HLA-G Molecules: from Maternal–Fetal Tolerance to Tissue Acceptance

ADVANCES IN IMMUNOLOGY, VOL. 81

HLA-G Molecules: from Maternal–Fetal Tolerance to Tissue Acceptance ¨ L LE MAOULT,* EDGARDO D. CAROSELLA,*,1 PHILIPPE MOREAU,* JOE MAGALI LE DISCORDE,* JEAN DAUSSET,y AND NATHALIE ROUAS-FREISS*

*Service de Recherches en He´mato-Immunologie, Direction des Sciences du Vivant, De´partement de Recherche Me´dicale, CEA Commissariat a` l’Energie Atomique, Institut Universitaire d’He´matologie, Hoˆpital Saint-Louis, 75010 Paris, France; and yFondation Jean Dausset, CEPH, Paris, France

Over the past few years, HLA-G, the non-classical HLA class I molecule, has been the center of investigations that have led to the description of its specific structural and functional properties. Although located in the HLA class I region of chromosome six, the HLA-G gene may be distinguished from other HLA class I genes by its low polymorphism and alternative splicing that generates seven HLA-G proteins, whose tissue-distribution is restricted to normal fetal and adult tissues that display a tolerogeneic function toward both innate and acquired immune cells. We review these points, with special emphasis on the role of HLA-G in human pathologies, such as cancer, viral infection, and inflammatory diseases, as well as in organ transplantation.

I. Introduction

Transcription of the non-classical human leukocyte antigen (HLA) HLA-G gene (Geraghty et al., 1987) and expression of the HLA-G protein (Kovats et al., 1990; McMaster et al., 1995) was initially described as restricted to the fetal–maternal interface on the extravillous cytotrophoblast. The presence of HLA-G at this immunologically privileged site was first proposed as a mechanism used by the fetal semi-allograft to avoid rejection by the mother’s immune system (Rouas-Freiss et al., 1997a, 1999). In this chapter, we review evidence supporting the ability of HLA-G to suppress immune cell functions, such as natural killer (NK) cell- and cytotoxic T lymphocyte (CTL)-mediated cytolysis, the T cell proliferative response, NK transendothelial migration, and dendritic cell (DC) maturation. HLA-G has been also shown to induce NK and T cell apoptosis, as well as a shift toward the Th2 cytokine profile (Carosella et al., 2000; Contini et al., 2003). These multiple suppressive effects 1 Service de Recherche en He´mato-Immunologie, CEA-DSV-DRM, Hoˆpital Saint Louis, Institut Universitaire d’He´matologie, 1, avenue Claude Vellefaux, 75475 Paris cedex 10, France. Tel.: þ 33(0) 1 53 72 22 27; fax: þ 33 (0) 1 48 03 19 60; E-mail: [email protected]

199 Copyright ß 2003 by Elsevier Inc. All rights of reproduction in any form reserved. 0065-2776/03 $35.00

200

E. D. CAROSELLA ET AL.

FIG. 1. Schematic Representation of alternative splicing of HLA-G transcripts and protein isoforms. Exon 1 (E1) encodes leader peptide, exons 2, 3, and 4 (E2, E3, and E4) encode the 1, 2, and 3 extracellular domains, respectively, exon 5 (E5) encodes the transmembrane region, and exon 6 (E6) encodes the reduced cytoplasmic domain of the HLA-G protein. Translation of HLA-G1, -G2, and -G4 transcripts give rise to membrane-bound forms of HLA-G proteins. Intron 4 is retained in both HLA-G5 and HLA-G6 transcripts, and intron 2 is retained in HLA-G7, thus generating soluble forms of HLA-G proteins. In these introns, open reading frames yield to a 21amino acid-specific tail for both HLA-G5, and HLA-G6 proteins, and a 2-amino acid-specific tail for the HLA-G7 isoform. (See Color Insert.)

in both innate and adaptive immunity are mediated by the ligation of HLA-G to several receptors, including ILT2, ILT4, KIR2DL4/p49, and CD8. The gene structure of HLA-G is highly homologous when compared to that of the other HLA class I genes (Ellis et al., 1986), analysis of its transcription has led to the identification of specific mechanisms of alternative mRNA splicing. The HLA-G primary transcript has been shown to generate seven alternative mRNAs able to encode four membrane-bound (HLA-G1, G2, G3, and G4) and three soluble (HLA-G5, G6, and G7) protein isoforms (Fig. 1) (Ishitani, 1992; Kirszenbaum et al., 1994; Paul et al., 2000a). We also review evidence that over the past few years, the tissue distribution of these HLA-G molecules has been found to be broader than originally thought, particularly in pathological processes. While the HLA-G gene is

HLA-G MOLECULES

201

transcribed in almost all tissues, HLA-G transcripts are translated into proteins in only a few non-pathological tissues. Indeed, HLA-G molecules are detected in oocytes and pre-implantation embryos (Jurisicova et al., 1996b; Menicucci et al., 1999; Fuzzi et al., 2002) and in amniotic cells and fluid (Houlihan et al., 1995; Hammer et al., 1997a; McMaster et al., 1998). In adults, thymus is the only non-pathological tissue in which HLA-G has been detected (Crisa et al., 1997; Mallet et al., 1999a). In contrast, HLA-G mRNA and protein upregulation have been described under pathological conditions in peripheral tissues, such as grafted organs, skin and muscle fiber inflammation, after viral infection, and in malignancies. We discuss ectopic HLA-G expression both with respect to its ‘‘positive’’ role in allowing better acceptance of organ grafts, analogous to its role in protecting the fetal semiallograft from maternal immune recognition, and to its ‘‘negative’’ role, when expressed by malignant or virus-infected cells, whose escape from the host’s immunosurveillance it favors. In inflammatory diseases, HLA-G may constitute a tissue-protective molecule against inflammatory aggression (Carosella et al., 2001). We also discuss the fact that besides the immunosuppressive properties exhibited by HLA-G, another possible function is acting as an antigen peptide-presenting molecule capable of being recognized by antigen-specific T cells.

II. The HLA-G Gene and Polymorphism

The HLA-G gene is located within the class I MHC locus, on the short arm of chromosome 6, approximately 230 kb telomeric from HLA-A and 100 kb centromeric to HLA-F. The overall HLA-G gene structure, previously designed HLA-6.0, was described in 1987 (Geraghty et al., 1987) and found to be highly homologous to that of classical HLA-class I genes. It presents
86% similarity with the consensus sequence of the HLA-A, -B, and -C genes and consists of eight exons and seven introns: Exon 1 encodes the signal peptide, exons 2, 3, and 4 respectively encode the 1, 2 and 3 extracellular domains, and exon 5 encodes the transmembrane domain. Exon 6 has a stop codon in its second codon, which results in a shorter cytoplasmic tail region, compared to classical HLA-class I gene products. Evolutionary trees obtained using complete genomic DNA sequences show that HLA-A, G, H, and J form a group of HLA-A-related loci that are readily distinguishable from HLA-B, C, E, and F (Messer et al., 1992). Recent data have also accumulated that show that the HLA-G gene is not confined to the human species and that the HLA-G gene and HLA-G protein present limited polymorphism, which however, may be relevant in the regulation of HLA-G function.

202

E. D. CAROSELLA ET AL.

A. HLA-G GENE HOMOLOGS

IN

NONHUMAN PRIMATES

No HLA-G gene homolog has been found in mice. Nevertheless, nonhuman primates that are phylogenetically close to humans express homologs of HLA-G. It has been shown that the cotton-tamarin (Saguinus oedipus—Saoe), which diverged from human lineage approximately 40 million years ago, expressed an HLA-G-‘‘like’’ gene (Watkins et al., 1990). More recently, data have shown that both New World and Old World monkeys have MHC-G genes (Corell et al., 1994; Arnaiz-Villena et al., 1997, 1999; Slukvin et al., 2000; Arnaiz-Villena et al., 2001). In monkeys, the degree of MHC-G gene polymorphism varies according to the species, being high in Saoe and low in Pongidae (Arnaiz-Villena et al., 2001). The MHC-G molecule in monkeys may exhibit a truncated cytoplasmic domain (Arnaiz-Villena et al., 2001), and alternative splicing of the primary transcript of MHC-G genes has also been reported (Castro et al., 2000b; Langat et al., 2002). Notably, a transcript retaining intron 4 and encoding a soluble Mamu-AG isoform (Manu-AG5) has been described in the rhesus monkey placenta, thereby confirming a homology between HLA-G in the human and Mamu-AG in the rhesus (Ryan et al., 2002). B. HLA-G GENE POLYMORPHISM Studies have converged on the existence of a low level of HLA-G allelic polymorphism compared with classical HLA-class I genes, and the absence of HLA-G gene imprinting (Kirszenbaum et al., 1997; Hiby et al., 1999a). To date, 15 HLA-G alleles have been assigned in the WHO nomenclature (Table I), including a null allele, known as HLA-G* 0105N. Among the ‘‘normal’’ alleles, the first allele that was sequenced, G* 01011 (recently renamed G* 010101), predominates in almost all populations studied to date (Asian, European, and African). Its frequency varies from 32% in the German/Croatian population (van der Ven et al., 1998) to 83% in the African Ghanaian population (Matte et al., 2000). The G* 01011 (G* 010101) and G* 01012 (G* 010102) alleles have been found with comparable frequencies among the German/Croatian (32% and 36%, respectively) (van der Ven et al., 1998) and Portuguese (37% and 31%, respectively) (Alvarez et al., 1999) populations. Nevertheless, the high frequency of the G* 01012 (G* 010102) allele seems to be confined to Caucasian populations and was found to be lower than that of the G* 01041 (G* 010401) allele, in both the Japanese (38%) and African (9.5–20.4%) populations (Alvarez et al., 1999). Some HLA-G alleles are found frequently only in certain specific populations. For example, G* 01013 (G* 010103) and G* 01018 (G* 010108) are found at frequencies of 17% in the Portuguese and 14.4% in the Shona population of Africa, respectively (Alvarez et al., 1999; Matte et al., 2000). Other alleles, such as

HLA-G ALLELES Alleles

Exon 2

Exon 3

30 UT

Exon 4

31

54

57

69

93

107

110

130

188

241

258

14 bp

ACG(T) ACG ACG ACG ACG ACG ACG ACG ACG TCG(S) ACG ACG ACG ACG ACG

CAG(Q) CAG CAG CAG CAG CAG CAG CAG CGG(R) CAG CAG CAG CAG CAG CAG

CCG(P) CCA CCA CCG CCG CCG CCA CCA CCG CCG CCA CCC CCG CCA CCA

GCC(A) GCC GCC GCT GCC GCC GCC GCC GCC GCC GCC GCC GCC GCC GCC

CAC(H) CAT CAC CAC CAC CAC CAT CAC CAC CAC CAC CAC CAC CAT CAT

GGA(G) GGA GGT GGA GGT GGA GGT GGA GGA GGA GGA GGA GGA GGA GGA

CTC(L) CTC CTC CTC CTC CTC CTC CTC CTC CTC ATC(I) ATC(I) ATC(I) CTC CTC

CTG(L) CTG CTG CTG CTG CTG CTG CTG CTG CTG CTG CTG CTG .TG(f) CTG

CAC(H) CAC CAC n.d. CAC CAT CAC CAC CAC n.d. CAC CAT CAC

TTC(F) TTC TCC(S) n.d. TTC TTC TTC TTC TTC n.d. TTC TTC TTC

ACG(T) ACG ACG n.d. ACG ACG ACG ACG ACG n.d. ACG ACG ACG

 þ þ

CAC

TTC

ATG(M)

 þ 

HLA-G MOLECULES

G* 010101 G* 010102 G* 010103 G* 010104 G* 010105 G* 010106 G* 010107 G* 010108 G* 0102 G* 0103 G* 010401 G* 010402 G* 010403 G* 0105N G* 0106

TABLE I PROTEIN POLYMORPHISM

AND

þ þ

A, Ala; F, Phe; G, Gly; H, His; I, Ile; L, Leu; M, Met; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; f, frameshift.

203

204

E. D. CAROSELLA ET AL.

G* 01015 (G* 010105), G* 01016 (G* 010106), G* 01017 (G* 010107), G* 0102, and G* 01043 (G* 010403) are very rare. Furthermore, the newly described G* 0106 allele (Hviid et al., 2001), which has not yet been extensively studied, was found at a frequency of 4% among Caucasians. Nevertheless, the existence of this allele was not confirmed by the studies of Geraghty’s group presented during the thirteenth Histocompatibility Workshop (Victoria, BC, Canada; May, 2002), and thus remains to be clarified. The G* 0105N null allele (Arnaiz-Villena et al., 1997) has been found at frequencies of 11% in African Shona, 7.4–8.3% among African-Americans, 7.3% in Nigerian and 5.3–6.3% in Cameroon populations, 4.8% among Ghanaians, 3% in Spaniards (in association with the HLA-A30-B13 haplotype), 2.9% among Mexican-Americans, 2.5% in Sardinians, 2.3% in mixed German-Croatian populations, 0.6% in Denmark, and absent in UK/USEuropean, Portuguese, and Japanese populations (Hviid et al., 1997; Suarez et al., 1997; Ober et al., 1998; van der Ven et al., 1998; Ishitani et al., 1999; Matte et al., 2000; Aldrich et al., 2002). This allele may have arisen in Africa and spread to Europe after being introduced in Spain on an HLA-A30 haplotype by Moorish conquerors during the eighth century, AD (Suarez et al., 1997). C. HLA-G PROTEIN POLYMORPHISM Although HLA-G alleles are essentially defined by non-synonymous substitutions in comparison with G* 01011 (G* 010101), single amino acid changes characterize six HLA-G proteins (Table I). They are localized in exon 2 at codon 31 (Thr ! Ser) and codon 54 (Gln ! Arg), in exon 3 at codon 110 (Leu ! Ile), and in exon 4 at codon 241 (Phe ! Ser) and codon 258 (Thr ! Met). Another HLA-G protein variant may be encoded by the G* 0105N allele, designated the ‘‘HLA-G null allele.’’ This allele has the same DNA sequence as G* 01012 (G* 010102), except for a cytosine deletion (1597
C) at the third position of codon 129 or the first position of codon 130, causing a frameshift mutation. Consequently, all amino acids encoded by the second half of exon 3 are different, and a stop codon generated at the beginning of exon 4 gives rise to a truncated HLA-G protein (Ober et al., 1998; Castro et al., 2000c). The reason for the unusually high frequency of the 1597
C mutation in African populations is unknown, but it has been suggested that it is a result of natural selection pressure (Aldrich et al., 2002). One hypothesis is that intrauterine pathogens may be the selective agents in populations from areas with a history of high pathogen load. Aldrich et al. hypothesized that reduced HLA-G1 expression in heterozygous placentas may result in an overall increase in the number of T cells available in the uterus to fight intrauterine infections. On the other hand, the existence of healthy individuals who are

205

HLA-G MOLECULES

homozygous for the G* 0105N allele raised doubts about the role of HLA-G1 in the maintenance of pregnancy (Ober et al., 1998). Although debated, it has been suggested that the HLA-G* 0105N allele can still produce the HLA-G2, G3, G6, and G7 isoforms that could substitute for HLA-G1 in fetal–maternal tolerance (Riteau et al., 2001c). Accordingly, the corresponding transcripts are produced in homozygous G* 0105N individuals, and HLA-G2 and G3 protein isoforms have been shown to inhibit NK function in vitro (Castro et al., 2000c; Riteau et al., 2001c). D. POLYMORPHISM

IN THE

NON-CODING REGION

OF THE

HLA-G GENE

Recent work has focused on HLA-G gene polymorphism in non-coding regions. First, a 14-bp deletion/insertion polymorphism has been localized in the 30 UT region of the HLA-G gene at respective frequencies of 58% for deletion and 42% for insertion (Harrison et al., 1993). HLA-G alleles containing a 14-bp insertion may undergo further alternative splicing, such that 92 bases of the 30 UT region are spliced out of a part of the transcripts (Hiby et al., 1999a). Analysis of this polymorphism in several other primates has shown that 14-bp polymorphism must be very recent, since the deletion is present in some human MHC-G alleles but not in other primate species. This suggests that the ‘‘long’’ MHC-G alleles are the oldest (Castro et al., 2000a). Such polymorphism may have functional consequences. One hypothesis is that the 92-base-pair deletion might influence the stability of the HLA-G transcripts. On the other hand, it is observed that the 14-bp polymorphism, in association with the T mutation at the third base of codon 93 (exon 3), may be associated with lower HLA-G transcript levels and be involved in the variation in alternative splicing profiles (for example, lack of HLA-G3 in pre-eclampsia) (O’Brien et al., 2001). Two studies have reported low polymorphism in the 1.4-kb promoter region of the HLA-G gene. One of the interesting polymorphisms is demonstrated in the Danish samples, located near the B2 site at position –201 from the gene, changing a G in the G* 01011 (G* 010101) allele to an A in the G* 01012 (G* 010102), G* 01013 (G* 010103), G* 0104, and G* 0105N alleles (Hviid et al., 1999). The other one is a nucleotide substitution (C ! T) at position –56 found in indigenous East-African populations and located in the putative binding site for polyomavirus enhancer-binding protein 2 (PEBP2) (Matte et al., 2002). The authors inferred that these polymorphisms might have an effect on the regulation of expression of some HLA-G alleles, a theory that requires further investigation. In conclusion, there is convergent data showing that in humans, the HLA-G gene exhibits little polymorphism, a property that contributes to supporting the HLA-G function during pregnancy. In some cases, the inheritance of specific HLA-G alleles, as well as the appearance of somatic mutations in the

206

E. D. CAROSELLA ET AL.

HLA-G gene, may have an impact on the modulation of HLA-G expression, consequently, on its function. Indeed, Rebmann et al. demonstrated the association of HLA-G alleles with soluble HLA-G levels in the plasma of 94 unrelated healthy individuals (Rebmann et al., 2001). High levels of soluble HLA-G were found to be associated with the presence of the HLA-G* 01041 (G* 010401) allele. Mean soluble HLA-G levels in individuals with the most common HLA-G alleles G* 01011 (G* 010101) and G* 01012 (G* 010102) were very similar and were approximately 63% of the levels found in HLAG* 01041 (G* 010401) individuals. In contrast, significantly reduced soluble HLA-G levels were observed in individual bearing G* 01013 (G* 010103) or G* 0105N alleles. In addition, recent data reveal that HLA-G polymorphism may influence some pathologies of pregnancy. Hviid et al. did not find significant differences in the distribution of HLA-G alleles between controls and Recurrent Spontaneous Abortion (RSA) couples. However, 15% of the RSA women studied carried the HLA-G* 0106 allele, compared to 2% of the control women (Hviid et al., 2002). Carreiras et al. recently showed that the presence of the HLA-G* 0104 and DRB1* 07/06 alleles, HCMV sequences, and the maternal inheritance of G* 0104 should be considered conditioning factors for the development of pre-eclampsia (Carreiras et al., 2002). Moreover Aikhionbare et al. (2001) demonstrated that mother-to-child discordance in exon 2 of HLA-G is associated with a reduced risk of perinatal HIV-1 transmission. Considering that HLA-G may be induced in some tumors (Cabestre et al., 1999) as well as in certain allogenic situations (Lila et al., 2001), identification of the HLA-G allelic forms involved should be taken into account. III. Regulation of HLA-G Gene Expression

HLA-G protein expression in non-pathological situations is very restricted. HLA-G is secreted by blastocysts (Fuzzi et al., 2002) and displays local cell surface expression in extravillous cytotrophoblasts, amnionic epithelial cells, chorionic villi endothelial cells, and adult thymic epithelial cells (McMaster et al., 1995; Blaschitz et al., 1997; Crisa et al., 1997; Hammer et al., 1997b). HLA-G is also activated in Human Cytomegalovirus (HCMV)- and HIV-1infected cells (Onno et al., 2000b; Lozano et al., 2002), tumors (Paul et al., 1999; Ugurel et al., 2001; Urosevic et al., 2001; Ibrahim et al., 2001b; Lefebvre et al., 2002; Ugurel et al., 2002; Urosevic et al., 2002a,b; Wiendl et al., 2002), inflammatory pathologies (Aractingi et al., 2001; Khosrotehrani et al., 2001), transplanted heart, and during the mixed lymphocyte reaction (Lila et al., 2000, 2001, 2002). It is usually reported that HLA-G protein expression correlates with high transcriptional activity, whereas HLA-G

