Trophoblast differentiation in vitro: establishment and

REPRODUCTIONRESEARCH

Trophoblast differentiation in vitro: establishment andcharacterisation of a serum-free culture model for murinesecondary trophoblast giant cells

A H K El-Hashash and S J Kimber

School of Biological Sciences, University of Manchester, 3.239 Stopford Building, Oxford Road,Manchester M13 9PT, UK

Correspondence should be addressed to S J Kimber; Email: [email protected]

Abstract

Differentiation of trophoblast giant cells is an early event during the process of murine embryo implantation. However, differ-

entiation of secondary trophoblast giant cells in the rodent is still only partially understood, probably because of the lack of

suitable in vitro models and cell markers. In order to advance our understanding of trophoblast differentiation, suitable in

vitro models and markers are required to study their development. The objectives of this study were to establish and charac-

terise a serum-free in vitro model for murine secondary trophoblast cells. Secondary trophoblast giant cells growing in vitro

and paraffin sections of day 8.5 postcoitum mouse embryos were processed for immunostaining to establish the expression of

potential markers using antibodies to blood group antigens, E-cadherin, a7 integrins and activator protein-g, as well as placen-

tal lactogen-II. Within 3 days in serum-free culture, ectoplacental cone-derived secondary trophoblast cells underwent simul-

taneous induction of both morphological and functional differentiation. Secondary trophoblasts grew in vitro as a monolayer

of cells with giant nuclei and expressed B and Le-b/Le-y blood group antigens, a7 integrins and placental lactogen-II, as well

as activator protein-g. Transcripts for activator protein-g and placental lactogen-II were detected in cultures by RT-PCR and

for placental lactogen-II by in situ hybridisation. At later time-points apoptosis increased. A fibronectin substrate significantly

increased secondary trophoblast cell numbers and surface area of outgrowth. The increase in cells with giant nuclei coincided

with induction of placental lactogen-II expression. A relationship was found between the nuclear area of secondary tropho-

blast cells and expression of placental lactogen-II.

Reproduction (2004) 128 53–71

Introduction

The outermost cells of the mammalian blastocyst, the epi-thelial trophectoderm, are composed of mural cells, awayfrom the inner cell mass and polar cells overlying it. Tro-phoblast giant cells (TGCs) differentiate from both polarand mural trophectoderm but at different times duringpregnancy in the mouse. Shortly after implantation (4.5days postcoitum; p.c.), mural trophectoderm differentiatesinto primary TGCs, which later form an anastomosingnetwork of blood sinuses at the periphery of the embryo.Acting as a part of the yolk sac placenta, TGC-derivedblood sinuses facilitate diffusion of O2 and nutrients fromthe maternal circulation into the embryo.

Secondary TGCs arise from the outermost layer of thepolar trophectoderm-derived ectoplacental cone (EPC),and later form the outermost layer of chorioallantoicplacenta (Copp 1979, Cross et al. 1994, Cross 2000).

The central cells of the EPC differentiate into thespongiotrophoblast that forms the intermediate layer of theplacenta and secondary TGCs (Zybina & Zybina 1996,Cross 2000). Trophoblast differentiation and invasion havebeen modelled in vitro using embryos or explants culturedon substrates of extracellular matrix molecules. In vitrocultured blastocysts attached to fibronectin (FN), lamininor collagen substrates, forming primary TGCs after the tro-phectoderm–trophoblast transition (Armant et al. 1986a,b,Farach et al. 1987, Sutherland et al. 1988). In other studies,isolated day 7.5 EPC explants were cultured on a laminin(Romagnano & Babiarz 1990) or an FN (Sutherland et al.1991, 1993) substrate and used as a model to study sec-ondary TGC adhesion, migration and invasion into theuterus in the mouse. Parast et al. (2001) used a similar cul-ture model to investigate the alterations of cytoskeletonorganisation during the differentiation of secondary TGCs.

q 2004 Society for Reproduction and Fertility DOI: 10.1530/rep.1.00149

ISSN 1470–1626 (paper) 1741–7899 (online) Online version via www.reproduction-online.org

The process of trophoblast attachment and outgrowthin vitro is thought to be analogous to the adhesion andinvasion stages of trophoblast development in utero(Sherman & Atienza-Samols 1978, Enders et al. 1981).During their growth in vitro, EPC explants first extend pro-cesses to attach to the culture substrate. Then they formsecondary TGCs that spread as a monolayer of flattenedcells (Romagnano & Babiarz 1990). TGCs are identified asthose cells that have characteristic large nuclei due toDNA endoreduplication (Zybina & Zybina 1996). More-over, secondary TGCs can be distinguished from primaryin vivo by their larger and irregular-shaped nuclei as wellas by being multinucleated. Primary TGCs are mononu-cleated and have regular and smaller nuclei.

A number of cytoplasmic and cell surface moleculeshave been used as markers to identify TGCs in vitro andin vivo. Secondary TGCs express activator protein-2g (AP-2g) in vitro (Richardson et al. 2001) and in vivo (Sapinet al. 2000, Auman et al. 2002). In the human, AP-2gregulates chorionic gonadotrophin-b (Johnston & Jameson1999) and placental lactogens (Richardson et al. 2000),and has been shown to control trophoblast proliferationand differentiation in the mouse (Werling & Schorle2002). Certain markers are selectively expressed on extra-embryonic cells of the peri-implantation embryo. Forinstance, blood group B antigen is expressed from the latemorula stage as the outer cells of the embryo differentiateas trophectoderm, on blastocyst trophectoderm and afterimplantation by primary and secondary TGCs and parietalendoderm (Brown et al. 1993), while Le-y is expressed bythe trophectoderm at and after implantation, and to a verylimited extent by ectoderm (Fenderson et al. 1986, Kimber1990). These cell surface carbohydrate epitopes may playa role in cell interactions during early development and inthe immune system (Kimber 1990, Springer 1990).Moreover, trophectoderm and TGCs express a7 integrins(laminin receptors) in the early postimplantation mouseembryo which may facilitate their invasion into the uterinestroma (Sutherland et al. 1993, Klaffky et al. 2001). Onthe other hand, E-cadherin, a member of the cadherinfamily of adhesion molecules, is expressed by embryonicectoderm, trophectoderm and labyrinthine trophoblastcells, but is absent on TGCs in the rat (Reuss et al. 1996).

Differentiation of TGCs is characterised by exit from thecell cycle leading to DNA endoreduplication and for-mation of characteristic giant nuclei (Barlow & Sherman1972, Snow & Ansell 1974, Cross 2000). DifferentiatedTGCs also produce metalloproteinases (Rinkenberger et al.1997) and several members of the placental prolactingene family, including placental lactogen (PL)-I (Faria et al.1990, Peters et al. 2000), PL-II (Campbell et al. 1989,Linzer & Fisher 1999) and proliferin (Linzer & Nathans1985).

Primary TGCs synthesise PL-I from early to mid preg-nancy in the mouse, while PL-II expression serves as anendocrine marker for secondary TGCs in the mid- to late-gestation period (Lin & Linzer 1998). PL-1 and PL-II target

the ovary and act to maintain the corpus luteum of preg-nancy and stimulate luteal progesterone production(Glosy & Talamantes 1995, Linzer & Fisher 1999), as wellas to promote mammary gland development and lactation(Colosi et al. 1988).

The PL-II is a non-glycosylated single-chain polypeptidewith a molecular size of about 25 kDa. It has significantamino acid sequence homology with PL-I (Duckworthet al. 1986, Jackson et al. 1986). In the mouse, PL-IIappears in the maternal circulation on day 9, increasesuntil about day 14 and then either remains relatively con-stant in some strains of mice or continues increasing untilthe end of pregnancy in others (Soares & Talamantes1983). Production of PL-II is controlled by several proteinssuch as epidermal growth factor (Yamaguchi et al. 1992),interleukin-1 and -6 (Yamaguchi et al. 1996) and decidua-derived calcyclin (Richard et al. 1998).

In this study, we have developed a serum-free model forsecondary trophoblast outgrowth and differentiation invitro based on those established by Romagnano & Babiarz(1990) and Sutherland et al. (1991, 1993). We have useda panel of molecular markers to identify secondary TGCsin vitro and in vivo and track TGC differentiation and therelationship between secondary trophoblast nuclear areaand PL-II expression. We have also used identification ofapoptotic cells to monitor cell survival in vitro.

Materials and Methods

Animals

All embryos were produced by natural mating of MF1female mice (7–12 weeks of age) with stud males of thesame strain (Harlan Olac Ltd, Bicester, Oxon, UK).Females were caged with males for mating and examinedthe following morning (day 0.5 of pregnancy) for the pre-sence of vaginal plugs to confirm that mating hadoccurred. Mice were kept under a standard 12 h light:12 hdark cycle with free access to water and food.

Culture of EPC explants

For culturing EPC explants, day 8.5 p.c. pregnant micewere killed by cervical dislocation. EPCs were dissectedas described previously (Hogan et al. 1994). Briefly,decidual capsules were dissected from the uterus andplaced in Dulbecco’s modified Eagles’ medium (DMEM)(Invitrogen, Paisley, Strathclyde, UK), with 10% (v/v) heat-inactivated foetal calf serum (Sigma, St Louis, MO, USA).The decidual capsules were split open to isolate thewhole embryos from the surrounding decidual tissues. Forseparation from the whole embryos, EPCs were dissectedat the junction with the extraembryonic ectoderm andseparated from attached TGCs with fine sterilised needles.

