Volume 122, Issue 2 pp. 179-188

Free Access

Synthesis of β₂-microglobulin-free, disulphide-linked HLA-G5 homodimers in human placental villous cytotrophoblast cells

Pedro J. Morales,

Pedro J. Morales

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Judith L. Pace,

Judith L. Pace

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Jeralyn Sue Platt,

Jeralyn Sue Platt

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Daudi K. Langat,

Daudi K. Langat

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Joan S. Hunt,

Joan S. Hunt

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Pedro J. Morales,

Pedro J. Morales

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Judith L. Pace,

Judith L. Pace

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Jeralyn Sue Platt,

Jeralyn Sue Platt

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Daudi K. Langat,

Daudi K. Langat

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

Joan S. Hunt,

Joan S. Hunt

Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS, USA

Search for more papers by this author

First published: 02 May 2007

https://doi.org/10.1111/j.1365-2567.2007.02623.x

Citations: 61

Dr J. S. Hunt, Department of Anatomy & Cell Biology, University of Kansas Medical Center, 3901 Rainbow Boulevard, Mail Stop 3050, Kansas City, KS 66160–7400, USA. Email: [email protected]
Senior author: Dr J. S. Hnt

Share a link

Email
Wechat
Bluesky

Summary

Human leucocyte antigen-G (HLA-G) is a natural immunosuppressant produced in human placentas that binds differently to the inhibitory leucocyte immunoglobulin-like receptors LILRB1 (ILT2) and LILRB2 (ILT4) according to its biochemical structure. To predict the binding functions of the HLA-G5 soluble isoform synthesized in placental villous cytotrophoblast (vCTB) cells, we investigated structural features of this protein. Biochemical and immunological studies showed that vCTB cell HLA-G5 heavy (H)-chain proteins are disulphide-bonded homodimers unassociated with β₂-microglobulin (β2m) light-chain proteins. Although comparatively low levels of β2m messenger RNA (mRNA) were identified by real-time reverse transcription–polymerase chain reaction, immunoprecipitation studies failed to detect β2m protein even when specific mRNA was doubled by transduction of a lentivirus-β2m complementary DNA into vCTB cells. No abnormalities were identified in the translational start codon of vCTB cell β2m mRNA and differentiation into syncytium did not promote β2m synthesis. The failure of vCTB cells to exhibit β2m in vitro was paralleled by a lack of detectable β2m in vCTB cells in vivo. Lack of the β2m protein could be the result of low levels of β2m transcripts or of as yet unidentified translational defects. Experiments with recombinant ectodomains of LILRB indicate that β2m-free HLA-G binds strongly to LILRB2, a receptor that is expressed by macrophages. This potentially immunosuppressive cell type is abundant in the pregnant uterus. Thus, our findings are consistent with the postulate that the natural β2m-free homodimeric form of HLA-G5 synthesized in primary vCTB cells could comprise a particularly effective tolerogenic molecule at the maternal–fetal interface.

Abbreviations:

β2m: β₂-microglobulin
H-chain: heavy-chain
HLA-G: human leucocyte antigen-G
LILRB: inhibitory leucocyte immunoglobulin-like receptors
mAb: monoclonal antibody
MC: mesenchymal cell
Treg: regulatory T cells
sTB: syncytial trophoblast
vCTB: villous cytotrophoblast.

Introduction

Reprogramming of immune cell functions into suppressive modes is believed to be a central feature of maternal tolerance to the semiallogeneic fetus.^1,2 Support for this idea is found in studies showing that high levels of human leucocyte antigen-G (HLA-G) are associated with successful implantation of in vitro cultured embryos and the finding that low levels produced from specific HLA-G alleles are associated with fertility disorders.^3–7 Recent studies indicate that suppression may be promoted by binding of soluble or membrane-bound HLA-G produced in placentas to inhibitory receptors on leucocytes known as LILRB1 (the inhibitory leucocyte immunoglobulin-like receptor 1; also called ILT2, CD85j) and LILRB2 (also called ILT4, CD85d).^1,2^,⁸ Immune cells driven into suppressive modes by HLA-G include CD8⁺ lymphocytes, natural killer (NK) cells, activated macrophages and CD4⁺ CD25⁺ cells.^9–12

The HLA-G gene differs in significant ways from other HLA class I genes. In particular, it is characterized by alternative splicing of its single transcript to yield seven different messenger RNAs (mRNAs), four of which encode membrane proteins (HLA-G1, -G2, -G3, -G4) and three of which encode soluble proteins (HLA-G5, -G6 and -G7).^1,2 HLA-G was first identified in placental trophoblast cells, the unique lineage of cells derived from the trophectoderm layer of the blastocyst that interfaces directly with maternal uterine and blood cells. Subsequently, specific HLA-G isoforms derived from the array of messages were discovered to be differentially distributed according to cell type and anatomic location. In particular, it has been learned that although membrane isoforms are present on invading cytotrophoblast cells, they are absent on both villous cytotrophoblast (vCTB) cells and syncytiotrophoblast (sTB) comprising the placental villi.^1,2

Regarding soluble isoforms, vCTB cells underlying the sTB synthesize one of these proteins, HLA-G5, but not a second, HLA-G6.¹⁰ Several specific biochemical features of recombinant HLA-G5 produced in HEK293 cells are known¹⁰ but those of the HLA-G5 produced in primary vCTB cells in normal placentas have not been reported. The question of whether the cells produce monomers or disulphide-bonded dimers may be critical; Shiroishi et al.¹³ have demonstrated that the avidity of binding of HLA-G1 dimers to LILRB1 and LILRB2 is significantly greater than the binding of HLA-G1 monomers. In the same study, binding of HLA-G1 dimers to LILRB1 enhanced signals leading to immunosuppression. HLA-G1 dimerization is achieved via disulphide bonding in the α1 region of the molecule.¹⁴ The same structure may occur in HLA-G5, which is identical to HLA-G1 in the α1, α2 and α3 domains^1,2 and is produced by cells in placental villi.¹⁵

