2 dahl 2012 hla early pregnancy loss.pdf

...........................................................................................................................

Human leucocyte antigen class Ibmolecules in pregnancy successand early pregnancy lossMetteDahl and Thomas Vauvert F. Hviid *

Centre for Immune Regulation and Reproductive Immunology (CIRRI), Department of Clinical Biochemistry, Copenhagen University Hospital(Roskilde), Region Sjlland, 7-13 Kgevej, DK-4000 Roskilde, Denmark

*Correspondence address. Tel: +45-4732-5622; Fax: +45-4632-1615; E-mail: [email protected] or [email protected] on March 20, 2011; resubmitted on September 1, 2011; accepted on October 6, 2011

table of contents

Introduction Search method and HLA nomenclature HLA-G expression at the materno-fetal interface Functions of HLA-G during pregnancy HLA-G polymorphism and pregnancy success Soluble HLA-G: A possible biomarker in human reproduction HLA-E alleles and expression in relation to pregnancy HLA-E in regulating NK cells during pregnancy Ambiguity of the basic role of HLA-F in reproduction New aspects of HLA-F function Signicance of HLA-C Conclusions and future perspectives

background: The human leucocyte antigen (HLA) class Ib molecules, HLA-E, -F and -G, are expressed at the materno-fetal interface.Because of the apparent immunoregulatory functions of these proteins, they may be involved in successful acceptance of the semi-allogenicfetus during pregnancy.

methods: The literature on polymorphisms of the three genes, expression patterns of the proteins, and interactions with immune cellreceptors have been evaluated to elucidate whether HLA-E, -F and -G are involved in the pathogenesis of some cases of recurrent miscar-riages and unexplained infertility.

results and conclusions: The HLA class Ib molecules seem to induce suppression of the maternal immune system, but are notnecessarily fundamental factors for pregnancy success. However, evidence points towards low expression of these proteins, especiallyHLA-G, being associated with reduced fertility. To clarify the functions of HLA-E, -F and -G future studies need to link investigations ofthe polymorphisms in these genes to measurements of protein levels, and examine the role of these proteins in the complex interplay ofimmune cells and cytokines at the materno-fetal interface.

Key words: MHC / HLA class Ib / pregnancy / recurrent miscarriages / assisted reproduction

IntroductionThe human leucocyte antigen (HLA) system has functions in antigenprocessing and presentation. The genes encoding these moleculesare located on the short arm of chromosome 6 and can be dividedinto HLA class I and class II.

The HLA class I genes are expressed on the surface of all nucleatedcells, and are responsible for displaying foreign peptides to the CD8+

T cells. This will typically elicit a specic T cell response with cytotoxicimmunological processes killing all cells displaying foreign peptide ontheir surfaces. The class I molecules consist of six different proteins,

& The Author 2011. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.For Permissions, please email: [email protected]

Human Reproduction Update, Vol.18, No.1 pp. 92109, 2012

doi:10.1093/humupd/dmr043

at Hacettepe U

niversity Library (HU) on February 23, 2012http://hum

upd.oxfordjournals.org/D

ownloaded from

of which the classical HLA-A, -B and -C are highly polymorphic andhave central functions in antigen processing and presentation asdescribed, whereas the non-classical HLA-E, -F and -G showdecreased polymorphism, restricted tissue distribution and haveother functions than HLA class Ia, as outlined in this review.

The class II genes encodes ve differentmolecules, of which HLA-DP,-DQ and -DR are polymorphic and function directly in antigen presen-tation, whereas the HLA-DM and HLA-DO show decreased poly-morphism and function in regulating the peptide load of the otherHLA class II molecules. TheHLA class II molecules aremainly expressedon dendritic cells, B lymphocytes and macrophages, which expressforeign peptides on the cell surface, giving the CD4+ T cells the possi-bility of initiating either a T helper 1 (Th1) or a Th2 response if pathogenrecognition occurs. This means that, depending on the cytokinespresent, and the type of cell activating the T cell, a Th1 response char-acterized by a cell-mediated immune response, or a Th2 response char-acterized by the production of antibodies, will be initiated (Parham,2005a, b). Balancing the Th1 and Th2 immune responses seems toserve an important role during pregnancy, as it is believed that the ma-ternal immune system shifts from a Th1 prole to a less cytotoxic Th2prole in response to fertilization (Wegmann et al., 1993; Billington,2003). However, the discovery that some cytokines, such as IL-10,are secreted by both Th1 and Th2 cells indicates that a simple Th1/Th2 shift is a simplied concept, not sufcient to explain how toleranceby the maternal immune system is necessary for fetal survival. Theincreased complexity has recently drawn focus towards regulatory Tcells, suggesting that this cell population is important in balancing theTh1 and Th2 responses.

Furthermore, the discovery that fetal cells are devoid of the highlypolymorphic HLA class Ia molecules, except for a low expression ofHLA-C, is believed to play a dominant role for the induction of toler-ance to the semi-allogenic fetus (Moffett-King et al., 2002). Interesting-ly, the fetal-derived tissue in placenta do express the less polymorphicHLA class Ib molecules, HLA-E, -F and -G, which has led to increasedinterest in the immunological role of these three proteins during preg-nancy (Fig. 1; Favier et al., 2007). Originally, HLA-G was thought to beexpressed only in fetal tissues, but recently the molecule has also beenfound in healthy human tissue such as thymic medulla, cornea and pan-creas. In addition, expression of HLA-G seems to be either induced orincreased in a number of pathological conditions such as cancer, auto-immune diseases and viral infections (Favier et al., 2007).

HLA-E is expressed in all human tissues where HLA class Ia mole-cules are expressed (Heinrichs and Orr, 1990). HLA-F expressionseems to be limited to the tonsils, spleen, thymic tissue, and placenta,and overall transcription of this gene appears to be higher in lymphoidcells compared with non-lymphoid cells (Lepin et al., 2000).

In addition, a part of the explanation of how the fetus avoids immuno-logical attack from thematernal immune system,might lie in the fact thatimmune cells that are in close contact with fetal tissues, display alteredfunctions comparedwith peripheral immune cells. In this regard, around90%of uterine natural killer cells (uNK cells) have an unusual phenotype,(CD56bright) expressing high levels of CD56, displaying decreased cyto-toxicity and an altered receptor expression with higher levels of bothCD94/NKG2 and IL2Rabg, and essentially absence of Killer-cellimmunoglobulin-like receptor (KIR; Starkey et al., 1991; Jacobs et al.,2001). These cells accumulate in large numbers at the implantationsite (Ashkar et al., 2003), and have been shown to be themain cytokine-

producing NK cells (Cooper et al., 2001), secreting cytokines, whichmight be important during implantation and subsequent angiogenesisand development of the placenta (Saito et al., 1993; Trundley andMoffett, 2004; Higuma-Myojo et al., 2005).

Decidual macrophages might also have immunoregulatory functions asthey express indolamine-2,3-dioxygenase (IDO), which metabolizes theamino acid tryptophan needed for T cell activation and displays decreasedamounts of theT lymphocyte co-stimulatorymoleculesCD80 andCD86.

Concerning T cells, approximately three times as many CD3+CD8+

cells as CD3+CD4+ cells are found in the decidua comparedwith periph-eral blood (Bulmer et al., 1991). In women experiencing recurrentmiscar-riages (RMs) and unexplained infertility, disturbances of this ratiomight beinvolved in the pathogenesis (Klentzeris et al., 1992; Lachapelle et al.,1996). Besides these T cells, a small subgroup of the CD3+CD56+

NKT cells called invariant Natural Killer T (iNKT) cells are found indecidua, and might have a role in the development of tolerance as theyare able to suppress antifetal recognition through interactions withCD1d on the surface of trophoblast cells when IL-10 is present (Scaifeet al., 2006; Uemura et al., 2008). A reduced number of these cellshave been correlated with RM (Yahata et al., 1998; Yamamoto et al.,1999). Recently, a new population of T cells has been identied as regu-latory T cells. These cells are characterized by the CD4+CD25+ pheno-type, and express high levels of FoxP3, which is known as a transcriptionalregulator important in development and function of cells. Among otherfunctions, these cells inhibit the proliferation and cytokine secretion ofCD3+ T cells, reduce the cytotoxicity of NK cells and prevent dendriticcell maturation (Sakaguchi et al., 2006). The regulatory T cells make up1015% of all CD4+ T lymphocytes and they seem to bemost abundantduring early pregnancy (Heikkinen et al., 2004; Sasaki et al., 2004). Theexact mechanisms of how tolerance is induced by these cells are notyet claried, but the cells have been shown to express IDO preventingT cell activation, thereby abolishing the immunological processescarried out by these cells (Heikkinen et al., 2004).

