Early Embryonic Death in Mice Lacking the {beta}-Catenin-Binding Protein Duplin

The Wnt signaling pathway plays a pivotal role in vertebrateearly development and morphogenesis. Duplin (axis duplicationinhibitor) interacts with ß-catenin and prevents itsbinding to Tcf, thereby inhibiting downstream Wnt signaling.Here we show that Duplin is expressed predominantly from early-to mid-stage mouse embryogenesis, and we describe the generationof mice deficient in Duplin. Duplin^–/– embryos manifestgrowth retardation from embryonic day 5.5 (E5.5) and developmentalarrest accompanied by massive apoptosis at E7.5. The mutantembryos develop into an egg cylinder but do not form a primitivestreak or mesoderm. Expression of ß-catenin targetgenes, including those for T (brachyury), Axin2, and cyclinD1, was not increased in Duplin^–/– embryos, suggestingthat the developmental defect is not simply attributable toupregulation of Wnt signaling caused by the lack of this inhibitor.These results suggest that Duplin plays an indispensable role,likely by a mechanism independent of inhibition of Wnt signaling,in mouse embryonic growth and differentiation at an early developmentalstage.

INTRODUCTION

Top

Introduction

Wnt proteins constitute a large family of cysteine-rich secretedligands that control development in organisms ranging from nematodesto mammals (46). The intracellular signaling pathway triggeredby Wnt is also evolutionarily conserved and regulates cellularproliferation and differentiation (5, 35, 46). In the absenceof Wnt, casein kinase I ${alpha}$ (CKI ${alpha}$ ) and glycogen synthase kinase 3ß(GSK-3ß) target cytoplasmic ß-catenin fordegradation (19, 27, 50). Axin forms a complex with GSK-3ß,CKI ${alpha}$ , ß-catenin, and the adenomatous polyposis coliprotein, and it stimulates the CKI ${alpha}$ -dependent and GSK-3ß-dependentphosphorylation of ß-catenin (12, 19, 22, 24, 27).Phosphorylated ß-catenin in turn forms a complex withFbw1 (also known as ß-TrCP/FWD1), a member of theF-box protein family, resulting in the degradation of ß-cateninby the ubiquitin/proteasome pathway (25).

The binding of Wnt to its cell surface receptor, consistingof Frizzled, as well as lipoprotein receptor-related protein(LRP) 5 and LRP6, triggers the accumulation of ß-cateninin the cytoplasm as a result of the inhibition by Dvl of ß-cateninphosphorylation. The accumulated ß-catenin is thentranslocated to the nucleus, where it binds to the transcriptionfactor T-cell factor, or Tcf (also known as lymphoid-enhancerfactor, or Lef), and thereby induces the expression of variousgenes (5, 18, 35). Several proteins in the nucleus that bindto Tcf regulate the formation of the ß-catenin-Tcf-DNAcomplex. ß-Catenin signaling is thus regulated inboth the cytoplasm and the nucleus.

Groucho binds to Tcf and represses the expression of Tcf targetgenes (8). Drosophila CREB-binding protein interacts with DrosophilaTcf and reduces its affinity for Armadillo (Drosophila ß-catenin)(43), but mammalian CREB-binding protein and the related proteinp300 act synergistically with ß-catenin to activategene expression (16, 41). COOH-terminal binding protein alsobinds to Tcf and inhibits its transactivation activity (6).NEMO-like kinase associates directly with and phosphorylatesTcf, resulting in inhibition of the binding of the ß-catenin-Tcfcomplex to DNA (20). Pontin52 and Reptin52, both of which aretranscription factors that bind directly to ß-cateninand the TATA-binding protein, also inhibit the transactivationactivity of the ß-catenin-Tcf complex (2, 3). In addition,XSox17 (Sox) binds to ß-catenin and inhibits the inductionof gene expression by the ß-catenin-Tcf complex (51).Inhibitor of ß-catenin and Tcf-4 (ICAT) binds to ß-catenin,inhibits formation of the ß-catenin-Tcf-DNA complex,and thereby negatively regulates Wnt signaling (40). Wnt signalingthrough Tcf thus appears to be inhibited by several mechanismsat the level of the ß-catenin-Tcf complex in the nucleus.