207

HLA-G MOLECULES

transcript levels are generally low or absent in situations in which the protein is absent (Onno et al., 1994). Notably, using RT-PCR analysis, the absence of HLA-G transcripts has been demonstrated in numerous cell-lines (Frumento et al., 1999, 2000), in CD34 þ and NK hematopoietic cells, and in firsttrimester fetal liver (Kirszenbaum et al., 1994, 1995; Amiot et al., 1996; Onno et al., 1997; Moreau et al., 1998). In spite of the fact that HLA-G mRNA accumulates during trophoblast cell differentiation prior to protein synthesis, suggesting tight post-transcriptional control (Copeman et al., 2000), apparently an important part of the regulation of HLA-G expression takes place at the transcriptional level. A. REGULATORY SEQUENCES IN THE PROMOTERS HLA-CLASS I AND HLA-G GENES

OF

CLASSICAL

In contrast to the promoter of classical HLA-class I genes, the HLA-G gene promoter is atypical since almost all conserved cis-regulatory elements are disrupted (van den Elsen et al., 1998a). They comprise two modules located 220 bp from the gene initiation codon (Fig. 2). 1. The Upstream Module of Classical HLA-Class I Gene Promoters The upstream module of classical HLA-class I gene promoters contains (i) the enhancer A, consisting of the B2 and B1 sites, which bind the NFB/rel and Sp1 DNA-binding proteins; and (ii) the interferon-stimulated response element (ISRE), which is a binding site for interferon-induced transcription factors such as interferon regulatory factor (IRF)-1 and interferon-stimulated gene factor (ISGF)-3, as well as repressors of HLA class I transcription, such as IRF-2 and interferon consensus sequence-binding protein (ICSBP) (Gobin et al., 1998a, 1999b). In the HLA-G promoter, enhancer A is unresponsive to NF-B, and the two B sites display binding affinity only for the P50 subunit of NF-B, as well as the Sp1 factor for B2. ISRE is partially deleted in the HLA-G promoter, thus has no binding affinity for proteins of the IRF1 family, and cannot mediate interferon-induced HLA-G expression (Gobin et al., 1998a, 1999b; Gobin and van den Elsen, 2000). 2. The Downstream SXY Module of Classical HLA-Class I Gene Promoters The downstream SXY module of the classical HLA-class I gene promoters has been identified by sequence comparison of HLA class I and HLA-class II promoters and the reduced level of MHC class I expression observed, in addition to the lack of HLA class II expression in the type III Bare Lymphocyte Syndrome (BLS) (van den Elsen et al., 1998b; Masternak et al., 2000). There is increasing evidence that this module is critical for constitutive and inducible expression of both HLA class I and HLA class II genes

208

E. D. CAROSELLA ET AL.

FIG. 2. Schematic representation of the HLA-G gene promoter. Numbers indicate location of regulatory boxes relative to ATG (bp). LCR?: putative locus control region; CRE: three functional CRE/TRE sites bound by CREB1, ATF1 and c-Jun. ISRE: interferon sequence responsive element. Note that only one ISRE site is functional; HSE: heat shock element that bind Heat Shock Factor 1; GAS: nonfunctional Interferon- activated site. B2 and B1 are referred to enhancer A within classical HLA class I promoter. They are disrupted within the HLA-G promoter and display affinity for P50 a subunit of NF-B; The conserved X1 half of X box associates to RFX and Sp1 in vitro; ‘‘?’’ indicates that RFX member factor is not yet identified in vivo. X2 and Y boxes are mutated thus avoiding CIITA induced transactivation of HLA-G gene.

(Gobin et al., 1998b, 2001), and consists of four cis-acting elements (Waldburger et al., 2000), referred to as the ‘‘S box’’ (also called W or Z), the X box, comprising the X1 and X2 (also called  site) halves, and the Y box, which is an inverted CCAAT box (also called enhancer B). The function of the S box is not fully understood; it could possibly play a role in promoter architecture, rather than as a transcription factor-binding site. The X1 box is the binding site for the ubiquitous regulatory factor X (RFX) complex, consisting of the three subunits RFX5, RFXB/ANK, and RFXAP. The X2 box is bound by the X2BP-CREB/ATF factor, and the Y box by a heterotrimer, the NF-Y factor (also referred to as CBF and CP1). Together, these factors are highly cooperative and are the results of a stable, higher-order enhanceosome

HLA-G MOLECULES

209

complex. HLA-class II expression requires the presence of the multi-protein complex in order to interact with the master co-activator CIITA, which is constitutively expressed only in APCs of the immune system, but which may be inducible in other cell types by IFN- (Waldburger et al., 2000; Masternak and Reith, 2002). Similar to HLA class II genes, the HLA class I gene promoter is transactivated by CIITA (Gobin et al., 1997a, 1998b; Lefebvre et al., 1999a). The S and X1 boxes are the only conserved motifs in the HLA-G gene promoter, suggesting a potential role for them in the regulation of this gene. The X2 box mutation was shown to affect the binding of the cAMP response element binding protein/activation transcription factor (CREB/ATF) family of factors in vitro (Rousseau et al., 2000), and the absence of an intact SXY module in the HLA-G promoter has been demonstrated to prevent CIITA recruitment in vivo (unpublished), and consequently, in CIITA-mediated activation of the gene (Gobin and van den Elsen, 2000). Although the conserved X1 box of the HLA-G promoter is a potential target for RFX5 and Sp1 factors, as revealed by Electrophoretic Mobility Shift Assay (EMSA) (Rousseau et al., 2000) chromatin immunoprecipitation experiments (ChIP) reveal this not to be the case in vivo (unpublished). Therefore, in the absence of regulatory pathways shared between the HLA-G and HLA-class I and class II genes, several groups have investigated other elements that may control HLA-G gene transcription. B. SPECIFIC REGULATORY SEQUENCES

OF

HLA-G GENE

One strategy for analyzing the potential site of HLA-G gene regulation has been the use of transient transfection experiments carried out on the HLA-Gpositive JEG-3 choriocarcinoma cell line with luciferase reporter constructs containing HLA-G promoter fragments of different lengths. This allowed the detection of a sole negative regulatory sequence in the region 450 to 220 bp upstream from the first exon of HLA-G gene. The search for other regulatory sequences outside the HLA-G gene promoter region yielded the detection of weakly induced activity in intron 2 (Gobin et al., 1999a). Another strategy was achieved in vivo, using HLA-G transgenic mice into which HLA-G transgenes of different lengths were introduced. This made it possible to show that a distal upstream regulatory element contained in a 244-bp HindIII/EcoR1 fragment located just over 1.1 kb from the first exon of the HLA-G gene was necessary to direct specific HLA-G transcription in transgenic mouse placenta and resulted in similar levels of extra embryonic HLA-G mRNA as those seen in the human (Schmidt et al., 1993; Yelavarthi et al., 1993; Schmidt et al., 1995). The authors suggested that the 244-bp fragment might contain a control locus region, but such a role remains to be clarified, because of the absence of correlation with results obtained with

210

E. D. CAROSELLA ET AL.

in vitro transient transfection experiments. However, several data support the contention that this region participates in the control of HLA-G gene transcription (Fig. 2). Indeed, EMSA carried out with the 244-bp HindIII/ EcoRI region incubated with nuclear extracts from cells and tissues exhibiting positive (JEG-3 and PBMC) and negative (YT2C2 and first-trimester fetal liver) HLA-G transcription revealed shared DNA/protein complexes (C2) and specific complexes (C1, C3, C5, C6, and C7), according to the HLA-G expression status (Moreau et al., 1997, 1998). More recently, computer-aided search within the 1438-bp fragment of the HLA-G gene promoter permitted the localization of three functional cAMP response element/TPA-response element (CRE/TRE) elements (CRE1380, CRE930, and CRE770) shown to have a binding affinity for CREB/ATF and Fos/jun proteins. Interestingly, in vivo binding of CREB-1 and c-Jun has been demonstrated around the CRE site located at position 1380 within the 244-bp HindIII/EcoR1 fragment (Gobin et al., 2002). In that case, the binding did not account for the tissue-specific HLA-G transcriptional activity, since it was also observed in the Tera-2 and Raji HLA-G-negative cell lines. This complex might therefore correspond to the C2 complex resulting from the EMSA analysis, since the binding was previously localized between 1438 and 1311, and was not tissue-specific. HLA-G transactivation by CREB-1 via three binding sites appears to be an alternative pathway to the conserved HLA-class I regulatory routes. Interestingly, the in vitro activation of HLA-G promoter by CREB has been shown to be strongly inhibited by inducible cAMP early repressor (ICER), whereas activation by CREB was enhanced by overexpression of the coactivators CBP and p300. In addition, the in situ nuclear localization of CREB/ATF and CREB binding protein (CBP) in extravillous cytotrophoblast strongly supports the contention these transcription factors play a role in the HLA-G transactivation observed in trophoblast cells (Gobin et al., 2002). On the other hand, several groups have investigated the impact of cytokines, stress, and hormones on the modulation of HLA-G gene expression. Despite the absence of a classical ISRE site, an increase in HLA-G protein and HLA-G mRNA after IFN treatment has been described in different studies, suggesting that another HLA-G specific sequence could be involved (Yang et al., 1996; Chu et al., 1999b; Lefebvre et al., 2001). The question is still being debated, since promoter activity of the region encompassing 1.4 kb from ATG appears to be modestly affected by IFN treatment, compared with the classical HLA class-I gene (Gobin et al., 2002). Nevertheless, although a candidate IFN- activated site (GAS) located at position 740 in the HLA-G promoter was shown to be nonfunctional (Chu et al., 1999a), an ISRE sequence located at position 746 is capable of transactivating HLA-G

HLA-G MOLECULES

211

following IFN- treatment (Lefebvre et al., 2001). HLA-G promoter activity was weakly induced in the JEG-3 cell line, but reproducible and clearly significant in the thymic epithelial cell LT-TEC2 line, with an average of 2.8-fold enhancement. Moreover, using EMSA, this work also provided evidence for the binding of IRF-1 to the functional HLA-G ISRE in response to IFN- (Lefebvre et al., 2001). The HLA-G gene may also be activated by stress conditions. Application of heat shock at 42 C or arsenite treatment for 2 h, reverses HLA-G gene repression in the M8 melanoma cell-line without affecting other MHC class I genes. Interestingly, HLA-G6 expression is induced prior to that of the other HLA-G transcripts early during the recovery time after stress treatment, which also indicates tight control of HLA-G alternative splicing. The study also identifies a heat shock element (HSE) located between positions 487 and 466 of the HLA-G gene promoter that binds to heat shock factor 1 (HSF1) in vitro under stress conditions (Ibrahim et al., 2000) (Fig. 2). C. EPIGENETIC MECHANISMS

MAY

REGULATE HLA-G GENE EXPRESSION

DNA methylation and histone modification are interrelated mechanisms known to play a key role in transcriptional control (El-Osta and Wolffe, 2000; Bird, 2002). CpG methylation has been analyzed in the JAR choriocarcinoma cell-line, revealing the activation of HLA-G transcription after treatment with demethylating agent 5-azacytidine. Conversely, no correlation was observed between HLA-G gene transcriptional activity and methylation of CpG islands in the 50 -part of the HLA-G gene in cells that either express HLA-G transcripts (trophoblasts, JEG-3 cells and CD2 þ lymphocytes) or that do not (syncytiotrophoblasts and CD34 þ cells) (Onno et al., 1997). However, these studies are limited to the proximal promoter region of the HLA-G gene, and questions remain concerning potential subsets of CpG sites that might be important in HLA-G gene silencing. A trans-acting demethylating process could also be envisaged that reverses repression of specific transcriptional factors, which in turn would activate HLA-G transcription. Using cell lines negative for HLA-G transcription that exhibits various phenotypes (B-EBV and tumor cell), recent data show that repression of the HLA-G gene by DNA methylation is a more general mechanism than expected. Indeed, although cell exposure to histone deacetylase inhibitors only activated HLA-G transcription in melanoma cells, cell exposure to the demethylating agent 5-aza-20 -deoxycytidine reversed HLA-G gene silencing in all cell-lines studied. In particular, strong HLA-G repression in Raji cells is reversed by 5-aza-20 -deoxycytidine treatment, giving rise to both HLA-G transcription and protein expression (Moreau et al., 2003). Therefore, despite

212

E. D. CAROSELLA ET AL.

in vivo binding of CREB1 factors to the HLA-G gene promoter in these cells (Gobin et al., 2002), methylation represses HLA-G gene transcription. Whether cis- or trans-acting mechanisms occur and involves particular methyl-binding or cofactors must be elucidated. D. HLA-G CAN BE UPREGULATED BY SOLUBLE MEDIATORS PLACENTAL OR TUMOR MICROENVIRONMENTS

OF

Several groups have investigated the effects of factors secreted locally in placental or tumor microenvironments on HLA-G transcriptional activity and HLA-G cell-surface expression. IL-10 is produced in placenta during all stages of gestation, as well as in tumor cells and may be secreted at high levels in the serum of cancer patients (Dummer et al., 1995; Roth et al., 1996; Bennett et al., 1997; Yue et al., 1997; Urosevic et al., 2002b). Serum IL-10 has also been shown to be elevated in patients with orthotopic liver transplants demonstrating non-acute rejection (Minguela et al., 1999). IL-10 exhibits a broad spectrum of biological activities, including anti-inflammatory and immunosuppressive ones (Goldman and Stordeur, 1997). This interleukin has been shown to selectively upregulate HLA-G transcription in cultured trophoblast explants, and to enhance both HLA-G transcription and cellsurface expression of HLA-G protein in monocytes (Moreau et al., 1999). Accordingly, CMV produces a biologically active IL-10 homolog that can upregulate HLA-G protein expression at the monocyte cell surface (Spencer et al., 2002). Glucocorticoid hormones are also widely expressed in placental tissues (Sun et al., 1998) and have been shown to upregulate HLA-G transcript levels in trophoblast explants treated with dexamethasone and hydrocortisone (Moreau et al., 2001). Transactivation of HLA-G transcription has also been demonstrated by JEG-3 cell exposure to Leukemia Inhibitory Factor (LIF) (Bamberger et al., 2000), a pleiotropic cytokine shown to be expressed at the maternal–fetal interface and which plays an essential role in implantation in mice (Stewart et al., 1992). Moreover, Granulocyte-Macrophage Colony Stimulating Factor (GM-CSF), in combination with IFN- , have been shown to stimulate HLA-G cell-surface expression on the U937 monohistiocytic cell line (Onno et al., 2000a). The mechanisms, by which IL-10, glucocorticoid, LIF, and GMCSF stimulate HLA-G transcription, or both HLA-G transcription and HLA-G protein expression, are still generally unknown, and some of them could involve both transcriptional and posttranscriptional processes. It has notably been observed that these latter factors generally have no effect on HLA-G gene transcription in cells in which the HLA-G gene is strongly repressed (Frumento et al., 2000). This reinforces the hypothesis of

HLA-G MOLECULES

213

the existence of epigenetic mechanisms that may control HLA-G gene activation. In conclusion, evolution has developed very specific and multifactorial pathways for the transcriptional regulation of the HLA-G gene, which, in combination with post-transcriptional mechanisms (mRNA stability and protein processing) contribute to the temporal regulation that occurs in placenta (Solier et al., 2001), and to the constitutive or inducible expression of the HLA-G molecule in other tissues. On the other hand, HLA-G expression may be upregulated in monocytes and T lymphocytes of HIV þ patients, in activated macrophages infected with Human Cytomegalovirus (HCMV), probably due to viral IL-10-like protein, and in HCMV-infected U-373 MG astrocytoma cells, where IE-pp72 and IE-pp86 HCMV proteins are involved (Onno et al., 2000b; Lozano et al., 2002). Therefore, some viruses seem to have developed a strategy for controlling HLA-G expression that consist in taking advantage of its specific immune tolerance function. IV. Processing and Transport of HLA-G Molecules

Folding, assembly, and peptide loading of HLA-class I molecules are critical for the regulation of their cell-surface expression, and involve intracellular protein complexes and chaperone molecules (Neefjes and Momburg, 1993). Schematically, cytosol and nuclear proteins are degraded by the proteasome, and then transported into the lumen of the endoplasmic reticulum (ER) by a heterodimeric peptide transporter, which consists of TAP1 and TAP2. At the same time, the class I chain and 2-microglobulin are assembled in the ER with the help of calnexin and calreticulin. The peptides are then loaded into the groove of the HLA class I heavy-chain, which in turn associates 2-microglobulin. The glycosylation pattern of the class I heavy-chain is modified once the complex has exited the lumen of the ER and the heterotrimer is transported through the golgi apparatus to the cell surface. Due to alternative splicing of the HLA-G primary transcript, HLA-G presents seven isoforms (Fig. 1) (Ishitani and Geraghty, 1992; Fujii et al., 1994; Kirszenbaum et al., 1994; Moreau et al., 1995; Paul et al., 2000a): HLA-G1 is the complete membrane-bound protein; HLA-G2, -G3, and -G4 are membrane-bound isoforms respectively devoid of the 2, 23, and 3 extracellular domains. The intracytoplasmic domains of these isoforms are all reduced, in comparison with those of classical HLA-class I molecules. In addition, the existence of transcripts in which intron 4 or 2 are retained generates three soluble isoforms, known as HLA-G5 (G1-like), -G6 (G2-like), and -G7 (G3-like), respectively (Fig. 1). The use of viruses that affect classical HLA class I protein expression, in combination with biochemical approaches, has revealed common pathways in

214

E. D. CAROSELLA ET AL.

the assembly and transport of HLA-G and HLA class I molecules (Schust et al., 1999). Nevertheless, the non-classical structural properties of HLA-G isoforms suggest the existence of specific HLA-G trafficking pathways. Most of them have been investigated focusing on HLA-G1 and HLA-G5, for which suitable reagents were available. A. HLA-G BINDS INTRACELLULAR PEPTIDES MECHANISM

WITH A

TAP-DEPENDENT

The question about the ability of the HLA-G molecule to present a peptide was first investigated in comparison with the HLA-A2 molecule. Similarities found in the 1 and 2 domains, which form the peptide pocket, suggested that some peptides presented by HLA-A2 might also be presented by HLA-G1 (Geraghty et al., 1987). Biochemical analysis of the HLA-G1 (membrane-bound) and the HLA-G5 (soluble) forms of the molecule has demonstrated that both proteins, which are expressed in the LCL721.221 lymphoblastoid cell line, consisted of heavy-chain/beta-2-microglobulin/ peptide in a 1 : 1 : 1 ratio. Peptide elution experiments confirmed that the peptides presented by HLA-G1 and HLA-G5 were of the same size as those carried by HLA-A2, and consisted of nine amino acids with the XI/ LPXXXXXL consensus sequence (Lee et al., 1995; Diehl et al., 1996). In accordance with the observed sequence similarities between HLA-G1 and HLA-A2, anchoring positions are located on amino acids 2 (isoleucine or leucine), 3 (proline), and 9 (leucine), although positions 2 and 9 are enough for efficient anchoring (Lee et al., 1995). In transfected cells, the diversity of peptides eluted from HLA-G1 and HLA-G5 was estimated to be around five-fold lower than with classical HLA class I, a property that may be related to the HLA-G molecule’s invariance in the peptide-binding groove region (Lee et al., 1995). On the other hand, in the absence of data obtained with cytotrophoblasts and thymic epithelial cells, the limited peptide diversity could be different from the repertoire of individual peptides bound in vivo. However, it was suggested that HLA-G could bind peptides derived from viruses, a limited number of which infect trophoblasts, and may be presented in order to allow HLA-G anti-viral function (Le Bouteiller and Blaschitz, 1999). Notably, one report described that alloreactive HLA-G-specific CD8 þ CTL clones could be obtained in vitro in a peptide-dependent fashion (Horuzsko et al., 1999). However, whether HLA-G can induce HLA-G restricted CTL responses against viral protein remains to be elucidated. Although there is strong evidence that HLA-G1 and HLA-G5 bind peptide, the question is still open as to whether there are peptides that bind to the other HLA-G isoforms, namely HLA-G2, -G3, -G6, and -G7. The 1 and 2 domains, which form the peptide groove, are present only in HLA-G4, which

