Soluble HLA-A,-B,-C and -G molecules induce apoptosis in T and NK CD8+ cells and inhibit cytotoxic T cell activity through CD8 ligation

There is convincing evidence that soluble HLA-A,-B,-C (sHLA-A,-B,-C) and soluble HLA-G (sHLA-G) antigens can induce apoptosis in CD8(+) activated T cells although there is scanty and conflicting information about the mechanism(s) by which sHLA-A,-B,-C antigens and sHLA-G antigens induce apoptosis. In this study we have compared the apoptosis-inducing ability of sHLA-A,-B,-C antigens with that of sHLA-G1 antigens in CD8(+) T lymphocytes and CD8(+) NK cells. Furthermore we have compared the inhibitory effect of sHLA-A,-B,-C antigens and of sHLA-G1 antigens on the activity of EBV-specific CD8(+) cytotoxic T lymphocytes (CTL). sHLA molecules were purified from serum and from the supernatant of HLA class I-negative cells transfected with one gene encoding either classical or non-classical HLA class I antigens. Both classical and non-classical sHLA class I molecules trigger apoptosis in CD8(+) T lymphocytes and in CD8(+) NK cells, which lack the T cell receptor, and their apoptotic potency is comparable. The binding of sHLA-A,-B,-C and sHLA-G1 molecules to CD8 leads to Fas ligand (FasL) up-regulation, soluble FasL (sFasL) secretion and CD8(+) cell apoptosis by Fas/sFasL interaction. Moreover, classical and non-classical sHLA class I molecules inhibit the cytotoxic activity of EBV-specific CD8(+) CTL. As the amount ofsHLA-G molecules detectable in normal serum is significantly lower than that of sHLA-A,-B,-C molecules, the immunomodulatory effects of sHLA class I molecules purified from serum are likely to be mainly attributable to classical HLA class I antigens. As far as the potential in vivo relevance of these findings is concerned, we suggest that classical sHLA class I molecules may play a major immunoregulatory role in clinical situations characterized by activation of the immune system and elevated sHLA-A,-B,-C serum levels. In contrast, non-classical HLA class I molecules may exert immunomodulatory effects in particular conditions characterized by elevated sHLA-G levels such as pregnancy and some neoplastic diseases.     

Direct Binding of Purified HLA Class I Antigens by Soluble NKG2/CD94 C-Type Lectins from Natural Killer Cells

Human natural killer (NK) cell cytotoxicity is inhibited following human leucocyte antigen (HLA) class I binding by killer cell inhibitory receptors belonging to the immunoglobulin or C-type lectin protein families. Of the latter family, CD94 and NKG2A or -B associate to inhibit NK cell cytotoxicity. We have constructed C-Myc epitope-tagged soluble NKG2A, -B, -C, -D and human NKR-P1 lectin domains, and studied their ability to associate with Flu-tagged soluble CD94 lectin domains. Furthermore, their ability to bind solubilized immunoaffinity-purified HLA class I antigens, either alone or following association with CD94 lectin domains, was evaluated using flow cytometry and Western blot analyses. We show that soluble NKG2A, -B and -C lectin domains interact with CD94 lectin domains to form complexes, whereas NKG2D and human NKR-P1 lectin domains do not. Soluble NKG2C, -D and CD94 lectin domains bind solubilized affinity purified HLA class I antigens on their own, whereas NKG2A and -B require association with CD94 lectin domains for binding. Soluble human NKR-P1 lectin domains do not bind solubilized HLA class I antigens in our system.     

Soluble HLA class I antigen secretion by normal lymphocytes: relationship with cell activation and effect of interferon-gamma

