Antigen presentation by class I major histocompatibility complex molecules: a context for thinking about HLA-G

HLA-G and the MHC Cusp Theory

Human leukocyte antigens (HLA) are significant genetic risk factors in a long list of diseases. However, the mechanisms underlying these associations remain elusive in many cases. The best-characterized function of classical major histocompatibility complex (MHC) antigens is to allow safe presentation of antigenic peptides via a self/non-self-discrimination process. Therefore, most hypotheses to date have posited that the observed associations between certain HLA molecules and human diseases involve antigen presentation (AP). However, these hypotheses often represent inconsistencies with current knowledge. To offer answers to the inconsistencies, a decade ago we have invoked the MHC Cusp theory, postulating that in addition to its main role in AP, the MHC codes for allele-specific molecules that act as ligands in a conformationally-conserved cusp-like fold, which upon interaction with cognate receptors can trigger MHC-associated diseases. In the ensuing years, we have provided empirical evidence that substantiates the theory in several HLA-Class II-associated autoimmune diseases. Notably, in a recent study we have demonstrated that HLA-DRB1 alleles known to protect against several autoimmune diseases encode a protective epitope at the cusp region, which activates anti-inflammatory signaling leading to transcriptional and functional modulatory effects. Relevant to the topic of this session, cusp ligands demonstrate several similarities to the functional effects of HLA-G. The overall goal of this opinion article is to delineate the parallels and distinctive features of the MHC Cusp theory with structural and functional aspects of HLA-G molecules.

Molecular mechanisms of MHC class I-antigen processing: redox considerations

Major histocompatibility complex (MHC) class I molecules present antigenic peptides to the cell surface for screening by CD8(+) T cells. A number of ER-resident chaperones assist the assembly of peptides onto MHC class I molecules, a process that can be divided into several steps. Early folding of the MHC class I heavy chain is followed by its association with beta(2)-microglobulin (beta(2)m). The MHC class I heavy chain-beta(2)m heterodimer is incorporated into the peptide-loading complex, leading to peptide loading, release of the peptide-filled MHC class I molecules from the peptide-loading complex, and exit of the complete MHC class I complex from the ER. Because proper antigen presentation is vital for normal immune responses, the assembly of MHC class I molecules requires tight regulation. Emerging evidence indicates that thiol-based redox regulation plays critical roles in MHC class I-restricted antigen processing and presentation, establishing an unexpected link between redox biology and antigen processing. We review the influences of redox regulation on antigen processing and presentation. Because redox signaling pathways are a rich source of validated drug targets, newly discovered redox biology-mediated mechanisms of antigen processing may facilitate the development of more selective and therapeutic drugs or vaccines against immune diseases.

Immunosurveillance is active in colorectal cancer as downregulation but not complete loss of MHC class I expression correlates with a poor prognosis

Many colorectal tumors lose or downregulate cell surface expression of MHC class I molecules conferring resistance to T-cell-mediated attack. It has been suggested that this phenomenon is due to in vivo immune-tumor interactions. However, evidence of the impact of MHC class I loss on outcomes from colorectal cancer is scarce. In our study of more than 450 colorectal cancers in tissue microarray format, we have shown that both high levels of MHC class I expression and absent MHC class I expression are associated with similar disease-specific survival times, possibly due to natural killer cell-mediated clearance of MHC class I-negative tumor cells. However, tumors with low level expression of MHC class I were found to confer a significantly poorer prognosis, retaining independent significance on multivariate analysis. The existence of these poor prognosis tumors, which may avoid both NK- and T-cell-mediated immune surveillance, has important implications for the design of immunotherapeutic strategies in colorectal cancer.

