Complementary roles of genes regulated by twopaternally methylated imprinted regions onchromosomes 7 and 12 in mouse placentation

Manabu Kawahara1, Qiong Wu1, Yukio Yaguchi2, Anne C. Ferguson-Smith3

and Tomohiro Kono1,*

1Department of BioScience and 2Electron Microscope Centre, Tokyo University of Agriculture, Setagaya-ku,

Tokyo 156-8502, Japan and 3Department of Physiology, Development and Neuroscience, University of Cambridge,

Downing Street, Cambridge CB2 3DY, UK

Received June 22, 2006; Revised and Accepted August 8, 2006

Imprinted genes have prominent effects on placentation; however, there is limited knowledge about themanner in which the genes controlled by two paternally methylated regions on chromosomes 7 and 12contribute to placentation. In order to clarify the functions of these genes in mouse placentation, we exam-ined transcription levels of the paternally methylated genes, tissue differentiation and development and thecirculatory system in placentae derived from three types of bi-maternal conceptuses that contained genomesof non-growing (ng) and fully grown (fg) oocytes. The genetic backgrounds of the ng oocytes were as fol-lows: one was derived from the wild-type (ngWT) and another from mutant mice carrying a 13 kb deletionin the H19 transcription unit including the germline-derived differentially methylated region (H19-DMR) onchromosome 7 (ngDch7). Another set of oocytes was derived from mutant mice carrying a 4.15 kb deletionin the intergenic germline-derived DMR (IG-DMR) on chromosome 12 (ngDch12). Although placental masswas lower in the ngWT/fg placentae compared with that in the WT placentae, it was recovered in thengDch7/fg placentae, but not in the ngDch12/fg placentae. The ngDch7/fg placental growth improvement wasassociated with severe dysplasia such as an expanded spongiotrophoblast layer and a malformed labyr-inthine zone. In contrast, the ngDch12/fg placentae retained the layer structures with expanded giant cells,but their total masses were smaller with a normal circulatory system in order. Our findings demonstratethat the genes controlled by the two paternally methylated regions, H19-DMR and IG-DMR, complementarilyorganize placentation.

INTRODUCTION

In mammals, imprinted genes, wherein only one of the twoparental chromosome copies is expressed, are regulated byepigenetic modifications, including DNA methylation. Acqui-sition of methylation at key regional controlling elementsoccurs in the parental germlines, during male or femalegametogenesis. Most imprinted genes are regulated by mater-nally derived methylation, whereas only three have beenidentified as genes regulated by paternally derived methyl-ation (1). The latter include H19-Igf2 and Gtl2-Dlk1, whichare regulated by a differentially methylated region (DMR)on chromosomes 7 (H19-DMR) and 12 (IG-DMR),

respectively, and Rasgrf1 on chromosome 9 (2–4).H19-Igf2 and Gtl2-Dlk1 are regulated by paternally derivedmethylation; however, Igf2 and H19 have oppositeexpression patterns as do Dlk1 and Gtl2. Imprinted genesalso have prominent effects on placentation (5). A series ofexperiments have demonstrated that in the mouse placenta,Igf2 and Peg1/Mest appear to be involved in growth enhance-ment, whereas Grb10 and Phlda2 (Ipl, Tssc3) have restrain-ing effects on growth processes. Imprinted genes onchromosome 12 have also been shown to play a role in regulat-ing placental size and organization (6,7). Additionally, Mash2and many genes linked to the X chromosome play criticalroles in placentation (8–14).

# The Author 2006. Published by Oxford University Press. All rights reserved.For Permissions, please email: journals.permissions@oxfordjournals.org

*To whom correspondence should be addressed. Tel/Fax: þ81 354772543; Email: tomohiro@nodai.ac.jp

Human Molecular Genetics, 2006, Vol. 15, No. 19 2869–2879doi:10.1093/hmg/ddl228Advance Access published on August 21, 2006

Normal parthenotes containing two maternal genomes frommatured oocytes showed poorly developed extraembryonictissues, contributing to their death by E9.5 (embryonic day9.5) (15,16). The failure of the placenta in parthenotes isdue to the inappropriate expression of imprinted genes. Thismainly occurs because the maternal epigenotype of theimprinted domain has been imposed on both the two haploidsets resulting in the absence of expression of paternallyexpressed imprinted genes or the overexpression of maternallyexpressed imprinted genes (17–19). However, when non-growing (ng) oocytes of WT mice (ngWT), which haveerased their maternal imprints and thus are considered to benaı̈ve with respect to most of the maternal imprintingprocess, were combined with fully grown (fg) oocytes, weobserved that ngWT/fg bi-maternal conceptuses coulddevelop to E13.5. Such ngWT/fg bi-maternal conceptusesformed placentae that consisted of three layers, namely thetrophoblastic giant cells, spongiotrophoblast and labyrinthinelayers. This suggests that failure to acquire maternal imprintsin the ng oocyte rescues the placenta to some extent. Further,we have shown that ng/fg bi-maternal conceptuses are infre-quently able to develop to term, when the embryos are recon-structed with ng oocytes (ngDch7) of mutant mice carrying a13 kb deletion in the H19 transcription unit with its DMR(20–23). This more successful outcome associated with theprovision of Igf2 indicates that its absence in normal parthe-notes is a contributory factor to their developmental failure.In addition, in these bi-maternal conceptuses, the IG-DMRon chromosome 12 of the ngDch7/fg bi-maternal conceptusesthat developed to term became unexpectedly methylated.Expression studies confirmed that the switch from thematernal to the paternal epigenotype in the imprinteddomain at the distal regions of chromosomes 7 and 12 resultedin the appropriate transcription of Igf2-H19 and Dlk1-Gtl2from the ng allele, which in turn enables further developmentof bi-maternal conceptuses. However, placentation in ng/fgbi-maternal conceptuses has not been investigated to date.