HLA-G MOLECULES

215

might therefore be a candidate peptide-presenter. Another hypothesis to explore would be the possible formation of HLA-G2, HLA-G4, and HLA-G3 homodimers, which might permit peptide loading similar to HLA-class II antigens (Ishitani and Geraghty, 1992). Is peptide loading of HLA-G1 and HLA-G5 a TAP-dependent mechanism? Studies seeking HLA-G1 and HLA-G5 expression in the transfected TAPnegative B-LCL 721.131 lymphoblastoid cell line showed that HLA-G1 expression is reduced to around 20%, with respect to expression in the B-LCL721.221 TAP-positive cell line, whereas no HLA-G5 protein could be detected in the supernatant of the TAP-negative cells. HLA-G therefore uses both TAP-dependent (HLA-G1 and HLA-G5) and TAP-independent (HLA-G1) pathways to bind peptide (Lee et al., 1995). Accordingly, a 10fold downregulation was observed with a remainder HLA-G1 cell-surface expression at the surface of HLA-G-LCL721.221 cells transfected with the ICP47 protein of the Herpes simplex virus (HSV), a viral protein that has been shown to block the TAP transporter (Munz et al., 1999). In the JEG-3 choriocarcinoma cell-line infected with (HSV) or transfected with ICP47 protein, HLA-G1 fails to acquire endoglycosidase-H (endo-H) resistance, thus is retained in the endoplasmic reticulum (Schust et al., 1996). In agreement with HSV-related experiments, HLA-G1 has been shown to be sensitive to endo-H treatment in JEG-3 cells transfected with human cytomegalovirus (HCMV)-US6, a gene that encodes a glycoprotein that also prevents peptide-loading of the HLA-class I molecule by inhibiting the TAP complex (Jun et al., 2000). The existence of an interaction between the TAP complex and HLA-G1 has been demonstrated by immunoprecipitation studies (Lee et al., 1995; Gobin et al., 1997b). However, the absence of a detectable association between TAP and HLA-G5 has led to the hypothesis of the possible binding of a cytosolic peptide without interaction. This also suggests that either the residues that interact with TAP are located in the cytoplasmic tail of HLA-G1, or that the unique 21-amino acids located in the C-terminal of HLA-G5 interfere with TAP association. The peptide-loading process thus remains to be investigated, and is apparently associated with either the soluble- or membrane-form state of the molecule. Whatever the loading mechanisms involved may be, the following observations strongly argue that the HLA-G heavy-chain requires peptide association for membrane expression: First, peptide-deficient HLA-G1 molecules are rapidly degraded in the ER of US6-expressing HeLa cells (Park et al., 2001) and, second, HLA-G1 cell-surface expression was entirely abolished in cells with both TAP and proteasome deficiencies (unpublished). The loading of HLA-G1 is not strictly TAP-dependent (as is the case for HLA-G5) and may vary according to cell type. This suggests the existence of

216

E. D. CAROSELLA ET AL.

an alternative pathway for reaching the cell surface, as described for classical HLA-class I molecules. B. SPECIFIC FEATURES OF THE HLA-G MOLECULE INFLUENCE TRANSPORT TO THE CELL SURFACE

ITS

Compared with classical HLA-class I molecules, HLA-G is likely to have evolved as a more resistant antigen, thus facilitating cell traffic and conferring resistance to some virus products: (a) HLA-G molecules produced by cytotrophoblasts and found in amniotic fluid bear unusual carbohydrate structures, leading to a broad molecular weight range of HLA-G-immunoreactive bands; (b) like classical HLA class I molecules, HLA-G contains a single N-linked glycosylation site. However, digestion of HLA-G proteins from placenta by endo- D-galactosidase suggests that these molecules specifically carry N-acetyllactosamine units, which might stabilize the molecule (McMaster et al., 1998); (c) in both JEG-3 and 2A2 porcine bone marrowderived stromal cells, HLA-G1 has been shown to be resistant to dislocation and degradation mediated by HCMV US2 and US11 gene products, sharing this property with HLA-Cw3, -Cw4. In vitro co-transcription/ translation of class I heavy chains with US2 and US11 gene products followed by immunoprecipitation experiments have demonstrated that these proteins do not associate with either HLA-G or HLA-C molecules (Schust et al., 1998). Another striking feature is the shortened cytoplasmic domain of the HLA-G molecule, due to a premature stop codon in exon 6 (Geraghty et al., 1987). This deletes potential endocytosis signals found in the cytoplasmic tails of all HLA-class I molecules. Consequently, spontaneous endocytosis of HLA-G1 was shown to be reduced, compared to classical class I molecules, resulting in the prolonged retention of molecules at the cell-surface (Davis et al., 1997). In addition, the dilysine residue motif positioned 2–3 amino acids from the C-terminal end of the cytoplasmic domain allows interaction with the coat protein complex (COP), which is known to mediate retrieval from post-ER compartments (Teasdale and Jackson, 1996; Davis et al., 1997). Demonstration that the dilysine motif function acts as a retrieval signal for HLA-G from the post-ER compartment to the endoplasmic reticulum and that it is responsible for the slow transport kinetics of HLA-G was made by Park et al., both in JEG-3 and LCL721.221 transfected cells, which stably express HLA-G1. Interestingly, the loading of HLA-G1 with high-affinity peptides (KIPAQFYIL) instead of low-affinity peptides (KGGAQFYIL), has been shown to avoid retrieval of the HLA-G molecule, resulting in increased cell-surface expression of HLA-G1 (Park et al., 2001). The authors therefore proposed that the loading of these high-affinity peptides into the HLA-G groove could induce conformational changes that might alter recognition of

HLA-G MOLECULES

217

the dilysine motif. This in turn might inactivate the retrieval motif and result in upregulation of HLA-G1 cell-surface expression. The shortened cytoplasmic domain of the HLA-G1 molecule thus plays a critical role in quality control of the HLA-G molecule. Finally, HLA-G may be expressed as three other membrane-bound isoforms, HLA-G2, G3, and G4, all of which are present in cytotrophoblasts (Menier et al., 2000). As will be mentioned in the following chapter, the question of their expression at the cell surface has been investigated by three groups and produced debatable results (Bainbridge et al., 2000b; Mallet et al., 2000; Riteau et al., 2001c). HLA-G may be also be secreted. In particular, the soluble HLA-G5 protein includes the three extracellular domains of HLA-G1, but is shorter than the membrane-bound molecule (Fujii et al., 1994). HLA-G5 thus does not possess a tail endowed with dilysin residues, which are involved in quality control for reaching the cell-surface. Second, HLA-G5 exhibits a unique, 21-amino acid carboxyl terminus that confers solubility to the molecule. It is therefore tempting to suggest that these structural properties could render certain viral protein/HLA-G1 interactions ineffective. In agreement with this hypothesis is the absence of interaction of HLA-G5 with TAP, as mentioned above. Finally, it is noteworthy that the soluble HLA-G6 isoform shares common C-terminal properties with HLA-G5, and may thus also enjoy favored expression. In particular, the HLA-G6 protein isoform has been shown to circulate in maternal blood during pregnancy (Hunt et al., 2000). Therefore, convergent data show that HLA-G1 exhibits the typical structure of an antigen-presenting element. The demonstration that the HLA-Grestricted specific CD8 þ T cell repertoire is selected in HLA-G transgenic mice argues in favor of HLA-G protein being a restricting element. Nevertheless, the primary function of HLA-G might not be peptide presentation to the TCR, since the molecule exhibits low peptide diversity and slow turnover, which would be inefficient in presenting exogenous peptides. As recently proposed, peptide-loading could be a key mechanism in the transport of HLA-G to the cell-surface, acting in a quality-control process (Park et al., 2001). V. Structural and Functional Properties of HLA-G Molecules

The HLA-G primary transcript is alternatively spliced, resulting into seven alternative mRNAs which encode four membrane-bound (HLA-G1, -G2, -G3, -G4), and three soluble (HLA-G5, -G6, -G7) protein isoforms (Fig. 1) (Ishitani, 1992; Kirszenbaum et al., 1994; Paul et al., 2000a). The structural and functional properties of these HLA-G isoforms are presented below.

218

E. D. CAROSELLA ET AL.

A. THE FULL-LENGTH HLA-G1 PROTEIN HLA-G1 mRNA encodes a 39-kDa full-length protein that contains 1, 2, and 3 extracellular domains linked to the transmembrane domain encoded by exon 5 and to a shortened cytoplasmic tail, due to a premature stop codon in exon 6. This isoform presents a structure similar to that of the other HLA class I molecules, in that it is non-covalently associated with 2-microglobulin and binds a nonapeptide (Lee et al., 1995). The N-linked glycosylation site (Asn 86) and the consensus cysteine pairs in the 2 and 3 domains are also conserved (Geraghty et al., 1987). Of note, the HLA-G1 heavy chain/ 2m/peptide heterotrimer has been recently described in a dimerized form expressed on the cell-surface. This dimerization is mediated through disulfide bonds of Cys-42 of the heavy chain (Boyson et al., 2002). The availability of monoclonal antibodies, such as 87G (Lee et al., 1995), 01G (Paul et al., 2000b), G233 (Loke et al., 1997), and MEM-G/09 (Lozano et al., 2002; Menier et al., 2003), able to specifically react with cell-surface HLA-G1, has widely permitted analysis of both the expression and function of this isoform in vitro and ex vivo. Several lines of evidence indicate that the primary function of HLA-G1 is to serve as an inhibitory ligand for immunocompetent cells, thus contributing to the establishment and maintenance of immune tolerance (Fig. 3): First, HLA-G1 inhibits the CD4 þ T cell proliferative response in allogeneic mixed lymphocyte reactions (Riteau et al., 1999; Bainbridge et al., 2000a). Second, HLA-G1 impairs the ability of HLA-A2-restricted, viral antigenspecific CD8 þ T cells to lyse target cells (Le Gal et al., 1999). Third, HLA-G1 expressed either spontaneously by trophoblasts or following transfection of the encoding cDNA by human cell lines (i.e., K562 erythroleucemic cells, lymphoblastoid LCL721.221 cells, or M8 melanoma cells), porcine endothelial cells, Chinese hamster ovarian cells, and J26 murine fibroblasts, inhibits both decidua and peripheral blood NK cell-mediated cytolysis (Rouas-Freiss et al., 1997a,b; Khalil-Daher et al., 1999; Navarro et al., 1999; Ponte et al., 1999; Sasaki et al., 1999, 1999a; Forte et al., 2000; Riteau et al., 2001a). These inhibitory effects of HLA-G1 are mediated through direct binding to inhibitory receptors, namely the immunoglobulin-like transcript (ILT2/CD85j) expressed by lymphoid and myelomonocytic cells (Colonna et al., 1997; Cosman et al., 1997; Saverino et al., 2000), ILT4 (CD85d) expressed by monocytes, macrophages, and dendritic cells (Colonna et al., 1998), and p49/KIR2DL4 (CD158d) expressed by NK and T cells (Ponte et al., 1999; Rajagopalan and Long, 1999). However, the nature of the HLA-G-KIR2DL4 interaction needs to be explored in more detail since divergent results have been obtained on this point (Boyson et al., 2002). Therefore, through these receptors the HLA-G1 protein can directly interact with T, B, NK, and APC

HLA-G MOLECULES

219

FIG. 3. Functional implications of interactions between HLA-G1 and immune cells such as, T cell, NK cell, and APC.

and exert its immunotolerant functions at different stages of the immune response. The biological relevance of HLA-G1 expression by APC, such as monocytes/macrophages and dendritic cells bearing ILT2 and ILT4 receptors, which have been described as binding soluble HLA-G1 tetrameric complexes (Allan et al., 1999), remains to be assessed in humans. However, we may hypothesize that such functions play an important role in blocking the host’s immune initiation, as recently proposed in a murine model (Horuzsko et al., 2001; Liang et al., 2002). HLA-G1 has been described to bind CD8 with an affinity (150 M) comparable to that of the CD8/HLA-A2 interaction (Gao et al., 2000). Therefore, it is possible that HLA-G1/CD8 interaction may facilitate HLAG-restricted antigen recognition by CTLs. However, although HLA-G1 binds nonamer peptides and CD8, it remains to be ascertained whether HLA-G1 is capable of presenting bacterial or viral peptides to T cells and elicit an HLA-G-restricted cytotoxic T cell-mediated response in a way similar to other HLA class I molecules. Finally, HLA-G1 expressed on transfected cells has been shown to modulate the release of cytokines from peripheral blood and decidual mononuclear cells. The amounts of TNF- and IFN- released from decidual and peripheral blood mononuclear cells were found to be decreased, while the amount of IL-4 was increased (Kanai et al., 2001a,b). Similarly, a Th2-type cytokine response (upregulation of IL-10 and IL-4 with downmodulation of

220

E. D. CAROSELLA ET AL.

TNF- and IFN- ) has been associated with exposure to high concentrations of HLA-G1 purified from first-trimester trophoblast tissue (Kapasi et al., 2000). However, HLA-G1 has recently been found to affect the general cytokine production of decidual large granular lymphocytes in a manner not consistent with the Th1/Th2 paradigm (i.e., downmodulation of IL-10, IL-13, TNF-, IFN- , and GM-CSF) (Rieger et al., 2002). Interestingly, HLA-G1 protein cell-surface expression may in turn be induced by several cytokines, such as interferons (Yang et al., 1995, 1996; Amiot et al., 1998; Chu et al., 1998, 1999b; Lefebvre et al., 1999b, 2000, 2001; Ugurel et al., 2001), IL-10 (Moreau et al., 1999; Spencer et al., 2002), GM-CSF (Amiot et al., 1998), and LIF (Bamberger et al., 2000). B. THE HLA-G2, -G3,

AND

-G4 TRUNCATED PROTEINS

HLA-G2, -G3, and -G4 transcripts, which respectively exclude exon 3, exons 3 and 4, and exon 4, generate truncated isoforms that retain only the 1 domain for HLA-G3, the 1 and 3 domains for HLA-G2, and the 1 and 2 domains for HLA-G4, joined to the transmembrane region. Their ‘‘non conventional’’ conformational structure, exhibiting either one or two extracellular domains (monomers, dimers, heterodimers, or polymers) remains to be elucidated. In contrast to HLA-G1, analysis of the expression pattern of these other HLA-G isoforms remains difficult, due to the absence of reagents able to specifically recognize each of them. However, the availability of the 4H84 monoclonal antibody, raised against peptide 61–83 of the HLA-G 1 domain common to all HLA-G isoforms (McMaster et al., 1998), has allowed characterization of these truncated isoforms. It should be noted that a new mAb exhibiting a similar pattern of recognition to that of 4H84 is now available (MEM-G/01) (Lozano et al., 2002; Menier et al., 2003). By compiling data on the expression of HLA-G2, -G3, and -G4 proteins, several conclusions may be drawn: (i) they are expressed physiologically in cytotrophoblast (Menier et al., 2000), pathologically in tumor cell-lines, such as choriocarcinoma (Adrian Cabestre et al., 1999b), and melanoma (Paul et al., 1998; Adrian Cabestre et al., 1999b), as well as in transfected cells (Mallet et al., 2000; Bainbridge et al., 2000c; Riteau et al., 2001b,c) as glycoproteins of 31-, 22-, and 29-kDa, for HLA-G2, -G3, and -G4, respectively; (ii) their cellsurface expression is likely to be dependent on the cell type in which they are expressed. Indeed, HLA-G2, -G3, and -G4 proteins linked to a tag molecule and expressed in HLA class I-negative cell lines could not be detected on the cell surface (Bainbridge et al., 2000c; Mallet et al., 2000). In contrast, each of these untagged isoforms was found to be expressed on the cell-surface of a transfected HLA-A, -B, -C, and -E-positive melanoma cellline as endoglycosidase-H (Endo-H)-sensitive (i.e., exhibiting an immature

HLA-G MOLECULES

221

glycosylation pattern) cell-surface glycoproteins after a 2 h chase period. It has been suggested that in the latter cell model, an optimal environment is provided, allowing the possible association of these isoforms with chaperone molecule(s), and their consecutive cell-surface expression. Although cellsurface proteins containing immature oligosaccharides is unusual (HLA-G1 and classical HLA class I are Endo-H-resistant cell-surface molecules), it has been reported for other proteins, such as for the HLA class I-like CD1d, and for the mouse cytomegalovirus gp34 protein which escapes retention from endoplasmic reticulum (ER) by associating with folded class I MHC molecules. Similarly, truncated HLA-G isoforms may associate with a chaperone protein, such as other HLA class I molecules, allowing their escape from ER retention, and their cell-surface expression. By using the HLA class I-positive melanoma cell line in which HLA-G2, -G3, and -G4 isoforms were expressed on the cell surface, functional properties of these isoforms could be investigated. These truncated isoforms were found to inhibit both NK and antigen-specific CTL cytolysis through an HLA-E-independent pathway (Riteau et al., 2001c). Indeed, the lytic activity of the YT2C2-PR NK clone, which does not express CD94/NKG2A, the HLA-E-specific inhibitory receptor, was inhibited by the HLA-G isoforms, and the HLA-G isoformmediated inhibition of polyclonal NK lysis was not reversed by blocking HLA-E/CD94NKG2A interactions. The extracellular 1 domain, which is shared by all HLA-G isoforms, is thought to mediate these inhibitory properties. These truncated isoforms might become relevant in situations in which HLA-G1 expression is hampered, such as in 2-m-deficient cells, since HLA-G2, -G3, and -G4 may not require association with 2-m for their expression (Riteau et al., 2001c), and in individuals homozygous for the HLAG* 0105N ‘‘null allele’’ (Ober et al., 1998; Castro et al., 2000c; Arnaiz-Villena et al., 2001; Moreau et al., 2002). The potential functionality of this allele, which might indeed produce functional truncated HLA-G molecules at the fetal–maternal interface, has recently been discussed (Moreau et al., 2002). C. THE SOLUBLE HLA-G PROTEINS The soluble HLA-G isoforms are devoid of transmembrane and cytoplasmic parts, due to the presence of a stop codon in intron 4 (-G5, and -G6) or intron 2 (-G7) leading to a C-terminal tail specific for these soluble forms (Fig. 1). The soluble full-length HLA-G5 isoform is a 37-kDa glycoprotein that retains identical leader, 1, 2, and 3 domains, but includes an intron 4 sequence, yielding a specific open reading frame that encodes 21-amino acids linked to the 3 domain and excludes the transmembrane domain (Lee et al., 1995). Similarly, the HLA-G6 isoform retains an intron 4 sequence, yielding a 27-kDa soluble protein that lacks the 2 domain (Paul et al., 2000a). HLA-G7,

