BIOLOGY OF REPRODUCTION 52, 903-912 (1995)
Monoclonal Antibody Directed to Le Oligosaccharide Inhibits
Implantation in the Mouse1
ZHENG M. ZHU, 3 4 NAOYA KOJIMA, 5 MARK R. STROUD, 5 SEN-ITIROH HAKOMORI, 5
and BRUCE A. FENDERSON 2' 3
Department of Pathology, Anatomy, and Cell Biology,3 Jefferson Medical College
Thomas Jefferson University, Philadelphia,Pennsylvania 19107
Department of Biochemistry,4 Dalian Medical University, Dalian 116023, People's Republic of China
The Biomembrane Institute and University of Washington,5 Seattle, Washington 98119
ABSTRACT
We investigated the role of carbohydrates in blastocyst attachment to the uterine epithelium. LeY (Fucal--2Gall--4tFucal-3J1GlcNAc) was localized by indirect immunofluorescence to the surface of the mouse blastocyst and uterine epithelium. Western blot analysis showed that Le is carried on many uterine glycoproteins in both pregnant and nonpregnant females;
however, new species were detected on Day 4 postcoitum (p.c.) coincident with the onset of uterine receptivity. The function
of LeYin implantation was tested by injecting monoclonal antibody (mAb) directly into the uterine lumen on Days 3-5 p.c. The
effects of intrauterine injections on implantation were scored by comparing the number of viable embryos to the number of CL
on Day 10 p.c. Injection of purified anti-LeY IgM into the uterine lumen on the afternoon of Day 4 significantly inhibited implantation. This effect was dose-dependent and was obtained during a narrow time window, from 87 to 93 h p.c. Inhibition of
implantation was not observed in contralateral uterine horns injected with saline, nor was it observed in uterine horns injected
with other anti-carbohydrate mAbs. We conclude that binding of anti-LeY to the blastocyst or luminal epithelium masks a ligand
involved in implantation. Although the mechanism of inhibition is unknown, we show that LeY can interact with another oligosaccharide (H) that has been described as a possible uterine ligand for blastocyst attachment. We hypothesize that LeY and H
form carbohydrate-carbohydrate interactions that promote close apposition of cell surface membranes during an early step in
implantation.
INTRODUCTION
cell adhesion molecules such as laminin, fibronectin, and
E-cadherin have not been identified on the apical surface
of the blastocyst or uterine epithelium (reviewed in [9]).
On the other hand, uterine receptivity to implantation is
known to be associated with numerous changes in cell surface carbohydrate expression, including changes in the glycoprotein composition of the surface epithelium [10-12],
the glycolipid composition of the surface endometrium
[13, 14], the size and charge of the apical glycocalyx [2,15,16],
and the profile of glycoproteins and proteoglycans secreted
into the luminal fluid [17, 18]. The embryo also undergoes
a series of glycosylation changes prior to implantation [19, 20].
Together, these changes in carbohydrate expression may
serve to regulate the time and place of blastocyst attachment within the uterus (reviewed in [21]).
Previously, we reported that the Le antigen (see Table
1 for structure) is present on the surface of peri-implantation mouse embryos and noted that Le expression is regulated by differentiation or coating factors within the uterine lumen [22]. In the present report, we have 1)
characterized the distribution and composition of LeY in the
mouse uterus during the estrous cycle and early pregnancy,
2) tested the effects of anti-Le antibody on implantation
using a novel bioassay, and 3) investigated the possibility
that Le mediates blastocyst attachment to the uterine epithelium by forming carbohydrate-to-carbohydrate interactions.
Implantation is a complex developmental process in which
the mammalian embryo attaches to the intact surface of the
uterine luminal epithelium (reviewed in [1]). In the mouse,
the embryo hatches from the zona pellucida at the 64-cell
stage and attaches to the anti-mesometrial surface of the
uterine wall on Day 4.5, approximately 100 h postcoitum
(p.c.) [2,3]. Ultrastructural studies have shown that this stage
of implantation is preceded by interdigitation of cell surface microvilli and close apposition of plasma membranes
[2, 4, 5]. Blastocyst attachment is known to trigger local degeneration of the uterine epithelium [6] and permit integrin receptors on trophectodermal cells to engage extracellular matrix molecules in the basement membrane of
the uterine epithelium [7, 8]. Within 24 h of contact, the
embryo is firmly embedded in the uterus and has begun
to form extra-embryonic membranes involved in nutrient
absorption and secretion.
The mechanisms that control blastocyst attachment to the
uterine luminal epithelium are largely unknown. Obvious
Accepted December 5, 1994.
Received September 8, 1994.
'This investigation received financial support from the Special Programme of
Research, Development, and Research Training in Human Reproduction, World Health
Organization (B.F.), and from N.I.H. Outstanding Investigator Grant CA42505 (S.H.).
2Correspondence: Dr. Bruce Fenderson, Department of Pathology, Anatomy, and
Cell Biology, Thomas Jefferson University, 1020 Locust St., Philadelphia, PA 19107.
FAX: (215) 955-8703.
903
904
ZHU ET AL.
MATERIALS AND METHODS
Animals
Mature (6-8 wk) female Swiss-Webster mice were purchased from Taconic Farms (Germantown, NY) and housed
in the Thomas Jefferson University animal quarters with an
automatic 12L:12D schedule. Fertilization in normally cycling females was assumed to take place around midnight
(5-7 h after dark). The day a copulatory plug was present
was considered to be Day 1 of pregnancy.
Embryos
Preimplantation embryos were flushed from the uterus
with phosphate-buffered saline (0.9%; PBS, pH 7.2). The effects of monoclonal antibodies (mAb) on embryo viability
and development were tested by culturing Day 3 and 4 embryos in 96-well microtiter plates (Falcon Plastics, Oxnard,
CA) in the presence of either sterile ascites or purified IgM.