HLA class I antigens are thought to be integral membrane proteins. However, soluble forms of these molecules have been detected. Our laboratory has recently shown that the predominant form of these soluble proteins present in human serum, spleen tissue and culture supernatant of activated lymphocytes exhibits molecular weight and structure similar to classical HLA class I antigens, but lacks HLA A or B polymorphic determinants. In the present study, the secretion of such soluble proteins by lymphocytes has been further explored. Phytohaemagglutinin-stimulated normal lymphocytes secrete considerable quantities of soluble HLA (sHLA) class I proteins. This secretion seems to be a general property of lymphocytes, since activation of T as well as B cells by appropriate mitogens equally induce sHLA I secretion. Lymphocytes require RNA and protein synthesis, but not DNA synthesis, for the secretion to occur. Kinetic studies reveal that maximal sHLA I secretion precedes the peak of DNA synthesis by 24 h. In vitro stimulation with antigens or alloantigens also provokes sHLA I secretion. Moreover, this phenomenon has also been detected for in vivo-activated lymphocytes, as enhanced spontaneous sHLA I secretion was observed in cultures of low-density blastic B and T cells, and of blood lymphocytes obtained from normal subjects who had received a booster immunization 5 days earlier. Interferon-gamma (IFN-gamma) increases the expression of membrane-bound class I antigens but does not induce any sHLA I secretion, suggesting that both molecules are under different regulatory mechanisms. Our results indicate that human lymphocytes, upon stimulation, actively secrete considerable amounts of a soluble form of these biologically relevant proteins.     

Detection and characterization of HLA class I molecules in the supernatant of an hepatocarcinoma cell line and of EBV-transformed B cell lines.

Human leukocyte antigens (HLA) class I molecules can be detected in "soluble" form in the supernatant of cultured cell lines and in serum and plasma of humans. These "soluble" HLA class I molecules are assumed to play a role in liver transplantation. In order to define the nature and composition of HLA class I molecules found in solution, we studied the HLA class I production of an hepatoma carcinoma cell line (HepG2) and of EBV-transformed B-cell lines. Based on molecular weight (MW) analysis, it was demonstrated that different forms of HLA class I molecules were produced by HepG2 cells and EBV B-cells. Monoclonal antibodies (mAbs) specific for HLA class I alleles were able to recognize the mature 45 kDa form, but failed to interact with the 42 kDa and 39 kDa MW forms of HLA class I. Of these different MW forms of HLA class I molecules the mature 45 kDa product was found predominantly to be associated with subcellular vesicles whereas the alternative MW forms of 42 kDa and 39 kDa exist as truly free entities in supernatants.     

HLA polymorphisms in Nigerians

The HLA class I and class II phenotypes of a panel of 114 unrelated Nigerians have been determined. The panel was tested for all the known class I antigens and comparisons of the HLA-A and -B frequencies with those of other African Negroid populations revealed some differences. Only limited comparisons could be made for the HLA-DR and -D frequencies as these are not available for any well-defined African Negroid population. The data concerning the class II antigens of this panel are the most interesting. Half of the DRw11-positive panel members are DQw3 negative and DQw1 positive. In addition, there is dissociation of some HLA-D and -DR specificities, a number of panel members are positive for an HLA-D specificity and are negative for the corresponding HLA-DR specificity. Our results show the value of population studies in the investigation of the relationship between the different HLA class II antigens.     

The HLA-DR phenotype of the responder is predictive of humoral response against HLA class I antigens

Recent studies suggest that the immunogenicity of an human leukocyte antigen (HLA) incompatibility should be considered in the context of the HLA phenotype of the recipient. The HLA-DR phenotype of the responder is thought to be predictive for the strength of the alloimmune response. In order to analyze the humoral response against HLA class I antigens in the context of the HLA-DR phenotype of the responder, we selected all HLA-DR homozygous Dutch patients that were present on the Eurotransplant waiting list between 1967 and 2000 (n=1,317 patients). By logistic regression it was determined whether antibody production against a specific HLA class I antigen is associated with a particular HLA-DR antigen in the patient. Furthermore, it was analyzed whether a patient, expressing a particular HLA-DR antigen, preferentially produces antibodies against particular HLA class I antigens. The results demonstrate that patients, homozygous for a certain HLA-DR antigen, cannot be considered high or low responders when analyzing the antibody response in terms of panel reactive antibody (PRA) value. However, a correlation can be found between the HLA-DR phenotype of the patient and the specific antibody response against HLA class I antigens. For example, antibodies against HLA-A10, -A11, -A19, and -B35 are produced more frequently by HLA-DR6 positive individuals, whereas antibodies against HLA-A3, -B5, -B7, -B8, and -B12 are produced more frequently by HLA-DR4 positive individuals. These data confirm that the HLA-DR phenotype of the responder plays a determinative role in the immunogenicity of mismatched HLA antigens. The results indicate that selection of HLA class I mismatches of the donor in the context of the HLA-DR phenotype of the responder might reduce the incidence of humoral graft rejection and minimize the sensitization grade of retransplant candidates.     

