Mol. Hum. Reprod. Advance Access originally published online on December 5, 2005
Molecular Human Reproduction 2005 11(10):711-713; doi:10.1093/molehr/gah211
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Human villous trophoblast and the lack of intron 4-retaining soluble HLA-G secretion: beware of possible methodological biases
INSERM U563, Hôpital Purpan, Toulouse, France
 |
Introduction |
|---|
 Top Introduction Types of cellsReal-time PCRWestern blottingELISAImmunohistochemistryReferences |
|---|
The straightforward Blaschitz et al. (2005)
manuscript title makes the reader think that villous trophoblast cells in the human placenta definitely not secrete intron 4-retaining soluble HLA-G5 (also called soluble HLA-G1) molecules. Things might not be as simple... as the results presented by these authors contradict a number of previous reports on this subject (Chu et al., 1998; Solier et al., 2002; Ishitani et al., 2003; Morales et al., 2003). Rather than trying to resolve this issue, I like to focus on several experimental procedures and materials used by Blaschitz and collaborators and make the suggestion that some of them might have introduced possible biases that would possibly account for these discrepancies. Let us thus have a look at the materials and methods used by the authors. |
Types of cells |
|---|
|
TopIntroductionTypes of cellsReal-time PCRWestern blottingELISAImmunohistochemistryReferences |
|---|
Besides placental tissue sections, the authors used several cell types to sustain their conclusions. In contrast to previous studies performed on either purified term villous cytotrophoblast (Solier et al., 2002
; Morales et al. 2003) or extravillous trophoblast (Morales et al., 2003), the authors based several of their experiments on the use of a ‘freshly isolated primary first trimester trophoblast’ cell population containing 2% of cytokeratin 7 (trophoblast) negative cells, and 40% of HLA-G negative cells (immunohistochemical analysis). No flow cytometry analysis measuring HLA-G1, HLA-E and HLA-C on the cell surface (extravillous cytotrophoblast phenotype) or showing an absence of cell surface HLA-class I expression (villous cytotrophoblast phenotype) allowed the reader to assess the composition of this trophoblast/non-trophoblast mixture. A detailed phenotypic analysis of basal plate, chorion laeve and villous chorion samples is not provided either. The use of such a mixture of trophoblast cells makes the results of this study very difficult to compare with the previous reports based on separated populations of villous and extravillous trophoblast cells (Solier et al., 2002; Morales et al., 2003). Blaschitz et al. (2005) also used untransfected trophoblast-derived cell lines for their demonstration. Although useful controls, these cell lines exhibit obvious limits and cannot be considered as pure villous or extravillous cytotrophoblast. The widely used JEG-3 cell is a clonally derived tumor cell line isolated from a choriocarcinoma more than 30 years ago (Kohler and Bridson, 1971). JEG-3 expresses HLA-E, HLA-C and both membrane-bound and soluble HLA-G1 (Barel et al., 2003; Le Bouteiller, unpublished data) and thus clearly not exhibit villous trophoblast phenotype. Moreover, JEG-3 differs in many ways from in situ trophoblast, including polyploidy and several other chromosomal abnormalities. Compared with cultured cytotrophoblast, JEG-3 cells divide and their extent of differentiation is limited. AC1-M59 is an immortalized cell line fusioned with JEG-3. Such a fusion implies that this cell line is also very likely to be polyploid. The fact that AC1-M59 expressed ‘a stronger HLA-G signal on a higher number of cells than JEG-3’, does not make it a better model to study term placental trophoblast.
The choice of control transfectants by the authors may also be a matter of discussion. 221-G5 transfectant is a lymphoblastoid cell line (Ishitani and Geraghty, 1992). When they compared the diversity of peptide ligands bound to HLA-G expressed by the same transfectant and by trophoblast from term placenta, Ishitani et al. (2003) discovered that the relative abundance of individual peptides found in .221-HLA-G5 was far lower than that found from the placental-derived cells and that a smaller number of distinct peptides did bind to the placental material. Although this should not alter western blotting analysis, it cannot be excluded that, when using enzyme-linked immunosorbent assay (ELISA), these differences could modify folding of HLA-G and thus affect the affinity of some of the HLA-G monoclonal antibodies (mAb) recognizing conformational determinants.