Immunological relationship between the mother and the fetus

The immunological relationship between the mother and the fetus is a bi-directional communication determined on the one hand by fetal antigen presentation and on the other hand by recognition of and reaction to these antigens by the maternal immune system. There is evidence now that immunological recognition of pregnancy is important for the maintenance of gestation, and that inadequate recognition of fetal antigens might result in failed pregnancy. In contrast to HLA-A and -B Class I genes that are downregulated in human trophoblast cells, nonpolymorphic Class I molecules, e.g., HLA-G Class Ib, are expressed in extravillous cytotrophoblast and also in endothelial cells of fetal vessels in the chorionic villi as well as in amnion cells and amniotic fluid. The trophoblast does not induce transplantation immunity and resists NK- and CTL-mediated lysis in vitro. According to our present knowledge, HLA-G presents antigens for gamma/delta T cells and at the same time defends the trophoblast from cytotoxic effector mechanisms. Since polymorphic MHC is absent from the trophoblast, presentation of fetally derived antigens is unlikely to be MHC restricted. gamma/delta T cells recognize a distinct group of ligands with a smaller receptor repertoire than alpha/beta T cells. Most gamma/delta T cells recognize unprocessed foreign antigens without MHC. In the decidua gamma/delta TCR-positive cells significantly increase in number and the majority of decidual gamma/delta T cells are in an activated form due to recognition of conserved mammalian molecules on the trophoblast. Following recognition of fetally derived antigens, the immune system reacts with the setting in of a wide range of protective mechanisms. Many observations suggest that pregnancy is associated with an altered TH1/TH2 balance. Maternal immune response is biased toward humoral immunity and away from cell-mediated immunity that could be harmful to the fetus. Cytokines of maternal origin act on placental development. On the other hand, antigen expression on the placenta determines maternal cytokine pattern. Normal human pregnancy is characterized by low peripheral NK activity, and increased NK activity seems to play a role in spontaneous abortions of unknown etiology. In early human pregnancy the majority of uterine lymphocytes are CD56(bright) granulated NK cells, which do not express CD16 or CD3. In rodents and humans, uterine NK cells are under hormonal control. In early pregnancy they are enriched at sites where fetal trophoblast infiltrates the decidua. The dynamics of the appearance of uterine NK cells suggest that one of the functions of these cells is control of placentation. Another protective mechanism operating in favor of pregnancy is progesterone-dependent immunomodulation. Due to stimulation by fetally derived antigens, pregnancy lymphocytes develop progesterone receptors and in the presence of progesterone produce a mediator (PIBF) that, through altering the cytokine balance, inhibits NK activity and exerts an antiabortive effect in mice.

Recognition of nonclassical HLA class I antigens by gamma delta T cells during pregnancy

The healthy trophoblast does not express classical HLA-A and HLA-B products; therefore, an MHC-restricted recognition of trophoblast-presented Ags is unlikely. In the decidua and also in peripheral blood of healthy pregnant women, gammadelta T cells significantly increase in number. We investigated the possible role of gammadelta T cells in recognition of trophoblast-presented Ags. PBL and isolated gammadelta T cells from healthy pregnant women as well as from those at risk for premature pregnancy termination were conjugated to choriocarcinoma cells (JAR) transfected with nonclassical HLA Ags (HLA-E, HLA-G). To investigate the involvement of killer-inhibitory/killer-activatory receptors in trophoblast recognition, we tested the effect of CD94 block on cytotoxic activity of Vdelta2(+) enriched gammadelta T cells to HLA-E- and/or HLA-G-transfected targets. Lymphocytes from healthy pregnant women preferentially recognized HLA(-) choriocarcinoma cells, whereas those from pathologically pregnant patients did not discriminate between HLA(+) and HLA(-) cells. Normal pregnancy Vdelta2(+) T cells conjugated at a significantly increased rate to HLA-E transfectants, whereas Vdelta2(+) lymphocytes from pathologically pregnant women did not show a difference between those and HLA(-) cells. Blocking of the CD94 molecule of Vdelta2(+) lymphocytes from healthy pregnant women resulted in an increased cytotoxic activity to HLA-E-transfected target cells. These data indicate that Vdelta2(+) lymphocytes of healthy pregnant women recognize HLA-E on the trophoblast, whereas Vdelta1 cells react with other than HLA Ags. In contrast to Vdelta2(+) lymphocytes from healthy pregnant women, those from women with pathological pregnancies do not recognize HLA-E via their killer-inhibitory receptors and this might account for their high cytotoxic activity.

Immunology of implantation

During implantation, the uterine decidua is invaded by extravillous trophoblast (EVT) cells whose function is to destroy the walls of the uterine spiral arteries in order to provide an adequate blood flow to the fetus. These EVT cells express an unusual combination of HLA class I molecules-HLA-C, HLA-E and HLA-G. The decidua is infiltrated by a population of natural killer (NK) cells with a distinctive phenotype of CD56(bright)CD16(-)CD3(-). These cells are particularly numerous in the decidua basalis at the implantation site where they come into close contact with invading EVT cells. These NK cells express a variety of receptors (CD94/NKG2, KIR and ILT) which are known to recognize HLA class I molecules. We believe that interaction between these NK cells and EVT cells provides the controlling influence for implantation.