Secondary TGC outgrowth in serum-free culture

Isolated EPC explants were incubated for 3 days either onbovine serum albumin (BSA)- or 100mg/ml FN-coated

54 A H K El-Hashash and S J Kimber

Reproduction (2004) 128 53–71 www.reproduction-online.org

coverslips (Invitrogen) in a 24-well plastic tissue cultureplate (Nunc, Roskilde, Denmark). Each well contained0.5 ml culture medium. The culture medium consisted ofDMEM supplemented with 20 mg/ml L-glutamine andpenicillin–streptomycin (60mg/ml and 50mg/ml) (ICNPharmaceutical, Basingstoke, Hampshire, UK), and 2%(v/v) Nutridoma-HU (a growth factor-free growth-support-ing nutrient; Roche GmbH, Mannheim, Germany).

The time and percentage of EPC attachment to sub-strates were recorded at regular intervals after plating, asdescribed for blastocyst outgrowth by Sherman (1988).Briefly, an EPC explant was considered to be attachedwhen it remained attached to the coverslip when the cul-ture plate was tapped gently or swirled in a circularmotion. A giant cell was defined as one with a nucleuswhich had a diameter at least double that of nuclei (8–10mm) from cells at plating. TGCs could also be recog-nised by the fact that they were nearly detached from theoriginal cell explant but still shared at least 20% of theirborders with other cells. Further identification of second-ary TGCs was carried out by staining for PL-II and B bloodgroup antigen antibodies (TGC and trophectoderm markerrespectively), and absence of staining for vimentin (markerfor endoderm and mesoderm cells). The number of sec-ondary TGCs was counted and the surface area of out-growth was measured from the explant centre at regulartime-intervals after attachment of the explant (every 6 h for3 days). Fifty percent of the medium was replaced by freshmedium after 24 h of attachment. The progress of the out-growth was recorded using a Leitz Dialux inverted phasemicroscope (Leitz, Wetzlar, Germany).

Morphological and statistical analysis

Outgrowth extension was measured as the distancebetween the farthest migrated, or pioneer, cells and thecentre of the explant. A graticule was photographed at thesame magnification and used to calibrate printed images.A cross was drawn through the centre of the outgrowthand the top right quadrant was used to measure the out-growth extension and to count the number of outgrowingcells on printed images in all explants. Extension of out-growth or cell number were calculated as the average offour explants for each time-point in a single experiment.Every experiment was carried out in triplicate giving 12explants in total for each time-point. For Fig. 1b, 73 and62 explants were used for attachment on FN and BSA sub-strate respectively. The data were analysed using eitherone- or two-way ANOVA on SPSS computer software(SPSS Inc., Illinois, IL, USA). Differences were consideredsignificant at P , 0.05.

TUNEL assay

The terminal deoxynucleotide transferase-mediated dUTP-biotin nick end labelling (TUNEL) was carried out using aTUNEL kit (Roche) and following the company’s protocol.Briefly, EPC trophoblast cells were fixed in 4% (w/v)

paraformaldehyde (PFA; Sigma) in phosphate-bufferedsaline (PBS; Oxoid Ltd, Basingstoke, Hampshire, UK), andwashed in a PBS/PVP wash (PBS, pH 7.4 containing3 mg/ml polyvinylpyrrolidone (PVP) (Sigma)). They werepermeablised in PBS containing 0.5% Triton X-100(Sigma) for 60 min at room temperature, then washed inPBS/PVP. The terminal deoxynucleotide transferase (TdT)reaction mixture was prepared by mixing 10ml TdT(enzyme solution) with 90ml dUTP-fluorescein isothiocya-nate (FITC) labelling solution. Outgrowths were incubatedwith the TdT reaction mixture for 60 min at 37 8C in thedark before washing in PBS containing 0.5% Triton X-100,then in PBS/PVP. They were then mounted in Vectashieldcontaining 4,6-diaminidino-2-phenylindole hydrochloride(DAPI; Vector Laboratories, Burlingame, CA, USA). Con-trols were performed as above but without adding enzymesolution to the reaction mix. A positive control was alsoincluded by incubating outgrowths with 3 units/ml DNAseI (Roche). A trophoblast cell was considered apoptoticwhen showing fragmented nuclei (with blue channel forDAPI) which were also green-fluorescence labelled. Thepercentage of apoptotic cells was calculated from threesets of experiments performed on different occasions usingthree explant cultures in each set (excluding controls).

Immunofluorescence of outgrowths

Trophoblast outgrowths were fixed in fresh 4% PFA for10 min and washed in PBS. For staining with antibody toa7 integrins, outgrowths were fixed in 1% PFA for 10 minfollowed by another fixation in methanol for 8 min at220 8C. For cytoplasmic antigens, outgrowths were per-meablised in 0.01% Tween X-100 in PBS and blocked innormal goat serum (1:20 in PBS; Sigma). Outgrowths wereincubated overnight at 4 8C with a primary antibody asdescribed in Table 1, or with irrelevant antibody of thesame isotype as a negative control. The following day, out-growths were washed and incubated with the appropriateFITC-conjugated secondary antibody for 1 h at room tem-perature. Outgrowths were washed, mounted in Vecta-shield containing DAPI and photographed using anOlympus Vannox microscope (Olympus Optical Ltd,Tokyo, Japan) with epifluorescence capacity.

Immunofluorescence of sections

Day 8.5 p.c. mated females were killed by cervical dislo-cation. Dissected uterine horns were fixed in fresh 4%PFA overnight at 4 8C. Dissected decidua were dehy-drated, cleared and then embedded in paraffin wax(BDH, Poole, Dorset, UK). Paraffin sections were cut at5mm thickness and mounted on cleaned slides. Sectionswere de-paraffinised in xylene (Genta Medical, York,Yorks, UK), hydrated through a series of ethanols andwashed in distilled H2O before processing for antigenretrieval using microwave treatment. Sections were left tocool at room temperature, then treated with the follow-ing: 50 mM NH4Cl (BDH) in PBS for 30 min and 1%

In vitro model for murine secondary trophoblast giant cell differentiation 55

www.reproduction-online.org Reproduction (2004) 128 53–71

(w/v) BSA (Sigma), 0.2% (w/v) gelatine (Sigma) and0.05% (w/v) saponin (Sigma) in PBS three times for10 min. Sections were then incubated overnight at 4 8Cwith the appropriate primary antibody (Table 1) or

irrelevant antibody. On the next day they were rinsed in0.1% BSA, 0.2% gelatine and 0.05% saponin in PBSthree times for 10 min before being incubated with theappropriate secondary antibody (1:125) diluted in 0.1%

Table 1 Antibodies used in this study.

Antibody Specification Dilution Antigens Source

AP-2g Rabbit IgG 1:250 AP-2g Active motif; Geneka, Rixensart, Belgiuma7 integrin Rabbit IgG 1:200 a7 integrins Dr U Mayer; University of Manchester, Manchester, UKBOO6 Mouse IgM 1:40 B-histo-blood BioCarb; Lund, Sweden

group antigensDecma-1 Rat IgG 1:50 E-cadherin SigmaHOO1 Mouse IgM 1:30 Le-b/Le-y-blood MonoCarb; Lund, Sweden

group antigensPL-II Rabbit IgG 1:250 PL-II Chemicon, Harrow, UK

AP, activator protein.

Figure 1 Characteristics of secondary TGC outgrowth on FN and control substrates. EPC explants were plated in 0.5 ml medium in 24-well plateson FN-coated coverslips. (a) Average time at which attachment was first observed on different FN and Nutridoma-HU concentrations. (b) Thepercentage of EPCs which showed initial attachment during different time-periods on FN (100mg/ml) and BSA substrates. On FN (100mg/ml),outgrowths were scored for number of (c) secondary TGCs and (d) other cells (with non-giant nuclei). (e) Surface area of outgrowth wasmeasured on FN (100mg/ml) and BSA substrates. For (c–e), 50% of the culture medium was replaced after 24 h of attachment. Results are givenas means^S.E.M. of three sets of experiments performed on different occasions. n (the total number of explant cultures examined) ¼ 12 for (a)and 24 for (c, d and e). For (b), a total of 73 explants on FN substrate and 62 explants on BSA were used to plot the time-course of initial attach-ment. *P , 0.05 compared with the BSA control; (a) one-way ANOVA–Dunnett test and (c–e) two-way ANOVA.



BSA, 0.2% gelatine and 0.05% saponin for 1 h at roomtemperature. Sections were washed first in 0.1% BSA,0.2% gelatine and 0.05% saponin three times for 10 min,then in PBS three times for 10 min and finally in distilledH2O. Sections were mounted in a hydrophilic mountingmedia, Vectashield.

RNA extraction

Total RNA was extracted from secondary TGCs growing inculture using Qiagen RNeasy total RNA mini kit (Qiagen,Crawley, West Sussex, UK) following the manufacturer’sprotocol. To ensure extraction of RNA from trophoblastoutgrowths but not from the central explants, the latterwere separated from the culture using sterile fine needles.Quantification of RNA was carried out by measuring theabsorbance at 260 nm in a u.v. spectrophotometer(GeneQuant DNA/RNA Calculator; Pharmacia Biotech,St Albans, Herts, UK). The purity of the RNA was assessedby measuring the ratio of the absorbance at 260:280 nm.RNA was considered pure when A260/A280 $2.0.

