HLA-G has a concentration-dependent effect on the generation of an allo-CTL response.

Human leucocyte antigen (HLA) -G is expressed on trophoblast cells during pregnancy, suggesting a role in protection of the semiallogeneic fetus. Published data suggest that HLA-G protects a cell against natural killer cell lysis. It has been hypothesized that HLA-G may also protect the fetus by preventing allo-cytotoxic T lymphocyte (CTL) responses. To test this hypothesis, we assayed the effects of various concentrations of purified HLA-G on CTL response in a mixed lymphocyte culture (MLC) system. We found that concentrations > or =0.1 microg/ml of HLA-G suppressed the allo-CTL response by 30-100% over the control, but, paradoxically, concentrations of 0.01-0.05 microg/ml of HLA-G augmented the allo-CTL response by 25-50% over the control. Concentrations < or = 0.001 microg/ml HLA-G had no effect. Addition of HLA-G to preprimed allo-CTL effector cells did not affect their killing ability. Allo-CTL suppressive doses of HLA-G induced a T helper type 2 (Th2) cytokine response, whereas allo-CTL-enhancing doses of HLA-G induced a Th1-type cytokine response. HLA-G purified from first-trimester placenta does not affect allo-proliferative responses nor does it alter the percentage of CD4+ or CD8+ T cells in MLCs. These findings support a potential role for HLA-G-mediated suppression of allo-CTL formation in normal pregnancies. In addition, the effects observed at lower concentrations of HLA-G may have interesting implications for the condition of pre-eclampsia in which concentrations of this HLA class I molecule are reduced.

Variability in the expression of trophectodermal markers beta-human chorionic gonadotrophin, human leukocyte antigen-G and pregnancy specific beta-1 glycoprotein by the human blastocyst.

Improved culture conditions that support the development of human embryos to the blastocyst stage in vitro led to the prospect of blastocyst transfer to increase pregnancy rates. Thus, there is a need for characterization of possible biochemical markers able to predict the implantation potential of human blastocysts. In this study, the expression of three placental markers that are expressed prior to implantation, beta-human chorionic gonadotrophin (HCG), human leukocyte antigen (HLA)-G and pregnancy specific beta-1 glycoprotein (SP-1), was investigated. beta-HCG transcript could be detected as early as the two-cell stage, which is one to two cleavage divisions earlier than previously reported. Both beta-HCG and HLA-G transcripts could be detected in the majority of blastocysts, but their levels were highly variable. No association could be found between the amount of transcript for these genes, total cell number or cell death rate. Interestingly, there was a highly positive correlation between accumulation of beta-HCG and HLA-G transcripts. SP-1 protein concentrations were assessed in the culture medium of blastocysts using enzyme-linked immunosorbent assay. There was a significant positive correlation between SP-1 concentrations and blastocyst cell numbers. Moreover, synthetic oviductal medium enriched with potassium resulted in an SP-1 concentration twice as high as that observed using human tubal fluid medium. These data suggest that SP-1 may be used to select blastocysts with higher cell number, possibly resulting in higher pregnancy rates.

HLA-G isoforms produced by placental cytotrophoblasts and found in amniotic fluid are due to unusual glycosylation.

The human placenta expresses HLA-G, a nonclassical (class Ib) MHC molecule that could play a central role in maternal tolerance of the semiallogeneic fetus. In this work, we report the production of a new mAb, 4H84, that specifically reacts with HLA-G in two formats: immunocytochemistry and immunoblotting. Immunolocalization experiments with 4H84 confirmed our previous finding that cytotrophoblasts within the uterine wall are the only cells in tissue sections of placenta that express the HLA-G protein. Additional experiments showed that both amniocytes and cytotrophoblasts in the amnion-chorion express this protein. Since multiple HLA-G transcripts have been described, we used immunoblotting to study the HLA-G isoforms produced by cytotrophoblasts in vitro and by the amnion-chorion in vivo. Cytotrophoblasts, their conditioned medium, and amniotic fluid samples contained heterodisperse immunoreactive bands (Mr 35,000-50,000). N-deglycosylation by peptide-N-glycosidase F digestion resolved these isoforms into two distinct bands. Cell samples contained primarily an Mr 37,000-42,000 protein, most likely encoded by the full-length mRNA. Conditioned medium and amniotic fluid contained a slightly smaller protein, most likely the secreted form lacking the transmembrane and cytoplasmic regions. Removal of polylactosamine chains by endo-beta D-galactosidase digestion significantly reduced the electrophoretic mobility of the immunoreactive bands, suggesting that HLA-G, unlike class Ib molecules studied to date, carries N-acetyllactosamine units. These data show that Mr heterogeneity of HLA-G is due to its novel glycosylation, rather than to the translation of alternatively spliced mRNAs. We postulate that the unusual carbohydrate structures this molecule carries could interact with maternal immune cells and/or stabilize the molecule.

ME1 epitope of HLA-B27 confers class I-mediated modulation of gram-negative bacterial invasion.