|
Real-time PCR |
|---|
|
TopIntroductionTypes of cellsReal-time PCRWestern blottingELISAImmunohistochemistryReferences |
|---|
Lack of details concerning the standard curves used by the authors (did they use genomic DNA or cDNA? did they make a comparison with a ß-actin standard curve?) and the absence of positive and negative control cell lines render the results rather difficult to evaluate. Other reports, using purified villous cytotrophoblast cells, not only did isolate transcripts encoding HLA-G5 by using RT–PCR with intron 4-specific primers, but most importantly, did sequence the corresponding cDNAs, definitely demonstrating they did contain intron 4 (Solier et al., 2002
; Morales et al., 2003). |
Western blotting |
|---|
|
TopIntroductionTypes of cellsReal-time PCRWestern blottingELISAImmunohistochemistryReferences |
|---|
The recombinant HLA-G molecules used as positive control by the authors might not be the most appropriate for this kind of study. Being made in Escherichia coli, this recombinant protein cannot be glycosylated like the soluble HLA-G proteins produced by eukaryote primary trophoblast or trophoblast transfectants. Moreover, the absence of recombinant HLA-G5 control molecule in each western blot experiment does not allow clear comparisons. The reactivity of the MEM-G/1 mAb on 221-HLA-G1 cell-culture supernatant is quite confusing: how a full length membrane-bound HLA-G1 molecule could ever be detected in the cell-culture supernatant? Does this reflect the presence of cells in this supernatant? Could this be due to cell apoptosis, possibly generated by the lack of FCS in the medium? The proofs of the presence of shed HLA-G1 molecules in cell-culture supernatants may also be a subject of discussion. I am not sure that the molecular weight single parameter is enough to assess they are shed forms. It would have been reassuring to demonstrate the appearance of such forms after a specific metalloprotease cleavage. The very low signals obtained in trophoblast culture supernatants make the western blotting data quite difficult to analyse. Immunoprecipitation of culture supernatants with 16G1 mAb followed by MEM-G/1 blotting would have probably help resolving this issue.
|
ELISA |
|---|
|
TopIntroductionTypes of cellsReal-time PCRWestern blottingELISAImmunohistochemistryReferences |
|---|
Once again the absence of total purity of the first trimester trophoblast population used by the authors renders the data difficult to evaluate. The time of culture of trophoblast is also an important issue as soluble HLA-G secretion levels measured by ELISA may increase over the time of culture (Solier et al., 2002
). In this article, the authors only used 24 h of culture which thus gives more limited information. The use of standard curves (for each single experiment performed) with calibrated conformational recombinant intron 4-retaining soluble HLA-G5 proteins associated with human ß2-microglobulin (for each of the HLA-G mAb combination) would have certainly improve ELISA quantification. The use of AC1-M59 cell line for term placenta will never replace purified villous trophoblast from term placenta and makes the data somewhat difficult to compare with previous results (Solier et al., 2002; Ishitani et al., 2003). |
Immunohistochemistry |
|---|
|
TopIntroductionTypes of cellsReal-time PCRWestern blottingELISAImmunohistochemistryReferences |
|---|
The authors showed a strong positive staining of acetone-fixed, cytocentrifuged HLA-G5 transfectants incubated with the 16G1 or 5A6G7 mAbs, both directed against intron-4 peptide, and concluded that the absence of staining of villous cytotrophoblast and syncytiotrophoblast by these mAbs on similarly fixed frozen placental tissue sections, meant that HLA-G5 was not expressed by villous trophoblast in situ. 221-HLA-G5 cells have been transfected with retroviral vectors expressing HLA-G5 cDNA and therefore overexpress the HLA-G5 product, more than 10-fold higher than naturally expressing JEG-3 (Ishitani and Geraghty, 1992
), which obviously does not reflect the physiological conditions occurring in situ. HLA-G5 transfectants expressing a single copy of the transfected gene and treated as described in Blaschitz et al. (2005) article would have probably been negatively stained by these antibodies, like the villous trophoblast. Results obtained by the authors on paraffin embedded, paraformaldehyde-fixed placental tissue sections with the same mAbs were quite different, staining being observed in uterine glands and some decidual cells. However, in these conditions, the 16G1 mAb strongly stained the syncytiotrophoblast. When using the same 16G1 mAb on paraformaldehyde-fixed frozen tissue sections, Ishitani et al. (2003) did found that it strongly stained villous cytotrophoblast and syncytiotrophoblast from first trimester pregnancy. These authors further provided a sound proof of the specificity of their immunohistochemical staining: reactivity of the 16G1 mAb generated against the C-terminal intron 4 peptide of HLA-G5 was specifically blocked by addition of the 20-mer peptide used to generate this mAb immediately before addition of 16G1 to the placental sections (Ishitani et al., 2003). Increasing concentrations of this peptide correlated inversely with the 16G1 signal, whereas control peptides had no effect, ruling out possible non-specific artefactual labelling. By using frozen, unfixed placental tissue sections but a different specific mAb directed against a recombinant HLA-G5 protein associated with human ß2-microglobulin, Morales et al. (2003) clearly demonstrated that, in addition to extravillous trophoblast, it stained both syncytiotrophoblast and villous cytotrophoblast of the chorionic villi from first, second and last trimester of pregnancy. They noticed that their mAb did not detect antigens in paraformaldehyde-fixed paraffin-embedded tissues. This mAb also labelled isolated cytospun purified villous and extravillous cytotrophoblast (Morales et al., 2003). The validity of these different findings, confirmed by three different groups (Solier et al., 2002; Ishitani et al., 2003; Morales et al., 2003) but challenged by those of Blaschitz et al. (2005), is further strengthened by the similar observations made on non-human primate placentas (Slukvin et al., 1998; Langat et al., 2002; Ryan et al., 2002). Immunohistochemical studies performed on paraformaldehyde-fixed frozen sections of rhesus monkey placenta with the 16G1 mAb directed against the intron 4-retaining human soluble HLA-G5 also showed a specific staining in the syncytiotrophoblast and villous trophoblast lining the chorionic villi (Ryan et al., 2002), confirming previous observations (Slukvin et al., 1998). This corresponded to the intron 4-retaining soluble Mamu-AG5 molecule. Similarly, an antibody specific for the intron 4 of Paan-AG Baboon soluble molecule, and which also recognizes HLA-G5, did strongly stain villous syncytiotrophoblast and, less intensively, villous cytotrophoblast in paraformaldehyde-fixed paraffin-embedded first trimester placental section of this non-human primate (Langat et al., 2002). A comparative analysis of intron 4 sequences of the three Paan-AG-, Mamu-AG5- and HLA-G5- soluble molecules showed that they were very similar (Langat et al., 2002).