Investigating placentation in ng/fg bi-maternal conceptusesmay help in elucidating the roles that paternally methylatedimprinted genes play in mouse placentation in three ways.First, our system provides the only way of manipulating theexpression of multiple imprinted genes, that can be comparedwith the extremes of full parthenogenesis and androgenesis.Secondly, we can monitor the development of placentae thatexclusively possessed the maternal genomes during gestation;that is, we can investigate the phenotypes of the placentae inwhich all paternally methylated imprinted genes are absentby default. In the case of the ngWT/fg placenta, maternalmethylation imprinting was globally modified, differing fromthe imprinting that is observed in the case of normal parthe-notes because one set of maternal imprints are erased andneither paternal methylation nor a second set of maternalmethylation imprints are imposed. Thirdly, the evaluation ofthe ngDch7/fg placentae enables an understanding of the rela-tive contributions of paternally methylated imprinted geneson chromosome 7 to the development of the placenta. Further-more, in this study, we examined the ngDch12/fg placentae byusing ng oocytes of mice heterozygously carrying a 4.15 kbdeletion of the IG-DMR on chromosome 12 (ngDch12) (4). Itis known that the deletion of the IG-DMR from the maternally

inherited chromosome causes loss of imprinting of all genes inthe 1 Mb cluster that carries the maternally repressed genes,Dlk1, Dio3 and Rtl1, and the maternally expressed non-codingRNAs, including Gtl2, and microRNAs (4). In the absence ofthe IG-DMR on the maternal chromosome, Dlk1, Dio3 andRtl1 are activated and the non-coding RNAs repressed. Asexpected, the ngDch12/fg placentae did not show appropriateexpression of imprinted genes on chromosome 7 and exhibitednormal expression of imprinted genes on chromosome 12. Onthe basis of this, it is be possible to address and compare thecontributions of paternally methylated imprinted genes onchromosomes 7 or 12 to mouse placentation in a ng/fgbi-maternal epigenetic background.

Morphometric and histological analyses revealed that cor-rection of the expression of paternally methylated imprintedgenes on chromosome 7 affected not only the increase in pla-cental mass but also the normal differentiation of giant cells.However, ngDch7/fg bi-maternal conceptuses formed placentaewith severe dysplasia such as an anomalously expanded spon-giotrophoblastic layer and a malformed labyrinthine layer withan anomalous circulatory system. This indicates that thesephenotypes are caused by the absence and/or overexpressionof other imprinted genes in these placentae. Through the cor-rection of the expression of paternally methylated imprintedgenes on chromosome 12, we demonstrated that ngDch12/fgbi-maternal conceptuses could develop into at least18.5-day-old bi-maternal conceptuses, with the placentae com-prising the three layers in order. We further observed that thecirculatory system of the ngDch12/fg placenta was comparableto that of the WT placenta. However, the placental weightbarely increased, and the giant cells were abnormallyexpanded. This suggests that the absence and/or overexpres-sion of chromosome 12 imprinted genes make a major contri-bution to the dysplasia in ngWT/fg placentae, but cannot rescuethe size defect or trophoblast giant cell phenotype that are nolonger evident when normal H19 and Igf2 levels are restored.Thus, we provide the first demonstration that the genes regu-lated by the two paternally imprinted methylated regions onchromosomes 7 and 12 contribute distinct and complementaryfunctions to mouse placentation.

RESULTS

The developmental ability of the ngWT/fg and ngDch7/fgbi-maternal conceptuses had previously been determined;however, the developmental ability of the ngDch12/fgbi-maternal conceptuses remained to be elucidated (17,20).We first examined the ngDch12/fg bi-maternal conceptuses car-rying the ng-oocyte genome that contained a deletion of theIG-DMR. The results clearly showed that the ngDch12/fgbi-maternal conceptuses could develop into at least E18.5(Kawahara et al., manuscript in preparation). Next, we exam-ined the masses of all the four placental types. The ngWT/fgplacentae showed severe growth retardation at E12.5(Fig. 1A and B). Until E18.5, the ngDch7/fg placentae consist-ently showed the highest masses among the placental typesderived from bi-maternal conceptuses, but nonetheless, theywere significantly lower than those of the WT placentae(Fig. 1B). Among the ngDch7/fg, ngDch12/fg and WT placentae,

2870 Human Molecular Genetics, 2006, Vol. 15, No. 19

the ngDch12/fg placentae maintained the lowest weight untilE18.5 (Fig. 1B). To understand the feature of each placentain detail, we carried out the following analyses: quantitativeanalysis of gene expression, in situ hybridization, histologicalanalysis and scanning electron microscopic (SEM) studies toobserve the placental circulatory system.

Gene expression of the paternally methylated imprintedgenes H19, Igf2, Gtl2 and Dlk1 in the placenta

We performed quantitative expression analysis of the H19,Igf2, Gtl2 and Dlk1 genes in individual WT, ngWT/fg,ngDch7/fg and ngDch12/fg placentae by using real-time PCR(Fig. 2A). As expected, the expression of the H19 and Igf2genes at E12.5 was corrected in the ngDch7/fg placenta, butnot in the ngWT/fg and ngDch12/fg placentae. As expected, inboth ngWT/fg and ngDch7/fg placentae, the Gtl2 RNAexpression level was approximately twice that of the meanvalue in the WT placentae, whereas the Dlk1 RNA expressionlevel remained reduced until E18.5. The ngDch12/fg placentaeshowed appropriate expression patterns of the Gtl2 and Dlk1genes, except significantly higher expression of the Gtl2gene at E12.5. Interestingly, the H19 RNA expression levelin the ngDch12/fg placentae was corrected to that in the WTplacentae at E18.5, but not at earlier stages. The Igf2 RNAexpression level was repressed in the ngDch12/fg placentae.Next, we analysed the localization signal of RNA expressionby using in situ hybridization (Fig. 2B). The intensities ofthe expression signals reflected the results of the quantitativeexpression analysis. Although, in previous published studies,the signals of Gtl2 RNA expression had not been detected ingiant cells (24), other studies have observed Gtl2 transcriptionin some though not all giant cells (da Rocha et al., manuscriptin preparation); signals were detected in the giant cells of thengWT/fg and ngDch7/fg placentae. Dlk1 RNA expressionsignals were distinctly recognized in the endothelial cells ofthe fetal blood vessels within the labyrinth of the WT and

ngDch12/fg placentae; however, no expression signals couldbe detected in the ngWT/fg and ngDch7/fg placentae. Thus,we confirmed that in the ngDch7/fg and ngDch12/fg placentae,the expression patterns of H19-Igf2 and Gtl2-Dlk1 RNAswere respectively corrected to approximately normal levelsseen in the WT placentae.

