assessment of in vitro embryo development was recorded daily until they reached the blastocyst stage. Cell Lines and Tissue Culture. A human choriocarcinoma.
Proc. Natl. Acad. Sci. USA Vol. 93, pp. 161-165, January 1996 Developmental Biology
HLA-G expression during preimplantation human embryo development (major histocompatibility complex/non-classical HLA)
ANDREA JURISICOVA*, ROBERT F. CASPER*, NEIL J. MACLUSKY*, GORDON B. MILLSt, *Division of Reproductive Science, CCRW Room 3-830, Department of Obstetrics and Gynecology, and Institute, 101 College Street, Toronto, ON Canada M5G 1L7
AND
CLIFFORD L. LIBRACH*t
tOncology Research, The Toronto Hospital Research
Communicated by Neil L. First, University of Wisconsin, Madison, WI, July 13, 1995
Immunocytochemical (ICC) studies of MHC expression during preimplantation human embryo development have given conflicting results. Lack of expression of class I and class II antigens was reported by Dohr et al. (8) and Desoye et al. (9). However, Roberts et al. (10) reported strong staining in some unfertilized human oocytes with the pan-class I monoclonal antibody (mAb) W6/32. Early literature on the expression of MHC antigens in embryos of other mammalian species was also controversial. However, employment of more sensitive techniques such as reverse transcription polymerase chain reaction (RT-PCR) or ELISA led to confirmation of the presence of classical and nonclassical class I antigens on the surface of preimplantation mouse embryos (11, 12). Recently, Sprinks et al. (13) reported detectable levels of transcript for the murine classical class I antigen H-2D as early as a few hours after syngamy of pronuclei. f32-Microglobulin (P32m) binds noncovalently to MHC class I molecules and ensures their proper folding and transport to the cell surface as well as participating in selection of class I associated peptides (14). Mice homozygous for a defective p32m gene have very little to nondetectable class I antigens present on their cell surface (15, 16). It is noteworthy that f32m of paternal origin has been shown to be synthesized by murine embryos as early as the two-cell stage (17). Meziou et al. (18) reported positive staining for f32m in the early blastocyst stage of porcine embryos. This study demonstrates the presence of protein and mRNA for the heavy chain of HLA-G and for P12m throughout the whole course of human preimplantation development from the oocyte to blastocyst stages. Detection of HLA-G was variable between embryos and was absent or present even among embryos from the same parental origin. f32m transcript was also found in embryos having HLA-G heavy chain transcripts. Interestingly, embryos expressing HLA-G mRNA had a significantly increased cleavage rate as compared to those with no detectable expression.
ABSTRACT HLA-G is a nonclassical class I major histocompatibility complex molecule with a restricted pattern of expression that includes the placental extravillus cytotrophoblast cells in direct contact with maternal tissues. Circumstantial evidence suggests that HLA-G may play a role in protection of the semiallogeneic human fetus. We examined whether HLA-G is expressed during the critical period of preimplantation human development and whether expression of this molecule could be correlated with the cleavage rate of embryos. Using reverse transcription PCR on surplus human embryos and unfertilized oocytes from patients undergoing in vitro fertilization we detected HLA-G heavy chain mRNA in 40% of 148 of blastocysts tested. The presence of HLA-G mRNA was also detected in unfertilized oocytes and in early embryos, but not in control cumulus oophorus cells. 132Microglobulin mRNA was also found in those embryos expressing HLA-G. In concordance with our mRNA data, a similar proportion of embryos stained positive for HLA-G utilizing a specific monoclonal antibody. Interestingly, expression of HLA-G mRNA was associated with an increased cleavage rate, as compared to embryos lacking HLA-G transcript. Thus, HLA-G could be a functional homologue of the mouse Qa-2 antigen, which has been implicated in differences in the rate of preimplantation embryo development. To our knowledge, the presence of HLA-G mRNA and protein in human preimplantation embryos and oocytes has not been reported previously. The correlation of HLA-G mRNA expression with cleavage rate suggests that this molecule may play an important role in human pre-embryo development.