T cell activity after dendritic cell vaccination is dependent on both the type of antigen and the mode of delivery

Previous work in both human and animal models has shown that CTL responses can be generated against proteins derived from tumors using either peptide-pulsed dendritic cells (DCs) or nucleic acids from the tumor transfected into autologous DCs. Despite the efficacy of this approach for vaccine therapy, many questions remain regarding whether the route of administration, the frequency of administration, or the type of Ag is critical to generating T cell responses to these Ags. We have investigated methods to enhance CTL responses to a peptide derived from the human proto-oncogene HER-2/neu using mice containing a chimeric HLA A2 and H2Kb allele. Changes in amino acids in the anchor positions of the peptide enhanced the binding of the peptide to HLA-A2 in vitro, but did not enhance the immunogenicity of the peptide in vivo. In contrast, when autologous DCs presented peptides, significant CTL activity was induced with the altered, but not the wild-type, peptide. We found that the route of administration affected the anatomic site and the time to onset of CTL activity, but did not impact on the magnitude of the response. To our surprise, we observed that weekly administration of peptide-pulsed DCs led to diminishing CTL activity after 6 wk of treatment. This was not found in animals injected with DCs every 3 wk for six treatments or in animals initially given DCs weekly and then injected weekly with peptide-pulsed C1R-A2 transfectants.

NK cell recognition of non-classical HLA class I molecules

NK cells recognize several HLA class Ib molecules employing both immunoglobulin-like (Ig-like) and C-type lectin receptors. The CD94/NKG2 and NKG2D lectin-like molecules, respectively, interact with HLA-E and MICA; CD94/NKG2A functions as an inhibitory receptor, while CD94/NKG2C and NKG2D trigger NK cell activity. HLA-E predominantly presents nonamers from the leader sequences of other class I molecules; a peptide derived from HLA-G1 constitutes the highest affinity ligand for both CD94/NKG2 receptors. Members of the Ig-like transcript (ILT) or leucocyte Ig-like receptor (LIR) family (ILT2 or LIR-1 and ILT4 or LIR-2), expressed by other leucocyte lineages, interact with a broad spectrum of HLA class Ia molecules and HLA-G1. Among Ig-like KIRs, the KIR2DL4 (p49) receptor has been shown to specifically recognize HLA-G1; this molecule displays an unusual hybrid structure, sharing features with inhibitory and triggering KIRs.

Lack of MHC class I surface expression on neoplastic cells and poor activation of the secretory pathway of cytotoxic cells in oral squamous cell carcinomas

Cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells use the secretory pathway of perforin/granzymes to kill their target cells. In contrast to NK cells, CTL responses are MHC class I restricted. In this study we analysed the relative activation of CTL and NK cells in relation with MHC class I expression on oral squamous cell carcinomas (OSCCs). MHC class I expression was investigated in 47 OSCCs by immunohistochemistry using HCA2, HC10 and beta2-m antibodies. The presence of CTLs, NK cells, and its activation, was investigated in 21 of these OSCCs using respectively, CD8, CD57 and GrB7 antibodies. The Q-Prodit measuring system was used for quantification of cytotoxic cells. All OSCCs showed weak or absent staining of beta2-m on the cell surface. The absence of beta2-m was significantly associated with absent expression of MHC class I heavy chain as detected by HC10 antibody (P = 0.004). In tumour infiltrates CTLs always outnumbered NK cells, as reflected by the ratio CD57/CD8 being always inferior to one (mean: 0.19; SD: 0.15). The proportion of activated cytotoxic cells as detected by granzyme B expression was generally low (mean: 8.6%; SD 8.9). A clear correlation between MHC class I expression and the relative proportion of NK cells/CTLs was not found. This study shows that the majority of OSCCs show weak or absent expression of MHC class I molecules on the cell surface, possibly due to alterations in the normal beta2-m pathway. The low proportion of granzyme B-positive CTLs/NK cells indicates that the secretory pathway of cytotoxicity is poor in these patients. The lack of correlation between MHC class I expression and CTL/NK cell activation as detected by granzyme-B expression suggests that, next to poor antigen presentation, also local factors seem to determine the final outcome of the cytotoxic immune response.