Not only is the question of heavy-chain : heavy-chain (H:H) dimer formation via disulphide bonding uninvestigated in HLA-G5-producing primary placental vCTB cells, there is also no information on whether these cells produce the HLA class I light-chain, β₂-microglobulin (β2m), which is required for forming heavy-chain : chain (H : L) heterodimers. Production of β2m-free H-chains would not be unique to placentas; free H-chains of other class I HLAs that include HLA-B27¹⁶ are found in activated and neoplastic cells. The question of association of β2m is important because very recently, Shiroishi et al.¹⁷ demonstrated that association of β2m with HLA-G1 is not required for recognition by LILRB2.

The experiments reported here aimed to establish the biochemical characteristics of the HLA-G5 produced in normal human villous placentas. The results show clearly that normal vCTB cells synthesize β2m-free, disulphide-bonded HLA-G5 H : H-chain homodimers. Taken together with reports of preferential binding of β2m-free H : H dimers to LILRB2, which are abundant on decidual macrophages,¹⁸ these findings strongly support the idea that HLA-G5 synthesized in the villous placenta may be particularly effective in driving mononuclear phagocytes¹¹ and other LILRB2-expressing cells into secretory profiles that benefit semiallogeneic human pregnancy.

Materials and methods

Antibodies and control reagents

Table 1 lists the antibodies used in this study. It includes concentrations utilized for immunohistology and immunoblotting, species of origin, target antigen, immunoglobulin subclass when appropriate and source or reference. Purified β2m protein from human urine (Sigma-Aldrich, St Louis, MO) was used as a positive control in some experiments.

Table 1. Antibodies used in this study

Antibodies	Species	Reference or source	Target	Concentration (μg/ml)
Antibodies	Species	Reference or source	Target	IHC	ELISA	IB
3H36	mouse IgG1	US Biologicals, Boston, MA	β₂m	6	–	1
2M2	mouse IgG1	BioLegend, San Diego, CA	β₂m	–	–	1
W6/32	mouse IgG2a	ref. 39	HLA class I	–	–	–
OV-TL 12/30	mouse IgG1	DAKO, Carpentaria, CA	cytokeratin-7	0·5	–	–
16G1	mouse IgG1	ref. 40	HLA-G5 and -G6	–	–	1
Anti-actin	rabbit polyclonal	Sigma-Aldrich, St. Louis MO	Actin	–	–	0·5
Anti-V5-HRP	mouse IgG2a	Invitrogen, Carlsbad, CA	V5 viral peptide	–	–	1
Anti-mouse-IgG-HRP	goat polyclonal	Jackson ImmunoResearch, West Grove, PA	mouse H + L IgG	–	–	0·04
Anti-rabbit-IgG-HRP	goat polyclonal	Jackson ImmunoResearch, West Grove, PA	Rabbit H + L IgG	–	–	0·04
Anti-mouse-IgG-biotin	horse polyclonal	Vector, Burlingame, CA	mouse H + L IgG	10	–	–
Control IgG₁	mouse IgG1	BD Biosciences, San Diego, CA	unknown	20	–	1
Control IgG_2a	mouse IgG2a	BD Biosciences, San Diego, CA	unknown	20	–	–

IHC, immunohistochemistry; ELISA, enzyme-linked immunosorbent assay; IB, immunoblot.

Acquisition of tissues, purification of cells and reagents

Samples of placentas were obtained from normal deliveries under a protocol approved by the Human Subjects Committee of this institution. First- and second-trimester samples were obtained from elective pregnancy terminations, and third-trimester (term) placentas were obtained from Caesarean sections performed to relieve or avoid fetal distress. For immunohistochemistry, which was performed as previously described,¹⁰ tissues were embedded in tissue-freezing medium (Triangle Biomedical Sciences, Durham NC), and stored at − 80° until sectioned by cryostat (5-μm sections onto glass slides). Villous CTB cells were purified using a previously described protocol¹⁹ that included removal of HLA class I-positive cells by magnetic bead technology (Miltenyi Biotec Inc., Auburn, CA) using the monoclonal antibody (mAb) W6/32. Purity was consistently > 95% as assessed by cytokeratin-7 positivity. In some experiments, W6/32⁺ placental mesenchymal cells (MCs) that had bound to the beads were released according to the manufacturer's instructions.

Immunoblots and reducing experiments

For immunoblots of cell proteins, the cells were solubilized in radioimmuno-precipitation assay (RIPA) buffer (phosphate-buffered saline containing 1% nonidet-P40 and a complete protease inhibitor cocktail; Sigma-Aldrich). Proteins were separated by polyacrylamide gel electrophoresis (PAGE) then transferred to nitrocellulose membranes. The blots were developed with anti-mouse immunoglobulin G–horseradish peroxidase (IgG–HRP; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) and SuperSignal^® West Dura (Pierce Biotechnology Inc., Rockford, IL), and visualized using Hyperfilm™ enhanced chemiluminescence (Amersham Biosciences, Little Chalfont, UK). To establish the effects of reducing and non-reducing conditions, the protein samples were diluted in Laemmli buffer in the absence or presence of 100 mm dithiothreitol and boiled for 5 min before loading onto sodium dodecyl sulphate (SDS)–PAGE as described previosuly for HEK293 cells.¹⁰

