It has been reported that preimplantation human embryos secrete HLA-G, and the levels may be predictive of their ability to implant. However, it is not known which of the membrane-bound (HLA-G 1–4) and soluble (HLA-G 5–6) alternatively spliced forms are present, nor the developmental stage at which they appear. Therefore, we have investigated HLA-G mRNA isoform expression on single embryos at the two-, four-, six-, and eight-cell, morula, and blastocyst stages. The percentage of embryos expressing each HLA-G isoform mRNA increased with developmental stage, but contrary to expectation, HLA-G5 mRNA was not detected in single two- to eight-cell embryos and was only expressed by 20% of morulae and blastocysts. Similarly, soluble HLA-G6 mRNA was not detected until the blastocyst stage and then in only one-third of embryos. In contrast, labeling with MEM G/9 Ab (specific for HLA-G1 and -G5) was observed in 15 of 20 two- to eight-cell embryos and 5 of 5 blastocysts. This disparity between mRNA and protein may be due to HLA-G protein remaining from maternal oocyte stores produced before embryonic genome activation and brings into question the measurement of soluble HLA-G for clinical evaluation of embryo quality. Although HLA-G is expressed in the preimplantation embryo, later it is primarily expressed in the invasive trophoblast of the placenta rather than the fetus. Therefore, we have investigated whether down-regulation of HLA-G first occurs in the inner cell mass (precursor fetal cells) of the blastocyst and, in support of this concept, have shown the absence HLA-G1 and -G5 protein and mRNA.

Successful implantation in the human is dependent on the early embryo avoiding immune recognition and destruction by the maternal immune system. One of the key protective mechanisms is thought to be the selective expression of the nonclassical HLA class I gene HLA-G by the trophoblast in the absence of classical class I (HLA-A and HLA-B) Ags. In contrast to the classical class I Ags, HLA-G is virtually nonpolymorphic with a small number of HLA-G alleles encoding only three different amino acid primary sequences (1). In addition, HLA-G mRNA can be alternatively spliced into at least six transcripts, which encode four membrane bound isoforms (G1, G2, G3, and G4) and two soluble isoforms (G5 and G6, otherwise known as soluble HLA-G1 and soluble HLA-G2) (1). HLA-G1 is the full-length isoform containing eight exons and seven introns. Exons 2, 3, and 4 encode the α1, α2, and α3 domains, respectively. Exon 5 encodes the transmembrane region, and exon 6 encodes the intracellular region. The other HLA-G isoforms are alternatively spliced, shorter transcripts lacking regions complementary to one or more entire exons. Thus, HLA-G2 lacks exon 3, corresponding to the α2 domain; HLA-G3 lacks exon 3 and 4 and thus only has the α1 domain; and HLA-G4 lacks exon 4 and hence the α3 domain. HLA-G5 and -G6 retain intron 4, which contains a stop codon that prevents the transcription of the transmembrane region, resulting in the expression of the soluble proteins. A further splice variant of HLA-G (HLA-G7) has also been reported (2). This isoform contains intron 2, which has a stop codon, so that the resulting G7 protein would be soluble HLA-G comprised of only the α1 region. However, while the authors were able to demonstrate the presence of G7 proteins in extracts of transfected cells by Western blot analysis and immunoprecipitation, it could not be detected as a secreted protein and is therefore not considered further here. Although the full range of functions of HLA-G remain to be elucidated, both full-length membrane-bound (HLA-G1) and soluble (HLA-G5) forms have been shown to have immunoregulatory functions, including the inhibition of T cell activation and stimulating decidual NK cells and macrophages to produce cytokines that are beneficial to implantation (3).

HLA-G mRNA and protein have previously been shown to be expressed by RT-PCR and immunofluorescence staining in a proportion of cleavage stage (two- to eight-cell, day 2/3) human embryos and blastocysts (day 5/6) created by in vitro fertilization (IVF)3 (4, 5). The primers used for the RT-PCR in these original studies amplified all isoforms of HLA-G, so it was not known which specific isoforms (membrane bound or soluble) were present. Interestingly, it was noted that sibling embryos from patients that became pregnant were more likely to express HLA-G than embryos from patients that did not conceive as a result of their IVF cycles (5), suggesting that HLA-G expression enhances the ability of embryos to implant. This is supported by several recent reports showing that a proportion of cleavage stage IVF embryos secrete the soluble form of HLA-G into the culture medium. It was found that, in general, embryos that produced soluble HLA-G were much more likely to produce pregnancies when transferred back into the uterus than those that did not (6, 7, 8, 9, 10). This is an important new finding, which, if confirmed, has significant implications for the selection of the best embryos for transfer in IVF. Currently, embryos are selected for transfer at the cleavage stage (day 2/3) principally on the basis of morphology, which is not always reliable. The availability of a noninvasive, quantitative assay for a marker of implantation potential could revolutionize IVF practice by increasing success rates and lowering the risk of multiple pregnancy and by reducing the number of embryos transferred. However, to be clinically useful, a better understanding of which HLA-G isoforms are expressed at each stage of development and in what proportion of embryos they are found is required, and this was one aim of the current study.

HLA-G appears to be expressed throughout the early embryo developmental stages (cleavage stages, morula, and blastocyst), but later on in development, its expression is predominantly confined to the invasive extravillous cytotrophoblast of the placenta. Thus, at some point, HLA-G expression must be switched off in the cells that are destined to form the fetus and the villous core of the placenta. The most likely time for this to happen is at the blastocyst stage when differentiation into the trophectoderm (which goes on to form the trophoblast) and the inner cell mass (ICM) (which forms the fetus) first occurs. Therefore, we have investigated the expression of HLA-G isoforms in isolated ICM to determine whether its down-regulation in fetal tissues begins at this stage.

