Implantation and placentation are key processes in mam- malian embryonic development. They physically connect the embryo to its mother and are critical for sufficient nutrient and gas exchange. The extraembryonic cell lineage is the first to differentiate in the developing conceptus, reflecting the importance of this cell type for the establish~ ment of fetal-maternal connections. During routine devel° opment, the outer layer of the blastocyst, the mural tro- phectoderm, begins to differentiate into primary trophoblast giant cells on day 5 of gestation (e5). These cells invade the uterine epithelium and penetrate deeply into the stroma. At the same time, the polar trophectoderm cells continue to proliferate and form the ectoplacental cone (EPC). On e7, the outer cells of the EPC begin to differen- tiate into secondary trophoblast giant cells. The invasion of the uterine stroma by these cells is critical for successful placentation ICross et aI., 1994).
Trophoblast invasion triggers secretion of proteinases
158
that degrade extracellular matrix molecules. Mouse tropho- blasts have been shown to synthesize and secrete serine proteinases (e.g., plasminogen activator), matrix metallo- proteinases (MMPs), and cysteine proteinases (cathepsin B and L). Invasion of the trophoblasts is a highly controlled process. The decidua restricts invasion by secreting protein- ase inhibitors in a specific spatial and temporal pattern. They include plasminogen inhibitor, tissue inhibitors of metalloproteinases (TIMPs), and cystatin C. Proteinases and proteJnase inhibitors have antagonistic functions in implantation and placentation which may be mirrored by the reciprocity of their expression patterns. The most striking examples are Mmp9, which is expressed in all terminally differentiated trophoblast giant cells, and TIMP3, present in directly adjacent cells of the differenti- ating maternal decidua (Alexander et aI., 1996). Similarly, both cathepsins B and L are localized to mature, invasive trophoblast giant ceils, whereas cystatin C is a major product of the decidualizing stroma (Afonso et aI., 1997). The critical role of proteinases and their inhibitors for
t rophoblas t invasion was also shown in vivo by adminis t ra- t ion of inhibi tors of MMPs or cys te ine proteil~ases and overexpression of TIMP1, which resul ts in abnormal em- bryonic deve lopment and u ter ine decidual iza t ion (Alex- ander et al., 1996; Afonso et al., 1997; Rinkenberger et al., 1997).
Dur ing tumor invas ion and metastas is , the degradat ion of basement membranes is often accompl ished by the same proteinases impl ica ted in implan ta t ion and normal tropho- blast invasion (Strickland and Richards, 1992; MacDougaU and Matris ian, 1995; Wilson et M., 1997; Edwards and Murphy, 19981. This is, for example, evident from increased levels of cathepsins that have been associated wi th t u m o r progression and aggressiveness in m a n y types of cancer (Garcia et al., 1990; Ren and Sloane, 1996; Kos and Lab, 1998; Herszenyi et al., 1999}. Uncont ro l led t rophoblas t invasion, l ike in chor iocarcinomas, resul ts in one of the most me tas ta t i c tumors k n o w n (Strickland and Richards, 1992). Therefore, i t is l ike ly tha t the same enzymat ic and cel lular mach ine ry is iuvolved in both processes. We here describe a c D N A subt rac t ion approach be tween invasive and noninvas ive stages of p lacenta l deve lopment which was a imed to e lucidate further factors involved in tropho- blast invasion. Several novel cDNAs tha t represent excel- len t candidates for having cri t ical roles in t rophoblas t invasion were identified.
MATERIALS AND METHODS
cDNA Subtraction Library and Screening
For cDNA subtraction, the upper parts of e8 conceptuses lcount~ ing day of vaginal plug as day 1 ), including the ectoplacental cone, invasive trophoblast giant cells, extraembryonic ectoderm, and chorion {abbreviated comprehensively as EPC} were collected from pregnant C57BL/6 inbred mice. The same mouse strain was used to isolate el8 placentas. These placentas were lifted from the uterus with blunt forceps and thus contained the maternally derived decidua in addition to labyrinthine trophoblast, spongiotropho- blast, and trophoblast giant cells. However, the yolk sac was removed. For RNA extraction, EPCs from 10 mice were pooled. Total RNA was extracted using Trizol Reagent (Gibco). cDNA synthesis and subtraction were performed with 500 ng of total RNA using a cDNA subtraction kit (Clontech). cDNA from e8 EPCs was used as tester and el8 placenta cDNA as driver, cDNA pools were digested with restriction endonuclease RsaI according to the manufacturer's protocol to enhance subtraction efficiency. After two rounds of PCR amplification, subtraction products were cloned into the pGEM T-Easy vector {Promegal. After transforma- tion, inserts of 384 randomly picked clones were PCR-amplified with M13-specific primers. PCR products were spotted twice onto Hybond N+ nylon membrane (Alnersham) to generate identical filter replicas, Complex hybridization was performed with a2p_ labeled, adaptor-free cDNA populations before and after subtrac~ tion. Plasmid DNA of 93 differentially hybridizing clones was sequenced from both ends on a LiCOR sequencer. All subtraction clones were called "Epcs" (ectoplacental cone*specific) in combi- nation with the further specifying ID.
5' and 3' Cloning of cDNA Ends (RACE) For the isolation of additional cDNA sequence of tile Epcs26 and
El)csSO transcripts, specific primers situated within the known sequence were synthesized: Epcs26-5', GGCATCTATTTGGCT- GACAG; Epcs2&5' nested, ACATCTGGGCACTATATGGG; Epcs26-3', CGACAGATTGGTTCATGGTC; F4)cs26-3' nested, TACCGAATGTGGCATCCGCA; Epcs5&5', TTGTCAGTGAA- GACTCCTCC; EpcsS&5' nested, GTCTGAGATAGCAGATG- G.AC; EpcsSO-3', AATAAGGGAAACCGTGCTGC; and Epcs504' nested, GAGTCTTCACTGACAAGAGC. RACE PCRs were per° formed with the e8 EPC-derived cDNA under identical conditions: 94°C for 3 ruin, followed by 40 cycles of 94°C for 5 s, $8°C for 10 s, and 68°C for 5 rain. For nested PCRs, 1/50 of the primary PCR products was used as template.
