heparan sulfate is involved in attachment and spreading of mouse embryos in vitro

DEVELOPMENTAL

BIOLOGY

123,401-410

(1087)

Heparin/Heparan Sulfate Is Involved in Attachment and Spreading of Mouse Embryos in Vitro M. C. FARACH,’

J. P. TANG,

G. L.

DECKER,

AND

D. D.

CARSON~

Department of Biochemistry and Molecular Biology, M. D. Anderson Hos&al and Tumor Institute, 6723 Bertner Avenue, Box 117,Houston, Texas 77080 Received May 13, 1986;accepted in revised form May 18, 1987 The involvement of embryonic cell surface proteoglycans in the attachment and outgrowth of cultured mouse embryos has been investigated. Several lines of evidence indicate that periimplantation stage blastocysts express heparin/heparan sulfate proteoglycans on their cell surfaces that can mediate embryo attachment and trophoblast outgrowth on a variety of matrices. First, in the presence of soluble heparin, the rate at which embryos attach and outgrow on laminin, fibronectin, or monolayers of uterine epithelial cells is reduced considerably. In the case of fibronectin, the rate of outgrowth in the presence of the heparin is slower than in the presence of the Arg-Gly-Asp-Ser-containing peptide that is recognized by a fibronectin receptor. Embryos also attach and exhibit a limited ability to outgrow on platelet factor IV, a heparin binding protein that does not possess the additional binding domains of laminin or fibronectin. Attachment on platelet factor IV is inhibited by heparin. Second, cell surface digestion of attachment-component embryos with heparinase, but not chondroitinase ABC, slows the rate of outgrowth on tissue culture plates in the presence of serum. Third, selective staining for sulfated molecules on the trophectoderm surface of periimplantation stage embryos indicates that such molecules are abundant and uniformly distributed on these cell surfaces. Last, heparin/heparan sulfate proteoglycans are detected as major cell surface components of embryos using vectorial labeling with lactoperoxidase and Nan% Collectively, these data indicate that heparin/heparan sulfate-bearing molecules have o 1987 Academic PWS, IIIC. a direct role in attachment and outgrowth of implantation stage blastocysts.

INTRODUCTION

During early pregnancy in mammals, embryos must successfully implant in the endometrium of the uterus. The interactions occurring between the cell surfaces of the embryo and the uterine epithelial cells during the periimplantation period are both componentially complex and temporally regulated (Sherman and Wudl, 1976). Although considerable effort has been spent to identify the specific molecules expressed by the blastocyst that can mediate attachment and subsequent outgrowth, the precise mechanism by which embryos implant has remained obscure. A number of in vitro models for implantation have been devised. In all of these, preimplantation stage embryos are removed from uteri and introduced to a variety of substrata on which they may implant. One model that should contain molecules the embryo would attach to in vivo uses monolayers of uterine epithelial cells maintained in serum-free medium (Sherman and Salomon, 1975). Limitations of this system include the difficulty in visualizing early stages of implantation and an in’ Present address: Department of Physiology and Molecular physics, Baylor College of Medicine, Houston, TX 77030. ’ To whom correspondence should be addressed.

Bio-

401

ability to distinguish between the embryonic and maternal cell types in the analysis of metabolic labeling patterns. These two problems are minimized in another model system that allows embryos to implant on tissue culture dishes in the presence of fetal calf serum (Sherman and Salomon, 1975). It is important to note that attachment factors that can support implantation in vitro may not be available to the embryo during implantation in vivo. For example, neither fibronectin (Wartiovaara et al., 1979; Grinnell et al., 1982) nor laminin or collagen (Leivo et ab, 1980) appear to be distributed at apical surfaces of uterine epithelia; however, all of these molecules have been reported to support embryo attachment and outgrowth in vitro (Jenkinson and Wilson, 1973; Armant et aZ.,1986). The number of defined molecules with multifunctional domains that will support attachment and outgrowth suggests that embryos contain multiple adhesion systems, although perhaps only one is normally utilized during the initial stages of implantation in vivo. Consequently, it is essential to demonstrate the existence of these attachment factors at the apical surface of the uterine epithelium if these factors are proposed to be involved in early stages of implantation in utero. Analysis of glycoconjugates synthesized by mouse embryos during development from the blastocyst stage 0012-1606/W Copyright All rights

$3.00

0 1987 by Academic Press, Inc. of reproduction in any form reserved.

402

DEVELOPMENTAL BIOLOGY

through the early somite stages revealed that postimplantation embryos synthesize various glycosaminoglycans including heparin/heparan sulfate (Heifetz et al., 1980). Furthermore, it has been reported that mouse embryos possess heparan sulfate proteoglycans on their surfaces starting as early as the two-cell stage (Dziadek et ak, 1985). Because of these reports, and because the heparin/heparan sulfate-binding proteins, laminin and fibronectin, support trophoblast outgrowth in vitro (Armant et aL, 1986), we investigated the involvement of heparin/heparan sulfate proteoglycans in attachment and outgrowth of mouse blastocysts. We compared implantation in several in vitro systems including epithelial cell monolayers, serum-containing medium, and dishes coated with defined extracellular matrix components in serum-free medium. In all cases, interactions involving heparin/heparan sulfate proteoglycans appear to function as important periimplantation stage adhesion systems involved in attachment and outgrowth. MATERIALS