The human embryos in this study were donated with informed consent by couples attending the Oxford Fertility Unit, John Radcliffe Hospital. Ethics approval for the study was obtained from the Central Oxford Research Ethics Committee and the Human Fertilization and Embryology Authority. Ovarian stimulation was achieved by a combination of pituitary desensitization and gonadotropin stimulation protocol. Oocyte retrieval was performed 35 h following human chorionic gonadotropin (hCG). Embryos were initially cultured singly in 1 ml of universal IVF medium (MediCult) in a humidified atmosphere of 5% CO2 in air at 37°C. On day 2 following insemination, embryos surplus to treatment and freezing, which had been donated for research, was transferred to culture medium 2 of the Blastassist System (MediCult) for an additional 4 days, with the medium being changed on day 4 (11). Embryos were cultured singly throughout to retain their identity for morphological assessment and were harvested at the two-, four-, six-, eight-cell, morula, and blastocyst stages. Day 6 blastocysts were graded according to the criteria of Gardner and Schoolcraft (12). Only grade A–C embryos and grade 4 blastocysts were used for mRNA isolation or immunofluorescence studies.

The HLA-G expressing choriocarcinoma cell line, JEG-3 (American Type Culture Collection), was used as a positive control for these studies. JEG-3 cells were grown in DMEM:Ham’s F-12 (Sigma-Aldrich) supplemented with 10% FCS, glutamine, and antibiotics.

The zona pellucida was removed by briefly exposing the embryo to Tyrode’s acid solution (Sigma-Aldrich). Individual human embryos were lysed in 300 μl of lysis/binding buffer (Dynabeads mRNA DIRECT kit; Dynal Biotech) and stored at −70°C until mRNA isolation. Embryo mRNA was isolated using the Dynabeads mRNA DIRECT kit following the manufacturer’s instructions (13). The mRNA was taken up in 20 μl of diethylpyrocarbonate-treated water.

Eight sets of specific primers (Table I) for HLA-G mRNA isoform-nested PCR were designed using software Primer3 (14) based on published sequences retrieved from GenBank (accession no. NM_002127). In the first round of the nested PCR, the forward primer was located in exon 1 and the reverse primer was across exons 5 and 6 so that all HLA-G mRNA isoforms could be amplified. In the second round of PCR, primers were designed to span the link between two neighboring exons to individually distinguish each HLA-G mRNA isoform, namely G1, G2, G3, G4, G5, and G6. An additional set of primers was designed to amplify all isoforms of HLA-G together (designated pan HLA-G). All primers were synthesized commercially by Invitrogen Life Technologies. Both cDNA synthesis and the first round of PCR were performed in a single tube using SuperScript One-Step RT-PCR with the Platinum Taq kit (Invitrogen Life Technologies) and conducted using a PerkinElmer DNA thermal cycler 480. mRNA was reverse transcribed and amplified in a 50-μl volume containing 2× reaction mix, 0.4 μM forward and reverse primer, and RT/Platinum Taq Mix. The conditions for first-strand cDNA synthesis and predenaturation were 1 cycle of 45°C for 45 min and 94°C for 2 min. PCR amplification was conducted for 40 cycles using the following conditions: denaturation for 30 s at 94°C, annealing for 30 s at 55°C, and extension for 2 min at 68°C. Amplification was completed by a final extension at 68°C for 5 min.

Table I.

Nested RT-PCR primers for amplification of HLA-G mRNA isoforms

Target GenePrimer Sequence (5′-3′)Position in SequenceProduct Size (bp)
First PCR primer AAC CCT CTT CCT GCT GCT CT Exon 1 1004 
 CTC CTT TTC AAT CTG AGC TCT TCT Exon 5–6  
HLA-G1 GAG CGA GGC CAG TTC TCA Exon 2–3 576 
 AGG GAA GAC TGC TTC CAT CTC Exon 4–5  
HLA-G2 ACC AGA GCG AGG CCA ACC Exon 2–4 304 
 AGG GAA GAC TGC TTC CAT CTC Exon 4–5  
HLA-G3 GCT CCC ACT CCA TGA GGT ATT Exon 2 276 
 ACT GCT TGG CCT CGC TCT Exon 2–5  
HLA-G4 GGC CAG TTC TCA CAC CCT CCA Exon 2–3 290 
 AAG ACT GCT CCG CGC GCT Exon 3–5  
HAL-G5 ATA CCT GGA GAA CGG GAA GG Exon 3 363 
 AGG CTC CTG CTT TCC CTA AC Intron 4  
HLA-G6 ACC AGA GCG AGG CCA ACC Exon 2–4 339 
 GGC TCC TGC TTT CCC TAA CAG Intron 4  
Pan HLA-G CTG ACC CTG ACC GAG ACC T Exon 1 291 
 CTC GCT CTG GTT GTA GTA GCC Exon 2  
Target GenePrimer Sequence (5′-3′)Position in SequenceProduct Size (bp)
First PCR primer AAC CCT CTT CCT GCT GCT CT Exon 1 1004 
 CTC CTT TTC AAT CTG AGC TCT TCT Exon 5–6  
HLA-G1 GAG CGA GGC CAG TTC TCA Exon 2–3 576 
 AGG GAA GAC TGC TTC CAT CTC Exon 4–5  
HLA-G2 ACC AGA GCG AGG CCA ACC Exon 2–4 304 
 AGG GAA GAC TGC TTC CAT CTC Exon 4–5  
HLA-G3 GCT CCC ACT CCA TGA GGT ATT Exon 2 276 
 ACT GCT TGG CCT CGC TCT Exon 2–5  
HLA-G4 GGC CAG TTC TCA CAC CCT CCA Exon 2–3 290 
 AAG ACT GCT CCG CGC GCT Exon 3–5  
HAL-G5 ATA CCT GGA GAA CGG GAA GG Exon 3 363 
 AGG CTC CTG CTT TCC CTA AC Intron 4  
HLA-G6 ACC AGA GCG AGG CCA ACC Exon 2–4 339 
 GGC TCC TGC TTT CCC TAA CAG Intron 4  
Pan HLA-G CTG ACC CTG ACC GAG ACC T Exon 1 291 
 CTC GCT CTG GTT GTA GTA GCC Exon 2  