RT-PCR

Relative changes in AP-2g and PL-II mRNA were exam-ined during 3 days of culture using RT-PCR. Changes inPCR products obtained for AP-2g and PL-II were normal-ised by comparison with two housekeeping genes: glycer-aldehyde-3-phosphate dehydrogenase (GAPDH) and S28.

First-strand cDNA synthesis

For annealing with the oligo-(dT) primer, 2mg templateRNA was incubated with 1mg oligo-(dT)12–18 primer(0.5mg/ml; Invitrogen) and sterile H2O to make a totalvolume of 10ml. The mixture was heated at 70 8C for10 min and chilled on ice. To allow cDNA synthesis, thefollowing reagents (all from Invitrogen) were added to themixture: 4ml 5 £ strength first-strand buffer (250 mMTris–HCl (pH 8.3), 375 mM KCl and 15 mM MgCl2), 2ml0.1 M dithiothreitol, 1ml dNTPs (10 mM each dNTP) and1ml moloney murine leukaemia virus reverse transcriptase(M-MLV RT; 200U). The reaction volume was made up to20ml by adding 2ml sterile H2O, then incubating at 37 8C

for 1 h. The RT reaction was inactivated by heatingsamples to 95 8C for 5 min, cooling on ice and storagewas at 220 8C. A negative control omitting the M-MLV RTwas run in parallel in all reactions. cDNA concentrationwas calculated by measuring the absorbance at 260 nm ina u.v. spectrophotometer.

PCR

An initial titration was carried out to determine that ampli-fications at the selected cycle number were within the lin-ear range. All primer sets were designed to span at leastone intron to ensure that cDNA, not genomic DNA, wasbeing amplified. Prior to PCR, primers were optimised toadjust their annealing temperatures and times followingQiagen’s protocols. PCR was carried out with an Eppen-dorf mastercycler gradient (Hamburg, Germany) using pri-mers and cycle numbers shown in Table 2. PCRamplification was performed using HotStarTaq DNA poly-merase kit (Qiagen) following the manufacturer’s protocol.Briefly, 5ml diluted cDNA (#1mg/reaction volume) wasadded to a mixture containing 5ml 10 £ strength PCRbuffer, 10ml 5 £ strength Q solution (a PCR additive thatfacilitates template amplification by modifying DNA melt-ing behaviour), 1ml dNTP mix (200mM of each dNTP),0.25mM appropriate primers (Table 2), 0.25ml HotStarTaqDNA polymerase (2.5 units/reaction) and distilled H2O toa total reaction volume of 50ml. This mixture was incu-bated in the thermal cycler under the following con-ditions: denaturation at 94 8C for 1 min, primer annealingat 60 8C for 1 min and primer extension at 72 8C for 1 min.PCR products were separated on 2% (w/v) agarose gelelectrophoresis using a 100 bp DNA ladder (Invitrogen)for size assessment. Gels were recorded using a Polaroiddual-intensity transilluminator camera (UVP Inc., SanGabriel, CA, USA).

DNA sequencing

To confirm identity, amplified DNAs were subcloned intoa plasmid TA cloning vector (pCRII; Invitrogen) andsequenced using an automated ABI DNA sequencer (ABIAnalytical, Ramsey, NJ, USA). The sequencing resultswere verified by a BLAST search.

Table 2 Primers used in this study.

Primera Primer sequence 50 ! 30 (base number of primer) Product size (bp) No. of cycles Gene Bank no.

AP-2g (f) GAGATGGCTCACCCCATAGA 181 28 BC003778(r) CAGGGACTGAGCAGAAGACC

PL-II (f) CACCAGACAACATCGGAAGA 243 28 M14647(r) TGACCATGCAGACCAGAAAG

GAPDH (f) TCTGAGGGCCCACTGAAG 200 28 NM008084(r) AGGGTTTCTTACTCCTTGGAGG

S28 (f) TAACCTGTCTCACGACGGTCT 180 28 BC010987(r) CAGGGATAACTGGCTTGTGG

a (f), forward primer; (r), reverse primer.



In situ hybridisation

Preparation of the probes

Mouse cDNA for PL-II was generated and amplified asdescribed above. The identity of the oligonucleotidesequence for PL-II cDNA was determined by DNA sequen-cing and BLAST search, which showed 98% homologywith the target gene. Further identification of the PL-IIcDNA was carried out by restriction enzyme digestionusing appropriate enzymes, which resulted in two frag-ments of the expected size. Purity of PL-II cDNA wasassessed using the u.v. spectrophotometer by measuringthe ratio of the absorbance at 260:280 nm. PL-II cDNA wasconsidered pure when A260/A280 ¼ 1.8–2.0. To furtherconfirm its purity, PL-II cDNA was run on a 2% agarose gelgiving a single distinct band. Before subcloning into theappropriate plasmid, PL-II cDNAs were purified by Gene-clean kit (Q.Biogene, Nottingham, Notts, UK), followingthe manufacturer’s protocol in order to clean PL-II cDNAfrom non-specific products.

PL-II cDNA was subcloned into pGEM-T plasmid(Promega, Southampton, Hants, UK), and transformedinto the E.coli host strain, XL-1 blue (Stratagene, SanDiego, CA, USA) as instructed by the manufacturer.PL-II-ligated plasmids were purified using a Qiagenplasmid miniprep purification kit and analysed for theirorientation and sizes by restriction digest. They werelinearised and used as templates for in vitro transcrip-tion. Generation of digoxygenin (DIG) dUTP-labelledsense and antisense riboprobes by in vitro transcriptionwas carried out using a DIG-RNA labelling kit (Roche)following the manufacturer’s protocol as adapted bySidhu & Kimber (1999). Labelling efficiency and probenormalisation was carried out using a dot/slot assayprotocol (Roche). Sense probes were used as controlsfor the specificity of hybridisation. Mouse b-actincDNA (a gift from Dr S Sidhu, University of California,San Francisco, CA, USA) was ligated to pGEM4Z(Promega) and processed for the generation of a posi-tive control riboprobe as mentioned for the PL-IIriboprobe.

Preparation of samples for in situ hybridisation

EPC culture EPC cultures were fixed on days 1, 2 and 3of attachment in 4% PFA in PBS at 4 8C overnight, thentreated with absolute methanol:30% hydrogen peroxide(5:1) before being processed directly for in situ hybridis-ation or stored in methanol at 220 8C until processed.

Paraffin-embedded tissues Uterine horns were fixed in4% PFA and embedded in paraffin wax (Sidhu & Kimber1999). After dehydration, decidua were transferred tochloroform:ethanol (1:1) and 100% chloroform for 30 minand 1 h respectively and then vacuum embedded. Sectionsof 7mm thickness were floated out on diethyl pyrocarbo-nate (DEPC; Sigma)-treated, warmed H2O to remove

wrinkles before mounting on 2% (v/v, acetone) aminopro-pyltriethoxysilane (Sigma)-coated slides.

Pre-hybridisation Paraffin sections were dewaxed, rehy-drated and refixed in 4% PFA in PBS for 20 min. Afterwashing in PBS, sections were treated with proteinase K(20mg/ml; Sigma) in Tris–EDTA (BDH) buffer for10 min, washed again in PBS, and then incubated in0.1 M triethanolamine HCl (pH 8.0; Sigma) containingacetic anhydride (0.25% v/v; Sigma). They were thenwashed in PBS and dehydrated through ethanols beforeair-drying.

In situ hybridisation (ISH)

Whole mount ISH on EPC culture The procedure fornon-radioactive whole amount in situ hybridisation wasadapted from Conlon & Rossant (1992) and Carney et al.(1993). Briefly, EPC outgrowths were rehydrated in amethanol:PBS (containing 0.1% Tween-20 (PBT)) series of75%/25%, 50%/50% and 25%/75% (v/v) before washingin PBT. Cultures were then treated with proteinase K(50mg/ml), washed in PBT containing freshly added gly-cine (2 mg/ml), then PBT before re-fixing in 0.2% (v/v) glu-taraldehyde (Sigma)/4.0% PFA. They were treated with0.1% (v/v) NaBH4 before pre-hybridisation by incubatingcultures for 4 h in the following buffer (PE): 50% deionisedformamide (Sigma), 0.75 M NaCl (BDH), 100mg/ml tRNA(Roche), 5 mg/ml sodium heparin (Sigma), 1 mg/ml fattyacid free-BSA (Roche) and 1% (w/v) SDS (Sigma) in10 mM PIPES (Sigma)/1mM EDTA (BDH), pH 6.8. Forhybridisation with the riboprobe, fresh buffer containingDIG-labelled PL-II or b-actin riboprobe (1mg/ml) wasadded. Cultures were incubated with the riboprobes over-night at 50 8C in humid chambers. The following day,three fresh hybridisation washes, A, B and C were pre-pared as follows: wash A ( ¼ 0.3 M NaCl, 1% (v/v) SDS inPE), wash B ( ¼ 0.05 M NaCl, 0.1% SDS in PE) and washC ( ¼ 0.5 M NaCl, 0.1% (v/v) Tween-20 in PE). Cultureswere incubated in 75%/25%, 50%/50% and 25%/75%(v/v) hybridisation buffer/wash A for 3 min each. Theywere then incubated twice for 30 min at 50 8C in wash A,then in wash B, equilibrated by incubation in RNAse buf-fer ( ¼ 0.5 M NaCl, 10 mM PIPES, pH 7.2, 0.1% (v/v)Tween-20) for 10 min, and then incubated with 0.1 mg/mlRNAse A in RNAse buffer for 30 min at 37 8C. Furtherwashings were carried out in wash A and wash B at 50 8Cfor 30 min each, followed by wash C for 20 min at 70 8C.Immunological detection of DIG was carried out by incu-bating EPC cultures with anti-DIG antibody conjugated toalkaline phosphatase (Roche) overnight at 4 8C, followedby colour development with nitro-blue-tetrazolium-chlor-ide (Sigma) and 5-bromo-4-chloro-3-indolyl-phosphate(Roche). The reaction was stopped by rinsing in DEPC-treated distilled water. Cultures were then dried andmounted in gelvatol (Fisher Scientific, Loughborough,



Leics, UK). Cells were examined under bright field (Olym-pus Vannox; Olympus, Tokyo, Japan).