Sequence A n a l y s i s
Sequence similarity searches were performed in the NCBI nor> redundant sequence database (nr) and EST database by the Btastn program and protein sequence database (SwissProt} by the Blastx program. Sequences that did not show significant similarity to any of these entries (score <200) were designated as being unknown. For EST assembly, only mouse ESTs with a score >200 were chosm~. Assemb}y of database ESTs was performed using the Tigem N e t s equence u t i l i t i e s { h t t p : / / w w w . t f g e m . i t / L O C A L / sequtils.html). Sequences of overlapping and identical subtraction clo~es were aligned to deduce the consensus sequence. All clones analyzed by RNA in situ hybridization were investigated for con- served sequence blocks using the BEAUTY search program [http:// www.hgsc.bcm.tmc.edu/SearchLauncher/), Extended open tea& ing frames were checked by b lock search th t tp : j /www, hgsc.bcm.tmc,edu/SearchLauncher/) and aligned using the ClustalW progrmn at NIH.
RATA in Situ Hybridization (ISH) Differential and cell type~specific expression was investigated by
RNA ISH. Pregnant C57BL/6 females were sacrificed at e8 and el6. At el6, embryos and their associated placentas were collected, At e8, whole uterine regions containing three to four conceptuses wme dissected. Tissues were fixed in Serra's fixative at 0°C overnight and further processed for routine paraffin histology.
Plasmids of subtraction clones were linearized and used for in vitro transcription in both directions. For hybridization with the Pll probe, a plasmid containing an g00-bp eDNA fragment was used (Colosi et al., 1987}~ Labeling of riboprobes was performed using [dsStUTP. ISH was performed as described elsewhere (Adam et al., 1996; Hemberger et al., 1998]. Slides were dipped into K.5 photoemutsion (Ilford) and exposed for 18 days. Counterstaining was performed using hemalaun.
Northern Analysis Multiple mouse tissue Northern blots of adult tissues lClontech)
were hybridized according to the manufacturer's instructions. For placenta Northern blots, 25/~g total RNA was electrophoresed on a 1.5% formaldehyde agarose gel and blotted onto Hybond N+ nylon membrane tAmersham) as described previously tSambrook et al., 1989), Ten microliters of the SMART eDNA synthesis reaction used for subtraction were electrophoresed on a 1% agarose get and blotted onto Hybond N+ membrane using 0.4 M NaOH to
generate virtual Northern blots. Hybridization with 32P-labeled insert preparations of the subtraction clones was performed using Ctontech's ExpressHyb solution at 68°C for 1 h. Autoradiography of filters was performed on Fuji Medical X-ray films.
Genetic Mapping
For three cDNAs, restriction fragment length variants (RFLVs) between Mus rnusculus strain C57BL/6 (B6) and the M. spretus strain LS were found: Epcs50 PCR products were differentiated with SpeI, Epcs24 with Csp6I, and Epcs68 with NdeI. For Epcs26 no RFLV was detected after sequencing of B6 and LS PCR products. However, a single base substitution that could be visualized by SSCP analysis was identified. Primer sequences were Epcs5OF, AATAAGGGAAACCGTGCTGC; Epcs5OR, GTCTGAGATAG- CAGATGGAC; Epcs24F, CGACAGATTGGTTCATGGTC; Epcs24R, GTCGAGGTATCCTTTCTGC; Epcs68F, TTATCAAC- CATTCTGTTCTGGTAG; Epcs68R, GCTTAAGTTACAGTATC- CATCATGTC; Epcs26F, CGACAGATTGGTTCATGGTC; Epcs26R, GGCATCTATTTGGCTGACAG. Genetic mapping was carried out on the EUCIB-BSS (B6 × LS} × LS backcross panel (Breen et al., 1994; Rhodes et aL, 1998). DNA from strain M. spretus LS and from EUCIB backcross mice was obtained from the HGMP in Hinxton, UK. Evaluation of genotyping data and map construction were done using the MAPMAKER 3,0 software (Lander et aL, 1987).
A
-" a ;d *
d g-<* d "
Gapd
1.35 1.1 0.87
0.6
FIG, 1, Subtraction control experiments. In (A), total e8 EPC- specific cDNA populations before and after subtraction against e 18 placenta are shown. The use of marker DNA as tester mixed with and subtracted against control (con) eDNA proved efficient re- moval of nondifferential DNA. (B} Transcript levels of the house- keeping gene Gapd were substantially reduced in cDNA after subtraction (lanes 2 and 4). Marker sizes are given in kilobases.
RESULTS
Screening for Di f ferent ial ly Expressed Genes in Trophoblast
To identify genes involved in trophoblast invasion, cDNA subtraction was performed between e8 EPCs and mature el8 placenta. High efficiency of subtractive hybrid- ization was confirmed in a control subtraction experiment and by proving substantial reduction of Gapd transcript levels (Figs. 1A and 1B). A total of 384 clones were screened and 93 of them exhibited differential hybridization signals. Sequence analysis revealed that they represented 42 differ- ent cDNA fragments (Table 1). Twenty of them matched previously described genes in the nr database, and 1 was contained within a genomic BAC sequence. Strikingly, placental lactogen-1 (Pll), a gene predominantly expressed in trophoblast giant cells during midgestation, was repre- sented by 24/93 clones. Additional cDNAs isolated in the subtraction that are known to possess characteristic func- tions in extraembryonic tissue at this developmental stage were adrenomedullin (Yotsumoto et al., 1998) and the ETn transposon (Brfilet et al., 1985). The other 21 sequences did not match any entries in the nr database, and 16 of thern were recognized by sequences in the EST database. Differ- ential expression of three sequences represented only once, Epcs32 (Nasp), Epcs3 (AA467609), and Epcs21 (AA273200), was confirmed by RNA ISH. Thereby, high levels of Nasp mRNA were detected in e8 embryo and lower levels in EPC, extraembryonic ectoderm, chorion, and amnion (Figs. 2Aand 2B). Several of the as yet uncharacterized sequences
were represented by 3 or more individual clones and were preferentially analyzed.