In this review, the possible role of the HLA class Ib molecules in thepathogenesis of infertility and RM is discussed in an immunologicalcontext. Infertility is dened as the inability of a couple to conceiveafter 1 year, and 1015% of couples in the reproductive age can becategorized as infertile. Interestingly, 1520% of infertility casescannot be accounted for by any known factor (Schorge, 2008). RMis dened as three or more consecutive pregnancy losses before the20th gestational week, and is estimated to affect 12% of allcouples (Coulam, 1991). Interestingly, in about 50% of RM cases noetiological factor can be identied (Lee and Silver, 2000). It can bespeculated that disturbances in the processes generating the charac-teristic immune tolerance during pregnancy might have a role in thepathogenesis of some cases of RM and unexplained infertility.

Search method and HLAnomenclatureA search of the scientic literature was performed using the NCBIdatabase. The search was focused on the polymorphisms, expressionpatterns and interactions with immune cells and cytokines of the HLAclass Ib molecules and HLA-C. The search was limited to studies pro-viding insights into the possible function of these molecules in thepathogenesis of RM and unexplained infertility.

HLA class Ib and pregnancy success 93

at Hacettepe U



ownloaded from

The HLA nomenclature system has recently been revised and in thisreview, both the nomenclature in use at the time where the specicstudies were performed and the corresponding new nomenclatureare used because the literature does not always offer a descriptionof the specic alleles and polymorphisms, as we know them currently.If any doubt exists on the transformation, all possible alleles corre-sponding to the new nomenclature are indicated.

HLA-G expression at thematerno-fetal interfaceOf the HLA class Ib molecules, HLA-G is the most intensively studied,and much effort has been put into specifying the role of this protein indeveloping tolerance during pregnancy.

At present, seven alternatively spliced variants of HLA-G have beenidentied, counting four membrane-bound and three soluble isoforms,as represented in Fig. 2A (Ishitani and Geraghty, 1992; Pascale et al.,2000a). Of these variants, only HLA-G1 and -G5 associate withb2-microglobulin (b2m). This is an important fact when designingassays that detects different HLA-G isoforms, as some antibodiesonly bind to HLA-G molecules when these are complexed with b2m.

The expression of splice variants in different trophoblast cellpopulations of placenta is an area of controversy. It is well-documented that invasive extravillous trophoblast (EVT) cellsexpress a membrane-anchored HLA-G protein at the cell surface, asrst shown by Ellis et al. (1986) and subsequently by several groups.Whether or not other trophoblast cell populations, such as syncytio-trophoblast (ST) cells and VT cells express HLA-G, is still a matter ofdiscussion, and this disagreement isprobablyprimarily caused bytechnical differences, such as selection of antibodies, between different

Figure 1 A schematic representation of the fetus, placenta, decidua and functional signicance of the threeHLA class Ibmolecules: HLA-F, -E and -G. Therectangular insert shows a magnication of the materno-fetal interface with localization of the trophoblast cell populations. The cytotrophoblast (CT)cells serve as progenitors for the differentiated trophoblast cell populations; the extravillous trophoblast (EVT) cells, which express HLA-E, HLA-G andprobably HLA-F on their surfaces; and the syncytiotrophoblast (ST) cells, which along with the villous trophoblast (VT) cells might express solubleHLA-G and possibly intracellularly located HLA-E and HLA-F. Arrows from the round insert display the interactions of HLA-E, -F and -G with differentreceptors on the surface of uNK cells, B19 monocytes, regulatory T cells (Treg) and CD4+/CD8+ T cells. See text for details.

94 Dahl and Hviid

at Hacettepe U



ownloaded from

Figure 2 (A) HLA-G gene structure showing the extracellular, transmembrane and intracellular regions. The localization of gene polymorphisms im-portant in pathological conditions of pregnancy is marked as well. Alternative splicing of HLA-G results in production of four membrane-bound and threesoluble isoforms of the protein. In addition to the seven isoforms, alternative HLA-Gmolecules have been demonstrated in transfection studies, howeverat present, there is not sufcient evidence that these molecules are found at the materno-fetal interface. No alternative splicing is known for HLA-E andHLA-F. (B) Schematic representations of HLA-G as a disulde-linked-b2m-associated dimer, a b2m-free disulde dimer and a b2m-free heavy chain. (C)Posttranslational modications specic for HLA-E and HLA-F, respectively. A nonamer sequence is needed for the surface expression of HLA-E, whereasHLA-F has been found to include a cytoplasmic tail for export from the ER, and a RxRmotif specic forHLA-F for localization to theGolgia complex, whereHLA-F protein has been shown to accumulate. Which route HLA-F takes to a possible surface expression is still unknown.


at Hacettepe U



ownloaded from

research groups (Hunt et al., 2005; Apps et al., 2008a). Initially, theonly antibody available for detection of HLA-G was W6/32, whichalso binds to other HLA class I molecules and only when these arein complex with b2m. This means that there were no methods to dis-tinguish between membrane-bound and soluble isoforms of HLA-G,and splice variants other than the full-length HLA-G1 and -G5 couldnot be detected due to the lack of association with b2m.

The soluble HLA-G5 variant has recently been thoroughly investi-gated. Ishitani et al. (2003) showed that all trophoblast cell popula-tions, including VT cells and ST cells secrete sHLA-G. This studywas carried out using three different antibodies: 87G detecting bothsoluble and membrane-bound forms, 16G1, which only binds thesoluble isoforms and o1G capturing solely the membrane-boundforms. The specicities of 87G and o1G, respectively, have been veri-ed, but problems such as lack of reactivity and non-specic bindinghave been reported when using the antibody 16G1 (Pascale et al.,2000b). Another study also detected HLA-G5 secreted from VTcells using the same antibody 16G1, and MEM-G/9, which detectsboth membrane-bound HLA-G1 and soluble HLA-G5, and has well-documented specicity. To make sure that the HLA-G observedwas in fact the HLA-G5 isoform, ow cytometry with MEM-G/9was carried out and because no surface expression could be detectedthe protein was stated to be HLA-G5 (Solier et al., 2002).

A null allele G*01:05N with a one base deletion mutaton in exon 3has been detected, which results in a shortened transcript lacking theability to produce HLA-G1 and HLA-G5. This null allele seems to be arisk factor in RM, but G*01:05N homozygotic adult individuals havebeen reported, proving that full-length HLA-G is not absolutely neces-sary for fetal survival (Suarez et al., 1997; Ober et al., 1998). There-fore, recent studies have focused on additional splice variants ofHLA-G, led by the hypothesis that in the case of HLA-G1 and -G5 de-ciency, other isoforms might carry out immunological functions andcompensate for the lack of these molecules. Morales et al. (2003)demonstrated, in addition to the expression of HLA-G5 in manytrophoblast cell populations, the expression of HLA-G2/HLA-G6 inEVT cells. This study used the antibodies 1-2C3 for detection ofHLA-G5, and 26-2H11 for detection of HLA-G2 and HLA-G6, andto date, no other group has veried these ndings.

Overall, interest in detecting alternative forms of HLA-G hasincreased recently, even though knowledge about these moleculeshas existed for many years. Ishitani and Geraghty described that in add-ition to the conventional heterotrimer and the several splice variants,HLA-G could also be expressed as a disulde-linked-b2m-associateddimer, a b2m-free heavy chain and a b2m-free disulde dimer(Fig. 2B; Ishitani and Geraghty, 1992). Currently transfection studiesexamining these molecules show contradictory results including detec-tion of both b2m-associated dimers and b2m-free monomers, dimersand trimers on the cell surface (Gonen-Gross et al., 2005; Apps et al.,2007). Furthermore a study on isolated VT cells found that solubleHLA-G5 was able to perform disulde-linked-b2m-free dimerization.This study has not been veried by any other group, and the capabilityof sHLA-G to perform dimerization in vivo remains questionable(Morales et al., 2007). In conclusion, it is still unclear which HLA-Gisoform is dominantly expressed in EVT cells and other trophoblastcell populations. Also, whether alternative HLA-G molecules serve im-portant functions in the placenta and the decidua is questionable, andfurther investigations need to be carried out.