222

E. D. CAROSELLA ET AL.

the most recently described isoform, is a 17-kDa soluble protein produced by a spliced variant in which the open reading frame continues in intron 2, which contains a stop codon. Thus, HLA-G7 is formed by the 1 domain linked to two specific amino acids encoded by intron 2 (Paul et al., 2000a). Due to the availability of antibodies, such as the monoclonal antibody 16G1 (Lee et al., 1995) and the polyclonal PAG5-6 antibody (Paul et al., 2000b) that are specific for the C-terminal part (i.e., intron-4-encoded residues) of both HLA-G5 and HLA-G6, the expression of these isoforms can now be investigated. Furthermore, the antibodies described above to recognize HLAG1 cell-surface molecules also react with the soluble HLA-G5 protein. In this regard, various assays based on the use of these antibodies have been described for measuring the level of soluble HLA-G in biological fluids. However, most of these assays do not distinguish between soluble HLA-G1 protein (i.e., HLA-G1s) released by the shedding of membrane-bound HLA-G1, and the soluble HLA-G5 protein, produced from a specific spliced transcript. HLA-G5 protein has been described in various body fluids, such as amniotic fluid and serum, from pregnant women (Puppo et al., 1999; Rebmann et al., 1999; Hamai et al., 1999a), cancer patients (Adrian Cabestre et al., 1999a; Ugurel et al., 2001), and transplanted patients (Lila et al., 2000, 2002). HLA-G5 has been also described in trophoblast (Chu et al., 1998; Fournel et al., 2000b), thymus (Mallet et al., 1999a,b), oocytes (Menicucci et al., 1999), pre-implantation embryo (Jurisicova et al., 1996b; Fuzzi et al., 2002), and during human cytomegalovirus reactivation (Onno et al., 2000b). It has been suggested that soluble HLA-G6 is present in maternal blood during pregnancy (Hunt et al., 2000), and until now has been clearly detected in vivo in serum from heart-transplanted patients (Lila et al., 2000, 2002). The presence of HLA-G7 as a secreted protein in both transfected cell supernatants and body fluids remains to be detected. While the functional properties of the HLA-G6 and -G7 soluble proteins remain to be characterized, those of the HLA-G5 soluble protein have been studied, using either HLA-G5 recombinant protein produced by prokaryotic (Marchal-Bras-Goncalves et al., 2001) or eukaryotic (Fournel et al., 1999, 2000a; Kanai et al., 2001a; Contini et al., 2003) cells, or HLA-G5 naturally produced by trophoblast cells (Kapasi et al., 2000) and T cells (Lila et al., 2001). Recombinant HLA-G5 exhibits inhibitory properties and binds to CD8 and inhibitory receptors, as does HLA-G1. Indeed, the HLA-G5 soluble form inhibits NK cell- and CD8 þ T cell-mediated lysis (Marchal-Bras-Goncalves et al., 2001; Contini et al., 2003) and the allogeneic CTL response (Kapasi et al., 2000). Moreover, HLA-G5 naturally produced by alloreactive CD4 þ T cells inhibits their proliferative response, thus also exerting its inhibitory effects through a feedback mechanism (Lila et al., 2001). On the other hand,

223

HLA-G MOLECULES

conflicting information has been reported regarding the ability of HLA-G5 to induce apoptosis of activated CD8 þ T cells. While recombinant HLA-G5 purified from transfected cells has been found to induce apoptosis of phytohemagglutinin-activated CD8 þ cells through binding to CD8 and in a Fas/FasL-dependent-manner (Fournel et al., 2000a; Contini et al., 2003), this soluble form naturally produced by alloreactive CD4 þ T cells during the mixed lymphocyte reaction did not enhance their apoptosis (Lila et al., 2001). Furthermore, an HLA-G-negative glioma cell-line transfected with HLA-G5 cDNA and co-cultured with freshly isolated CD4 þ or CD8 þ lymphocytes did not induce apoptosis in these effector cell populations (Wiendl et al., 2002). Finally, recombinant HLA-G5 has been found to influence the release of cytokines from peripheral blood mononuclear cells by stimulating the release of IL-10, TNF-, and IFN- (Kanai et al., 2001a). Such a soluble protein may have particular relevance in vivo, as it exerts its functions in the environment near its site of origin, and/or at distant sites, because of distribution via the circulatory system. D. RELATIONS BETWEEN HLA-G

AND

HLA-E EXPRESSION

AND

FUNCTION

In addition to its direct inhibitory role in interacting with the abovementioned immune receptors, HLA-G may also exert its immunosuppressive effects via an indirect pathway by favoring the expression of another non-classical HLA class I molecule, HLA-E. Indeed, peptides derived from the leader sequences of several HLA class I molecules, including HLA-G, bind to HLA-E and stabilize its conformation, allowing its stable cell-surface expression. Thus, when HLA-G is expressed in cells or tissues, co-expression of cell-surface HLA-E molecules may also occur. HLA-E inhibits both NK and T cells by interacting with the CD94/NKG2A inhibitory receptor, present on both these immune cells (Braud et al., 1998; Lee et al., 1998a,b). The contributions of the direct and indirect pathways to HLA-G-mediated inhibition have been evaluated in the context of the co-expression of HLA-G and naturally expressed HLA-E and the classical HLA class I molecules. Experiments carried out with polyclonal NK cells and CTL have demonstrated that when HLA-G1 is co-expressed with HLA-E, the blockade of HLA-E/CD94/NKG2A interaction does not reverse the inhibition of NK and CTL cytolysis (Le Gal et al., 1999; Riteau et al., 2001a,c). These experiments provide evidence that, in this context, HLA-G1 becomes the major NK and CTL inhibitory ligand. Furthermore, the leader peptide of HLA-G2, -G3, and -G4 might also permit stable HLA-E cell-surface expression, thus constituting an indirect inhibitory pathway by blocking lysis by CD94/NKG2A þ effector cells. In this regard, the triple transfection of human 2-microglobulin, HLA-E, and HLA-G3 genes into swine endothelial cells leads to HLA-E cell-surface expression, as well as inhibition

224

E. D. CAROSELLA ET AL.

of human polyclonal NK cell-mediated lysis of such transfected cells (Matsunami et al., 2002). However, such a hypothesis for HLA-G2 and HLAG4 remains to be verified. Interestingly, HLA-E is also able to interact with specific activating receptors, such as CD94/NKG2C, present on NK cells. Of particular interest is that recognition of HLA-E by both CD94/NKG2A and CD94/NKG2C receptors is influenced by the HLA class I leader peptide sequence bound to HLA-E (Vale´s-Gomez et al., 1999). HLA-E loaded with the HLA-G-derived nonamer can interact with both receptors; it very efficiently inhibits the cytolytic activity of CD94/NKG2A þ NK clones and promotes a strong cytotoxic response by CD94/NKG2C þ NK clones by interacting with this activating receptor with a high affinity (Llano et al., 1998). Although the in vivo functional role of such an HLA-E/HLA-G peptide complex remains to be elucidated, it should be noted that target cells exhibiting both HLA-E and HLA-G cell-surface molecules are strongly protected from the lytic activity of polyclonal peripheral blood NK cells, showing a dominant inhibitory activity. In this case, as mentioned above, the major NK lysis inhibitory effect is mediated by the interaction of HLA-G with ILT-2 present on peripheral blood NK cells. E. HLA-G, CYTOKINES,

AND INFLAMMATION

Cytokine-interleukins are regulators of the host response during the inflammatory processes occurring during immune reactions, infections, and trauma. Pro-inflammatory cytokines and anti-inflammatory cytokines intervene during this process; the first acts mainly on TNF, IL1 , IFN , IL5, IL8, IL12, IL16, IL17, IL18, MIP1, MIP1 , and G-CSF; the second on IL4, IL6, IL10, IL11, IL13, TGF , and EGF. The majority of these molecules play a major role in the immune response and intervene in the lymphocyte balance Th1-Th2 (Dinarello, 2000; Opal and DePalo, 2000). Th1 cells are implicated in cellular immunity and macrophage-dependent inflammation. A dominant reply of this T cell subpopulation is thought to be involved in the pathogenicity of auto-immune specific organ diseases, of acute graft reject, of inflammatory dermatoses (psoriasis and atopic dermatitis), and of spontaneous recurrent miscarriage. On the other hand, Th2 cells are implicated in the antibody response (including IgE) and in macrophageindependent inflammations (accumulation of eosinophiles and inhibition of phagocytic action). Th2 cells are also responsible for allergies and for a faster evolution of HIV infection (Shurin et al., 1999). Studies carried out on inflammatory diseases, such as psoriasis (Aractingi et al., 2001), atopic dermatitis (Khosrotehrani et al., 2001), myopathic inflammations (Wiendl et al., 2000), and cutaneous lymphomas (Urosevic et al., 2002b), as well as on the inflammatory centers of other tumors (Urosevic et al.,

HLA-G MOLECULES

225

2001), and even on HIV-positive patients (Lozano et al., 2002), show the accrued presence of HLA-G protein (soluble or membranous). The expression of this molecule is principally induced by IL10, contrary to that of class I and II antigens, whose expression is inhibited by IL10 and increased by proinflammatory interleukins such as TNF, IFN , and IL1 (Moreau et al., 1999). These observations, both ex vivo and in vitro, lead to the proposal that the HLA-G molecule probably acts as a tissue defense mechanism in cytolysis against target cells (Carosella et al., 2001). The HLA-G molecule would therefore block the reaction of T effector and APC cells, thus avoiding the destruction of healthy tissues, and would play a regulatory role in limiting inflammatory processes. VI. Role of HLA-G in Normal and Pathological Pregnancies

Human reproduction begins by the union of paternal and maternal gametes, yielding a fetus that is semiallogenic with respect to the maternal body in which it will develop. The immunogenetically different fetus should elicit immune responses by the potentially hostile maternal organism. In fact, pregnancy is a prime example of immunotolerance. While at first glance, this maternal tolerance to the semiallogenic fetus appears to violate the classical rules of immunology, when fetal cells are confronted with maternal immunity, classical immune responses are subdued or suppressed. Thorough knowledge of the structure of the fetal–maternal interface and of HLA antigen distribution on maternal and fetal cells will provide insight into understanding the establishment of this privileged-immune status, as well as of its disturbances. A. THE FETAL–MATERNAL INTERFACE A fertilized human egg, called an embryo during the first eight weeks, and then a fetus, divides in two parts. One becomes the embryo proper and the other forms the trophoblast, which develops into all the cell types found in the human placenta. The placenta is now acknowledged to be a selective filter between the mother and the developing embryo/fetus. The trophoblast differentiates from a proliferative, undifferentiated, mononuclear cytotrophoblast—the stem cells of the placenta—into three distinct, differentiated forms of trophoblast cells: villous syncytiotrophoblasts, extravillous anchoring trophoblasts, and invasive trophoblasts. The trophoblast layer of human placental villi infiltrates the uterine lining (endometrium and myometrium) and penetrates the maternal arteries; it is the extension of fetal tissue into the mother. The essential roles of trophoblast cells in the fetal–maternal exchanges implicated in successful pregnancy occur during implantation, placentation,

226

E. D. CAROSELLA ET AL.

hormone production, invasion of maternal blood vessels into the placenta, and protection of the fetus against attack by infiltrating maternal immune cells. Disturbance of any one of these functions can lead to pregnancy failure. One would expect contact by fetal cells with the maternal immune system to lead to rejection and destruction of the foreign fetus. To understand the absence of rejection, which is an atypical immune reaction, it is important to define the cell population present in the gravid uterus and the antigenic status of the placental cells (Johnson et al., 1999). Lymphomyeloid populations in the endometrium essentially consist of CD56 þ CD16 NK cells (distinct from peripheral NK cells), as well as a small number of CD8 þ T cells, CD16 þ NK cells, macrophages, and mast cells. The ubiquity of classical class I HLA-A, -B, and -C molecule expression should be restricted to non-trophoblastic somatic cells. Indeed, trophoblast tissues displays a particular HLA protein distribution, characterized by the absence of HLA class I molecules, and the presence of HLA-G (Kovats et al., 1990; McMaster et al., 1995; Hutter et al., 1996; Tarrade et al., 2001), of HLA-E (King et al., 2000; Menier et al., 2003), and to a lesser extent, of HLA-C (King et al., 1996). B. PROTECTIVE EFFECT

OF

HLA-G TOWARD MATERNAL CYTOTOXIC CELLS

During pregnancy, HLA-G expression is spatially restricted to trophoblast cells and temporally restricted to the first trimester (as well as to the third trimester at a decreased level) (Hiby et al., 1999b). A molecular construction that includes the HLA-G promoter in transgenic mouse models confirms this temporal expression, attributing it to control at its promoter level (Solier et al., 2001). Thus, the extinction of classical HLA class I antigens, the predominant expression of HLA-G in fetal trophoblast cells, and the presence of HLA-G receptors on decidual immune cells makes this gene an excellent candidate for explaining immune tolerance observed at the fetal– maternal interface. To examine this hypothesis, we conducted (for the first time) a study on the immunological role of HLA-G, carrying out in vitro cytotoxicity assays that confronted peripheral blood NK cells (obtained from 50 donors) and the K562 cell line, which does not express either classical or non-classical HLA class I molecules, transfected by HLA-G1. Lysis of the K562 cell line by NK cells was inhibited when the target K562 cell line expressed HLA-G1. The correlation between the presence of HLA-G1 on the target and inhibition of NK lysis was strengthened by the restoration of NK lysis after blocking of HLA-G1 with anti-HLA-G1 antibodies. In order to confirm these results in the physiological context of pregnancy, we carried out ex vivo experiments, using fetal and maternal tissues obtained from voluntary first-trimester terminations of normal pregnancies. Cytotoxic assays were carried out

HLA-G MOLECULES

227

under semiallogenic conditions (uterine NK cells and trophoblast cells obtained from the same mother), as well as under allogenic conditions (uterine NK cells and trophoblast cells from a different mother). The inhibition of cytolysis under both conditions demonstrated the ability of trophoblast cells to protect themselves against the lytic activity of decidual NK cells of the mother or of another pregnant woman. Taken together, these experiments demonstrate the protective role of the HLA-G molecule present on the surface of cytotrophoblast cells towards lysis carried out by uterine NK cells. HLA-G was thus shown to be responsible for the absence of maternal rejection during the time the fertilized egg implants into the uterus (RouasFreiss et al., 1997a). The natural physical barrier between fetal tissues and maternal organs is not impermeable; trophoblast cells encounter maternal immune blood cells. It is possible for maternal cells to enter the fetus, and fetal cells are able to cross the trophoblast and enter the maternal circulation (Hahn and Holzgreve, 2002). The presence of HLA-G on these fetal cells could favor their escape from maternal immunosurveillance, and to persist in maternal tissues. In this context, a study was done concerning the polymorphic eruptions of pregnancy (PEP), a cutaneous gestational pathology that occurs during the third trimester of pregnancy, when peripheral blood chimerism is very high. Interestingly, by studying samples of skin from women with PEP who were carrying male fetuses, Y-chromosome DNA was detected in dermis that also expressed HLA-G protein. In contrast, male DNA was not detected in any non-pregnant women. Moreover, expression of HLA-G by the cytotrophoblasts allows them to migrate to the maternal circulation and their infiltration into epidermal tissue (Aractingi et al., 1998). Similarly, a soluble HLA-G isoform not associated with light-chain 2-microglobulin was detected in a serum sample from pregnant women at higher level than in serum from non-pregnant women (Hunt et al., 2000). Moreover, the recent attribution of the increased HLA-G level encountered in peripheral maternal blood to fetal trophoblast cells detectable at the ninth week of gestation has permitted noninvasive prenatal diagnosis of numeric chromosomal aberrations (van Wijk et al., 2001). C. FAILURES

OF

PREGNANCY

During pregnancy, fetus and mother live in symbiosis, and the outset of this relationship is most important. Its disruption often ends in the failure of pregnancy. During this period, immunological interactions between fetal trophoblast cells and maternal decidual leukocytes at the fetal–maternal interface are primordial. Since HLA-G is the most abundant HLA protein expressed by fetal trophoblast, invading the decidua and the walls of spiral arteries, and because HLA-G is a powerful immunosuppressor (Menier et al.,

228

E. D. CAROSELLA ET AL.

2000; Riteau et al., 2001a), research aimed at explaining fetal–maternal tolerance—and even successful pregnancy—focuses principally on the HLA-G gene. An interruption or failure in reproduction can take place at many—but not random—temporal and spatial levels during pregnancy. Indeed, abnormal development may occur either within the embryo or in the extra-embryonic placenta at the beginning of pregnancy. When a fertilized egg implants outside the uterine cavity, placental formation is wrongly located and the pregnancy is said to be ectopic. A welltargeted implantation can be disrupted during placentation, which is the invasion of trophoblast tissue into a well-structured, functional placenta. These dysfunctions often end in miscarriage. The three main complications of pregnancy related to the trophoblast and placenta are hyper- and hypoinvasion of trophoblast, immunological rejection, and infection by microorganisms. 1. Normal and Ectopic Implantation The detection of HLA-G mRNA from blastocysts suggested an important role for HLA-G in embryo viability and implantation (Jurisicova et al., 1996a,b). This suggestion was confirmed by the detection of soluble HLA-G proteins deriving from blastocysts (Menicucci et al., 1999), then by the identification of this soluble form as deriving from HLA-G5 (Fuzzi et al., 2002). Furthermore, the authors of this study correlated successful implantation and detection of HLA-G5 from pre-implantation embryos after in vitro fertilization or intracytoplasmic sperm injection. In fact, pregnancy was achieved when in vitro transfers were obtained with growing embryos that secreted HLA-G5 in their supernatant, and pregnancy did not occur with preimplantation embryos that did not secrete HLA-G5. In contrast, chromosome abnormalities, such as trisomy, do not induce deficient HLA-G expression in trophoblast (Rabreau et al., 2000). Study of HLA-G expression during pregnancy is carried out with cells that are in contact with the maternal decidua, but these two tissues could not be dissociated in any intrauterine implantation, which would have made it possible to determine whether these tissues were dependent on HLA-G expression. In ectopic tubal pregnancies, which represent a decidual layer-free tissue, although disturbed trophoblast differentiation is visible, all extra-villous trophoblast cells express HLA-G (Proll et al., 2000; Emmer et al., 2002). As in the case for implantation in the uterus, HLA-G mRNA and protein were found to be expressed in ectopic pregnancies and in extra-villous trophoblast from placental accreta, which develops without decidua (Goldman-Wohl et al., 2000; Rabreau et al., 2000). These results suggest that HLA-G expression in trophoblast cells does not depend on specific

HLA-G MOLECULES

229

environmental factors present in the decidua that permit implantation in other tissues. 2. Recurrent Miscarriage Recurrent spontaneous abortion (RSA) remains unexplained in a large proportion of cases. Numerous biological causes could lead to this result, but studies of HLA-G in RSA showed it to be involved at different levels in cellular interactions, in the cytokine pattern, and in genetic predisposition (Hviid et al., 2002). HLA-G immunostaining at the fetal–maternal interface in the columns of proliferating cytotrophoblast cells of recurrent miscarriage tissues revealed— as in the normal situation—increasing staining from the embryo to the decidua, but this staining appeared to be weaker in recurrent miscarriage tissues. It has been suggested that this decreased HLA-G expression could be related to an abnormal maternal NK cell phenotype in RSA cases, notably in the appearance of a CD16 population in pathological tissues (Emmer et al., 2002). Lower HLA-G expression in RSA could influence cytokine release. Cytokine assay in co-cultures reveals that cytokine release by peripheral mononuclear cells after contact with HLA-G-expressing cells is perturbed, whether mononuclear cells come from women with recurrent abortion, compared with mononuclear cells from fertile women (Hamai et al., 1998). Allele frequency studies of HLA-G and RSA in Finnish and Japanese populations did not detect significant differences (Karhukorpi et al., 1997; Yamashita et al., 1999). Other authors obtained the same results when they analyzed data from all RSA couples without distinction, but it was possible to establish a correlation between HLA-G genotype and RSA in women who had undergone five or more RSAs, with a significant increase in the HLA-G* 01013 and 0105N allele distribution (Pfeiffer et al., 2001). The presence of either HLA-G* 0104 or HLA-G* 0105N has also been associated with unexplained recurrent miscarriage (Aldrich et al., 2001). 3. Gestational Trophoblastic Disease In contrast with pre-eclampsia, gestational trophoblastic disease represents increased and uncontrolled trophoblast invasion. This exaggerated proliferation may not have pathological consequences, but can result in placentalsite trophoblastic tumors, molar pregnancy, or even choriocarcinoma. The deeper trophoblast cells invade the decidua, the more HLA-G they express. Shallow trophoblast invasion, encountered in pre-eclampsia, is associated with reduced HLA-G expression. Molar pregnancies develop morphologically and genetically in two different forms: complete and partial. A complete hydatidform mole (CHM) appears when paternal chromosomes dominate the genome in the absence of the maternal component. In CHM, there is no