Paraffin-embedded tissues In situ hybridisation on paraf-fin-embedded sections was carried out following themethods of Wilkinson & Nieto (1993) and Sidhu & Kimber(1999). Briefly, sections were incubated overnight at 65 8Cwith either PL-II or b-actin riboprobes at 1mg/ml inhybridisation buffer containing 50% (v/v) formamide(Sigma), 1% (v/v) SDS, 50mg/ml tRNA (Roche), 5 £

strength SSC (pH 4.5) and 50mg/ml heparin (Sigma).Hybridisation was performed in sealed humidifiedchambers to prevent sections from drying. The followingday, sections were given a series of stringent washes toreduce background. Immunological detection was carriedout as described for EPC cultures.

Image analysis

To examine the relationship between nuclear size andPL-II expression, the nuclear area of PL-II expressing andnon-expressing cells were measured with the assistanceof Image-Pro Plus analysis software (Media Cybernetics,Silver Spring, MD, USA). Nuclear area distributions weredetermined by measuring the nuclear area of cells inthe top right quadrant of outgrowths on each day of cul-ture. Within each field the nuclear area of PL-II-positivecells and unstained cells were measured. For each

culture day, 12 outgrowths were analysed in each exper-iment over six sets of experiments performed on differ-ent occasions.

Results

Cells of the growing EPC undergo final cell-fatedecisions around day 7.5 p.c. in the mouse in vivo (Yanet al. 2001). Therefore, day 8.5 p.c. was chosen forexcision of EPC explants to be used as progenitors forsecondary TGCs.

Establishment of an in vitro serum-free culture modelfor secondary TGC outgrowth

Our first goal was to determine the substrate requirementsfor secondary TGCs in order to establish an in vitro modelto study their development. When EPC explants were pla-ted on a low concentration FN substrate (10mg/ml) in thepresence of Nutridoma-HU (1%) in the medium, initialattachment was observed after an average time of 17 h(Fig. 1a). A small trophoblast outgrowth was formed onlyafter 2 days of attachment (Fig. 2b). At 1% Nutridoma-HU, increasing the FN concentration to 100mg/ml did notincrease trophoblast outgrowth (Fig. 2c), or significantlyreduce the time of initial attachment of the explants(Fig. 1a), suggesting that increasing the FN concentrationis not enough to produce further trophoblast outgrowth.

Figure 2 Phase contrast micrographs of secondary TGCs on (a) BSA (control)-, (b) 10mg/ml FN- and (c) 100mg/ml FN-coated coverslips with 1%Nutridoma-HU in the medium after 2 days of attachment. (d–h) Outgrowths on 100mg/ml FN substrate were with 2% Nutridoma-HU in themedium: (e) after 12 h of attachment, (d and f) after 30 h of attachment and (g and h) after 2 days of attachment. Arrowheads refer to (f) smallrounded cells growing in clusters, (g) elongated/spindle-shape TGCs and (h) multinucleated cells and giant nucleus insert is higher magnification(£400) of (h). Scale bars ¼ (a, b, c and d) 250mm, (e and f) 50mm and (g and h) 30mm.



However, using FN substrate (100mg/ml) and increasingNutridoma-HU to 2% in the medium significantly reducedthe average time taken for initial attachment of EPCexplants to occur, compared with attachment on BSA(control) (Fig. 1a). The majority of EPC explants attachedwithin 26 h: 70% initiated attachment in the period 10.5–26 h after plating (Fig. 1b). In addition, at FN (100mg/ml)and Nutridoma-HU (2%), a greater number of secondaryTGCs as well as an increased surface area of outgrowth(Figs 1c and e and 2d) was observed, compared with out-growth on BSA-coated coverslips (Figs 1c and e and 2a).

After 30 h of attachment, the rate of trophoblast out-growth was greatly reduced, probably due to consumptionof essential nutrients from the medium (data not shown).Replacing 50% of the culture medium after 24 h of culturemaintained a continuous increase of trophoblast out-growth (Fig. 1c–e).

Morphology of secondary TGCs in culture

EPC outgrowths exhibited different cell morphologies andpatterns of growth during their development in culture.After about 12 h of attachment, EPC explants extendedlong processes on the surface of the culture substrate (FN;100mg/ml) (Fig. 2e). Outgrowing cells formed a radialoutgrowth (Fig. 2d), which appeared as a monolayer sheetof cohesive cells after approximately 30 h of attachment(Fig. 2f). Regions on the periphery of the EPC explant didnot appear to produce outgrowing cells equally at the startof outgrowth (Fig. 2e). However, trophoblast outgrowthappeared to be formed from all around the EPC explantby approximately 30 h of attachment (Fig. 2d and f).

The morphology of secondary TGC-like cells was diverseduring their development in culture. By 24 h of EPC attach-ment, cells were small, rounded and growing in clusters(data not shown). At 30 h of attachment, cells became flat-tened with a pavement-like morphology (Fig. 2f). Somecells also grew in clusters at the periphery of the outgrowth(Fig. 2f). Fibroblast-like trophoblast cells were shown inculture after 2 days of attachment (Fig. 2g). Many outgrow-ing cells showed large cell size and characteristic, easilydistinguished giant nuclei (Fig. 2h), a characteristic ofTGCs. Some cells were also multinucleated (Fig. 2g and h)and aggregated together to form cell islands (Fig. 2h),suggesting that these cells were TGCs.

Trophoblast cell death in culture

To examine the differentiation and survival of trophoblastcells after a longer time in culture, we grew EPC out-growth for 6 days after attachment. Trophoblast cells con-tinued outgrowing during this period but with nosignificant change in total cell number (data not shown).In addition, the morphology of the cells changed dramati-cally during days 5 and 6 after attachment. At day 6, EPCoutgrowth appeared as a flattened sheet of trophoblastcells which were mainly in close cell–cell contact withan epithelial appearance (i.e. epithelial pavement-like

morphology) (Fig. 3a). These cells have large nuclei(Fig. 3a). However, the number of detached, probablyapoptotic cells, greatly increased during days 4-6 postattachment.

EPC cultures grown for 1–3 days showed very fewTUNEL-positive cells. After 1 day of attachment, the per-centage of TUNEL-positive cells was 2% (Fig. 3b) while at2–3 days, the percentage of apoptotic cells had increasedto 6% (Fig. 3b). After 3 days, however, TUNEL-positivecells increased markedly reaching 10%, 12.2% and15.3% of the total cell population at days 4, 5 and 6 afterattachment respectively (Fig. 3b). Consequently, we usedday 1–3 EPC cultures in this study.

Characterisation of secondary TGCs in vitro

TGCs were morphologically identified by their giantnuclear size. To further identify secondary TGC-like cells,we established a panel of molecular markers to the EPCoutgrowth in vitro. We examined the expression of B and

Figure 3 (a) Phase contrast micrograph of secondary TGCs after 6days of attachment. Note the pavement-like morphology of the cells.Scale bar ¼ 30mm. (b) Histogram showing the percentage of TUNEL-labelled cells at different days after attachment. Each bar representsthe means^S.E.M. of three independent experiments, each using threeexplants (i.e. a total of nine explants for each day).



Le-b/Le-y blood group antigens, a7 integrins (lamininreceptors) and E-cadherin (Figs 4 and 5). mRNA and pro-tein expression of AP-2g and secondary TGC-specific pro-tein PL-II were also examined (Figs 5, 6 and 7).

To ensure that no other types of cell were growing inthe EPC culture, EPC explants were dissected and isolatedfrom attached tissues with care. The absence of stainingfor vimentin confirmed the lack of mesoderm or endo-derm cells in the culture. EPC culture cells stained posi-tively for Le-b/Le-y blood group antigens (Fig. 4a), butwere negative for E-cadherin (Fig. 4c). Intense hom-ogenous staining for Le-b/Le-y blood group antigens wasmainly on the peripheral cell surface between spreadingcells (Fig. 4a). a7 integrin was expressed strongly, in apunctate pattern, at the periphery of cells containing giantnuclei (Fig. 4b). The pattern of expression of B bloodgroup antigens on the trophoblast cell surface changedwith time in culture (Fig. 5). In early EPC outgrowths (12 hafter attachment), B blood group antigens were expressedat cell surface peripheries between crowded cells in con-tact (Fig. 5a). After 2 days of attachment, B blood groupantigens appeared as punctate staining on the cell surface(Fig. 5b). With increasing cell spreading, stainingappeared to have increased and was homogenous at thecell surface periphery with punctate/patch-like stainingabundant on the rest of cell surface of more develop-mentally advanced outgrowths (Fig. 5e). These resultssuggested that changes in the pattern of distribution of B

blood group antigens may be correlated with thereduction of cell–cell adhesion and increasing cell

spreading. Control immunocytochemical reactions (irrele-

vant primary) showed the absence of non-specific staining(Fig. 5f).