AND

METHODS

Materials. CF-1 Mice (males and females) were obtained locally (Timco, Houston, TX). Pregnant mare serum gonadotropin was obtained from Calbiochem (LaJolla, CA), and human chorionic gonadotropin was from Sigma Chemical Co. (St. Louis, MO). Tissue culture media were purchased from Grand Island Biological Co. (Grand Island, NY) and Irvine Scientific (Santa Ana, CA), and tissue culture supplies from Falcon Division, Be&on-Dickinson and Co. (Oxnard, CA). Polyvinyl alcohol (PVA) was obtained from Eastman-Kodak Co. (Rochester, NY). Chromatography resins, heparin, chondroitinase ABC, and pronase were obtained from Sigma. Heparinase and heparan sulfate were from Miles Laboratories (Naperville, IL). The Arg-Gly-Asp-Ser synthetic peptide was purchased from Peninsula Laboratories (Belmont, CA). Fibronectin was purchased from Bethesda Research Laboratories, Inc. (Gaithersburg, MD), and laminin was purchased from Collaborative Research, Inc. (Lexington, MA). Platelet factor IV was prepared as described by Laterra et al. (1983b) from outdated human platelets obtained from the Gulf Coast Regional Blood Center (Good Texas Blood; Houston, TX). [6-3H]Glucosamine (25 Ci/mmole), Nalz51(carrier-free), and H235S04 (carrier-free) were purchased from ICN Chemical and Radioisotope Division (Irvine, CA). Embryo culture and labeling. Embryos were obtained from superovulated CF-1 females as described (Armant et ak, 1986). Blastocysts were flushed from uteri on the morning of the fourth day after mating. Blastocyst stage embryos were cultured in a 37°C humidified incubator in CMRL 1066 plus 50 units/ml penicillin and 50 pug/ml

VOLUME 123. 1987

streptomycin supplemented according to experimental design (see individual experiments). Isotopes were dried down in a Speed Vat concentrator (Savant Instrument, Inc., Hicksville, NY) and resuspended in the appropriate labeling medium. To label with H235S04, sulfate-free RPM1 1640 medium containing 10% (v/v) dialyzed fetal calf serum, 1 mCi/ml H235S04,2.5 units/ml penicillin, and 2.5 pg/ml streptomycin sulfate was used. The streptomycin sulfate served as the source of nonradioactive sulfate in these experiments. In some cases, double labeling was performed by adding 200 &i/ml of [3H]glucosamine to this same medium. Embryos were incubated in this medium overnight after at least a day in sulfate-containing medium. Under these conditions, we found that the embryos remained healthy as indexed by their ability to form outgrowths with the same efficiency as control embryos cultured in sulfate-replete medium. Although the amount of label incorporated was substantially lower, embryos labeled with 35S04in sulfate-replete medium produced sulfated molecules with the same ion exchange characteristics as those produced by embryos cultured in low sulfate medium (see below). Radioiodinatim of embryo cell surfies. The procedure used was essentially that of Morrison (1981). Briefly, embryos were rinsed several times in ice-cold phosphatebuffered saline (PBS). The embryos were radioiodinated in 200 ~1 of ice-cold PBS containing 0.2 mCi of Nalz51 (carrier-free) and 20 pug/ml of lactoperoxidase. The incubation was maintained on ice for 30 min and 2 ~1 of 1 mM hydrogen peroxide was added at the outset and an additional 2 ~1 was added subsequently at 5-min intervals to support the reaction. At the end of the incubation the embryos were transferred through several drops of PBS containing 1 mM NaI. The embryos then were extracted in guanidine-Chaps and desalted on Sephadex G-50 as described below. The void volume (macromolecular fraction) from Sephadex G-50 was used for further analysis. Culture of uterine epithelial cells. Primary cultures of uterine epithelial cells were prepared as described previously (Dutt et ab, 1986). Coating of dishes. To coat dishes, 50 ~1 of a 1 mg/ml solution of laminin, fibronectin, or platelet factor IV in PBS was added to wells of a 96-well tissue culture plate. After incubation overnight at 3’7°C the wells were extensively washed with PBS and used immediately for embryo outgrowth assays. Synchronous blastocyst populations. There is substantial variability in the time at which blastocysts attach and outgrow on an artificial substratum following their removal from the uterus. In addition, it can take as many as 3 days for embryos to attach and outgrow in vitro even when they appear to be morphologically normal. For this reason, it was critical to both synchronize and

FARACH ET AL.