The second round of the nested PCR was performed using the Platinum PCR SuperMix High Fidelity kit (Invitrogen Life Technologies). Two microliters of the first-round PCR products were added to the second PCR Mastermix to a total volume of 50 μl containing Platinum PCR SuperMix HF and 0.4 μM forward and reverse primers. PCR was conducted for one cycle of 3 min at 94°C and then 40 cycles of 30 s at 94°C, 30 s at 58°C, and 1 min at 68°C. All PCR products were separated on 1.5% agarose gels and visualized with ethidium bromide. As a control, β-actin was amplified in parallel in all embryos using nested RT-PCR. The corresponding primer pairs are shown in Table I.

The JEG-3 choriocarcinoma cell line was used to validate the HLA-G nested RT-PCR. Amplicons of the nested PCR from JEG-3 cells were sequenced for confirmation using the BigDye Terminator version 3.1 Cycle Sequencing kit (Applied Biosystems).

The sensitivity of the nested RT-PCR was examined in two ways. First, total RNA was isolated from JEG-3 cells using the RNeasy Mini kit (Qiagen) according to the manufacturer’s instructions. Serial dilutions of 106, 105, 104, 103, 102, 101, and 1 pg of total RNA were subjected to HLA-G mRNA isoform nested RT-PCR, as above. Second, a suspension of 105 JEG cells was serially diluted to 104, 103, and 102, and micromanipulation was used to obtain suspensions containing 10 cells and 1 cell. The cells were transferred to lysing/binding buffer (Dynabeads mRNA DIRECT kit; Dynal Biotech), and messenger RNA was isolated using Dynabeads mRNA DIRECT kit (Dynal Biotech) and subjected to HLA-G mRNA isoform-nested RT-PCR.

ICM were isolated from human blastocysts using a standard immunosurgery technique (15). The blastocysts were briefly exposed to 0.5% protease (Sigma-Aldrich) to remove the zona pellucida. The zona-free blastocyst was then incubated with 50% broad-specificity rabbit anti-human serum (Sigma-Aldrich) for 30 min, which binds to Ags on the trophectoderm. The addition of 10% guinea pig complement (Sigma-Aldrich) for an additional 30 min causes Ab-dependent complement-mediated lysis of the trophectoderm. Lysed trophectoderm cells were removed by sucking the embryo through a small-bore glass pipette. The intact isolated ICM were either stained by immunofluorescence or lysed in 300 μl of lysis/binding buffer and stored at −70°C until mRNA isolation as above.

Human embryos with intact zona pellucidae or ICM were fixed in 3% paraformaldehyde and permeabilized with 0.1% Triton X-100. They were then incubated with the 25 μg/ml primary anti-HLA-G mAb MEM-G/9 (Serotec) overnight at 4°C, followed by a 1/75 dilution of the secondary donkey anti-mouse IgG Ab conjugated to FITC (The Jackson Laboratory) for 45 min at room temperature. Control embryos were stained using mouse IgG (DakoCytomation) in place of the MEM-G/9 Ab. JEG-3 cells were used as positive controls. The embryos and ICM were mounted in Vectashield (Vector Laboratories) mounting medium containing 4′,6′-diamidino-2-phenylindole to counterstain the nuclei, and images were captured using a Leitz DMRBE microscope (Leica Microsystems) and Openlab imaging software (Improvision).

The JEG-3 cell line was shown to express all HLA-G alternatively spliced mRNA variants (Fig. 1) with the PCR products corresponding to the predicted sizes shown in Table I. Complementary DNA sequencing confirmed that the nested RT-PCR amplicons were identical to the regions spanned by the appropriate forward and reverse primers (data not shown).

FIGURE 1.

HLA-G mRNA isoforms in the JEG-3 cell line. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control.

FIGURE 1.

HLA-G mRNA isoforms in the JEG-3 cell line. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control.

Close modal

Using JEG-3 cells, all HLA-G isoforms could be detected at RNA concentrations as low as 10 pg, and HLA-G3 and -G4 were still detectable in 1 pg of total RNA (Fig. 2). All HLA-G mRNA isoforms were amplified successfully from single JEG-3 cells.

FIGURE 2.

Determination of RT-PCR sensitivity. A, HLA-G mRNA isoforms from different concentrations of total RNA from JEG-3 cells. B, HLA-G mRNA isoforms from different numbers of JEG-3 cells.

FIGURE 2.

Determination of RT-PCR sensitivity. A, HLA-G mRNA isoforms from different concentrations of total RNA from JEG-3 cells. B, HLA-G mRNA isoforms from different numbers of JEG-3 cells.