Morulae were cultured in Brinster's BMOC-3 medium (Gibco
Labs., Grand Island, NY). Blastocysts were cultured in Ham's
F12/DMEM (Gibco Labs.) supplemented with 10% fetal bovine serum (Hyclone Labs., Logan, UT). In some experiments, blastocysts were allowed to attach to glass coverslips
precoated with 10 g/ml human fibronectin (BoehringerMannheim, Indianapolis, IN). Embryos were observed with
a Leitz (Wetzlar, Germany) inverted-phase microscope with
a Nikon (Tokyo, Japan) camera.
Antibodies
The following mAbs were used: FE-J1 directed to terminal GlcNAc [23]; FE-A5 directed to lacto-series type 2 chain
[24]; MNH-1 directed to lacto-series type 1 chain (Stroud
and Hakomori, unpublished data); BE2 directed to H type
2 chain [25]; MC-480 directed to Lex [19]; AH-6 directed to
Le [26]; and VESP 6.2 directed to sulfated galactose [27].
The carbohydrate structures recognized by these mAbs are
shown in Table 1. Ascites fluids were prepared in pristanetreated Balb/c mice or Balb/c x C57B1/6 F1 mice (for FE-J1).
Before injection, ascites fluids were dialyzed exhaustively
against Hanks' balanced salt solution (HBSS), centrifuged at
10 000 x g for 5 min, sterilized by passage through a 0.22,Im filter, and stored in 100-l aliquots at -80 0C. IgM concentration in ascites fluids was estimated by serial dilution
and dot-blot analysis; alkaline phosphatase-conjugated goat
anti-mouse IgM (Hyclone Labs.) was used as a detecting antibody (diluted 1:3000), and purified AH-6 was used as an
IgM standard. Each ascites tested contained approximately
4 mg/ml IgM. Monoclonal antibody AH-6 was purified by
ammonium sulfate precipitation and gel exclusion chromatography on Sephacryl S-300 (Pharmacia, Uppsala, Sweden) according to the method of Harlow and Lane [28]. The
first peak off the column was dialyzed extensively against
HBSS, and anti-Le activity was verified by indirect immunofluorescence on uterine frozen sections. IgM purity was
assessed by SDS-PAGE according to the method of Laemmli
[29]. Protein concentration was determined through use of
the Bradford protein assay (Bio-Rad Labs., Melville, NY) with
BSA (Sigma Chemical Co., St. Louis, MO) as protein standard.
IntrauterineInjection
Pregnant mice were anesthetized by i.m. injection of 60
mg/kg ketamine (Aveco, Fort Dodge, A) and 6 mg/kg xylazine (Rugby Labs., Rockville Center, NY). Bilateral incisions were made in the skin on the dorsal surface (1 cm
from the spine at the level of the last rib), and a small incision was made in the peritoneum near the position of the
ovarian fat pad. The fat pad, with accompanying ovary and
uterus, was gently pulled from the body cavity, and 20 l
of sterile test solution was injected (30 ° angle) into the lumen of the proximal portion of each uterine horn by means
of a Hamilton (Fisher Scientific, Pittsburgh, PA) repeating
syringe and 30-gauge needle. Sterile PBS was injected into
the left uterine horn (control), and sterile mAb was injected
into the right uterine horn (experimental). After injection,
the peritoneum was closed with a single suture and the
skin was closed with a single wound-clip. Mice were killed
on Day 10 p.c., and the numbers of embryos and CL were
counted under a dissecting microscope. For each uterine
horn, an implantation rate was calculated as the ratio of
viable embryos to CL x 100. Solutions having a blocking
effect on implantation were expected to lower this implantation rate by reducing the number of embryos without affecting the number of CL.
StatisticalAnalysis
Differences in implantation rates between control (PBSinjected) and experimental (mAb-injected) uteri were analyzed for significance using a paired t-test. Differences between AH-6- and other mAb-injected groups were analyzed
for significance using the Kruskal-Wallis chi-square test.
Immunofluorescence
Uterine frozen sections (5 jim) representing different
reproductive stages were fixed with acetone for 1 min, rehydrated with 5% (w/v) BSA, and reacted with mAb for 1
h at 4C. Bound antibody was detected with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgM or antimouse immunoglobulins (Dako, Carpinteria, CA). In some
experiments, the retention time of mAbs injected into the
uterine lumen was verified by direct application of FITCconjugated second antibody to uterine frozen sections. Results were photographed via a fluorescence microscope with
epi-illumination (Carl Zeiss, Inc., Thornwood, NY). The histology of PBS- and mAb-injected uterine horns was examined for evidence of pathologic changes by hematoxylin and
eosin staining of 6-jim paraffin sections. Zona-free mouse
embryos were examined for surface carbohydrate expres-
905
SACCHARIDES INVOLVED IN IMPLANTATION
TABLE 1. Carbohydrate differentiation antigens of the mouse uterine epithelium.