HeparidHeparan

increase the rate of outgrowth in order to perform accurate kinetic studies. This was done by culturing embryos for 48 to 72 hr in CMRL 1066 containing 1% (w/ v) PVA. Under these conditions, embryos emerge normally from their zonae pellucidae but are unable to attach and outgrow on the tissue culture plates. These “delayed” blastocysts are extremely sticky and efficient manual transfer can be accomplished only through the use of siliconized pipets. At the start of the experiment, attachment-competent embryos are transferred to implantation medium containing 10% (v/v) fetal calf serum. To ensure that only hatched embryos were scored in kinetic assays, two alternate protocols were used. In the first, embryos were treated briefly with pronase to dissolve the zona pellucida at the time they were harvested from the uterus (Sherman and Salomon, 1975). In the second, embryos were allowed to hatch normally and those that had not hatched by the start of the experiment were discarded. Using either procedure, embryos typically began to attach and outgrow within 2-4 hr after transfer to serum-containing medium. Assays of trophoblast outgrowth. All assays of attachment and outgrowth were performed by visual inspection of embryo populations at various times after transfer to 96-well plates prepared according to experimental design. When inhibitors such as heparin (0.5 mg/ml) or the synthetic peptide, Arg-gly-Asp-Ser (0.5 mg/ml), were tested, they were added to the medium prior to the embryos. For kinetic assays of the rate of outgrowth following cell surface digestion with various enzymes, blastocysts were synchronized as described above. They then were divided into four groups of approximately equal numbers. Two of these groups were incubated for 1 hr at 37°C in the presence of either heparinase (125 mu/ml) or chondroitinase ABC (250 mu/ml) plus a mixture of protease inhibitors (Carson et ak, 1985). The third group was incubated for 1 hr in the presence of protease inhibitors only. The last group was treated for 5 min with 0.5% (w/v) pronase. At the start of the experiment, the embryos from all four groups were washed sequentially through three drops of CMRL 1066 containing 4 mg/ml bovine serum albumin (BSA) to remove any active enzyme and then transferred to wells of serum-containing medium. Analysis of glycoconjugates. For analysis of total metabolically labeled material, embryos and medium were removed from 96-well plates with a siliconized Pasteur pipet. The wells then were washed twice with Hanks’ balanced salt solution (HBSS), and both washes were combined with the embryos. To detach outgrown embryos and to extract material possibly attached to the plates, the wells were washed with 4 M urea and 1% Triton X-100 in 50 M Tris-acetate, pH 7.0, and the final wash was combined with the previous ones. Complete

&&ate

in Embrgo

Attachment

403

removal of cellular material from the wells was verified by visual inspection. Embryos labeled with H235S0, were rinsed several times with HBSS and then extracted for 1 hr at room temperature in 1 ml of 4 M guanidine HCl, 1% (w/v) Chaps, 20 mM Tris-acetate, pH 7.0,25 mM EDTA, and a mixture of protease inhibitors (Dutt et al, 1986). The extract was desalted on a 0.75 X 22-cm column of Sephadex G-50 (fine) eluted with 4 M urea, 0.1% (w/v) octylglucoside, 20 mM Tris-acetate, pH 7.0. The material eluting in the void volume of Sephadex G-50 was subjected to anion exchange liquid chromatography on a 0.5 X 5-cm Mono Q column (Pharmacia, Uppsala, Sweden) eluted with 0.5 M urea, 20 mM Tris-acetate, pH 7.0, 0.01% (w/v) octylglucoside, and a gradient of 0 to 4 M NaCl. The column was maintained at room temperature and was eluted at a flow rate of 1 ml/min. The back pressure was approximately 550 psi. Certain Mono Q fractions were chromatographed on a 1 X 30-cm column of Superose 12 (Pharmacia) eluted with 2 M guanidine-HCl, 20 mM Tris-acetate, pH 7.0 and 0.01% (w/ v) octylglucoside. This column was maintained at room temperature and eluted at a flow rate of 0.7 ml/min. The back pressure was approximately 350 psi. The chromatography system (Beckman Instruments, Berkeley, CA) was driven by two Model 1OOApumps controlled by a Model 421A controller and equipped with a Model 163 flow through variable wavelength detector interfaced with a Model 427 integrator. Heparinase digestions of column fractions were performed by addition of heparinase (0.1 units/ml) in the presence of a battery of protease inhibitors (Dutt et ak, 1986). After incubation for 3 hr at 37”C, heparinase was added again and the incubation was continued for 1 additional hr. Samples then were chromatographed on Superose 12 as described above. The conditions for nitrous acid treatment were those of Hart and Lennarz (1978). Pronase digestions were performed as described previously (Dutt et al, 1986). Histochemical staining of sulfated molecules. Blastocysts were rinsed three times in Hanks’ balanced salt solution containing 0.1% (w/v) bovine serum albumin and then fixed for 1 hr at room temperature with 3% (v/v) glutaraldehyde, 0.3 M magnesium chloride, 25 mM sodium acetate, pH 5.6, and 0.05% (w/v) alcian blue 8GX (Scott and Dorling, 1965). The embryos were rinsed in the latter buffer minus the alcian blue and postfixed in this same buffer containing 1% (w/v) osmium tetroxide on ice for 30 min. The embryos were rinsed with phosphate-buffered saline and dehydrated. They were impregnated with Spurr resin by incubation in a 1:l (v/v) mixture of absolute ethanol and Spurr resin for 30 min at room temperature and then with 100% Spurr resin for 1 hr. Finally, fresh Spurr resin was added and the

404

DEVELOPMENTAL BIOLOGY

incubation continued at 60°C overnight. Sections were then prepared for and examined by electron microscopy as described (Decker and Lennarz, 1979).