Duplin (axis duplication inhibitor) was identified by yeasttwo-hybrid screening of a rat brain cDNA library with the PDZdomain of Dvl-1 as the bait (34). Duplin is a protein of 749amino acids whose COOH-terminal region (residues 482 to 749)contains several clusters of basic residues and a nuclear localizationsignal. Duplin does not form a complex with Dvl in COS cells,but it binds directly to the region of ß-catenin thatincludes the Armadillo repeats, resulting in inhibition of theinteraction of ß-catenin with Tcf-4. Although overexpressionof Duplin does not inhibit Wnt-3a-dependent accumulation ofß-catenin in the cytoplasm, it does block the activationof Tcf-4 by Wnt-3a. Dorsal injection of Duplin mRNA into Xenopusembryos results in loss of the head, as well as inhibition ofthe expression of siamois, whose product mediates the effectsof the Wnt signaling pathway on axis formation. Furthermore,Duplin inhibits Wnt-8-dependent or ß-catenin-dependentformation of a secondary dorsal axis in Xenopus. It also inhibitssiamois-dependent axis duplication and blocks the Wnt signalingpathway at an additional point downstream by affecting the expressionof ß-catenin target genes (26). The physiologicalfunction of Duplin has remained largely unclear, however.

To examine the biological importance of Duplin in mammaliandevelopment, we disrupted the corresponding gene by homologousrecombination in the mouse. The development of Duplin-deficientanimals was arrested at gastrulation, with the embryos manifestingmassive apoptosis. Unexpectedly, the expression of ß-catenintarget genes was not increased in the mutant embryos, suggestingthat the lack of Duplin did not result in the constitutive activationof Wnt signaling during embryogenesis. Our data suggest thatDuplin is indispensable for normal mouse development probablyas a result of its function independent of inhibition of Wntsignaling pathway.

MATERIALS AND METHODS

Top

Materials and Methods

Immunoblot analysis. Embryonic stem (ES) cells and embryos from embryonic day 8.5(E8.5) to birth were lysed and subjected to immunoblot analysisas described previously (13) with polyclonal antibodies to Duplinand a monoclonal antibody to ${alpha}$ -tubulin (TU-01; Zymed).

Whole-mount in situ hybridization. Embryos were fixed with 4% paraformaldehyde in phosphate-bufferedsaline and then exposed to H₂O₂. Whole-mount in situ hybridizationwas performed as described previously (45). Duplin riboprobes(sense and antisense) corresponding to the entire open readingframe of the cDNA were synthesized with a DIG RNA labeling kit(Roche). Embryos were examined with a dissection microscope.

Construction of a targeting vector and generation of Duplin^–/– mice. Cloned DNA corresponding to the Duplin locus was isolated froma 129/Sv mouse genomic library (Stratagene). The targeting vectorwas constructed by replacing a 13-kb BamHI-HindIII fragmentof genomic DNA containing all Duplin exons with a PGK-lox-neo-poly(A)cassette. The vector thus contained 6.5- and 1.2-kb regionsof homology located 5' and 3', respectively, relative to theneomycin resistance gene (neo). A PGK-tk-poly(A) cassette wasligated at the 5' end of the targeting construct. The maintenance,transfection, and selection of ES cells were performed as describedpreviously (31). The recombination event was confirmed by Southernblot analysis with a 0.5-kb PstI-XbaI fragment of genomic DNAthat flanked the 3' homology region as the probe (see Fig. 2A).The expected sizes of hybridizing fragments after digestionwith BglII and EcoRV were 5.4 and 8.0 kb for the wild-type andmutant Duplin alleles, respectively. Mutant ES cells were microinjectedinto C57BL/6 mouse blastocysts, and the resulting male chimeraswere mated with C57BL/6 females. Germ line transmission of themutant allele was confirmed by Southern blot analysis. Heterozygousoffspring were intercrossed to produce homozygous mutant animals.For genotyping of embryos, DNA was extracted from whole embryosat E3.5 to E9.5 or from corresponding embryonic tissue removedfrom sections on microscope slides; the extracted DNA was thenanalyzed by the nested PCR with the primers PJL (5'-TGCTAAAGCGCATGCTCCAGACTG-3'),MN5 (5'-TATAGATTTCCTGTTTGATTTTCC-3'), RV1 (5'-AACTCCGTAACCATTTGTCTATTC-3'),KN11 (5'-ATGCTCCAGACTGCCTTGGGAAAA-3'), MN11 (5'-AAAGAATCACACTAGATCTAATCC-3'),and RV2 (5'-GAAACAATGTAAAACAGGCAAATG-3'). The present studyconformed with the guidelines of Kyushu University for the careand use of laboratory animals.