Murine T cell determination of pregnancy outcome

At the fetomaternal interface, maternal effector cells come in intimate contact with fetal trophoblast cells which express paternal antigens. Failure of fetal trophoblast cells to activate maternal Th1 immune responses has been attributed in part to the absence of classical Class I and Class II major histocompatibilty complex (MHC) antigen expression and elaboration of factors which reduce TcR expression and shift any immune responses which may occur to Th2. Classical TcR alphabeta(+) T cells have not been found to be able to respond to trophoblasts. Recently, TcR gammadelta(+) T cells have been characterized in the low-abortion-rate pregnant C57Bl/10 mouse decidua, and the Vgamma1(+) subset may be able to respond to trophoblasts in a non-MHC-dependent manner. Trophoblast-recognizing T cells with Vgamma1 receptors are also present in the decidua of CBA/J mice pregnant by DBA/2, an abortion-prone mating combination. To test the role of the Vgamma1 subset of decidual gammadelta T cells in abortion-prone pregnancies, we altered this subset by injecting monoclonal anti-Vgamma1.1 antibody on gestation day 5.5, 1 day after implantation. This reduced detectability of a Vgammadelta subset producing TNF-alpha and reduced the abortion rate. Anti-Vgamma2, which reacts with a similar proportion of decidual gammadelta T cells as anti-Vgamma1.1, failed to prevent abortions. Vdelta6.3(+) cells are prominent at the fetomaternal interface, and anti-Vdelta6 antibody injected on day 5.5 prevented abortions. TGF-beta2(+) gammadelta cells first appear on day 8.5 of pregnancy; anti-Vgamma1.1 antibody injection on day 8.5 depleted these cells and boosted abortions; anti-Vdelta6.3 given on day 8.5 boosted abortions to the same level. These results suggest that two populations of Vgamma1.1(+)delta6.3(+) T cells may arise in the decidua: an early population that is Th1, abortogenic, and present during the time of implantation, and a Th2/3 cell subset that is present in the decidua later during pregnancy and which is pregnancy-protective.

Differential surface expression of MICA by endothelial cells, fibroblasts, keratinocytes, and monocytes

MICA is a new, highly divergent and polymorphic HLA-related gene that has a similar intron-exon organization as the HLA class I genes. It functions as a restriction element for intestinal gammadeltaT cells and it behaves as a cell stress molecule. It is likely that the polymorphic MICA molecule may be target for specific antibodies and T cells in solid organ grafts or in graft versus host disease (GVHD). Previously, we generated three MICA-specific sera in rabbits, which were used for Western blot, flow cytometry and immunoprecipitation. We demonstrated that MICA is expressed in endothelial cells, keratinocyes and monocytes, but not in CD4+, CD8+ or CD19+ lymphocytes. We also found that MICA is expressed on the cell surface in HeLa cells. In the present work, performing peptide neutralization assays, we further confirmed the specificity of the reactivity of these sera against MICA. Also, by Western blot we demonstrate that freshly isolated human skin-derived fibroblasts express MICA. We also investigated the surface expression of MICA in different, freshly-isolated cells. The results show that endothelial cells and fibroblasts express MICA at the cell surface. Although expressing the 62 kDa MICA band, as detected by Western blots, keratinocytes and monocytes do not seem to express this antigen on the cell membrane. This differential surface expression of MICA by endothelial cells and fibroblasts vs. keratinocytes and monocytes, may indicate that the levels of surface MICA are differentially regulated in different cells. Moreover, the expression of MICA on the surface of endothelial cells makes this polymorphic molecule a possible target during the immune response of graft rejection in organ transplantation.

The influence of the maternal uterine immune response on placentation in human subjects

The immunological relationship between the mammalian fetus and its mother during pregnancy has been considered similar to that between a transplanted allograft and its recipient ever since Medawar (1953) first proposed the concept of the 'fetus as an allograft' in the early 1950s. Based on this analogy, it has been assumed that implantation of the fetal placenta in the uterus would be controlled similarly by a maternal immune response mediated by T-cells recognizing paternally-derived alloantigens expressed by the placenta. Surprisingly, recent evidence suggests that implantation might involve predominantly a novel allogeneic recognition system based on natural killer cells rather than T-cells (Loke & King, 1995). The cellular and molecular basis of this local immune interaction between the fetal placenta and maternal uterus is now the focus of intense research interest. Since aberrant implantation can cause a variety of clinical problems, including miscarriage, intrauterine growth retardation and pre-eclampsia, an understanding of the immunological mechanism by which this process is controlled could lead to the development of regimens to improve fetal growth and development.