230

E. D. CAROSELLA ET AL.

fetus and the trophoblast is hyperplastic. Partial hydatidform moles (PHM) contain both abnormal trophoblastic cells, and often a fetus with severe defects and a triploid karyotype consisting of one paternal and two maternal chromosomes. In PHM, the trophoblast is generally not hyperplastic (Paradinas et al., 1996). Hyperplastic clusters of extravillous trophoblast cells from both CHM and PHM express a greater number of HLA-G molecules than a normal egg does. In PHM, a less invasive trophoblast population is associated with HLA-G expression lower than that of CHM (Rabreau et al., 2000). This reduction of HLA-G expression in PHM compared with CHM can go as far as the complete absence of HLA-G when PHM is associated with pre-eclampsia (Goldman-Wohl et al., 2001). Thus, HLA-G expression is greatly enhanced in molar pregnancies, presenting deeper infiltration of extravillous trophoblast cells in the decidua. Furthermore, the exclusively paternal genetic status of CHM confirms that the expression of HLA-G molecules—whether of paternal or maternal origin—confer trophoblast cells with the capacity to protect themselves from NK cells. Choriocarcinoma is a rare form of primary placental cancer. The JEG-3, BeWo, and JAR human choriocarcinoma cell lines are widely used as models to study trophoblast cells and HLA-G (Kovats et al., 1990; Copeman et al., 2000; Sivori et al., 2000; Easterfield et al., 2001). Only one study has reported the results of biopsies of choriocarcinoma and trophoblastic tumors. Positive HLA-G immunoreactivity was specific for intermediate trophoblast (a subpopulation of extravillous trophoblast cells) in choriocarcinoma and trophoblastic tumors, and the authors suggest using HLA-G as a marker in the differential diagnosis of gestational trophoblastic disease (Singer et al., 2002). 4. Pre-eclampsia Pre-eclampsia (PE) is a materno-placental disease in which utero-placental blood pressure increases dramatically. In pre-eclampsia, invasion by trophoblast from maternal spiral arteries is decreased or absent, resulting in inadequate perfusion of the placental bed and a reduced supply of oxygen and nutrients to the fetus (Khong et al., 1986; Genbacev et al., 1997). These clinical and anatomic observations are consequences whose causes remain to be identified and explained (Roberts and Redman, 1993). PE appears related to an enhancement of maternal immunological defenses, preventing invasion into the spiral arteries. Indeed, at the pre-eclamptic fetal–maternal interface, trophoblasts do not progress beyond the endometrium and myometrium (Taylor, 1997). Familial predisposition to PE suggests genetic control and inheritance in this pathology (Arngrimsson et al., 1990). Expression of several genes outside the HLA complex has been linked to disorders in PE, but the cause of inadequate trophoblast invasion was not elucidated. HLA-G seems to be an

HLA-G MOLECULES

231

ideal candidate gene to explain the ability of trophoblast cells to differentiate and proliferate in maternal tissue. A preliminary study observed a lower mRNA HLA-G level in the pre-eclamptic population, but the authors attribute this lower expression to a reduced number of trophoblast cells in this pathological tissue (Colbern et al., 1994). The development of specific antibodies against HLA-G protein have allowed improved experiments and to confirm impaired HLA-G expression in pre-eclamptic trophoblast (Hara et al., 1996; Lim et al., 1997). Immunohistochemical characterization of trophoblast and lymphocyte populations in the pre-eclamptic fetal–maternal interface has led to the observation of CD56 þ NK/CD8 þ T cell imbalance with enhancement of CD8 þ T cells, and led to the description of two different trophoblast types; one in the decidua, observed in normal placenta, and the other unable to migrate among NK cells (Stallmach et al., 1999). These two different trophoblast behaviors were distinguished by their HLA-G expression. Indeed, molecular analysis by RNA in situ hybridization in normal placentae correlates increased HLA-G expression with increased trophoblast invasiveness, and allow distinguishing two trophoblast populations in pre-eclamptic placentae, one consisting of HLA-G-negative trophoblasts that failed to invade maternal tissue, and another consisting of a small quantity of HLA-G-positive trophoblasts that had reached the decidua. An evaluation of mRNA HLA-G expression in PE and normal placental tissue revealed the significant overall deficit of HLA-G transcription and protein expression to be associated with a disturbed HLA-G isoform expression profile and the absence or very low expression of the spliced HLA-G3 mRNA (O’Brien et al., 2001). The cytokine environment is also changed at the pre-eclamptic fetal– maternal interface, with a pro-inflammatory Th1/anti-inflammatory Th2 cytokine imbalance observed in patients with predominant Th1-type immunity, an increase in IFN- secreting cells (Th1), and a decrease in IL10- and IL4-secreting cells (Th2) (Hennessy et al., 1999; Saito et al., 1999). A reciprocal effect is now known to exist between the release of cytokines and HLA-G expression. In fact, cytokines can raise HLA-G expression, and conversely. A co-culture of decidual lymphocytes with HLA-G-expressing cells reduced the release of cytokines, including IFN- and IL10, indicating an immune-suppressive effect of HLA-G on decidual lymphocyte activity (Rieger et al., 2002). Genotyping of HLA-G in various populations has led to a change in its initial description as a non-polymorphic gene to that of a low polymorphic one. The relationship between HLA-G genotype and susceptibility to the development of diseases of pregnancy has since been investigated in the search for possible predominance among the 15 known HLA-G alleles. Deletion or insertion of a 14-nucleotide (D/I-14) sequence in the untranslated region of exon 8 seems to play a role in its transcriptional and

232

E. D. CAROSELLA ET AL.

post-transcriptional functions. A study of persons affected by pre-eclampsia (PE patients, their husbands, and individuals born of a PE pregnancy) revealed no significant difference in the polymorphism of these 14 nucleotides in the 30 UT region (Humphrey et al., 1995), although placenta from mild preeclampsia, compared with normal placenta revealed significant differences in the silent CAC-CAT polymorphism at codon 93 (C ! T-93), in exon 3 with an excess of T-93, and in exon 8, with an excess of I-14 alleles in PE samples (O’Brien et al., 2001). Conversely, distribution of the G* 0106 allele (T-93 and I-14) screened for the presence of polymorphism in codon 258 showed no obvious association between PE and recurrent spontaneous abortion (Hviid et al., 2001). Furthermore, a recent study suggests an association between PE and the HLA-G* 0104 allele, which has the C-93 and D-14 genotype (Carreiras et al., 2002). Therefore, it is now clear that there is a relationship between PE and HLA-G polymorphism, but the identity of the implicated alleles remains to be clarified. Discrepancies observed among the various studies may be due to the fact that their authors did not take the degrees of seriousness of the PE diagnosis into account. In fact, it is possible to distinguish degrees of pre-eclampsia, from mild to severe, the latter often being associated with retarded fetal development. It seems important to not dissociate the importance of polymorphism from the level of progression of the disease. By labeling invading trophoblast cells, the HLA-G distribution pattern reflects the heterogeneity of the trophoblast at the fetal–maternal interface. Through its ability to prevent allo-recognition by inhibiting maternal NK and CTL activity, HLA-G plays a key role in successful implantation and placentation. Alteration of HLA-G expression and HLA-G allele inheritance has often been correlated with major disturbances and diseases of pregnancy. The impact of HLA-G molecule expression on the success and failure of pregnancy is now studied in relation to full-length HLA-G1 and its soluble HLA-G5 isoform. It would be interesting to introduce analysis of the other spliced isoforms, but specific antibodies against these HLA-G molecules do not exist. Finally, these results considered together can reveal the involvement of HLA-G in disease states only if pathologies are meticulously diagnosed and when patients are divided into groups, according to the disease state. VII. HLA-G in Organ Transplantation

The role of HLA-G in, and its potential use for the prevention of graft rejection are being increasingly investigated. The reasons for this interest are that while HLA-G differs from classical HLA Class I molecules in being of low polymorphism and so far described as non-allo-stimulatory, HLA-G is a ligand

HLA-G MOLECULES

233

of inhibitory receptors. What makes it even more attractive is that all cell subsets involved in graft rejection bear at least one receptor for HLA-G. HLAG therefore has the potential to be tolerogenic, by exerting its inhibitory functions on all cells responsible for graft rejection without triggering an allogeneic response. The potential involvement of HLA-G in the establishment and maintenance of graft tolerance has been investigated in vitro and in vivo, in the context of allo- and xeno-transplantation, in the human and in the animal system, and when HLA-G is expressed by various cell subsets. We summarize here the latest data on this topic. A. IN VIVO RELEVANCE

OF

HLA-G

HLA-G is a molecule that under normal circumstances is not expressed, except at the feto-maternal interface and by thymic epithelial cells. However, ectopic HLA-G expression has been demonstrated in heart-transplanted patients. Two studies by Lila et al. (2000, 2002) showed that in the absence of inflammation, HLA-G was expressed by myocardiac cells of 18% of hearttransplanted patients. It was systematically possible to detect the soluble HLA-G5 and HLA-G6 isoforms in the serum of these HLA-G-positive patients. Finally, HLA-G expression by myocardiac cells of heart-transplanted patients proved to be stable over time. Most interestingly, expression of HLAG by heart-transplanted patients might be linked to better graft acceptance. HLA-G expression significantly correlated with a reduced number of acute rejection episodes (1.2  1.1 in HLA-G positive patients vs 4.5  2.8 in HLAG-negative patients) with no chronic rejection. HLA-G-negative patients occurred and had a higher number of acute rejection episodes and chronic rejection in 27% of the cases. The mechanisms by which HLA-G neo-expression occurs in some transplanted patients are still unknown, but there are data suggesting that an allo-reaction itself might trigger it. Indeed, Lila et al. (2001) showed that in vitro, HLA-G5 was produced and secreted by alloreactive CD4 þ T cells from some mixed lymphocyte reactions. Furthermore, this alloreactive T cellderived soluble HLA-G was functionally active and capable of inhibiting CD4 þ T cells during in vitro allo-proliferation. These data were the first and remain the only linking natural HLA-G expression with long-term allograft acceptance in vivo in humans. They are strengthened by recent work in the murine system (Horuzsko et al., 2001; Liang et al., 2002), which shows that human HLA-G can bind to the murine ILT4 homolog PIR-B receptor, and that this engagement of PIR-B by human HLA-G significantly improved skin allograft survival. In the first report, Horuzsko et al. showed that HLA-G-transgenic mice exhibited reduced cellular immune responses. Indeed, allograft survival in HLA-G mice was

234

E. D. CAROSELLA ET AL.

prolonged around two-fold, compared with non-transgenic controls, even though the T cell function in HLA-G-transgenic mice was normal. The authors demonstrated that the prolonged survival of allografts was due to an HLA-G-induced maturation defect of the antigen-presenting cells of HLA-G transgenic mice, leading to impaired stimulation of allo-reactive T cells. Following up on these findings and working with non-transgenic animals, Liang et al. (i) confirmed the HLA-G-induced APC maturation defect, and (ii) demonstrated that immunization of recipient animals with HLA-G-coated latex microbeads 1 day prior to, and again 4 days after skin allo-transplantation increased graft survival by about two-fold as well. These latter data are of crucial importance because they show for the first time that (i) human HLA-G can bind to the murine homolog of a human HLA-G receptor on APCs, even though no murine HLA-G homolog that has been found to date, and (ii) in vivo immunization with human HLA-G is effective and improves allograft survival. The mechanisms by which HLA-G can alter allo-transplantation outcome have not been investigated in vivo. However, in vitro studies produced results clear enough to shed light on what might happen, indicating that HLA-G might improve graft acceptance by acting on multiple parameters and cellular actors in graft rejection. B. POTENTIAL MECHANISMS

OF

GRAFT PROTECTION

BY

HLA-G

Rejection of an allo- or a xeno-graft is a complex, multi-stepped event. If the potentially reactive cells are NK cells, they will need to reach the graft, adhere to the tissue, and finally lyse it. However, if the potentially reactive cells are T cells, an allogeneic response must be initiated through (i) graft allo-antigen uptake and presentation by infiltrating APCs; (ii) maturation of these APC; (iii) mutual stimulation of allo-specific CD4 þ , CD8 þ T cells and APCs; (iv) migration of the cytolytic cells to the graft; and finally, (v) lysis. In vitro data seem to indicate that HLA-G is capable of interfering with each of these steps, providing multi-level protection to the graft. 1. Inhibition of Adhesion and Transendothelial Migration Rolling, and firm adhesion of NK cells to their targets is a prerequisite to their cytolytic function. The influence of HLA-G on rolling adhesion and transendothelial migration of NK cells has been investigated in the context of xeno-recognition, since NK seem to play a crucial role in the rejection of xenografts. Forte et al. (2000, 2001) have shown that expression of HLA-G at the surface of porcine endothelial cells (PEC) monolayers inhibits the rolling adhesion of activated human NK cells. In these experiments, rolling adhesion of activated NK cell lines to HLA-G-transfected PEC monolayers was

HLA-G MOLECULES

235

inhibited 50%, compared with their rolling adhesion to non-transfected PEC. The effect was not completely mediated by the HLA-G receptor ILT2, since masking ILT2 on NK cells produced no, or only partial reversal of the HLA-G-mediated inhibition of rolling adhesion. This might suggest that NK cell-expressed KIR2DL4 might have been involved principally in the inhibition of NK rolling adhesion, or that a yet-unknown HLA-G receptor was. Using a similar system, Dorling et al. demonstrated HLA-G-mediated inhibition of the transendothelial migration of human NK cells (Dorling et al., 2000). Kinetics experiments showed that HLA-G delayed human NK transendothelial migration across PEC monolayers for up to half an hour. In this study, the inhibitory effect of HLA-G could be completely reversed by blocking ILT2 on NK cells. These results indicate that HLA-G might improve pig-to-human xenograft survival, by limiting human NK–pig tissue interactions. The same experiments have not been carried out in the context of allo-recognition, or for cytotoxic T lymphocytes. However, since the basic requirements for cytolysis by NK cells or CTLs with respect to migration and adhesion are similar, these results may very well hold true outside the xeno-recognition system. 2. Inhibition of CD4 þ , CD8 þ , APC Mutual Activation/Differentiation and Effector Functions The generation of an allo-response involves at least three different cell subsets: alloreactive CD4 þ T cells, alloreactive CD8 þ T cells, and antigenpresenting cells. Proper activation and functional maturation of APCs and allo-specific CD4 þ T cells is necessary for generation of allo-specific CTLs. HLA-G can act on CD4 þ T cells and APCs, leading to the inhibition of in vitro allo-responses. The inhibition of CD4 þ T cell allo-proliferation by membrane-bound HLA-G1 was investigated in two studies (Riteau et al., 1999; Bainbridge et al., 2000a) that demonstrate that HLA-G1-transfected MHC class II-positive cells inhibited the allo-proliferation of CD4 þ T cells by 80%, compared to nontransfected controls. Similar results were obtained when HLA-G1 was presented by HLA-negative, HLA-G1-transfected K562 cells used as third party cells in mixed lymphocyte reactions (Riteau et al., 1999), or when soluble HLA-G5 was present (Lila et al., 2001). It should be noted that in these three types of experiments, HLA-G could potentially bind to CD4 þ T cells (through ILT2) as well as to antigen-presenting cells (through ILT2 and/or ILT4). It is therefore difficult to determine whether HLA-G acted on CD4 þ T cells and prevented them from being activated, on APCs and prevented them from fully mature and properly stimulate CD4 þ T cells, or on both cell subsets.

236

E. D. CAROSELLA ET AL.

The question of the direct effect of HLA-G on CD4 þ T cells is still unanswered, but as stated above, recent work has shed light on the effects of HLA-G on APCs. Binding to murine PIR-B receptors, HLA-G induces a defect in murine dendritic cell functional maturation in vivo in HLA-Gtransgenic mice, as well as in vitro in non-transgenic mice (Horuzsko et al., 2001; Liang et al., 2002). This functional effect of HLA-G on APCs is further strengthened by a recent study by Chang et al. (2002), which shows that ILT4 over-expressing APCs do not induce and support allo-specific CD4 þ T cell proliferation. If data from these two groups are put together in an all-human system, it seems that by binding to ILT4 receptors on dendritic cells, HLA-G should be capable of preventing APC functional maturation and the generation of an allogeneic response. This effect might even be increased by a direct effect of HLA-G on T lymphocytes. On the other hand, the inhibition of NK function by HLA-G-transfected xenogeneic cells is now well documented. Human NK cells can lyse allogeneic, as well as xenogeneic targets. HLA-G, being a human HLA-Class I molecule of low polymorphism, can act as a ligand for the NK ILT2 and/or KIR2DL4 inhibitory receptors, and protect allo- and xenogeneic cells from NK cell lysis, as will be reviewed next. Transfection of HLA-G1 cDNA or genomic HLA-G DNA into porcine endothelial cells (PEC) or Chinese hamster ovarian cells (CHO) lead to a ‘‘partial’’ to complete inhibition (ranging from 30 to 100%) of the cytolytic function of human polyclonal NK cells and of that of most human NK cell lines (Sasaki et al., 1999a,b; Forte et al., 2000, 2001; Matsunami et al., 2001). Furthermore, two studies have shown a correlation between the level of HLA-G1 cell-surface expression by transfected PEC and the extent of NK functional inhibition (Forte et al., 2001; Matsunami et al., 2001). In conclusion, all these experiments demonstrate that membrane-bound and soluble HLA-G are expressed in the context of allo-transplantation and are capable of inhibiting NK cytolytic function, as well as CD4 þ and CD8 þ T cell allo-responses (Fig. 4). It seems that HLA-G may participate in the down-modulation of a rejection episode by acting at all levels of the allo-response. It is not known which effect of HLA-G is functionally the most important, but the fact that in vivo immunization with HLA-G induces prolonged allograft survival in mice is certainly highly promising.

VIII. HLA-G in Malignancies

Over the past few years, HLA-G has been proposed as playing a potential role in the molecular mechanisms underlying immune escape strategies

HLA-G MOLECULES

237

FIG. 4. Potential inhibitory effect of HLA-G on allogeneic response. Center: Schematic representation of the sequence of events involved in transplant rejection. In the course of a graft rejection reaction, antigens from the transplant (Ag) are captured and processed for presentation by immature APCs in the context of HLA class I and class II molecules. APCs mature and interaction with antigen-specific CD4 þ and CD8 þ T cells leads to the activation of all three cell subsets and to the differentiation of CD4 þ T cells and CD8 þ T cells into CD4 þ Th and CD8 þ CTLs effectors respectively. CTLs then migrate to the graft and may lyse graft cells. NK cells migrate to the graft and may lyse graft cells, based on the compatibility between graft HLA Class I antigens and NK cells inhibitory receptors. Sides: Summary of HLA-G functions in the context of transplantation. G represents HLA-G. A: Human HLA-G interacts with murine APCs and inhibit their maturation in vitro and in vivo. B: Allo-specific CD4 þ T cells express and secrete soluble HLA-G5 in vitro. C: HLA-G inhibits CD4 þ T cells proliferation in vitro and may inhibit concomitant maturation. D: HLA-G inhibits the generation of allo-specific CTLs in vitro. E: HLAG inhibits rolling adhesion and transendothelial migration of NK cells in the xenogeneic context. F: HLA-G inhibits the cytolytic function of NK cells in the allo- and in the xenogeneic context, and inhibits antigen-specific CTL cytolytic function.

utilized by tumor cells. Although divergent information has been provided about HLA-G expression by some types of tumors, many investigations have shown HLA-G mRNA and protein expression to be associated with the malignant transformation of cells. We discuss below the possible reasons for such conflicting results and present studies emphasizing that HLA-G, the engagement of which generates inhibitory signals in various immune cells, has functional significance in the host’s immunosurveillance of tumors.