Epcs26
The clone Epcs26 was identified three times within the subtraction library. The deduced consensus sequence of 1110 bp contained a poly(dA) stretch at its 3' end and exhibited homologies in a 306-bp region between bp 333 and 639 with the human X-chromosomal clone AC004383 (78% identity) in the nr database. The whole sequence of Epcs26, however, was recognized by 10 partially overlap- ping mouse ESTs, 8 of which were derived from 7.5-dpc ectoplacental cone libraries. One of these ESTs, AA409600, could be used for extension of the 5' end by an additional 117 bp. However, the 5' end of the eDNA identified by RACE PCR was exactly identical to that of the other 3 EST sequences that spanned the 5' region of the Epcs26 se- quence. This may indicate that the additional 117 bp are derived from a larger transcript (see below). The Epcs26 sequence did not exhibit any homology to known genes or proteins. Hybridization of the isolated eDNA fragment revealed specific expression in the parietal yolk sac endoderm of e8 conceptuses (Figs. 2C-2E). Hybridization signals were also observed in undifferentiated cells of the ectoplacental cone and a few trophoblast giant cells. The decidualizing stroma was completely devoid of transcripts. Differential expression could be proven, as only very -weak hybridization signals were detectable in the labyrinthine
Note. Sequences of Epcs24 and Epcs71, Epcs68 and Epcs70, as well as Epcs50 and Epcs5 could be combined and were called Epcs24, Epcs68, and Epcs50, respectively, thereafter.
trophoblast of e16 and e18 placentas. Northern analysis with total placenta RNA at e l0 and el8 also confirmed down-regulation of Epcs26 expression at later stages of development (Fig. 2F).
In e l0 placenta, two bands were identified by Nor thern analysis, a more prominent band at approximately 1.3 kb and a minor band at 2.6 kb (Fig, 2F}. This indicates that the extended sequence of Epcs26 may represent the full-length cDNA of the smaller transcript. At el8, only the 1.3&b band was present though at m u c h reduced levels. In the adult mouse, the 1.3-kb Epcs26 transcript was detected only in testis (Fig. 2G)
Epcs26 was mapped to the mouse X chromosome with. a LOD score of 10.8 to marker DXMitS. Using mul t ipoint linkage analysis, Epcs26 was placed between markers DXMit50 and DXMitl05 on the proxim.al X chromosome (Fig. 2H), which corresponds to the interval between 12.7 and 14.5 cM on the consensus map for this chromosome (Boyd et el., 1999).
EpcsSO
EpcsbO was identified by 11 individual clones. By 5' and 3' RACE, an additional 622 bp were isolated. The resulting assembled cDNA sequence of 3015 bp contained an ATG start codon, an extensive open reading frame, a polyadenyl- ation signal, and a poly(A) tail. Presence of both A TG start codon and the poly(A) tail indicates that the fulLlength sequence has been isolated. Five regions of the consensus sequence exhibited significant hom.ology to the murine Mpsl gent (Acc, No. L20315). These regions spanned bp 274-492 (89% identity), bp 538-843 (87% identity), bp 990- t269 (83% identity), bp t392---1762 (83% identity), and bp 1835-1918 (92% identity) (Fig. 3A). Both Mpsl and Epcs50 showed weak similarity to the same region of the murine perforin 1 precursor protein (P1; Acc. No. P10820). For Epcs50, this region of homology spanned bp 513M 121, matching to aa 130-352 of the P1 protein. This stretch of the P1 molecule exhibits similarity to various complem.ent component precursor proteins and contains the membrane attack region complex, h~ situ hybridization revealed high expression levels in a subset of trophoblast giant cells in the EPC and the parietal yolk sac (Figs, 3B and 3D). However, transcripts were conspicuously absent from other tropho- blast giant cells tFig. 3C). Expression was also detected in the parietal endodermal cells (Fig. 3D). No other cell type within the e8 conceptus exhibited hybridization signals. Moreover, Epcs50 was not expressed in any tissue of em- bryonic or extraembryonic origin at later stages of develop- ment (e16).
Nor thern analysis with el0 and e18 total placenta RNA showed a prominent transcript with a size of approximately 3 kb at el0. In addition, a m u c h weaker signal of approxi- mately 1.4 kb was detected on both e l0 and e l8 (Fig. 3E). In adult mouse, EpcsbO was strongly expressed in liver. Lower expression was observed in heart and several other tissues. However, transcript sizes differed from that observed in e l0 placenta as the Epcs50 band closely comigrated wi th t h e 4.4d<b size standard in adult tissues (Fig. 3F).
Two-point linkage analysis indicated linkage of Epcs50 to mouse chromosome 19 with a LOD score of 13.37 to
Copyright ~92000 by Academic Press. All rights of rcproductlon in any form reserved.
162 Hemberger et al.
- 2 . 6
- 1 . 9
- 1 , 4
- 0 . 9 5
G
4 . 4
2 . 4
1 . 3 5
Eoos2 + DXMitl05 +2.3 cM _+
1,4 cM
1,6 cM
FIG. 2. ISH with subtraction clones Epcs32 (Nasp) and Epcs26 on e8 conceptuses. (A and B) Bright-field {A) and dark-field (B) i l lumination after mRNA ISH with Epcs32 indicating high levels of Nasp expression in the embryo (e) and lower levels in the ectoptacental cone (epc), chorion (oh), and amnion (a) Hybridization signals are absent from decidual tissue (de). Note that light reflections of fetal blood (bl) are not caused by silver grains (bar, 200 bcm), (C and D) The abembryonic pole of the conceptus is shown, including the mesometrial zone of the
marker Fth. Multipoint linkage analysis localized EpcsSO to the proximal part of this chromosonre between D19Mit22 and D 19Mit28. No recombination with markers D19Mit79, D 19Mit31, and D 19Pas6 was detected (Fig. 3 G). The follow- ing marker order and recombination distances were evalu- ated: centromere-DI9Mit22-3.5 + 2.0 cM-Epcs50-2.3 + 1.6 cM-D19Mit28-telomere.