Increased serum concentration of soluble HLA-DR antigens in HIV infection and following transplantation

Abstract: HLA class I and class II antigens circulate in serum as soluble molecules. Increased concentrations of soluble HLA class I molecules have been demonstrated in viral diseases, in rejection episodes following organ transplantation and in graft versus host disease. To explore the possibility of a variation of the serum concentrations of soluble HLA class II molecules in the same pathologic conditions we developed a double determinant immune assay that detects whole soluble HLA-DR molecules (sHLA-DR). The mean level of sHLA-DR antigens in sera from 23 healthy individuals was 0.64±0.72 μg/ml. Elevated serum concentrations of sHLA-DR molecules were detected in sera from HIV infected patients in CDC2/3 and in CDC4C1 stages (2.0±1.7 g/ml and 4.6±1.7 g/ml, respectively), in sera from patients affected by acute rejection after liver transplantation (5.3±3.7 μg/ml) and in sera from patients affected by severe acute graft versus host disease following bone marrow transplantation (8.8±3.1 μg/ml). The increase of sHLA-DR molecules in these sera significantly correlated with the elevation of soluble HLA class I antigens (P=0.0004). The reported data suggest that both soluble HLA class I and class II molecules serum levels increase during viral infections and strong immune reactions and could suggest the involvement of these molecules in immunoregulation.     

Human mumps virus-specific cytotoxic T lymphocytes: quantitative analysis of HLA restriction

To obtain quantitative information about the use of HLA antigens as restriction element by antiviral cytotoxic T lymphocytes (CTL), we have analyzed precursors of human mumps virus-specific CTL by limiting dilution. CTL generated by restimulation of peripheral blood T lymphocytes with autologous mumps virus (MV)-infected stimulator cells were restricted by autologous HLA class I antigens, and derived from the T4-8+ population. They were specific for MV and did not lyse autologous target cells infected with other viruses. Frequencies of MV-specific CTL precursors ranged from 1/500 to 1/8000. HLA restriction was analyzed by split-well analysis of individual CTL colonies. CTL recognizing HLA-A or B antigens were unequally distributed: HLA-B7, -B13, and -B27 were found to function as predominant, in some cases as exclusive, restriction elements, whereas other antigens such as HLA-A24 were never or rarely used. In several combinations, there was no evidence for antigenic variants of HLA molecules as reason for the failure to be recognized. The proportion of CTL precursors recognizing HLA-A2 and -B8 seemed to be dependent on the presence or absence of "dominant" restriction elements. We conclude that CTL precursors recognizing certain virus-HLA combinations are preferentially expanded during an infection, but that low responsiveness to a given combination is not necessarily absolute.     

Xenogeneic recognition of soluble and cell surface HLA class II antigens by proliferative murine T cells

In order to characterize the murine anti-human xenogeneic mixed lymphocyte reactions (MLR), we studied T cell proliferative responses against various human lymphoid cells by immunization of mice either with cellular or purified HLA-DR antigens. Data presented here indicated that small amounts of soluble HLA-DR antigen were able to prime mice, and that the xenogeneic MLR depends on the expression of HLA class II antigens on the stimulating cells. Experiments using a mutant cell line clearly showed that HLA-DP molecules were also sufficient in eliciting a primary or a secondary xenogeneic MLR while no secondary proliferative response was obtained with cells expressing only HLA class I molecules. Using a large panel of human cells with various haplotypes, our results also showed that (a) nonpolymorphic determinants of HLA class II antigens trigger dominantly the murine T cells and (b) the xenogeneic response required I-E and L3T4 accessory molecules and was not inhibited with anti I-A and monomorphic anti-HLA class II antigen monoclonal antibodies. Altogether these results suggest that HLA class II antigens act as nominal antigens in triggering a murine anti-human proliferative response.     