Expression'
Antibody
Antigen
FE-J1
FE-A5
MNH-1
BE-2
MC-480
GlcNAc
Lacto (type 2)
H (type 1)
H (type 2)
X
Le
AH-6
Ley
VESP 6.2
Sulfatide
Carbohydrate structure
Proestrus
Estrus
Day 4
GlcNAc31-6Gal
Galpl-4GIcNAcpl--3Gal
Fucal1-2Gall1--3GlcNAcpl-3Gal
Fuca1-2Galp1--4GlcNAcp1-3Gal
Gal1-4GlcNAcP1-3Gal
3
-
-
-
+++
++
+
+
+
+
+ +b
+++
++
+
Fucal
Fucal- 2G all14GIcNAc 1--3Gal
3
Fuca1
SO3-3Gal1--Ceramide
-
'Indirect immunofluorescence analysis of carbohydrate antigen expression in uterine frozen sections. Scale is strong
positive (+++) to negative (-) staining.
bThe lumenal fluid is strongly positive at this stage.
sion by indirect immunofluorescence as described previously [30]. Information concerning the carrier molecules of
carbohydrate antigens was obtained by pretreating embryos for 15 min at room temperature with 1 M NaC in
PBS, 0.05% trypsin/10 mM EDTA (Gibco Labs.), or 1% octylglucoside in PBS.
Western Blotting
Uteri were removed at different reproductive stages and
immersed in ice-cold HBSS containing 1 mM PMSF (Sigma
Chemical Co.). Uterine segments were opened by longitudinal incision, placed on an ice-cold frosted-glass surface,
and scraped with a surgical blade to remove endometrium.
Endometrial cells were extracted with 1% (v/v) Nonidet P40 in HBSS containing 1 mM PMSF for 1 h at 4°C. Samples
were centrifuged at low speed to remove cellular debris
and concentrated by acetone precipitation (drop-wise addition of 10x each sample volume with ice-cold acetone).
Uterine glycoproteins were dissolved in Laemmli buffer
containing 4% SDS, boiled for 2 min, and separated by SDSPAGE. Proteins were identified in the gel using Coomassie
brilliant blue or were transferred to nitrocellulose membranes and labeled with anti-carbohydrate mAbs, as described previously [24]. After electrophoretic transfer, nitrocellulose membranes were stained with 0.1% Fast green
(Sigma Chemical Co.) to verify protein transfer, destained,
blocked with 5% (w/v) BSA in Tris-buffered 0.9% saline
(pH 7.6), and incubated with primary antibody for 2 h at
4°C. Bound antibody was detected through use of alkaline
phosphatase-conjugated goat anti-mouse IgM (Hyclone Labs.)
diluted 1:3000. Color reaction was obtained using 5-bromo4-chloro-3-indolyl phosphate and Nitroblue tetrazolium
(Sigma Chemical Co.) according to the procedure of Harlow and Lane [28]. Molecular mass standards were from BioRad Labs.
Liposome Adhesion
Liposome binding to glycolipid targets was performed
as described previously [31, 32]. In brief, liposomes containing a glycolipid standard were labeled with '4 C-cholesterol and incubated in plastic wells precoated with increasing quantities of various glycolipid targets. Liposomes were
prepared with 25 ,Iug dimyristoyl phosphatidylcholine, 25
Iug '4C-labeled cholesterol (25 000 cpm/mg), and 12.5 xg
glycolipid. Lipid mixtures were dissolved in 100 ill diethyl
ether and injected slowly into 2 ml warm Tris-buffered saline (TBS) with vortex mixing. Ether was removed by rotary
evaporation, and liposome concentration was adjusted to
25 000 cpm '4C-cholesterol/100 Ipl. For adhesion assays, flatbottomed 96-well plates (Falcon Plastics) were precoated
with glycolipids in ethanol, dried at 37°C, and incubated for
1 h with TBS containing 1% BSA. Glycolipid-liposomes were
added to each well (100 1 l) and incubated with target glycolipids on a rotary shaker at room temperature. After 16
h, each well was washed seven times with TBS-BSA through
use of a thin capillary tip. Liposomes remaining were extracted with isopropyl alcohol/hexane/water (55:25:20), and
radioactivity was counted. Lipids were from Sigma Chemical Co.; ' 4C-cholesterol (53 mCi/mmol) was from DuPontNew England Nuclear (Boston, MA). To obtain glycolipids,
Le was prepared by enzymatic fucosylation of H type 2 chain
[33]; H type 2 chain and paragloboside were isolated from
human blood group O erythrocytes; H type 1 chain was
isolated from pig intestine; and Lex was isolated from human colonic adenocarcinoma at The Biomembrane Institute.
RESULTS
DifferentiationAntigens
Carbohydrate differentiation antigens of the mouse uterus
were investigated by indirect immunofluorescence using
well-characterized mAbs (Table 1). Antigens involving terminal N-acetylglucosamine and N-acetyllactosamine were not
906
ZHU ET AL.
FIG. 1. Indirect immunofluorescence analysis of LeY expression in the mouse uterus through use of mAb AH-6. During estrus, LeY is detected on the
epithelial cell surface and in the luminal fluid (a, b). On Day 4 p.c., LeYis detected in the uterine epithelium, primarily on the apical cell surface (c, d).
Photographs were obtained via phase (a, c) or fluorescence (b, d) microscopy. Abbreviations: E, epithelium; S. stroma. Magnification x200.
detected in the uterus at any phase of the estrous cycle.
Type 1 chain H was detected in the glandular and luminal
epithelium throughout the estrous cycle, but was weakly
detected on Day 4 p.c. Antigens involving sulfated galactose
were detected in the epithelium primarily during estrus. In
contrast, antigens involving od1--3 fucosylated type 2 chain
(LeX and LeY) were detected in the glandular and luminal
epithelium during all phases of the estrous cycle and early
pregnancy. Le was also detected in the luminal fluid during
estrus (Fig. 1).
Uterine glycoproteins that carry Le antigens were identified by Western blotting with mAb AH-6 (Fig. 2). LeY was
detected on most high-molecular-mass glycoproteins in the
endometrium of both pregnant and nonpregnant females.