Epcs24 and Epcs68 Epcs24 and Epcs68 were represented hy nine and six
clones, respectively, which could be assembled after data- base search revealed an overlapping EST for each eDNA, AA408092 for Epcs68 and AA408231 for Epcs24. Sequence assembly resulted in 996~ and 1395~bp cDNAs for Epcs24 and Epcs68, respectively. An RsaI site was detected in the overlapping region of both ESTs that presumably led to eDNA fragmentation. Both contigs shared 72% overall similarity and exhibited extensive ORFs. In addition, the Epcs68 sequence contained a translation start, a TAA stop codon, two putative polyadenylation signals, and a 3' poly(A) stretch. Searches against the SwissProt database revealed homology of both sequences to cathepsin L precur- sor proteins. Epcs2d showed highest similarity (68 %) to rat cathepsin L precursor/major excreted protein (Ace. No. P07154), Epcs68 was recognized by human cathepsin L precursor (Ace. No. P07711) with 71% similarity. Both Epcs24 and Epcs68 shared the characteristic features of cysteine proteinases in that active site residues and com- mon blocks of eukaryotic thiol proteases were highly con- served (Fig. 4A). The 333-aa open reading frame of Epcs68 fitted exactly within the size range of cathepsin L precursor proteins. Northern analysis was carried out using subtraction-derived clones of 890 (Epcs24) and 666 bp (Epcs68). Analysis of el0 and el8 total placenta RNA revealed silmlar transcript sizes of 1.45 kb for Epcs24 and Epcs68. Thus, Epcs68 most likely represented the full- length eDNA, and only minor parts of the 3' end were missing in the Epcs24 sequence.
Differential expression of Epcs24 and Epcs68 was shown by RNA ISH. Very strong expression was observed in trophoblast giant cells of the parietal yolk sac and EPC (Figs. 5A, 5B, 5F, and 5H}. Fo~ Epcs24 and Epcs68, expres- sion patterns were comparable in that only a subset of trophoblast giant cells abundantly expressed these mRNAs. In contrast, some giant cells showed only low transcript levels or no expression at all. Control hybridization with
the giant cell marker Pll on adjacent sections marked giant cells not expressing Epcs24 or Epcs68 (Figs. 5C and 5G). In mature placentas at el6, widespread expression was con- fined to trophoblast cells within the labyrinthine region (Figs. 5D and 5E), whereas hybridization signals were absent from spongiotrophoblast, trophoblast giant cells, and de- cidua. The widespread expression of Epcs24 and Epcs68 excludes the possibility that these genes are expressed in the few trophoblast giant cells that colonize the labyrinth at e16 (Hall and Talamantes, 1984). Compared to e8, expres- sion levels were significantly lower at this stage of devel- opment. An absence of transcripts was observed for all enrbryonic tissues at e8 and el6. In contrast, Epcs68 tran- scripts were detected in adult tissues by Northern analysis (Fig. 4C). A prominent band of 1.45 kb was present in liver; minor amounts of transcripts were observed in testis, kidney, and heart; and only very faint bands could be detected in brain and lung. With Epcs24, several weak bands were detected in liver and testis. Only in heart was a more prominent band observed at 1.35 kb (Fig. 4B).
Both Epcs2d and Epcs68 were mapped to mouse chromo~ some 13 at high LOD scores (12.6 and 10.9) to marker D13MitlO. Linkage distances to flanking map anchors were calculated (Fig. 4D) and used to integrate the chromosomal positions of Epcs24 and Epcs68 with the consensus map of chromosome 13 (Stephenson and Lueders, 1999). As shown in Fig. 4E, Epcs24 and Epcs68 closely coloealize with two more genes encoding cysteine proteinases described and mapped previously (Deussing et al., 1997; Tisljar et al., 1.999).
DISCUSSION
In the present study, we describe the construction of a eDNA subtraction library that was aimed to reveal tran- scripts up-regulated in invasive trophoblast. Several genes and sequences of unknown function that exhibited striking expression differences in extraembryonic tissues between early and late stages of development were isolated. Twenty- six percent of the clones analyzed represented Pll, a gene known to be highly expressed in trophoblast giant cells around midgestation (Colosi et al., 1987), thus serving as an excellent proof for the efficiency of our subtractive hybrid- ization. Several sequences isolated from the subtraction library were represented by more than one clone. As repre- sentation frequencies are correlated with the degree of
decidua (de), chorion (ch), parietal endoderm (pe), and trophoblast giant cells (gi). Dark-field illumination in D reveals predominant expression of Epcs26 in parietal endoderm cells (bar, 50/,m). (E) Higher magnification Of the parietal yolk sac shown in C and D showing higher density of silver grains in the parietal endodexm (pc 1 {bar, 20 ~m). (F) Northern analysis of Epcs26 on total el0 (lane 1) and el8 (lane 2) placenta RNA. Two transcripts were detected, the more prominent at approximately 1.3 kb and a weaker band at 2.6 kb. Marker sizes are given in kilobases, (G) Northern analysis of various adult tissues demonstrates expression of Epcs26 only in testis. Marker sizes are given in kilobases (sin, skeletal muscle). (H) Chromosomal mapping of Epcs26 to the mouse X chromosome on the EUCIB backcross. The number of recombination events found, and calculated genetic distances in centimorgans including standar d error are indicated.
FIG. 3. (A) Schematic representation of regions of homology of Epcs50 to the murine Mpsl gene (Acc. No. L20315). Regions of homology between the givell base pairs are shaded. (B and D) Expression of Epcs50 in e8 conceptuses is highly abundant in parietal endoderm cells and a subset of trophoblast giant cells. The arrows point to giant cells exhibiting strikingly different expression levels. The core of the ectoplacental cone containing the more undifferentiated cells does not exhibit any hybridization signal (bar~ 200 bin1). (C) Ctose-up view of a nonexpressing giant cell marked by the arrowhead (bar, 25 ~m). (E) Northern analysis of Epcs50 on total el0 (lane 1) and e18 (lane 2) placenta RNA. A prominent transcript is detected at approximately 3 kb in el0 RNA only, a much weaker signal is present at 1.4 kb. Marker sizes are given in kilobases. (F) Northern analysis of various adult tissues indicates predominant expression in liver. Marker sizes are given in kilobases (sin, sketetat muscle}. (G} Chromosomal mapping of Epcs50 to mouse chromosome 19 using the EUCIB backcross. Number of recombination events and calculated genetic distances in centimorgans including standard error are indicated, amz, antimesometrial zone; ee, embryonic ectoderm; epc, ectoplacental cone; gi, trophoblast giant cells; mz, mesometrial zone; pe, parietal endodenn; ue, uterine epithelium.
differential expression, these sequences were mos t l ike ly to be p r edominan t ly expressed in e8 EPC. In fact, none of the cDNAs ident i f ied by mul t ip l e clones was found to be false posit ive. Unambiguous differential expression was also
observed for three cDNAs ident i f ied only once among the clones analyzed. The subt rac t ion cloning ident i f ied genes previously no t k n o w n to be expressed in ex t raembryonic t issues. This was shown for the nuclear autoant igenic
Novel Genes Expressed in Invasive 7h'ophoblast 165
sperm, protein NASP, a histone-binding protein identified in primary spermatocytes and round spermatids of human testis (Welch and O'Rand, 1990; O'Rand et al., 1992). Consequently, the number of differentially expressed clones identified in our subtraction approach is comparable to that obtained in studies using similar experimental strategies (Jin et aL, 1997; Heller et al., 1998; Kuang et aL, 1998; Shimono et aI., 1999).