Why the maternal immune system does not reject the developing fetus remains a fundamental mystery of reproductive biology. The series of experiments presented here test the hypothesis that HLA-G, a nonclassical class I major histocompatibility complex (MHC) antigen with limited polymorphism (1), may be expressed in the human preimplantation embryo. This molecule is the only class I molecule expressed by human placental trophoblast cells (2). Expression of HLA-G has also been found in some choriocarcinoma (trophoblast-derived tumor) cell lines (2, 3). The most striking features of this molecule, as compared to classical MHC molecules, are (i) that it has limited polymorphism (4) and (ii) that its expression is restricted to immunologically privileged tissues, the placenta and the human eye (5, 6). Preliminary evidence suggests that expression of HLA-G protects cells against natural killer (NK) cell lysis and that it does not stimulate an allogeneic response by peripheral blood T cells (7). These features suggest that expression of this molecule could be a crucial factor for fetal survival in the face of a potentially hostile maternal immune
MATERIALS AND METHODS In Vitro Fertilization (IVF) and Embryo Culture. Human unfertilized oocytes, abnormally fertilized oocytes, and spare preimplantation embryos were donated by couples attending the IVF Program of the Division of Reproductive Sciences at The Toronto Hospital. Ovarian stimulation was carried out using a modified flare-up protocol as described (19). Sixteen to 18 hr after insemination, oocytes were examined for the presence of two pronuclei. Spare embryos were either frozen for future transfers or donated for research after Abbreviations: MHC, major histocompatibility complex; RT-PCR, reverse transcription polymerase chain reaction; FITC, fluorescein isothiocyanate; 932m, f32-microglobulin; mAb, monoclonal antibody; ICC, immunocytochemistry (immunocytochemical); ICM, inner cell mass; NK, natural killer. ITo whom reprint requests should be addressed.
system. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
161
162
Developmental Biology: Jurisicova et al.
informed consent was obtained. This research was approved by the human ethics committee of The Toronto Hospital. Spare embryos from normally fertilized oocytes with two pronuclei were cultured in Ham's F10 medium (GIBCO/ BRL) supplemented with 10% human serum or cocultured with ovarian cancer cells of epithelial origin. Morphological assessment of in vitro embryo development was recorded daily until they reached the blastocyst stage. Cell Lines and Tissue Culture. A human choriocarcinoma cell line, JEG-3 (American Type Culture Collection), served as a positive control for expression of HLA-G in our experiments. The only HLA class I molecule that these cells express is HLA-G (2). The JEG-3 cells were cultured in Dulbecco's modified Eagle medium (DMEM)/F12 medium (GIBCO/ BRL) supplemented with 10% fetal calf serum (FCS) (Sigma). Cumulus oophorus cells, obtained together with oocytes from aspirated follicles, served as a negative control for HLA-G expression. These cells were cultured for 2 days in DMEM/F12 plus 10% FCS before they were used in experiments. HLA A,B,C null LCL.221 (American Type Culture Collection) and HLA-G transfected LCL.221-G cells (a gift from R. DeMars) were grown in RPMI medium with 10% FCS (20). RT-PCR. RT-PCR was performed on nucleic acid isolated from (i) human blastocysts, (ii) pools of five embryos at each cleavage stage, (iii) five to eight unfertilized oocytes, and (iv) '100 control cells (JEG or cumulus) using a protocol developed for nucleic acid preparation from small numbers of cells (21). The zona pellucida was removed prior to nucleic acid isolation using acid tyrode solution (pH 2.5). First-trimester placental tissue was obtained from elective pregnancy terminations. Total RNA was extracted using a standard phenolchloroform extraction method (22). cDNA synthesis was carried out using a 1:1 mixture of Moloney and avian reverse transcriptases (GIBCO/BRL). The RT reaction was performed at 37°C for 15 min. Each first strand cDNA product was divided into two portions for amplification of HLA-G heavy chain cDNA and 132M cDNA from the same specimen. Forty cycles of PCR, which included denaturation at 95°C for 1 min, annealing at 61°C for 2 min, and extension at 72°C for 3 min, were performed in a Thermal Cycler (PerkinElmer). For each reaction, water and cumulus cells were used as negative controls. JEG and first-trimester placental RNAs were used in each reaction as positive controls. PCR was carried out in a 20-,l volume. Primer Pm corresponded to a sequence located in exon 5 (transmembrane region; bases 2805-2826) and primer