Real-time polymerase chain reaction

Total RNA was isolated from HeLa cells (American Type Culture Collection, Cat. No. CCL-2, Rockville, MD), purified MCs were isolated from term placentas and matching harvests of vCTB cells were made from the same placentas. HeLa cells were cultured in Eagle's minimal essential medium with 2 mm l-glutamine and Earle's basic salt solution adjusted to contain 1·5 g/l sodium bicarbonate, 0·1 mm non-essential amino acids, 1·0 mm sodium pyruvate, and 10% fetal bovine serum (all from Cellgro^®, Mediatech, Inc., Herndon, VA) at 37° in 5% CO₂. For these experiments, vCTB cells were affinity column-purified twice. Complementary DNA was synthesized from 1·0 μg RNA as previously described²⁰ and was subjected to real-time polymerase chain reaction (PCR; Applied Biosystems, Foster City, CA) to determine the levels of β2m mRNA in the samples. The relative quantification method was used to estimate the levels of β2m in each sample, with glyceraldehyde-6-phosphate dehydrogenase (GAPDH) mRNA serving as the endogenous normalization control and HeLa cell mRNA serving as the calibrator. Before using this method, a validation study was conducted using serial dilutions of HeLa complementary DNA (cDNA). The results showed that the amplification efficiencies of GAPDH and β2m mRNA messages were equal and that a dilution of 1 : 25 of cDNA was optimal. Real-time PCR was performed in triplicate in 96-well MicroAmp™ optical reaction plates (Applied Biosystems, Foster City, CA). Each 25-μl reaction contained 1 × 6-carboxylfluorescein (FAM)-labelled β2m Taqman^® Gene Expression Assay or GAPDH endogenous control, Taqman^® Universal PCR Master Mix (Applied Biosystems) and cDNA diluted 1 : 25 in water. The PCR was performed using a 7500 Real-Time PCR System (Applied Biosystems) with an initial AmpErase uracyl-N-glycosylase (UNG) activation step conducted at 50° for 2 min followed by a 10-min incubation at 95° to activate the Amplitaq Gold^® polymerase enzyme. This was followed by 40 cycles of denaturation at 95° for 15 seconds and annealing/extension at 60° for 1 min. The data were collected and analysed using the ABI Sequence Detection Software, version 1·3.1 (Applied Biosystems).

Immunoprecipitation

Purified vCTB or peripheral blood mononuclear cells (PBMCs) (20 × 10⁶ cells per immunoprecipitation reaction) were lysed in RIPA buffer as previously described.²⁰ Approximately 5 mg total protein from a single harvest of vCTB or 0·5 mg total protein from PBMCs from a single donor, each contained in 0·5 ml RIPA buffer, were incubated with a mixture of 1·0 μg control immunoglobulin (IgG1 and/or IgG2a) or a mixture of two primary antibodies, 16G1, anti-HLA-G intron 4, a gift of D. Geraghty, Fred Hutchinson Cancer Research Center, Seattle, WA, and W6/32, anti-HLA class I, for 1 hr at 4°. Thereafter, 50 μl of beads (EZview™ Red Protein G Affinity Gel, Sigma-Aldrich) was added and the mixture was incubated for an additional 1 hr at 4°. The mixture was centrifuged for 30 seconds at 8200 g, unbound protein solution was discarded and beads were washed three times with 750 μl RIPA lysis buffer. The bound antigen–antibody complexes were eluted by adding 25 μl non-reducing Laemmli buffer and boiling the samples for 5 min. The entire sample was subjected to SDS–PAGE and transferred to nitrocellulose membranes (Schleicher & Schuell, Keene, NH). Immunoblots were performed to identify β2m as described above but employing a second mouse anti-human β2m mAb (2M2, BioLegend, San Diego, CA).

Lentivirus transduction and RT-PCR

For primary cell transductions, reagents were purchased from Invitrogen (Carlsbad, CA). To obtain a β2m lentiviral construct, the full-length human β2m cDNA was cloned into pENTRY™/D-TOPO^®, then inserted into pLenti6/V5-DEST by recombination using the Gateway^® LR Clonase II enzyme mix. The viral stocks were generated using HEK293FT cells and the ViraPower™ Lentiviral Expression System. The vCTB cells were seeded into six-well plates (2 × 10⁶ cells/well) in culture medium. After 24 hr the cells were rinsed and infected with LacZ-V5 or β2m-lentivirus (10⁻¹ to 10⁻⁵ dilutions) containing 6 μg/ml Polybrene^® (Sigma-Aldrich). After 24 hr the medium was exchanged with fresh medium and the cells were cultured for 3 days to allow protein expression. To distinguish endogenous from lentivirus-derived β2m mRNA the following primer sets were designed (Lasergene, DNAStar Inc., Madison, WI). For endogenous β2m, the forward primer was 5′-CGC TGG CGG GCA TTC CTG-3′ and the reverse primer was 5′-TGC GGC ATC TTC AAA CCT CCA T-3′ (57·8° annealing temperature, 431-base-pair product). For the lentivirus-derived β2m, the forward primer, which followed the transcriptional start site of the vector, was 5′-CCA CGC TGT TTT GAC CTC CAT AGA-3′ and the reverse primer was within the β2m coding region, 5′-GTT CAC ACG GCA GGC ATA CTC ATC-3′ (57·4° annealing temperature, 434-base-pair product). The PCR primers and conditions for HLA-G5 and HLA-G6 mRNA published previously²⁰ were used in these and all other experiments where HLA-G mRNAs were identified. PCR products were analysed by electrophoresis in agarose gels and were stained using ethidium bromide. All PCR products were sequenced for authenticity. The amplicons were subjected to densitometric evaluation.