View larger version (33K):
[in this window]
[in a new window]
FIG. 2. Gene targeting of the mouse Duplin locus. (A) Schematic representation of the wild-type Duplin allele, the targeting vector, and the mutant allele after homologous recombination. A 13-kb genomic fragment including all exons of Duplin was replaced by a loxP-neo cassette. Exons and the probe used for hybridization are denoted by open and filled boxes, respectively. Restriction sites: B, BamHI; B2, BglII; E5, EcoRV; H3, HindIII; P, PstI; S, SalI; X, XbaI. (B) Southern blot analysis with the probe shown in panel A of tail genomic DNA from the offspring of heterozygote crosses after digestion with BglII and EcoRV. The 5.4- and 8.0-kb bands corresponding to the wild-type and mutant alleles, respectively, are indicated.

Histopathology and immunostaining. Embryos within their deciduae were fixed with 4% paraformaldehydein phosphate-buffered saline, embedded in paraffin, sectionedwith a cryostat at a thickness of 3 µm, and stained withhematoxylin-eosin. Immunostaining was performed as describedwith a monoclonal antibody to ß-catenin (C19220; TransductionLaboratories), polyclonal antibodies to brachyury (N-19; SantaCruz Biotechnology), conductin/Axin2 (S-19; Santa Cruz Biotechnology),cyclin D1 (553; MBL), cyclin D2 (M-20; Santa Cruz Biotechnology),or cyclin D3 (C-16; Santa Cruz Biotechnology). Immune complexeswere detected with a HistoMouse-Plus kit (Zymed) or with biotinylatedsecondary antibodies (1/400 dilution), an ABC kit (Vector),and diaminobenzidine (Sigma). Sections were examined with adifferential interference contrast microscope.

Culture of preimplantation embryos. Heterozygous male and female mutant mice were bred to obtainwild-type (Duplin^+/+), heterozygous (Duplin^+/–), and homozygousmutant (Duplin^–/–) embryos. The morning of the dayon which a vaginal plug was detected was designated E0.5. Embryosat E3.5 were collected by flushing oviducts or the uterus withHEPES-buffered medium 2 (M2; Sigma). Blastocysts were culturedfor 6 days in tissue culture dishes containing cES medium withoutleukemia inhibitory factor (Chemicon), and outgrowths were inspecteddaily and photographed to monitor their development.

Detection of apoptosis. For the TUNEL (terminal deoxynucleotidyltransferase-mediateddUTP-biotin nick end labeling) assay, paraffin-embedded sectionswere treated with H₂O₂, permeabilized for 15 min at 37°Cwith proteinase K (20 µg/ml; Sigma), and then incubatedfor 1 h at 37°C with a reaction mixture containing terminaldeoxynucleotidyl transferase (Gibco) and biotinylated dUTP (Boehringer).Labeled DNA was visualized with an ABC kit and diaminobenzidine.Sections were examined with a differential interference contrastmicroscope. Sections first treated with DNase (Wako) were usedas a positive control. For a negative control, the transferasewas omitted from the reaction mixture.

RESULTS

Top

Results

We first examined the expression of Duplin during mouse developmentby immunoblot analysis. Duplin was abundant both in ES cells(equivalent to E3.5) and in embryos at E8.5 and E12.5, but itsexpression was reduced at E16.5 and virtually undetectable innewborns (Fig. 1A). Duplin was also expressed in various adultmouse tissues at low levels (data not shown). The expressionof Duplin during early mouse development was also examined bywhole-mount in situ hybridization. Duplin mRNA was detectedthroughout embryos from E7.5 to E9.5 but was localized predominantlyin the brains, faces, branchial arches, limb buds, and tailbuds of embryos at E10.5 (Fig. 1B, and data not shown). Theseresults suggest that Duplin function is largely restricted tothe early to middle stage of mouse embryonic development.

View larger version (72K):
[in this window]
[in a new window]
FIG. 1. Duplin expression during mouse development. (A) Immunoblot analysis of Duplin expression in wild-type ES cells (equivalent to E3.5), embryos (E8.5, E12.5, and E16.5), and newborn animals. The expression of ${alpha}$ -tubulin was examined as a loading control. (B) Whole-mount in situ hybridization of wild-type embryos at E7.75, E8.5, and E10.5 with a Duplin antisense riboprobe. The result of a control experiment with a sense riboprobe is shown for an E7.75 embryo. Scale bars, 100 µm.