The Epcs26 eDNA was shown to be predominantly expressed in the endoderm layer of the parietal yolk sac at early developmental stages. The parietal endoderm cells are characterized by the synthesis of large amounts of base- ment membrane components that constitute Reichert's membrane (Hogan et al., 1994). Temporal and spatial dis- tribution of Epcs26 transcripts might indicate the involve- lnent of this gene in the establishment, differentiation, and function mainly of parietal endoderm cells. Interestingly, similar sites of expression in extraembryonic tissues have been described for the throlnbin receptor thrombomodulin (Weiler-Guettler et al., 1996), whose targeted mutation causes an early lethal phenotype (Healy et al., 1995). Due to lack of any homology to known genes and proteins, there is at present no evidence for the functional role of EPCS26 in early mouse development°
Epcs50 exhibited intervals of sequence homology to the murine Mpsl gene. The Mps l gene, initially termed Mpg-1, was isolated in an approach to isolate genes preferentially expressed in mature macrophages (Spilsbury et al., 1995). Mpsl and EpcsSO exhibited weak homology to perforins, which might indicate a distant relationship to these pro- teins. Perforin is a pore-forming protein crucial for the lytic action of cytotoxic T cells and natural killer cells (Kfigi et aL, 1996). Interestingly, amino acid homology was confined to the region critical for the lytic activity of perforin. During pregnancy, perforin is abundantly expressed in granular metrial gland cells of the decidualizing uterine stroma (Parr et al., 1990; Zheng et aI., 1991) and a role in maternal defensive response against trophoblast invasion has been ascribed to this protein (Parr et aL, 1990). The Mpsi-related sequence Epcs50 was shown to be transcribed in trophoblast giant cells and parietal endoderm cells. It is tempting to speculate that Epcs50 has the same lytic capacity as perforin and contributes to destruction of ma- ternal tissues during invasion of giant cells into the uterine stroma.
Cathepsin L, which is related to Epcs2d and Epcs68, is a major lysosomal cysteine proteinase produced by mouse placenta and fibroblasts (Nilsen-Hamilton et al., 199l). Cathepsin L is generated by rapid autocatalysis of its precursor major excreted protein (MEP) at low pH (Gal and Gottesman, 1986a; Mason et aI., 1987). MEP was first identified as one of the most abundant proteins secreted by mouse fibroblasts upon malignant transformation (Gottes- man, 1978). Its synthesis is also induced by growth factors, tumor promoters, and cyclic AMP (Dong et al., 1989; Prence et al., 1990; Troen eta]., 1991). Extracellular matrix proteins have been shown to be among the substrates of
cathepsin L (Gal and Gottesman, 1986b) which may facili- tate invasion of tumor cells. Interaction of cathepsins with the mannose 6-phosphate/insulin-like growth factor recep- tor 2 (IGF2R) (Dong at a]., 1989; Dong and Sahagian, 1990; Prence at al., 1990; Stearns et aI., 1990) leads to its inter- nalization (Edwards and Murphy, 1998). As IGF2R acts as a sink for IGF2 (Czech, 1989; Filson et al., 1993), cathepsins support the mitogenic activity of this important growth factor. Cathepsins have been suggested to modulate im- mune responses and may be involved in avoiding maternal immune rejection of fetal tissues (McCoy et ~d., 1988; Riese eta]., 1998). In addition, members of the cathepsin protein- ase family have been shown to induce apoptosis (Moallem and Hales, 1995; Deiss et al,, 1996). In conclusion, cathep- sins have important roles in tissue invasion and cell prolif- eration, that is, processes which are also essential for trophoblast invasion. Therefore, it is likely that EPCS24 and EPCS68 contribute to the invasive properties of tontine trophoblast giant cells. This notion is supported by the fact that both EPCS24 and EPCS68 possess the functional pro- teolytic domains of cathepsins. The presence of two other cathepsins (cathepsin L and J) in the same region of mouse chromosome 13 suggests that these genes have arisen by gene duplication from a common ancestral gene (Deussing et aL, 1997; Tisljar et al., 1999). It is noteworthy that all of these genes are expressed in extraembryonic tissues. How- ever, expression of cathepsin L is not restricted to tropho- blast giant cells, indicating that these genes are regulated differentially.
A remarkable phenomenon of all clones that were ana- lyzed in the present study was their expression in only a subset of trophoblast giant cells at e8. This is in striking contrast to genes like PI1 (Colosi et al., 1987), M m p 9 (Alexander et aL, 1996; Das et aL, 1997), or Hand l (Cross et al , 1995) that are expressed in all trophoblast giant cells at this stage. To date, only a few genes that exhibit different expression levels within, the same cell population in ex- traembryonic tissues have been described. This was re- ported for thrombomodulin expression in parietal endoderm and secondary giant cells (Weiler-Guettler et aL, 1996), for Mrj in giant cells (Hunter et aI., 1999), and, most recently, for G c m l in trophoblast cells of the chorion and the labyrinth at later stages (Basyuk et aI., 1999). Thus, specific cell populations formerly regarded as uniform might be subdivided into several subpopulations due to their cell-specific expression patterns. Presumably, the dif- ferential gene expression reflects various differentiation stages of single cells within a given cell population. Alter- natively, diverse functions of subpopulations may account for the heterogeneous expression patterns.
