Endothelial cells in chorionic fetal vessels of first trimester placenta express HLA-G

Melanomas and melanoma cell lines do not express HLA-G, and the expression cannot be induced by gammaIFN treatment

HLA-G is an effective ligand of natural killer (NK) inhibitory receptors, HLA-G transcripts have been detected in several human tumors, and cytokines like gamma interferon (IFN) enable HLA-G molecules to be expressed. These findings are particularly upsetting in case of melanomas: IFN treatment is frequently included in melanoma therapeutic protocols, and downregulation of classical class I molecules occurs in nearly half of these tumors. Therefore, a melanoma cell downregulating classical class I and de novo expressing HLA-G, either constitutively or upon IFN treatment, is probably a stealthy target for the immune system, having inhibited both the cytotoxic T lymphocyte (CTL) and the NK activity. To elucidate this point we have investigated the expression of HLA-G molecules in 45 melanoma cell lines before and after gammaIFN treatment. Analysis was performed by immunofluorescence and flow cytometry, using the anti-HLA-G MoAbs 87G and G233, by Western blot, using the anti-HLA-G MEM/G1 MoAb and PAG1 antiserum, and by RT-PCR analysis. In addition, 8 melanoma tissues from patients free from therapy and 6 nevi were studied by immunohistochemistry using the 87G MoAb. No evidence was gathered of HLA-G expression, neither constitutive nor, in cell lines, after gammaIFN treatment. We therefore conclude that HLA-G expression is an uncommon event in melanomas, and that a therapy including IFNs cannot harm the patient by inducing the de novo expression of HLA-G molecules at least in its G1 isoform.

Heat shock and arsenite induce expression of the nonclassical class I histocompatibility HLA-G gene in tumor cell lines

The nonclassical histocompatibility class I gene HLA-G has a tissue-restricted expression. To explore mechanisms involved in HLA-G transcriptional regulation, we have investigated the effect of stress, including heat shock and arsenite treatment, on HLA-G expression in tumor cell lines. We show that stress induces an increase of the level of the different HLA-G alternative transcripts without affecting other MHC class I HLA-A, -B, -E, and -F transcripts. A heat shock element (HSE) that binds to heat shock factor 1 (HSF1) on stress conditions was further identified within the HLA-G promoter. Considering the ability of HLA-G to modulate the function of immunocompetent cells, we hypothesize a new feature of HLA-G as a signal regulating the immune response to stress.

Cutting edge: soluble HLA-G1 triggers CD95/CD95 ligand-mediated apoptosis in activated CD8+ cells by interacting with CD8

The nonpolymorphic soluble HLA-G1 (sHLA-G1) isoform has been reported to be secreted by trophoblast cells at the materno-fetal interface, suggesting that it may act as immunomodulator during pregnancy. In this paper, we report that affinity-purified beta2-microglobulin-associated sHLA-G1 triggered apoptosis in activated, but not resting CD8+ peripheral blood cells. We demonstrate by Western blotting that sHLA-G1 enhanced CD95 ligand expression in activated CD8+ cells. Cytotoxicity was inhibited by preincubation of the cells with a CD95 antagonist mAb (ZB4) or a soluble recombinant CD95-Fc, indicating that apoptosis is mediated through the CD95/CD95 ligand pathway. Finally, we show that such sHLA-G1-induced apoptosis depends on the interaction with CD8 molecules, with cell death being blocked by various CD8 mAbs.

Posttranscriptional regulation of human leukocyte antigen G during human extravillous cytotrophoblast differentiation

Human maternal tolerance to a semiallogenic fetus may be maintained, in part, by the unusual expression pattern of antigen-presenting molecules in placental trophoblast cells. Extravillous cytotrophoblast (EVC) cells, which invade the maternal decidua, express high levels of human leukocyte antigen G (HLA-G), a nonclassical, major histocompatibility complex (MHC) class I molecule. HLA-G transcripts have been detected in tumors and other tissues, yet protein accumulation is rare. We show that, within EVC cells themselves, the mRNA is more broadly expressed than the protein. Specifically, accumulation of HLA-G protein was markedly delayed during EVC cell differentiation. To elucidate this mechanism, we performed a comprehensive analysis comparing the expression of HLA-G and proteins essential for MHC class I expression at the cell surface. The transporter for antigen processing proteins TAP1 and TAP2, as well as tapasin and beta(2)-microglobulin, appeared to be coordinately expressed throughout EVC cell columns. Strikingly, they all accumulated well in advance of the HLA-G protein but concurrently with its mRNA. A similar delay in the accumulation of the HLA-G protein was observed in vitro, using cultures of chorionic villi. We conclude that posttranscriptional regulation of HLA-G is fundamental to EVC cell development and is achieved independently of the peptide loading system. This represents a novel mechanism of MHC class I regulation.