To further identify secondary TGCs in culture, we exam-

ined gene and protein expression of AP-2g and PL-II. Tran-scripts for AP-2g, a gene acting to regulate TGC

proliferation and differentiation, were expressed by sec-

ondary TGCs on all 3 days of culture (Fig. 6a). Tropho-blast-like cells with giant nuclei showed positive staining

for AP-2g protein (Fig. 6b). AP-2g stain was localisedmainly to the nuclei, with a diffuse reticulated pattern.

However, less intense cytoplasmic staining was also

observed (Fig. 6b). Control outgrowths showed theabsence of non-specific staining (Fig. 6c). A population of

trophoblast-like cells with giant nuclei expressed PL-II pro-tein in vitro (Fig. 7). At day 1 of attachment, trophoblast

cells spread as cell islands and expressed PL-II mainly atthe periphery of the cytoplasm (Fig. 7a). By day 3 of

attachment, cytoplasmic staining of PL-II had increased;

however, the highest staining intensity was still observedat the cell periphery (Fig. 7b). Cells with small, non-giant

nuclei showed no staining for PL-II (Fig. 7a and b). Out-growths incubated with irrelevant antibody of the same

species/isotype showed the absence of non-specific stain-

ing (Fig. 7c).

Figure 4 Immunostaining of secondary TGCs from EPC outgrowths cultured on FN (100mg/ml) substrate for 2 days after attachment. (a) Le-b/Le-yblood group antigens, (b) a7 integrin, (c) E-cadherin and (d, e and f) double staining with DAPI. Arrowheads refer to high stain intensity on cellsurface peripheries. Scale bars ¼ 10mm.



Expression of PL-II transcripts

Changes in the expression of transcripts for PL-II from day1 to day 3 in culture were examined by semi-quantitativeRT-PCR. The expression of PL-II RNA appeared relativelysimilar at all 3 days of culture as compared with GAPDHand S28 mRNA (Fig. 8a). Expression of PL-II RNA by sec-ondary TGCs was further analysed by in situ hybridisationboth in vitro (Fig. 8) and in vivo (Fig. 9). Using a ribo-probe specific for PL-II RNA transcripts, PL-II were loca-lised exclusively to secondary TGCs in vitro (Fig. 8) and

in vivo (Fig. 9). The sense probe for PL-II showed no

detectable hybridisation (Figs 8c and 9a and d). Second-

ary TGC-like cells spreading as cell clusters/islands

exhibited a heterogeneous pattern of staining for PL-II,

with signals concentrated in the cytoplasm (Fig. 8b). In

contrast, all cells were hybridised with the riboprobe for

b-actin (Fig. 8d). PL-II message was detected in secondary

TGCs at days 7.5 and 8.5 of pregnancy (Fig. 9b, c, e and

f). PL-II signals were more abundant in peripheral tropho-

blast cells than those growing near the EPC (Fig. 9b and

Figure 5 Expression of B blood group antigens on secondary TGCs cultured on FN (100mg/ml) substrate after (a) 1, (b) 2 and (e) 3 days of attach-ment. (f) Staining with mouse IgM (negative control) and (c, d, g and h) double staining with DAPI. Arrowheads refer to high stain intensity oncell surface peripheries. Scale bars ¼ 10mm.



c), suggesting that PL-II expression increases as tropho-blast cells are displaced peripherally. A strong hybridis-ation signal was also observed in secondary TGCislands/clusters (Fig. 9e and f).

Relationship between nuclear size and PL-IIexpression

In preliminary experiments, we observed that the increasein the number of cells expressing PL-II, a marker of func-tional differentiation of secondary TGCs, coincidedwith the elevation of the number of cells with giantnuclei, a marker for morphological differentiation of sec-ondary TGCs. Consequently, we assessed the relationshipbetween PL-II expression and giant cell nuclear sizeduring secondary TGC development in culture (Fig. 10).

A computer software-based measurement of nucleararea was used to determine whether nuclear enlargement,the main morphological character of differentiated TGC,was related to the expression of PL-II. The two-dimen-sional area covered by nuclei was therefore examined inrelation to expression of PL-II (Fig. 10). Expression of PL-IIwas restricted to cultured cells containing large/giantnuclei (Fig. 10). At all culture days, there was a highly sig-nificant difference (P # 0.0001) in nuclear area betweenPL-II-expressing and non-expressing cells. The observedarea covered by nuclei in PL-II-expressing cells was 306–440mm2, while nuclei in non-expressing cells covered asmaller area (30–304mm2) (Fig. 10).

Moreover, examination of EPC cultures by image analy-sis indicated an obvious shift in the nuclear area of cellsover the 3 days of culture (Fig. 10). This shift in nucleararea was associated with the increase of PL-II-positivelystained cells with time in culture (Fig. 10).

Expression of TGC markers in vivo

The expression of TGC markers, B, Le-b/Le-y blood groupantigens, a7 integrins and AP-2g as well as PL-II in themouse uterus was examined to see whether these proteinswere expressed by specific cells at day 8.5 of pregnancy(Fig. 11). Multicellular islands of secondary TGCs showedintense staining for all these markers (Fig. 11), but noimmunoreactivity was observed in the uterine cells(Fig. 11). No staining was observed with irrelevant primarycontrols (data not shown).

Discussion

In the present study, we have investigated the develop-ment of secondary TGCs in culture using EPC explantsisolated from day 8.5 p.c. mouse embryos as a source ofsecondary TGCs. The current in vitro culture model, likesimilar secondary TGC culture systems (Romagnano &Babiarz 1990, Sutherland et al. 1991, 1993), showed thatEPC explants could attach and form secondary TGC out-growths on an FN substrate. Outgrowth occurred in theabsence of serum or medium conditioned by other cells.

Figure 6 Expression of AP-2g by secondary TGCs in vitro. (a) RT-PCR analysis of relative AP-2g mRNA levels in EPC culture (top panel). Changesin AP-2g cDNA were compared with two endogenous standards, GAPDH cDNA (middle panel) and S28 cDNA (bottom panel). PCR was per-formed in triplicate on different cDNA samples. Relative AP-2g mRNA levels remained nearly similar between culture days. (b) Expression ofAP-2g protein after 2 days of attachment. (c) Double staining with rabbit IgG (negative control) and DAPI. (d) Same explant as (b) stained withDAPI. M, 100 bp ladder; -ve, control reaction (no RT). Scale bars ¼ 10mm.



High concentration of FN stimulates secondary TGCoutgrowth

Trophoblasts growing on an extracellular matrix (ECM)substrate such as FN in culture can mimic the pro-gression of secondary TGCs along the endometrial base-ment membrane in vivo (Sherman & Atienza-Samols1978, Enders et al. 1981, Armant et al. 1986a). FN hasalso been shown to effectively promote blastocystattachment and outgrowth of primary (Armant et al.1986a) and secondary TGCs in vitro (Romagnano &Babiarz 1990). Secondary TGC outgrowth was thereforeassessed on FN substrates produced using two differentconcentrations of FN, and two different concentrationsof the serum substitute, Nutridoma-HU, were alsotested.

Our study established that the highest concentrationsof FN (100mg/ml) and Nutridoma-HU (2%) stimulaterapid attachment of a high percentage (86%) of EPCexplants in vitro. The number of secondary TGCs anddegree of trophoblast spreading was also high underthese conditions. FN is present in the serum and the

basement membranes of the endometrial epithelium andinterstitial matrix of uterine stroma (Wartiovaara et al.1979, Grinnell et al. 1982). Attachment and outgrowthon an FN substrate in vivo and in vitro required tropho-blast cells to differentiate and express FN receptors(Armant et al. 1986a). Trophoblast outgrowth on FN sub-strate in vitro has been suggested to represent a modelfor trophoblast invasion in vivo where trophoblast cellsattach and degrade matrix FN in order to penetrate thestroma (Armant et al. 1986a).

As a cell attachment molecule (Yamada 1983), FNpromotes the migration of many cell types, includingparietal endoderm (Behrendtsen et al. 1995), neural crest(Bronner-Fraser 1986), primordial germ cells (Ffrench-Constant et al. 1991) and avian precardiac mesoderm(Linask & Lash 1988). Inhibiting FN interaction withthese cells using FN-inhibitory antibodies, RGD-contain-ing peptides or antibodies to FN receptors results in lossof cell migration and reduced outgrowth (Bronner-Fraser1986, Linask & Lash 1988, Ffrench-Constant et al.1991).

Figure 7 Expression of PL-II protein by secondary TGCs in vitro after(a) 1 and (b) 3 days of EPC explant attachment, which were doublestained with DAPI. (c) Double staining with rabbit IgG (negative con-trol) and DAPI. Arrowheads refer to high stain intensity at cell per-ipheries. Arrows refer to unstained cells with non-giant nuclei. Scalebars ¼ 10mm.