Our future efforts lie in the identification of additional mRNAs that are highly specific to the routine invasive trophoblast. In addition, the functions of the corresponding proteins will have to be elucidated using in vivo and in vitro approaches.
atL prec. human z CatL prec. rat zz Ca tLp rec . bov ine CatL2prec . human z
EPCS68 6 z EPCS24 s 1 CatL prec. mouse s z CatL prec. human s z CatL prec. rat s z Ca tLprac , bov ine s z Ca tL2prec . human 61
EPCS68 z 2 z EPCS24 1 z 9 CatL prec. mouse ~ 2 CatL prec. human z 2 z CatL prec. rat z 2 z CatLDrec. bov ine z 2 z Ca tL2prec , human 1 21
EPCS68 1 8 z EPCS24 1 7 CatL prec. mouse a s z CatL prec. human z 81 CarL prec. rat z 8 z Ca tLp rec . bov ine z 8 CatL2prec. human 1 8
EPCS68 2 3 9 EPCS24 2 37 CarL prec. mouse 2 3 9 CatL prec. human 2 3 9 C a r l prec, rat 2 3 9 Ca tLprec . bov ine 2 4 o CatL2 prec, human 2 4 o
EPCS68 2 99 EPCS24 2 97 CatL prec, m o u s e 2 9 9 CatL prec, human 2 9 s CatL prec. rat 2 9 9 C a r l prec. b o v i n e 3 0 o CatL2 prec, human 3 o 0
B
Epcs24
u ) ~ =. (9 C CEp >, e. =_ ® =
CS68 a v " E => _= } ~a ~
D
D 13Mit10 ~
Epcs24 Epcs68
EUCIB-BSS D13Mit47' MMU 13
9/88 10.2 cM ¢ 3 . 2 c M
4/90 4.4 cM ± 2.2 cM
22/8"7 25 .3 cM • 4 .7 c M
C
-4.4
"2.4
-1.35
E Ctsl i
D13Mi t lO
_ Ctsi" ~-pcs2;~ Epcs68 ~
Consensus Map t [ MMU13 D13Mit47
30.0 31 .0
36 .0 3 7 . 1 3 9 . 7
55.0
FIG. 4. (A) Protein similarit ies between amino acids encoded by the open reading frames in EPCS24 and EPCS68 and cathepsin L precursor proteins of different species. Black shading represents identical amino acids, indicating homology between the sequences. Arrows point to highly cot~served active site residues of eukaryotie cysteine proteinases. Boxed amino acids mark the four conserved blocks in cysteine proteinases. (B) Norther11 blot analysis of several adult mouse tissues reveals no prominent transcript of Epcs24 except for a faint band at
FIG. 5. (A and B) Expression of Epcs24 in e8 embryos is highly abundant in a subset of trophoblast giant cells in the parietal yolk sac and at the outside of the ectoplacental cone. Arrows mark giant cells exhibiting different expression levels. No expression is detected in decidual and embryonic tissues (bar, Z00/~m). (C) Expression of Pli on a section consecutive to that shown in B demarcates all differentiated giant cells (bar, 200/~m). (D and E) At el6, Epcs24 transcripts are detected exclusively in labyrinthine trophoblasts, although at a much lower level. The arrowhead points to a cluster of spongiotrophoblast cells within the labyrinth, where complete absence of hybridization signals is observed in the dark-field i l lumination (bar, 200 i~m). (F and H) ISH with Epcs68 on e8 conceptuses. Arrows point to trophoblast giant cells exhibiting differential hybridization signals (bar, 200 i~m). (G) In sitn hybridization with the PII probe on a section serial to F, indicating the presence of differentiated trophobtast giant cells (bar, 200/~m). a, amnion; amz, antim.esometrial zone; bl, maternal blood cells; ee, embryonic ectoderm; epc, ectoplacental conei gi, trophoblast giant cells; mz, mesometrial zone; pe, parietal endoderm; ue, uterine epithelium.
1.35 kb in heart. (C) Epes68 is predominantly expressed in adult liver with a single transcript of approximately 1.45 kb. Lower expression is detectable in testis, kidney, and heart, and only faint bands are present in lung and brain. Marker sizes are given in kilobases (sin, skeletal muscle 1. (D 1 Chromosomal mapping of Epcs24 and Epcs68 to mouse chromosome 13 on the EUCIB backcross. Observed recombination events to flanking EUCIB anchor markers and calculated genetic distances including standard errors are indicated. (El Position of Epcs24 and Epcs68 on the consensus map of chromosome 13. Two additional genes, CtsI and Ctsj, encoding cysteine proteinases colocalize with Epcs24 and Epes68 in the same chrom.osomat interval.
We are grateful to Drs. H.-G. Nothwang and S. Lenzner for helpful discussions and advice and to Professor H. H. Ropers for his continuous support. We are also grateful to Dr. J. C. Cross for critically reading the manuscript. This work was supported by the Max-Planck Gesellschaft.
REFERENCES
Adam, G., Cui, H., Miller, S., Flare, F., and Ohlsson, R. (1996). Allele-specific in situ hybridization (ASISH) analysis: A novel technique which resolves differential allelic usage of HI 9 within the same cell lineage during human placental development. Development 122, 839-847.
Arouse, S., Romagnano, L., and Babiarz, B. (1997). The expression and function of cystatin C and cathepsin B and cathepsin L during mouse embryo implantation and placentation. Develop- meat 124, 3415-3425.
Alexander, C. M., Hansell, E. J., Behrendtsen, O., Flannery, M. L., Kishnani, N. 8., Hawkes, S. P., and Werb, Z. (1996). Expression and function of matrix metalloproteinases and their inhibitors at the maternal-embryonic boundary during mouse embryo im- plantation. Develol)ment 122, 1723-1736.
Basyuk, E., Cross, J. C., Corbin, J., Nakayama, H., Hunter, P,, NaitOumesmar, B., and Lazzarini, R. A. (1999). Murine Gcrnl gene is expressed in a subset of placental trophoblast cells. Dev. Dyn. 214, 303-,311.
Boyd, Y., Denny, P., Masson, W., Reed, V., and Elliott, R. (1999). Mouse X chromosome. Mature. Genome 10, 961.
Breen, M. (1994). Towards high-resolution maps of the mouse and human genomes--A facility for ordering markers to 0.1 cM resolution. Hum. Mol. Genet. 3, 521-627.
Brfilet, P., Condamine, H., and Jacob, F. (1985). Spatial distribution of transcripts of the long repeated ETn sequence during early mouse embryogenesis. Prec. Natl. Acad. Sci. USA 82, 2054- 2058.