Functions of HLA-G duringpregnancyOf special interest, regarding the functions of HLA-G, is the interactionwith the inhibitory receptors, immunoglobulin-like transcript-2 and -4(ILT-2 and ILT-4), which is expressed by a wide variety of immunecells and selectively on the surface of monocytes, macrophages anddendritic cells, respectively (Colonna et al., 1997, 1998). It has beenshown that the binding afnity of these two receptors for HLA-G ishigher than for other HLA class I molecules, indicating specic func-tions as a result of the interactions (Fig. 1; Shiroishi et al., 2003).Both these receptors have been shown to compete with CD8 inbinding HLA-G, which could prevent activation of cytotoxic CD8+

T cells by preventing an interaction between HLA-G and CD8,thereby contributing to the induction of tolerance (Shiroishi et al.,2003). Furthermore, HLA-G has been shown to up-regulate bothILT-2 and ILT-4, along with the killer-cell immunoglobulin-likereceptor-2DL4 (KIR2DL4), on the surface of both antigen presentingcells, NK cells, and CD4+ T cells without preceding antigenicco-stimulation. This provides an interesting possibility that thesereceptors function at the materno-fetal interface by raising the thresh-old of the maternal immune system activation, thereby serving toinduce tolerance (LeMaoult et al., 2005). In continuation of this,both soluble and membrane-bound HLA-G proteins have beenshown to induce inhibition of T cell alloproliferation through bothILT-2 and ILT-4. This was validated by the abolishment of inhibitionwhen adding antibodies against these receptors (Naji et al., 2007).

The signicance of HLA-G dimerizationfor receptor interactionsExperiments have shown that the HLA-G dimer has higher afnity forILT-2 and ILT-4 than the HLA-G monomer has, and this has drawnfocus towards which effects dimerization might have (Shiroishi et al.,2006). A transfection study stated that sHLA-G dimers inhibited allor-ecognition by reducing proliferation of both CD4+ and CD8+ T cells,and that this was accompanied by an up-regulation of ILT2 on theCD8+ cells, probably leading to increased inhibition (Zhang et al.,2009). Apps et al. recently found that homodimeric complexes ofHLA-G associated with b2m binds to ILT-2 expressed on all decidualdendritic cells and macrophages. This interaction results in alteredcytokine expression with increased levels of IL-6 and IL-10, indicatinga shift towards the Th2 cytokine prole. In addition, they found thatthe interaction resulted in an inhibition of the indirect T cell prolifer-ation carried out by the antigen presenting cells, as these encounterednon-self tissue antigens (Apps et al., 2007). As free heavy chains ofHLA-G have been detected on the surface of trophoblast cells,these have also been examined, and the necessity of b2m associationfor the interaction with ILT-2 has been conrmed (Gonen-Gross et al.,2005). Conversely, ILT-4 has been shown to interact with bothb2m-associated and free HLA-G (Shiroishi et al., 2006).

Specic interactions of HLA-G with ILT-2and ILT-4In addition, transfection studies have shown that ILT-2 interact withHLA-G1 and prevent lysis by NK cells. By adding antibodies againstthe ILT-2 receptor, lysis was restored, providing evidence that a

96 Dahl and Hviid

at Hacettepe U



ownloaded from

specic interaction between these two molecules can inhibit cell lysis(Navarro et al., 1999). Recently, it has also been shown that HLA-Gdown-regulates IFN-g produced by NK cells as a result of inter-action with ILT-2, but under certain conditions, up-regulation ofthis cytokine is observed as well, indicating the necessity of tightregulation (Morel and Bellon, 2008). A recent study proved thatthe interaction between HLA-G and ILT-2 resulted in inhibition ofNK cell cytolytic functions and impairment of the release of IFN-g(Favier et al., 2010). However, it is still unclear whether uNK cellsactually do display cytotoxicity in vivo, which makes the signicanceof these ndings difcult to evaluate. At this point it is more likelythat uNK cells are cytokine-producing upon activation rather thancytotoxic. Whether cytotoxicity to trophoblast cells carried out byperipheral NK cells during the invasion of the trophoblast into thedecidua could interfere with the generation of tolerance is stillunknown.

The effects of HLA-G interaction with ILT-4 on the surface ofdendritic cells have also been an area of focus. A study by Groset al. showed that HLA-DR and CD80 were reduced in dendriticcells in the presence of sHLA-G, and in addition, a reduction ofIL-12 secretion was observed. This resulted in a decreased activationof NK cells, pointing towards a role for HLA-G in inhibiting dendriticcells and preventing activation of effector cells in the immunesystem (Gros et al., 2008). Recently, IL-10 dependent dendritic cellsDC-10 has been shown to induce differentiation of naive CD4+ Tcells into type 1 regulatory T cells, as a result of the interactionbetween ILT-4 and HLA-G (Gregori et al., 2010). These Tr1 cellsproduce high levels of IL-10, transforming growth factor (TGF)-band IL-5, low levels of IFN-g and IL-2, and no IL-4, and have shownto actively suppress immune responses mainly via IL-10 and TGF-b(Groux et al., 1997).

Specic interactions between HLA-Gand KIR2DL4An additional receptor for HLA-G is KIR2DL4, which has been shownto bind HLA-G both in transfection studies and when expressed onthe surface of JEG-3 cells (Rajagopalan and Long, 1999; Yu et al.,2006). The receptor is present on a wide subset of NK cells, especiallyin decidua, and seems to be HLA-G specic, indicating that the inter-action is immunologically important during pregnancy (Ponte et al.,1999). Giving this specicity, determination of the fundamental signal-ling during pregnancy seemed one step closer, but one study foundthat a woman with a history of several successful pregnancies washomozygous for a genotype not encoding KIR2DL4 (Gomez-Lozanoet al., 2003). Evidence indicates that this receptor might not onlyresult in inhibition of NK cells but has the ability to provide activatingsignals as well (Lanier et al., 1998). However, currently there is noknowledge on whether these activating signals take place in uNKcells when KIR2DL4 encounters HLA-G, and what the possibleoutcome of this activating signal might be, although secretion of cyto-kines has been suggested. Supporting this is a study showing that a cellline transfected with HLA-G could not reverse lysis carried out byuNK cells but did induce a release of cytokines. In addition, it wasshown that the protein level of KIR2DL4 in uNK cells isolated fromfertile controls was much higher than in uNK cells from womenexperiencing RM (Wei-hua et al., 2007).

Validation of the receptor interaction studiesMost of the studies have been carried out on peripheral immune cells,and therefore, it will be interesting to perform the same studies on Tcells from decidua and isolated trophoblast cells expressing HLA-G, toassess whether the interactions and functions described here occur atthe materno-fetal interface. Studies of receptor interactions in vivo aresparse, but a recent, remarkable study managed to knock downHLA-G in human rst trimester EVT cells by RNA interference. Theresults were intriguing, since NK cells showed considerable cytotox-icity to the HLA-G-knock-downed cells compared with thenon-knocked-down controls, indicating that the interactionsbetween HLA-G and the receptors ILT-2, ILT-4 and in additionKIR2DL4 on the surface of uNK cells are important for fetal survival(Chen et al., 2010).

HLA-G polymorphism and pregnancy successIn the HLA-G gene, several polymorphisms have been identied,which might have a role in the pathogenesis of infertility and RM.Studies have focused on polymorphisms in the coding region, in the5 upstream regulatory region (5URR) and in the 3 untranslatedregion (3UTR), respectively (Fig. 2B).

To date, 15 polymorphisms at the protein level and two null alleleshave been identied, and most studies have focused on the null alleleG*01:05N, which do not encode the full-length HLA-G1 or HLA-G5(http://www.HLA-alleles.org). Discovery of this allele led to specula-tions on whether fetal homozygosity of G*01:05N could be an etio-logical factor in couples experiencing RM and in unexplainedinfertility, but this was partially ruled out by the detection ofG*01:05N homozygous adults (Suarez et al., 1997; Ober et al.,1998). Despite this, several studies have found an increased frequencyof this allele in couples experiencing RM (Aldrich et al., 2001; Pfeifferet al., 2001), dening it as a risk factor. Surprisingly, another studyfailed to show this correlation, and conversely found an increased fre-quency of this allele in the population, challenging the postulation thatfull-length HLA-G is crucial for successful pregnancy (Abbas et al.,2004). The explanation behind this paradox might be found in theunique splice pattern of HLA-G, where other isoforms might compen-sate for the lack of HLA-G1 and HLA-G5. In support of this hypoth-esis, it has been shown that in cells transfected with G*01:05N,full-length primary transcript was detected, but at a decreased levelcompared with cells transfected with G*010102 (G*01:01:02:01/02). Furthermore, protein expression of the full-length HLA-G1 andHLA-G5 could not be detected in the G*01:05N transfected cellsand neither could HLA-G4. In contrast, expression of the membrane-bound HLA-G2 and -G3 and/or soluble HLA-G6 and G7 wasdetected, but could not be distinguished due to the lack of specicantibodies. Because inhibition of NK cell-mediated lysis was observedin these cells lacking full-length protein, evidence pointed towards thefact that other HLA-G isoforms than HLA-G1 and -G5 could providesufcient immunological protection (Le Discorde et al., 2005).