Mouse Duplin is located on chromosome 14, spans $~$ 13 kb, and consistsof nine exons (Fig. 2A). The binding site for ß-cateninis encoded by exons 5 to 9. Mouse Duplin cDNA contains an openreading frame of 2,256 bp and shares 96 and 91% sequence identitywith the corresponding rat and human cDNAs, respectively. Toelucidate the function of Duplin during mouse development, wegenerated mice deficient in this protein by gene targeting.The targeting construct for the disruption of mouse Duplin wasdesigned to delete all exons of the gene (Fig. 2A). ES cellswere transfected with the linearized targeting vector, and recombinantclones were selected and injected into C57BL/6 blastocysts.Chimeric males that transmitted the mutant allele to the germline were mated with C57BL/6 females, and the resulting heterozygoteswere intercrossed to yield homozygous mutant mice. However,Southern blot analysis of tail DNA from 3-week-old mice revealedthe absence of animals homozygous for the mutation (Fig. 2B).To date, no homozygous mutants have been detected among 88 newbornanimals from heterozygote crosses, whereas heterozygous offspringappeared normal and fertile. To determine the time at whichthe Duplin mutation becomes lethal, we examined embryos fromDuplin^+/– intercrosses at various developmental stages.Most Duplin^–/– embryos had been resorbed by E9.5;they were still recoverable at E8.5 (Table 1), however, althoughtheir growth appeared arrested (see Fig. 3).

View this table:
[in this window]
[in a new window]
TABLE 1. Genotype of embryos from Duplin heterozygous intercrosses^a

View larger version (179K):
[in this window]
[in a new window]
FIG. 3. Histopathology of Duplin^–/– embryos. The development of Duplin^+/+ (A, D, G, and J) and Duplin^–/– (B, E, H, and K) embryos is shown at E5.5, E6.5, E7.5, and E8.5, respectively. Higher-magnification views of the Duplin^–/– embryos are also shown (C, F, I, and L). Arrowheads in panel I indicate apoptotic cells with condensed nuclei. Abbreviations: ec, ectoderm; me, mesoderm; pen, parietal endoderm; ven, visceral endoderm; xec, extraembryonic ectoderm. Scale bars, 100 µm.

In contrast to wild-type embryos (Fig. 3A, D, G, and J), thegrowth of Duplin^–/– embryos appeared retarded asearly as E5.5 (Fig. 3B and C) and almost arrested after E6.5(Fig. 3E, F, H, I, K, and L), although the wild type and thehomozygous mutant appeared to be indistinguishable at the blastocyststage (E3.5) (data not shown). Duplin^–/– embryosdeveloped into an egg cylinder consisting of two layers of tissue,ectoderm and visceral endoderm, at E7.5 (Fig. 3H and I), butformation of the primitive streak and mesoderm was not observed.Cells with condensed nuclei (Fig. 3I) and cells positive forTUNEL staining (Fig. 4D) were apparent in the ectoderm-likelayer of Duplin^–/– embryos at E7.5. Wild-type embryosat this stage contained few TUNEL-positive cells (Fig. 4C).Furthermore, we performed TUNEL assay in the wild type and inhomozygous mutants at E6.0 when the first morphological changescan be detected (Fig. 4A and B). However, we failed to detectTUNEL-positive cells at this stage in the wild-type and mutantembryos (Fig. 4A and B), suggesting that the pronounced apoptosisapparent in Duplin^–/– E7.5 embryos is a secondaryeffect of the developmental arrest.

View larger version (157K):
[in this window]
[in a new window]
FIG. 4. Pronounced apoptosis in early Duplin^–/– embryos. Sections of Duplin^+/+ (A and C) and Duplin^–/– (B and D) embryos at E6.0 (A and B) and E7.5 (C and D) were subjected to the TUNEL assay. Scale bars, 100 µm.