Survival of trophoblast cells in culture

Very few cells were lost from the EPC culture systemduring the first 72 h after attachment. The percentage ofapoptotic cells was low and these were located close tothe EPC explants. They were absent in the peripheral (pio-neer) trophoblast cells that have previously been identifiedas differentiated TGCs (Guillemot et al. 1994, Cross 2000,Scott et al. 2000). A possible explanation may be theproduction of anti-apoptotic factor(s) such as parathyroidhormone-related protein (PTHrP) which has been reportedto be produced by differentiated secondary TGCs (Karper-ien et al. 1996, A K H El-Hashash S J Kimber, unpublishedobservations), or the presence of other apoptosis-inhibi-tory factor such as transforming growth factor-a (Brison &Shultz 1998), while EPC outgrowth central cells lack boththese factors. Furthermore, cell death may be triggered ifcells fail to enter either the mitotic cell cycle or the pro-gramme of endoreduplication (Goncalves et al. 2003).

Characteristic morphology of secondary TGCsgrowing in vitro

In the present study, secondary TGCs grew in vitro as amonolayer sheet or meshwork of flattened cells, in agree-ment with Romagnano & Babiarz (1990). Multinucleatedcells with large and irregular-shaped nuclei were observed

as previously reported for secondary TGCs (Zybina &Zybina 1996). The characteristic increase in the nuclearsize was reported by Barlow & Sherman (1972) in themouse to be due to DNA endoreduplication.

Characterisation of the secondary trophoblastoutgrowth model

In this study, we have described the expression of a num-ber of antigens that have previously been reported to beexpressed by primary and/or secondary trophoblast cellsat postimplantation development in the mouse. There arelimited specific markers for secondary TGCs. Only PL-IIhas been used extensively as a specific molecular markerfor secondary TGCs. Therefore, we used known generalmarkers for trophoblast cells, e.g. blood group antigens,a7 integrins and AP-2g, after ensuring that cultures werefree from other embryonic cell types. Generating a culturemodel for secondary TGCs free of other cell types wasachieved by careful isolation of EPC explants fromattached embryonic and extraembryonic tissues, and con-firmed by lack of staining of EPC culture cells for vimentinand negligible E-cadherin.

The distribution patterns of B and Le-b/Le-y blood groupantigens and a7 integrins were fairly similar. Blood groupantigens and a7 integrins were localised mainly to cellperipheries with some punctate staining on cell surfaces.

Figure 8 Expression of PL-II mRNA by secondary TGCs in vitro. (a) RT-PCR showing the relative levels of PL-II mRNA on days 1–3 of EPC culture(top panel). Changes in PL-II cDNA were compared with two endogenous standards, GAPDH cDNA (middle panel) and S28 cDNA (bottompanel). PCR was performed in triplicate on different cDNA samples. (b) In situ hybridisation of day 3 EPC outgrowths hybridised with DIG-labelled PL-II antisense riboprobe. (c) Control outgrowth hybridised with sense probe for PL-II, negative control or (d) DIG-labelled b-actin anti-sense, positive control. Note the strong staining over the cytoplasm corresponding to the site of PL-II mRNA localisation, and cells spreading inclusters. Arrowheads refer to the hybridisation in the cytoplasm. M, 100 bp ladder; -ve, control reaction (no RT). Scale bars ¼ 10mm.



Le-b/Le-y blood group antigens have been shown to beexpressed on the surface of TGCs at pre- and postimplan-tation stages, while B blood group antigens are expressedat day 6–7 p.c. by both primary and secondary TGCs aswell as parietal endoderm (Kimber 1990, Brown et al.1993). A change in the distribution of B blood group anti-gens was observed from mainly restricted to the cell mar-

gins to covering the entire cell surface. Whether thisrelates to the reduction of cell–cell adhesion and increasein cell spreading remains to be determined.

Sutherland et al. (1993) and Klaffky et al. (2001) pre-viously reported that a7b1 integrin receptors are expressedon trophectoderm cells at late blastocyst and on bothprimary and secondary TGCs until day 8.5 p.c. They areessential for trophoblast–laminin interactions that facili-tate trophoblast invasion into the uterine stroma duringimplantation. In agreement with these authors, we foundsecondary TGCs expressed intense a7 integrins bothin vivo and in our cultures. Our study demonstrated thatE-cadherin protein expression was extremely low or

undetectable in secondary trophoblast outgrowth in vitro.Similar results were shown in other mammalian embryoswhere downregulation of E-cadherin was coincident

with the differentiation of TGCs in the rat (Reuss et al.1996), and with the differentiation of non-invasivecytotrophoblast into invasive trophoblast in the human(Coutifaris et al. 1991, Floridon et al. 2000).

Secondary TGCs cultured from day 8.5 p.c. EPCexplants expressed PL-II in their cytoplasm in keeping

with previous in vivo and in vitro studies. PL-II has beenused as a specific marker for secondary TGCs (Linzer &Nathans 1985, Soares et al. 1991). PL-II produced by sec-ondary TGCs has a lactogenic effect and promotes mam-mary gland growth (Robertson et al. 1982). It enhancesthe expression of both oestrogen receptors a and b (Tell-eria et al. 1998), and stimulates luteal progesterone pro-duction (Linzer & Fisher 1999). We also demonstratedthe expression of AP-2g in the nuclei of secondary TGCsin vitro. AP-2g protein has been shown to be present inmurine secondary TGCs in vivo (Sapin et al. 2000, Aumanet al. 2002), and detected at the RNA level in human cyto-trophoblast cells in vitro (Richardson et al. 2001). As anupstream regulator of several trophoblast-specific genes,

AP-2g has been suggested to regulate the genetic pro-grammes controlling murine trophoblast proliferation anddifferentiation (Werling & Schorle 2002). In other studies,

Figure 9 Cellular localisation of PL-II mRNA within (a–c) day 7.5 and (d–f) day 8.5 mouse embryos. PL-II mRNA was detected by in situ hybrid-isation of cross sections of the embryo and the surrounding decidua with PL-II antisense riboprobe. (a and d) Absence of specific hybridisationsignals in representative sections exposed to DIG-labelled PL-II sense riboprobe, negative control. (c) Peripheral cells far from the EPC expresshigher PL-II signals than (b) central trophoblast near the EPC. (e) Trophoblast cell clusters express PL-II RNA. (f) Higher magnification of the boxin (e) showing cytoplasmic staining of PL-II. Arrows refer to PL-II-positively stained cells. Am, amnion. Scale bars ¼ 10mm.



AP-2g has also been shown to control murine adenosinedeaminase gene expression (Shi et al. 1997, Shi &Kellems 1998), and human placental lactogen (Richardsonet al. 2000), as well as chorionic gonadotrophin-b (John-son & Jameson 1999).

Morphological and functional differentiation ofsecondary TGC in vitro

EPC-derived secondary TGCs differentiate in a serum-freemedium in vitro and the degree of their differentiationincreased with time after attachment. This was shownmorphologically by increasing numbers of cells with giantnuclei, and functionally by increasing numbers of PL-II-expressing cells. Our results relating cell morphologyand nuclear area to PL-II expression confirm in vitro

differentiation of secondary TGCs. If secondary TGCsdifferentiate normally in culture, one can expect PL-II tobe expressed only by cells with giant nuclear size, but notby small nuclei cells. This expectation was investigatedquantitatively using image analysis to measure the nucleararea of EPC trophoblast cells. PL-II-expressing cells werefound to have significantly (P , 0.0001) larger nuclearareas than non-expressing cells. These results were inkeeping with those reported in similar studies by Carneyet al. (1993) on cultured isolated EPC explants, and byFaria & Soares (1991) on Rcho1 cells. Although our datashowed a distinct relationship between nuclear area andPL-II expression by secondary TGCs, they did not demon-strate cause and effect. Increasing cell DNA content ofTGCs reported here and by others in vitro (Goncalves et al.2003) and in vivo (Barlow & Sherman 1972) will result

Figure 10 Comparison of the nuclear size distribution for the total trophoblast cells and PL-II-positive secondary TGCs at days 1, 2 and 3 afterattachment. Note the increase in the number of cells with larger nuclear area from day 1 to day 3 in culture and that PL-II-positive cells wererestricted to cells possessing larger nuclei ($306mm2). Results are cumulative values for nuclear size over 12 explants calculated from six sets ofexperiments performed on different occasions on each day tested. The total cells analysed were 709 on day 1, 548 on day 2 and 627 on day 3.



not only in nuclear enlargement but changing nuclear:cy-toplasmic volume and surface ratio (Ilgren 1980, Kuhnet al. 1991). These changes have been suggested toalter the availability or effective concentrations ofdifferentiation-regulatory molecules within the trophoblastcell (Carney et al. 1993).

Several reports identified the cells of trophoblast out-growth growing close to the EPC explant as proliferativestem cells and the peripheral cells as differentiated sec-ondary TGCs as evidenced by their expression of specifictranscription factors (Guillemot et al. 1994, Cross 2000,Scott et al. 2000). In this study, we found that PL-II RNAexpression was more abundant in peripheral trophoblastcells than those growing near the EPC in day 7.5 embryosin vivo. These results support a previous suggestion thatEPC trophoblast cells differentiate into secondary TGC asthey are displaced peripherally (Hoffman & Wooding1993, Cross 2000). We have also detected a change in theexpression of cyclin B1 and cyclin D1 correlating with anincrease in the nuclear size as trophoblast cells spreadaway from the original EPC explant in vitro (A H K El-Hashash & S J Kimber, unpublished observations).