Another HLA-G null allele G*01:13N has been described, in whichexpression of all splice variants is compromised. This allele has onlybeen found in heterozygous individuals and the frequency in the popu-lation seems to be limited (Mendes-Junior et al., 2010).

An overview of the studies examining the association of HLA-Galleles and RM is shown in Table I. Some groups nd no association


at Hacettepe U



ownloaded from

at all (Penzes et al., 1999; Yamashita et al., 1999; Yan et al., 2006a),and others report an increased frequency of G*010103(G*01:01:03:01/02) (Pfeiffer et al., 2001; Abbas et al., 2004). In astudy by Aldrich et al., it was shown that in RM couples who continu-ously failed to achieve pregnancy, G*0104 (G*01:04:01/02/03/04/05) and G*0105N (G*01:05N) were more frequent than in RMcouples, who became pregnant at last (Aldrich et al., 2001). In add-ition, an increased frequency of G*0106 (G*01:06) in RM patientshas been reported, but not at a signicant level (Hviid et al., 2002).Taken together, there is no clear picture on whether presence ofany of the alleles besides the G*01:05N null allele possesses anincreased risk for RM.

Studies examining polymorphisms in the non-coding regions of theHLA-G gene are shown in Table II. In the 5URR, Ober et al. (2003)found an increased frequency of a 2725G variation in couples withsporadic fetal loss. Currently, this nding has not been reproducedin other studies. Furthermore, it cannot be ruled out that the2725G variant might be linked to other polymorphisms in theHLA-G gene region associated with pregnancy complications, provid-ing an explanation for these ndings.

Of interest is also the 14 bp del/ins polymorphism in the 3UTR ofthe HLA-G gene (Harrison et al., 1993). Several studies have observeda trend that more RM women than normal fertile women are homo-zygous for the 14 bp insertion sequence, although in some studies thiscorrelation is not signicant (Hviid et al., 2002, 2004a; Yan et al.,2006b; Zhu et al., 2010). Other studies have found an increasednumber of heterozygotic individuals among RM women (Tripathiet al., 2004; Xue et al., 2007), and yet others have reported no cor-relation at all (Sipak-Szmigiel et al., 2008; Suryanarayana et al.,2008). However, the latter study reported an increased number ofindividuals in the RM group carrying both the 14 bp insertion sequenceand a novel SNP in exon 8, a T C mutation at position 1570. Com-parison and evaluation of these studies are compromised by the ethnicdifferences in distribution of the polymorphism, and possible linkagedisequilibrium with other HLA variants that could contribute to theaetiology. The presence of the 14 bp polymorphism has been

correlated with reduced levels of HLA-G mRNA, possibly providingan explanation for the increased frequency in women with RM(Hviid et al., 2003). This study also demonstrated an association ofthe 14 bp insertion sequence with a deletion of 92 bp of exon 8. Ithas been shown that deletion of the 92 bp is associated with higherstability of HLA-G mRNA than the +14 bp transcripts (Rousseauet al., 2003). To date it is unknown whether deletion of the 92 bpresults in a lower stability of the transcript, than deletion of the14 bp sequence.

Soluble HLA-G: a possiblebiomarker in humanreproduction

Soluble HLA-G during pregnancySoluble HLA-G has been detected in blood plasma and amniotic uidsand as a consequence, much effort has been put into developingmethods to determine the concentration of sHLA-G in various bodyuids, and to understand what kind of functions this immunomodula-tory molecule might have in these compartments (Rebmann et al.,1999).

An overview of studies linking sHLA-G in blood plasma to preg-nancy success is listed in Table III. Early studies used the antibodyW6/32, and could not distinguish sHLA-G from other soluble HLAclass I proteins, such as HLA-A, -B, -C and -E, making evaluationsof these studies difcult. Yet, sera from 20 women experiencingfetal loss of unknown aetiology were shown to contain lower levelsof sHLA-G than sera from women in rst trimester of intact pregnan-cies (Athanassakis et al., 1999). By using an antibody specic for thesHLA-G isoforms, 16G1, Hunt et al. (2000) tried to eliminate thecross reactivity, and showed that levels of sHLA-G were approximate-ly twice as high in pregnant women as in non-pregnant women. Toelucidate the possible immunological functions of sHLA-G in bloodplasma during pregnancy, Pfeiffer et al. examined infertile women

.............................................................................................................................................................................................

Table I Association between HLA-G polymorphisms in the coding regions and RM.

Study Number ofcases

Inclusion criteria forcases

Number ofcontrols

Inclusion criteria forcontrols

Results

Penzes et al. (1999) 21 couples RM - not specied 72 individuals Healthy individuals No association

Yamashita et al.(1999)

20 couples RM - not specied 54 individuals Not informed No association

Pfeiffer et al. (2001) 78 couples 3 RMs 52 couples 1 normal pregnancy G*010103 (G*01:01:03:01/02)No history of fertilityproblems

G*0105N (G*01:05N)

Aldrich et al. (2001) 113 couples 3RMs No controls G*0104 (G*01:04:01/02/03/04/05)G*0105N (G*01:05N)

Hviid et al. (2002) 61 couples 3 RMs 47 fertile couples 1 normal pregnancy Trend for G*0106 (G*01:06)Abbas et al. (2004) 120 women 3 RMs 120 women 3 normal pregnancy Trend for G*010103

(G*01:01:03:01/02)No history of fertilityproblems

Yan et al. (2006a, b) 69 women 3 RMs 146 women 1 normal pregnancy No association

98 Dahl and Hviid

at Hacettepe U



ownloaded from

.............................................................................................................................................................................................

Table II Association between HLA-G polymorphisms in the non-coding regions and RM


Inclusion criteria for cases Number ofcontrols

Inclusion criteriafor controls

Results

2725 SNP in 5URR:

Ober et al. (2003) 42 huteritecouples

Association of sporadic fetal losswith 725G/2725G genotype

Sipak-Szmigiel et al.(2008)

58 couples 3 RMs 58 couples 2 normalpregnancies

No association

Roussev and Coulam(2007)

25 couples History of RM Not informed Not informed 2725G mutation

14 bp ins/del polymorphism

Hviid et al. (2002,2004a, b)

61 women 3 RMs 93 women 2 normalpregnancies

+14 bp/+14 bp genotype

No history of fertilityproblems

Tripathi et al. (2004) 120 women 3 RMs 120 women 3 normalpregnancies

214 bp/+14 bp genotype


Yan et al. (2006a, b) 79 women 3 RMs 109 women 2 normalpregnancies

Trend for +14 bp allele

Xue et al. (2007) 24 women 3 RMs 88 women One normalpregnancy

214 bp/+14 bp genotype


58 couples 3 RMs 58 couples 2 normalpregnancies

No association.

Suryanarayana et al.(2008)

169 women 3 RMs 92 women One normalpregnancy

Association with genotypesincluding the +14 bp allele and a1570T C mutation

Berger et al. (2010) 238 women 2 RMs 233 women One normalpregnancy

Trend for +14 bp/+14 bpgenotypeAssociation with two SNP denedhaplotypes

Zhu et al. (2010) 188 women 3 RMs (Additionally 138women with 2 RMs)

251 women 1 normal pregnancy Trend for +14 bp/+14 bpgenotype. Association only inwomen 4 RMs


.............................................................................................................................................................................................

Table III Association between soluble HLA-G levels in blood plasma and pregnancy success.