Put-2 was located in exon 8 (untranslated region; bases 3878-3998) of the HLA-G gene (1). This primer pair produced an amplified product of 292 bp. The f32m primers, Pf3l (bases 326-345) and Pf32 (bases 10361045 plus 2290-2299) (GenBank accession no. M17987; HUMB2M2), demarcated a 123-bp nucleotide sequence in the coding region of f32m. To test the specificity of our RT-PCR amplifications, we performed heminested primer PCRs using a second HLA-G downstream primer, PUt-1 (bases 3768-3789), located within the region of the primary PCR product. These primers enclosed a 183-bp region. Heminested primer PCR was performed under the same conditions as described for primary PCR using 2 gl of the amplified product as a template for a second PCR amplification with PTm and PUt-1 primers. For all reactions, a water control was used. The specificity of the amplified products was also confirmed by sequencing. HLA-G-Specific mAb (1B8). A mAb (1B8) recognizing HLA-G heavy chain was produced utilizing a 21-amino acid peptide corresponding to amino acids in positions 61-83 (23). The specificity of 1B8 was confirmed by comparing the ICC reaction observed with class I null cells (+0.221) with that of cells expressing HLA-G (HLA-G plus LCL.221, JEG) and cells expressing other class I molecules (ovarian cumulus cells) as well as by immunoprecipitation of the molecule (23).
Proc. Natl. Acad. Sci. USA 93
(1996)
ICC. ICC studies were carried out using the following mAbs: (i) W6/32 (DAKO), an IgG mouse mAb directed against a nonpolymorphic determinant of all human class I MHC antigens; (ii) 1B8, anti-HLA-G, an IgM mouse mAb, and (iii) anti-/32m rabbit anti-human IgG mAb (Dako). The JEG choriocarcinoma cell line and cumulus cells were grown on gelatin-coated coverships. Cells were fixed in phosphate-buffered saline (PBS) containing 3% paraformaldehyde, washed in PBS/0.1 M glycine, and incubated in PBS/1% human serum albumin (HSA) for 10 min. The concentrations of W6/32 (stock was 93 ,ug/ml) and 1B8 in ascites were adjusted by dilution with PBS/1% HSA to 1:100 and 1:150, respectively. The optimal dilution of the 1B8 antibody was determined from a preliminary series of experiments. Incubation with primary antibody was followed by four washes and then incubation with the appropriate fluorescein-conjugated secondary antibody. For 1B8, the secondary antibody was goat anti-mouse IgM (Zymed), at a dilution of 1:50, and for W6/32, the secondary antibody was goat anti-mouse IgG (Western Blotting Enterprises, Ontario) at a dilution of 1:15. The secondary antibody alone was used as a negative control for each experiment. LCL.221 and LCL.221-G cells grown in suspension were cytospun onto slides for analysis and fixed in the same way as described above. These slides were preincubated with 3% H202 solution. Preblocking was carried out in the Tris buffer containing 10% goat serum, followed by application of the 1B8 antibody at a dilution of 1:150 for 1 hr at room temperature. After washing, the biotinylated goat anti-mouse IgM (Zymed) was applied at a dilution of 1:100. Final detection was performed with streptavidin-peroxidase conjugate (Dako) and the reaction was terminated by diaminobenzidine, producing a brown staining on the surface of HLA-G-positive cells. A
, b-
12 3 4
B
FIG. 1. RT-PCR for detection of HLA-G heavy chain and 132m RNA expression in human first-trimester placenta, embryos, oocytes, and JEG choriocarcinoma cells. (A) Primary PCR containing HLA-G primers Pm and Put-2 (amplified fragment is 292 bp). Lanes: 1, placenta; 2, JEG-3 human choriocarcinoma cells; 3, cumulus oophorus cells; and 4-7, individual human blastocysts. (B) Two microliters of the first-round PCR products shown in A was used as a template for a second round of PCR with nested primers (PTm and PUt-1) amplifying a fragment of 183 bp. Lane assignments are the same as in A. Blastocysts in lanes 4-6 showed detectable expression of HLA-G transcript. However, no expression was observed in the blastocyst in lane 7. (C) The second half of the cDNA product was used as a PCR template with 132m-specific primers (Pl31 and P,32; amplified fragment is 123 bp). Positive expression of 132m was detected in all blastocysts as well as in placenta, JEG, and cumulus cells. Lane assignments are the same as in A. (D) RT-PCR amplified DNA fragment, using HLA-G-specific primers Pm and PUt-1. Lanes: 1, placenta; 2, JEG-3; 3, cumulus cells; and 4-9, sibling pools of five to eight unfertilized oocytes from six individual patients. Oocytes in lane 4-7 showed positive expression; however, no specific fragment could be seen in oocytes in lanes 8 and 9 and in cumulus cells (lane 3). Arrows indicate location of specific amplified fragment in each gel.