Comparative reactivity of different HLA-G monoclonal antibodies to soluble HLA-G molecules

Different HLA-G monoclonal antibodies (mAbs) were first evaluated for their capability to identify soluble HLA-G (sHLA-G) in ELISA. Three of them, namely 87G, BFL.1 and MEM-G/9, when used as coating mAbs together with W6/32 capture mAb, identified beta2-microglobulin (beta2m)-associated-sHLA-G but not soluble HLA-B7 (sHLA-B7) in cell culture supernatants from transfected cells. By comparison, the anti-HLA class I mAb 90 did recognize both sHLA-G and sHLA-B7. By using these HLA-G mAbs, sHLA-G was identified in amniotic fluids as well as in culture supernatants of first trimester and term placental explants but not in cord blood. Intron 4-retaining sHLA-G isoforms were identified in some amniotic fluids by the use of an intron 4-specific mAb (16G1). Reactivity of these different HLA-G mAbs was then compared to determine their respective binding sites on soluble and membrane-bound HLA-G. Using both ELISA and flow cytometry analysis, we showed that they did not compete with each other, which suggested that they did not recognize the same determinants. Finally, we report that two mAbs directed against the alpha1 domain of HLA class I heavy chain (mAb 90 and YTH 862) did compete with 87G, therefore demonstrating that this latter mAb recognized an epitope localized on this external domain of HLA-G.

The full length HLA-G1 and no other alternative form of HLA-G is expressed at the cell surface of transfected cells

In contrast to HLA class Ia, the HLA-G class Ib transcripts can be alternativeley spliced to yield several isoforms including four potentially membrane-bound variants, namely HLA-G1, -G2, -G3 and G4. It is so far unclear whether each of these splice variants lacking one or two external domains is properly translated and expressed at the cell surface. We used targeted Enhanced Green Fluorescence Protein (EGFP)-HLA-G fusion cDNA to track HLA-G isoform expression in living murine (L-human beta2m) and human (JAR) transiently transfected cells. It was demonstrated that the four HLA-G1, -G2, -G3, and -G4 isoforms were translated in these transfectants by the means of (i) Western blotting analysis, using an anti-EGFP mAb; (ii) intracellular double labeling flow cytometry analysis, using the EGFP natural fluorescence and phycoerythrin-labeled HCA2 anti-HLA-G mAb; and (iii) immunocytochemistry on isolated acetone fixed transfectants with the use of different anti-HLA-G mAbs. Cell surface flow cytometry analysis using the HCA2 mAb revealed that only the HLA-G1 isoform was expressed as a membrane-bound protein. Two color confocal microscopy performed on fixed, permeabilized cells further showed that the EGFP green fluorescence co-localized with anti-calnexin rhodamine fluorescence in the four HLA-G isoform transfectants but only in HLA-G1 transfectant was the green EGFP fluorescence also detectable at the outer part of the cells, suggesting that the HLA-G2, -G3, and G4 were retained in the endoplasmic reticulum. Such intracellular retention of the three shorter forms of HLA-G suggest that they may play a role in regulating cell surface expression either of the full length HLA-G1 form or of HLA-E.

Studies of mesenchymal cells from 1st trimester human placenta: expression of cytokeratin outside the trophoblast lineage

A fibroblast cell strain that expressed cytokeratins 8 and 18 was isolated from explanted first trimester placenta. Its properties were consistent with an origin in the villous fibroblast-myofibroblast lineage, including expression of vimentin, smooth muscle alpha-actin and fibroblast surface protein. The cells grew rapidly in vitro, and exhibited tightly aligned bipolar morphology at confluence, absence of multi-nucleated cells, lack of secreted chorionic gonadotrophin, and absence of HLA G, placental alkaline phosphatase and pregnancy-specific beta-1 glycoprotein (SP1). On these criteria it was concluded they were not trophoblasts. On the basis of their morphology, cytokeratin expression and absence of CD34 and endoglin it was concluded they were not endothelial cells. They lacked desmin and smooth muscle myosin and so were not vascular smooth muscle cells. Further mesenchymal cell isolates were studied to determine the generality of these findings. Phenotypic heterogeneity was a consistent characteristic, but cytokeratin-positive cells were always present both in first trimester and term strains. Desmin was absent from almost all the cells isolated using the protocols employed, despite its occurrence in a significant subpopulation of cells in the villous stroma. Cytokeratins 8 and 18 can be observed in the stromal compartment of both first trimester and term placental villi. Cytokeratin has hitherto been regarded as a highly reliable marker for cells of the trophoblast lineage in vitro. These observations suggest that care should be taken in characterization of placental cell isolates; trophoblasts should be identified by the presence of cytokeratin 7 in preference to cytokeratin 8/18. The functional significance of cytokeratin expression in placental mesenchymal cells remains to be established.