Study Method Number ofcases

Inclusion criteria forcases

Results

Athanassakis et al.(1999)

Detection of HLA-G with W6/32 20 women Fetal loss of unknownaetiology

Decreased level of sHLA-G in womenwith fetal loss compared with controls

Pfeiffer et al. (2000) Depletion of HLA class Ia and HLA-Ewith mAb TP25.99

65 women Undergoing IVFtreatment

Pre-ovulatory sHLA-G levels and sHLA-Glevels in early pregnancy lower in womenwith fetal loss, than in women with intactpregnancies

Detection of HLA-G with W6/32 andanti-b2m

Hviid et al. (2004a, b) Detection with MEM-G/9 and W6/32 85 individuals Undergoing IVFtreatment

No detectable sHLA-G in individuals with:

+14 bp/+14 bp genotype 14 bp/214 bp genotype and 2725G

mutation


Not described 20 women 3 consecutive IVFfailures

Deceased level of sHLA-G in women withIVF failure compared with controls


at Hacettepe U



ownloaded from

undergoing IVF treatment in a pilot study. They showed that pre-ovulatory levels of sHLA-G in 20 IVF treated women with subsequentmiscarriages were much lower than in 45 women with intact pregnan-cies (Pfeiffer et al., 2000). A difference in sHLA-G levels between thewomen with intact pregnancies and those experiencing early miscar-riages was also observed during the following 9 weeks of gestation, in-dicating that higher levels of sHLA-G might be correlated withsuccessful treatment in couples with fertility problems. Anotherstudy found that in subjects homozygous for the +14 bp insertion se-quence, which has also been associated with RM, no sHLA-G in bloodplasma could be detected. In addition, the 2725G polymorphism inthe 5URR accompanied by two 14 bp deletion alleles also resultedin lack of detectable sHLA-G levels in blood plasma (Hviid et al.,2004b). Here, MEM-G/9 and W6/32 were used. These antibodieshave been stated as the most specic combination in detectingsHLA-G, when a possibility of cross reactivity with other HLA classI molecules is present (Sageshima et al., 2007). Finally, anotherstudy reported that levels of sHLA-G in peripheral blood were signi-cantly lower in women having experienced at least three IVF failuresthan in fertile controls. No signicant reduction was seen in womenexperiencing RM, although absolute numbers pointed towards alower concentration of sHLA-G in this group as well (Sipak-Szmigielet al., 2007).

The concentrations of sHLA-G in maternal blood has been shownto increase signicantly during the rst trimester of pregnancy, andfrom this point, levels decline during second and especially third tri-mester based upon single measurements on blood samples from dif-ferent groups of pregnant women (Steinborn et al., 2007). Yet, thisexpression pattern is similar to that of other biomarkers during preg-nancy, and therefore does not necessarily reect specic functions ofsHLA-G during different stages of pregnancy. In contrast to this study,Rizzo et al. managed to discern detection of the genuine HLA-G5 andthe sHLA-G1, which can be shed from the cell membrane. This wasobtained by detecting both isoforms with the specic MEM-G/9 anti-body and selectively capturing HLA-G5 with the intron-4 specic5A6G7 antibody, and subsequently determine the amount ofsHLA-G1 as the difference between these two values (Rizzo et al.,2009). The study found that high or detectable values of sHLA-Gare not fundamental for pregnancy, since sHLA-G could not bedetected in maternal plasma from all women with normal, healthypregnancies. However, a lower level of sHLA-G1 in late pregnancywas linked to an increased risk of pre-eclampsia, while high levels ofHLA-G5 seemed to be associated with more severe pre-eclampsiaand other pregnancy complications such as intrauterine growth retard-ation. This will not be discussed further in this review, but provides thebasis for carrying out similar experiments in women with a history ofRM and women undergoing infertility treatment.

The detection of sHLA-G in plasma also led to the question ofwhether antibodies against HLA-G were secreted into the blood ofpregnant women, and whether these could provide immunologicalchallenges in subsequent pregnancies if preformed antibodies boundto sHLA-G prevent the protein from carrying out fundamental im-munological functions in the peripheral blood of these pregnantwomen. Hunt et al. showed that antibodies against sHLA-G couldbe detected in sera from some multigravid women, but not fromwomen who never had children, or from men. Yet, the six women,who were found to secrete anti-HLA-G antibodies all delivered

healthy babies, leading to the conclusion that any production ofanti-HLA-G antibodies cannot provide explanation for any of thepregnancy complications discussed here (Hunt et al., 2003).

The use of soluble HLA-G in assistedreproductionAs it has become evident that at least some trophoblast cell popula-tions secrete sHLA-G, much effort has been put into determining howearly in fetal development it is possible to detect this protein. In par-ticular, many studies have tried to link the detection of sHLA-G to ob-tainment of clinical pregnancy in assisted reproductive treatments(ARTs), as summarized in Table IV. From this, it is clear that severalstudies have correlated detection of sHLA-G in oocyte culture super-natant to successful pregnancies in ART procedures such as ICSI andIVF. While some studies report that no clinical pregnancies wereachieved from negative sHLA-G cultures at all (Fuzzi et al., 2002;Noci et al., 2005), others simply nd that pregnancy rate is decreasedwhen using these embryos compared with embryos from sHLA-Gpositive cultures (Sher et al., 2004; Yie et al., 2005; Desai et al.,2006). Despite the great number of reports, questions have beenraised on whether the detection of sHLA-G is an artefact (Menezoet al., 2006). A problem posed here, is that no standardized recombin-ant sHLA-G protein has been available, and therefore, studies areforced to rely on estimations of protein levels detected. In a studyby Rebmann et al., a puried sHLA-G standard for measuringprotein levels was used, and it was only possible to detect sHLA-Gin 20% of the embryo cultures. In addition, it was found thatsHLA-G correlated well with obtainment of clinical pregnancy afterIVF treatment and especially after ICSI (Rebmann et al., 2007). Inthe search for optimal experimental conditions, including the bestcombination of antibodies for ELISA, a study by Sageshima et al.(2007) failed to detect sHLA-G in any culture supernatants of fertilizedeggs at Day 23 and Day 46. However, it was reported that some ofthe supernatants showed intensities between the background and thecut-off value of 1 ng/ml, and therefore it was concluded that a farmore sensitive ELISA method is needed to conrm production ofsHLA-G in early embryos. Another study found that only in one ofthree fertility treatment centres, sHLA-G could be associated withbetter implantation rates (Tabiasco et al., 2009). At this centre, onlyICSI and not IVF treatment was carried out, supporting the work byRebmann et al. described above. Since a higher number of sHLA-Gpositive embryo culture supernatants were found after IVF thanafter ICSI, and implantation rates were generally higher after ICSIthan after IVF the study concluded that sHLA-G cannot predict preg-nancy success alone, but multiple parameters such as type of ARTprocedure and oocyte culture medium also plays an important role.However, the detection level of this assay was evaluated to be 7+2 ng/ml, and evidence indicates that lower detection levels areneeded in order for sHLA-G to be a good predictor of pregnancysuccess. In addition, this study only measured implantation rates,whereas other studies examined the obtainment of clinical pregnan-cies, making comparison difcult. Overall, it is unlikely, that sHLA-Galone could be responsible for the obtainment and maintenance ofpregnancy but it is more likely a factor among several that might beimportant. A study by Shaikly et al. analysed the sHLA-G expressionin early embryos and showed that sHLA-G expression was variable,

100 Dahl and Hviid

at Hacettepe U



ownloaded from

but mainly localized to the trophoectoderm layer and projections fromthis, indicating a role during implantation, where fetal and maternaltissues are in close apposition (Shaikly et al., 2008). Evidence pointstowards the fact that sHLA-G can be one of the useful tools whenselecting embryos for transfer in fertility procedures, and for thisreason resources are put into developing ELISA methods of higherspecicity and sensitivity for the future goal of sHLA-G serving as anon-invasive biomarker, increasing the chances of successful pregnan-cies in couples undergoing ART procedures. A recent meta-analysisstates that in samples of good embryo quality, detection of sHLA-Gis a much better prognostic factor for obtaining clinical pregnancythan is the case when examining embryos of mixed quality (Vercam-men et al., 2008). A German multicenter study supports this nding,showing that morphological grading is the most powerful predictorof pregnancy, but that detection of sHLA-G is also a powerful tool,with better prognostic value than factors such as age, prior ARTsand number of embryos transferred (Rebmann et al., 2010).