The mechanisms responsible for TGC differentiation arestarting to be understood, particularly for primary TGCs.However, factors that trigger secondary TGC differentiationare still poorly defined. Morphological differentiationof TGCs is accompanied by arrest of cell proliferation

but continued DNA synthesis (Snow & Ansell 1974,Varmuza et al. 1988, Hoffman & Wooding 1993,Goncalves et al. 2003). DNA endoreduplication andformation of secondary TGCs with giant nuclei are associ-ated with their functional differentiation, i.e. production ofPL-II (Soares & Glasser 1987) and progesterone (Glasseret al. 1987). The arrest of cell proliferation may be the firststep of trophoblast cell differentiation (Faria & Soares1991). In the present work, the in vitro induction ofsecondary TGC differentiation may be due to autocrinefactors produced by the trophoblast cells themselves. Onepotential factor that may induce in vitro differentiation ofsecondary TGC is PTHrP. PTHrP is produced by secondaryTGCs, acts to increase DNA synthesis in vitro (Centrellaet al. 1989), and stimulates primary TGC outgrowth invitro (Nowak et al. 1999). In parallel work, we are studyingthe role of PTHrP and other factors in regulating secondarytrophoblast outgrowth and differentiation using the modelsystem described here (A H K El-Hashash & S J Kimber,unpublished observations).

In conclusion, we have established a serum-freemodel in which secondary TGCs express differentiatedphenotypes with a characteristic time-course on a FNsubstrate. The differentiated phenotypes include bothmorphological and functional (i.e. PL-II expression) par-ameters. In this model, PL-II, AP-2g and a7 integrinsproved to be characteristic markers for secondary TGCs

Figure 11 Expression of different secondary TGCs markers in day 8.5 mouse embryos in vivo. (a) Haematoxylin and eosin staining. Paraffin sec-tions were stained for (b) B blood group antigens, (c) Le-b/Le-y blood group antigens, (d) a7 integrins, (e) PL-II and (f) AP-2g. Arrowheads refer tointense staining of secondary TGC islands. Am, amnion. Scale bars ¼ 10mm.



in vitro, and PL-II protein expression was related to cellnuclear area. Apoptosis was shown to increase in parallelwith the differentiation of cells. This culture model canbe used to study the factors and mechanisms regulatingsecondary TGC differentiation.

Acknowledgements

The authors would like to acknowledge Dr Ulrike Mayer,University of Manchester, UK for providing antibody for a7

integrins, Dr S Sidhu, University of California, San Francisco,USA for generously furnishing cDNA for murine b-actin andMr Ian Miller, University of Manchester, UK for assistancewith image processing. We also thank Mr Mark Lunt, Epide-miology Department, University of Manchester, UK foradvice on statistical analysis.

References

Armant DR, Kaplan HA & Lennarz WJ 1986a Fibronectin andlaminin promote in vitro attachment and outgrowth of mouseblastocysts. Developmental Biology 116 519–523.

Armant DR, Kaplan HA, Mover H & Lennarz WJ 1986b The effect ofhexapeptides on attachment and outgrowth of mouse blastocystscultured in vitro: evidence for the involvement of the cell recog-nition tripeptide Arg-Gly-Asp. PNAS 83 6751–6755.

Auman HJ, Nottoli T, Lakiza O, Winger Q, Donaldson S & WilliamsT 2002 Transcription factor AP-2g is essential in the extra-embryo-nic lineages for early postimplantation development. Development129 2733–2747.

Barlow BW & Sherman MI 1972 The biochemistry of differentiationof mouse trophoblast: studies on polyploidy. Journal of Embryologyand Experimental Morphology 27 447–465.

Behrendtsen O, Alexander CM & Werb Z 1995 Cooperative inter-actions between extracellular matrix, integrins and parathyroidhormone-related peptide regulate parietal endoderm differentiationin mouse embryos development. Development 121 4137–4148.

Brison DR & Schultz RM 1998 Increased incidence of apoptosis intransforming growth factor a-deficient mouse blastocyst. Biology ofReproduction 59 136–144.

Bronner-Fraser M 1986 An antibody to a receptor for fibronectin andlaminin perturbs cranial neural crest development in vivo. Devel-opmental Biology 117 528–536.

Brown DG, Warren VN, Pahlesson P & Kimber SJ 1993 Carbohydrateantigen expression in murine embryonic stem cells and embryos.I. Lacto and neo-lacto determinants. Histochemical Journal 25452–463.

Campbell WJ, Deb S, Kwok SCM, Joslin J & Soares MJ 1989 Differ-ential expression of placental lactogen-II and prolactin-like pro-tein-A in the rat chorioallantoic placenta. Endocrinology 1251565–1574.

Carney EW, Prideaux V, Lye SJ & Rossant J 1993 Progressiveexpression of trophoblast-specific genes during formation of mousetrophoblast giant cells in vitro. Molecular Reproduction andDevelopment 34 357–368.

Centrella M, Canalis E, McCarthy T, Stewart A, Orloff J & Insogna K1989 Parathyroid hormone-related protein modulates the effect oftransforming growth factor beta on deoxyribonucleic acid and col-lagen synthesis in fetal rat bone cells. Endocrinology 125199–208.

Colosi P, Ogren L, Southard JN, Thordarson G, Linzer DIH & Tala-mantes F 1988 Biological, immunological, and binding propertiesof recombinant mouse placental lactogen-1. Endocrinology 1232662–2667.

Conlon RA & Rossant J 1992 Exogenous retinoic acid rapidly inducesanterior ectopic expression of murine Hox-2 genes in vivo. Devel-opment 116 357–368.

Copp AJ 1979 Interaction between inner cell mass and trophecto-derm of the mouse blastocyst. II. The fate of the polar trophecto-derm. Journal of Embryology and Experimental Morphology 51109–120.

Coutifaris C, Kao L-C, Sehdev H, Chin U, Babalola G, Blaschuk O &Strauss JF 1991 E-cadherin expression during differentiation ofhuman trophoblast. Development 113 767–777.

Cross JC 2000 Genetic insight into trophoblast differentiation andplacental morphogenesis. Seminars in Cellular and DevelopmentalBiology 11 105–113.

Cross JC, Werb Z & Fisher SJ 1994 Implantation and the placenta:key pieces of the developmental puzzle. Science 266 1508–1518.

Duckworth ML, Kirk KL & Friesen HG 1986 Isolation and identifi-cation of cDNA clone of rat placental lactogen II. Journal of Bio-logical Chemistry 261 10871–10878.

Enders AC, Chavez DJ & Schlafke S 1981 Comparison of implan-tation in utero and in vitro. In Cellular and Molecular Aspects ofImplantation, pp 365–382. Eds SR Glasser & BW Bullock. NewYork: Plenum.

Farach MC, Tang JP, Decker GL & Carson DD 1987 Heparin/heparinsulfate is involved in attachment and spreading of mouse embryosin vitro. Developmental Biology 123 401–410.

Faria TN & Soares MJ 1991 Trophoblast cell differentiation: establish-ment, characterization, and modulation of a rat trophoblast cellline expressing members of the placental prolactin family. Endo-crinology 129 2895–2906.

Faria TN, Deb S, Kwok SCM, Talamantes F & Soares MJ 1990 Onto-geny of placental lactogen-I and placental lactogen-II expressionin the developing rat placenta. Developmental Biology 141279–291.

Farnsworth RL & Talamantes F 1998 Calcyclin in mouse decidua:expression and effects on placental lactogen secretion. Biology ofReproduction 59 546–552.

Fenderson BA, Holmes EH, Fukushi Y & Hakomori S 1986 Coordi-nate expression of X and Y haptens during murine embryogenesis.Developmental Biology 114 12–21.

Ffrench-Constant C, Hollingsworth A, Heaasman J & Wylie CC 1991Response to fibronectin of mouse primordial germ cells before,during and after migration. Development 113 1365–1373.

Floridon C, Nielson O, Holund B, Sunde L, Westergaard JG, Thom-sen SG & Teisner B 2000 Localization of E-cadherin in villous,extravillous and vascular trophoblast during intrauterine, ectopicand molar pregnancy. Molecular Human Reproduction 6943–950.

Glasser S, Julian J, Munir MI & Soares M 1987 Biological markersduring early pregnancy: trophoblast signals of the periimplantationperiod. Environment and Health Perspectives 74 120–147.

Glosy SS & Talamantes F 1995 Luteotropic actions of placental lacto-gens at midpregnancy in the mouse. Endocrinology 1363993–4003.

Goncalves CR, Antonini S, Vianna-Morgante AM, Machado-SantelliGM & Bevilacqua E 2003 Developmental changes in the ploidy ofmouse implanting trophoblast cells in vitro. Histochemistry andCell Biology 119 189–198.

Grinnell F, Head JR & Hoffpauir J 1982 Fibronectin and cell shape invivo: studies on the endometrium during pregnancy. Journal ofCell Biology 94 597–606.

Guillemot F, Nagy A, Auerbach A, Rossant J & Joyner AL 1994Essential role of Mash-2 in extraembryonic development. Nature371 333–336.

Hoffman L & Wooding FBP 1993 Giant and binucleate trophoblastcells in mammals. Journal of Experimental Zoology 266 559–577.

Hogan B, Beddington R, Constantini F & Lacy E 1994 Manipulatingthe Mouse Embryo: A Laboratory Manual, 2nd edn, pp 112–115.New York: Cold Spring Harbour Laboratory Press.