Primary cultured human thymic epithelial cells express both membrane-bound and soluble HLA-G translated products

This report demonstrates that both membrane-bound and soluble HLA-G isoforms are present in primary cultured human thymic epithelial cells (TEC). HLA-G transcriptional isoforms have been detected by RT-PCR, using different sets of HLA-G specific primers. A flow cytometry analysis, using two anti-HLA-G mAbs, namely 87G and BFL.1, revealed the presence of HLA-G translated products at the cell surface of a subpopulation of TEC. Finally, it was shown that HLA-G soluble forms were secreted in TEC culture supernatant, using a sandwich ELISA with BFL.1 and W6/32 mAbs. These results confirm and extent those previously described showing that HLA-G expressing cells were detectable by immunohistochemistry in thymic medullary epithelial cells.

Soluble HLA-G: purification from eukaryotic transfected cells and detection by a specific ELISA

PROBLEM

The detection of soluble forms of human leukocyte antigen-G molecule (sHLA-G) at the maternal-fetal interface suggest that sHLA-G may play a role during pregnancy. To study the potential functions of sHLA-G, we developed a procedure to detect and produce such HLA-G isoforms.

METHOD OF STUDY

Transfected cell lines expressing either sHLA-G1s cDNA (JAR-G1s) or an sHLA-G monochain DNA (Fox-G-mono) containing extracellular domains of HLA-G linked to the human beta 2-microglobulin were used. Specific sHLA-G enzyme-linked immunosorbent assay (ELISA), using anti-HLA-G monoclonal antibodies (mAbs) (87G and BFL.1) as coating antibodies and the biotinylated HLA class I mAb, W6/32, to reveal the bound molecules, was then developed.

RESULTS

To assess the specificity of the ELISA, we tested cell culture supernatants from the trophoblast-derived JEG-3 cell line and the HLA-G1s-transfected JAR cells, and we detected sHLA-G in both supernatants. sHLA-G monochain was also detected by ELISA in transfected cell supernatants using the conformational mAb, W6/32, showing that the conformation of sHLA-G monochain was proper. Using the same ELISA, sHLA-G was detected in various samples of amniotic fluid. To test the potential role of sHLA-G, sHLA-G has been purified by immunoaffinity columns, using W6/32 mAb, from culture supernatants of HLA-G1s or sHLA-G monochain-transfected cells.

CONCLUSION

These important tools will be useful both for the detection of sHLA-G in various biological fluids and in functional tests.

First trimester human endovascular trophoblast cells express both HLA-C and HLA-G

PROBLEM

In human pregnancies, trophoblasts, in contrast to placental connective tissue and the fetus itself, come into direct contact with the maternal allorecognizing system at special sites. Villous syncytiotrophoblasts washed around by maternal blood lack HLA class I proteins, whereas extravillous trophoblasts, which deeply invade maternal uterine tissues, express high amounts of HLA-G and also HLA-C, the latter to a lesser degree, however. A subpopulation of extravillous trophoblasts, the endovascular trophoblast, enters maternal spiral artery lumen and, like syncytiotrophoblast, comes into direct contact with maternal blood. Less is known about HLA class I distribution on this endovascular trophoblast subpopulation.

METHOD OF STUDY

A comparative immununohistochemical analysis was done on decidual cryo-sections containing trophoblast-invaded spiral arteries using different anti-HLA class I monoclonal antibodies (mAbs) and a peroxidase-labeled streptavidinbiotin detection system.