Close modal

A total of 20 individual cleavage stage human embryos was examined for HLA-G mRNA isoform expression by nested RT-PCR (five two-cell, five four-cell, five six-cell, and five eight-cell). Fig. 3,A shows a typical HLA-G mRNA isoform RT-PCR from an eight-cell embryo, and Fig. 3,B shows representative results of embryos from each cleavage stage. Only 7 of 20 embryos expressed HLA-G as determined by the pan-HLA-G primers (summarized in Fig. 5). However, not all isoforms were expressed by the one embryo. One two cell embryo expressed G3 and G4, one four-cell embryo expressed G1, G3, and G4, and one six-cell expressed G3 and G4 while another expressed G1, G3, and G4. Three eight-cell embryos were positive, with two expressing G1, one expressing G2, three expressing G3, and two expressing G4. No soluble G5 and G6 were detected in any of these embryos.

FIGURE 3.

HLA-G mRNA isoforms in human embryos from 2-cell to morula stage. A, HLA-G mRNA isoform expression in a typical human 8-cell embryo. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control. B, HLA-G mRNA isoforms from representative 2-cell (2c), 4-cell (4c), 6-cell (6c), and 8-cell (8c) embryos and morulas (Mo).

FIGURE 3.

HLA-G mRNA isoforms in human embryos from 2-cell to morula stage. A, HLA-G mRNA isoform expression in a typical human 8-cell embryo. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control. B, HLA-G mRNA isoforms from representative 2-cell (2c), 4-cell (4c), 6-cell (6c), and 8-cell (8c) embryos and morulas (Mo).

Close modal
FIGURE 5.

Summary of HLA-G mRNA isoform expression in human preimplantation embryos. 1, HLA-G1; 2, HLA-G2; 3, HLA-G3; 4, HLA-G4; 5, HLA-G5; 6, HLA-G6; and P, Pan HLA-G.

FIGURE 5.

Summary of HLA-G mRNA isoform expression in human preimplantation embryos. 1, HLA-G1; 2, HLA-G2; 3, HLA-G3; 4, HLA-G4; 5, HLA-G5; 6, HLA-G6; and P, Pan HLA-G.

Close modal

In addition, we made two pools of two- to four-cell human embryos (each containing mRNA from 10 embryos). G4 was found in one pool and G5 in the other (data not shown). Two similar mRNA pools of 10 six- to eight-cell embryos were made and HLA-G1, -G3, and -G4 expression was found in both (data not shown).

Five individual day 4 human morulas were next examined. All of them expressed HLA-G as detected by the pan-G primers. G1 was detected in three morulas, G2 in one, G3 in all five, and G4 in four. Soluble HLA-G5 was found in one morula. Fig. 3,B shows two typical results. Finally, 25 individual day 6-expanded blastocysts (grade 4) were studied. All of them were found to express HLA-G mRNA (Fig. 4, A and B), but the isoforms were differentially expressed. G1 was detected in 20 of 25 (80.0%), G2 in 4 of 25 (16.0%), G3 in 25 of 25 (100%), G4 in 24 of 25 (96.0%), G5 in 5 of 25 (20%), and G6 in 8 of 25 (32%).

FIGURE 4.

HLA-G mRNA isoforms in human blastocysts. A, HLA-G mRNA isoform expression in a typical day 6 human blastocyst. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control. B, HLA-G mRNA isoforms from 10 representative blastocysts.

FIGURE 4.

HLA-G mRNA isoforms in human blastocysts. A, HLA-G mRNA isoform expression in a typical day 6 human blastocyst. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control. B, HLA-G mRNA isoforms from 10 representative blastocysts.

Close modal

Overall, there was an increase in HLA-G mRNA expression with developmental stage with the most abundant isoforms being the truncated G3 and G4 forms (Fig. 5). The full-length G1 isoform was expressed in the majority of embryos by the blastocyst stage, but the full-length soluble form G5 was not expressed until the morula stage and was poorly expressed even at the blastocyst stage.

Twenty two- to eight-cell human embryos were stained with MEM-G/9, a mAb specific for HLA-G1 and G5 (Fig. 6). HLA-G labeling was found in 15 of 20 two- to eight-cell human embryos. Two of two two-cell embryos, five of eight four-cell embryos, two of three six-cell embryos, and six of seven eight-cell embryos were HLA-G positive. HLA-G-positive labeling was observed in all three morulas and five blastocysts examined (Fig. 6). No staining was found when nonspecific mouse IgG was used as a control for the primary Ab (Fig. 6).

FIGURE 6.

Immunofluorescence labeling for HLA-G protein in different developmental stages of human preimplantation embryos and ICM. Embryos and ICM were stained using MEM-G/9 (an HLA-G-specific mAb) and donkey anti-mouse IgG conjugated with FITC. Negative control embryos were stained with mouse IgG and donkey anti-mouse IgG conjugated with FITC. The nuclei were labeled with 4′,6′-diamidino-2-phenylindole.

FIGURE 6.

Immunofluorescence labeling for HLA-G protein in different developmental stages of human preimplantation embryos and ICM. Embryos and ICM were stained using MEM-G/9 (an HLA-G-specific mAb) and donkey anti-mouse IgG conjugated with FITC. Negative control embryos were stained with mouse IgG and donkey anti-mouse IgG conjugated with FITC. The nuclei were labeled with 4′,6′-diamidino-2-phenylindole.

Close modal

Three ICM were isolated from day 6 human blastocysts and subjected to mRNA extraction and nested RT-PCR for HLA-G mRNA isoforms. The only isoforms found were HLA-G3 and -G4, which were present in all three ICM. Fig. 7 shows a typical nested RT-PCR from one ICM. No immunofluorescence staining of the ICM with the MEM-G/9 was seen (n = 2) (Fig. 6).

FIGURE 7.