A limited number of HLA epitopes are recognized from HLA class I-specific antibodies detected in the serum of sensitized patients

The goal of this study was to evaluate the epitope specificity of HLA class I-specific antibodies detected in the serum of sensitized patients awaiting retransplantation. The study group consisted of 22 sensitized from previous graft patients, who produced stable IgG HLA class I-specific antibodies. A total of 60 serum samples were screened and analyzed by two techniques in parallel: the antihuman globulin augmented CDC (AHG-CDC) technique and an ELISA technique. All recipients and donors were typed for class I HLA antigens by a standard lymphocytotoxicity technique. The epitope identification was based on class I HLA antigens sequencing, where the multiple immunogenic epitopes are differentially shared among various HLA antigens. The unique epitope configuration on one HLA antigen represents the private epitope of the specific HLA antigen while epitopes shared by more than one HLA antigen represent public determinants. In some HLA antigens (HLA-A1), more than one private epitope has been defined, while in others (HLA-B35, -B51), the private epitopes are not yet known. In a total of 36 antibody reactivity patterns, the majority of the definable IgG HLA class I-specific antibodies corresponded to the A-locus (75%), and only 25% had specificities against the B-locus antigens, although the number of incompatibilities concerning both loci were almost identical (29 for the HLA antigens of the A-locus and 26 for those of B-locus). All patients produced HLA class I-specific antibodies with specificities against the private epitopes of the immunogenic mismatched HLA antigen(s). In 6/21 cases (28.6%), HLA class I alloreactivity spreading to nongraft HLA antigens was detected and 9 public (shared) immunogenic alloepitopes were recognized. In conclusion, appling the epitope analysis of HLA class I-specific antibodies produced by sensitized from previous graft patients, we were able to define the immunogenic alloepitopes. We consider that the immunogenic alloepitopes, during transplantation course, are mainly private epitopes of mismatched HLA antigens and, in certain cases, shared epitopes between the donor alloantigens and other HLA antigens. This knowledge may offer the potential of transplanting sensitized patients through improved donor selection.     

Associated expression of HLA class I and class II antigens on melanoma cells in surgically removed metastases

Malignant transformation of melanocytes and further neoplastic progression may be associated with qualitative and/or quantitative changes in expression of HLA class I and class II antigens. Since previous immunohistochemical studies of surgically removed melanoma lesions have suggested a relationship in the expression of HLA class I and class II antigens, we have investigated the expression of these antigens at the single cell level. Double immunofluorescence staining of frozen sections of melanoma metastases and immunoelectron microscopic double labelling of melanoma cell suspensions prepared from three of these lesions has detected three HLA phenotypes on the large majority of melanoma cells: either both HLA class I and class II antigens, neither HLA antigen or only HLA class I antigens. In four out of the 11 lesions a few melanoma cells were found to express HLA class II antigens and to lack HLA class I antigens. A relationship was also found in the level of expression of HLA class I and class II antigens, as estimated by the intensity of staining with monoclonal antibodies. The level of expression of HLA class II antigens appeared to be similar to or lower than that of HLA class I antigens on the large majority of melanoma cells. This coordinated heterogeneity in the expression of HLA class I and class II antigens by melanoma cells may have implications in the interactions of tumour cells with the host's immune system.     

Calreticulin associates with non-HLA-A,-B class I proteins in the human choriocarcinoma cell lines JEG-3 and BeWo.

Human placental trophoblast expresses as unusual repertoire of major histocompatibility complex (MHC) class I products that appears to reflect the unique role of this epithelium in mediating feto-maternal relations during pregnancy. Trophoblast is devoid of human leucocyte antigen (HLA)-A,-B antigens but can express one or more non-HLA-A,-B class I proteins. The human choriocarcinoma cell lines JEG-3, BeWo and JAR are widely used as models to study trophoblast. During attempts to isolate non-HLA-A,-B class I from JEG-3 and BeWo by immunoaffinity chromatography using a monoclonal antibody to beta 2-microglobulin we observed a 55,000 MW protein co-purifying with class I. N-terminal amino acid sequencing and immunoblotting using a specific antiserum identified this product as calreticulin, a molecule recently shown to be involved in the assembly of classical class I in human B-lymphoblastoid cells. In our hands JEG-3 and BeWo were found to express 45,000 MW non-HLA-A,-B class I proteins while the 40,000 MW HLA-G product was identified only in JEG-3. Our data suggest that calreticulin associates with non-HLA-A,-B class I heterodimers and with free 45,000 MW non-HLA-A,-B class I H chains in JEG-3. JAR was found to be devoid of detectable class I H chains but contained beta 2-microglobulin and calreticulin. However, calreticulin-beta 2-microglobulin complexes were not detected in JAR. Calreticulin and class I were apparently co-localized within the endoplasmic reticulum of JEG-3 cells whereas only class I was expressed at the cell surface. These studies demonstrate that calreticulin is associated with non-HLA-A,-B class I products in human choriocarcinoma cells.     