RESULTS

MAbs W6/32 (anti-HLA-A, -B, -C, -G), HCA2 (anti-HLA-A, -G) G233 and 87G (both anti-HLA-G) resulted in strong positivity on endovascular trophoblasts. L31 (anti-HLA-C) and HC10 (anti-HLA-B, -C) revealed clear positivity, whereas TU149 (anti-HLA-B, -C, some -A) produced a heterogeneous staining pattern, faintly positive on some endovascular trophoblastic cells and negative on others. MAb LA45 (anti-HLA-A, -B) did not bind to any endovascular trophoblast, neither did BFL.1 (anti-HLA-G) nor 16G1 (anti-HLA-G, soluble).

CONCLUSION

This study shows that trophoblastic cells belonging to the endovascular subpopulation express considerable amounts of HLA-G and slightly less HLA-C.

HLA-G in the human thymus: a subpopulation of medullary epithelial but not CD83(+) dendritic cells expresses HLA-G as a membrane-bound and soluble protein

The human MHC class Ib gene HLA-G is transcribed and translated in different placental cell subpopulations during pregnancy. In addition to this restricted tissue distribution, HLA-G proteins were also recently detected in the thymus of HLA-G transgenic mice, as well as in some human thymic epithelial cells (TEC). There was a need to further define the phenotype of the HLA-G-expressing cells in the human thymus as well as the type of translated forms that they produce. Using several HLA-G-specific mAb and immunohistochemistry performed on cryosections of human thymi at different ages, we found that the HLA-G-expressing cells are present on medullary cells exhibiting the epithelial morphological type 6. Co-localization experiments performed by double or triple immunofluorescence staining demonstrate that these HLA-G-expressing cells express various cytokeratins, epithelial cell markers but not the CD83 dendritic cell marker. We further show by ELISA measurements that a subset of primary cultured human TEC also expresses soluble HLA-G. Therefore, HLA-G protein tissue distribution is not restricted solely to placental cells. A subpopulation of medullary TEC also expresses HLA-G both at their cell surface and in secreted form, raising the question of the functional significance of such MHC class Ib molecules. Whether thymic soluble and/or membrane-bound HLA-G contribute to inhibit NK cells or to a negative selection of autoreactive T cells which could be harmful in case of pregnancy and/or to a positive selection of viral peptides/HLA-G-restricted CD8(+) T cells remains to be demonstrated.

Expression of HLA G in human tumors is not a frequent event

Expression of HLA G may be a way for tumor cells to escape immuno-surveillance. HLA G is selectively expressed by extravillous trophoblast in the human placenta, a tissue that does not express HLA A or B molecules. It is tempting to propose that tumor cells resemble this unique HLA class I phenotype as they frequently lose classical HLA A, B and C class I expression. Such peculiar HLA class I distribution would in theory allow tumor cells to escape from T- and NK-cell cytotoxicity. To determine whether HLA G is expressed on tumor cells, we studied HLA G mRNA levels using RT-PCR and HLA G cell-surface expression by immunohistological techniques in a panel of 50 human solid tumor tissues, 31 tumor cell lines of different origin, 4 autologous mucosa samples and 3 peripheral white cell samples. We found mRNA transcripts of different HLA G isoforms in most of the samples studied. However, we did not detect cell-surface expression of HLA G using 3 specific monoclonal antibodies (MAbs; 87G, 01G and G223). HLA G was detected only in the U937 myelomonocytic cell line after stimulation with IFN-gamma. We favor the hypothesis that HLA G plays a minor role, if any, in providing an inhibitory signal to NK cells to escape immunosurveillance. We cannot, however, exclude the possibility that some other HLA G isoforms may be expressed in some tumors.

IL-10 selectively induces HLA-G expression in human trophoblasts and monocytes

HLA-G plays an essential role in feto-maternal tolerance by inhibiting lysis by maternal NK cells. The factors that allow tissue-specific activation of HLA-G gene expression in trophoblasts remain to be characterized. We investigated the potential effect of IL-10, a cytokine which is secreted in placenta, on HLA-G gene transcription in trophoblasts. Using Northern blot, RNase protection assay and RT-PCR analysis, we demonstrated that IL-10 enhances steady-state levels of HLA-G transcription in cultured trophoblast cells. We further tested the effect of IL-10 on HLA-G gene transcription and protein expression in peripheral blood monocytes, showing that IL-10 can up-regulate HLA-G cell surface expression in this cell type. This effect of IL-10 is selective, since classical MHC class I products and MHC class II are down-regulated in monocytes following IL-10 treatment. Induction of HLA-G expression by IL-10 on monocytes may thus play a role in down-regulation of the immune response. We propose that IL-10 secretion by trophoblasts during pregnancy may also influence the HLA class I expression pattern at the feto-maternal barrier, thus protecting the fetus from rejection. This should be taken into consideration in the design of treatment for pathologies of pregnancy.