HLA-G polymorphisms: ethnic differences and implications for potential molecule function

PROBLEM

Human leukocyte antigen (HLA)-G is uniquely expressed on extravillous cytotrophoblasts of the placenta and is postulated to be a mediator of maternal immune tolerance. Although it was originally considered to be nonpolymorphic, variations of the HLA-G DNA sequence have been reported, and a limited number of HLA-G alleles been defined.

METHOD OF STUDY

The HLA-G wild-type sequence was compared with HLA-A2 with regard to the conservation of functionally essential parts of classical HLA-I molecules. HLA-G polymorphisms were analyzed under the aspect of ethnic differences, site, and consequences for postulated molecule functions.

RESULTS

HLA-G exhibits a high degree of conservation relative to HLA-A2 in functionally relevant sites of HLA-class I molecules. However, polymorphic sites in HLA-G and classical HLA loci are not congruent.

CONCLUSION

The type and localization of HLA-G polymorphisms suggest that different parts of HLA-G molecule underlie different selective constraints.

HLA-G polymorphisms and allele frequencies in Caucasians

HLA-G, a nonclassical class I MHC molecule, is uniquely expressed on extravillous cytotrophoblasts of the maternal-fetal interface and is suggested to be essential for establishment of maternal-fetal immune tolerance. Although the level of polymorphism in HLA-G has originally been considered low, number, nature and site of polymorphisms seem to vary between different ethnic populations. We investigated HLA-G polymorphisms in a population of German and Croatian origin by SSCP-analysis and direct sequencing as well as RFLP analysis for presence of the 1597delC mutation. HLA-A alleles associated with the different HLA-G alleles were determined by SSP PCR-typing. In Caucasians, HLA-G exhibits a low degree of polymorphism on the amino-acid level and only slightly higher variability on the nucleotide level. In 264 independent chromosomes, 4 HLA-G alleles on the level of amino acid polymorphisms and an additional 6 variations of nucleotide sequences could be identified. The null-allele G*0105N was present at an allele frequency of 2.3%, which is higher than initially suggested for Caucasians but lower than in Hispanics and African-Americans. Furthermore, some HLA-G alleles exhibit strong linkage disequilibrium with HLA-A.

HLA-G polymorphisms: neutral evolution or novel function?

HLA-G is a non-classical class Ib gene with many unusual features. Because of its unique expression pattern, which is primarily limited to fetal cells at the maternal fetal interface, this gene has gained the attention of many investigators. In this paper we review some of the novel features of HLA-G, with particular reference to polymorphic variants in the gene, and discuss the implications of these features for the potential function and evolutionary history of HLA-G.

Immunology of human placental implantation: clinical implications of our current understanding

In the early 1950s, Medawar proposed the concept of "the fetus as an allograft". Since then, the immunological relationship between the mammalian fetus and its mother during pregnancy has been considered to be similar to that between a transplanted allograft and its recipient. Because of this analogy, it has been assumed that implantation of the fetal placenta in the uterus would similarly be controlled by a maternal immune response mediated by T cells recognizing paternally derived alloantigens expressed by the placenta. Surprisingly, recent evidence suggests that implantation might predominantly involve a novel allogeneic recognition system based on natural killer cells rather than T cells. The cellular and molecular basis of this local immune interaction between the fetal placenta and maternal uterus is now the focus of intense research interest. Because aberrant implantation can cause a variety of clinical problems including miscarriage, intrauterine growth retardation and pre-eclampsia, an understanding of the immunological mechanism by which this process is controlled could lead to the development of regimes to treat these important obstetric conditions.

Amnion epithelial cells, in contrast to trophoblast cells, express all classical HLA class I molecules together with HLA-G

PROBLEM

The expression of the non-classical HLA-G gene has been shown at the protein level on trophoblast-derived embryonic tissue, like the extravillous cytotrophoblast. However, the presence of HLA-G on embryoblast-derived cells is currently controversial. The amnion epithelium is an embryoblast-derived cell layer covering the amnion cavity and is the main source for the amnion fluid.

METHOD

The expression of HLA class I molecules was investigated by immunohistochemical, biochemical, and molecular biological methods in amnion membranes and amnion fluid.