238

E. D. CAROSELLA ET AL.

A. EXPRESSION OF HLA-G CANCER PATIENTS

IN

TUMORS OBTAINED FROM

By carrying out immunohistochemical analysis of tumor biopsies obtained from cancer patients, upregulation of HLA-G protein expression has been found in various types of malignant lesions, such as melanoma (Paul et al., 1998, 1999; Carosella et al., 2000; Wagner et al., 2000), cutaneous lymphoma (Urosevic et al., 2002b), non-Hodgkin lymphomas (Dre´nou et al., 2002), glioblastoma (Wiendl et al., 2002), breast cancer (Lefebvre et al., 2002), lung carcinoma (Pangault et al., 1999; Urosevic et al., 2001; Pangault et al., 2002), colorectal carcinoma (Fukushima et al., 1998), renal cell carcinoma (Ibrahim et al., 2001a), bladder carcinoma (Carosella et al., 2000), and hydatidform moles (Rabreau et al., 2000; Goldman-Wohl et al., 2001). According to type of tumor, the expression of HLA-G was found in tumor cells or/and in infiltrating cells (predominantly monocytes/macrophages, but also lymphocytes), but not in the corresponding healthy tissue. Studies of interest describing ex vivo expression of HLA-G at the tumor site are reviewed below. Among them, one deserves particular attention: Sequential biopsies of healthy skin, primary cutaneous tumor, lymph node metastasis, and a tumor regression site within the skin primary tumor were all obtained from one patient. HLA-G protein was detected in both primary and metastatic tumor sites, but not in either healthy skin that had been resected from the vicinity of the tumor or at a tumor regression site in which an efficient immune response may develop, in the absence of HLA-G (Paul et al., 1999). Another study described a correlation between poor prognosis of melanoma treated with interferon alpha (IFN-)-2b and the presence of HLA-G in melanoma cells before the treatment (Wagner et al., 2000). This may be important for selection of cancer patients likely to benefit from IFN- therapy. Recently, various types of malignant epithelial tumors and benign cutaneous lesions arising in kidney transplant recipients have been investigated. The results revealed HLA-G expression in malignant and pre-malignant cutaneous diseases (i.e., 35% of squamous cell carcinomas, 47% of Bowen disease, 27% of actinic keratoses, 14% of basal-cell carcinomas) but not in the 24 benign lesions studied (Aractingi et al., 2003). Interestingly, when both benign and malignant cutaneous lesions developed at distinct sites in the same transplanted patient, upregulation of HLA-G expression was only found at the malignant site. This study confirms the link between ectopic HLA-G expression and tumor development in patients with cancer. The HLA-G gene was found to be transcribed in 100% of 45 cases of primary cutaneous lymphoma and the HLA-G protein was detected in 70% and 45% of cutaneous B-cell and T-cell lymphomas, respectively (Urosevic et al., 2002b). In T-cell cutaneous lymphomas, HLA-G protein expression

HLA-G MOLECULES

239

was associated with high-grade histology and an advanced stage of the disease. Two populations of cells expressing HLA-G were identified: lymphoid cells (i.e., neoplastic and infiltrating T and B lymphocytes) and cells of myeloid origin (i.e., intra-lesional macrophages and dendritic cells). Interestingly, the presence of the immunosuppressive IL-10 cytokine, which is known to be secreted by cutaneous lymphomas and to induce HLA-G expression, could be correlated with HLA-G protein expression in these cutaneous tumors. Analysis of tumor and adjacent normal renal tissue samples showed tumorspecific HLA-G expression in 61% of renal cell carcinomas (Ibrahim et al., 2001a). HLA-G has been shown to be expressed in 26–33% of lung carcinoma lesions, either in tumor cells and/or intratumoral-infiltrating cells, such as activated macrophages and dendritic cells (Pangault et al., 1999; Urosevic et al., 2001; Pangault et al., 2002). HLA-G protein was recently detected in brain tumors, such as glioblastomas and anaplastic oligoastrocytoma (Wiendl et al., 2002), as well as in breast cancer (38% of cases) (Lefebvre et al., 2002). In the latter study, both HLA-G and its inhibitory receptor, ILT2, were found to be expressed in tumor tissue samples. Interestingly, in vivo expression of the ILT2 inhibitory receptor has already been reported on tumor-specific CTLs in melanoma (Ikeda et al., 1997; Bakker et al., 1998; Speiser et al., 1999), and renal cell carcinoma (Guerra et al., 2000; Gati et al., 2001) leading to the downmodulation of the lytic activity of such CTLs. In conclusion, the number of lesions analyzed in the various types of malignancies ranges from about 10 lesions in basal cell carcinoma to more than 100 lesions in melanoma. The percentage of positive lesions varies markedly among the various types of malignancies. It is noteworthy that, although HLA-G is activated at the transcriptional level in most of the tumor specimens analyzed, no systematic correlation with protein expression is found. This may be explained by several mechanisms, including mutations in the HLA-G transcript, blocking protein translation and strong posttranscriptional regulatory mechanisms. Moreover, not all HLA-G isoforms are usually detected at the transcriptional level in tumors, as is the case in trophoblasts. An illustrative example is the differential transcription of HLA-G1 and HLA-G5 in melanoma biopsies. This may reflect an adaptive tumor mechanism that enables selective expression of membrane-bound or soluble HLA-G isoforms (Paul et al., 1999). The generation of soluble HLA-G molecules has also been associated to certain HLA-G alleles, implying that the levels of soluble HLA-G secretion are under genetic control (Rebmann et al., 2001). However, divergent information has been provided concerning the expression of HLA-G protein in malignant tumors. Indeed several authors

240

E. D. CAROSELLA ET AL.

could not detect HLA-G expression in a variety of solid tumors (Pangault et al., 1999; Real et al., 1999; Frumento et al., 2000; Davies et al., 2001; Hurks et al., 2001; Palmisano et al., 2002). These contrasting results may be explained by the fact that most of these studies were conducted on frozen tissue sections, allowing less discrimination of cell morphology, obtained from a small number of biopsies from a tumor-type analyzed, and using mAbs that react only with the HLA-G1, and -G5 isoforms. Sharing of reagents as well as standardization of techniques will avoid the generation of conflicting data in future literature. Studies along this line will benefit from comparative analysis of the reactivity of anti-HLA-G antibodies in frozen versus paraffin-embedded tissues and from the use of anti-HLA-G (Fab0 )2 fragments in immunohistochemical experiments. In this regard, the second international workshop on HLA-G, which will be held in Paris in 2003, will allow pursuing our efforts in the standardization of tools and protocols. B. DETECTION PATIENTS

OF

SOLUBLE HLA-G

IN

SERUM OBTAINED

FROM

CANCER

An increase in the serum HLA-G level has been described in patients with melanoma that was enhanced upon treatment with IFN-. The serum HLA-G level was found to be correlated with advanced disease stage and tumor load. However, it was not associated with recurrence-free or overall survival. Interestingly, cell-surface expression of HLA-G1 was detected on peripheral blood monocytes of two melanoma patients with an elevated soluble HLA-G serum level. Such expression was upregulated after systemic IFN- immunotherapy. Monocytes may thus constitute a possible source of elevated serum soluble HLA-G levels in these patients (Ugurel et al., 2001). Such serum HLA-G antigens, which may be derived from the release of membrane-bound HLA-G1 and/or from the secretion of soluble HLA-G5 may affect anti-tumor immune both locally at the tumor site, as well as systemically by distribution via the circulation (see above-described functional properties of HLA-G). C. HLA-G EXPRESSION

BY

TUMOR CELL-LINES

With the exception of choriocarcinoma cell-lines (i.e., JEG-3 and BeWo), only a few tumor cell-lines have been found to express HLA-G on their cell membranes (Pangault et al., 1999; Real et al., 1999; Polakova and Russ, 2000; Frumento et al., 2000; Davies et al., 2001; Hurks et al., 2001). The discrepancy between the data obtained with tumor cell-lines and surgically removed lesions may reflect the lack of factors in tissue culture media that

HLA-G MOLECULES

241

induce HLA-G expression in the tumor microenvironment. In fact, tumor cells may definitively lose their initial HLA-G expression after long-term in vitro cell culture. Factors that participate in ectopic activation of HLA-G gene transcription and protein expression in tumor cells remain to be defined and may implicate several cytokines, including IL-10, IFN-, , and . IL-10 is often secreted by malignant cells, and according to its immunosuppressive properties, also constitutes a way for tumor cells to evade tumor antigen-specific immune responses (Botti et al., 1998). More relevant is the demonstration that IFN- treatment induces or enhances HLA-G expression in several tumor cell lines, such as the U937 leukemia monohistiocytic cell-line (Yang et al., 1996; Amiot et al., 1998), the T98G glioblastoma cellline (Maier et al., 1999; Viendl et al., 2002), the JEG-3 (Yang et al., 1995), and BeWo (Hamai et al., 1999b) choriocarcinoma cell lines. It should be noted that the U937 monohistiocytic cell-line expresses HLA-G1 when incubated with IFN- , IFN- plus IL-2, or IFN- plus GM-CSF (Amiot et al., 1998). However, constitutive expression of HLA-G has been described in few tumor cell-lines. First, melanoma cell-lines have been described to express HLA-G isoforms (Paul et al., 1998; Adrian Cabestre et al., 1999b). Also, HLA-G1 cell-surface expression and HLA-G5 secretion were maintained in a tumor cell line established from an HLA-G-positive renal cell carcinoma lesion (Ibrahim et al., 2001a). Of note, IFN-(, , and ) treatment of this renal tumor cell line enhances HLA-G1 cell-surface expression (Ibrahim et al., 2001a). Recently, several glioma cell-lines were found to highly transcribe the HLA-G gene and to express the HLA-G1 protein on their cell-surfaces (Wiendl et al., 2002). Interestingly, HLA-G1 cell-surface expression could be induced on initially HLA-G protein-negative glioma cell lines upon IFN- treatment (Wiendl et al., 2002). Short-term tumor cell-lines from patients with advanced ovarian carcinoma were described as expressing an HLA-G protein whose level is enhanced after IFN- treatment (Malmberg et al., 2002). Analysis of MHC abnormalities in non-Hodgkin lymphoma showed that three out of 50 lymphomas tested expressed surface HLA-G1. These HLA-G1positive cases correspond to HLA class I-defective lymphomas (Dre´nou et al., 2002). In the same way, HLA-G1 cell-surface expression was analyzed in 40 leukemia samples of various subtypes (i.e., acute lymphoblastic, acute myeloblastic, chronic myeloblastic, and chronic lymphocytic leukemia). Although none of the leukemia cells spontaneously expressed surface HLA-G1, 21% of them became HLA-G1-positive upon IFN- stimulation (Mizuno et al., 2000). Taken together, these results show that, although it is a rare event, HLA-G may be found in tumor cell-lines, allowing investigation of its functional role in human cancer.

242

E. D. CAROSELLA ET AL.

D. FUNCTIONAL SIGNIFICANCE

OF

HLA-G EXPRESSION

IN

TUMORS

The immunosuppressive properties of HLA-G, as described above, has led to the hypothesis that malignant cells could be protected by HLA-G expression from anti-tumor immune responses. In vitro studies with melanoma cell-lines that express various HLA-G isoforms were first found to be resistant to lysis from the NK cell-line YT2C2-PR (Paul et al., 1998; Adrian Cabestre et al., 1999b). This NK cell-line does not express known inhibitory receptors that interact with HLA-E and/or classical HLA class I molecules, but bears the KIR2DL4 receptor, which specifically binds HLA-G (Rouas-Freiss et al., 1999; Riteau et al., 2001a). In addition, several glioma cell-lines expressing the full-length HLA-G1 isoform on their cell-surfaces, either constitutively or after induction by IFN- , were found to be protected from alloreactive cytolysis (Wiendl et al., 2002). This protective effect could be reversed by blocking HLA-G1 with an anti-HLA-G1 mAb. Finally, these results were confirmed by utilizing HLA-G-negative tumor cell-lines, such as melanoma (Riteau et al., 2001c) or glioma (Wiendl et al., 2002) cell-lines in which HLA-G1, -G2, -G3, -G4, or -G5 cDNA has been transfected. In both cases, such gene transfer rendered the tumor cells highly resistant to lysis by NK- and Ag-specific CTL cells (Riteau et al., 2001c), inhibited the alloproliferative response, and prevented efficient priming of cytotoxic T cells (Wiendl et al., 2002). Of particular interest, analysis of the balance between the activating signal delivered by the stress-inducible molecule MICA and the inhibitory signal generated by HLA-G1 on NK cell-mediated lysis of a melanoma cell-line expressing both molecules, showed that HLA-G1 counteracts the triggering signal of MICA (Menier et al., 2002). This finding suggests that in vivo, overexpression of inhibitory ligands, such as HLA-G, by tumor cells may bypass activating signal(s), such as that mediated by MICA, thereby favoring tumor progression. Interestingly, the tumor tissue-distribution of MICA is common with that of HLA-G (Groh et al., 1999; Pende et al., 2001). Another tumor situation in which HLA-G expression would have a biological relevance is 2-microglobulin-deficient tumors. Indeed, while HLA-A, -B, -C, -E, -G1, and -G5 molecules will not be properly folded and expressed on the cell-surface of 2-microglobulin-deficient tumors, expression of the truncated HLA-G isoforms might still occur. In such cases, these isoforms would facilitate escape from NK cell immunosurveillance of these otherwise NK-susceptible tumor targets. Finally, cytokines, such as IFN and IL-10, which have been described to upregulate HLA-G1 cell-surface expression, are also detected in tumor sites (Knoefel et al., 1997; Asadullah et al., 1998; Botti et al., 1998) and found to correlate with HLA-G protein expression by malignant cells (Urosevic et al.,

HLA-G MOLECULES

243

2001, 2002b). It should be stressed that some of these cytokines are used for immunotherapy of malignant diseases (Agarwala and Kirkwood, 1996) and may contribute to the induction of HLA-G expression at the tumor site, thus conferring immune-protection to tumor cells and facilitating their spreading in the host (Wagner et al., 2000; Ugurel et al., 2001). REFERENCES Adrian Cabestre, F., Lefebvre, S., Moreau, P., Rouas-Freiss, N., Dausset, J., Carosella, E. D., and Paul, P. (1999a). Sem. Cancer Biol. 9, 27. Adrian Cabestre, F., Moreau, P., Riteau, B., Ibrahim, E. C., Le Danff, C., Dausset, J., RouasFreiss, N., Carosella, E. D., and Paul, P. (1999b). J. Reprod. Immunol. 43, 183. Agarwala, S. S., and Kirkwood, J. M. (1996). Curr. Opin. Oncol. 8, 167. Aikhionbare, F. O., Hodge, T., Kuhn, L., Bulterys, M., Abrams, E. J., and Bond, V. C. (2001). Aids 15, 2196. Aldrich, C., Wambebe, C., Odama, L., Di Rienzo, A., and Ober, C. (2002). Hum. Immunol. 63, 405. Aldrich, C. L., Stephenson, M. D., Karrison, T., Odem, R. R., Branch, D. W., Scott, J. R., Schreiber, J. R., and Ober, C. (2001). Mol. Hum. Reprod. 7, 1167. Allan, D. S. J., Colonna, M., Lanier, L., Churakova, T., Abrams, J. S., Ellis, S. A., McMichael, A. J., and Braud, V. M. (1999). J. Exp. Med. 189, 1149. Alvarez, M., Santos, P., Martinho, A., Simoes, O., Abade, A., and Breda-Coimbra, H. (1999). Transplant Proc. 31, 1829. Amiot, L., Onno, M., Drenou, B., Monvoisin, C., and Fauchet, R. (1998). Hum. Immunol. 59, 524. Amiot, L., Onno, M., Renard, I., Drenou, B., Guillaudeux, T., Le Bouteiller, P., and Fauchet, R. (1996). Tissue Antigens 48, 609. Aractingi, S., Berkane, N., Bertheau, P., Le Goue´, C., Dausset, J., Uzan, S., and Carosella, E. D. (1998). Lancet 352, 1898. Aractingi, S., Briand, N., Le Danff, C., Viguier, M., Bachelez, H., Michel, L., Dubertret, L., and Carosella, E. D. (2001). Am. J. Pathol. 159, 71. Aractingi, S., Kanitakis, J., Euvrard, S., Le Danff, C., and Carosella, E. D. (2003). Int. J. Cancer 106, 232. Arnaiz-Villena, A., Martinez-Laso, J., Alvarez, M., Castro, M. J., Varela, E., Gomez-Casado, E., Suarez, B., Recio, M. J., Vargas-Alarcon, G., and Morales, P. (1997). Immunogenetics 46, 251. Arnaiz-Villena, A., Martinez-Laso, J., Castro, M. J., Morales, P., Moscoso, J., Varela, P., GomezCasado, E., and Allende, L. M. (2001). Immunol. Rev. 183, 65. Arnaiz-Villena, A., Morales, P., Gomez-Casado, E., Castro, M. J., Varela, P., Rojo-Amigo, R., and Martinez-Laso, J. (1999). J. Reprod. Immunol. 43, 111. Arngrimsson, R., Bjornsson, S., Geirsson, R. T., Bjornsson, H., Walker, J. J., and Snaedal, G. (1990). Br. J. Obstet. Gynaecol. 97, 762. Asadullah, K., Docke, W. D., Volk, H. D., and Sterry, W. (1998). Exp. Dermatol. 7, 314. Bainbridge, D. R., Ellis, S. A., and Sargent, I. L. (2000a). J. Reprod. Immunol. 48, 17. Bainbridge, D. R., Ellis, S. A., and Sargent, I. L. (2000b). J. Reprod. Immunol. 47, 1. Bainbridge, D. R. J., Ellis, S. A., and Sargent, I. L. (2000c). J. Reprod. Immunol. 47, 1. Bakker, A., Phillips, J., Figdor, C., and Lanier, L. (1998). J. Immunol. 160, 5239. Bamberger, A. M., Jenatschke, S., Schulte, H. M., Loning, T., and Bamberger, M. C. (2000). J. Clin. Endocrinol. Metab. 85, 3932. Bennett, W. A., Lagoo-Deenadayalan, S., Whitworth, N. S., Brackin, M. N., Hale, E., and Cowan, B. D. (1997). Early Pregnancy 3, 190.