The presence of HLA-G in the malereproductive systemDue to the ndings that sHLA-G can be detected in early embryos,studies have been performed to examine whether HLA-G couldhave paternal origin and serve important functions during fertilizationand implantation. HLA-G has been detected in human prostatictissue in one study, and even though HLA-G1 mRNA was the most

prominent transcript found, immunohistochemistry only revealed ex-pression of HLA-G5 using the antibody 1-2C3 (Langat et al., 2006).On this basis, interest has also been directed towards whetherseminal plasma could contain sHLA-G protein. There is evidencethat sHLA-G is found in seminal plasma and that large individual vari-ation in sHLA-G levels exist (Larsen et al., 2011). This study also foundHLA-G expression in normal testis and epididymal tissues, but it is stillunclear where in the male reproductive system, the sHLA-G detectedin seminal plasma originates from. However, detection of the proteinin this uid indicates that sHLA-G may serve immunoregulatory rolesduring fertilization and implantation. In the future, it will be interestingto determine sHLA-G levels in seminal plasma from men with fertilityproblems, to observe whether there is an association betweensHLA-G concentrations in seminal plasma and pregnancy outcomein couples with unexplained infertility and RM.

Regarding possible functions carried out by sHLA-G, a study bySelmani et al. showed that HLA-G5 was required to expand the popu-lation of CD25+CD4+FoxP3 cells, and that lymphocytes subsequentlyexhibited hypo-responsiveness to allogenic stimuli, supporting the hy-pothesis that HLA-G5 induces tolerance through these regulatory Tcells (Selmani et al., 2008). In addition, it has been observed that re-combinant HLA-G5 and HLA-G6 molecules are able to induce a re-duction in IL-10 secretion and a signicant increase in TGF-b1production in myelomonocytic cells (Mclntire et al., 2004). This sup-ports that sHLA-G present in seminal plasma could alter the immuno-logical functions carried out by macrophages, leading to a decreased

.............................................................................................................................................................................................

Table IV Association between soluble HLA-G levels in early embryo cultures and pregnancy success.

Study Antibodies fordetection ofsHLA-G

Numberof patients

Procedure Overallpregnancyrate (%)

Pregnancy ratein sHLA-Gpositive embryocultures

Pregnancy ratein sHLA-Gnegativeembryo cultures

Signicantassociation

Fuzzi et al.(2002)

MEM-G/9 andW6/32

101 IVF/ICSI 18 24% 0% Yes (P 0.0026)

Sher et al.(2004)

MEM-G/1 andW6/32

201 ICSI 51 Women ,39 years:71%; Women .39years: 52%

Women ,39 years:22%; Women .39years: 15%

Yes; (,39 years:P, 0.0001); (.39years: P, 0.003)

Noci et al.(2005)

MEM-G/9 andW6/32

66 IVF/ICSI 14 23% 0% Slight (P 0.045)

Yie et al.(2005)

4H84 and 3C/G4 137 IVF 37 48% 17% Yes (P 0.0026)

Desai et al.(2006)

Not informed 83 ICSI 55 64% 36% Yes (P, 0.05)

Rebmannet al.(2007)

MEM-G/9 andanti-b2m

313 IVF/ICSI 25 40% 19% (P, 0.001). Yet,after ICSI(P, 0.0001)

After ICSI: 42% After ICSI: 18%

Tabiascoet al.(2009)

MEM-G/9 andW6/32

355 IVF/ICSI 33 34% 19% Not overall, onlyimplantation rate inone centre(P 0.034)(Only ICSIperformed here)

Rebmannet al.(2010)

MEM-G/9 andanti-b2m.

2364 IVF/ICSI (36TESE/96 notspecied)

31 41% 26% Yes (P, 0.001)

IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; TESE, testicular sperm extraction.


at Hacettepe U



ownloaded from

immunological response. In addition, TGF-b has been shown toinduce differentiation of naive CD25+CD4+ T cells into regulatoryT cells expressing FoxP3 molecules, to induce proliferation of theseregulatory T cells and to inhibit target cell lysis carried out by NKcells through these regulatory T cells, leading to the development oftolerance (Chen et al., 2003; Ghiringhelli et al., 2005a, b). In clinicalstudies, it has been shown that women experiencing RM have areduced number of CD25+CD4+FoxP3 cells in peripheral blood,and in addition, these cells were shown to be functionally decient,as a higher number of cells from these women were required toinduce suppression of the immune response (Arruvito et al., 2007).In addition, expression of FoxP3 transcripts in endometrial tissuewas reduced in a group of women with unexplained primary infertilitycompared with control subjects (Jasper et al., 2006). In a recent study,RM women were shown to have lower amounts of IL-2-inducibleFoxP3 regulatory T cells in peripheral blood (Arruvito et al., 2010).Furthermore, RM women proved to have reduced amounts of IL-2and TGF-b in sera compared with controls, but elevated levels ofIL-6. The study showed that induction of these regulatory T cellswere carried out through IL-2 signalling and STAT-5 leading to expres-sion of FoxP3, and stimulation with IL-2 resulted in a much lowerFoxP3 expression in RM women than in fertile controls, indicating adeciency in this signalling pathway. A reduced immune suppression,in agreement with previous observations, was also detected (Arruvitoet al., 2010). In summary, evidence point towards a complex interplaybetween both decidual and peripheral immune cells and the cytokinessecreted, and further studies are needed to obtain a more conciseunderstanding of these interactions.

HLA-E alleles and expressionin relation to pregnancyTo assess whether HLA-E is important during pregnancy, studies havefocused on the expression of this molecule at the materno-fetal inter-face. Both mRNA and protein products have been identied in avariety of cells, but surface expression of the protein is restricted tocells that express other HLA class I molecules able to provide anonamer signalling sequence needed for directing the protein to themembrane, as shown in Fig. 2C (Lee et al., 1998a). Using rst trimes-ter placental and decidual tissues, King et al. (2000) found that tropho-blast cells express HLA-E on the cell surface. Subsequently, Ishitaniet al. (2003) found that HLA-E protein could be detected in all cellpopulations expressing either membrane-bound or soluble HLA-G in-cluding EVT, VT and ST cells. Another group only found HLA-Elocated intracellularly in VT cells, and only sporadic cytoplasmic ex-pression of the protein in ST. In addition, HLA-E was only detectedin some of the EVT cells, and this expression also seemed to belimited to the cytoplasm (Bhalla et al., 2006). This study also triedto identify a difference between placental expression of HLA-E inwomen with RM and fertile controls, but failed to show a signicantcorrelation. Recent work has conrmed the expression of HLA-E onthe surface of EVT cells but has failed to identify the protein in anyother subpopulation at the materno-fetal interface (Apps et al., 2009).

Lately, a study by Shaikly et al. (2010) has shown that HLA-E isco-expressed with HLA-G at the surface of the trophoectoderm ofDay 6 blastocyst preimplantation embryos, suggesting that both

these molecules are important during implantation. However, thesurface expression of HLA-E needs to be veried by other groupsto validate this nding.

A few studies have also focused on the expression of HLA-E in themale reproductive system, but have only managed to identify tran-scripts in spermatocytes, spermatides and most recently in ejaculatedsperm, whereas no HLA-E protein could be detected (Guillaudeuxet al., 1996; Avendano et al., 2009). The latter study showed thatthe level of HLA-E mRNA was higher in a group of fertile men thanin a group of men undergoing consultation for infertility. Furthermore,the mRNA products were detectable 24 h after ICSI, possibly indicat-ing that the mRNA products may have a function during early embryodevelopment.

To date, nine different alleles of the HLA-E gene have been described(http://www.HLA.allels.org). Only three variants can be distinguishedat the protein level, and several groups have examined the frequencyof these in different populations. This has led to the conclusion thatthe E*0104 (E*01:04) allele is very rare, and therefore, only twoalleles are of practical importance: E*0101 (E*01:01:01:01/02/03)and E*0103 (E*01:03:01:01/02 or E*01:03:02:01 or E*01:03:03/04).These two alleles are separated on the basis of a non-synonymous sub-stitution at codon 107 from arginine (HLA-ER) to glycine (HLA-EG)(Ohya et al., 1990; Geraghty et al., 1992; Gomez-Casado et al., 1997;Grimsley andOber, 1997). Several studies have focused on the distribu-tion of these two alleles in women experiencing RM (Table V). Only onestudy reported that the frequency of HLA-ER was signicantly higherin women with RM than in fertile controls. The correlation was particu-larly strong in HLA-ER homozygous RM women (Tripathi et al., 2006).Pfeiffer et al. (2001) also reported an increased frequency of theHLA-ER variant in RM women, but the difference was not signicant.Other studies however, failed to show this correlation but differed inselection of control individuals and by inclusion criteria for RMwomen, as seen in Table V (Steffensen et al., 1998; Kanai et al., 2001).