VOLUME 123, 1987

decreased dramatically in the presence of heparin. In the case of fibronectin, the inhibition of outgrowth formation observed in the presence of heparin was even greater than that observed in the presence of the synRESULTS thetic peptide, Arg-Gly-Asp-Ser, which contains the cell Outgrowth of mouse embryos on various matrices. Pre- recognition sequence of fibronectin (Ruoslahti and vious reports have indicated that mouse blastocysts can Pierschbacher, 1986). Incubation with heparin plus the outgrow on a number of matrices including the extra- peptide did not produce any more inhibition than that cellular matrix proteins laminin and fibronectin (Ar- observed with heparin alone. Heparin also severely inmant et aZ.,1986). One common feature of laminin and hibited attachment on platelet factor IV. After 24 hr, fibronectin is their ability to bind heparin/heparan sul- less than 10% of the embryos cultured in the presence fate (Aplin and Hughes, 1982, Woodley et al, 1983). As of heparin in wells coated with platelet factor IV atshown in Table 1, embryos attached and outgrew to a tached to the substratum and, instead, remained freesimilar extent on plates coated with laminin or fibro- floating. Perhaps of greatest significance was the obsernectin as they did on epithelial cell monolayers or in vation that heparin considerably reduced the rate at serum-containing medium. We found that another hep- which embryos produced visible outgrowths on epithelial arin-binding protein, platelet factor IV, effectively sup- cells (Fig. 2C). Other glycosaminoglycans including ported embryo attachment; however, outgrowth for- commercial heparan sulfate did not inhibit the rate of mation was limited and irregular on this substrate (see outgrowth formation on any of these matrices. It was Fig. 1). By comparison to outgrowths formed on fibro- very difficult to identify the early stages of outgrowth nectin, on laminin, or in the presence of serum (Fig. lA), on uterine lawns using phase-contrast microscopy. the outgrowths formed on platelet factor IV in serum- Nonetheless, within 3-5 hr after noticeable outgrowth began, there were considerable differences in the extent free medium were usually very small and asymetrical (Fig. lB-1D). It seems, therefore, that platelet factor IV of outgrowth and displacement of uterine cells by emsupported blastocyst attachment much more effectively bryos in the presence versus the absence of heparin. By than trophoblast outgrowth. The other substrata tested 72 hr, these differences were largely overcome and the majority of embryos had formed trophoblast outsupported both processes well. We also examined the ability of soluble heparin to growths. Heparinase treatment slows implantation rate. To furinhibit embryo attachment and outgrowth in these systher test the idea that cell surface heparin/heparan sultems. As shown in Figs. 2A and 2B, we found that the fate may be involved in attachment and outgrowth of rate of outgrowth on both laminin and fibronectin was mouse blastocysts, attachment-competent embryos were subjected to cell surface digestion with a variety of enzymes including heparinase. As shown in Fig. 3, only TABLE 1 digestion with heparinase significantly slowed the rate HEPARIN/HEPARAN SULFATE BINDING PROTEINS SUPPORT at which attachment-competent blastocysts attached OUTGROWTH OF MOUSE EMBRYOS IN VITBO and outgrew on dishes in the presence of serum. This result was seen consistently in four separate experiPercentage embryos Total no. displaying outgrowth ments. In contrast, chondroitinase ABC-treated embryos Coating agent of embryos or attachmenta never outgrew later than untreated controls. Because these digestions were performed in the presence of a None 0 54 variety of protease inhibitors it is unlikely that protease Platelet factor IV l31+ 10 59 contamination of heparinase was responsible for this Laminin 52 83f6 Fibronectin 92f6 effect. Furthermore, a 5-min exposure of hatched em26 Uterine epithelial bryos to pronase only affected the rate of implantation cell monolayer &I+8 137 in one of the four experiments. We also attempted to Serum 231 9324 assess the effect of heparinase on the rate of embryo outgrowth formation on lawns of uterine epithelial cells 0 Synchronous attachment-competent blastocysts were transferred from CMRL 1066 containing 1% (w/v) PVA to dishes coated with the in serum-free medium. We were unable to detect an inindicated agents, and the number of embryos displaying outgrowths hibition of outgrowth formation in the heparinasewere scored after 72 hr. The data show the averages and ranges obtreated embryos; however, as noted above, it was very tained from at least two separate experiments in each case. In the difficult to see the early stages of outgrowth formation case of platelet factor IV, values are shown for the percentage of emin these assays. As shown in Fig. 3, embryos recover bryos that displayed attachment because outgrowth formation on this substrate was limited and irregular (see Fig. 1 and text). from heparinase digestion within a few hours. Conse-

FARACH ET AL.

Heparin/Heparan

Sulfate in E-0

Attachment

405

FIG. 1. Trophoblast outgrowths produced by embryos attached to I ilatelet factor IV. (A) Blastocyst outgrowth in the presence of slerum. Embryos cultured in serum-free medium on fibronectin- or laminin-co: ited dishes displayed similar morphology (not shown). (B-D) Blasl tocyst (B, C) or outgrowths on dishes coated with platelet factor IV. In almost all cases, outgrowths were either highly asymmetric or irregular extremely small (D). Incubations with platelet factor IV-coated dishes 1were performed in serum-free medium. Photographs were taken at 72 hr after transfer to these wells.