Class I antigens of the major histocompatibility complex on cytotrophoblast of human chorion laeve

Twenty amniochorions from normal, term pregnancies were studied immunohistologically with the use of well-characterized monoclonal antibodies to beta-2-microglobulin (beta 2M) and to common determinant of HLA-A, -B, -C. In the same study, polyclonal antiserum to trophoblast antigens (TA) were employed with double fluorochrome labels to determine if cytotrophoblast in the chorion laeve expressed TA, HLA or both. The results showed that, in the majority of cases, HLA was not identified on trophoblast, but TA were. However, some cytotrophoblast within the amniochorionic mantle were non-reactive with anti-TA serum and were positive with monoclonal antibodies to HLA and beta 2M antigens. Since these structures were identified as cytotrophoblast by a battery of techniques, in this report they are tentatively designated as metatrophoblast. The role of these cells in the materno-trophoblastic relationship in normal human pregnancy has yet to be determined. This is not envisaged as a simple matter, for metatrophoblast can be recognized by anti-beta 2M and anti-HLA (W6/32), but not by anti-HLA (61D2), suggesting that their expression of class I histocompatibility complex (MHC) antigens may be incomplete, or that there may be a contribution of antigens from an extra-embryonic MHC.     

Establishment of a monoclonal anti-pan HLA class I antibody suitable for immunostaining of formalin-fixed tissue: unusually high frequency of down-regulation in breast cancer tissues

A novel monoclonal anti-pan human leukocyte antigen (HLA) class I heavy chain antibody, EMR8-5, was established. It could detect HLA-A, -B, and -C antigens in formalin-fixed paraffin embedded tissues. By immunohistochemical staining using the EMR8-5 antibody, various cancer tissues from 246 cases were examined for HLA class I expression. It was found that HLA class I expression was decreased in 20% to 42% of the cases of lung cancer, hepatocellular carcinoma, colon cancer, renal cell carcinoma, and urothelial carcinoma. In contrast, 85% of breast cancer cases had loss of or decreased HLA class I expression. Of the 35 breast cancer cases that had decreased HLA class I heavy chain expression, 33 (94%) also had decreased beta2-microglobulin expression detected by immunohistochemical staining. It was suggested that HLA class I down-regulation might be a common characteristic of breast cancer mostly caused by the down-regulation of beta2-microglobulin expression.     

Expression of HLA class I antigens in transporter associated with antigen processing (TAP)-deficient mutant cell lines

Abstract: Cells lacking expression of the transporter associated with antigen processing (TAP) are deficient in surface HLA class I, yet express reduced levels of HLA-A2 antigen through TAP-independent processing pathways. We have analysed the expression of HLA-A, -B and -C antigens on the 721.174 and T2 TAP-deficient mutant cell lines using a panel of monoclonal antibodies specific for the HLA antigens encoded by the genotype of these cells. Our study has shown the constitutive expression of HLA-Cwl molecules on the cell surface of both T2 and 721.174 cells and has confirmed that HLA-A2 and HLA-B51 are expressed at low levels. Transfection of 721.174 cells with cDNAs encoding TAP1 and TAP2 proteins did not fully restore HLA class I antigen expression on these cells, which appeared to be mainly due to a deficiency in expression of the HLA-B51-associ-ated Bw4 epitope. This suggests that additional antigen-processing genes may be required for optimal generation of HLA-B-binding peptides. Our results indicate that TAP-independent pathways of antigen-processing provide peptides for functional expression of all three classical HLA class I molecules.     

HLA-A33 and -B44 and susceptibility to postherpetic neuralgia (PHN)

HLA class I and class II alleles of 32 Japanese patients with postherpetic neuralgia (PHN) and 136 healthy controls were analyzed by serological (class I) and DNA (class II) typing for any significance in the susceptibility to varicella-zoster virus (VZV). We recognized positive associations of the development of PHN with the HLA class I antigens HLA-A33 and -B44, and the HLA-A33-B44 haplotype. This haplotype is tightly linked to DRB1*1302 in a Japanese healthy population. However, no significant association between PHN and HLA class II alleles was observed with no linkage of the HLA haplotype HLA-A33-B44 to HLA-DRB1*1302 in the patients with PHN. These findings suggest that HLA class I gene may genetically control the immune response against VZV in the pathogenesis of PHN.