RESULTS

Immunohistochemically, HLA-C and occasionally also-B molecules as well as HLA-A and/or -G molecules have been identified on amnion epithelial cells. These results were extended by Western blotting with purified amnion epithelial cells where HLA-B and/or -C, HLA-A and HLA-G antigens have been detected. As expected HLA-G mRNA was detected in amino epithelial cells. Furthermore, classical HLA molecules as well as HLA-G were found in amnion fluid.

CONCLUSION

These results show that the amnion epithelium frequently expresses classical HLA class I molecules as well as HLA-G. The expression of HLA-G antigens on amnion epithelial cells and their presence in the amnion fluid, which is continually ingested by the fetus, may be particularly relevant for the induction of peripheral tolerance.

Population genetic studies of HLA-E: evidence for selection

HLA-E is a nonclassical, class I gene (Ib) of unknown function. The study was initiated to determine the amount and nature of the variation in the class Ib gene HLA-E in diverse ethnic groups. A single base-pair substitution (A-->G at 382, exon 3) resulting in a change from an arginine (R) to a glycine (G) at codon 107 was found. A glycine was present at position 107 in individuals from four ape species, suggesting that EG107 is the older of the two alleles. The two human alleles were present in all samples studied. The alleles were in linkage disequilibrium with HLA-A (W = 0.58), HLA-B (W = 0.59) and HLA-C (W = 0.55) in the Hutterites. The frequencies of the two HLA-E alleles were more equal than expectations based on neutrality in inbred and outbred Caucasian samples (Watterson's F = 0.506, p = 0.02 and F = 0.512, p = 0.047, respectively) and nearly significant in African-American and Hispanic samples (F = 0.513, p = 0.063 and F = 0.508, p = 0.053). These data suggest that this polymorphism arose before the expansion of Homo sapiens and has been maintained in diverse populations by stabilizing selection.

Allelic repertoire of the human MHC class I MICA gene

The hallmark of the classical major histocompatibility complex (MHC) class I molecules is their astonishing level of polymorphism, a characteristic not shared by the nonclassical MHC class I genes. A distinct family of MHC class I genes has been recently identified within the human MHC class I region. The MICA (MHC class I chain-related A) gene in this family is a highly divergent member of the MHC class I family and has a unique pattern of tissue expression. We have sequenced exons encoding the extracellular alpha1, alpha2, and alpha3 domains of the MICA gene from twenty HLA homozygous typing cell lines and four unrelated individuals. We report the identification of eleven new alleles defined by a total of twenty-two amino acid substitutions. Thus, the total number of MICA alleles is sixteen. Interestingly, a tentative superimposition of MICA variable residues on the HLA-A2 structure reveals a unique pattern of distribution, concentrated primarily on the outer edge of the MICA putative antigen binding cleft, apparently bordering an invariant ligand binding site.

Detection of membrane-bound HLA-G translated products with a specific monoclonal antibody

A monomorphic anti-HLA-G monoclonal antibody (mAb) was obtained by immunization of HLA-B27/human beta 2-microglobulin double-transgenic mice with transfected murine L cells expressing both HLA-G and human beta 2-microglobulin. This mAb, designated BFL.1, specifically recognizes, by flow cytometry analysis, the immunizing HLA-G-expressing cells, whereas it does not bind to parental untransfected or to HLA-B7- and HLA-A3-transfected L cells, suggesting that it distinguishes between classical HLA-A and -B and nonclassical HLA-G class I molecules. This was further assessed by the absence of BFL.1 reactivity with a number of human cell lines known to express classical HLA class I proteins. In addition, we showed that the BFL.1 mAb also labels HLA-G-naturally-expressing JEG-3 and HLA-G-transfected JAR human choriocarcinoma cell lines as well as a subpopulation of first-trimester placental cytotrophoblast cells. Further biochemical studies were performed by immunoprecipitation of biotinylated membrane lysates: BFL.1, like the monomorphic W6/32 mAb, immunoprecipitated a 39-kDa protein in HLA-G-expressing cell lines, a size corresponding to the predicted full-length HLA-G1 isoform. However, in contrast to W6/32, which immunoprecipitates both classical and nonclassical HLA class I heavy chains, BFL.1 mAb does not recognize the class Ia products. Such a mAb should be a useful tool for analysis of HLA-G protein expression in various normal and pathological human tissues and for determination of the function(s) of translated HLA-G products.