Developmental Biology: Jurisicova et al.
Proc. Natl. Acad. Sci. USA 93 (1996)
163
Table 1. Results of nested primer RT-PCR with human blastocysts Total Blastocysts No. % From 3PN From 1PN From 2PN PCR results 13.4 22 1 0 21 032m(-)/HLA-G(-) 43.3 71 2 1 68 932m(+)/HLA-G(-) 43.3 71 7 5 59 ,32m(+)/HLA-G(+) from developed blastocysts human mRNA in preimplantation 132m mRNA and Expression of HLA-G normally fertilized oocytes (2PN, two pronuclei), from abnormally fertilized (3PN, three pronuclei) oocytes, or from activated oocytes (1PN, one pronucleus).
A similar protocol was used for the rabbit anti-human P32m antibody at a dilution of 1:100 for labeling of JEG-3 cells, LCL 221, and LCL 221-G cells as well as human oocytes and embryos. The secondary antibodies used were biotinylated swine anti-goat, anti-mouse, and anti-rabbit immunoglobulins (Multilink, Dako) at a dilution of 1:100 and streptavidinperoxidase labeling was performed as described previously. All cells, whole-mount embryos, and oocytes were examined using a Zeiss Axioplan Universal microscope, and photographs were taken with a Zeiss microscope camera (MC 100).
FIG. 2. ICC for HLA antigens on whole-mount human oocytes and embryos. The first of each pair of photographs represents a phasecontrast micrograph; the second is the same preparation viewed using fluorescein isothiocyanate (FITC) fluorescence. (A and B) Unfertilized oocyte with cumulus oophorus cells after ICC using a primary pan class I mAb (W6/32) and goat anti-mouse IgG FITC secondary antibody. Positive staining was observed in oocyte membrane and in cumulus oophorus cells. (C and D) Unfertilized oocyte with cumulus oophorus cells after ICC with 1B8 anti-HLA-G mAb. Oocyte shows positive staining, but cumulus cells are negative. (E and F) Two-cell prembryo stained with 1B8 antibody. (G and H) Expanded human blastocyst stained with 1B8 antibody. Staining is restricted to trophectodermal cells. (I and J) Expanded human blastocyst stained with pan class I W6/32 antibody. Staining is localized on trophectodermal cells as well as the ICM. (A-D, x70; E and F, x 140; G-J, x 160.)