However, certain low-molecular-mass species, with molecular masses of 12 kDa and 16 kDa, appeared in the endometrium on Day 4 p.c. coincident with the onset of uterine
receptivity. It is possible that these glycoproteins represent
proteolytic fragments of higher-molecular-mass components; however, identical results were obtained in five in-
FIG. 2. Western blot analysis of uterine endometrial glycoproteins that carry LeY antigen. Glycoproteins were derived as explained in Materials and
Methods. Lanes are identified as follows: S, molecular mass standards (x 10-3); N, nonpregnant; 3-6, Days 3, 4, 5, and 6 p.c.
907
SACCHARIDES INVOLVED IN IMPLANTATION
TABLE 2. Stage-specific expression of LeY on mouse embryos.
Day p.c.
3
4
5
6
Developmental stage
8 cell from oviduct
16-64 cell from uterus
Hatched in vitro
Trophectoderm outgrowth in vitro
%Positive (number)'
33
100
100
0
(40/120)
(92/92)
(10/10)
(0/21)
'Results of indirect immunofluorescence using mAb AH-6. Data for Days 3
and 4 of development are from Fenderson et al. 1221.
hydrophobic, as they were not detected on blots pretreated
with 1% (v/v) Tween 20 (results not shown).
The presence of LeY on peri-implantation mouse embryos was reported previously [22]. Here, we investigated
the fate of Le expression during hatching and trophoblast
outgrowth. Le was clearly detected by indirect immunofluorescence on blastocysts that hatched in vitro (Fig. 3, a and
b). Embryo dissociation revealed that both trophectoderm
and inner cell mass (ICM) express this antigen (Fig. 3, c
and d). However, LeY was not detected on the exposed surface of trophectodermal cells after contact with fibronectincoated plastic (Fig. 3, e and f). These results establish LeY
as a stage-specific antigen of 16- to 64-cell mouse embryos
(Table 2). Expression of this antigen on blastocysts was stable to high salt and resistant to trypsin/EDTA but was readily
extractable with nonionic detergent (Table 3). These results suggest that the LeY antigen of early embryos is carried
either on a lipid (glycosphingolipid) or on a trypsin-resistant integral membrane protein.
FIG. 3. Indirect immunofluorescence analysis of LeY expression on
mouse blastocysts using mAb AH-6. LeY is present on the surface of embryos flushed from the uterus and allowed to hatch in vitro (a, b). DissoY
ciation with trypsin/EDTA reveals the presence of Le
on all cells of the 32cell blastocyst, including those of trophectodermal and ICM lineage (c, d).
Le is not present on trophectodermal cells after attachment to fibronectincoated plastic (e, f). Photographs were obtained via phase (a, c, e) or fluorescence (b, d, f) microscopy. Abbreviations: ICM, inner cell mass; T, trophectoderm. Magnification x200.
Effects of mAbs on Implantation
An in vivo assay was developed to test the function of
Le and other surface antigens in implantation. Monoclonal
antibodies (either sterile ascites or purified IgM) were injected into the uterus of anesthetized mice on Days 3-5 p.c.
Injection of PBS into the contralateral uterine horn served
as an internal negative control. The implantation rate in mAband PBS-injected uterine horns was determined on Day 10
p.c. by comparison of the number of viable embryos in the
uterus to the number of CL in the ovary (see Materialsand
Methods). Intrauterine administration of mAb AH-6 on the
afternoon of Day 4 (87 h p.c.) was found to cause a dramatic decline in implantation rate, from 80% (PBS control)
to 40% (Fig. 4). Inhibition of implantation was obtained with
both AH-6 ascites and purified AH-6 IgM, but was not obtained with four other anti-carbohydrate mAbs, including
TABLE 3. Stability/resistance of LeY antigen on mouse blastocysts.
Treatment
dependent assays using extracts from four different pregnant mice. The 12-kDa and 16-kDa glycoproteins of
pregnancy were also detected in pseudopregnant mice on
Day 4 p.c. (results not shown). These glycoproteins may be
1 M NaCI
0.05% trypsin/EDTA
1% octylglucoside
%Positive (number)b
100 (6/6)
100 (8/8)
0 (0/2)
aEmbryos were treated for 15 min at room temperature, washed in PBS
containing 1% BSA, and then reacted with mAb AH-6 at 4C.
bResults of indirect immunofluorescence.
908
ZHU ET AL.
[Conl
100
Experimental I
100
75
75
!
i
50
.q
50
25
25
0
AH-6
n=20
*AH-6
n=13
FE-JI
n=11
FE-A5
n=5
MNH-I
n=6
VESP 6.2
n=6
FIG. 4. Effects of intrauterine injection of mAb on implantation. Pregnant mice were injected on Day 4 (-87-90 h p.c.) with 20 I sterile PBS
(left uterine horn) or 20 I1sterile hybridoma ascites (right uterine horn).
Purified AH-6 IgM was also tested (*AH-6; 80 g per injection). The effects
of mAbs on implantation were determined by comparing the total number
of viable embryos to the total number of CL in each uterine horn on Day
10 p.c. Implantation rate is defined as number of embryos/number of CL
x 100. Results represent the mean
SE of replicate assays (see abscissa
for n values). Differences in implantation rate between PBS- and AH-6-injected uterine horns were highly significant (paired t-test; p < 0.0001). Differences in implantation rate between AH-6- and FE-J1-injected mice were
also highly significant (Kruskal-Wallis chi-square test; p < 0.0001).
two that bind to the uterine surface epithelium (MNH-1 and
VESP 6.2).
