417–1430.

Flenniken, A.M., Gale, N.W., Yancopoulos, G.D., Wilkinson, D.G., 1996. Distinct andoverlapping expression patterns of ligands for Eph-related receptor tyrosinekinases during mouse embryogenesis. Dev. Biol. 179, 382–401.

Frantz, G.D., Weimann, J.M., Lvin, M.E., McConnell, S.K., 1994. Otx1 and Otx2 definelayers and regions in developing cerebral cortex and cerebellum. J. Neurosci. 14,5725–5740.

Fujii, T., Pichel, J.G., Taira, M., Toyama, R., Dawid, I.B., Westphal, H., 1994. Expressionpatterns of the murine LIM class homeobox gene lim1 in the developing brain andexcretory system. Dev. Dyn. 199, 73–83.

Hill, R.E., Jones, P.F., Rees, A.R., Sime, C.M., Justice, M.J., Copeland, N.G., Jenkins, N.A.,Graham, E., Davidson, D.R., 1989. A new family of mouse homeo box-containinggenes: molecular structure, chromosomal location, and developmental expressionof Hox-7.1. Genes Dev. 3, 26–37.

Hirata, T., Suda, Y., Nakao, K., Narimatsu, M., Hirano, T., Hibi, M., 2004. Zinc finger genefez-like functions in the formation of subplate neurons and thalamocortical axons.Dev. Dyn. 230, 546–556.

Kimura, S., Hara, Y., Pineau, T., Fernandez-Salguero, P., Fox, C.H., Ward, J.M., Gonzalez, F.J.,1996. The T/ebpnullmouse: thyroid-specific enhancer-bindingprotein is essential forthe organogenesis of the thyroid, lung, ventral forebrain, and pituitary. Genes Dev. 10,60–69.

Kimura, C., Takeda, N., Suzuki, M., Oshimura, M., Aizawa, S., Matsuo, I., 1997. Cis-actingelements conserved between mouse and pufferfish Otx2 genes govern theexpression in mesencephalic neural crest cells. Development 124, 3929–3941.

Kimura, C., Yoshinaga, K., Tian, E., Suzuki, M., Aizawa, S., Matsuo, I., 2000. Visceralendoderm mediates forebrain development by suppressing posteriorizing signals.Dev. Biol. 225, 304–321.

Kimura, C., Shen, M.M., Takeda, N., Aizawa, S., Matsuo, I., 2001. Complementaryfunctions of Otx2 and Cripto in initial patterning of mouse epiblast. Dev. Biol. 235,12–32.

Kimura, J., Suda, Y., Kurokawa, D., Houssain, Z.M., Nakamura, M., Takahashi, M., Hara, A.,Aizawa, S., 2005. Emx2 and Pax6 function in cooperation with Otx2 and Otx1 todevelop caudal forebrain primordium that includes future archipallium. J. Neurosci.25, 5097–5108.

Kimura-Yoshida, C., Kitajima, K., Oda-Ishii, I., Tian, E., Suzuki, M., Yamamoto, M., Suzuki,T., Kobayashi, M., Aizawa, S., Matsuo, I., 2004. Characterization of the pufferfish Otxcis-regulators reveals evolutionally conserved genetic mechanisms for vertebratehead specification. Development 131, 57–71.

Kimura-Yoshida, C., Nakano, H., Okamura, D., Nakao, K., Yonemura, S., Belo, J.A., Aizawa,S., Matsui, Y., Matsuo, I., 2005. Canonical Wnt signaling and its antagonist regulateanterior–posterior axis polarization by guiding cell migration in mouse visceralendoderm. Dev. Cell 9, 639–650.

Kimura-Yoshida, C., Tian, E., Nakano, H., Amazaki, S., Shimokawa, K., Rossant, J., Aizawa,S., Matsuo, I., 2007. Crucial roles of Foxa2 in mouse anterior–posterior axispolarization via regulation of anterior visceral endoderm-specific genes. Proc. Natl.Acad. Sci. U. S. A. 104, 5919–5924.

Korinek, V., Barker, N., Willert, K., Molenaar, M., Roose, J., Wagenaar, G., Markman, M.,Lamers, W., Destree, O., Clevers, H., 1998. Twomembers of the Tcf family implicatedin Wnt/β-catenin signaling during embryogenesis in the mouse. Mol. Cell. Biol. 18,1248–1256.

Kurokawa, D., Takasaki, N., Kiyonari, H., Nakayama, R., Kimura-Yoshida, C., Matsuo, I.,Aizawa, S., 2004a. Regulation of Otx2 expression and its functions in mouse epiblastand anterior neuroectoderm. Development 131, 3307–3317.