Statistical Analysis. The Mann-Whitney U test was used to compare the number of blastomeres per embryo in HLA-Gpositive and -negative blastocysts 24 and 48 hr after the presence of two pronuclei. A P value of 20 women were consistently negative for HLA-G since these cells express only classical HLA molecules. A f32m-specific product (123 bp) was amplified from JEG-3, placental, and cumulus cell mRNA and variably from blastocyst (Fig. 1C). Occasionally we observed additional bands after amplification with 12m primers, as in lanes 4 and 5 (Fig. 1C), despite our efforts for optimization of conditions for PCR. Of a total of 164 blastocysts from 49 couples, 71 (43.3%) were positive for HLA-G heavy chain and /32m mRNA (Table 1). In 71 cases (43.3%) only f32m mRNA was detected. In 22 blastocysts (13.4%) neither HLA-G nor f32m mRNA was detected. RT-PCR Detection of HLA-G and j32m in Human Oocytes and During Preblastocyst Developmental Stages. After establishing the presence of HLA-G and P32m mRNA in human blastocysts, we concentrated on preblastocyst development. For each RT-PCR, several embryos at the same developmental stage were pooled. Most of these were mixtures from unrelated patients. Embryos at the 2- to 4-cell, 5- to 8-cell, 9- to 16-cell, and morula stages were screened (data not shown). We were able to detect expression of HLA-G heavy chain mRNA and f32m through the whole course of preimplantation development of human embryos. To determine if there was expression of HLA-G mRNA prior to fertilization, we evaluated unfertilized oocytes. Twenty-one pools of unfertilized oocytes were screened. Eleven pools contained oocytes of the same maternal origin and 10 were mixtures from different patients. In 7 of 11 sibling pools, HLA-G and 932m mRNA were detected and in the remaining 4 of these 11 pools we could detect only j32m mRNA (Fig 1). All 10 multiple patient oocyte mixtures had detectable HLA-G and 32m mRNA. ICC Detection of HIA-G Protein During Human Preimplantation Development. W6/32, an IgG mouse mAb (Dako) directed against all human class I MHC antigens, showed positive membrane staining on the surface of 6 of 13 unfertilized oocytes. As expected, cumulus cells had positive W6/32 staining representing classical class I antigens (Fig. 2A and B). In two of five blastocysts, the trophectoderm and the inner cell mass (ICM) stained with W6/32 (Fig. 2 I and J). In blastocysts where no trophectoderm
164
**vi,>-.o;4@kI .;: . l.i*.ad'S"w ;,,
Developmental Biology: Jurisicova A.
et
al.
*
Proc. Natl. Acad. Sci. USA 93
(1996)
f.l
:. -.: -s;:
_
..
%, ..
* -.'.:.:-,
..: b
wF
.-.F,,,.;
@-s
sf
-'
@
of
*
..
tz,_r
*:
,-,,..;:
::. r: .S
.,
..
*,:::
te,.-; ,rf.
B,¢. .....
........
0
4
Ah
,.
#
V-
¢ ..
,4b0
j . w
f-
,.,,.
*i;.'. !. -.
:. :,
0
k
a
0::
FIG. 3. Immunoperoxidase staining using 1B8 HIA-G-specific mAb (cells are counterstained with hematoxylin). (A) HLA A,B,C null lymphoblastoid cell line LCL.221 showing no membrane staining. (B) LCL.221-G transfected cells showing positive expression of HLA-G determined by brown surface staining. (C) Photographs showing light field of human choriocarcinoma cells JEG-3 These cells often grow in clusters, seen at the bottom of the print. (D) Positive membrane staining of the same field as in C using indirect ICC with 1B8 (HLA-G) mouse anti-human IgM mAb and FITC-coupled goat anti-mouse IgM. Positive staining is seen in cells growing in the monolayer, with stronger staining of the cell clusters. (A, x90; B, x120; C and D, x155.) .
staining was observed (HLA-G negative), W6/32 staining was still found in the ICM (not shown). HLA-G protein was detected utilizing the mouse mAb 1B8 raised against the al region of HLA-G. The specificity of 1B8 staining was confirmed by positive HLA-G staining of LCL.221-G cells transfected with the HLA-G gene (Fig. 3B) and JEG-3 choriocarcinoma cells (Fig. 3D) as compared to negative staining of the class I null LCL.221 parent cell line (Fig. 3A). Of 25 oocytes, 15 (60%) stained positive with 1B8. No HLA-G staining was observed in cumulus cells (Fig. 2 C and D), although these cells stained with W6/32 (Fig. 2A and B). Positive staining with 1B8 was also observed in 2- (Fig. 2F), 4-, and 16-cell embryos (not shown) and in seven of nine expanded blastocysts. ICM staining was never observed with 1B8 (Fig. 2H) as opposed to W6/32 (Fig. 2J) The trophectodermal localization of staining was identical for both antibodies. No membrane staining was observed in oocytes or embryos stained with the secondary antibody alone. Positive f32m staining was observed in oocytes as well as in early cleavage stage embryo and for all cells examined. No staining was observed in controls with secondary antibody only.