Several control experiments were performed to test the
validity and specificity of our results. First, the localization
and fate of AH-6 in the uterine lumen were determined by
direct treatment of frozen sections with FITC-conjugated antimouse IgM. AH-6 was detected along the apical surface of
the luminal epithelium at 6, 12, and 18 h postinjection (results not shown). Second, we determined the effect of antibody dosage on implantation rate. Inhibition was maximal
at 8 g IgM per injection but was not observed at 0.8 Ipg
IgM per injection (Fig. 5). Thus, the inhibitory effect of AH6 on implantation was dose-dependent. Third, the time
window of inhibition was investigated by performing intrauterine injections at various times postfertilization (Fig. 6).
There was no effect of intrauterine injection on Day 3 (60
h p.c.), a stage when embryos are still in the oviduct. Surprisingly, intrauterine injections of either PBS or AH-6 at 80
h p.c. caused inhibition. At this stage, embryos have just
entered the uterus and may be susceptible to nonspecific
loss via mechanical flushing through the cervix. Specific inhibition was observed at 87 and 93 h p.c., but this effect
was dramatically reduced at 99 and 105 h p.c. Thus, intrauterine administration of mAb AH-6 caused a specific inhibition of implantation only during a narrow time window, from 87 to 93 h p.c.
0
0
0.08
IgM (g/20
8
0.8
80
.1l injected nuid)
FIG. 5. Dose response of the inhibitory effect of mAb AH-6 on implantation. Pregnant mice were injected on Day 4 (-87 h p.c.) with either
20 pI sterile PBS (left uterine horn) or 20 Ijl of different amounts of purified
AH-6 IgM antibody (right uterine horn). Results represent the mean implantation rate + SE in uterine horns injected with PBS (X) or AH-6 (circles)
(8-9 mice per data point).
.,
L
-
75
-
I
C
Ia150
Is
25
0
I
48
60
72
84
96
108
120
Hours
FIG. 6. Time window for inhibitory effect of mAb AH-6 on implantation. Pregnant mice were injected with 20 i sterile PBS (left uterine horn)
or 20 pi sterile AH-6 ascites (right uterine horn) at specific times p.c. It is
assumed that fertilization takes place at about midnight (t = 0 h). Results
represent the mean implantation rate - SE in uterine horns injected with
PBS (X) or AH-6 (circles) (7-10 mice per data point). Critical steps in implantation include 1) blastocyst entry into the uterus (-78 h p.c.); 2) apposition of trophectodermal and uterine epithelial cell membranes (84-100
h p.c.); and 3) blastocyst adhesion and penetration of the uterine epithelium (100-120 h p.c.).
909
SACCHARIDES INVOLVED IN IMPLANTATION
CP
0,
W
4)-
c
en c
V
E x
c
04
o
.
CX
a
E x
_
oa
CLU C
>j
IJ
x
J
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Glycolipid Coated (jig)
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Glycolipid Coated (jig)
Y
FIG. 7. Evidence for specific interaction between Le glycolipid and blood group H antigen. A) LeY-containing liposomes labeled with 4C-cholesterol
were incubated in 96-well plates precoated with increasing quantities of H type 1 chain (open circles), H type 2 chain (solid circles), Lex (open squares),
Ley (open triangles), or paragloboside (solid triangles). Specific interaction between Ley and both H type 1 and H type 2 chains was observed. B) Lecontaining liposomes labeled with 4C-cholesterol were incubated in 96-well plates, precoated with increasing quantities of H type 1 chain (open circles),
H type 2 chain (solid circles), Le (open squares), or Ley (open triangles). Lex liposomes interact with LeX, but do not bind H type 1 or H type 2 chain.
Results represent the mean
SD of triplicate assay wells.
Additional control experiments were performed to clarify the mechanism of AH-6-induced inhibition of implantation. First, uteri were examined at various times following
AH-6 injection for evidence of pathologic changes; light microscopic analysis of paraffin sections at 6, 12, and 24 h
revealed no evidence of necrosis, edema, inflammation, or
acute cell injury (results not shown). Second, we compared
the implantation rates in uterine horns that had been injected previously with PBS or AH-6. Ten days after intrauterine injections, females were housed with males; all became pregnant within 6 days (n = 12), and the implantation
rate in PBS-injected and AH-6-injected uterine horns was
identical (90%). Thus, there were no long-term, ill effects
of AH-6 injection on uterine function. Third, we investigated the effects of mAb AH-6 on embryo viability. Neither
AH-6 ascites (diluted 1:5) nor purified AH-6 IgM (8-160 pag/
ml) was toxic to blastocysts cultured for 48 h in vitro. Identical results were obtained in three separate experiments
(n = 20). Furthermore, embryos recovered from uterine
horns 4-24 h after AH-6 injection were viable (i.e., refractile and able to exclude Trypan blue; n = 12).
Carbohydrate-CarbobydrateInteractions
Cell surface carbohydrates interact with other cell surface and extracellular matrix carbohydrates to mediate cell
adhesion. Therefore, we tested the possible interaction of
Y
Le
glycolipids with other glycolipids whose oligosaccharide head groups are known to be present on the embryo
Y
or the uterine epithelium. Le
glycolipids in liposomes were
found to bind H type 1 and H type 2 chain, but not Le,
Lex, or paragloboside (Fig. 7, panel A). Binding was proportional to the quantity of H glycolipid coated on the plastic surface. In contrast, LeX glycolipids in liposomes bound
Lex, but not H type 1 or type 2 chain (Fig. 7, panel B).
DISCUSSION
Our results suggest that Le oligosaccharides play a functional role in implantation in the mouse. This conclusion
is based on the following experimental findings: 1) Le is
a developmentally regulated carbohydrate antigen of the
mouse embryo and the uterine epithelium (Figs. 1-3, Tables 1 and 2); 2) intrauterine administration of anti-Le inhibits implantation in a dose- and time-dependent manner
(Figs. 4-6); 3) inhibition of implantation is not due to
pathologic changes in the uterine epithelium or to embryo
toxicity. We hypothesize that intrauterine administration of
anti-LeY mAb causes steric hindrance of receptors involved
in attachment of the blastocyst to the uterine luminal epithelium.