Kurokawa, D., Kiyonari, H., Nakayama, R., Kimura-Yoshida, C., Matsuo, I., Aizawa, S.,2004b. Regulation of Otx2 expression and its functions in mouse forebrain andmidbrain. Development 131, 3319–3331.

Kurokawa, D., Sakurai, Y., Inoue, A., Nakayama, R., Takasaki, N., Suda, Y., Miyake, T.,Amemiya, C.T., Aizawa, S., 2006. Evolutionary constraint on Otx2 neuroectodermenhancers-deep conservation from skate to mouse and unique divergence inteleost. Proc. Natl. Acad. Sci. U. S. A. 103, 19350–19355.

Kurokawa, D., Ohmura, T., Ogino, H., Takeuchi, M., Inoue, A., Inoue, F., Suda, Y., Aizawa,S., 2010. Evolutionary origin of the Otx2 enhancer for its expression in visceralendoderm. Dev. Biol. 342 (1), 110–120.

Li, J.Y.H., Joyner, A.L., 2001. Otx2 and Gbx2 are required for refinement and not inductionof mid-hindbrain gene expression. Development 128, 4979–4991.

Martinez-Barbera, J.P., Signore, M., Boyl, P.P., Puelles, E., Acampora, D., Gogoi, R.,Schubert, F., Lumsden, A., Simeone, A., 2001. Regionalisation of anterior

neuroectoderm and its competence in responding to forebrain and midbraininducing activities depend on mutual antagonism between OTX2 and GBX2.Development 128, 4789–4800.

Matsuo, I., Kuratani, S., Kimura, C., Takeda, N., Aizawa, S., 1995. Mouse Otx2 functions inthe formation and patterning of rostral head. Genes Dev. 9, 2646–2658.

McMahon, A.P., Bradley, A., 1990. The Wnt-1 (int-1) proto-oncogene is required fordevelopment of a large region of the mouse brain. Cell 21, 1073–1085.

Millet, S., Campbell, K., Epstein, D.J., Losos, K., Harris, E., Joyner, A.L., 1999. A role forGbx2 in repression of Otx2 and positioning the mid/hindbrain organizer. Nature401, 161–164.

Miyamoto, T., Kawahara, A., Teufel, A., Mukhopadhyay, M., Zhao, Y., Dawid, I.B.,Westphal, H., 2002. Mbx, a novel mouse homeobox gene. Dev. Genes Evol. 212,104–106.

Ohtoshi, A., Nishijima, I., Justice, M.J., Behringer, R.R., 2002. Dmbx1, a novelevolutionarily conserved paired-like homeobox gene expressed in the brain ofmouse embryos. Mech. Dev. 110, 241–244.

Oliver, G., Mailhos, A., Wehr, R., Copeland, N.G., Jenkins, N.A., Gruss, P., 1995. Six3, amurine homologue of the sine oculis gene, demarcates the most anterior border ofthe developing neural plate and is expressed during eye development. Develop-ment 121, 4045–4055.

Parr, B.A., Shea, M.J., Vassileva, G., McMahon, A.P., 1993. Mouse Wnt genes exhibitdiscrete domains of expression in the early embryonic CNS and limb buds.Development 119, 247–261.

Pleasure, S.J., Anderson, S., Hevner, R., Bagri, A., Marin, O., Lowenstein, D.H., Rubenstein,J.L., 2000. Cell migration from the ganglionic eminences is required for thedevelopment of hippocampal GABAergic interneuron. Neuron 28, 727–740.

Puelles, L., Rubenstein, J.L., 2003. Forebrain gene expression domains and the evolvingprosomeric model. Trends Neurosci. 26, 469–476.

Puelles, E., Acampora, D., Lacroix, E., Signore, A., Annino, A., Tuorto, F., Filosa, S., Corte, G.,Wurst, W., Ang, S.L., Simeone, A., 2003. Otx dose-dependent integrated control ofantero-posterior and dorso-ventral patterning of midbrain. Nat. Neurosci. 6,453–460.

Rhinn, M., Dierich, A., Shawlot, W., Behringer, R., Meur, M.L., Ang, S.-L., 1998. Sequentialroles for Otx2 in visceral endoderm and neuroectoderm for forebrain and midbraininduction and specification. Development 125, 845–856.

Roelink, H., Nusse, R., 1991. Expression of two members of the Wnt family duringmouse development-restricted temporal and spatial patterns in the developingneural tube. Genes Dev. 5, 381–388.

Rowitch, D.H., McMahon, A.P., 1995. Pax-2 expression in the murine neural plateprecedes and encompasses the expression domains of Wnt-1 and En-1. Mech. Dev.52, 3–8.

Rubenstein, J.L., Shimamura, K., Martinez, S., Puelles, L., 1998. Regionalization of theprosencephalic neural plate. Annu. Rev. Neurosci. 21, 445–477.