Providing a possible explanation for the increased frequency ofHLA-ER in RM women is the functional differences at the proteinlevel between this allele and HLA-EG. Strong et al. (2003) found,using ow cytometry, that the surface expression of HLA-E wasmuch lower in cells transfected with HLA-ER than in cells transfectedwith HLA-EG, while the overall protein level did not differ. Becausesurface expression of HLA-E depends on a nonamer sequence, asdescribed previously, peptide afnities of HLA-EG and HLA-ER havebeen compared. Using ELISA, it was shown that peptide afnity fora nonamer sequence derived from HLA-G was higher for HLA-EG

than for HLA-ER. This correlated with the nding that thermal stabilitywas higher in HLA-EG, and supported the hypothesis of a decreasedsurface expression on cells from individuals expressing the HLA-ER

allele (Strong et al., 2003). Overall, this shows that decreasedsurface expression of HLA-E in persons carrying the HLA-ER variantcould be a risk factor in the aetiology of RM, indicating that HLA-Ehas an important role in the maintenance of pregnancy.

However, at present there is no clear evidence that HLA-E isinvolved in the pathogenesis of RM, and further studies that includea higher number of patients are needed to clarify whether or notthe HLA-ER variant occurs with an increased frequency in RMwomen. Furthermore, a consensus needs to be reached about the in-clusion criteria used in these studies and the selection of controlindividuals.

102 Dahl and Hviid

at Hacettepe U



ownloaded from

HLA-E in regulating NK cellsduring pregnancyThe possible role of HLA-E as an immunomodulator during pregnancycame into view, as it was shown that cells transfected with HLA-Ewere able to inhibit cytolysis carried out by NK cells through theCD94/NKG2A receptor (Fig. 1; Lee et al., 1998b; Borrego et al.,2002). This receptor is expressed by the CD56bright uNK cells andthe inhibition requires surface expression of HLA-E and thereforealso expression of another HLA class I molecule (Verma et al.,1997; Ashkar et al., 2003). Because HLA-G, and to a lesser extentHLA-C, are expressed on trophoblast cells, studies have focused onthe co-expression of HLA-E with these two molecules (Moffett-Kinget al., 2002).

Performing cytotoxicity experiments, King et al. showed that anti-bodies against HLA-G could partially restore lysis, while this was notthe case with antibodies against HLA-C, indicating that HLA-G isthe most important molecule in providing the nonamer sequenceneeded for surface expression of HLA-E. Yet, these results wereobtained from transfectant cell lines, and the study failed to demon-strate lysis of trophoblast cells carried out by uNK cells (King et al.,2000). While the CD94/NKG2A receptor was clearly shown to beinvolved in inhibition of cytotoxicity, it was shown that the HLA-E/G-nonamer peptide induced cytotoxicity very efciently through theCD94/NKG2C receptor (Llano et al., 1998). However, whetherthis interaction takes place between HLA-E on trophoblast cells andthe NKG2C receptor on uNK cells is still unknown. The inductionof cytotoxicity through NKG2C might seem conicting with the ideaof HLA-E as an immunosuppressive molecule during pregnancy, andbecause no specic antibodies for the receptors have been availableuntil recently, this paradox remains to be elucidated. Yet, severalstudies provide possible explanations. First, it has been reportedthat the inhibitory NKG2A receptor has much higher afnity forHLA-E than the activating NKG2C receptor, which could accountfor the overall response being inhibitory (Vales-Gomez et al., 1999).

Second, it has been shown that 98% of CD56bright uNK cellsexpress the NKG2A receptor, while only 20% express the NKG2C re-ceptor. In addition, all CD56bright uNK cells in the uterus, that expressNKG2C also express NKG2A, in contrast to peripheral NK cells,where no simultaneous expression is observed, indicating that NKcells in decidua might have special regulatory functions on this basis(Kusumi et al., 2006). Third, it has been shown that NKG2A andNKG2C compete for binding to CD94 (Labonte and Letvin, 2004),and therefore also for surface expression, and this balance could bea reection of the total lytic activity. In addition, the function ofCD56bright uNK cells in general might be cytokine secretion and regu-lation rather than cytotoxicity, since these cells have been shown to bethe main cytokine-producing cells in the decidua (Cooper et al., 2001).

Peptide substitutions at P5, P6 and P8 of the HLA-G nonamer se-quence have been shown to impair recognition by HLA-E and there-fore also impair surface expression (Miller et al., 2003; Hoare et al.,2008). On this basis, it would be interesting to try to identify possibleamino acid substitutions in the nonamer sequence provided by HLA-Gin couples experiencing fertility problems, to assess whether the inhib-ition of cytotoxicity through the interaction of CD94/NKG2A on uNKcells and HLA-E on trophoblast cells is a fundamental factor in obtain-ing pregnancy.

Ambiguity of the basic roleof HLA-F in reproductionHLA-F was discovered in 1990 and remains the least studied of theHLA class Ib molecules (Geraghty et al., 1990). This gene also exhibitslow polymorphism, and to date only four variants at the protein levelhave been identied (http://www.HLA.alleles.org). In addition, a non-sense mutation in exon 3, resulting in a HLA-F null allele has beenfound, but only in heterozygous individuals, and no functional import-ance of this allele has yet been reported (Uchigiri et al., 1997). Further-more, a synonymous mutation has been found in exon 2 but since this

.............................................................................................................................................................................................

Table V Association between HLA-E allelic variants and RM.


Inclusioncriteria forcases

Number ofcontrols

Inclusion criteria forcontrols

Results

Steffensenet al. (1998)

82 women 5 RMs 150 healtycontrols

None specied No association

Kanai et al.(2001)

30 couples 3 RMs 38 couples One normal pregnancy No association

Pfeiffer et al.(2001)

78 couples 3 RMs 52 healthycontrols

One normal pregnancy No association, but in women with 5 RMs:Trend for E*0101 and E*0102*(E*01:01:01:01/02/03)


Tripathi et al.(2006)

120 women 3 RMs 120 women 3 normal pregnancies E*0101 (E*01:01:01:01/02/03) (P 0.043)No history of fertilityproblems

Bhalla et al.(2006)

45 placentas fromRM women

Not informed 17 gestationmatched placentas

From women undergoingelective terminations ofpregnancy

No association in EVT

E*0102 has shown to be identical with E*01:01:01:01.


at Hacettepe U



ownloaded from

mutation does not result in any amino acid substitution, the functionalsignicance is questionable as well (Kunishima et al., 1999). At present,no polymorphisms in the HLA-F gene have been associated with RMor infertility. Systematical examination of the 21 detected variations inthe coding and non-coding regions would, however, be interesting tocarry out in these patients, as the studies described below pointtowards the fact that HLA-F might have a role in modulating theimmune system during pregnancy.

Regarding the expression of HLA-F, in vivo data are very contradict-ory. Ishitani et al. (2003) were the rst to detect surface expression ofthe protein in cells from term placenta. They observed a low expres-sion of HLA-F located intracellularly in ST, VT and EVT cells, that werenot in contact with maternal decidua, however, in the EVT cells em-bedded within the decidua, substantially higher levels of proteincould be detected. In contrast, Nagamatsu et al. were only able todetect HLA-F within EVT, ST and VT, but not on the surface ofthese cells. This difference could be due to the fact that they onlyexamined rst trimester placenta for surface expression, and nottissues from later stages of gestation (Nagamatsu et al., 2006). Subse-quently, Shobu et al. found that HLA-F was only expressed in the cyto-plasm of rst trimester EVT cells at low levels, but during the second,and particularly the third trimester, the protein level was increasedand surface expression observed. Furthermore, they found a weak ex-pression of HLA-F in VT and ST cells with no differences in expressionlevels during the course of gestation. By ow cytometry, it was shownthat only 16% of the cells expressed HLA-F on their surface (Shobuet al., 2006). Unexpectedly, Apps et al. examined decidual and placen-tal tissue at term and from elective terminations of normal pregnan-cies, and detected HLA-F mRNA, but failed to conrm any proteinexpression. They used a different antibody against HLA-F than theother groups, and they suggest that the weak expression of HLA-Fthat has been observed in previous studies was in fact not expressionby trophoblast cells, but possibly uNK cells (Apps et al., 2008b). It isapparent that the confusion about the expression of HLA-F at thematerno-fetal interface could be due to insufcient characterizationof the cells, or cross reactivity of some of the antibodies with other(non-classical) HLA molecules.