Sequencing to test for mutations in the β2m gene

Endogenous and lentivirus β2m mRNA derived from vCTB were investigated for mutations. Corresponding cDNAs were extracted from agarose gels, purified (Qiagen, Valencia, CA) and TA-cloned into pGEM-Teasy vector (Promega, Madison WI). Plasmid DNA was extracted from positive E. coli DH5α transformants and both strands were sequenced using the ABII PRISM XL sequencing system (Biotechnology Support Facility, University of Kansas Medical Center). For analyses, the nucleotide sequences obtained from vCTB cells were aligned to human β2m mRNA (accession number NM_004048) using Lasergene (DNAStar).

Epidermal growth factor (EGF) experiments

To test for the effects of EGF, 6 × 10⁶ vCTB cells harvested as described above were seeded into 60-mm dishes in 3 ml Iscove's Dulbecco's modified Eagle's medium containing antibiotics, 2 mm glutamine and 10% fetal bovine serum. Cultures were incubated for 4 hr to allow for adherence, washed to remove non-adherent cells and cultured for the indicated times in the presence or absence of 100 ng/ml EGF (PeproTech, Rocky Hill, NJ) as described earlier.²¹ The methods for performing semiquantitative reverse transcription (RT) PCR and immunoblots of cell lysates are described above.

Results

vCTB cells produce HLA-G5 disulphide-bonded H-chain dimers

In the first set of experiments we investigated the ability of primary vCTB cells from term placentas to synthesize monomeric and dimeric forms of HLA-G5 and tested disulphide bonding. Proteins were obtained by lysis of vCTB cells that had been maintained in culture medium for 6 days. Figure 1(a) shows that the vCTB cells produced dimers migrating to a molecular weight (MW) of ∼ 74 000 under non-reducing conditions. Disulphide bonding was demonstrated by conducting the experiment under reducing conditions. Under these conditions, the dimers readily dissociated to yield monomers at ∼ 37 000 MW (Fig. 1b).

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Villous CTB cell HLA-G5 consists of disulphide-bonded H-chains. Villous CTB cells were cultured for 6 days to promote synthesis of HLA-G5. (a) Non-reducing conditions. HLA-G5 in vCTB cell lysates consists of ∼74 000 MW dimers. (b) Reducing conditions. vCTB cell HLA-G5 consists of ∼37 000 MW dimers. Fifty micrograms of cell lysate protein was loaded into each lane. Signals were detected by immunoblotting with the mAb to HLA-G intron 4 sequences present only in HLA-G5 and HLA-G6, 16G1, as described in the Materials and methods.

These results indicated that vCTB cells produce primarily HLA-G5 dimers under non-reducing conditions and that the H-chains forming the dimers are disulphide-bonded.

vCTB cells transcribe but do not translate β2m mRNA

Next, we investigated the ability of vCTB cells to produce the light-chains (β2m) required for generating H : L heterodimers. Highly purified vCTB cells from term placentas were tested for β2m mRNA using a semiquantitative RT-PCR and for β2m protein using immunoblotting.

As shown in Fig. 2(a), vCTB cell lysates subjected to semiquantitative RT-PCR contained β2m-specific mRNA. No RT-PCR product was seen when water was substituted for cDNA (Neg). In Fig. 2(b), Real-time PCR was used to determine levels of β2m mRNA in vCTB cells, comparing those levels with levels in matching harvests of placental MCs. Average levels of β2m mRNA in vCTB cells were 13-fold lower than in MCs. As illustrated in Fig. 2(c), proteins from three harvests of placental stromal cells residual from the W6/32 affinity columns that had been used to purify vCTB cell harvests demonstrated readily detectable β2m in immunoblots using a mouse mAb to β2m, 3H36. By contrast, lysates from the three matching harvests of purified vCTB cells failed to demonstrate detectable β2m protein. The positive control consisting of commercially available β2m yielded the ∼ 12 000 MW protein that is expected of authentic β2m protein (Fig. 2c).

To verify the lack of HLA-G5-associated β2m protein in vCTB cells, immunoprecipitation studies were performed (Fig. 3). Proteins were immunoprecipitated with a combination of anti-HLA class I mAb, W6/32, and anti-HLA-G intron 4, 16G1, then subjected to SDS–PAGE and blotted with a second mouse anti-β2m mAb (2M2). PBMCs served as a positive control (lane 4) and non-specific mouse IgG1 and IgG2a (lanes 1, 2 and 5) served as negative controls. PBMCs contained readily detectable β2m migrating to approximately 12 000 MW, as expected, whereas negative controls did not contain detectable β2m. As with the negative controls, β2m was undetectable in vCTB cells (lane 6).

These experiments showed that vCTB cells synthesize β2m mRNA at comparatively low but readily detectable levels, but do not demonstrate any immunohistochemically detectable β2m protein associated with their HLA-G5 H-chains.

β2m protein in vCTB cells cannot be induced by doubling vCTB cell β2m RNA

Because lack of β2m protein in vCTB cells could be the result of the low levels of β2m mRNA shown in Fig. 2(b), a lentivirus transduction system was used to increase their β2m mRNA. As illustrated in Fig. 4(a), transduction was successful, indicated by the immunoblotting of V5-β-galactosidase. Signal intensity was proportional to the input dose of infectious virus. Figure 4(b) shows that both transduced β2m mRNA (upper panel) and endogenous β2m mRNA (lower panel) were readily detected in the vCTB cells by RT-PCR. Signals were not the result of recognition of β2m genomic DNA, as shown by negative results when the mock reverse-transcribed samples (no reverse transcriptase used) were subjected to PCR. Mock transduction did not yield any signals in the Lenti6-β2m-transduced cells whereas endogenous β2m mRNA was present and unchanged by mock transduction. Comparisons of the strength of the signals of Lenti6-β2m mRNA and endogenous β2m mRNA indicated that levels of β2m mRNA were more than doubled at a virus dilution of 1 : 10 and were doubled at a dilution of 1 : 100.