The functionality of HLA-G is emerging

In view of the recently published data, the HLA-G class Ib gene appears to be a functional locus. This is based on the following observations: 1) HLA-G is capable of presenting nonamer peptides and of exerting antigen-presenting functions; 2) HLA-G is a ligand for at least three natural killer (NK) and other cell inhibitory receptors of the immunoglobulin superfamily, namely leukocyte immunoglobulin-like receptor-1/immunoglobulin-like transcript (ILT)-2, ILT-4 and p49; 3) in addition to the extravillous cytotrophoblast cells, HLA-G proteins have been detected in endothelial cells of placental chorionic villi, as well as in amniotic fluid and in some medullary thymic epithelial cells; 4) major histocompatibility complex (MHC) class Ib genes that share the unique characteristics of HLA-G, including a high expression in placenta, have been reported in other mammalian species. In addition to the classical MHC class I roles (antigen presentation and ligation to NK receptors inducing inhibitory and/or activatory signals), HLA-G is likely to exert other, novel functions: first, HLA-G was shown to be involved in the control of HLA-E expression by furnishing the appropriate class I leader sequence nonamer peptide; second, we hypothesize that HLA-G could be a regulator of placental angiogenesis; third, soluble HLA-G isoforms may act as specific immunosuppressors during pregnancy. Such functional properties, although incompletely understood, are likely to be important in the outcome of human pregnancies but also in normal adult life.

Molecular studies of trophoblast HLA-G: polymorphism, isoforms, imprinting and expression in preimplantation embryo

There is considerable interest in human HLA-G arising from the observation that it is expressed selectively on the surface of extravillous trophoblast, the fetal cell population directly in contact with the mother. We investigated several aspects of the molecular biology of this unusual molecule. Limited polymorphism at the nucleotide level, and even more restricted variation at the amino acid level, was found in our Caucasian population. A further unusual aspect of HLA-G is the occurrence of alternatively spliced mRNAs. Spliced messages that could give rise to either membrane-bound or soluble proteins have been reported and six of these alternative forms were detected in all first trimester and term placentae, highly purified villous and extravillous trophoblast and the cell lines, JEG-3 and 221-G. An additional novel splice variant involving loss of part of the 3'-untranslated region was observed with two alleles. Using a sensitive RNase protection assay higher levels of the membrane-bound RNAs as compared to the soluble forms were detected in first trimester and term placentae as well as in JEG-3. Contrary to previous findings our term samples taken from the maternal aspect showed higher levels of both mRNA species when compared to first trimester placenta. The question of imprinting was addressed through the detection of heterozygotes both in placental tissue and, more tellingly, in the purified trophoblast cells. There was no evidence of imprinting. In addition we did not find mRNA for HLA-G in human two to eight-cell embryos or in blastocyst or in sperm samples.

The maternal immune system in health and disease

Recent studies have clarified some characteristics of trophoblast and immune cells in the decidua and peripheral blood during pregnancy and the roles they play in the maternal immune system, and have added further evidence to the theory that maternal T helper cell type 1/2 cytokine balance shifts towards T helper cell type 2 dominance during pregnancy.

Soluble HLA-G in human placentas: synthesis in trophoblasts and interferon-gamma-activated macrophages but not placental fibroblasts

The HLA class Ib antigen, HLA-G, is highly expressed in early gestation placentas where it is believed to modulate maternal-fetal immunological interactions. In this study, soluble isoforms (sHLA-G) encoded by intron 4-retaining transcripts were identified in first trimester placentas by immunohistochemistry using a mAb specific for the C-terminus of sHLA-G. Immunoreactive sHLA-G protein was localized to trophoblast cells and to villous mesenchymal cells with the morphological features of macrophages. Reverse transcriptase polymerase chain reaction analysis which used primers specific for intron 4 and the 3' untranslated region of the HLA-G gene showed that transcripts encoding sHLA-G were present in the trophoblast-derived Jeg-3 cells as well as interferon-gamma-activated myelomonocytic U937 cells but were absent and uninducible in placental fibroblasts. These results indicate that placental sHLA-G is synthesized in trophoblast cells and activated placental macrophages and support the postulate that placenta-derived sHLA-G modulates maternal and fetal immune cell functions during pregnancy.