244

E. D. CAROSELLA ET AL.

Bird, A. (2002). Genes Dev. 16, 6. Blaschitz, A., Lenfant, F., Mallet, V., Hartmann, M., Bensussan, A., Geraghty, D. E., Le Bouteiller, P., and Dohr, G. (1997). Eur. J. Immunol. 27, 3380. Botti, C., Seregni, E., Ferrari, L., Martinetti, A., and Bombardieri, E. (1998). Int. J. Biol. Markers, 13, 51. Boyson, J. E., Erskine, R., Whitman, M. C., Chiu, M., Lau, J. M., Koopman, L. A., Valter, M. M., Angelisova, P., Horejsi, V., and Strominger, J. L. (2002). Proc. Natl. Acad. Sci. USA 99, 16180–16185. Braud, V. M., Allan, D. S., O’Callaghan, C. A., Soderstrom, K., D’Andrea, A., Ogg, G. S., Lazetic, S., Young, N. T., Bell, J. I., Phillips, J. H., Lanier, L. L., and McMichael, A. J. (1998). Nature 391(6669), 795. Cabestre, F. A., Lefebvre, S., Moreau, P., Rouas-Friess, N., Dausset, J., Carosella, E. D., and Paul, P. (1999). Semin. Cancer Biol. 9, 27. Carosella, E. D., Moreau, P., Aractingi, S., and Rouas-Freiss, N. (2001). Trends Immunol. 22, 553. Carosella, E. D., Paul, P., Moreau, P., and Rouas-Freiss, N. (2000). Immunol. Today 21, 532. Carreiras, M., Montagnani, S., and Layrisse, Z. (2002). Am. J. Reprod. Immunol. 48, 176. Castro, M. J., Morales, P., Martinez-Laso, J., Allende, L., Rojo-Amigo, R., Gonzalez-Hevilla, M., Varela, P., Moreno, A., Garcia-Berciano, M., and Arnaiz-Villena, A. (2000a). Hum. Immunol. 61, 1157. Castro, M. J., Morales, P., Martinez-Laso, J., Allende, L., Rojo-Amigo, R., Gonzalez-Hevilla, M., Varela, P., Moscoso, J., Garcia-Berciano, M., and Arnaiz-Villena, A. (2000b). Hum. Immunol. 61, 1164. Castro, M. J., Morales, P., Rojo-Amigo, R., Martnez-Laso, J., Allende, L., Varela, P., GarciaBerciano, M., Guillen-Perales, J., and Arnaiz-Villena, A. (2000c). Tissue Antigens 56, 232. Chang, C. C., Ciubotariu, R., Manavalan, J. S., Yuan, J., Colovai, A. I., Piazza, F., Lederman, S., Colonna, M., Cortesini, R., Dalla-Favera, R., and Suciu-Foca, N. (2002). Nat. Immunol. 3, 237. Chu, W., Fant, M. E., Geraghty, D. E., and Hunt, J. S. (1998). Hum. Immunol. 59, 435. Chu, W., Gao, J., Murphy, W. J., and Hunt, J. S. (1999a). Hum. Immunol. 60, 1113. Chu, W., Yang, Y., Geraghty, D. E., and Hunt, J. S. (1999b). J. Reprod. Immunol. 42, 1. Colbern, G. T., Chiang, M. H., and Main, E. K. (1994). Am. J. Obstet. Gynecol. 170, 1244. Colonna, M., Navarro, F., Bellon, T., Llano, M., Samaridis, J., Angman, L., Cella, M., and LopezBotet, M. (1997). J. Exp. Med. 186, 1809. Colonna, M., Samaridis, J., Cella, M., Angman, L., Allen, R. L., O’Callaghan, C., Dunbar, R., Ogg, G. S., Cerundolo, V., and Rolink, A. (1998). J. Immunol. 160, 3096. Contini, P., Ghio, M., Poggi, A., Filaci, G., Indiveri, F., Ferrone, S., and Puppo, F. (2003). Eur. J. Immunol. 33, 125. Copeman, J., Han, R. N., Caniggia, I., McMaster, M., Fisher, S. J., and Cross, J. C. (2000). Biol. Reprod. 62, 1543. Corell, A., Morales, P., Martinez-Laso, J., Martin-Villa, J., Varela, P., Paz-Artal, E., Allende, L. M., Rodriguez, C., and Arnaiz-Villena, A. (1994). Hum. Immunol. 41, 52. Cosman, D., Fanger, N., Borges, L., Kubin, M., Chin, W., Peterson, L., and Hsu, M.-L. (1997). Immunity 7, 273. Crisa, L., McMaster, M. T., Ishii, J. K., Fisher, S. J., and Salomon, D. R. (1997). J. Exp. Med. 186(2), 289. Davies, B., Hiby, S., Gardner, L., Loke, Y. W., and King, A. (2001). Am. J. Reprod. Immunol. 45, 103. Davis, D. M., Reyburn, H. T., Pazmany, L., Chiu, I., Mandelboim, O., and Strominger, J. L. (1997). Eur. J. Immunol. 27, 2714. Diehl, M., Munz, C., Keilholz, W., Stevanovic, S., Holmes, N., Loke, Y. W., and Rammensee, H. G. (1996). Curr. Biol. 6, 305.

HLA-G MOLECULES

245

Dinarello, C. A. (2000). Chest 118, 503. Dorling, A., Monk, N., and Lechler, R. (2000). Transplant Proc. 32, 938. Dre´nou, B., Le Friec, G., Bernard, M., Pangault, C., Grosset, J. M., Lamy, T., Fauchet, R., and Amiot, L. (2002). Br. J. Haematol. 119, 417. Dummer, W., Becker, J. C., Schwaaf, A., Leverkus, M., Moll, T., and Brocker, E. B. (1995). Melanoma. Res. 5, 67. Easterfield, A. J., Austen, B. M., and Westwood, O. M. (2001). J. Reprod. Immunol. 50, 19. Ellis, S. A., Sargent, I. L., Redman, C. W. G., and McMichael, A. J. (1986). Immunology 59, 595. El-Osta, A., and Wolffe, A. P. (2000). Gene Expr. 9, 63. Emmer, P. M., Steegers, E. A., Kerstens, H. M., Bulten, J., Nelen, W. L., Boer, K., and Joosten, I. (2002). Hum. Reprod. 17, 1072. Forte, P., Matter-Reissmann, U. B., Strasser, M., Schneider, M. K., and Seebach, J. D. (2000). Hum. Immunol. 61, 1066. Forte, P., Pazmany, L., Matter-Reissmann, U. B., Stussi, G., Schneider, M. K., and Seebach, J. D. (2001). J. Immunol. 167, 6002. Fournel, S., Aguerre-Girr, M., Campan, A., Salauze, L., Berrebi, A., Lone, Y. C., Lenfant, F., and Le Bouteiller, P. (1999). Am. J. Reprod. Immunol. 42, 22. Fournel, S., Aguerre-Girr, M., Huc, X., Lenfant, F., Alam, A., Toubert, A., Bensussan, A., and Le Bouteiller, P. (2000a). J. Immunol. 164, 6100. Fournel, S., Huc, X., Aguerre-Girr, M., Solier, C., Legros, M., Praud-Brethenou, C., Moussa, M., Chaouat, G., Berrebi, A., Bensussan, A., Lenfant, F., and Le Bouteiller, P. (2000b). Tissue Antigens 55, 510. Frumento, G., Franchello, S., Geraghty, E., and Ferrara, G. B. (1999). Transplant Proc. 31, 1847. Frumento, G., Franchello, S., Palmisano, G. L., Nicotra, M. R., Giacomini, P., Loke, Y. W., Geraghty, D. E., Maio, M., Manzo, C., Natali, P. G., and Ferrara, G. B. (2000). Tissue Antigens 56, 30. Fujii, T., Ishitani, A., and Geraghty, D. E. (1994). J. Immunol. 153, 5516. Fukushima, Y., Oshika, Y., Nakamura, M., Tokunaga, T., Hatanaka, H., Abe, Y., Yamazaki, H., Kijima, H., Ueyama, Y., and Tamaoki, N. (1998). Int. J. Mol. Med. 2, 349. Fuzzi, B., Rizzo, R., Criscuoli, L., Noci, I., Melchiorri, L., Scarselli, B., Bencini, E., Menicucci, A., and Baricordi, O. R. (2002). Eur. J. Immunol. 32, 311. Gao, G. F., Willcox, B. E., Wyer, J. R., Boulter, J. M., O’Callaghan, C. A., Maenaka, K., Stuart, D. I., Jones, E. Y., Van Der Merwe, P. A., Bell, J. I., and Jakobsen, B. K. (2000). J. Biol. Chem. 275, 15232. Gati, A., Guerra, N., Giron-Michel, J., Azzarone, B., Angevin, E., Moretta, A., Chouaib, S., and Caignard, A. (2001). Cancer Res. 61, 3240. Genbacev, O., Zhou, Y., Ludlow, J. W., and Fisher, S. J. (1997). Science 277, 1669. Geraghty, D. E., Koller, B. H., and Orr, H. T. (1987). Proc. Natl. Acad. Sci. USA 84, 9145. Gobin, S. J., Biesta, P., De Steenwinkel, J. E., Datema, G., and Van Den Elsen, P. J. (2002). J. Biol. Chem. 277, 39525. Gobin, S. J., Keijsers, V., Cheong, C., van Zutphen, M., and Van den Elsen, P. J. (1999a). Transplant Proc. 31, 1857. Gobin, S. J., Keijsers, V., van Zutphen, M., and van den Elsen, P. J. (1998a). J. Immunol. 161, 2276. Gobin, S. J., Peijnenburg, A., Keijsers, V., and van den Elsen, P. J. (1997a). Immunity 6, 601. Gobin, S. J., Peijnenburg, A., van Eggermond, M., van Zutphen, M., van den Berg, R., and van den Elsen, P. J. (1998b). Immunity 9, 531. Gobin, S. J., and van den Elsen, P. J. (2000). Hum. Immunol. 61, 1102. Gobin, S. J., van Zutphen, M., Westerheide, S. D., Boss, J. M., and van den Elsen, P. J. (2001). J. Immunol. 167, 5175.

246

E. D. CAROSELLA ET AL.

Gobin, S. J., van Zutphen, M., Woltman, A. M., and van den Elsen, P. J. (1999b). J. Immunol. 163, 1428. Gobin, S. J., Wilson, L., Keijsers, V., and Van den Elsen, P. J. (1997b). J. Immunol. 158, 3587. Goldman, M., and Stordeur, P. (1997). Eur. Cytokine Netw. 8, 301. Goldman-Wohl, D., Ariel, I., Greenfield, C., Hochner-Celnikier, D., Lavy, Y., and Yagel, S. (2001). Am. J. Obstet. Gynecol. 185, 476. Goldman-Wohl, D. S., Ariel, I., Greenfield, C., Hanoch, J., and Yagel, S. (2000). Mol. Hum. Reprod. 6, 535. Groh, V., Rhinehart, R., Secrist, H., Bauer, S., Grabstein, K. H., and Spies, T. (1999). Proc. Natl. Acad. Sci. USA 96, 6879. Guerra, N., Guillard, M., Angevin, E., Echchakir, H., Escudier, B., Moretta, A., Chouaib, S., and Caignard, A. (2000). Blood 95, 2883. Hahn, S., and Holzgreve, W. (2002). Clin. Obstet. Gynecol. 45, 649. Hamai, Y., Fujii, T., Miki, A., Geraghty, D. E., Harada, I., Takai, Y., Kozuma, S., Tsutsumi, O., and Taketani, Y. (1999a). Am. J. Reprod. Immunol. 41, 293. Hamai, Y., Fujii, T., Yamashita, T., Miki, A., Hyodo, H., Kozuma, S., Geraghty, D. E., and Taketani, Y. (1999b). Am. J. Reprod. Immunol. 41, 153. Hamai, Y., Fujii, T., Yamashita, T., Miki, A., Kozuma, S., Geraghty, D. E., and Taketani, Y. (1998). Am. J. Reprod. Immunol. 40, 408. Hammer, A., Hutter, H., Blaschitz, A., Mahnert, W., Hartmann, M., Uchanska-Ziegler, B., Ziegler, A., and Dohr, G. (1997a). Am. J. Reprod. Immunol. 37, 161. Hammer, A., Hutter, H., and Dohr, G. (1997b). Am. J. Reprod. Immunol. 38, 150. Hara, N., Fujii, T., Yamashita, T., Kozuma, S., Okai, T., and Taketani, Y. (1996). AJRI 36, 349. Harrison, G. A., Humphrey, K. E., Jakobsen, I. B., and Cooper, D. W. (1993). Hum. Mol. Genet. 2, 2200. Hennessy, A., Pilmore, H. L., Simmons, L. A., and Painter, D. M. (1999). J. Immunol. 163, 3491. Hiby, S., King, A., Sharkey, A., and Loke, Y. W. (1999a). Tissue Antigens 53, 1. Hiby, S. E., King, A., Sharkey, A., and Loke, Y. W. (1999b). Tissue Antigens 53, 1. Horuzsko, A., Lenfant, F., Munn, D., and Mellor, A. (2001). Int. Immunol. 13, 385. Horuzsko, A., Portik-Dobos, V., Hansen, K. A., Markowitz, R. B., Helman, S. W., and Mellor, A. L. (1999). Transplant Proc. 31, 1860. Houlihan, J. M., Biro, P. A., Harper, H. M., Jenkinson, H. J., and Holmes, C. H. (1995). J. Immunol. 154, 5665. Humphrey, K. E., Harrison, G. A., Cooper, D. W., Wilton, A. N., Brennecke, S. P., and Trudinger, B. J. (1995). Br. J. Obstet. Gynaecol. 102, 707. Hunt, J. S., Jadhav, L., Chu, W., Geraghty, D. E., and Ober, C. (2000). Am. J. Obstet. Gynecol. 183, 682. Hurks, H. M., Valter, M. M., Wilson, L., Hilgert, I., van den Elsen, P. J., and Jager, M. J. (2001). Invest. Ophthalmol. Vis. Sci. 42, 3081. Hutter, H., Hammer, A., Blaschitz, A., Hartmann, M., Ebbesen, P., Dohr, G., Ziegler, A., and Uchanska-Ziegler, B. (1996). Cell Tissue Res. 286, 439. Hviid, T. V., Christiansen, O. B., Johansen, J. K., Hviid, U. R., Lundegaard, C., Moller, C., and Morling, N. (2001). Immunogenetics 53, 48. Hviid, T. V., Hylenius, S., Hoegh, A. M., Kruse, C., and Christiansen, O. B. (2002). Tissue Antigens 60, 122. Hviid, T. V., Meldgaard, M., Sorensen, S., and Morling, N. (1997). J. Reprod. Immunol. 35, 31. Hviid, T. V., Sorensen, S., and Morling, N. (1999). Hum. Immunol. 60, 1237. Ibrahim, E., Guerra, N., Lacombe, M., Angevin, E., Chouaib, S., Carosella, E., Caignard, A., and Paul, P. (2001a). Cancer Res. 61, 6838.

HLA-G MOLECULES

247

Ibrahim, E. C., Guerra, N., Lacombe, M. J., Angevin, E., Chouaib, S., Carosella, E. D., Caignard, A., and Paul, P. (2001b). Cancer Res. 61, 6838. Ibrahim, E. C., Morange, M., Dausset, J., Carosella, E. D., and Paul, P. (2000). Cell Stress Chaperones 5, 207. Ikeda, H., Lethe´, B., Lehmann, F., Van Baren, N., Baurain, J.-F., DeSmet, C., Chambost, H., Vitale, M., Moretta, A., Boon, T., and Coulie, P. (1997). Immunity 6, 199. Ishitani, A., Kishida, M., Sageshima, N., Yashiki, S., Sonoda, S., Hayami, M., Smith, A. G., and Hatake, K. (1999). Immunogenetics 49, 808. Johnson, P. M., Christmas, S. E., and Vince, G. S. (1999). Hum. Reprod. 2(14 Suppl), 26. Jun, Y., Kim, E., Jin, M., Sung, H. C., Han, H., Geraghty, D. E., and Ahn, K. (2000). J. Immunol. 164, 805. Jurisicova, A., Casper, R. F., MacLusky, N. J., and Librach, C. L. (1996a). Fertil. Steril. 65, 997. Jurisicova, A., Casper, R. F., MacLusky, N. J., Mills, G. B., and Librach, C. L. (1996b). Proc. Natl. Acad. Sci. USA 93, 161. Kanai, T., Fujii, T., Kozuma, S., Yamashita, T., Miki, A., Kikuchi, A., and Taketani, Y. (2001a). Mol. Hum. Reprod. 7, 195. Kanai, T., Fujii, T., Unno, N., Yamashita, T., Hyodo, H., Miki, A., Hamai, Y., Kozuma, S., and Taketani, Y. (2001b). Am. J. Reprod. Immunol. 45, 94. Kapasi, K., Albert, S. E., Yie, S.-M., Zavazava, N., and Librach, C. L. (2000). Immunol. 101, 191. Karhukorpi, J., Laitinen, T., and Tiilikainen, A. S. (1997). Br. J. Obstet. Gynaecol. 104, 1212. Khalil-Daher, I., Riteau, B., Menier, C., Sedlik, C., Paul, P., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (1999). J. Reprod. Immunol. 43, 175. Khong, T. Y., De Wolf, F., Robertson, W. B., and Brosens, I. (1986). Br. J. Obstet. Gynaecol. 93, 1049. Khosrotehrani, K., Le Danff, C., Reynaud, B., Dubertret, L., Carosella, E. D., and Aractingi, S. (2001). J. Invest. Dermatol. 117, 750. King, A., Allan, D. S., Bowen, M., Powis, S. J., Joseph, S., Verma, S., Hiby, S. E., McMichael, A. J., Loke, Y. W., and Braud, V. M. (2000). Eur. J. Immunol. 30, 1623. King, A., Boocock, C., Sharkey, A. M., Gardner, L., Beretta, A., Siccardi, A. G., and Loke, Y. W. (1996). J. Immunol. 156, 2068. Kirszenbaum, M., Djoulah, S., Hors, J., Le Gall, I., de Oliveira, E. B., Prost, S., Dausset, J., and Carosella, E. D. (1997). Hum. Immunol. 53, 140. Kirszenbaum, M., Moreau, P., Gluckman, E., Dausset, J., and Carosella, E. (1994). Proc. Natl. Acad. Sci. USA 91, 4209. Kirszenbaum, M., Moreau, P., Teyssier, M., Lafon, C., Gluckman, E., Dausset, J., and Carosella, E. (1995). Hum. Immunol. 43, 237. Knoefel, B., K., N., Steiner, T., Junker, K., Kosmehl, H., Rebstock, K., Reinhold, D., and Junker, U. (1997). J. Interferon Cytokine Res. 17, 95. Kovats, S., Main, E. K., Librach, C., Stubblebine, M., Fisher, S. J., and DeMars, R. (1990). Science 248, 220. Langat, D. K., Morales, P. J., Fazleabas, A. T., Mwenda, J. M., and Hunt, J. S. (2002). Immunogenetics 54, 164. Le Bouteiller, P., and Blaschitz, A. (1999). Immunol. Rev. 167, 233. Le Gal, F. A., Riteau, B., Sedlik, C., Khalil-Daher, I., Menier, C., Dausset, J., Guillet, J. G., Carosella, E. D., and Rouas-Freiss, N. (1999). Int. Immunol. 11, 1351. Lee, N., Goodlett, D. R., Ishitani, A., Marquardt, H., and Geraghty, D. E. (1998a). J. Immunol. 160, 4951. Lee, N., Llano, M., Carretero, M., Ishitani, A., Navarro, F., Lopez-Botet, M., and Geraghty, D. E. (1998b). Proc. Natl. Acad. Sci. USA 95(9), 5199.

248

E. D. CAROSELLA ET AL.

Lee, N., Malacko, A. R., Ishitani, A., Chen, M. C., Bajorath, J., Marquardt, H., and Geraghty, D. E. (1995). Immunity 3, 591. Lefebvre, S., Adrian, F., Moreau, P., Gourand, L., Dausset, J., Berrih-Aknin, S., Carosella, E. D., and Paul, P. (2000). Hum. Immunol. 61, 1095. Lefebvre, S., Antoine, M., Uzan, S., McMaster, M., Dausset, J., Carosella, E. D., and Paul, P. (2002). J. Pathol. 196, 266. Lefebvre, S., Berrih-Aknin, S., Adrian, F., Moreau, P., Poea, S., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (2001). J. Biol. Chem. 276, 6133. Lefebvre, S., Moreau, P., Dausset, J., Carosella, E. D., and Paul, P. (1999a). Placenta 20, 293. Lefebvre, S., Moreau, P., Guiard, V., Ibrahim, E. C., Adrian-Cabestre, F., Menier, C., Dausset, J., Carosella, E. D., and Paul, P. (1999b). J. Reprod. Immunol. 43, 213. Liang, S., Baibakov, B., and Horuzsko, A. (2002). Eur. J. Immunol. 32, 2418. Lila, N., Amrein, C., Guillemain, R., Chevalier, P., Latremouille, C., Fabiani, J. N., Dausset, J., Carosella, E. D., and Carpentier, A. (2002). Circulation 105, 1949. Lila, N., Carpentier, A., Amrein, C., Khalil-Daher, I., Dausset, J., and Carosella, E. D. (2000). Lancet 355, 2138. Lila, N., Rouas-Freiss, N., Dausset, J., Carpentier, A., and Carosella, E. D. (2001). Proc. Natl. Acad. Sci. USA 98, 12150. Lim, K. H., Zhou, Y., Janatpour, M., McMaster, M., Bass, K., Chun, S. H., and Fisher, S. J. (1997). Am. J. Pathol. 151, 1809. Llano, M., Lee, N., Navarro, F., Garcia, P., Albar, J. P., Geraghty, D. E., and Lopez-Botet, M. (1998). Eur. J. Immunol. 28, 2854. Loke, Y. W., King, A., Burrows, T., Gardner, L., Bowen, M., Hiby, S., Howlett, S., Holmes, N., and Jacobs, D. (1997). Tissue Antigens 50, 135. Lozano, J. M., Gonzalez, R., Kindelan, J. M., Rouas-Freiss, N., Caballos, R., Dausset, J., Carosella, E. D., and Pena, J. (2002). Aids 16, 347. Maier, S., Geraghty, D., and Weiss, E. (1999). Transplant Proc. 31, 1849. Mallet, V., Blaschitz, A., Crisa, L., Schmitt, C., Fournel, S., King, A., Loke, Y. W., Dohr, G., and Le Bouteiller, P. (1999a). Int. Immunol. 11, 889. Mallet, V., Fournel, S., Schmitt, C., Campan, A., Lenfant, F., and Le Bouteiller, P. (1999b). J. Reprod. Immunol. 43, 225. Mallet, V., Proll, J., Solier, C., Aguerre-Girr, M., DeRossi, M., Loke, Y. W., Lenfant, F., and Le Bouteiller, P. (2000). Hum. Immunol. 61, 212. Malmberg, K. J., Levitsky, V., Norell, H., Matos, C. T., Carlsten, M., Schedvins, K., Rabbani, H., Moretta, A., Soderstrom, K., Levitskaya, J., and Kiessling, R. (2002). J. Clin. Invest. 110, 1515. Marchal-Bras-Goncalves, R., Rouas-Freiss, N., Connan, F., Choppin, J., Dausset, J., Carosella, E., Kirszenbaum, M., and Guillet, J. G. (2001). Transplant Proc. 33, 2355. Masternak, K., Muhlethaler-Mottet, A., Villard, J., Peretti, M., and Reith, W. (2000). Rev. Immunogenet. 2, 267. Masternak, K., and Reith, W. (2002). EMBO J. 21, 1379. Matsunami, K., Miyagawa, S., Nakai, R., Murase, A., and Shirakura, R. (2001). Clin. Exp. Immunol. 126, 165. Matsunami, K., Miyagawa, S., Nakai, R., Yamada, M., and Shirakura, R. (2002). Transplantation 73, 1582. Matte, C., Lacaille, J., Zijenah, L., Ward, B., and Roger, M. (2000). Hum. Immunol. 61, 1150. Matte, C., Lacaille, J., Zijenah, L., Ward, B., and Roger, M. (2002). Hum. Immunol. 63, 495. McMaster, M., Zhou, Y., Shorter, S., Kapasi, K., Geraghty, D., Lim, K. H., and Fisher, S. (1998). J. Immunol. 160, 5922. McMaster, M. T., Librach, C. L., Zhou, Y., Lim, K. H., Janatpour, M. J., DeMars, R., Kovats, S., Damsky, C., and Fisher, S. J. (1995). J. Immunol. 154, 3771.