Despite doubling of vCTB cell β2m mRNA, no β2m protein was detectable by immunoblotting at any level of virus input (Fig. 4c). In the same experiment, purified β2m used as a positive control yielded a positive signal at 2 ng, indicating that should any β2m protein be produced, its concentration was likely to be below this level.

These experiments suggested that conditions other than low levels of β2m are required to explain the lack of β2m protein in vCTB cells.

Sequencing of β2m mRNA

Impaired β2m translation as a result of mutations has been reported in tumour cells.^22–24 When endogenous and lentivirus-introduced vCTB cell β2m mRNA were sequenced and compared with the known sequence, no mutations were identified (data not shown).

Culture of vCTB cells with EGF does not promote synthesis of β2m

To ascertain whether or not stimulation of vCTB cells into syncytium²¹ might be required for translation of β2m mRNA, vCTB cells were cultured for 6 days in medium alone or in medium containing EGF to allow for protein synthesis, then were tested by semiquantitative RT-PCR and immunoblotting. The effects of EGF on induction of syncytialization were observed using an inverted microscope (data not shown).

Figure 5(a) shows RT-PCR results and Fig. 5(b) shows scanning densitometer values for mRNAs compared to β-actin mRNA. Culture of vCTB cells in medium containing EGF slightly diminished both β2m and HLA-G5 message levels when comparisons were drawn with β-actin mRNA (Fig. 5b). By contrast, HLA-G6 mRNA dramatically declined. No β2m signal appeared on immunoblots of vCTB cells incubated for 6 days in medium alone or medium containing EGF (upper panel, Fig. 5c) whereas β-actin was readily detected in both lysates (lower panel, Fig. 5c).

Thus, syncytialization achieved by EGF did not result in detectable β2m.

Absence of detectable β2m chain in human placental trophoblast cells in situ

The final group of experiments was designed to learn whether vCTB cells and/or sTB in situ contained immunohistochemically detectable β2m. Whether or not the HLA-G5 H-chain is associated with β2m-chain in vivo was uncertain as the scientific literature contains conflicting evidence on the presence of β2m in sTB.^25,26

Immunohistology was used to investigate the expression of β2m in placental villous trophoblast cells in situ during early (Fig. 6a), middle (Fig. 6b) and late (Fig. 6c) stages of pregnancy. Villous CTB cells did not contain detectable protein at any stage although light staining that could have been ingested β2m or neonatal FcR-associated β2m was occasionally observed in sTB (not shown). By contrast, the placental stromal cells were positive for β2m and the intensity of staining increased as gestation proceeded to term. No immunostaining signals were detected in the isotype-specific controls (Fig. 6d).

These data demonstrated that in situ, villous trophoblast cells uniquely fail to contain detectable levels of β2m for associating with HLA-G5 or other HLA H-chains in H : L-chain heterodimers.

Discussion

The results of the experiments described here show definitively that primary vCTB cells produce β2m-free, disulphide-linked HLA-G5 H : H homodimers. To the best of our knowledge, this study is the first to identify a type of cell that produces HLA-G5 class I H-chains but does not produce the light-chains. We present evidence indicating that one potential reason for production of H : H homodimers rather than H : L heterodimers in this trophoblast cell subpopulation is low transcription of the β2m gene, and a second may be an inability to translate β2m mRNA into protein. Our preliminary conclusions are in accord with those of Capps et al. for other types of cells lacking β2m²⁷ and the results of experiments on HLA-B27 conducted by Allen et al.¹⁶

The HLA-G5 protein produced in vCTB cells is a disulphide-bonded dimer that is dissociated under reducing conditions when disulphide bonds are interrupted. This finding was in agreement with the results of a previous study conducted on supernatant culture media from HEK293 cells stably transfected with HLA-G5.¹⁰ Disulphide bonding of the H-chains to form dimers was not entirely unexpected as site-directed mutagenesis has indicated that the two extra cysteine residues that characterize HLA-G, cys 42 and cys 147, are associated with disulphide-linked oligomerization.^8,14 However, these studies were limited to transfected cells and cell lines. Since there are no cysteines within the 21 amino acids encoded by HLA-G5 intron 4, it is highly likely that cys 42 and cys 147 form the disulphide bonds identified in this study.

The HLA-G5 produced in vCTB cells is also unusual in failing to have detectable β2m associated with H-chains. Biochemical experiments performed earlier on the products of the HLA-G5-transfected HEK293 cells¹⁰ indicated that their HLA-G5 H-chains were associated with endogenous β2m, strongly suggesting that HLA-G5 H-chains are fully capable of binding endogenous β2m. To determine whether vCTB cells similarly produced β2m for constructing H : L heterodimers, we examined vCTB cell β2m message and protein. Surprisingly, β2m mRNA was readily detected but protein was undetectable by immunoblotting and immunoprecipitation followed by immunoblotting, suggesting interference between transcription and translation. Since it seemed possible that vCTB cells might produce levels of β2m mRNA that were insufficient for generating detectable protein, levels were increased by delivering additional β2m mRNA using a lentivirus system. Doubling of β2m mRNA in vCTB cells also did not result in any detectable β2m protein, indicating that lack of β2m mRNA may not be the reason for lack of β2m protein in these cells. Whether the HLA-G5 H-chains produced in vCTB cells associate with β2m in sTB in an additional processing step remains to be seen as the question of production of β2m in the syncytium is unresolved.