Immunohistochemical analysis revealed that ß-cateninwas expressed predominantly at the intercellular adhesive surfacein tissues of both wild-type and mutant embryos at E7.5 (Fig.5A to C). We also examined the expression of the products oftarget genes of Wnt-ß-catenin signaling, includingT (brachyury), Axin2 (also known as Conductin or Axil), andD-type cyclins (D1, D2, and D3), by immunohistochemistry atE7.5. Brachyury is a transcription factor that contains a Tbox and regulates the balance between mesodermal and neuralcell fates in the primitive streak during gastrulation and anteroposterioraxis development (48). Brachyury was not detected in Duplin^–/–embryos, whereas it was expressed in the primitive streak ofwild-type embryos (Fig. 5D to F). Axin2 is a direct target anda major negative regulator of Wnt signaling pathway, which promotesthe phosphorylation and degradation of ß-catenin (4,21, 29, 49). Axin2 was expressed in ectodermal and mesodermalcells in the wild-type embryos, whereas it was expressed inthe ectoderm-like layer of Duplin^–/– embryos (Fig.5G to I). In wild-type embryos, cyclin D1 was expressed predominantlyin ectodermal and mesodermal cells (Fig. 5J), whereas the expressionof cyclin D2 was restricted to the mesoderm (Fig. 5M) and cyclinD3 expression was not detected in embryonic tissue but was apparentin the ectoplacental cone and extraembryonic endoderm (Fig.5P). In contrast, none of the three D-type cyclins was detectedin embryonic tissue of Duplin^–/– mutants (Fig. 5K, L, N, O, Q, and R),although a low level of cyclin D3 expressionwas apparent in the ectoplacental cone and extraembryonic endoderm.The expression of downstream effectors of Wnt-ß-cateninsignaling was thus not increased, as might have been expected,in the absence of the inhibitor Duplin.

View larger version (116K):
[in this window]
[in a new window]
FIG. 5. Reduced expression of the products of ß-catenin target genes in Duplin^–/– embryos. Sections of Duplin^+/+ (A, D, G, J, M, and P) and Duplin^–/– (B, E, H, K, N, and Q) embryos at E7.5 were subjected to immunohistochemistry with antibodies to ß-catenin (A and B), to brachyury (D and E), to Axin2 (G and H), to cyclin D1 (J and K), to cyclin D2 (M and N), or to cyclin D3 (P and Q). Higher-magnification views of the Duplin^–/– embryo sections are also shown (C, F, I, L, O, and R). Scale bars, 100 µm.

To assess directly the growth capability of Duplin^–/–embryos, we collected E3.5 blastocysts from heterozygote intercrossesand cultured them individually in vitro for 6 days, during whichtime they formed outgrowths. Both Duplin^+/+ and Duplin^–/–blastocysts hatched, attached to the culture dish, and producedapparently normal trophoblast giant cells (Fig. 6), suggestingthat Duplin^–/– blastocysts can implant in the uterus.However, the inner cell mass of Duplin^–/– embryosdeveloped poorly and formed colonies incompletely after 4 to5 days of culture (equivalent to E7.5 to E8.5) (Fig. 6F and H);it had collapsed and disappeared after 6 days of culture(equivalent to E9.5) (Fig. 6J), whereas the inner cell massof Duplin^+/+ blastocysts appeared to grow normally. This developmentalphenotype of Duplin^–/– embryos in vitro is consistentwith that of mutant postimplantation embryos in vivo.

View larger version (72K):
[in this window]
[in a new window]
FIG. 6. Defective outgrowth of Duplin^–/– blastocysts in vitro. (A to J) Blastocysts from heterozygote intercrosses were cultured for 6 days. The development of Duplin^+/+ (A, C, E, G, and I) and Duplin^–/– (B, D, F, H, and J) blastocysts was examined at E5.5 (A and B), E6.5 (C and D), E7.5 (E and F), E8.5 (G and H), and E9.5 (I and J). (K to N) Outgrown Duplin^+/+ (K and M) and Duplin^–/– (L and N) blastocysts at E9.5 were subjected to immunofluorescence analysis with antibodiesto ß-catenin (K and L) or to cyclin D1 (M and N). Abbreviations: icm, inner cell mass; tr, trophoectoderm; zp, zona pellucida. Scale bars, 100 µm.

Immunostaining for ß-catenin showed that this proteinwas localized to the perinuclear region and cytoplasm of trophoectodermalepithelial cells in both Duplin^–/– and wild-typeembryos after 6 days in culture (Fig. 6K and L). Furthermore,immunostaining for cyclin D1 revealed that this protein waslargely restricted to the nucleus of trophoectodermal epithelialcells of both wild-type and Duplin^–/– embryos atthis time (Fig. 6M and N). These results also support the notionthat the lack of Duplin does not simply result in abnormal activationof the Wnt signaling pathway.