HLA-G mRNA isoforms in the ICM of human blastocyst. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control.

FIGURE 7.

HLA-G mRNA isoforms in the ICM of human blastocyst. The lanes are marked as follows: M, 1-kb DNA marker; G1, HLA-G1; G2, HLA-G2; G3, HLA-G3; G4, HLA-G4; G5, HLA-G5; G6, HLA-G6; PG, Pan HLA-G; and -ve, negative control.

Close modal

This study has confirmed that human preimplantation embryos express HLA-G mRNA and has shown that the proportion of embryos expressing it increases with developmental stage. To our knowledge, this is the first report investigating the expression profiles of alternatively spliced transcripts of HLA-G in human preimplantation embryos. Studying single embryos, these were found to be differentially expressed with the predominant forms being HLA-G3 and -G4 throughout. The full-length membrane bound (G1) and soluble forms (G5) and the truncated G2 and G6 were more varied in their expression, with G1 mRNA being present in 80% of blastocysts, while the soluble G5 was only present in 20% and soluble G6 in 32%.

Jurisicova et al. (4, 5) reported the expression of HLA-G mRNA in 43.3% of 148 of human blastocysts tested, which was much lower than in our study. They also reported that HLA-G mRNA was found in all preblastocyst development stages, including 2- to 4-, 5- to 8-, and 9- to 16-cell embryos and morulas. Surprisingly, no data on the proportions of embryos expressing HLA-G in each group was given, so comparison with our study is not possible. In another report, Hiby et al. (16) found no HLA-G mRNA in 11 preimplantation embryos ranging from the 2 cell to the blastocyst stage using nested primers for full-length HLA-G. Their methodology differed from ours in that they isolated RNA from zona intact embryos rather than removing the zona first and used standard phenol-chloroform extraction instead of the magnetic bead method, both of which may affect the yields of RNA. The primers used in both the studies of Jurisicova and Hiby would have amplified all the different isoforms, giving the total HLA-G mRNA expression, but the hemi-nested RT-PCR system used by Jurisicova et al. (4, 5) was probably less sensitive than our specific nested RT-PCR. Hiby et al. (16) also used nested RT-PCR with the outside forward primer located at exon 3 and the reverse primer located at 3′ untranslated region. The inside forward primer was located at exon 5, and the reverse primer was located at 3′ untranslated region. This primer set cannot amplify HLA-G2 and -G3 because G2 lacks exon 3 and G3 lacks exon 3 and 4. Another possible reason is the differences in the quality of the embryos used as only better grade embryos were included in our study.

Using immunofluorescence staining with the HLA-G-specific mAb MEM-G/9, we found that all blastocysts and 75% (15 of 20) of the preblastocyst human embryos examined expressed HLA-G protein. This was much higher than HLA-G mRNA expression for the preblastocyst embryos (7 of 20, 35.0%). The Ab MEM-G/9 specifically recognizes full-length HLA-G1 and soluble HLA-G5 in association with β2-microglobulin (β2m) and does not react with HLA-G2, -G3, or -G4 (17). The relative absence of G1 and G5 mRNA, compared with protein in early embryos is unlikely to be due to a lack of sensitivity of the amplification because our dilution experiments showed that it is possible to amplify all HLA-G isoforms from a single JEG-3 cell and from as little as 10 pg of total RNA. Although there are no data for humans, the average total RNA in mouse 2-, 4-, 8- to 16-, and 32-cell embryos is 350, 240, 690, and 1470 pg, respectively (18), and in addition, the volume of the human embryo is four times larger than that of the mouse. HLA-G5 mRNA was found in two- to four-cell embryos when pools of 10 embryos each were used (thereby increasing the cell numbers to between 20 and 80), but only in one of the four pools studied. Although, it is not possible to know how many embryos in each pool were producing HLA-G5 mRNA, this is consistent with a low frequency of expression in these early developmental stages.

There is conflicting evidence from other studies of class I MHC protein expression on human embryos. Desoye et al. (19) were unable to detect staining with the pan class I Ab W6/32 (which recognizes both HLA-G1 and -G5) staining on three unfixed polyploid embryos at the two-, five-, and eight-cell stages. Using the same Ab, Roberts et al. (20) found no staining on three blastocysts. In contrast, Jurisicova et al. (4) used both W6/32 and the anti-HLA-G monoclonal 1B8 (raised against the α1 region) and found W6/32 to stain all preblastocyst development stages, including 2- to 4-, 5- to 8-, and 9- to 16-cell embryos, morulas, and two of five blastocysts. Similarly, the 1B8 Ab stained 2-, 4-, and 16-cell embryos and seven of nine blastocysts. Possible reasons for these discrepancies are differences in methodology and the quality of the embryos used. Previous studies have used unfixed (19), acetone fixed (20), and paraformaldehyde fixed embryos (4) with primary Ab incubation times between 30 min and 1 h. In our study, we used the HLA-G-specific mAb MEMG/9 on fixed and permeabilized embryos, which will reveal both cytoplasmic and surface expression, together with an overnight incubation with the Ab to increase the sensitivity of the technique. Furthermore, all of the embryos studied were good-quality diploid embryos (grade A–C embryos and grade 4 blastocysts), which was not the case in all the previous studies (19).