Histological analyses of the placenta

Histological analysis revealed that the ngWT/fg placentaeshowed disproportionate expansion of the spongiotrophoblastlayer along with a distorted and ambiguous boundary betweenthe spongiotrophoblast layer and the labyrinth; they alsoshowed enlarged giant cells (Fig. 3A and B). However, in thengDch7/fg placentae, enlargement of the trophoblast giant cellswas not detected, in contrast to the morphology of the boundarythat was not restored. In contrast, the boundary in the ngDch12/fgplacenta was entirely restored, but enlargement of giant cellsseen in the ngWT/fg placenta was not entirely corrected(Fig. 3A and B). Further, morphometric analysis revealed thatthe labyrinth of the ngWT/fg placentae was reduced in size,and the ratio of the spongiotrophoblast layer to the labyrinthshowed an increase of greater than 4-fold (Fig. 3B). This dis-proportion appeared to be corrected in the ngDch7/fg placentaeat E12.5. However, the ratio of the spongiotrophoblast layerto the labyrinth in the ngDch7/fg placentae was 1.5-fold higherthan that in the WT placentae. The ngDch7/fg placenta consist-ently showed disproportionate expansion of the spongiotropho-blast layer and a reduction in the labyrinthine layer. In contrast,the ngDch12/fg placentae never showed this type of dispropor-tionate expansion of the spongiotrophoblast layer, and theratio of the spongiotrophoblast layer to the labyrinth was iden-tical to that in the WT placenta up to E18.5 (Fig. 3B). To gainfurther insight into the disproportionate expansion of the spon-giotrophoblast layer in the ngWT/fg and ngDch7/fg placentae, weexamined the expression of the Phlda2 gene by real-time PCRand in situ hybridization in the four placental types (Fig. 4A

Figure 1. (A) Four types of placentae, namely the WT, ngWT/fg, ngDch7/fg and ngDch12/fg at E12.5, E15.5 and E18.5. (B) Graphical representation of placentalmasses (n ¼ 5). Values are represented as means+ s.d. (indicated by error bars). (a and b) Values with different superscripts are significantly different within thesame gestational age (P , 0.05).

Human Molecular Genetics, 2006, Vol. 15, No. 19 2871

and B). This gene is located on chromosome 7, is exclusivelyexpressed from the maternal allele and encodes a cytoplasmicprotein with a pleckstrin-homology domain. It controls placen-tal size via a mechanism that is independent of Igf2 signalling,

and its knockout results in global hyperplasia of placental tissueswith disproportionate expansion of the spongiotrophoblast layer(11,25). The expression of the Phlda2 gene was repressed in allplacental types derived from bi-maternal conceptuses, namely

Figure 2. Gene expression of the paternally methylated imprinted genes H19, Igf2, Gtl2 and Dlk1 in placentae. (A) Graphical representation of gene expressionin the WT, ngWT/fg, ngDch7/fg and ngDch12/fg placentae at E12.5, E15.5 and E18.5 (n ¼ 3). Values are represented as means+ s.d. (indicated by error bars).(a–c) Values with different superscripts are significantly different within the same gestational age (P , 0.05). (B) In situ hybridization analysis of the paternallyimprinted genes H19, Igf2, Gtl2 and Dlk1 to E12.5 placentae (Spo, spongiotrophoblastic layer; Lab, labyrinthine layer). Anti-sense probes were hybridized to thecryosections of the WT, ngWT/fg, ngDch7/fg and ngDch12/fg placentae. Note that the signals of Gtl2 mRNA expression are evident in the giant cells of the ngWT/fgand ngDch7/fg placenta, but not of the WT and ngDch12/fg placenta.

2872 Human Molecular Genetics, 2006, Vol. 15, No. 19

Figure 3. Histological and morphometric analyses of the WT, ngWT/fg, ngDch7/fg and ngDch12/fg placentae. (A) Midline placental sections (H&E) at E12.5, E15.5and E18.5. The blue arrowheads indicate the border between the spongiotrophoblast layer and the labyrinth. High-magnification views of giant cells adjacent tothe spongiotrophoblast layer at E12.5 are shown in the bottom row. (B) Spongiotrophoblastic layer (Spo), labyrinthine layer (Lab) and Spo/Lab ratios in the WT,ngWT/fg, ngDch7/fg and ngDch12/fg placentae (n ¼ 4). The average areas of giant cells at E12.5 was also calculated (n ¼ 4). (a and b) Values with different super-scripts are significantly different within the same gestational age (P , 0.05).

Human Molecular Genetics, 2006, Vol. 15, No. 19 2873

ngWT/fg, ngDch7/fg and ngDch12/fg placenta (11, 14 and 22%,respectively). This was in accordance with the results of the insitu hybridization (Fig. 4A and B). Hence, these results indicatedthat the ngWT/fg and ngDch7/fg placentae show disproportionateexpansion of the spongiotrophoblast layer, independently ofthe repression of the Phlda2 gene. However, the ngDch7/fg pla-centae showed normal giant cells. Furthermore, the ngDch12/fgplacentae showed no histological defects except the enlargedgiant cells. These results indicate that the spongiotrophoblastexpansion and its defective interface with the labyrinthine zonein ngWT/fg is rescued by corrected expression of imprintedgenes on chromosome 12 and the trophoblast giant cell expan-sion by corrected Igf2-H19 expression. More precise geneexpression analysis using each placenta tissue might providefurther insight into a various placental abnormalities in thebi-maternal conceptuses.