quently, it seems likely that embryos had recovered from the heparinase treatment (see Fig. 3) before we could begin to distinguish outgrowth formation on cell layers. Identi$cation of sulfated surface of periimplantation

macromolecules on the cell stage embryo. As one ap-

proach to determine if proteoglycans were distributed on embryonic cell surface, we used a histochemical staining technique successfully used by others (Scott and Dorling, 1965; Bolender et al., 1980; Reale et al., 1983) to

detect sulfated molecules on the surfaces of periimr )lantation stage embryos. We utilized a cationic dye, allcian blue, which at specific pH and divalent cation conicentrations selectively stains sulfated molecules (Scott ; and Dorling, 1965). As shown in Fig. 4, the stain was uniformly deposited at the apical trophectodermal SUI *face on all sides of the embryo including polar (A) and mural (B and C) trophectoderm. These observations indic ated that sulfated molecules were abundant at these cell sur-

406

DEVELOPMENTAL BIOLOGY A

0

60 40 20 2 ; 2 a

O

L6

60

.-F $ 40 E: i5 g 20 t i f

Cl

0 60 40 20 0

0 ‘AL C

L. 0

5

10

Time lhoursl

FIG. 2. Heparin inhibits embryo spreading on various matrices. Groups of zonae-free blastocyst stage embryos (approximately 30 in each case) were transferred to individual wells of 96-well plates that had been coated with the indicated agents as described under Materials and Methods. Rates of embryo outgrowth then were monitored by visual inspection. Values at each time point are the composite of trophoblast outgrowths recorded by three independent observers. (A) Laminin-coated dishes either with (@) or without (0) 0.5 mg/ml of heparin added to the incubation medium. (B) Fibronectin-coated dishes with no additions (0), with 0.5 mg/ml of heparin in the medium (O), with 0.5 mg/ml of the synthetic peptide of the sequence Arg-GlyAsp-Ser added (A), or with 0.5 mg/ml each of heparin and the synthetic peptide (A). (C) Monolayers of uterine epithelial cells in serum-free medium in the presence (0) or absence (0) of soluble heparin at 0.5 mg/ml. Note that in these experiments the rate of outgrowth formation on the artificial matrices laminin and fibronectin was similar to that observed on the epithelial cells.

faces, i.e., at areas of potential embryo-substratum contact. Hepatin/heparan sulfate is expressed on the trophectodermal cell surface. It was important to demonstrate that heparin/heparan sulfate-bearing molecules were among the sulfated components of the embryo cell surface. To do this we used the lactoperoxidase method (Morrison, 1974) to selectively radioiodinate the cell surface, protein-containing components of the embryos. Labeled embryos were extracted and analyzed by liquid chromatography as described under Materials and Methods. As shown in Fig. 5A, most of the radioiodinated material was detected in the 3-4 M NaCl elute of the anion exchange column. Analysis of radioiodinated embryos by sodium dodecyl sulfate-polyacrylamide gel

VOLUME 123, 1987

electrophoresis and autoradiography indicated that almost all the labeled material remained in the stacking gels (data not shown). Consequently, we used molecular exclusive liquid chromatography in dissociative solvents to analyze these components. This procedure indicated that the major radioiodinated components were of relatively high molecular weight, ca. 250,000 (Fig. 5B). When this same material was treated with heparinase prior to application to the Superose 12 column, its migration pattern was shifted to a lower molecular weight, ca. 110,000. Consequently, it appeared that heparin/heparan sulfate proteoglycans were major cell surface components of periimplantation embryos. Periimplantation stage blastocysts synthesize heparin/ heparan sulfate proteoglycans. We investigated whether we could identify newly synthesized heparin/heparan sulfate-bearing proteoglycans in total solubilized extracts prepared from periimplantation stage blastocysts. Embryos were labeled with H235S04overnight and then extracted in guanidine-Chaps as described above. After desalting, the extracts were analyzed by anion exchange chromatography. As shown in Fig. 6A, the majority of the labeled material elutes at 3-4 M NaCl, a migration position similar to that observed for the iodinated material from the trophectoderm surface. When the high salt fraction eluted from the anion exchange column was analyzed by molecular exclusion liquid chromatography under dissociative conditions it migrated as a single peak ranging from h& 180,000 to 650,000 with a median mo-

60 -

7

t II , 4 12 6 20 Hours after Transfer to Implantation Medium

FIG. 3. Heparinase treatment inhibits the rate of in vitro attachment and outgrowth. Synchronized attachment-competent blastocysts were digested with various enzymes as described under Materials and Methods and then transferred to individual wells of 96-well plates containing medium supplemented with 10% (v/v) fetal calf serum. The rate of attachment and outgrowth was monitored as described in the legend to Fig. 2. A range of 23 to 32 embryos was used in each group in these experiments. (0) Control embryos incubated in the presence of protease inhibitors (see text) for 1 hr; (0) embryos treated for 5 min with predigested pronase; (A) embryos treated 1 hr with chondroitinase ABC in the presence of protease inhibitors; (A) embryos treated similarly but with heparinase for 1 hr in the presence of protease inhibitors. In four separate experiments, only heparinase treatment delayed the rate of trophoblast outgrowth. All incubations were at 37°C.

FARACH ET AL.