A possible explanation for the variance between HLA-G mRNA and protein expression in the early embryos is that the HLA-G protein at this stage may be of maternal origin. It is known that fertilized oocyte transcription is silenced in the early stages of embryo development and ∼90% oocyte maternal mRNAs degrade in the 2 cell stage (18). The store of proteins in the fertilized oocyte is sufficient to support embryo development to the 8 cell stage (21). Embryonic gene activation occurs at the 4-8 cell stage in human embryos. Consistent with this, our results show that the expression rate of HLA-G mRNA increased from the six- to eight-cell stage onward. Thus, before the four- to eight-cell stage, HLA-G protein may come from the oocyte stores while during and after the four- to eight-cell stage, new HLA-G transcription and translation occurs with embryonic genome activation. There is however conflicting evidence concerning HLA-G protein expression by oocytes. Dohr et al. (22) found no staining with the pan class I Ab W6/32 used immunofluorescence on three unfixed oocytes which had either failed to fertilize or had fertilized abnormally, while Roberts et al. (20) using the same Ab on acetone fixed oocytes found 2 of 11 to be positive. Jurisicova et al. (4) used both W6/32 and the anti-HLA-G monoclonal 1B8 (raised against the α1 region) on paraformaldehyde-fixed oocytes and found positive staining on 6 of 13 and 15 of 20 oocytes, respectively. Their findings were supported by parallel studies that showed the presence of HLA-G mRNA in 17 of 21 pools of 5-8 unfertilized oocytes (4). Unfortunately, it is not currently possible to measure HLA-G mRNA and membrane-bound protein in the same embryos, which would help to clarify these issues, but further studies measuring mRNA and secreted soluble HLA-G5 are in progress in our laboratory.

These findings have implications for studies of soluble HLA-G secretion as a marker of human embryo developmental potential. One study (6) has reported that soluble HLA-G was detectable in day 2–3 culture supernatants of preimplantation (6–10 cells) human embryos. Supernatants from 285 embryo cultures from 101 IVF treatment cycles were examined. As some embryos were cultured in groups, it was not possible to study individual embryos, and therefore, the results were divided into patients who had embryos that secreted soluble HLA-G (n = 75) and those that did not (n = 26). No pregnancies were obtained in the patients whose embryos did not secrete HLA-G, but 18 of 75 of the patients whose embryos produced soluble HLA-G became pregnant, suggesting that soluble HLA-G secretion is a prerequisite for successful implantation. In more recent studies, only embryos that had been cultured singly were examined, making the data easier to interpret (7, 8). In one study, 72 of 101 (71%) of women under 39 years old who had at least one HLA-G secreting embryo transferred achieved a pregnancy, compared with 13 of 58 (22%) in the soluble HLA-G-negative group (7). In the other, no pregnancies were obtained in 26 of 66 women whose transferred embryos did not secrete soluble HLA-G, while nine pregnancies occurred in the 40 of 66 women whose embryos did secrete soluble HLA-G (8). It has also been reported that soluble HLA-G secretion was independent of embryo grade, but the cleavage rate of embryos secreting soluble HLA-G was significantly higher than that of those lacking it, as was the live birth rate (9). A prospective study, in which at least one embryo known to be producing soluble HLA-G was transferred to the mother, has also shown significantly improved implantation and pregnancy rates (10).

The majority of these studies used an ELISA based on the Ab MEM-G/9 and W6/32 (which recognizes all class I H chains in association with β2m) and hence would only specifically measure HLA-G5 secretion or membrane-bound HLA-G1 cleaved from the cell surface. However, two other reports using different ELISA systems do not support these findings. Van Lierop et al. (23) used the mAb G233 (raised against mouse L-cells transfected with both human β2m and HLA-G) as a capture Ab and Ab 56B (raised against the α2 domain of HLA-G) as the reporter. Although they could detect recombinant HLA-G and native HLA-G in trophoblast culture supernatants and amniotic fluid, they were unable to detect soluble HLA-G in follicular fluid or embryo culture supernatants from eight-cell, blastocyst, and late blastocyst stages. More recently, Noriko et al. (24) have used an ELISA based on the Abs 87G and MEMG/9 and were similarly unable to detect soluble HLA-G in 106 culture supernatants from day 3 embryos and blastocysts. The combinations of Abs used in both studies should detect both HLA-G1 and soluble HLA-G5.

Our results are more consistent with the latter findings as they show that the majority of cleavage stage embryos do not express HLA-G5 mRNA. It is only at the morula and expanded blastocyst stages that it is detectable and, even then, in no more than 20% of the embryos. These results also bring into question the specificity of the mAbs used, particularly MEMG/9, which we found to stain almost all embryos despite the absence of HLA-G1 and -G5 transcripts. Although the numbers studied are low due to the small numbers of human embryos available for research, as mentioned above, it is possible that HLA-G in 2-8 stage embryos is not from de novo transcription and translation but from stores in the oocytes, which are expressed in the early stages of embryo development. Thus, HLA-G expressed at this stage may be more a marker of oocyte rather than embryo quality.

A consistent finding in the current study was that at all stages of embryo development G3 and G4 were the predominant mRNA isoforms expressed. Ulbrecht et al. (25) report a similar finding in transfectants of the six different isoforms, with the amount of transcripts of G2, G3, G4, G5, and G6 exceeding by far that of G1. There is considerable debate about the function of the truncated isoforms and in particular whether they are expressed on the cell surface. Riteau et al. (26) have reported that HLA-G2, -G3, and -G4 can be expressed on the surface of transfected cells, thereby protecting these cells from lysis by both NK and CTL effector cells, and in this way may contribute to fetal survival (27) However, we (28) and others (25, 29) have found that although these truncated forms may be expressed inside the cell, there is no evidence for them reaching the cell surface. Because no mAbs specific for these isoforms are currently available, it is not possible to determine whether they are expressed on the surface of human embryos and have a functional role.