910
ZHU ET AL.
Fucosylated, lacto-series glycans such as Lex, LeY, and H
have been identified previously in the endometrium of the
mouse [34,35] and rabbit [14]. These antigens are found in
the luminal and glandular epithelium and in uterine secretions (Fig. 1). Hormone replacement studies in ovariectomized mice have shown that the expression of these antigens is controlled by estrogen and progesterone [36]. Our
results indicate that Le is carried on numerous, high-molecular-mass glycoproteins throughout the estrous cycle and
early pregnancy, and that new glycoproteins (or patterns of
glycosylation) appear during the period of uterine receptivity (Fig. 2). The appearance of these low-molecular-mass
glycoproteins in the uteri of both pregnant and pseudopregnant females suggests that their expression is regulated
by maternal hormones or growth factors. In this connection, leukemia inhibitory factor (LIF) is known to play an
essential role in preparing the uterine epithelium for blastocyst implantation [37, 38]. It will be important to test the
effects of LIF on the expression of uterine glycoproteins,
including the 12-kDa and 16-kDa glycoproteins identified
here.
In addition to their presence in the uterus, LeX and LeY
are stage-specific embryonic antigens, with LeY appearing
on intrauterine embryos as Lex (SSEA-1) expression declines ([22]; reviewed in [39]). The biosynthesis of LeY may
involve activation of an embryonic a1-->2Gal fucosyltransferase (H transferase). Alternatively, Le may be acquired
passively from coating factors present in the uterine lumen
[22]. In either case, LeY does not persist on trophectodermal
cells after blastocyst outgrowth (Table 2, Fig. 3). These results suggest that early embryos have a mechanism for coupling cell adhesion to patterns of glycosylation and/or surface carbohydrate expression that is independent of overt
cellular differentiation. The fate of Le on trophectodermal
cells during contact with epithelial cells in utero remains
to be studied.
Intrauterine administration of mAb directed to Le on Day
4 p.c. significantly inhibited implantation (Fig. 4). This effect was both dose-dependent (Fig. 5) and stage-specific (Fig.
6). Moreover, inhibition was not observed in contralateral
uterine horns injected with saline, nor was it observed with
four other anti-carbohydrate IgM antibodies tested. A simple interpretation of our results is that AH-6 binds to the
blastocyst and uterine epithelium and masks a ligand
(probably Le) involved in implantation. More complicated
interpretations of our results are also possible. For example, intrauterine administration of AH-6 could 1) mask
binding sites on the uterine wall for embryo-derived trophic
factors [40]; 2) mask binding sites for intrauterine growth
factors such as LIF [37,38]; 3) precipitate components of the
uterine fluid that are critically involved in mediating implantation; or 4) inhibit trophectodermal cell migration or
differentiation. It should be possible to test these alternatives through an in vitro blastocyst-to-endometrium attach-
ment assay [41] in which the composition of uterine- and
embyro-derived factors is defined.
It is not clear why the implantation rate in AH-6-injected
uterine horns was decreased only to 40%, a mere 50% reduction in implantation rate. However, we believe that these
results accurately reflect the biology of implantation. First,
our results indicate that AH-6 blocks an early step in implantation. Maximal inhibition was obtained at 87 and 93 h
p.c.; there was little or no effect of intrauterine injection
after 99 h p.c. (Fig. 6). These results indicate that mAb AH6 cannot displace embryos once they have made contact
with the uterine epithelium. Second, pregnant females typically contain embryos of different developmental stages,
and blastocyst adhesion to the luminal epithelium occurs
over a considerable span of time, from 90 to 120 h p.c.
[2, 3]. Hence, implantation of "advanced" embryos in contact with the uterine wall at 87 h p.c. would not be inhibited. Similarly, embryos developing more slowly might approach the uterine wall only after the local concentration
of IgM had significantly declined. Thus, the 40% implantation rate we observed in AH-6-injected uterine horns may
represent a complete blockade of embryo attachment at a
critical step in implantation.
Ultrastructural studies have shown that the initial stage
of implantation in rodents is characterized by close apposition of plasma membranes and interdigitation of cell surface microvilli [2,4,5]. Recent evidence for "pre-attachment" of the mammalian embryo was provided by Shiotani
and co-workers [41], who noted that mouse blastocysts adhere to uterine fragments in culture without direct, membrane-membrane contact. Their observation of a 0.3-0.7Lim gap between trophectodermal and uterine epithelial cells
at the site of implantation suggests that cell surface carbohydrates (including glycolipids, glycoproteins, and proteoglycans) play a central role in mediating initial cell adhesion. The results presented here support this hypothesis.
Indeed, carbohydrates are known to be involved in many
biological recognition events, including sperm-egg attachment [42], leukocyte recirculation [43], and embryonic development [39,44].
Two basic mechanisms have been proposed for recognition of cell surface carbohydrates: 1) protein-carbohydrate interactions involving selectins (reviewed in [43]) and
glycosyltransferases (reviewed in [45]) and 2) carbohydratecarbohydrate interactions involving glycolipids, glycoproteins, and glycosaminoglycans (reviewed in [46]). Previously, we demonstrated a specific interaction between LeX
and Lex that may contribute to membrane apposition during compaction of the morula [31]. Here, we demonstrate
a specific interaction between LeY and H (Fig. 7) that may
contribute to membrane apposition during implantation. H
>3GlcNAc) has been identitype 1 chain (Fuctl-2Gallfied previously as a possible ligand for blastocyst attachment [47]. On the basis of inhibition studies using oligosaccharides, Lindenberg and co-workers hypothesized that
SACCHARIDES INVOLVED IN IMPLANTATION
a carbohydrate-binding protein on trophectodermal cells
mediates implantation via lock-and-key recognition of H
present on the apical surface of the uterine epithelium. Our
results suggest a modification of their hypothesis in which
trophectoderm-associated Le and uterus-associated H form
multivalent, carbohydrate-carbohydrate links. This cell
adhesion process could be triggered by a decrease in the
net-negative charge of the uterine surface epithelium, which
is known to occur at the time of implantation [15, 16].