Ilgren EB 1980 Polyploidization of extraembryonic tissues duringmouse embryogenesis. Journal of Embryology and ExperimentalMorphology 59 103–111.

Jackson LL, Colosi P, Talamantes F & Linzer DIH 1986 Molecularcloning of mouse placental lactogen cDNA. PNAS 83 8496–8500.

Johnson W & Jameson JL 1999 AP-2 (Activating protein 2) and Sp1(selective promoter factor1) regulatory elements play distinct rolesin the control of basal activity and cyclic adenosine 30,50-mono-phosphate responsiveness of the human chorionic gonadotropin-beta promoter. Molecular Endocrinology 13 1963–1975.

Karperien M, Lanser P, De Laat SW, Boonstra J & Defize LHK 1996Parathyroid hormone related peptide mRNA expression duringmurine postimplantation development: evidence for involvementin multiple differentiation processes. International Journal ofDevelopmental Biology 40 599–608.

Kimber SJ 1990 Glycoconjugates and cell surface interactions in pre-and peri-implantation mammalian embryonic development. Inter-national Review of Cytology 120 153–167.

Klaffky E, Williams R, Yao C, Ziober B, Kramer R & Sutherland A2001 Trophoblast-specific expression and function of the integrina7 subunit in the peri-implantation mouse embryo. DevelopmentalBiology 239 161–175.

Kuhn EM, Therman E & Susman B 1991 Amitosis and endocycles inearly cultured mouse trophoblast. Placenta 12 251–261.

Lin J & Linzer DIH 1998 Regulation of erythropoiesis by mouse pla-cental prolactin-like proteins. Proceedings of the 80th Meeting ofthe Endocrine Society, New Orleans, LA. Abstract P3–463.

Linask KK & Lash JW 1988 A role for fibronectin in the migration ofavian precardiac cells. I. Dose-dependent effects of fibronectinantibody. Developmental Biology 129 315–323.

Linzer DIH & Nathans D 1985 A new member of the prolactin-growth hormone family expressed in mouse placenta. EMBO Jour-nal 41 419–1423.

Linzer DIH & Fisher SJ 1999 The placenta and the prolactin family ofhormones: regulation of the physiology of pregnancy. MolecularEndocrinology 13 837–840.

Nowak RA, Haimovici F, Biggers JD & Erbach GT 1999 Transforminggrowth factor-b stimulates mouse blastocyst outgrowth through amechanism involving parathyroid hormone-related protein.Biology of Reproduction 60 85–93.

Parast MM, Aeder S & Sutherland AE 2001 Trophoblast giant celldifferentiation involves changes in cytoskeleton and cell motility.Developmental Biology 230 43–60.

Peters TJ, Chapman BM, Wolfe MW & Soares MJ 2000 Placental lac-togen-1 gene activation in differentiation trophoblast cells: extrin-sic and intrinsic regulation involving mitogen-activated proteinkinase signalling pathways. Journal of Endocrinology 165443–456.

Reuss B, Hellmann P, Dahl E, Traub O, Butterweck A, Grummer R &Winterrhager E 1996 Connexins and E-cadherin are differentiallyexpressed during trophoblast invasion and placenta differentiationin the rat. Developmental Dynamics 205 172–182.

Richardson BD, Langland RA, Bachurski CJ, Richards RG, KesslerCA, Cheng YH & Handwerger S 2000 Activator protein-2 regulateshuman placental lactogen gene expression. Molecular and CellularEndocrinology 160 183–192.

Richardson BD, Cheng Y-H, Langland RA & Handwerger S 2001Differential expression of AP-2g and AP-2a during human tropho-blast differentiation. Life Sciences 69 2157–2165.

Rinkenberger JL, Cross JC & Werb Z 1997 Molecular genetics ofimplantation in the mouse. Developmental Genetics 21 6–20.

Robertson MC, Gillespie B & Friesen HG 1982 Characterization oftwo forms of rat placental latogen (rPL): rPL-I and rPL-II. Endocrin-ology 111 1862–1866.

Romagnano L & Babiarz B 1990 The role of murine cell surfacegalactosyltransferase in trophoblast: laminin interaction in vitro.Developmental Biology 141 254–261.

Sapin V, Bouillet P, Oulad-Abdelghani M, Dastugue B, Chambon P &Dolle P 2000 Differential expression of retinoic-acid inducible

(Stra) genes during mouse placentation. Mechanism of Develop-ment 92 295–299.

Scott IC, Anson-Cartwright L, Riley P, Reda D & Cross JC 2000 TheHand1 basic helix-loop-helix transcription factor regulates tropho-blast giant cell differentiation via multiple mechanisms. Molecularand Cellular Biology 20 530–541.

Sherman MI 1988 Implantation of mouse blastocyst in vitro. InMethods in Mammalian Reproduction, pp 247–257. Ed. JCDaniels Jr. New York: Academic Press.

Sherman MI & Atienza-Samols SB 1978 In vitro studies on the sur-face adhesiveness of mouse blastocysts. In Human Fertilisation, pp179–183. Eds H Ludwig & P Tauber. Boston, MA: PGS.

Shi D & Kellems RE 1998 Transcription factor AP-2 gamma regulatesmurine adenosine deaminase gene expression during placentaldevelopment. Journal of Biological Chemistry 273 27331–27338.

Shi D, Winston JH, Blackburn MR, Datta SK, Hanten G & KellemsRE 1997 Diverse genetic regulatory motifs required for murineadenosine deaminase gene expression in the placenta. Journal ofBiological Chemistry 272 2334–2341.

Sidhu SS & Kimber SJ 1999 Hormonal control of H-type a(1-2) Fuco-syltransferase messenger ribonucleic acid in the mouse uterus.Biology of Reproduction 60 147–157.

Snow MH & Ansell JD 1974 The chromosomes of giant trophoblastcells of the mouse. Proceedings of the Royal Society B 187 93–98.

Soares MJ & Talamantes F 1983 Genetic and litter size effects onserum placental lactogen in the mouse. Biology of Reproduction29 165–171.

Soares M & Glasser SR 1987 Placental lactogen production and func-tional differentiation of rat trophoblast cells in vitro. Journal andReproduction and Fertilization 79 335–341.

Soares MJ, Faria TN, Roby KF & Deb S 1991 Pregnancy and the pro-lactin family of hormones: coordination of anterior pituitary, uter-ine and placental expression. Endocrinology Review 12 402–423.

Springer TA 1990 The sensation and regulation of interaction withthe extracellular environment – the cell biology of lymphocyteadhesion receptors. Annual Review of Cell Biology 6 359–402.

Sutherland AE, Calarco PG & Damasky CH 1988 Expression andfunction of cell surface extracellular matrix receptors in mouseblastocyst attachment and outgrowth. Journal of Cell Biology 1061331–1348.

Sutherland AE, Sanerson RD, Mayes M, Seibert M, Calarco PG,Bernfield M & Damsky DH 1991 Expression of syndecan, a puta-tive low affinity fibroblast growth factor receptor, in the earlymouse embryo. Development 113 339–351.

Sutherland AE, Calarco PG & Damsky CH 1993 Developmentalregulation of integrin expression at the time of implantation in themouse embryo. Development 119 1175–1186.

Telleria CM, Ou J, Sugino N, Ferguson S & Gibori G 1998 Theexpression of interleukin-6 in the pregnant rat corpus luteum andits regulation by progesterone and glucocorticoid. Endocrinology139 3597–3605.

Varmuza S, Prideaux V, Kothary R & Rossant J 1988 Polytenechromosomes in mouse trophoblast giant cells. Development 102127–134.

Wartiovaara J, Leivo I & Vaheri A 1979 Expression of the cell-surfaceassociated glycoprotein, fibronectin, in the early mouse embryo.Developmental Biology 69 247–257.

Werling U & Schorle H 2002 Transcription factor gene AP-2g essen-tial for early murine development. Molecular and Cellular Biology22 3149–3156.

Wilkinson DG & Nieto MA 1993 Detection of messenger RNA by insitu hybridisation to tissue sections and in whole mounts. Methodsin Enzymology 225 361–373.

Yamada KM 1983 Cell surface interactions with extracellularmaterials. Annual Review of Biochemistry 52 761–799.

Yamaguchi M, Ogren L, Endo H, Thordarson G, Bigsby G &Talamantes F 1992 Production of mouse placental lactogen-I andplacental lactogen-II by the same giant cell. Endocrinology 1311595–1602.



Yamaguchi T, Iwano M, Kubo A, Hirayama T, Akai Y, Horii Y,Fujimoto T, Hamaguchi T, Kurumatani N, Motomiya Y & Dohi K1996 IL-6 mRNA synthesis by peripheral blood mononuclear cells(PBMC) in patients with chronic renal failure. Clinical and Exper-imental Immunology 103 279–284.

Yan J, Tanaka S, Oda M, Makino T, Ohgane J & Shiota K 2001 Reti-noic acid promotes differentiation of trophoblast stem cells to agiant cell fate. Developmental Biology 235 422–432.

Zybina EV & Zybina TG 1996 Polytene chromosomes in mammaliancells. International Review of Cytology 165 53–119.

Received 19 December 2003First decision 4 March 2004Revised manuscript received 14 April 2004Accepted 24 April 2004



Documents

Trophoblast differentiation in vitro: establishment and