As mentioned above, only soluble HLA-G5 secretion has been measured in human embryos. However, the soluble HLA-G6 isoform has been reported to be secreted into the circulation of women throughout pregnancy and to be produced exclusively by extravillous cytotrophoblast cells (30). In the current study, HLA-G6 mRNA was not detected until the blastocyst stage, and therefore, it will be of great interest to determine whether HLA-G6 protein is secreted by human embryos and at what stage.

In later gestation, HLA-G is expressed predominantly by extravillous cytotrophoblast cells but not in general by fetal cells, which express classical HLA-A and B class I Ags. It is not known when this differential expression of HLA first occurs, but the most likely time is at blastocyst formation when fetal cells first develop as the ICM, while trophoblast cells develop from the trophectoderm. The immunofluorescence staining shown here and in other studies (4, 5) shows the trophectoderm to be HLA-G positive. However, it is difficult to distinguish whether the ICM expresses HLA-G using this method. Therefore, we isolated the ICM from blastocysts using immunosurgery and analyzed HLA-G expression. We found that the inner cell mass still contained HLA-G mRNA, but only the G3 and G4 isoforms were detectable. This was consistent with finding that the ICM did not stain with MEM-G/9, which only recognizes G1 and G5. These results support the concept of a down-regulation of HLA-G in the ICM as differentiation occurs, with perhaps the more abundant HLA-G3 and -G4 transcripts being the slowest to degrade.

If HLA-G was not down-regulated initially in the inner cell mass, then all the cells of the developing fetus would be HLA-G positive, which is not the case. However, it has been shown recently that fetal erythroid progenitor cells and endothelial cells label with a mAb5A6G7 (raised against a peptide sequence specific for HLA-G5 and -G6) from as early as day 32 of pregnancy (31). Therefore, it must be assumed that the HLA-G gene is reactivated in these distinct cell populations as they differentiate. It will be of great interest to also determine at what stage of development and in which fetal and placental tissues up-regulation of classical class I MHC HLA-A, -B, and -C Ags occurs.

We thank Drs. Enda McVeigh and Karen Turner and the patients and staff of the Oxford Fertility Unit for providing the human embryos for this study and Professor Helen Mardon and Janet Carver for their assistance with imaging.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

3

Abbreviations used in this paper: IVF, in vitro fertilization; ICM, inner cell mass; β2m, β2-microglobulin; hCG, human chorionic gonadotrophin.