HLA-G Expression Correlates with an Increased in Vitro Embryo Cleavage Rate. Blastocysts with detectable HLA-G mRNA had a significantly (P < 0.001) greater mean number of blastomeres per embryo at 24 hr and 48 hr after fertilization than was observed in HLA-G-negative blastocysts (Table 2).
DISCUSSION Medawar, in 1953, suggested that antigenic immaturity of human embryos might protect them from allogeneic recognition by T cells (24). However, when class I antigen expression on a cell surface is reduced or absent, recognition and lysis of these cells by NK cells is increased (25). Therefore, simple lack of classical class I molecules on trophoectoderm and later on invading cytotrophoblast would seem to be inadequate to explain their relative protection from T-cell activation and immunologic rejection. The presence of a nonpolymorphic nonclassical class I antigen, HLA-G, on the surface of invasive cytotrophoblast cells may be a partial explanation for this conundrum.
Table 2. Association between HLA-G/932m mRNA expression and embryo cleavage rate 24 and 48 hr after the presence of two pronuclei No. of blastomeres PCR results on No. of blastocysts 24 hr 48 hr embryos 3.05 ± 0.00 5.71 + 2.21 21 032m(-)/HLA-G(-) 3.31 ± 0.97 6.06 ± 1.62 68 32m(+)/HLA-G(-) 32m(+)/HLA-G(+) 3.86 ± 0.99* 8.66 ± 2.01* 59 HLA-G expression data are grouped according to whether RT-PCR detected HLA-G and ,12m mRNA [032m(+)/HLA-G(+)], I32m but not HLA-G [,B2m(+)/HLA-G(-)], or neither mRNA [,B2m(-)/HLAG(-)]. The numbers of blastomeres recorded at 24-hr and 48-hr after the appearance of two pronuclei in these embryos studied by RT-PCR at the blastocyst stage are presented as means ± SD. *Significantly greater number of blastomeres than in 132m(-)/HLA-G(-) embryos at the same time point (Mann-Whitney U test; P < 0.001).
Developmental Biology: Jurisicova et aL The series of experiments presented in this report demonstrate that the heavy and light chains of HLA-G, a nonclassical class I surface antigen, are expressed at the mRNA and protein levels by some human oocytes and some, but not all, preimplantation embryos at each cleavage stage up to the blastocyst. The earliest human embryonic tissue previously reported to express HLA-G mRNA was an 8-week-gestation placenta (26). The lack of HLA-G ICC immunoreactivity in the embryonic ICM is intriguing, since these cells stain with W6/32. It is important to note that, in contrast to trophectoderm, the ICM is not in direct contact with maternal tissues and thus not exposed to maternal immune cells. Our results suggest that ICM differentiation may be marked by suppression of HLA-G, concomitant to initiation of classical class I HLA synthesis. This would be consistent with the findings that HLA-G is observed in the placenta, but not the fetus, later in gestation (26), and that many blastocysts exhibit f32m transcript, but not HLA-G heavy chain mRNA expression. 