Aberrant vasculature of the labyrinth of placentaderived from bi-maternal conceptuses

The labyrinth plays a crucial role in the exchange of nutrients,gases and waste between maternal and fetal blood. The numberof blood vessels was counted in the WT and ngDch12/fg placen-tae, and this value was found to increase with gestational age(Fig. 5B). SEM studies of casts additionally revealed that inthe ngDch7/fg placenta, both the maternal sinusoids and thefetal blood vessels within the labyrinth were defective(Fig. 5A, B and D). The ngWT/fg anfd the ngDch7/fg placentashowed irregular dilation of the maternal sinusoids(Fig. 5Bb). At E12.5 and later, on examining the vascularcasts of the maternal sinusoids in the WT placentae, weobserved that the maternal sinusoids become smaller andmore intricate, and the ring-like space at the base of the labyr-inth (embryonic side) becomes enlarged as gestation proceeds(Fig. 5Ba, e and h and C). However, in the ngDch7/fg placentae,the sinusoidal pattern was disordered and the degree of arbori-zation was reduced, and the development of the ring-like spaceof the labyrinth was inadequate (Fig. 5Bc, f and i and C). In con-trast, the maternal sinusoids of the ngDch12/fg placentae showed

none of the serious defects that were shown by the ngDch7/fgplacentae, except that their entire structure was very small(Fig. 5Bd, g and j and C). To visualize the feto-placental circu-lation, we prepared casts of the fetal circulation by injecting thecasting compound into the umbilical cords (Fig. 5D). The umbi-lical cord of the ngDch12/fg placenta was so narrow and thin thatthe casting compound could not be injected into them and weobtained casts only for the WT and ngDch7/fg placentae. Wefound that in the ngDch7/fg placenta, the fetal blood vesselswithin the labyrinth were strikingly disordered, particularlynear the spongiotrophoblast border. Moreover, the bloodvessels were remarkably dilated (centre of Fig. 5D). In themiddle of the labyrinth, we observed that the fetal bloodvessels were abnormally aggregated (right side of Fig. 5D).These results indicate that both the maternal sinusoids andfetal blood vessels in the ngDch7/fg placentae are larger andhave lower density than those in the WT placentae, whereasthe ngDch12/fg placentae retain the intact vascular architecture.

DISCUSSION

The various analyses reported here of the placental defectscaused by the disturbance of paternally methylated imprintedgenes located on chromosomes 7 and 12 facilitate the under-standing of the manner in which the genes in these regionscontribute to mouse placentation. We found that many pheno-types displayed by the ngDch7/fg and ngDch12/fg placentae wereheterogeneous and complementary. This suggests that the twoimprinted regions complementarily contribute to mouse pla-centation and the respective regions have multiple functionsin mouse placentation (Fig. 6). Appropriate expression ofthe two imprinted genes on chromosome 7 results in some cor-rections to the ngWT/fg placental phenotypes, namely someincrease in placental size and normalization of the expandedgiant cells arising in the ngWT/fg placentae. Furthermore,when the imprinted genes on chromosome 12 were appropri-ately expressed in the ngDch12/fg placentae, these placentaehad the three layers in order and a labyrinthine zone withintricate vasculature.

Figure 4. Investigation of Phlda2 RNA expression in the WT, ngWT/fg, ngDch7/fg and ngDch12/fg placentae. Phlda2 RNA transcription was investigated by(A) real-time PCR and (B) in situ hybridization (Spo, spongiotrophoblastic layer; Lab, labyrinthine layer) (n ¼ 3) at E12.5. Values are represented asmeans+ s.d. (indicated by error bars). (a and b) Values with different superscripts are significantly different within the same gestational age (P , 0.05).

2874 Human Molecular Genetics, 2006, Vol. 15, No. 19

Roles of the paternally methylated imprinted geneson chromosome 7

Themasses of the ngDch7/fg placentaewere higher than thoseof thengWT/fg and ngDch12/fg placentae. Previous studies reported thatthe placental phenotype that was associated with the lack of Igf2exhibits a reduced mass (8,26,27). On the basis of this, we

surmise that these corrections are due to the retrieval of Igf2expression in the ngDch7/fg placentae. However, the masses ofthe ngDch7/fg placentae were never equal to those of the WT pla-centae. This might be due to multiple histological defects in thengDch7/fg placentae (Fig. 3). Indeed, further correction of the dis-proportionate growth of both the spongiotrophoblastic layer andlabyrinthine zone by the appropriate transcription of the imprinted

Figure 5. Analyses of placental vasculature. (A) In the morphometric analysis, the number of blood vessels in the WT, ngDch7/fg and ngDch12/fg placentae wasinvestigated by tracing the surrounding area of each maternal and fetal blood vessel within 114 550 mm2 of the labyrinth (n ¼ 4). (a–c) Values with differentsuperscripts are significantly different within the same gestational age (P, 0.05). (B) Structure of the maternal sinusoids of the WT (a, e and h), ngDch7/fg (c, fand i) and ngDch12/fg (d, g and j) placentae at E12.5, E15.5 and E18.5. Structure of the maternal sinusoid of the ngWT/fg placenta at E12.5 is shown in (b). In eachpanel, the upper parts present an overview of the maternal sinusoids (bar ¼ 1 mm) and the lower parts present high-magnification views of the sinusoids in thelabyrinth (bar ¼ 50 mm). Note that the WT and ngDch12/fg placentae show a decrease in the size of the sinusoids with an increase in gestational age. In the ngDch7/fg placenta, all these changes were observed to a lesser extent, and the maternal sinusoids were irregularly dilated, as shown by the yellow arrow heads. ThengWT/fg placenta also showed irregular dilation of the maternal sinusoids. (C) The ring-like space at the base of the labyrinth in the WT, ngDch7/fg and ngDch12/fgplacentae at E12.5 and E18.5. The WT and ngDch12/fg placentae showed an increase in the size of the ring-like space at the base, but the ngDch7/fg placentae didnot. (D) Fetal side casts of the labyrinth of the WT (a) and ngDch7/fg (b) placentae at E15.5. An overview of the fetal blood vessels is shown on the left(bar ¼ 1 mm); the construction of the vasculature up to the spongiotrophoblast border is shown in the centre (bar ¼ 100 mm) and the fetal blood vessels inthe middle of the labyrinth are shown on the right (bar ¼ 100 mm). Note that the organization of blood spaces was disordered; the fetal blood vessels weredilated and aggregated in the ngDch7/fg placentae as shown by the blue or yellow arrowheads.

Human Molecular Genetics, 2006, Vol. 15, No. 19 2875

genesonchromosome12maybenecessary.This could also ensurethat the placental masses of the ng/fg placentae equal those of theWT placentae. Analysis of the placental phenotypes of doublemutant ng/fg oocytes will provide further insight into this.