1
Bainbridge, D., S. A. Ellis, P. Le Bouteiller, I. L. Sargent.
2001
. HLA-G remains a mystery.
Trends Immunol.
22
:
548
-552.
2
Paul, P., F. A. Cabestre, E. C. Ibrahim, S. Lefebvre, I. Khalil-Daher, G. Vazeux, R. M. Quiles, F. Bermond, J. Dausset, E. D. Carosella.
2000
. Identification of HLA-G7 as a new splice variant of the HLA-G mRNA and expression of soluble HLA-G5, -G6, and -G7 transcripts in human transfected cells.
Hum. Immunol.
61
:
1138
-1149.
3
Le Bouteiller, P., N. Pizzato, A. Barakonyi, C. Solier.
2003
. HLA-G, pre-eclampsia, immunity and vascular events.
J. Reprod. Immunol.
59
:
219
-234.
4
Jurisicova, A., R. F. Casper, N. J. MacLusky, G. B. Mills, C. L. Librach.
1996
. HLA-G expression during preimplantation human embryo development.
Proc. Natl. Acad. Sci. USA
93
:
161
-165.
5
Jurisicova, A., R. F. Casper, N. J. MacLusky, C. L. Librach.
1996
. Embryonic human leukocyte antigen-G expression: possible implications for human preimplantation development.
Fertil. Steril.
65
:
997
-1002.
6
Fuzzi, B., R. Rizzo, L. Criscuoli, I. Noci, L. Melchiorri, B. Scarselli, E. Bencini, A. Menicucci, O. R. Baricordi.
2002
. HLA-G expression in early embryos is a fundamental prerequisite for the obtainment of pregnancy.
Eur. J. Immunol.
32
:
311
-315.
7
Sher, G., L. Keskintepe, M. Nouriani, R. Roussev, J. Batzofin.
2004
. Expression of sHLA-G in supernatants of individually cultured 46-h embryos: a potentially valuable indicator of “embryo competency” and IVF outcome.
Reprod. Biomed. Online
9
:
74
-78.
8
Noci, I., B. Fuzzi, R. Rizzo, L. Melchiorri, L. Criscuoloi, S. Dabizzi, R. Biagiotti, S. Pellegrini, A. Menicucci, O. R. Baricordi.
2004
. Embryonic soluble HLA-G as a marker of developmental potential in embryos.
Hum. Reprod.
20
:
138
-146.
9
Yie, S. M., H. Balakier, G. Motamedi, C. L. Librach.
2005
. Secretion of human leukocyte antigen-G by human embryos is associated with a higher in vitro fertilization pregnancy rate.
Fertil. Steril.
83
:
30
-36.
10
Sher, G., L. Keskintepe, J. Batzofin, J. Fisch, B. Acacio, P. Ahlering, M. Ginsburg.
2005
. Influence of early ICSI-derived embryo sHLA-G expression on pregnancy and implantation rates: a prospective study.
Hum. Reprod.
20
:
1359
-1363.
11
Martin, K. L., D. H. Barlow, I. L. Sargent.
1998
. Heparin-binding epidermal growth factor significantly improves human blastocyst development and hatching in serum-free medium.
Hum. Reprod.
13
:
1645
-1652.
12
Gardner, D. K., W. B. Schoolcraft.
1999
. In vitro culture of human blastocysts. R. Jansen, and D. Mortimer, eds.
Towards of Reproductive Certainty: Fertility and Genetics Beyond
378
-388. Parthenon Press, Carmforth.
13
Yao, Y. Q., J. S. Xu, W. M. Lee, W. S. Yeung, K. F. Lee.
2003
. Identification of mRNAs that are up-regulated after fertilization in the murine zygote by suppression subtractive hybridization.
Biochem. Biophys. Res. Commun.
304
:
60
-66.
14
Rozen, S., H. J. Skaletsky.
2000
. Primer3 on the WWW for general users and for biologist programmers. S. Krawetz, and S. Misener, eds.
Bioinformatics Methods and Protocols: Methods in Molecular Biology
365
-372. Humana Press, Totowa.
15
Solter, D., B. B. Knowles.
1975
. Immunosurgery of mouse blastocyst.
Proc. Natl. Acad. Sci. USA
72
:
5099
-5102.
16
Hiby, S. E., A. King, A. Sharkey, Y. W. Loke.
1999
. Molecular studies of trophoblast HLA-G: polymorphism, isoforms, imprinting and expression in preimplantation embryo.
Tissue Antigens
53
:
1
-13.
17
Menier, C., B. Saez, V. Horejsi, S. Martinozzi, I. Krawice-Radanne, S. Bruel, C. Le Danff, M. Reboul, I. Hilgert, M. Rabreau, et al
2003
. Characterization of monoclonal antibodies recognizing HLA-G or HLA-E: new tools to analyze the expression of nonclassical HLA class I molecules.
Hum. Immunol.
64
:
315
-326.
18
Piko, L., K. B. Clegg.
1982
. Quantitative changes in total RNA, total poly(A), and ribosomes in early mouse embryos.
Dev. Biol.
89
:
362
-378.
19
Desoye, G., G. A. Dohr, W. Motter, R. Winter, W. Urdl, H. Pusch, B. Uchanska-Ziegler, A. Ziegler.
1988
. Lack of HLA class I and class II antigens on human preimplantation embryos.
J. Immunol.
140
:
4157
-4159.
20
Roberts, J. M., C. T. Taylor, G. C. Melling, C. R. Kingsland, P. M. Johnson.
1992
. Expression of the CD46 antigen, and absence of class I MHC antigen, on the human oocyte and preimplantation blastocyst.
Immunology
75
:
202
-205.
21
Nothias, J. Y., S. Majumder, K. J. Kaneko, M. L. DePamphilis.
1995
. Regulation of gene expression at the beginning of mammalian development.
J. Biol. Chem.
270
:
22077
-22080.
22
Dohr, G..
1987
. HLA and TLX antigen expression on the human oocyte, zona pellucida and granulosa cells.
Hum. Reprod.
2
:
657
-664.
23
Van Lierop, M. J., F. Wijnands, Y. W. Loke, P. M. Emmer, H. G. Lukassen, D. D. Braat, A. van der Meer, S. Mosselman, I. Joosten.
2002
. Detection of HLA-G by a specific sandwich ELISA using monoclonal antibodies G233 and 56B.
Mol. Hum. Reprod.
8
:
776
-784.
24
Noriko, S., H. Horotsugu, Y. Masanori, S. Takanori, O. Motoko, H. Katsuhiko, D. E. Geraghty, I. Akiko.
2004
. Are in vitro fertilized eggs able to secrete soluble HLA-G?.
Am. J. Reprod. Immunol.
52
: (Suppl. 1):
P8
25
Ulbrecht, M., S. Maier, V. Hofmeister, C. S. Falk, A. G. Brooks, M. T. McMaster, E. H. Weiss.
2004
. Truncated HLA-G isoforms are retained in the endoplasmic reticulum and insufficiently provide HLA-E ligands.
Hum. Immunol.
65
:
200
-208.
26
Riteau, B., N. Rouas-Freiss, C. Menier, P. Paul, J. Dausset, E. D. Carosella.
2001
. HLA-G2, -G3, and -G4 isoforms expressed as nonmature cell surface glycoproteins inhibit NK and antigen-specific CTL cytolysis.
J. Immunol.
166
:
5018
-5026.
27
Menier, C., B. Riteau, J. Dausset, E. D. Carosella, N. Rouas-Freiss.
2000
. HLA-G truncated isoforms can substitute for HLA-G1 in fetal survival.
Hum. Immunol.
61
:
1118
-1125.
28
Bainbridge, D. R. J., S. A. Ellis, I. L. Sargent.
2000
. The short forms of HLA-G are unlikely to play a role in pregnancy because they are not expressed at the cell surface.
J. Reprod. Immunol.
47
:
1
-16.
29
Mallet, V., J. Proll, C. Solier, M. Aguerre-Girr, M. DeRossi, Y. W. Loke, F. Lenfant, P. Le Bouteiller.
2000
. The full length HLA-G1 and no other alternative form of HLA-G is expressed at the cell surface of transfected cells.
Hum. Immunol.
61
:
212
-224.
30
Morales, P. J., J. L. Pace, J. S. Platt, T. A. Phillips, K. Morgan, A. T. Fazleabas, J. S. Hunt.
2003
. Placental cell expression of HLA-G2 isoforms is limited to the invasive trophoblast phenotype.
J. Immunol.
171
:
6215
-6224.
31
Menier, C., M. Rabreau, J. C. Challier, M. Le Discorde, E. D. Carosella, N. Rouas-Freiss.
2004
. Erythroblasts secrete the nonclassical HLA-G molecule from primitive to definitive hematopoiesis.
Blood
104
:
3153
-3160.