DISCUSSION

Top

Discussion

Although several proteins that participate in the Wnt signalingpathway have been identified, the precise roles of these moleculesin mammalian embryogenesis have remained unclear. Biochemicaland genetic studies in Xenopus have shown that Duplin repressesß-catenin-mediated transactivation and inhibits theWnt signaling pathway. We thus expected that Wnt-ß-cateninsignaling would be constitutively activated in mice lackingDuplin, resulting in a phenotype characterized by developmentalabnormality or carcinogenesis. Duplin^–/– mice indeeddied in utero at E7.5, suggesting that Duplin is essential forgastrulation during mouse development. In contrast to our expectations,however, Wnt signaling was not activated in the Duplin-deficientembryos, and these animals manifested massive apoptosis at death.Our data suggest two possibilities: (i) Duplin might not contributeto the Wnt signaling pathway but rather to some other functionsor (ii) Duplin might function in both the Wnt signaling pathwayand other ways, but the lack of the latter function resultsin an earlier or more severe phenotype than that of the former.Generation of conditional gene-targeted mice should provideinsight into this question.

Vertebrate gastrulation begins with the formation of the primitivestreak, which provides the first definition of the anteroposterioraxis. Subsequently, some cells intercalate through the streakto emerge as a layer of mesoderm. Wnt signaling is importantfor definition of the anteroposterior axis. Inhibition of posteriorlylocalized Wnt signaling by anteriorly localized Wnt antagoniststhus induces the formation of anterior structures, includingthe forebrain and heart, from neural ectoderm and mesoderm (47).Mice that lack ß-catenin manifest a defect in anteroposterioraxis formation at E5.5, do not form mesoderm or head structures,and show pronounced apoptosis at the time of death, which occursat approximately E7.5 (17). These characteristics indicate thatß-catenin is essential for formation of the anteroposterioraxis. Of the various Wnt genes, Wnt3 is expressed earliest duringdevelopment, its expression being apparent immediately beforegastrulation (E6.25) in the proximal epiblast of the egg cylinder(28). Wnt3^–/– mice develop a normal egg cylinderbut do not form a primitive streak, mesoderm, or node. BetweenE6.5 and E8.5, the Wnt3^–/– embryos continue to growas an egg cylinder consisting of two layers of tissue: ectodermand visceral endoderm. The developmental abnormalities of ß-catenin-deficientembryos are apparent earlier and are more severe than thoseof the Wnt3^–/– mutant, suggesting that another unknownWnt protein might function earlier than does Wnt3 to establishthe anteroposterior axis in mice. Alternatively, the ß-cateninsignaling pathway may be activated in mice by a process thatdoes not require Wnt ligands, as is thought to be the case inamphibians (7). The phenotype of Duplin^–/– mice,which do not form a primitive streak or mesoderm and arrestdevelopment at E7.5, is similar to that of ß-catenin-deficientembryos, suggesting that Duplin might regulate ß-cateninsignaling in a Wnt-independent manner. Although ß-catenin-nullembryos show a defect in anteroposterior axis formation at E5.5,as visualized by marker gene expression, they are morphologicallyindistinguishable from wild-type embryos at this time (17).In contrast, Duplin^–/– embryos already display adistinct morphology at E5.5, indicating that Duplin plays anindispensable role at a developmental stage earlier than thatat which Wnt-ß-catenin signaling regulates anteroposterioraxis formation.

Many target genes of the Wnt-ß-catenin signaling pathway,including those for T (brachyury) (1, 48), Axin2 (also knownas Conductin or Axil) (4, 21, 29, 49), cyclin D1 (36, 42), c-Myc(15), c-Jun, Fra-1 (30), and PPAR ${delta}$ (14), have been identifiedin mammals. In Duplin^–/– mice, we expected thatWnt-ß-catenin signaling would be constitutively activated,resulting in increased expression of these target genes. Previousanalysis revealed that overexpression of Duplin did not affectthe abundance or subcellular distribution of ß-catenin(34). We also did not detect any obvious difference in the amountor subcellular localization of ß-catenin between Duplin^–/–and wild-type embryos, indicating that Duplin does not affectthe stability or transport of ß-catenin. With regardto ß-catenin-mediated transactivation, expressionof brachyury, Axin2, or D-type cyclins was not increased inDuplin^–/– embryos at E7.5, suggesting that Duplindoes not function as a negative regulator of Wnt-ß-cateninsignaling. The expression of brachyury, a mesodermal marker,and mesoderm formation were not observed in the mutant embryos,although it remains unclear whether these abnormalities arethe cause or the result of embryonic death at this stage ofdevelopment.