We additionally found that among the three placental typesof bi-maternal placentae, only the ngDch7/fg placentae hadnormal giant cells. Middleton et al. (28) also reported thepossibility that Igf2 plays an important role in the growthand development of the giant cell tumours of the bone. Ourresults are consistent with the idea that the level of Igf2RNA expression is important for the normal formation of tro-phoblastic giant cells in mouse placentae.

Giant cells contribute to invasiveness and the process ofpregnancy by expressing the placental lactogen 1 and proliferingenes (29,30). In addition, oligonucleotide microarray analysisrevealed that, at E12.5, the expression of both these genes in thengWT/fg placentae was greater than twice that in WT placentae(Kawahara et al., unpublished data). In contrast, in the ngDch7/fgplacentae, the expression of these genes was normal. Takentogether, these results and the present findings indicate thatthe expression of Igf2 and H19 on chromosome 7 contributesto the differentiation of functional giant cells.

Roles of the imprinted genes on chromosome 12

ThengDch7/fgplacentae showedabnormal expression levelsof thepaternally methylated imprinted genes on chromosome 12; theseplacentae evidently had many gross abnormalities. These anom-alous phenotypeswere entirely restored in thengDch12/fgplacenta.The ngDch12/fg placenta had a restored boundary between thespongiotrophoblast and labyrinthine zone and the ratio of thespongiotrophoblast layer to the labyrinth totally correlated withthat of the WT placentae. The trophoblastic giant cells of thengDch12/fg placentae, however, remained abnormally expanded.

Consequently, these findings suggest that the imprinted genesondistalmouse chromosome 12 regulate a balance between spon-giotrophoblastic growth and labyrinthine growth.

Dlk1 is a paternally expressed, protein-coding gene onchromosome 12; this gene encodes a transmembrane proteincontaining epidermal growth factor repeats and is a memberof the Notch/Delta/Serrate family of developmental signallingmolecules. This gene is involved in several differentiation pro-cesses and is expressed in the fetal endothelial cells of themurine placenta; however, its precise function in the placentais yet to be determined (5,31–35). We examined the subcellularlocalization of Dlk1 RNA in the mouse placenta. The Dlk1RNA was specifically transcribed in the endothelial cells ofthe fetal blood vessels (Fig. 2B). Therefore, the disruptedfetal blood vessels observed in the ngDch7/fg placenta mightbe due to the absence of Dlk1 RNA. In fact, we observed thatthe ngDch12/fg placental vasculature was comparable to theWT placental vasculature. However, we cannot rule out thepossibility that other genes located on chromosome 12, suchas Rtl1 or Dio3, might be involved in placental vasculogenesis.Although placental defects in Dlk1 knockout animals were notdescribed, conceptuses with paternal uniparental disomy forchromosome 12 have defects in the fetal capillary componentof the labyrinthine zone (6,36). It is known that Rtl1 is aretrotransposon-like gene with an open-reading frame ofunknown function and that Dio3 is a negative regulator ofthyroid hormone metabolism; however, the particular roles ofboth these genes in placentation has not been reported thus far.

Low level of Phlda2 RNA expression in both thengDch7/fg and ngDch12/fg placentae

To understand the reason behind the anomalous expansionof the spongiotrophoblastic layer in the ngWT/fg and ngDch7/fg

Figure 6. Summary of the multiple roles played by the paternally methylated imprinted genes on chromosomes 7 and 12 in mouse placentation. Illustrationspresent the WT placenta (top), the ngWT/fg placenta (bottom), ngDch7/fg placenta (left) and ngDch12/fg placenta (right). The roles of the imprinted genes thatwere elucidated in the present study are listed in the table on the immediate right.

2876 Human Molecular Genetics, 2006, Vol. 15, No. 19

placentae, we investigated the expression pattern of thePhlda2 gene in the four placental types (Fig. 4). It is knownthat overgrowth along with the expansion of the spongiotro-phoblast layer is detected in the Phlda2-null placentae, inde-pendent of Igf2 signalling (11). However, all the placentaltypes exhibited low Phlda2 activity, including the ngDch12/fgplacentae, which did not show the expansion of the spongio-trophoblastic layer.

Phlda2 is a maternally expressed imprinted gene regulatedby the Kvdmr1 located in the 10th intron of the Kcnq1gene; we have confirmed that this methylation begins postna-tally in oocytes that have attained a diameter greater than40 mm (19). Hence, we had expected the level of Phlda2 tran-scription in the all placental types to be appropriately modifiedbecause the oocytes with a diameter less than 20 mm wereselected as ng oocytes for nuclear transfer. We have confirmedthe mono-allelic expression of the Phlda2 gene from thefg-oocyte genome in the ngWT/fg placentae at E12.5 (Ogawaet al., submitted for publication). Therefore, the disproportion-ate expansion of the spongiotrophoblastic layer could not beexplained based on the low level of Phlda2 transcription.The reasons for the repression of the Phlda2 gene in ng/fgbi-maternal placentae remain to be elucidated.

From the present study, we could derive the following con-clusions (Fig. 6). We confirmed that the deletion of theIG-DMR on chromosome 12 which causes paternalization ofthe maternal chromosome alone could restore some imbalanceimposed by two maternal genomes and facilitate the develop-ment of bi-maternal conceptuses to at least E18.5. A combi-nation of this new system and our previous bi-maternalconceptuses production system facilitated a detailed under-standing of the contribution of the paternally methylatedimprinted genes on chromosomes 7 and 12 towards mouse pla-centation. The present study provides evidence that imprintedgenes transcribed from both the regions, H19-Igf2 andGtl2-Dlk1, complementarily contribute to mouse placentation.It is probable that the appropriate expression of theseimprinted genes would enable ng/fg bi-maternal conceptusesto undergo definitive placentation.