In conclusion, results presented by Blaschitz et al. (2005) definitely contribute to the discussion as to whether human villous trophoblasts do produce intron 4-retaining soluble HLA-G5 molecules. However, in view of the several possible methodological biases I focused on, their conclusions, summarized in their self-assured manuscript title, perhaps might be considered as premature... I feel that the so-called ‘substantial evidence’ these authors claim they have provided might need quite a few additional reliable investigations performed on really purified, exclusively villous trophoblast from human placental tissue combined with all of the appropriate controls. Obviously the discussion is not over!
|
References |
|---|
|
TopIntroductionTypes of cellsReal-time PCRWestern blottingELISAImmunohistochemistryReferences |
|---|
Astrid Blaschitz, Herbert Juch, Armin Volz, Heinz Hutter, Christine Daxboeck, Gernot Desoye, Gottfried Dohr (2005) The soluble pool of HLA-G produced by human trophoblast does not contain the intron 4-derived peptide. Mol Hum Reprod 00,00–00.
Barel M, Ressing M, Pizzato N, van Leeuwen D, Le Bouteiller P, Lenfant F and Wiertz E (2003) Human cytomegalovirus-encoded US2 differentially affects surface expression of MHC class I locus products and targets membrane-bound, but not soluble HLA-G1 for degradation. J Immunol 171,6757–6765.
Chu W, Fant ME, Geraghty DE and Hunt JS (1998) Soluble HLA-G in human placentas: synthesis in trophoblasts and interferon-gamma-activated macrophages but not placental fibroblasts. Hum Immunol 59,435–442.[CrossRef][ISI][Medline]
Ishitani A and Geraghty DE (1992) Alternative splicing of HLA-G transcripts yields proteins with primary structures resembling both class I and class II antigens. Proc Natl Acad Sci USA 89,3947–3951.
Ishitani A, Sageshima N, Lee N, Dorofeeva N, Hatake K, Marquardt H and Geraghty D (2003) Protein expression and peptide binding suggest unique and interacting functional roles for HLA-E, F, and G in maternal-placental immune recognition. J Immunol 171,1376–1384.
Kohler P and Bridson W (1971) Isolation of hormone-producing clonal lines of human choriocarcinoma. J Clin Endocrinol Metab 32,683–687.[ISI][Medline]
Langat DK, Morales PJ, Fazleabas AT, Mwenda JM and Hunt JS (2002) Baboon placentas express soluble and membrane-bound Paan-AG proteins encoded by alternatively spliced transcripts of the class Ib major histocompatibility complex gene, Paan-AG. Immunogenetics 54,164–173.[CrossRef][ISI][Medline]
Morales P, Pace J, Platt J, Phillips T, Morgan K, Fazleabas A and Hunt J (2003) Placental cell expression of HLA-G2 isoforms is limited to the invasive trophoblast phenotype. J Immunol 171,6215–6224.
Ryan AF, Grendell RL, Geraghty DE and Golos TG (2002) A soluble isoform of the rhesus monkey nonclassical MHC class I molecule Mamu-AG is expressed in the placenta and the testis. J Immunol 169,673–683.
Slukvin II, Boyson JE, Watkins DI and Golos TG (1998) The rhesus monkey analogue of human lymphocyte antigen-G is expressed primarily in villous syncytiotrophoblasts. Biol Reprod 58,728–738.
Solier C, Aguerre-Girr M, Lenfant F, Campan A, Berrebi A, Rebmann V, Grosse-Wilde H and Le Bouteiller P (2002) Secretion of pro-apoptotic intron 4-retaining soluble HLA-G1 by human villous trophoblast. Eur J Immunol 32,3576–3586[CrossRef][ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. van der Meer, H.G.M. Lukassen, B. van Cranenbroek, E.H. Weiss, D.D.M. Braat, M.J. van Lierop, and I. Joosten Soluble HLA-G promotes Th1-type cytokine production by cytokine-activated uterine and peripheral natural killer cells Mol. Hum. Reprod., February 1, 2007; 13(2): 123 - 133. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||