Axin2 and its ortholog Axin are negative regulators of the Wntsignaling pathway. During mouse embryogenesis, Axin is expressedubiquitously, but Axin2 is expressed in a restricted patternthat overlaps with sites of Wnt signaling (21). In wild-typeembryos, Axin2 was restricted to the ectoderm and the mesoderm,whereas it was expressed in the ectoderm-like layer of Duplin^–/–embryos. Given that mesoderm formation is not observed in themutant embryos, there is no substantial difference in the expressionpattern between wild-type and mutant embryos. Furthermore, quantitativePCR revealed that the amount of Axin2 mRNA was not increasedin Duplin^–/– embryos (data not shown).

Three D-type cyclins have been identified and are expressedin different tissues in mammals (11, 37, 38). In addition tocyclin D1, cyclin D2 also controls cellular proliferation byacting downstream of Wnt-ß-catenin signaling (23).Whereas the tissue-specific expression of D-type cyclins wasapparent in wild-type embryos at E7.5, none of these proteinswas expressed in Duplin^–/– embryos at this time,suggesting that the cell cycle of individual cells was arrestedand the embryos were already dead at this stage. We also didnot detect the expression of cyclin D by immunohistochemistryfrom E5.5 to E6.5 in the mutant embryos (data not shown), suggestingthat the failure to detect cyclin D was not simply the resultof embryonic death.

The earlier onset, as well as the greater severity, of the developmentaldefect of Duplin^–/– embryos compared to that ofß-catenin-deficient mice suggested the possibilitythat Duplin might function to regulate basic cellular activitiessuch as the cell cycle or apoptosis at an early developmentalstage. Given also that Duplin^–/– embryos died atE7.5 manifesting massive apoptosis, we performed TUNEL assayat earlier developmental stage, but we detected few apoptoticcells. These results suggest that apoptosis is induced afterthe stage when the first morphological changes can be detected.The apoptosis observed in the mutant embryos is likely the consequenceof developmental arrest in Duplin^–/– embryos. Withregard to a link between Wnt signaling and apoptosis, crosstalk between Wnt signaling and the p53 signaling pathway hasbeen described (9, 10, 33). Overexpression of ß-cateninthus resulted in accumulation of p53, whereas overexpressionof p53 resulted in downregulation of ß-catenin. Incontrast, WISP-1 (Wnt-1-induced secreted protein), a targetof Wnt-ß-catenin signaling, has been shown to activatethe antiapoptotic signaling pathway mediated by Akt (proteinkinase B), as well as to prevent cells from p53-dependent apoptosisthrough inhibition of the release of cytochrome c from mitochondriaand upregulation of the antiapoptotic protein Bcl-x_L (39). Thephenotype of Duplin-deficient mice is more severe than thatof mice lacking other regulators of apoptosis, such as caspasesor members of the Bcl-2 family of proteins; the lack of theseregulators often affects specific tissues at late stages ofembryogenesis or even in adulthood (32, 44). What function eitheralone or together with defective Wnt signaling is responsiblefor the severe developmental abnormalities of Duplin^–/–mice therefore remains to be determined.

ACKNOWLEDGMENTS

We thank S. Hatakeyama, T. Kamura, M. Shirane, Y. A. Minamishima,M. Yada, T. Hara, and Y. Kotake for valuable discussions andY. Yamada and K. Shimoharada for technical assistance. We alsothank A. Ohta and M. Kimura for help in preparing the manuscript.

This study was supported in part by a grant from the Ministryof Education, Science, Sports, and Culture of Japan and by theYasuda Medical Research Foundation.

FOOTNOTES

* Corresponding author. Mailing address: Department of Molecular and Cellular Biology, Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Fukuoka 812-8582, Japan. Phone: 81-92-642-6815. Fax: 81-92-642-6819. E-mail: nakayak1{at}bioreg.kyushu-u.ac.jp.

REFERENCES

Top