Colosi, P., Talamantes, F., and Linzer, D. L (1987). Molecular cloning and expression of mouse placental lactogen I comple- mentary deoxyribonucleic acid. Mol. EndocrinoL 1, 767-776.
Cross, J. C., Flannery, M. L., Blanar, M. A., Steingrimsson, E., Jenkins, N. A., Copeland, N. G., Rutter, W. J., and Werb, Z. (1995). Hxt encodes a basic helix-loop-helix transcription factor that regulates trophoblast cell development. Development 121, 2513---2523.
Cross, J. c., Werb, Z., and Fisher, S. J. (1994). hnplantation and the placenta: Key pieces of the development puzzle. Science 266, 1508-1518.
Czech, M. P. (1989). Signal transmission by the insulinqike growth factors. Ceil 59, 235-238.
Das, S. K., Yano, 8, Wang, J., Edwards, D. R., Nagase, H., and Dey, S. K. (1997). Expression of matrix metalloproteinases and tissue inhibitors of metalloproteinases in the mouse uterus during the peri-implantation period. Dev. Ge~et. 21, 44-54.
Deiss, L. P., Galinka, H., Berissi, H., Cohen, O., and Kimchi, A. (1996)~ Cathepsin D protease mediates programmed cell death induced by interferon-% Fas/APO-1 and TNF-c~. EMBO J. 15, 386t-3870.
Deussing, J., Roth, W., Rommerskirch, W., Wiederanders, B., yon Fignra, K., and Peters, C. (1997). The gene of the lysosomal
cysteine proteinases cathepsin B, H, L, and S map to different mouse chromosomes. Mature. Genome 8, 241-245.
Dong, J. M., Prence, E. M., and Sahagian, G. G. (1989). Mechanism for selective secretion of a lysosomal protease by transformed mouse fibroblasts. L Biol. Chem. 264, 7377-7383.
Dong, 1- M., and Sahagian, G. G. {1990]. Basis for low affinity binding of a lysosomat cysteine protease to the cation- independent mannose 6-phosphate receptor. J. Biol. Chem. 265, 4210-4217.
Edwards, D. R., and Murphy, G. (1998). Cancer. Proteases-- lnvasion and more. Nature 394, 527-528.
Filson, A. J., Louvi, A., Efstratiadis, A , and Robertson, E. J. (1993). Rescue of the T-associated maternal effect in 1nice carrying null mutations in tgf-2 and Igf2r, two reciprocally imprinted genes. Development 118, 731-736.
Gal, S., and Gottesman, M. M. (1986a). The major excreted protein of transfonned fibroblasts is an activable acid-protease. [ Biol. Chem. 261, 1760-1765.
Gal, S.1 and Gottesman, M. M. (1986b). The major excreted protein (MEP 1 of transformed mouse cells and cathepsin L have similar protease specificity. Biochem. Biophys. Res. Commun. 139, 156-1.62.
Garcia, M., Deroq, D., Pujol, P., and Rochefort, H. (1990). Overex- pression of transfected cathepsin D in transformed cells increases their malignant phenotype an d metastatic potency. Oncogene 5, 1809-1814.
Gottesman, M. M. (1978). Transformation-dependent secretion of a low molecular weight protein by murine fibroblasts. Prec. Natl. Acad. Sci. USA 75, 2767-2771.
Hall, J., and Talamantes, F. (1984). Immunocytochemical localiza- tion of mouse placental tactogen in the mouse placenta. J. His- tochem. Cytochem. 32, 379-382.
Healy, A. M., Raybum, H. B., Rosenberg, R. D., and Weiler, H. (1995). Absence of the blood-clotting regulator thrombomodulin causes embryonic lethality in mice before development of a functional cardiovascular system. Proc. Natl. Acad. 8ci. USA 92, 850-854.
Heller, S, Sheane, C. A., laved, Z., and Hudspeth, A. J. (1998). Molecular markers for cell types of the inner ear and candidate genes for hearing disorders. Prec. Natl. Acad. ScL USA 95, 11400-1 I405.
Hemberger, M., Redies, C., Krause, R., Oswat G J., Walter, J., and Fundele, R. H. (1998}. H19 and Igf2 are expressed and differen- tially imprinted in neuroectoderm~derived cells in the mouse brain. Dev. Genes EvoL 208, 393-402.
Herszenyi, L., Plebani, M., Carraro~ P., De Paoli, M., Roveroni, G, Cardin, R., Tulassay, Z., Naccarato, R., and Farinati, F. (1999). The role of cysteine and serine proteases in colorectal carcinoma. Cancer 86, 1135-1142.
Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (t994). "Manipulating the Mouse Embryo (A Laboratory Manual)." Cold Spring Harbor Laboratory Press, New York.
Hunter, P. J., Swanson, B. J., Haendel, M. A., Lyons, G. E., and Cross, J. C. (1999). Mrj encodes a DnaJ-related co-chaperone that is essential for routine placental development. Development 126, 1247--1258.
Jin, H., Cheng, X., Diatchenko, L., Siebert, P. D., and Huang, C.-C. (1997). Differential screening of a subtracted cDNA library: A method to search for genes preferentially expressed in multiple tissues. Biotechniques 23, 1984~1086.
Kagi, D., Ledermann, B,, Burki, K., Zinkernagel, R. M., and Hen- gartner, H. (1996). Molecular mechanisms of lymphocyte-
Nove l Genes Expressed in h~vasive Trophoblas t 169
mediated cytotoxicity and their rote in immunological protec~ tion and pathogenesis in vivo. AT?nu. Rev. hnmu17ol. 14, 207-- 232.
Kos, J., and Lah, T. T. (1998). Cysteine proteinases and their endogenous inhibitors: Target proteins for prognosis, diagnosis and therapy in cancer. Oz~col. Rep. 5, 1349-1361.
Kuang, W. W., Thompson, D. A., Hoch, R. V., and Weigel, R. J. (1998). Differential screening and suppression subtractive hy~ bridization identified genes differentially expressed in an estro- gen receptor-positive breast carcinonqa celt line. Nucleic Acids Res. 26, 1[t6-1123.