MATERIALS AND METHODS

Production of ng/fg bi-maternal conceptuses

Fg germinal vesicle (GV) oocytes were collected into M2medium from the ovarian follicles of B6D2F1 (C57BL/6N�DBA) female mice, 44–48 h after they were injectedwith equine chorionic gonadotrophin (37). Ovulated MIIoocytes were also collected from superovulated B6D2F1mice, 16 h after they were injected with human chronic gon-adotrophin. We collected ng oocytes that were in the diplotenestage of the first meiosis from the ovaries of 1-day-oldnewborn mice. Serial nuclear transfer was then performedusing a previously described method (17,18,22). Ng oocytesderived from WT B6D2F1 (ngWT) females, H19D13 nullmutants (ngDch7) or IG-DMRD4.5 heterozygous mutantswere reconstructed with enucleated GV oocytes. After fusionwith inactivated Sendai virus, the reconstructed oocyteswere cultured for 14 h in a-MEM medium (GIBCO, GrandIsland, NY, USA). A spindle from the reconstructed oocytes

was again transferred into ovulated MII oocytes, followedby treatment with 10 mM SrCl2 in Ca

2þ-free M16 mediumfor 2 h. These embryos were cultured in M16 medium for3.5 days in an atmosphere of 5% CO2, 5% O2 and 90% N2at 378C (38). The embryos that developed to the blastocyststage were transferred into the uterine horns of recipientfemale mice at 2.5 days of pseudopregnancy. The placentaewere recovered from the pregnant mice at E12.5, E15.5 andE18.5 and used in the subsequent analyses. In order to dis-tinguish the ngDch12/fg bi-maternal conceptuses from thengWT/fg bi-maternal conceptuses, we confirmed the deletionin IG-DMR by genotyping yolk sacs isolated from all recov-ered bi-maternal conceptuses.

Quantitative gene expression analysis

Total RNA was extracted using an RNAeasy Mini Kit(QIAGEN K.K., Tokyo, Japan) from the four types of wholeplacentae: WT placentae derived from fertilized embryos(B6D2F1�C57BL/6N), ngWT/fg, ngDch7/fg and ngDch12/fg.The cDNAs were then synthesized using the SuperScript

TM

.II RnaseH reverse transcriptase kit (Invitrogen, Carlsbad,CA, USA) in a reaction solution (20 ml) containing the totalRNA (1 mg) prepared from each placenta. Finally, we per-formed a quantitative analysis of the gene expression byusing real-time PCR (LightCyclerTM System, Roche Molecu-lar Biochemicals, Mannheim, Germany) after preparing areaction mixture (LightCycler FirstStart DNA Master SYBRGreen I, Roche Molecular Biochemicals). The primers usedfor the analysis were as described in Wu et al. (22). Theprimers for Phlda2 gene expression analysis were used asfollows, sense: 50-CTT CGA AAA CCG TGA AGA CC-30

and anti-sense primer: 50-CCT TGT AAT AGT TGG TGACGA TG-30.

mRNA in situ hybridization

In order to prepare cryosections, we collected and fixed thefour types of placentae at E12.5 with 4% paraformaldehyde,incubated them at 48C overnight in 20% sucrose in PBS andsubsequently embedded them in OCT compound for 10 minat room temperature. This was followed by freezing at2808C until use. The sections were allowed to dry and wereimmediately processed for RNA in situ hybridization afterslicing. For in situ hybridization, digoxigenin (DIG)-labelledRNA probes were prepared according to manufacturer’sinstructions by using a DIG RNA labelling kit (Roche Diag-nostics GmbH, Mannheim, Germany). To prepare RNAprobes for each gene, a 27emsp14;kb mouse H19 cDNAclone (39) was used. For other probes, the following geneswere amplified: Igf2 1346 bp (sense primer 50-GCT GACCTC ATT TCC CGA TAC-30 and anti-sense primer 50-AAAATT TTG GGT CCC CTT CCT-30), Gtl2 (sense primer50-AAC CCA CTA CCA TAC AGA GGA-30 and anti-senseprimer 50-CGA GAG AAT GGT TGA GAC ACA-30) andDlk1 (sense primer 50-CCTCTTGC TCCTGCTGGCTTTC-30

and anti-sense primer 50-GAT GTG TTG CTC GGG CTGCTG A-30). Following amplification, these genes were insertedinto a pGEMw-T Easy Vector (Promega). The Phlda2 probe

Human Molecular Genetics, 2006, Vol. 15, No. 19 2877

was prepared as described previously (25). None of thesesense control probes produced significant signals.

Histological and morphometric analyses

In order to measure the areas of the labyrinthine and spongio-trophoblast layers, we captured digitized images of the midlineparaffin sections stained with haematoxylin/eosin (H&E). Pla-cental entire images were saved as high-quality tif-files andanalysed by using the MetaMorph software (UniversalImaging Co., Downingtown, PA, USA). The total number ofpixels in each layer was calculated by using the ‘measure-ment’ tool of the MetaMorph software. Additionally, we ana-lysed the average areas of giant cells at E12.5 and the numberof blood vessels within the labyrinth by using PALM RoboSoftware 2.2-0103 (PALM Microlaser Technologies, AG).We calculated the number of blood vessels by tracing thearea surrounding each blood vessel, that is, within114 550 mm2 of the labyrinth. In the case of giant cells,average areas of five giant cells per H&E section from eachplacental type were calculated.

Scanning electron microscopy of vascularcorrosion casts

Vascular corrosion casts of the placentae were prepared basedon methods described previously (29,40). To prepare casts ofthe maternal vasculature, pregnant mice at E12.5, E15.5 andE18.5 were anaesthetized and the left ventricle of thebeating heart was injected with a heparinized saline solution.An incision that served as an exit point for perfusion wasmade in the right atrium. Subsequently, the placentae wereperfused with Mercox solution (Okenshoji, Tokyo, Japan), acasting compound, by infusion via the left ventricle.

For preparing fetal side casts at E15.5, the pregnant micewere sacrificed by cervical dislocation, and the uterus wasexcised and immersed in PBS. An implantation site wasexcised from the uterus and placed in a Petri dish under astereomicroscope. The uterus was cut to expose the yolksac, and the embryo and placenta were then exposed. Theembryo and placenta were transferred onto a dried paper,and a 32G needle was introduced into the umbilical artery;this was followed by injection with heparinized saline sol-ution, followed by perfusion with Mercox solution. Anincision that served as an exit point for perfusion was madein the umbilical vein.