Scholpp, S., Foucher, I., Staudt, N., Peukert, D., Lumsden, A., Houart, C., 2007. Otx1l, Otx2and Irx1b establish and position the ZLI in the diencephalon. Development 134,3167–3176.

Shibata, M., Kurokawa, D., Nakao, H., Ohmura, T., Aizawa, S., 2008. MicroRNA-9modulates Cajal-Retzius cell differentiation by suppressing Foxg1 expression inmouse medial pallium. J. Neurosci. 28, 10415–10421.

Shimamura, K., Hartigan, D.J., Martinez, S., Puelles, L., Rubenstein, J.L.R., 1995.Longitudinal organization of the anterior neural plate and neural tube. Develop-ment 121, 3923–3933.

Shinozaki, K., Miyagi, T., Yoshida, M., Miyata, T., Ogawa, M., Aizawa, S., Suda, Y., 2002.Absence of Cajal–Retzius cells and subplate neurons associated with defects oftangential cell migration from ganglionic eminence in Emx1/2 double mutantcerebral cortex. Development 129, 3479–3492.

Shinozaki, K., Yoshida, M., Nakamura, M., Aizawa, S., Suda, Y., 2004. Emx1 and Emx2cooperate in initial phase of archipallium development. Mech. Dev. 121,475–489.

Simeone, A., Acampora, D., Gulisano, M., Stornaiuolo, A., Boncinelli, E., 1992. Nestedexpression domains of four homeobox genes in developing rostral brain. Nature358, 687–690.

Simeone, A., Acampora, D., Mallamaci, A., Stornaiuolo, A., D'Apice, M.R., Nigro, V.,Boncinelli, E., 1993. A vertebrate gene related to orthodenticle contains ahomeodomain of the bicoid class and demarcates anterior neuroectoderm in thegastrulating mouse embryo. EMBO J. 12, 2735–2747.

Suda, Y., Matsuo, I., Aizawa, S., 1997. Cooperation between Otx1 and Otx2 genes indevelopmental patterning of rostral brain. Mech. Dev. 69, 125–141.

Suda, Y., Nakabayashi, J., Matsuo, I., Aizawa, S., 1999. Functional equivalency betweenOtx2 and Otx1 in development of the rostral head. Development 126, 743–757.

Suda, Y., Houssain, Z.M., Kobayashi, C., Hatano, O., Yoshida, M., Matsuo, I., Aizawa, S.,2001. Emx2 directs the development of diencephalon in cooperation with Otx2.Development 128, 2433–2450.

Svingen, T., Spiller, C.M., Kashimada, K., Harley, V.R., Koopman, P., 2009. Identification ofsuitable normalizing genes for quantitative real-time RT-PCR analysis of geneexpression in fetal mouse gonads. Sex. Dev. 3, 194–204.

Tian, E., Kimura, C., Takeda, N., Aizawa, S., Matsuo, I., 2002. Otx2 is required to respondto signals from anterior neural ridge for forebrain specification. Dev. Biol. 242,204–223.

Wakasugi, S., Maeda, S., Shimada, K., Nakashima, H., Migita, S., 1985. Structuralcomparisons between mouse and human prealbumin. J. Biochem. 98, 1707–1714.

Walther, C., Gruss, P., 1991. Pax-6, a murine paired box gene, is expressed in thedeveloping CNS. Development 113, 1435–1449.

Wilkinson, D.G., 1993. In situ hybridization. In: Stern, C.D., Holland, P.W.H.(Eds.), Essential developmental biology: A practical approach. IRL Press,Oxford, pp. 257–274.

403Y. Sakurai et al. / Developmental Biology 347 (2010) 392–403

Yagi, T., Tokunaga, T., Furuta, Y., Nada, S., Yoshida, M., Tsukada, T., Saga, Y., Takeda, N.,Ikawa, Y., Aizawa, S., 1993a. A novel ES cell line, TT2, with high germline-differentiating potency. Anal. Biochem. 214, 70–76.

Yagi, T., Tokunaga, T., Furuta, Y., Nada, S., Yoshida, M., Tsukada, T., Saga, Y., Takeda, N.,Ikawa, Y., Aizawa, S., 1993b. A novel negative selection for homologousrecombinants using diphtheria toxin A fragment gene. Anal. Biochem. 214, 70–76.

Yoshida, M., Suda, Y., Matsuo, I., Miyamoto, N., Takeda, N., Kuratani, S., Aizawa, S., 1997.Emx1 and Emx2 functions in development of dorsal telencephalon. Development124, 101–111.

Yoshida, M., Assimacopoulos, S., Jones, K.R., Grove, E.A., 2006. Massive loss of Cajal–Retzius cells does not disrupt neocortical layer order. Development 133,537–545.