Lander, E. S, Green, P., Abraharnson, J., Barlow, A., Daty, M. J,, Lincoln, S. E., and Newburg, L. (1987). Mapmaker, an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. GePomdcs 1, 174-181.
MacDougalt, J. R., and Matrisian, L. M. (1995). Contributions of tmnor and stromal matrix metaltoproteinases to tmnor progres- sion, invasion and metastasis. Cancer Metastasis Rev. 14, 351- ,362.
Mason, R. W., Gal, S., and Gottesman, M. M. (1987}. The identifi~ cation of the major excreted protein (MEP) from a transformed mouse fibroblast cell line as a catalytically active precursor form of cathepsin L. Biochem. J. 248, 449-454.
McCoy, K., Gal, S., Schwartz, R. H., and Gottesman, M. M. (1988). An acid protease secreted by transformed cells interferes with antigen processing. ]; Cell Biol. 106, 1879-t884.
Moallem, S. A., and Hales, B. F. (1995). Induction of apoptosis and cathepsin D in limbs exposed in vitro to an activated analog of cyclophosphamide. Teratology 52, 3-14.
Nilsen-Hamilton, M., Jang, Y. J., Delgado, M., Shim, J, K., Bruns, K., Chiang, C. P., ~ang, Y., Parfett, C. L., Denhardt, D. T., and Hamilton, R. T. (1991). Regulation of the expression of mitogen~ regulated protein (MRP; proliferin) and cathepsin L in cultured cells and in the murine placenta. Mol. Cell. Endocri±~ol. 77, 115-122.
O'Rand, M. G~, Richardson, R. T., Zimmerman, L. J., and Widgren, E. E. (1992). Sequence and localization of human NASP: Conser- vation of a Xenopus histone-binding protein. Dev. Biol. 154, 37-44.
Parr, E. L., Young, L. H., Parr, M. B., and Young, J. D. (1990). Granulated metfial gland cells of pregnant mouse uterus are natural killer-like cells that contain perforin and serine esterases. J. Immm~ol. 145, 2365-2372.
Prence, E. M., Dong, J. M., and Sahagian, G. G. {19901. Modulation of the transport of a lysosomal enzyme by PDGF. J. Cell Biol. 110, 319-~-326.
Ren, W. P., and Stoane, BI F. (1996). Cathepsins B and D in breast cancer. In "Mammary Tm'uor Cell Cycle, Differentiation and Metastasis" (R. B. Dickson and M. E. Lippman, Eds.), pp. 326~ 352. Academic Press, San Diego.
Rhodes, M., Straw, R., Fernando, S., Evans, A., Lacey, T., Dearlove, A., Greystrong, J., Walker, J., Watson, P., Weston, P., Kelly, M., Taylor, D., Gibson, K., Mundy, C., Bourgade, F., Poirier, C., Shnon, D., Brunialti, A . L., Montagutelli, X., Gudnet, J.-L.,
Haynes, A., and Brown, S. D. (t998). A high-resolution microsat- ellite map of the mouse genome. Genome Res. 8, 531-542.
Riese, R. J., Mitchdl, R. N., Villadangos, J. A., Shi, G. P., Pahner, J. T., Karp, E. R., De Sanctis, G T., Ploeghl H. L., and Chapman, H. A. (1998) Cathepsin S activity regulates antigen presentation and immunity./. Clin. h~vest. 101, 2351~2363.
Rinkenberger, J. L., Cross, J. C., and Werb, Z. (1997). Molecular genetics of implantation in the mouse. Dev. Gez)et. 21, 6-20.
Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). "Molecular Cloning: A Laboratory Manual." Cold Spring Harbor Laboratory Press, New York.
Shimono, A., and Behringer, R. R. (1999}. Isolation of novel cDNAs by subtractions between the anterior mesendoderm of single mouse gastrula stage embryos. Dev. Biol. 209, 369-380.
Spilsbury, K, O'Mara, M. A., Wu, W. M., Rowe, P. B., Sylnonds, G., and Takayama, Y. (1995}. Isolation of a novel macrophageo specific gene by differential eDNA analysis. Blood 85, 1620- 1629.
Stearns, N. A., Dong, J. M, Pan, J. X., Brenner, D. A., and Sahagian, G. G. (1990). Comparison of eathepsin L synthesized by normal and transformed ceils at the gene, message, protein, and oligo- saccharide levels Arch. I~iochem. Biophys. 283, 447-457.
Stephenson, D. A., and Lueders, K. K. (1999). Mouse chromosome 13. Mature. Genorne 10, 954.
Strickland, S., and Richards, W. G. (1992). Invasion of the tropho- blasts. Cell 71, 355--357.
Tisljar, K., Deussing, J., and Peters, C. (1999). Cathepsin J, a novel mufine cysteine protease of the papain family with placenta- restricted expression. FEBS Lett. 459, 299 --304.
Troen, B. R., Chauhan, S. S., Ray, D., and Gottesman, M. M. (1991). Downstream sequences mediate induction of the mouse eathep- sin L promoter by phorbol esters. Cell Growth Differ. 2, 23-31.
Weiler-Guettler, H., Aird, W. C., Rayburn, H., Husain, M., and Rosenberg, R. D. (1996). Developmentally regMated gene expres- sion of throlnbolnodulin in postirnplantation mouse embryos. Devalopment 122, 2271-2281.
Welch, J. E., and O'Rand, M. G. {1990). Characterization of a spernbspecific nuclear autoantigenic proteim II. Expression and localization in the testis. Biol. Repro& 43, 569-578.
Wilson, C. L., Heppner, K. J., Labosky, P. A., Hogan, B. L., and Matrisian, L, M. (1997). Intestinal tumorigenesis is suppressed in mice lacking the metalloproteinase matfilysin. Proc. Natl. Acad. Sci. USA 94p 1402-1407.
Yotsumoto, S., Shimada, T., Cui, C. Y., Nakashima, H., Fujiwara, H., and Ko, M. S. (19981. Expression of adrenomedullin, a hypotensive peptide, in the trophoblast giant cells at the embryo implantation site in mouse. Dev. Biol. 203~ 264-275.
Zheng, L. M., Liu, C. C., Ojcius, D. M., and Young, J. D. (1991). Expression of lymphocyte perforin in the mouse uterus during pregnancy. J. Cell Sci, 99, 317-323.
