Detection of HLA-G by a specific sandwich ELISA using monoclonal antibodies G233 and 56B

M.J.C.van Lierop 2 F.Wijnands 2 Y.W.Loke 3 P.M.Emmer 1 H.G.M.Lukassen 0 D.D.M.Braat 0 A.van der Meer 1 S.Mosselman 2 I.Joosten 1 0 Gynaecology and Obstetrics, University Medical Centre Nijmegen , Nijmegen , The Netherlands 1 Bloodtransfusion and Transplantation Immunology 2 Department of Pharmacology, NV Organon , 5342 CC Oss , Departments of 3 Research Group in Human Reproductive Immunobiology, Department of Pathology, University of Cambridge , Cambridge , UK Human leukocyte antigen (HLA)-G, which is mainly expressed at the maternal-fetal interface, may play a role in the immune tolerance of the semi-allogenic fetus by the mother. Functional studies have shown that HLA-G is indeed a potential modulator of different immune responses. Therefore, it is of interest to study the level of expression of soluble HLA-G in several biological fluids derived from women with and without fertility problems. In order to measure soluble HLA-G, a reliable and sensitive HLA-G specific sandwich ELISA is required. Here, we describe such an ELISA in which G233 is used as the coating antibody and 56B as the detecting antibody. In comparison with two other assays, this assay shows highest responses to recombinant HLA-G and native HLA-G in primary trophoblast culture supernatant and high responses to HLA-G in amniotic fluid. No HLA-G in follicular fluid or preimplantation embryo culture supernatant could be detected. - The non-classical major histocompatibility complex (MHC) class I molecule, human leukocyte antigen (HLA)-G, has gained a lot of interest in reproductive immunology because of its predominant expression in trophoblast cells. HLA-G differs from other MHC class I molecules by its low polymorphism, truncated cytoplasmic tail and the existence of seven splice variants (Geraghty et al., 1987; Ishitani and Geraghty, 1992; Fujii et al., 1994; Kirszenbaum et al., 1994; Paul et al., 2000a). In particular, the full length HLA-G isoform (HLA-G1) has clearly been shown to be expressed at the cell surface or expressed and secreted in a soluble form (HLA-G5) (Bainbridge et al., 2000a; Mallet et al., 2000; Paul et al., 2000a). Cell surface expression of the other membrane-bound isoforms (HLA-G2, -G3 and -G4) has recently also been shown on transfected cells (Riteau et al., 2001). However, whether these shorter isoforms of HLA-G are really expressed under physiological conditions is still a subject of debate (Bainbridge et al., 2000a; Mallet et al., 2000). The restricted expression of HLA-G at the maternalfetal interface suggests that it may play a major role in the immune tolerance of the semi-allogenic fetus by the mother. Functional studies have shown that HLA-G is a potential modulator of different immune responses. It has the capacity to function as a restriction element for mouse CD8 cytotoxic T cells (Horuzsko et al., 1997). HLA-G has also been shown to form a ligand for several killer inhibitory receptors present on natural killer cells, myelomonocytic cells and/or T cell subsets (Cantoni et al., 1998; Navarro et al., 1999; Pazmany et al., 1999). Through these receptors HLA-G might be able to modulate cytotoxicity, cytokine production and proliferation, as has been shown in vitro (Maejima et al., 1997; Riteau et al., 1999; Bainbridge et al., 2000b; Kapasi et al., 2000). Furthermore, in-vitro studies with soluble HLA-G have shown induction of apoptosis of blast-like cells (Fournel et al., 2000a), an effect on the release of cytokines from peripheral blood mononuclear cells (Kanai et al., 2001) and inhibition of proliferative allo-responses (Lila et al., 2001). If HLA-G is indeed essential for successful embryo implantation and early pregnancy, it is of interest to study the level of expression of soluble HLA-G in several biological fluids derived from women with and without fertility problems. In order to measure soluble HLA-G, different groups have reported the use of ELISA techniques using different monoclonal antibodies. However, for most of these antibodies cross-reactivity with other HLA molecules can not completely be excluded. Alternatively, when no HLA-G specific antibodies were available, indirect methods, including immunodepletions, have been used (Athanassakis et al., 1999; Fournel et al., 1999; Puppo et al., 1999; Rebmann et al., 1999). A study has been described in which different sandwich ELISA methods to detect soluble HLA-G were compared (Fournel et al., 2000b). However, the well characterized, highly HLA-G specific antibody G233 (Loke et al., 1997) was not used in this study. Here, we describe a reliable and sensitive HLA-G specific sandwich ELISA in which G233 is used as the coating antibody and 56B as the detecting antibody. In comparison with two other assays, this assay shows highest responses to recombinant HLA-G and native HLA-G in primary trophoblast culture supernatant and high responses to HLA-G in amniotic fluid. No HLA-G in follicular fluid or preimplantation embryo culture supernatant could be detected. A different assay, in which BFL.1 is used as the detecting antibody, shows a response in amniotic fluid only. Furthermore, when testing a large number of amniotic fluids by this assay and by the combination of G233 and 56B-biotin, the outcomes do not correspond. Materials and methods G233 (IgG2a, ) is an HLA-G specific antibody generated by immunization of HLA-A2.1/human 2m double transgenic mice with murine L-cells transfected with both human 2m and HLA-G (Loke et al., 1997). W6/32 (IgG2a) recognizes 2m-associated HLA class I (Barnstable et al., 1978). BFL.1 Figure 5. HLA-G specific sandwich ELISA methods tested on amniotic fluid. (A) Responses to one pool of amniotic fluid using different capturing monoclonal antibodies (2 g/ml W6/32, 2 g/ml G233, 2 g/ml 56B or no antibody) in combination with BFL.1-biotin (1.25 g/ml) or a biotinylated isotype control (1.25 g/ml) as the detecting antibody. (B) Responses to one pool of amniotic fluid using different capturing monoclonal antibodies: W6/32 (2 g/ml), G233 (2 g/ml), 56B (2 g/ml) or no antibody in combination with 56B-biotin (0.5 g/ml), W6/32-biotin (10 g/ml) or G233-biotin (3 g/ml) as the detecting antibody. All responses have been corrected for background (assay buffer control). Results shown are representative for all assays performed (at least twice). (IgG2a; Coulter Immunotech) is an HLA-G specific antibody (Bensussan et al., 1995). Antibody 56B was generated and characterized as described below. Antibody 56B was generated by immunization of Balb/c mice with a peptide corresponding to amino acids 138158 of the 2 domain of HLA-G coupled to keyhole limpet haemocyanin. The 2 domain of HLA-G (TAAQISKRKCEAANVAEQRRA) represents a region of very low homology to other HLA class I molecules and within the structure of HLA-G it is well accessible for antibody binding. Mice were immunized with 100 g of the peptide in Complete Freunds Adjuvant s.c.; booster-immunizations were given 3 and 6 weeks la (...truncated)