For complete hardening, each cast was polymerized in hotwater at 608C for 2 hours, corroded in 20% KOH, washedovernight in water and air-dried. The air-dried samples werefrozen, cracked with a cooled razor blade to observe theinternal structure of the labyrinth and then sputter-coatedwith platinum. Using a SEM (Hitachi S-4000, Tokyo) at lowvoltage, we examined five to 12 casts of the maternal vascula-ture and three to five casts of the fetal vasculature obtainedfrom several pregnant mice.

Statistical analyses

Statistical analyses of all data for comparison were carried outusing analysis of one-way analysis of variance and Fisher’s

PLSD test by using the statistical analysis software Statview(Abacus Concepts, Inc., Berkeley, CA, USA). A P-value of,0.05 was considered significant. When depicting statisticalsignificance in the figures, the use of a, b and c indicatesthat within a gestational stage, a is significantly differentfrom b and c and b is significantly different from c. A valuemarked as ab is not significantly different from the valuesmarked a or b.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online.

ACKNOWLEDGEMENTS

We thank Dr Shirley Tilghman, Princeton University, for pro-viding the H19D13 mutant mice and Dr Tom Moore, Univer-sity College Cork, for the discussions. We would also like tothank Dr Eimei Sato, Tohoku University, Dr Yosuke Shiraishiof Shinjuku Acupuncture, Moxibustion and Judo TherapySpecial School, Dr Miya Kobayashi, Kinjo Gakuin Universityand Dr Kazuyo Suzuki, Nagoya University, for their advice oncorrosion casts and morphometry. This study was supportedby a grant from the Bio-oriented Technology ResearchAdvancement Institution (BRAIN), Japan.

Conflict of Interest statement. No conflicts declared.

REFERENCES

1. Kobayashi, H., Suda, C., Abe, T., Kohara, Y., Ikemura, T. andSasaki, H. (2006) Bisulfite sequencing and dinucleotide content analysisof 15 imprinted mouse differentially methylated regions (DMRs):paternally methylated DMRs contain less CpGs than maternallymethylated DMRs. Cytogenet. Genome Res., 113, 130–137.

2. Tremblay, K.D., Saam, J.R., Ingram, R.S., Tilghman, S.M. andBartolomei, M.S. (1995) A paternal-specific methylation imprint marksthe alleles of the mouse H19 gene. Nat. Genet., 9, 407–413.

3. Pearsall, R.S., Plass, C., Romano, M.A., Garrick, M.D., Shibata, H.,Hayashizaki, Y. and Held, W.A. (1999) A direct repeat sequence at theRasgrf1 locus and imprinted expression. Genomics, 55, 194–201.

4. Lin, S.P., Youngson, N., Takada, S., Seitz, H., Reik, W., Paulsen, M.,Cavaille, J. and Ferguson-Smith, A.C. (2003) Asymmetric regulation ofimprinting on the maternal and paternal chromosomes at the Dlk1-Gtl2imprinted cluster on mouse chromosome 12. Nat. Genet., 35, 97–102.

5. Coan, P.M., Burton, G.J. and Ferguson-Smith, A.C. (2005) Imprintedgenes in the placenta—a review. Placenta, 26 (Suppl. A), S10–S20.

6. Georgiades, P., Watkins, M., Surani, M.A. and Ferguson-Smith, A.C.(2000) Parental origin-specific developmental defects in mice withuniparental disomy for chromosome 12. Development, 127, 4719–4728.

7. Georgiades, P., Watkins, M., Burton, G.J. and Ferguson-Smith, A.C.(2001) Roles for genomic imprinting and the zygotic genome in placentaldevelopment. Proc. Natl Acad. Sci. USA, 98, 4522–4527.

8. Lopez, M.F., Dikkes, P., Zurakowski, D. and Villa-Komaroff, L. (1996)Insulin-like growth factor II affects the appearance and glycogen contentof glycogen cells in the murine placenta. Endocrinology, 137, 2100–2108.

9. Lefebvre, L., Viville, S., Barton, S.C., Ishino, F., Keverne, E.B. andSurani, M.A. (1998) Abnormal maternal behaviour and growth retardationassociated with loss of the imprinted gene Mest. Nat. Genet., 20,163–169.

10. Charalambous, M., Smith, F.M., Bennett, W.R., Crew, T.E.,Mackenzie, F. and Ward, A. (2003) Disruption of the imprinted Grb10gene leads to disproportionate overgrowth by an Igf2-independentmechanism. Proc. Natl Acad. Sci. USA, 100, 8292–8297.

2878 Human Molecular Genetics, 2006, Vol. 15, No. 19

11. Frank, D., Fortino, W., Clark, L., Musalo, R., Wang, W., Saxena, A.,Li, C.M., Reik, W., Ludwig, T. and Tycko, B. (2002) Placentalovergrowth in mice lacking the imprinted gene Ipl. Proc. Natl. Acad. Sci.USA, 99, 7490–7495.

12. Rossant, J., Guillemot, F., Tanaka, M., Latham, K., Gertenstein, M. andNagy, A. (1998) Mash2 is expressed in oogenesis and preimplantationdevelopment but is not required for blastocyst formation. Mech. Dev., 73,183–191.

13. Rodriguez, T.A., Sparrow, D.B., Scott, A.N., Withington, S.L., Preis, J.I.,Michalicek, J., Clements, M., Tsang, T.E., Shioda, T., Beddington, R.S.et al. (2004) Cited1 is required in trophoblasts for placental developmentand for embryo growth and survival. Mol. Cell. Biol., 24, 228–244.

14. Rossant, J. and Cross, J.C. (2001) Placental development: lessons frommouse mutants. Nat. Rev. Genet., 2, 538–548.

15. Surani, M.A., Barton, S.C. and Norris, M.L. (1984) Development ofreconstituted mouse eggs suggests imprinting of the genome duringgametogenesis. Nature, 308, 548–550.

16. Surani, M.A. and Barton, S.C. (1983) Development of gynogenetic eggsin the mouse: implications for parthenogenetic embryos. Science, 222,1034–1036.

17. Kono, T., Obata, Y., Yoshimzu, T., Nakahara, T. and Carroll, J. (1996)Epigenetic modifications during oocyte growth correlates with extendedparthenogenetic development in the mouse. Nat. Genet., 13, 91–94.