Because impaired β2m translation has been associated in tumour cells with mutations that affected either the translational start codon or structural codons or with production of a defective message from an intact gene,^22–24β2m messages transcribed in vCTB cells were sequenced. No mutations that could have accounted for lack of β2m protein in vCTB cells were identified in either endogenous or lentivirus-introduced β2m mRNA. Presumably, the entirely normal vCTB cells tested in this study are less likely than tumour cells to have mutations that cause failure of translation of β2m mRNA.

Failure of production of β2m in vCTB cells was also unrelated to a requirement for syncytialization. Syncytialization naturally occurs when vCTB cells are placed in culture, and is increased when the culture medium contains EGF. Light-chain, β2m, has been reported in syncytium²⁶ although this remains controversial. However, culturing vCTB cells in EGF again failed to result in production of detectable β2m protein. Potential mechanisms underlying failure of vCTB cells to produce β2m that remain to be investigated include dysregulation of the target of rapamycin (ToR) pathway and interference by micro-RNAs.^28–30 Interestingly, lack of β2m synthesis in mammalian trophoblast cells has long been known; mouse placental giant trophoblast cells express H-2 class I H-chains but not light-chains.³¹ Collectively, the results suggest that that the β2m-free structure has selective functional advantages for trophoblast cells, perhaps in evading immune recognition.

Early studies on HLA-G in placentas concluded that production is entirely confined to the migrating CTB cells that attach to and infiltrate the decidua.³² Even at present, this view is considered viable by some³³ despite the fact that several independent laboratories^15,34–36 have identified HLA-G5 in the villous placenta. The underlying reason for error may be that the antibody most commonly used for identification of HLA-G in trophoblast cells is W6/32, which preferentially binds to HLA class I H-chains when they are associated with light-chain, β2m. Our observation that the vCTB cell layer produces a form of HLA-G5 that is not associated with light-chain may be the reason for the widespread misunderstanding³³ of sources of HLA-G in human placentas.

It seems reasonable to suggest that the disulphide-bonded homodimers of HLA-G5 H-chains produced in the villous placenta are secreted into maternal blood because soluble HLA-G has been identified in pregnancy sera.³⁷ This could be of importance biologically because (1) dimers bind with higher affinity to LILRB than monomers, and (2) association of the H-chain dimers with β2m may also change LILR binding affinities.^13,38 It will be of considerable interest to learn whether the homodimers of HLA-G5 produced in the vCTB cell layer reach maternal blood and how their affinity for the LILRB receptors compares with the HLA-G isoforms produced in the trophoblast cell columns and chorion membrane.

In summary, we here document for the first time the novel quaternary structure of HLA-G5 synthesized in primary human vCTB cells. Although the underlying cause for lack of β2m protein in vCTB cells is as yet unexplained, the data collected to date are consistent with the postulate that an intrinsic developmental programme prevents vCTB cells from synthesizing H : L dimers. The results raise several important questions: could recombinant HLA-G5 in its natural homodimeric structure be used to promote tolerance in women with reduced fertility? Are the enzyme-linked immunosorbent assays currently used to identify soluble HLA-G appropriately constructed for maximum sensitivity? These and other questions of translational application remain to be evaluated.

Acknowledgements

These studies are supported by grants from the National Institutes of Health to J.S.H. (PO1 HD39878 Project III; U54 HD33994 Project IV; PO1 HD40984 Project I). The authors appreciate assistance from S. Fernald of the University of Kansas School of Medicine Image Analysis Core, J. S. Platt and the Kansas IDeA Network for Biomedical Research Excellence (P20 RR16475, J. S. Hunt, P.I).