032m in these embryos is probably associated with classical class I HLA expression on the membranes of the ICM. Data from many IVF clinics show that embryos coming from the same pool of mature fertilized oocytes are not necessarily synchronized in their rate of development. As early as 48 hr following fertilization, one can see differences in the number of blastomeres per embryo, ranging from two to six. In the course of the present study, we recorded differences in the rate of human embryo cleavage that were associated with HLA-G expression. It has been proposed (27, 28) that the cleavage rate of embryos may reflect their capacity to result in a successful pregnancy. The existence of nonclassical class I antigens is not unique to the human. Similarities at the DNA level (especially to murine genes Q7 or Q9) and its apparent lack of polymorphism suggest that HLA-G may be a structural homologue of the murine Qa-2 antigen (1). These molecules can be detected on murine oocytes and early-cleavage and blastocyst-stage embryos (29). Interestingly, the cleavage rate of mouse embryos has been shown to be linked to a gene designated "Ped" (preimplantation embryo development gene; ref. 30), which has been linked to the MHC and correlated with the presence or absence of the Qa-2 antigen (31). The Ped phenotype is an intrinsic property of the embryo, independent of the uterine environment (32). Removal of one form of Qa-2 antigen from the embryonic surface of fast-cleaving embryos results in slowing the rate of blastomere cleavage, thus mimicking the slow phenotype of the Ped gene (33). A recent report suggests that the Q9 gene may be involved, at least partially, in the Ped phenotype (34). The Qa-2 antigen may be a mediator of the transmission of mitogenic signals within the embryo and/or embryo and uterine environment in mouse (35). Since the Qa-2 antigen and HLA-G share some structural similarities, they may also function similarly, which could explain our findings. Previous studies have shown that HLA-G expression does not trigger generation of cytotoxic T lymphocytes (7) and confers resistance to NK cell lysis (20). Since NK-like cells are among the most abundant immune cells in the decidua (36), HLA-G might be necessary to protect the semiallogeneic embryo from the potentially hostile maternal immune system. Interestingly, in a previous study, the conditioned medium of 43% of embryos suppressed in vitro mitogen-induced lymphocyte proliferation (37). Moreover, the level of suppression in medium from the pre-embryos that later implanted was greater than that in embryos not resulting in a pregnancy (38, 39). HLA-G has a distinct acidic 37-kDa form that is secreted (2). It is possible therefore that there could be a connection between the embryonic HLA-G expression that we have observed and the immunosuppression previously observed with conditioned medium. In conclusion, the data presented in this report show that HLA-G is variably expressed during the critical period of
Proc. Natl. Acad. Sci. USA 93 (1996)
165
preimplantation embryo development. These findings suggest that HLA-G expression may play an important role in survival of the human embryo. We thank Dr. R. DeMars for his gift of the HLA-G transfected LCL.221 cells and Dr. Susan Fisher for the supply of 1B8 antibody. This work was supported by grants from the Medical Research Council of Canada, the Connaught Research Foundation, and The Genesis Research Foundation. 1. Geraghty, D. E., Koller, B. H. & Orr, H. T. (1987) Proc. Natl. Acad. Sci. USA 88, 9145-1499. 2. Kovats, S., Main, E. K., Librach, C. L., Stubblebine, M., Fisher, S. J. & DeMars, R. (1990) Science 248, 220-223. 3. Ellis, S. (1990) Am. J. Reprod. Immunol. 23, 84-86. 4. Morales, P., Corell, A., Martinez-Laso, J., Martin-Villa, J. M., Varela, P., Paz-Artal, E., Allende, L. M. & Arnaiz-Villena, A. (1993) Immunogenetics 38, 323-331. 5. Shukla, H., Swaroop, A., Srivastava, R. & Weismann, S. M. (1990)