HeparidHeparan

Sulfate

in Embryo

Attachment

FIG. 4. Alcian blue staining reveals the presence of sulfate groups on the trophectoderm surface. Periimplantation stage embryos were stained with alcian blue and prepared for electron microscopy as described under Materials and Methods. (A) An area at the polar trophectoderm (note the underlying cells of the inner cell mass). (B and C) Areas of the mural trophectoderm. In all cases, a dense uniform pattern of staining is evident at the apical trophectodermal cell surface.

lecular weight of 350,000 (Fig. 6B). Incubation of the material in this peak with pronase before application to the Superose 12 column quantitatively shifted its migration position to a significantly lower molecular weight (30,000), indicating a linkage to protein. Subsequently we found that 80-90% of the material in the high salt fraction could be degraded by nitrous acid to a form that migrated between the partially to fully included volume of Sephadex G-50, i.e., less than 10,000 M, (Fig. SC), suggesting that most of the ?‘S04 was in the form of heparin or heparan sulfate. We also found that heparinase digestion of 35S04-labeled embryos released 20 to 50% of the embryo-associated radioactivity, indicating that some of this material was exposed at the cell surface (data not shown). Collectively, these data indicated that proteoglycans containing heparin/heparan sulfate

chains were major sulfated macromolecules by periimplantation stage mouse embryos.

synthesized

DISCUSSION

In this study, we investigated the possibility that periimplantation stage blastocysts expressed heparin/heparan sulfate on the trophectoderm cell surface that could participate in cell surface interactions during implantation. Because implantation, like other cell-substratum adhesion processes, is exceedingly complex, we tested the role of heparinjheparan sulfate in attachment and outgrowth in vitro using a series of diverse approaches, as well as using a number of in vitro implantation model systems. We found that addition of heparin to embryo culture medium severely inhibited the rate of embryo outgrowth

408

DEVELOPMENTAL BIOLOGY VOLUME123,1987

6 4

2:

7

2

1

5! x c

0

z

0 4

8

12

16

20

24

26

32

factor IV formed only tight focal contacts and exhibited altered morphology. Our results imply that during attachment and outgrowth mouse embryos also may form multiple contacts with complex matrix molecules such as laminin or fibronectin, but that at least one of these required contacts involves heparin/heparan sulfate. This heparin/heparan sulfate-dependent association apparently is sufficient for attachment and prerequisite for outgrowth formation. A second approach used to assess the role of trophectoderm cell surface heparin/heparan sulfate in attachment and spreading in vitro was to test whether digestion of the cell surfaces of implantation-competent blastocysts with heparinase could delay attachment and

3 10

14

18

22

26

30

34 2

4-i

FIG. 5. Analyses of radioiodinated

trophectoderm cell surface components of periimplantation stage embryos. Twenty-five periimplantation stage embryos were radioiodinated by the lactoperoxidase procedure and extracted with guanidine-Chaps as described under Materials and Methods. Macromolecules were selected from the extract by desalting on Sephadex G-50 as described under Materials and Methods. (A) Profile obtained when desalted fractions were applied to a Mono Q anion exchange liquid chromatography column and then eluted with NaCl as detailed in the text. The material eluting at 3-4 M NaCl was pooled and then analyzed by molecular exclusion liquid chromatography on a Superose 12 column (B). (B) Elution patterns are presented for high salt fractions either untreated (0) or predigested with heparinase (0) as described under Materials and Methods. Molecular weights were estimated by comparison with the indicated markers. These markers were blue dextran (I’,), thyroglobulin (670K), catalase (230K), phosphorylase b (92K), carbonic anhydrase (30K), aprotinin (7K), and potassium dichromate (VJ.

L u

NI s!

1

21

0

0

z

5

10

20

15

30

25

Mmutes

;p

3

1

16 “0 ,I I I

i?iY ii

i

-After

Pronese

HN02

12

16 12 8

4

4 x

0

formation on laminin, fibronectin, and monolayers of uterine epithelial cells. Furthermore, tissue culture dishes coated with platelet factor IV, a protein that has been demonstrated to recognize both heparin and heparan sulfate (Handin and Cohen, 1976; Laterra et al., 1983a,b), supported embryo attachment and spreading. Although embryos were able to attach to platelet factor IV, they did not spread as effectively on this substrate as they did on fibronectin and laminin. This may be due to the additional binding domains of the latter two molecules. Similar results were reported for mouse fibroblasts adhering to various heparan sulfate-binding substrata including platelet factor IV (Laterra et ah, 1983a). In these studies, it was suggested that cells adhering to fibronectin were able to form both close contacts and tight focal contacts, whereas those spreading on platelet

10

14

18

22 Minutes

26

30

34

20

40 Frectm

60 No.

FIG. 6. Analysis of sulfate-labeled components of embryos cultured in vitro. (A) Populations of cultured blastocysts were incubated in low sulfate medium containing H2%01 (see Materials and Methods), and the labeled macromolecular fraction was analyzed by Mono Q anion exchange chromatography under conditions identical to those described in Fig. 5. (B) The high salt eluate was chromatographed on a Superose 12 column under dissociative conditions described under Materials and Methods. The molecular weight markers indicated are V,, blue dextran; 1, thyroglobulin (670 kDa); 2, catalase (230 kDa); 3, 1,G (160 kDa); 4, bovine serum albumin (68 kDa); 5, ovalbumin (43 kDa); 6, carbonic anhydrase (30 kDa); 7, cytochrome c (15 kDa); 8, aprotinin (6 kDa); and Vi, potassium dichromate. The profiles are shown for material obtained before (0) or after (0) pronase digestion. (C) Sephadex G50 chromatography of high salt fraction obtained as described in (A) following degradation with nitrous acid as described under Materials and Methods. The V, and V, markers were blue dextran and potassium dichromate, respectively.