In summary, intrauterine administration of various agents,
including antibodies, receptors, and ligands, should provide a useful method for testing the functional role of specific membrane molecules in implantation. Our results
demonstrate that saline and ascites fluids can be safely administered to the uterine lumen of pregnant mice without
adversely affecting implantation or embryonic development. Our results also focus attention on Le, a cell surface
antigen of embryos and uterine epithelial cells, as a possible ligand involved in implantation.
ACKNOWLEDGMENTS
The authors are grateful to Andrea and Ivan Damjanov for the generous use of
their laboratory and animal facilities, and for their kind help. The authors also thank
Yuan Gen Zhu, Thomas Jefferson University Center for Research in Medical Education, for expert advice on methods of statistical analysis.
REFERENCES
1. Damjanov I, Horvat B, Fenderson BA. Uterine environment during the implantation of the embryo (review). In: Bitar EE (ed.), Principles of Medical Biology.
Chicago: JAI Press; 1994: (in press).
2. Schlafke S, Enders AC. Cellular bases of interaction between trophoblast and
uterus at implantation. Biol Reprod 1975; 12:41-65.
3. Theiler K The House Mouse. Atlas of Embryonic Development. New York: SpringerVerlag; 1989.
4. Reinius S. Ultrastructure of blastocyst attachment in the mouse. Z Zellforsch 1967;
77:257-266.
5. Morris JE, Potter SW, Rynd LS, Buckley PM. Adhesion of mouse blastocysts to
uterine epithelium in culture: a requirement for mutual surface interactions. J
Exp Zool 1983; 225:467-479.
6. Parr EL, Tung HN, Parr MB. Apoptosis as the mode of uterine epithelial cell death
during embryo implantation in mice and rats. Biol Reprod 1987; 36:211-225.
7. Armant DR, Kaplan HA, Lennarz WJ. Fibronectin and laminin promote in vitro
attachment and outgrowth of mouse blastocysts. Dev Biol 1986; 116:519-523.
8. Sutherland AE, Calarco PG, Damsky CH. Expression and function of cell surface
extracellular matrix receptors in mouse blastocyst attachment and outgrowth. J
Cell Biol 1988; 106:1331-1348.
9. Carson DD, Raboudi N,Jacobs AL.Glycoprotein expression and function in embryo-uterine interactions. In: Lavia LA (ed.), Cellular Signals Controlling Uterine
Functions. New York: Plenum Press; 1991: 107-118.
10. Anderson , Olson GE, Hoffman LH. Stage specific alterations in the apical
membrane glycoproteins of endometrial epithelial cells related to implantation
in rabbits. Biol Reprod 1986; 34:701-720.
11. Horvat B, Damjanov . Two-dimensional electrophoretic analysis of endometrial
cellular proteins synthesized during early pregnancy in the mouse. Gynecol Obstet Invest 1993; 35:69-75.
12. Valdizan MC, JulianJ, Carson DD. WGA-binding, mucin glycoproteins protect the
apical cell surface of mouse uterine epithelial cells. J Cell Physiol 1992; 151:451465.
13. Zhu ZM, Deng H, Fenderson BA, Nudelman ED, Tsui Z. Glycosphingolipids of
human myometrium and endometrium and their changes during the menstrual
cycle, pregnancy and ageing. J Reprod Fertil 1990; 88:71-79.
14. Zhu ZM, Tsui ZC, Hakomori S, Fenderson BA Glycosphingolipids of rabbit endometrium and their changes during pregnancy. J Reprod Fertil 1992; 95:813823.
911
15. Hewitt K, Beer AE, Grinnell F. Disappearance of anionic sites from the surface
of the rat endometrial epithelium at the time of
blastocyst implantation. Biol
Reprod 1979; 21:691-707.
16. Anderson TL, Hoffman LH. Alterations in epithelial glycocalyx
of rabbit uteri during early pseudopregnancy and pregnancy, and following
ovariectomy. Am J Anat
1984; 171:321-334.
17. Bell SC, Patel SR, Kirwan PH, Drife JO. Protein synthesis
and secretion by the
human endometrium during the menstrual cycle and
the effect of progesterone
in vitro. J Reprod Fertil 1986; 77:221-231.
18. Dutt A, Carson DD. Lactosaminoglycan assembly, cell
surface expression and release by mouse uterine epithelial cells. J Biol Chem 1990;
265:430-438.
19. Solter D, Knowles BB. Monoclonal antibody defining
a stage-specific mouse embryonic antigen (SSEA-1). Proc Natl Acad Sci USA 1978;
75:5565-5569.
20. Shevinsky LH, Knowles BB, Damjanov I, Solter D. Monoclonal
antibody to murine embryos defines a stage-specific embryonic antigen on
mouse embryos and
human teratocarcinoma cells. Cell 1982; 30:697-705.
21. Fenderson BA. Saccharides involved in implantation.
Trends Glycosci Glycotech
1993; 5:271-285.
22. Fenderson BA, Holmes EH, Fukushi Y, Hakomori S. Coordinate
expression of X
and Y haptens during murine embryogenesis. Dev Biol 1986;
114:12-21.