Nucleic Acids Res. 18, 2189. 6. Ishitani, A. & Geraghty, D. E. (1992) Proc. Natl. Acad. Sci. USA 89, 3947-3951. 7. Kovats, S., Librach, C. L., Fisch, P., Main, E. K., Sondel, P. M., Fisher, S. J. & DeMars, R. (1991) Cell. Mol. Biol. Materno-Fetal Relat. 212, 21-29. 8. Dohr, G. A., Motter, W., Leitinger, S., Desoye, G., Urdl, W., Winter, R., Wilders-Truschnig, M., Uchanska-Ziegler, B. & Ziegler, A. (1987) J. Immunol. 138, 3766-3770. 9. Desoye, G., Dohr, G. A., Motter, W., Winter, R., Urdl, W., Pusch, H., Uchanska-Ziegler, B. & Ziegler, A. (1988) J. Immunol. 140, 4157-4159. 10. Roberts, J. M., Taylor, C. T., Melling, G. C. Kingsland, C. R. & Johnson, P. M. (1992) Immunology 75, 202-205. 11. Goldbard S. B., Gollnick S. 0. & Warner C. M. (1984) J. Immunol. Methodol. 68, 137-146. 12. Warner, C. M., Gollnick, S. 0. (1993) Biol. Reprod. 48, 1082-1087. 13. Sprinks, M. T., Sellens, M. H., Dealtry, G. B. & Fernandez, N. (1993) Immunogenetics 38, 35-40. 14. Perarnau, B., Siegrist, C. A., Gillet, A., Vincent, C., Kimura, S. & Lemonnier, F. A. (1990) Nature (London) 346, 751-754. 15. Zijlstra, M., Bix, M., Simister, N. E., Loring, J. M., Raulet, D. H. & Jaenisch R. (1990) Nature (London) 344, 742-746. 16. Koller B. H., Marrack P., Kappler J. W. & Smithies, 0. (1990) Science 248, 1227-1230. 17. Sawicki, J. A., Magnuson, T. & Epstein, C. J. (1981) Nature (London) 294, 450-451. 18. Meziou, W., Chardon, P., Flechon, J. E., Kalil, J. & Vaiman, M. (1983) J. Reprod. Immunol. 5, 73-80. 19. Gonen, Y., Jacobson, W. & Casper, R. F. (1990) Fertil. Steril. 53, 282-287. 20. Shimizu, Y. & DeMars, R. (1989) J. Immunol. 142, 3320. 21. Brady, G. & Iscove, N. N. (1993) Methods Enzymol. 225, 611-623. 22. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, Plainview, NY). 23. McMaster M., Librach, C. L., Zhou, Y., Lim, K., Janatpour, M., DeMars, R., Kovats, S., Damsky C. & Fisher, S. J. (1995) J. Immunol. 154, 3771-3778. 24. Medawar, P. B. (1953) Symp. Soc. Exp. Biol. 44, 320-338. 25. Shimizu, Y. & DeMars, R. (1989) Eur. J. Immunol. 19, 447-455. 26. Yelavarthi, K. K., Fishback, J. L. & Hunt, J. S. (1991)J. Immunol. 146, 2847-2854. 27. Sauer, M. V., Bustillo, M., Rodi, I. A. & Gorill, M. J. (1987) Hum. Reprod. 2, 701-703. 28. Bolton, V. N., Hawes, S. M., Taylor, C. T. & Parsons, J. H. (1989) J. In Vitro Fertilization Embryo Transfer 6, 30-35. 29. Warner, C. M., Gollnick, S. O., Flaherty, L. & Goldbard, S. B. (1987) Biol. Reprod. 36, 611-615. 30. Goldbard, S. B. & Warner, C. M. (1982) Biol. Reprod. 27, 419-424. 31. Warner, C. M., Gollnick, S. 0. & Goldbard, S. B. (1987) Biol. Reprod. 36, 606-610. 32. Brownell, M. S. & Warner, C. M. (1988) Biol. Reprod. 39, 806-811. 33. Tian, Z., Xu, Y. & Warner, C. M. (1992) Biol. Reprod. 47, 271-276. 34. Xu., Y., Jin P., Mellor A. L. & Warner C. M. (1994) Biol. Reprod. 51, 695-699. 35. Xu, Y., Jin, P. & Warner, C. M. (1993) Biol. Reprod. 48, 1042-1046. 36. Manaseki, S. & Searle, R. F. (1989) Cell Immunol. 121, 166-173. 37. Daya, S. & Clark, D. A. (1986) Am. J. Reprod. Immunol. 11, 98-101. 38. Clark, D. A., Lee, S., Fishell, S., Mahadevan, M., Goodall, M., Ah Moye, M., Stedronska-Clark, J., Daya, S., Underwood, J., Craft, I. & Mowbray, J. (1989) J. In Vitro Fertilization Embryo Transfer 6, 51-58. 39. Bose, R. & Mahadevan, M. M. (1992) Immunol. Invest. 21, 25-38.