18. Obata, Y., Kaneko-Ishino, T., Koide, T., Takai, Y., Ueda, T., Domeki, I.,Shiroishi, T., Ishino, F. and Kono, T. (1998) Disruption of primaryimprinting during oocyte growth leads to the modified expression ofimprinted genes during embryogenesis. Development, 125, 1553–1560.

19. Hiura, H., Obata, Y., Komiyama, J., Shirai, M. and Kono, T. (2006)Oocyte growth-dependent progression of maternal imprinting in mice.Genes Cells, 11, 353–361.

20. Kono, T., Obata, Y., Wu, Q., Niwa, K., Ono, Y., Yamamoto, Y.,Park, E.S., Seo, J.S. and Ogawa, H. (2004) Birth of parthenogenetic micethat can develop to adulthood. Nature, 428, 860–864.

21. Moore, T. and Ball, M. (2004) Kaguya, the first parthenogeneticmammal—engineering triumph or lottery winner? Reproduction, 128,1–3.

22. Wu, Q., Kumagai, T., Kawahara, M., Ogawa, H., Hiura, H., Obata, Y.,Takano, R. and Kono, T. (2006) Regulated expression of two sets ofpaternally imprinted genes is necessary for mouse parthenogeneticdevelopment to term. Reproduction, 131, 481–488.

23. Leighton, P.A., Ingram, R.S., Eggenschwiler, J., Efstratiadis, A. andTilghman, S.M. (1995) Disruption of imprinting caused by deletion of theH19 gene region in mice. Nature, 375, 34–39.

24. Tanaka, S., Oda, M., Toyoshima, Y., Wakayama, T., Tanaka, M.,Yoshida, N., Hattori, N., Ohgane, J., Yanagimachi, R. andShiota, K. (2001) Placentomegaly in cloned mouse concepti caused byexpansion of the spongiotrophoblast layer. Biol. Reprod., 65, 1813–1821.

25. Frank, D., Mendelsohn, C.L., Ciccone, E., Svensson, K., Ohlsson, R. andTycko, B. (1999) A novel pleckstrin homology-related gene familydefined by Ipl/Tssc3, TDAG51 and Tih1: tissue-specific expression,chromosomal location, and parental imprinting. Mamm. Genome, 10,1150–1159.

26. DeChiara, T.M., Efstratiadis, A. and Robertson, E.J. (1990) Agrowth-deficiency phenotype in heterozygous mice carrying aninsulin-like growth factor II gene disrupted by targeting. Nature, 345,78–80.

27. Ferguson-Smith, A.C., Cattanach, B.M., Barton, S.C., Beechey, C.V. andSurani, M.A. (1991) Embryological and molecular investigations ofparental imprinting on mouse chromosome 7. Nature, 351, 667–670.

28. Middleton, J., Arnott, N., Walsh, S. and Beresford, J. (1996) Theexpression of mRNA for insulin-like growth factors and their receptor ingiant cell tumors of human bone. Clin. Orthop. Relat. Res., 224–231.

29. Adamson, S.L., Lu, Y., Whiteley, K.J., Holmyard, D., Hemberger, M.,Pfarrer, C. and Cross, J.C. (2002) Interactions between trophoblast cellsand the maternal and fetal circulation in the mouse placenta. Dev. Biol.,250, 358–373.

30. Faria, T.N., Ogren, L., Talamantes, F., Linzer, D.I. and Soares, M.J.(1991) Localization of placental lactogen-I in trophoblast giant cells of themouse placenta. Biol. Reprod., 44, 327–331.

31. Smas, C.M. and Sul, H.S. (1997) Molecular mechanisms of adipocytedifferentiation and inhibitory action of pref-1. Crit. Rev. Eukaryot. GeneExpr., 7, 281–298.

32. Baladron, V., Ruiz-Hidalgo, M.J., Nueda, M.L., Diaz-Guerra, M.J.,Garcia-Ramirez, J.J., Bonvini, E., Gubina, E. and Laborda, J. (2005) dlkacts as a negative regulator of Notch1 activation through interactions withspecific EGF-like repeats. Exp. Cell Res., 303, 343–359.

33. Bauer, S.R., Ruiz-Hidalgo, M.J., Rudikoff, E.K., Goldstein, J. andLaborda, J. (1998) Modulated expression of the epidermal growthfactor-like homeotic protein dlk influences stromal-cell-pre-B-cellinteractions, stromal cell adipogenesis, and pre-B-cell interleukin-7requirements. Mol. Cell. Biol., 18, 5247–5255.

34. Van Limpt, V.A., Chan, A.J., Van Sluis, P.G., Caron, H.N.,Van Noesel, C.J. and Versteeg, R. (2003) High delta-like 1 expressionin a subset of neuroblastoma cell lines corresponds to a differentiatedchromaffin cell type. Int. J. Cancer, 105, 61–69.

35. Yevtodiyenko, A. and Schmidt, J.V. (2006) Dlk1 expression marksdeveloping endothelium and sites of branching morphogenesis in themouse embryo and placenta. Dev. Dyn., 235, 1115–1123.

36. Moon, Y.S., Smas, C.M., Lee, K., Villena, J.A., Kim, K.H., Yun, E.J. andSul, H.S. (2002) Mice lacking paternally expressed Pref-1/Dlk1 displaygrowth retardation and accelerated adiposity. Mol. Cell. Biol., 22,5585–5592.

37. Quinn, P., Barros, C. and Whittingham, D.G. (1982) Preservation ofhamster oocytes to assay the fertilizing capacity of human spermatozoa.J. Reprod. Fertil., 66, 161–168.

38. Whittingham, D.G. (1971) Culture of mouse ova. J. Reprod. Fertil.Suppl., 14, 7–21.

39. Poirier, F., Chan, C.T., Timmons, P.M., Robertson, E.J., Evans, M.J. andRigby, P.W. (1991) The murine H19 gene is activated during embryonicstem cell differentiation in vitro and at the time of implantation in thedeveloping embryo. Development, 113, 1105–1114.

40. Jiang, J.Y., Macchiarelli, G., Tsang, B.K. and Sato, E. (2003) Capillaryangiogenesis and degeneration in bovine ovarian antral follicles.Reproduction, 125, 211–223.

Human Molecular Genetics, 2006, Vol. 15, No. 19 2879