We have previously demonstrated that the presence of the MHC class I molecule, HLA-B27, on the surface of transfected fibroblasts differentially alters Gram-negative bacterial invasion as compared with class I alleles that are not implicated in the seronegative spondyloarthropathies. We have now extended this analysis to show that fibroblasts transfected with HLA-B7, a cross-reactive allele with HLA-B27, also demonstrate a similar altered bacterial invasion phenotype. The decrease in the ability of the bacteria to penetrate the HLA-B27 and HLA-B7 transfectants is an invasion-mediated event, as demonstrated by differential invasion events using Escherichia coli transfected with the inv gene of Yersinia enterocolitica. The lysine at position 70, although unique to the HLA-B27 subtypes, is shown to be not involved in mediating the decrease in invasion. However, the ME1 epitope is the critical factor in determining allele-specific alteration in invasion on the basis of the following: 1) ME1 mAb preincubation reverses the decrease; 2) ME1-binding alleles act like HLA-B27; 3) a class I allele that is intermediate in ME1 binding (HLA-B14) also demonstrates a relative decrease in invasion; and 4) mutation at residue 67 (C-->Y) in HLA-B27, which eliminates the ME1 epitope, normalizes the decreased invasion seen in the native HLA-B27-transfected cells. Thus, the ME1 epitope relates to the disease susceptibility for reactive arthritis that is conferred by both HLA-B27 and cross-reactive group Ags.

HLA-B27/microbial mimicry: an in vivo analysis.

The association between three major spondyloarthritic diseases, ankylosing spondylitis, Reiter's syndrome, and reactive arthritis, and the major histocompatibility complex (MHC) class 1 antigen HLA-B27 is well documented. The hypothesis of cross-reactivity between HLA-B27 and the antecedent infection-causing Gram-negative pathogens such as Salmonella, Shigella and Yersinia has been suggested by in vitro studies employing monoclonal antibodies. We have examined the possibility of such cross-reactivity in vivo using various rabbit immune sera and patient sera as the source of cross-reacting antibody. Mouse L cells were transfected with HLA-A3 or HLA-B27 and used as a source of antigen. Western blot analysis employing denatured antigen, FACS analysis employing native antigen and immunoprecipitation studies were undertaken to detect cross-reacting antibodies generated in vivo to HLA-B27 antigen. Antibodies generated in vivo by infection in patients or immunization in animals against arthritogenic bacteria did not demonstrate any cross-reactivity with HLA-B27 by any of the methods used. As defined by the humoral immune response, molecular mimicry appears unlikely to explain the role of B27 in the pathogenesis of reactive arthritis.

HLA-B27 expression modulates gram-negative bacterial invasion into transfected L cells.

The mechanism by which HLA-B27 confers genetic susceptibility to the seronegative spondyloarthropathies ankylosing spondylitis, Reiter's syndrome, and reactive arthritis, is not well understood. The current concept of an extraarticular bacterial infection functioning as the triggering event in a genetically susceptible host suggests the possibility of direct microbial-MHC interaction. We have addressed the role of HLA-B27 in microbial-host cell interaction by examining invasion by putatively arthritogenic gram-negative bacteria. Target cells used were murine L cells transfected with HLA-B27, HLA-A3, HLA-A2, HLA B44, HLA B18, or pSV2neo vector alone. Relative to the pSV2neo control and the HLA-A3 transfectant, HLA-B27-transfected cells demonstrated a consistent decrease in invasion for each of the following pathogens: Salmonella typhimurium (45 +/- 2% decrease), Shigella sonnei (53 +/- 13% decrease), Shigella flexneri (45 +/- 5% decrease), and enteroinvasive Escherichia coli (57 +/- 8% decrease). This decrease was specific for the HLA B27-transfected L cells and was not observed in the other B allele transfectants. The decreased invasion in the HLA-B27 transfectants is not the result of either altered endogenous mouse class I expression as a result of human class I transfection or increased intracellular bacterial killing within the B27 transfectants. There was an inverse relationship between the amount of surface expression of HLA-B27, as measured by FACS, and the degree of invasion. Blocking of surface B27 Ag with anti-B27 mAb augmented bacterial invasion in the B27 transfectants. These studies demonstrate a novel bacterial-B27 interaction that may have relevance to the pathogenesis of B27-related arthritis.

Cytomegalovirus infection of peripheral blood mononuclear cells: effects on interleukin-1 and -2 production and responsiveness.

Cytomegalovirus suppresses the proliferative response of peripheral blood mononuclear cells to phytohemagglutinin. In these experiments, we identified which mononuclear cell subpopulation might be responsible for the suppression. We found that prior infection of either lymphocytes or monocytes followed by reconstitution with monocytes or lymphocytes, respectively, would abrogate the proliferative response in a subsequent culture with phytohemagglutinin. Infection of either cell type also reduced both the production of interleukin-1 (IL-1) and IL-2 and the proliferative response to exogenously supplied IL-1 or IL-2. We did not find evidence for an IL-2 antagonist. These experiments suggest that cytomegalovirus causes a metabolic derangement in lymphocytes and monocytes and impairs their ability both to produce and to respond to physiological mediators of the immune response.