23. Symington FW, Fenderson BA, Hakomori S. Fine specificity
of a monoclonal antitesticular cell antibody for glycolipids with terminal N-acetyl-D-glucosamine
structure. Mol Immunol 1984; 21:877-882.
24. Fenderson BA, Nichols EJ, Clausen H, Hakomori S. A monoclonal
antibody defining a binary N-acetyllactosaminyl structure in lactoisooctaosylceramide
(IV6Gall -4G4Gl
cNAcnLc6):a useful probe for determining differential glycosylation patterns between normal and transformed human fibroblasts.
Mol Immunol 1986; 23:747-754.
25. Young WW, Portoukalian J, Hakomori S. Two monoclonal
anticarbohydrate antibodies directed to glycosphingolipids with a lacto-N-glycosyl type 2 chain.
J Biol
Chem 1981; 256:10967-10972.
26. Abe K, McKibbin JM, Hakomori S. The monoclonal antibody directed
to difucosylated type 2 chain (Fuced-2Gall--4[Fuc1---3]
GlcNAc; Y determinant).
J Biol Chem 1983; 258:11793-11797.
27. Petry K, Nudelman E, Eisen H, Hakomori S. Sulfated lipids represent
common
antigens on the surface of Trypanosoma cruzi and mammalian tissues.
Mol
Biochem Parasit 1988; 30:113-122.
28. Harlow E, Lane D. Antibodies. A Laboratory Manual. Cold Spring Harbor:
Cold
Spring Harbor Press; 1988.
29. Laemmli UK. Cleavage of structural proteins during the assemble of
the head of
Bacteriophage T4. Nature 1970; 227:680-684.
30. Fenderson BA, Hahnel AC, Eddy EM. Immunohistochemical
localization of two
monoclonal antibody-defined carbohydrate antigens during early murine
embryogenesis. Dev Biol 1983; 100:318-327.
31. Eggens , Fenderson BA, Toyokuni T, Dean B, Stroud MR, Hakomori
S. Specific
interaction between Lex and Le' determinants: a possible basis for cell recognition in preimplantation embryos and in embryonal carcinoma cells. J Biol
Chem
1989; 264:9476-9484.
32. Kojima N, Hakomori S. Specific interaction between gangliotriaosylceramide
(Gg3)
and sialosyllactosylceramide
(GM3) as a basis for specific cellular recognition
between lymphoma and melanoma cells. J Biol Chem 1989; 264:20159-20162.
33. Salyan MEK, Stroud MR, Levery SB. Differentiation of type and type
2 chain
linkages of native glycosphingolipids by positive-ion fast-atom bombardment mass
spectrometry with collision-induced dissociation and linked scanning.
Rapid
Commun Mass Spect 1991; 5:456-462.
34. Babiarz B, Hathaway HJ. Hormonal control of the expression of antibody-defined
lactosaminoglycans in the mouse uterus. Biol Reprod 1988; 39:699-706.
35. Kimber SJ, Lindenberg S, Lundblad A Distribution of some Gal1-3(4)GcNAc
related carbohydrate antigens on the mouse uterine epithelium in relation to
the peri-implantational period. J Reprod Immunol 1988; 12:297-313.
36. Kimber SJ, Lindenberg, S. Hormonal control of a carbohydrate epitope
involved
in implantation in mice. J Reprod Fertil 1990; 89:13-21.
37. Bhatt H, Brunet LJ, Stewart CL. Uterine expression of leukemia inhibitory
factor
coincides with the onset of blastocyst implantation. Proc Natl Acad Sci USA 1991;
88:11408-11412.
38. Stewart CL, Kaspar P, Brunet LJ, Bhatt H, Gadi , Kontgen F, Abbondanzo
S.
Blastocyst implantation depends on maternal expression of leukemia inhibitory
factor. Nature 1992; 359:76-79.
39. Fenderson BA, Eddy EM, Hakomori S. Glycoconjugate expression during embryogenesis and its biological significance. Bioessays 1990; 12:173-179.
40. Weitlauf HM. Embryonic signaling at implantation in the mouse. Prog Clin Biol
Res 1989; 294:359-376.
912
ZHU ET AL.
41. Shiotani M, Noda Y, Mori T. Embryo-dependent induction of uterine receptivity
assessed by an in vitro model of implantation in mice. Biol Reprod 1993; 49:794801.
42. Florman HM, Wassarman PM. O-linked oligosaccharides of mouse egg ZP3 account for its sperm receptor activity. Cell 1985; 41:313-324.
43. Gahmberg CG, Kotovuori P, Tontti E. Cell surface carbohydrate in cell adhesion.
Sperm cells and leukocytes bind to their target cells through specific oligosaccharide ligands. In: Clausen H, Dabelsteen E (eds.), Carbohydrate Pathology.
APMIS Suppl 1992; 100: 39-52.
44. Varki A, Hooshmand F, Diaz S, Varki NM, Hedrick SM. Developmental abnormalities in transgenic mice expressing a sialic acid-specific 9-O-acetylesterase.
Cell 1991; 65:65-74.
45. Hathaway HJ, Shur BD. Novel cell surface receptors during mammalian fertilization and development. Bioessays 1988; 9:153-158.
46. Kojima N. Glycosphingolipid-glycosphingolipid interaction: a model for a new
type of cell recognition system. Trends Glycosci Glycotechnol 1992; 4:491-503.
47. Lindenberg S, Sundberg K, Kimber SJ, Lundblad A. The milk oligosaccharide,
lacto-N-fucopentaose I, inhibits attachment of mouse blastocysts on endometrial
monolayers. J Reprod Fertil 1988; 83:149-158.