References

1 Hunt JS, Petroff MG, McIntire RH, Ober C. HLA-G and immune tolerance in pregnancy. FASEB J 2005; 19: 681–93.
10.1096/fj.04-2078rev
CAS PubMed Web of Science® Google Scholar
2 Hunt JS. Stranger in a strange land. Immunol Rev 2006; 213: 36–47.
10.1111/j.1600-065X.2006.00436.x
CAS PubMed Web of Science® Google Scholar
3 Pfeiffer KA, Rebmann V, Passler M, Van Der Ven K, Van Der Ven H, Grosse-Wilde H. Soluble HLA levels in early pregnancy after in vitro fertilization. Hum Immunol 2000; 61: 559–64.
10.1016/S0198-8859(00)00123-3
CAS PubMed Web of Science® Google Scholar
4 Yie S, Balakier H, Motamedi G, Librach C. Secretion of human leukocyte antigen-G by human embryos is associated with a higher in vitro fertilization pregnancy rate. Fert Steril 2005; 83: 30–6.
10.1016/j.fertnstert.2004.06.059
CAS PubMed Web of Science® Google Scholar
5 Aldrich CL, Stephenson MD, Karrison T et al. HLA-G genotypes and pregnancy outcome in couples with unexplained recurrent miscarriage. Mol Hum Reprod 2001; 7: 1162–72.
10.1093/molehr/7.12.1167
Web of Science® Google Scholar
6 O'Brien M, McCarthy T, Jenkins D, Paul P, Dausset J, Carosella ED, Moreau P. Altered HLA-G transcription in pre-eclampsia is associated with allele specific inheritance: possible role of the HLA-G gene in susceptibility to the disease. Cell Mol Life Sci 2001; 58: 1943–9.
10.1007/PL00000828
CAS PubMed Web of Science® Google Scholar
7 Sher G, Keskintepe L, Batzofin J, Fish J, Acacio B, Ahlering P, Ginsburg M. Influence of early ICSI-derived embryo sHLA-G expression on pregnancy and implantation rates: a prospective study. Hum Reprod 2005; 20: 1359–63.
10.1093/humrep/deh758
CAS PubMed Web of Science® Google Scholar
8 Shiroishi M, Tsumoto K, Amano K et al. Human inhibitory receptors Ig-like transcript 2 (ILT2) and ILT4 compete with CD8 for MHC class I binding and bind preferentially to HLA-G. Proc Natl Acad Sci USA 2003; 100: 8856–61.
10.1073/pnas.1431057100
CAS PubMed Web of Science® Google Scholar
9 Fournel S, Aguerre-Girr M, Huc X, Lenfant F, Alam A, Toubert A, Bensussan A, Le Bouteiller P. Soluble HLA-G1 triggers CD95/CD95 ligand-mediated apoptosis in activated CD8+ cells by interacting with CD8. J Immunol 2000; 164: 6100–4.
10.4049/jimmunol.164.12.6100
CAS PubMed Web of Science® Google Scholar
10 Morales P, Pace JL, Platt JS, Phillips TA, Morgan K, Fazleabas AT, Hunt JS. Placental cell expression of HLA-G2 isoforms is limited to the invasive trophoblast phenotype. J Immunol 2003; 171: 6215–24.
10.4049/jimmunol.171.11.6215
CAS PubMed Web of Science® Google Scholar
11 McIntire RH, Morales PJ, Petroff MG, Hunt JS. Soluble isoforms of HLA-G drive differential expression of anti-inflammatory cytokines by the U937 human macrophage cell line. J Leukocyte Biol 2005; 76: 1220–8.
10.1189/jlb.0604337
PubMed Web of Science® Google Scholar
12 Le Rond S, Azema C, Krawice-Radanne I, Durrbach A, Guettier C, Carosella ED, Rouas-Freiss N. Evidence to support the role of HLA-G5 in allograft acceptance through induction of immunosuppressive/regulatory T cells. J Immunol 2006; 176: 3266–76.
10.4049/jimmunol.176.5.3266
CAS PubMed Web of Science® Google Scholar
13 Shiroishi M, Kuroki K, Ose T et al. Efficient leukocyte Ig-like receptor signaling and crystal structure of disulfide-linked HLA-G dimer. J Biol Chem 2006; 281: 10439–47.
10.1074/jbc.M512305200
CAS PubMed Web of Science® Google Scholar
14 Boyson JE, Erskine R, Whitman MC et al. Disulfide bond-mediated dimerization of HLA-G on the cell surface. Proc Natl Acad Sci USA 2002; 99: 16180–5.
10.1073/pnas.212643199
CAS PubMed Web of Science® Google Scholar
15 Solier C, Aguerre-Girr M, Lenfant F, Campan A, Berrebi A, Rebmann V, Gross-Wilde H, Le Bouteiller P. Secretion of pro-apoptotic intron 4-retaining soluble HLA-G1 by human villous trophoblast. Eur J Immunol 2002; 32: 3576–86.
10.1002/1521-4141(200212)32:12<3576::AID-IMMU3576>3.0.CO;2-M
CAS PubMed Web of Science® Google Scholar
16 Allen RL, O'Callaghan CA, McMichael AJ, Bowness P. HLA-B27 can form a novel β2-microglobulin-free heavy chain homodimer structure. J Immunol 1999; 162: 5045–8.
CAS PubMed Web of Science® Google Scholar
17 Shiroishi M, Kuroki K, Rasubala L et al. Structural basis for recognition of the non-classical MHC molecule HLA-G by the leukocyte Ig-like receptor B2 (LILRB2/LIR2/ILT4/CD85d). Proc Natl Acad Sci USA 2007; 103: 16412–17.
10.1073/pnas.0605228103
PubMed Web of Science® Google Scholar
18 Petroff MG, Sedlmayr P, Azzola D, Hunt JS. Decidual macrophages are potentially susceptible to inhibition by class Ia and class Ib HLA molecules. J Reprod Immunol 2002; 56: 3–17.
10.1016/S0165-0378(02)00024-4
CAS PubMed Web of Science® Google Scholar
19 Phillips TA, Ni J, Hunt JS. Death-inducing tumor necrosis factor (TNF) superfamily ligands and receptors are transcribed in human placentas, cytotrophoblasts, placental macrophages, and placental cell lines. Placenta 2001; 22: 663–72.
10.1053/plac.2001.0703
CAS PubMed Web of Science® Google Scholar
20 Pace JL, Morales PJ, Phillips TA, Hunt JS. Analysis of the soluble isoforms of HLA-G. In: MJ Soares, JS Hunt, eds. Placenta and Trophoblast Methods and Protocols 2005122: 181–203.