FARACH ET AL.

HeparidHeparan

outgrowth. The strategy of synchronizing populations of blastocysts by continuous culture in medium that can support growth, but not attachment and outgrowth, proved to be very useful in assessing the effect of various treatments such as cell surface digestion with heparinase. When kinetic experiments are attempted on embryo populations just as they are harvested from uteri of pregnant animals, the inherent asynchrony in the time at which embryos are capable of forming trophoblast outgrowths obscures rate effects and makes interpretation difficult. On the other hand, we found that although delayed embryo populations showed some variability in the lag time between transfer to serumcontaining medium and the onset of outgrowth formation, groups of embryos consistently performed these processes with internal synchrony. This internal group synchrony made it feasible to make relatively rapid measurements of the kinetics of outgrowth. These assays were performed on blastocysts attaching and outgrowing on plates in the presence of serum. In these experiments, we consistently found that digestion of the trophectodermal surface with heparinase, but not with chondroitinase ABC, would significantly delay the rate of implantation by 2 to 4 hr. Given the rapid kinetics of outgrowth formation once initiated and the observed ability of the embryos to recover from such cell surface digestions within several hours after treatment, we find the 2- to 4-hr delay of outgrowth to be highly significant. In order to demonstrate directly the presence of sulfated molecules on the surface of the trophectoderm as well as to examine the distribution of these groups, we stained trophectodermal cell surfaces with an electrondense cationic dye, alcian blue. We took advantage of the observation that, under specific buffer conditions, such dyes selectively stain sulfated molecules (Scott and Dorling, 1965). This technique has been employed by others as evidence for the presence of cell surface glycosaminoglycans (Bolender et al, 1980; Reale et al, 1983). In fact, we found that this dye uniformly stained the trophectodermal surface of periimplantation stage embryos. These observations indicate that such molecules are present at areas of potential embryo-uterine contact. We also radioactively tagged the cell surface components of periimplantation stage embryos and analyzed those components biochemically. It was found that the majority of the cell surface components labeled by this technique have the characteristics of heparin/heparan sulfate proteoglycans. Consequently, it appears that heparin/heparan sulfate proteoglycans are available at the embryo’s surface to interact with the various substrata that the embryo may encounter. The proteoglycans detected by labeling with either 35S04or ‘%I had apparent molecular weights of 250,000 to 300,000.The radiolabeled material obtained following

Sulfate in Embryo Attachment

409

heparinase digestion had a molecular weight of approximately 110,000,while the glycosaminoglycan-containing glycopeptide obtained following exhaustive pronase digestion had a molecular weight of approximately 30,000. Consequently, it appears that heparinase digestion resulted in the loss of four to six glycosaminoglycan chain equivalents. Since residual carbohydrates may remain following heparinase digestion, we cannot say more regarding the molecular weight or number of core proteins present. The patterns of glycosaminoglycan degradation obtained following nitrous acid digestion indicate the presence of a large number of O-sulfated residues, a characteristic more similar to heparin than heparan sulfate (Gallagher and Walker, 1985). Furthermore, we observed that although the embryonic proteoglycans were sensitive to nitrous acid or heparinase digestion, they were resistant to heparitinase digestion (data not shown). From a biological standpoint, we found that commercial preparations of heparin, but not heparan sulfate, effectively inhibited embryo attachment and outgrowth on a variety of matrices. Thus, it appears that the structural features of heparin that are less abundant in heparan sulfate are of key importance in these embryonic interactions. The extent to which the embryonic glycosaminoglycans resemble heparin remains to be demonstrated. Nonetheless, these observations prompt the use of the term heparin/heparan sulfate proteoglycans in reference to these embryonic glycoconjugates. In conclusion, this study presents the first evidence that heparinjheparan sulfate proteoglycans are synthesized by mouse embryos during the periimplantation period and can mediate attachment and outgrowth on a variety of substrata including uterine epithelial cells. This finding is particularly interesting in light of the accumulating evidence that implicates heparan sulfate proteoglycans in a number of cell-substratum interactions (Aplin and Hughes, 1982; Laterra et al., 1983a,b; Woods et al., 1985). The ubiquitous interactions of heparan sulfate may explain, in part, the ability of blastocyst stage embryos to attach to and outgrow upon a variety of matrices and cell types in vitro and in a number of diverse tissues in vivo (Sherman and Salomon, 1975; Sherman and Wudl, 1976). In view of the promiscuous ability of embryos to attach to and invade so many tissues, the interesting question remains as to how the mammalian uterus prevents implantation during most of the reproductive cycle. The authors express their gratitude to Mrs. Gertrude Willis for her careful typing of this manuscript. While this work was performed, M.C.F. was the recipient of a post-doctoral fellowship (No. HD 06509) from the National Institutes of Health. This work was supported by grants awarded to D.D.C. from the March of Dimes (No. l-958) and the American Cancer Society (No. BC-503).