HLA-G MOLECULES

249

Menicucci, A., Noci, I., Fuzzi, B., Criscuoli, L., Scarselli, G., Baricordi, O., and Mattiuz, P. (1999). Hum. Immunol. 60, 1054. Menier, C., Riteau, B., Carosella, E. D., and Rouas-Freiss, N. (2002). Int. J. Cancer 100, 63. Menier, C., Riteau, B., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (2000). Hum. Immunol. 61, 1118. Menier, C., Saez, B., Horejsi, V., Martinozzi, S., Krawice-Radanne, I., Bruel, S., Le Dannff, C., Reboul, M., Hilgert, I., Rabreau, M., Larrad, M. L., Pla, M., Carosella, E. D., and Rouas-Freiss, N. (2003). Hum. Immunol. 64, 315. Messer, G., Zemmour, J., Orr, H. T., Parham, P., Weiss, E. H., and Girdlestone, J. (1992). J. Immunol. 148, 4043. Minguela, A., Torio, A., Marin, L., Muro, M., Villar, L. M., Diaz, J., Ramirez, P., Parrilla, P., Garcia-Alonso, A. M., and Alvarez-Lopez, M. R. (1999). Hum. Immunol. 60, 500. Mizuno, S., Emi, N., Kasai, M., Ishitani, A., and Saito, H. (2000). Br. J. Haematol. 111, 280. Moreau, P., Adrian-Cabestre, F., Menier, C., Guiard, V., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (1999). Int. Immunol. 11, 803. Moreau, P., Carosella, E., Teyssier, M., Prost, S., Gluckman, E., Dausset, J., and Kirszenbaum, M. (1995). Hum. Immunol. 43, 231. Moreau, P., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (2002). Biol. Reprod. 67, 1375. Moreau, P., Faure, O., Lefebvre, S., Ibrahim, E. C., O’Brien, M., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (2001). Transplant Proc. 33, 2277. Moreau, P., Lefebvre, S., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (1998). Hum. Immunol. 59, 751. Moreau, P., Paul, P., Gourand, L., Prost, S., Dausset, J., Carosella, E., and Kirszenbaum, M. (1997). Hum. Immunol. 52, 41. Moreau, P., Mouillot, G., Rousseau, P., Marcou, C., Dausset, J., and Carosella, E. D. (2003). Proc. Natl. Acad. Sci. USA 100, 1191. Munz, C., Stevanovic, S., and Rammensee, H. G. (1999). J. Reprod. Immunol. 43, 139. Navarro, F., Llano, M., Bellon, T., Colonna, M., Geraghty, D. E., and Lopez-Botet, M. (1999). Eur. J. Immunol. 29, 277. Neefjes, J. J., and Momburg, F. (1993). Curr. Opin. Immunol. 5, 27. Ober, C., Aldrich, C., Rosinsky, B., Robertson, A., Walker, M. A., Willadsen, S., Verp, M. S., Geraghty, D. E., and Hunt, J. S. (1998). Placenta 19, 127. O’Brien, M., McCarthy, T., Jenkins, D., Paul, P., Dausset, J., Carosella, E. D., and Moreau, P. (2001). Cell. Mol. Life Sci. 58, 1943. Onno, M., Amiot, L., Bertho, N., Drenou, B., and Fauchet, R. (1997). Tissue Antigens 49, 356. Onno, M., Guillaudeux, T., Amiot, L., Renard, I., Drenou, B., Hirel, B., Girr, M., Semana, G., Le Bouteiller, P., and Fauchet, R. (1994). Hum. Immunol. 41, 79. Onno, M., Le Friec, G., Pangault, C., Amiot, L., Guilloux, V., Drenou, B., Caulet-Maugendre, S., Andre, P., and Fauchet, R. (2000a). Hum. Immunol. 61, 1086. Onno, M., Pangault, C., Le Friec, G., Guilloux, V., Andre, P., and Fauchet, R. (2000b). J. Immunol. 164, 6426. Opal, S. M., and DePalo, V. A. (2000). Chest 117, 1162. Palmisano, G. L., Pistillo, M. P., Fardin, P., Capanni, P., Nicolo, G., Salvi, S., Spina, B., Pasciucco, G., and Ferrara, G. B. (2002). Hum. Immunol. 63, 969. Pangault, C., Amiot, L., Caulet-Maugendre, S., Brasseur, F., Burtin, F., Guilloux, V., Drenou, B., Fauchet, R., and Onno, M. (1999). Tissue Antigens 53, 335. Pangault, C., Le Friec, G., Caulet-Maugendre, S., Lena, H., Amiot, L., Guilloux, V., Onno, M., and Fauchet, R. (2002). Hum. Immunol. 63, 83. Paradinas, F. J., Browne, P., Fisher, R. A., Foskett, M., Bagshawe, K. D., and Newlands, E. (1996). Histopathology 28, 101.

250

E. D. CAROSELLA ET AL.

Park, B., Lee, S., Kim, E., Chang, S., Jin, M., and Ahn, K. (2001). Immunity 15, 213. Paul, P., Adrian Cabestre, F., Le Gal, F.-A., Khalil-Daher, I., Le Danff, C., Schmid, M., Mercier, S., Avril, M.-F., Dausset, J., Guillet, J.-G., and Carosella, E. D. (1999). Cancer Res. 59, 1954. Paul, P., Cabestre, F. A., Ibrahim, E. C., Lefebvre, S., Khalil-Daher, I., Vazeux, G., Quiles, R. M., Bermond, F., Dausset, J., and Carosella, E. D. (2000a). Hum. Immunol. 61, 1138. Paul, P., Rouas-Freiss, N., Khalil-Daher, I., Moreau, P., Riteau, B., Le Gal, F. A., Avril, M. F., Dausset, J., Guillet, J. G., and Carosella, E. D. (1998). Proc. Natl. Acad. Sci. USA 95, 4510. Paul, P., Rouas-Freiss, N., Moreau, P., Adrian Cabestre, F., Menier, C., Khalil-Daher, I., Pangault, C., Onno, M., Fauchet, R., Martinez-Laso, J., Morales, P., Arnaiz-Villena, A., Giacomini, P., Natali, P. G., Frumento, G., Ferrara, G. B., McMaster, M., Fisher, S., Schust, D., Ferrone, S., Dausset, J., Geraghty, D., and Carosella, E. D. (2000b). Hum. Immunol. 61, 1177. Pende, D., Cantoni, C., Rivera, P., Vitale, M., Castriconi, R., Marcenaro, S., Nanni, M., Biassoni, R., Bottino, C., Moretta, A., and Moretta, L. (2001). Eur. J. Immunol. 31(4), 1076. Pfeiffer, K. A., Fimmers, R., Engels, G., van der Ven, H., and van der Ven, K. (2001). Mol. Hum. Reprod. 7, 373. Polakova, K., and Russ, G. (2000). Neoplasma 47, 342. Ponte, M., Cantoni, C., Biassoni, R., Tradori-Cappai, A., Bentivoglio, G., Vitale, C., Bertone, S., Moretta, A., Moretta, L., and Mingari, M. C. (1999). Proc. Natl. Acad. Sci. USA 96, 5674. Proll, J., Bensussan, A., Goffin, F., Foidart, J. M., Berrebi, A., and Le Bouteiller, P. (2000). Tissue Antigens 56, 479. Proll, J., Blaschitz, A., Hutter, H., and Dohr, G. (1999). Am. J. Reprod. Immunol. 42, 30. Puppo, F., Costa, M., Contini, P., Brenci, S., Cevasco, E., Ghio, M., Norelli, R., Bensussan, A., Capitanio, G. L., and Indiveri, F. (1999). Transplant Proc. 31, 1841. Rabreau, M., Rouas-Freiss, N., Landi, M., Le Danff, C., and Carosella, E. D. (2000). Hum. Immunol. 61, 1108. Rajagopalan, S., and Long, E. O. (1999). J. Exp. Med. 189, 1093. Real, L. M., Cabrera, T., Collado, A., Jimenez, P., Garcia, A., Ruiz-Cabello, F., and Garrido, F. (1999). Int. J. Cancer 81, 512. Rebmann, V., Pfeiffer, K., Passler, M., Ferrone, S., Maier, S., Weiss, E., and Grosse-Wilde, H. (1999). Tissue Antigens 53, 14. Rebmann, V., van der Ven, K., Passler, M., Pfeiffer, K., Krebs, D., and Grosse-Wilde, H. (2001). Tissue Antigens 57, 15. Rieger, L., Hofmeister, V., Probe, C., Dietl, J., Weiss, E. H., Steck, T., and Kammerer, U. (2002). Mol. Hum. Reprod. 8, 255. Riteau, B., Menier, C., Khalil-Daher, I., Martinozzi, S., Pla, M., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (2001a). Int. Immunol. 13, 193. Riteau, B., Menier, C., Khalil-Daher, I., Sedlik, C., Dausset, J., Rouas-Freiss, N., and Carosella, E. D. (1999). J. Reprod. Immunol. 43, 203. Riteau, B., Moreau, P., Menier, C., Khalil-Daher, I., Khosrotehrani, K., Bras-Goncalves, R., Paul, P., Dausset, J., Rouas-Freiss, N., and Carosella, E. D. (2001b). Transplant Proc. 33, 2360. Riteau, B., Rouas-Freiss, N., Menier, C., Paul, P., Dausset, J., and Carosella, E. D. (2001c). J. Immunol. 166, 5018. Roberts, J., and Redman, C. (1993). Lancet 341, 1447. Roth, I., Corry, D. B., Locksley, R. M., Abrams, J. S., Litton, M. J., and Fisher, S. J. (1996). J. Exp. Med. 184, 539. Rouas-Freiss, N., Goncalves, R. M., Menier, C., Dausset, J., and Carosella, E. D. (1997a). Proc. Natl. Acad. Sci. USA 94, 11520. Rouas-Freiss, N., Khalil-Daher, I., Riteau, B., Menier, C., Paul, P., Dausset, J., and Carosella, E. D. (1999). Sem. Cancer Biol. 9, 3.

HLA-G MOLECULES

251

Rouas-Freiss, N., Marchal, R. E., Kirszenbaum, M., Dausset, J., and Carosella, E. D. (1997b). Proc. Natl. Acad. Sci. USA 94, 5249. Rousseau, P., Paul, P., O’Brien, M., Dausset, J., Carosella, E. D., and Moreau, P. (2000). Hum. Immunol. 61, 1132. Ryan, A. F., Grendell, R. L., Geraghty, D. E., and Golos, T. G. (2002). J. Immunol. 169, 673. Saito, S., Sakai, M., Sasaki, Y., Tanebe, K., Tsuda, H., and Michimata, T. (1999). Clin. Exp. Immunol. 117, 550. Sasaki, H., Xu, X. C., and Mohanakumar, T. (1999). J. Immunol. 163, 6301. Sasaki, H., Xu, X. C., Smith, D., Shenoy, S., Lowell, J., Howard, T., and Mohanakumar, T. (1999a). Transplant Proc. 31, 953. Sasaki, H., Xu, X. C., Smith, D. M., Howard, T., and Mohanakumar, T. (1999b). Transplantation 67, 31. Saverino, D., Fabbi, M., Ghiotto, F., Merlo, A., Bruno, S., Zarcone, D., Tenca, C., Tiso, M., Santoro, G., Anastasi, G., Cosman, D., Grossi, C. E., and Ciccone, E. (2000). J. Immunol. 165, 3742. Schmidt, C. M., Chen, H. L., Chiu, I., Ehlenfeldt, R. G., Hunt, J. S., and Orr, H. T. (1995). J. Immunol. 155, 619. Schmidt, C. M., Ehlenfeldt, R. G., Athanasiou, M. C., Duvick, L. A., Heinrichs, H., David, C. S., and Orr, H. T. (1993). J. Immunol. 151, 2633. Schust, D. J., Hill, A. B., and Ploegh, H. L. (1996). J. Immunol. 157, 3375. Schust, D. J., Tortorella, D., and Ploegh, H. L. (1999). Semin. Cancer Biol. 9, 37. Schust, D. J., Tortorella, D., Seebach, J., Phan, C., and Ploegh, H. L. (1998). J. Exp. Med. 188, 497. Shurin, M. R., Lu, L., Kalinski, P., Stewarts-Akers, A. M., and Lotze, M. T. (1999). Springer Semin. Immunopathol. 21, 339. Singer, G., Kurman, R. J., McMaster, M. T., and Shih Ie, M. (2002). Am. J. Surg. Pathol. 26, 914. Sivori, S., Parolini, S., Marcenaro, E., Millo, R., Bottino, C., and Moretta, A. (2000). Hum. Immunol. 61, 1055. Slukvin, II, Lunn, D. P., Watkins, D. I., and Golos, T. G. (2000). Proc. Natl. Acad. Sci. USA 97, 9104. Solier, C., Mallet, V., Lenfant, F., Bertrand, A., Huchenq, A., and Le Bouteiller, P. (2001). Immunogenetics 53, 617. Speiser, D. E., Pittet, M. J., Valmori, D., Dunbar, R., Rimoldi, D., Lienard, D., MacDonald, H. R., Cerottini, J.-C., Cerundolo, V., and Romero, P. (1999). J. Exp. Med. 190, 775. Spencer, J. V., Lockridge, K. M., Barry, P. A., Lin, G., Tsang, M., Penfold, M. E., and Schall, T. J. (2002). J. Virol. 76, 1285. Stallmach, T., Hebisch, G., Orban, P., and Lu, X. (1999). Virchows Arch. 434, 207. Stewart, C. L., Kaspar, P., Brunet, L. J., Bhatt, H., Gadi, I., Kontgen, F., and Abbondanzo, S. J. (1992). Nature 359, 76. Suarez, M. B., Morales, P., Castro, M. J., Fernandez, V., Varela, P., Alvarez, M., Martinez-Laso, J., and Arnaiz-Villena, A. (1997). Immunogenetics 45, 464. Sun, K., Yang, K., and Challis, J. R. (1998). Placenta 19, 353. Tarrade, A., Lai Kuen, R., Malassine, A., Tricottet, V., Blain, P., Vidaud, M., and Evain-Brion, D. (2001). Lab. Invest. 81, 1199. Taylor, R. N. (1997). Am. J. Reprod. Immunol. 37, 79. Teasdale, R. D., and Jackson, M. R. (1996). Annu. Rev. Cell. Dev. Biol. 12, 27. Ugurel, S., Rebmann, V., Ferrone, S., Tilgen, W., Grosse-Wilde, H., and Reinhold, U. (2001). Cancer 92, 369. Ugurel, S., Reinhold, U., and Tilgen, W. (2002). Onkologie 25, 129. Urosevic, M., Kurrer, M. O., Kamarashev, J., Mueller, B., Weder, W., Burg, G., Stahel, R. A., Dummer, R., and Trojan, A. (2001). Am. J. Pathol. 159, 817.

252

E. D. CAROSELLA ET AL.

Urosevic, M., Trojan, A., and Dummer, R. (2002). J. Pathol. 196, 252. Urosevic, M., Willers, J., Mueller, B., Kempf, W., Burg, G., and Dummer, R. (2002). Blood 99, 609. Vale´s-Gomez, M., Reyburn, H. T., Erskine, R. A., Lopez-Botet, M., and Strominger, J. L. (1999). EMBO J. 18, 4250. van den Elsen, P. J., Gobin, S. J., van Eggermond, M. C., and Peijnenburg, A. (1998a). Immunogenetics 48, 208. van den Elsen, P. J., Peijnenburg, A., van Eggermond, M. C., and Gobin, S. J. (1998b). Immunol. Today 19, 308. van der Ven, K., Skrablin, S., Engels, G., and Krebs, D. (1998). Hum. Immunol. 59, 302. van Wijk, I. J., Griffioen, S., Tjoa, M. L., Mulders, M. A., van Vugt, J. M., Loke, Y. W., and Oudejans, C. B. (2001). Am. J. Obstet. Gynecol. 184, 991. Wagner, S. N., Rebmann, V., Willers, C. P., Grosse-Wilde, H., and Goos, M. (2000). Lancet 356, 220. Waldburger, J. M., Masternak, K., Muhlethaler-Mottet, A., Villard, J., Peretti, M., Landmann, S., and Reith, W. (2000). Immunol. Rev. 178, 148. Watkins, D. I., Chen, Z. W., Hughes, A. L., Evans, M. G., Tedder, T. F., and Letvin, N. L. (1990). Nature 346, 60. Wiendl, H., Behrens, L., Maier, S., Johnson, M. A., Weiss, E. H., and Hohlfeld, R. (2000). Ann. Neurol. 48, 679. Wiendl, H., Mitsdoerffer, M., Hofmeister, V., Wischhusen, J., Bornemann, A., Meyermann, R., Weiss, E. H., Melms, A., and Weller, M. (2002). J. Immunol. 168, 4772. Yamashita, T., Fujii, T., Tokunaga, K., Tadokoro, K., Hamai, Y., Miki, A., Kozuma, S., Juji, T., and Taketani, Y. (1999). Am. J. Reprod. Immunol. 41, 159. Yang, Y., Chu, W., Geraghty, D. E., and Hunt, J. S. (1996). J. Immunol. 156, 4224. Yang, Y., Geraghty, D. E., and Hunt, J. S. (1995). J. Reprod. Immunol. 29, 179. Yelavarthi, K. K., Schmidt, C. M., Ehlenfeldt, R. G., Orr, H. T., and Hunt, J. S. (1993). J. Immunol. 151, 3638. Yue, F. Y., Dummer, R., Geertsen, R., Hofbauer, G., Laine, E., Manolio, S., and Burg, G. (1997). Int. J. Cancer 71, 630.