10.1385/1-59259-989-3:181
Google Scholar
21 Maruo T, Matsuo H, Otani T, Mochizuki M. Role of epidermal growth factor (EGF) and its receptor in the development of the human placenta. Reprod Fertil Dev 1995; 7: 1465–70.
10.1071/RD9951465
CAS PubMed Web of Science® Google Scholar
22 Rosa RF, Berissi H, Weissenbach J, Maroteaux L, Fellous M, Reve M. The beta2-microglobulin mRNA in human Daudi cells has a mutated initiation codon but is still inducible by interferon. EMBO J 1983; 2: 239–43.
10.1002/j.1460-2075.1983.tb01412.x
CAS PubMed Web of Science® Google Scholar
23 Wang Z, Cao Y, Albino AP, Zeff RA, Houghton A, Ferrone S. Lack of HLA class I antigen expression by melanoma cells SK-MEL-33 caused by a reading frameshift in β₂-microglobulin messenger RNA. J Clin Invest 1993; 91: 684–92.
10.1172/JCI116249
CAS PubMed Web of Science® Google Scholar
24 Chang CC, Campoli M, Restifo NP, Wang X, Ferrone S. Immune selection of hot-spot β2-microglobulin gene mutations, HLA-A2 allospecificity loss, and antigen-processing machinery component down-regulation in melanoma cells derived from recurrent metastases following immunotherapy. J Immunol 2005; 174: 1462–71.
10.4049/jimmunol.174.3.1462
CAS PubMed Web of Science® Google Scholar
25 Faulk WP, Temple A. Distribution of β2 microglobulin and HLA in chorionic villi of human placentae. Nature 1976; 262: 799–802.
10.1038/262799a0
CAS PubMed Web of Science® Google Scholar
26 Leitner K, Ellinger A, Zimmer KP, Ellinger I, Fuchs R. Localization of β2-microglobulin in the term villous syncytiotrophoblast. Histochem Cell Biol 2002; 117: 187–93.
10.1007/s00418-001-0366-y
CAS PubMed Web of Science® Google Scholar
27 Capps GG, Robinson BE, Lewis KD, Zuniga MC. In vivo dimeric association of class I MHC heavy chains. Possible relationship to class I MHC heavy chain-beta 2-microglobulin dissociation. J Immunol 1993; 151: 159–69.
CAS PubMed Web of Science® Google Scholar
28 Martin PM, Sutherland AE. Exogenous amino acids regulate trophectoderm differentiation in the mouse blastocyst through an mTOR-dependent pathway. Dev Biol 2001; 240: 182–93.
10.1006/dbio.2001.0461
CAS PubMed Web of Science® Google Scholar
29 Inoki K, Corradetti MN, Guan KL. Dysregulation of the TSC-mTOR pathway in human disease. Nature Genet 2005; 37: 19–24.
10.1038/ng1494
CAS PubMed Web of Science® Google Scholar
30 Kuersten S, Goodwin EB. The power of the 3′ UTR. translational control and development. Nat Rev Genet 2003; 4: 626–37.
10.1038/nrg1125
CAS PubMed Web of Science® Google Scholar
31 Jaffe L, Robertson EJ, Bikoff EK. Distinct patterns of expression of MHC class I and beta 2-microglobulin transcripts at early stages of mouse development. J Immunol 1991; 147: 2740–9.
CAS PubMed Web of Science® Google Scholar
32 McMaster MT, Librach CL, Zhou Y et al. Human placental HLA-G expression is restricted to differentiated cytotrophoblasts. J Immunol 1995; 154: 3771–8.
10.4049/jimmunol.154.8.3771
CAS PubMed Web of Science® Google Scholar
33 Blaschitz A, Juch H, Volz A, Hutter H, Daxboeck C, Desoye G, Dohr G. The soluble pool of HLA-G produced by human trophoblasts does not include detectable levels of the intron 4-containing HLA-G5 and HLA-G6. Mol Hum Reprod 2005; 11: 699–710.
10.1093/molehr/gah185
CAS PubMed Web of Science® Google Scholar
34 Hunt JS, Geraghty DE. Soluble HLA-G isoforms: technical deficiencies lead to misinterpretations. Mol Hum Reprod 2005; 11: 715–7.
10.1093/molehr/gah223
PubMed Web of Science® Google Scholar
35 Le Bouteiller P. Human villous trophoblasts and the lack of intron 4-retaining soluble HLA-G secretion: beware of possible methodological biases. Mol Hum Reprod 2005; 11: 711–13.
10.1093/molehr/gah211
PubMed Web of Science® Google Scholar
36 LeMaoult J, Rouas-Freiss N, Carosella ED. HLA-G5 expression by trophoblast cells: the facts. Mol Hum Reprod 2005; 11: 719–22.
10.1093/molehr/gah224
PubMed Web of Science® Google Scholar
37 Hunt JS, Jadhav L, Chu W, Geraghty DE, Ober C. Soluble HLA-G circulates in mothers during pregnancy. Am J Obstet Gynecol 2000; 183: 682–8.
10.1067/mob.2000.106762
CAS PubMed Web of Science® Google Scholar
38 Gonen-Gross T, Achdout H, Arnon TI et al. The CD85J/leukocyte inhibitory receptor-1 distinguishes between conformed and β2-microglobulin-free HLA-G molecules. J Immunol 2005; 175: 4866–74.
10.4049/jimmunol.175.8.4866
CAS PubMed Web of Science® Google Scholar
39 Barnstable CJ, Bodmer WF, Brown G, Galfre G, Milstein C, Williams AF, Ziegler A. Production of monoclonal antibodies to group A erythrocytes, HLA and other human cell surface antigens: new tools for genetic analysis. Cell 1978; 14: 9–20.
10.1016/0092-8674(78)90296-9
CAS PubMed Web of Science® Google Scholar
40 Lee N, Malacko AR, Ishitani A, Chen MC, Bajorath J, Marquardt H, Geragthy DE. The membrane-bound and soluble forms of HLA-G bind identical sets of endogenous peptides but differ with respect to TAP association. Immunity 1995; 3: 591–600.
10.1016/1074-7613(95)90130-2
CAS PubMed Web of Science® Google Scholar

Citing Literature

Volume122, Issue2

October 2007

Pages 179-188

Synthesis of β₂-microglobulin-free, disulphide-linked HLA-G5 homodimers in human placental villous cytotrophoblast cells