410

DEVELOPMENTALBIOLOGY

VOLUME 123, 1987

REFERENCES

LATERRA, J., NORTON,E. K., IZZARD,C. S., and CULP, L. A. (1983a). Contact formation by fibroblasts adhering to heparan sulfate-binding APLIN, J. D., and HUGHES,R. C. (1982). Complex carbohydrates of the substrata (fibronectin or platelet factor 4). Exp. Cell Res. 146, 1527. extracellular matrix: Structures, interactions and biological roles. Biochim Biophys. Acta 694,375-418. LATERRA,J., S&BERT,J. E., and CULP,L. A. (1983b).Cell surface heparan ARMANT,D. R., KAPLAN, H. A., MOVER,H., and LENNARZ,W. J. (1986). sulfate mediates some adhesive responses to glycosaminoglycanThe effect of hexapeptides on attachment and outgrowth of mouse binding matrices, including fibronectin. J. Cell BioL 96, 112-123. embryos in vitro: Evidence for the involvement of the cell recognition LEIVO, I., VAHERI, A., TIMPL, R., and WARTIOVAARA,J. (1980). Aptripeptide Arg-Gly-Asp. Proc. Nat1 Acad. Sci USA 83,6751-6’755. pearance and distribution of collagens and laminins in the early BOLENDER,D. L., SELIGER,W. G., and MARKWALD,R. R. (1980). A hismouse embryo. Dev. BioL 76,100-114. MORRISON,M. (1974). The determination of exposed proteins on memtochemical analysis of polyanionie compounds found in the extracellular matrix encountered by migrating cephalic neural crest cells. branes by the use of lactoperoxidase. In “Methods in Enzymology” Anat. Rec. 196,401-412. (S. Fleischer and L. Packer, Eds.), Vol. 32, pp. 103-109. Academic CARSON,D. D., FARACH,M. C., EARLES,D. S., DECKER,G. L., and LENPress, New York. NARZ,W. J. (1985). A monoclonal antibody inhibits calcium accu- REALE, E., LUCIANO,L., and KUHN, K.-W. (1983). Ultrastructural armulation and skeleton formation in cultured embryonic cells of the chitecture of proteoglycans in the glomerular basement membrane. sea urchin. Cell 41,639-648. J. Histochem. Cytochem 31,662-668. DECKER,G. L., and LENNARZ,W. J. (1979).Sperm binding and formation RUOSLAHTI,E., and PIERSCHBACHER,M. D. (1986). Arg-Gly-Asp: A of the fertilization envelope by surface complexes isolated from sea versatile cell recognition signal. Cell 44, 517-518. urchin eggs. J. Cell BioL 81.92-103. SCOTT,J. E., and DORLING,J. (1965). Differential staining of acid glyDUTT, A., TANG, J.-P., WELPLY,J. K., and CARSON,D. D. (1986). Regcosaminoglycans (mucopolysaccharides) by alcian blue in salt soulation of N-linked glycoprotein assembly in uteri by steroid horlution. Histochemistry 5,221-233. mones. Endocrinol. 118,661-673. SHERMAN,M. I., and SALOMON,D. S. (1975). The relationship between DZIADEK, M., FUGIWARA,S., PAULSSON,M., and TIMPL, R. (1985). Imthe early mouse embryo and its environment. In “The Developmental munological characterization of basement membrane types of heBiology of Reproduction” (C. L. Markert and J. Papaconstantinou, paran sulfate proteoglycan. EMBO .I 4,905-912. Eds.), pp. 277-309. Academic Press, New York. GALLAGHER,J. T., and WALKER, A. (1985). Molecular distinctions be- SHERMAN,M. I., and WUDL, L. R. (1976). The implanting mouse blastween heparan sulfate and heparin. B&hem J. 230,665-674. tocyst. In “The Cell Surface in Animal Embryogenesis and DevelGRINNELL,F., HEAD, J. R., and HOFFPAUIR,J. (1982). Fibronectin and opment” (G. Poste and G. L. Nicolson, Eds.), pp. 81-125. Elsevier/ cell shape in viva: Studies on the endometrium during pregnancy. North-Holland, Amsterdam. J. Cell Biol 94, 597-606. WARTIOVAARA,J., LEIVO, I., and VAHERI, A. (1979). Expression of the HANDIN, R. I., and COHEN,H. J. (1976). Purification and binding propcell surface-associated glycoprotein, fibronectin, in the early mouse erties of human platelet factor four. J. BioL Chem 251,4272-4282. embryo. Dev. Biol 69,247-257. HEIFETZ,A., LENNARZ,W. J., LIBBUS,B., and Hsu, Y.-C. (1980).Synthesis WOODLEY,D. T., RAO, C. N., HASSEL, J. R., LIOT~A, L. A., MARTIN, of glycoconjugates during the development of mouse embryo in vitro. G. R., and KLEINMAN, H. K. (1983). Interactions of basement memDew. Biol 80,98-408. brane components. B&hem. Biqvhys. Acta 761,278-283. JENKINSON,E. J., and WILSON, I. B. (1973). In vitro studies on the WOODS,A., COUCHMAN,J. R., and HOOK, M. (1985). Heparan sulfate control of trophoblast outgrowth in the mouse. J. Embryol. Exp. proteoglycans of rat embryo fibroblasts. J BioL Chem 260,10,872Mwphol.

30,21-30.

10,879.