New aspects of the functionof HLA-F

Possible interactions of HLA-F with relevantreceptorsThe synthesis and posttranslational modications of HLA-F have beenshown to be different from those of other HLA molecules. As shownin Fig. 2C, it has been found that the cytoplasmic tail of HLA-F is es-sential for exportation from the endoplasmic reticulum (ER), and thatthe HLA-F unique RxR motif is responsible for localizing the protein tothe Golgia complex where it seems to accumulate (Boyle et al., 2006).Contrary to other HLA-class I molecules, HLA-F surface expression invitro seems to be independent of TAP and partially independent oftapasin, even though the molecules have the ability to bind peptidein the ER (Lepin et al., 2000; Wainwright et al., 2000; Lee and Ger-aghty, 2003). This indicates that surface expression of HLA-F mighttake alternative routes compared with other HLA class I molecules.

Characterization of possible receptor interactions of HLA-F havebeen sparse, but the protein has been demonstrated to interactwith ILT-2 and ILT-4 on the surface of monocytes and CD19+ Bcells, but not on CD56+ NK cells or CD3+ T cells (Fig. 1). The inter-action could only be inhibited partially by adding antibodies againstILT-2 and ILT-4, suggesting that HLA-F might also interact withother receptors on the surface of these cells. The tetramers used inthis study were prepared without peptide ligands, and at presentthere is no knowledge to whether specic peptides binding toHLA-F may induce surface expression or may result in interactionswith different sets of receptors, as it is the case for HLA-E (Lepinet al., 2000).

A novel role of HLA-F: possible implicationsfor pregnancyIn conclusion, the data on HLA-F are contradictory and while it seemslikely that the protein is expressed on the surface of cells at thematerno-fetal interface, a possible role in developing toleranceduring pregnancy is still in question. A recent important study byLee et al. (2010) used a different approach in trying to elucidate thefunctions of this protein. They examined the distribution of HLA-Fin lymphocytes and found that all resting B cells, T cells, NK cellsand monocytes express HLA-F within the cells, but upon activation,surface expression is induced, as shown by ow cytometry. An intri-guing fact is that surface expression is not up-regulated in the regula-tory T cells. This led the authors to hypothesize that HLA-F might beinvolved in regulating the immune response during pregnancy byserving as a marker of maternal activated lymphocytes, interactingwith the regulatory T cells making them secrete inhibitory cytokinesand induce suppressive signalling, which will result in the generationof tolerance (Fig. 1). In this regard, studies have shown that thenumber of regulatory T cells might be reduced in women experiencingspontaneous miscarriage compared with women with successful preg-nancies (Sasaki et al., 2004). It is mandatory to examine whetherHLA-F might bind to any receptors on the surface of these regulatoryT cells.

Signicance of HLA-CAlthough HLA-C belongs to the classical HLA class Ia molecules it hassimilarities with the non-classical HLA molecules in relation to preg-nancy success. HLA-C is a highly polymorphic molecule, which isexpressed ubiquitously like HLA-E.

HLA-C is expressed on the surface of EVT cells both asb2m-associated molecules and as free heavy chains. The expressionof HLA-C on EVT generally seems to be low, but can be up-regulatedby IFN-g (King et al., 1996, 2000). To date, no expression of the poly-morphic HLA-C in VT or ST cells has been reported.

Regarding receptor interactions, HLA-C has been shown to interactwith KIR2D receptors, which are expressed on the surface of uNKcells. Given the high degree of polymorphism of both KIR andHLA-C, many different receptor interactions are possible and willvary with each pregnancy. Basically, HLA-C molecules can bedivided into two groups regarding interaction with KIR receptors.HLA-C1 molecules have serine at codon 77 and aspargine at codon80 and include HLA-Cw1, -Cw3, -Cw7, -Cw8, -Cw12, -Cw14 and

104 Dahl and Hviid

at Hacettepe U



ownloaded from

-Cw1601. HLA-C2 molecules have aspargine at codon 77 and lysine atcodon 80 and includes HLA-Cw2, -Cw4, -Cw5, -Cw6, -Cw15,-Cw1602, -Cw17 and -Cw18. HLA-C1 molecules interact with the in-hibitory KIR2DL2 and KIR2DL3 and potentially the activatingKIR2DS2, while HLA-C2 molecules interact with the inhibitory2DL1 and the activating 2DS2 (Parham, 2005a, b; Moesta et al.,2008) The HLA-C2 allotypes have been found to be increased incouples experiencing RM, although only signicant among the menwhen compared with controls, which were women with uncompli-cated pregnancies (Hiby et al., 2008). In a previous study by Christian-sen et al. (1997) RM couples were compared with control couples andno association between RM and HLA-C1 or -C2 was observed. In thestudy by Hiby et al., KIR genotypes were examined as well. A signi-cant increase of the dominantly inhibitory KIR AA genotype was foundin RM women, whereas the genotype BB has more activating KIRs. Inparticular, lack of KIR2DS1 was highly signicant among RM women.The same results have been obtained by two other groups (Wittet al., 2004; Flores et al., 2007). A conicting result has been reportedin a Chinese study, showing an overrepresentation of KIR2DS1 inpatients compared with controls (Wang et al., 2007). They did,however, report that in a group of couples where both partnerscarried an HLA-C2 allele, KIR2DS1 was signicantly lower comparedwith controls. In addition, similar studies have focused on the relation-ship between HLA-C alleles, KIR genotypes and pre-eclampsia, butthis will not be discussed in detail here. In conclusion, it seems thatcertain combinations of HLA-C alleles and KIR genotypes possessan increased risk for RM, but further studies need to be carried out.

Conclusions and futureperspectivesAt this point it is clear that the HLA class Ib molecules along withHLA-C are expressed at the materno-fetal interface. At the sametime many transfection studies and studies on interactions with periph-eral immune cells show that these three proteins and HLA-C interactwith different receptors on immune cells resulting in immunomodula-tion. Because of this, it is apposite to believe that these proteins carryout immunoregulatory functions that serve to induce tolerance to thesemi-allogenic fetus during pregnancy. However, at present, studiesexamining how isolated trophoblast cells and immune cells fromdecidua interact are sparse, and more experiments need to becarried out to clarify the specic underlying immunoregulatorymechanisms that take place during pregnancy.

Clinical studies have examined the different HLA alleles and poly-morphisms and how these might be linked to pathological distur-bances in pregnancy such as RM and unexplained infertility. Atpresent, no consensus has been reached in this area, but results indi-cate that some HLA-G alleles and polymorphisms are risk factors inRM, and that there is a correlation between these gene variants andnegative results in ART procedures in couples experiencing unex-plained infertility. On this basis, it is clear that HLA-G has the potentialof serving as an important diagnostic tool when examining couplesundergoing infertility treatment. Today, different experimentalstudies agree that presence of sHLA-G is a positive prognosticfactor when transferring embryos in IVF treatments. As soon as agood quantitative method for measuring sHLA-G levels in IVF

medias is available, this can be carried out as a routine procedureserving to optimize the results in ART treatments. Newfound interestalso lies in determining which role paternal sHLA-G levels in seminalplasma might have for the development of pregnancy complications,and in the future it will be interesting to examine both male andfemale levels of sHLA-G in couples with fertility problems to try toclarify the importance of this factor in the pathogenesis of theseconditions.

A challenge posed in determining the role of HLA-G and the otherclass Ib molecules in the pathogenesis of RSA is the two parallel linesof studies in this eld. On one hand, basic research on immune cellsand cytokines try to clarify the receptor interactions and mechanismsresulting from these proteins, and on the other hand, clinical studiesexamine HLA alleles, polymorphisms and protein levels in couples ex-periencing RSA and unexplained infertility, to try to determine a cor-relation. Although it is not possible to quantify immune cells andcytokines in decidua in these patients to obtain a more integratedimage of the complex immunological interplay during pregnancy, onemust always remember to use the knowledge from basic research inclinical studies and vice versa.

Authors rolesM.D. conceived and designed the study, identied the articles,acquired and analysed the data, drafted the rst version of the manu-script and revised the manuscript. T.V.F.H. conceived and designedthe study, identied the articles, acquired and analysed the data andrevised the manuscript.

FundingM.D. was supported by a grant from Region Zealand Health SciencesResearch Foundation.

ReferencesAbbas A, Tripathi P, Naik S, Agrawal S. Analysis of human leukocyte antigen

(HLA)-G polymorphism in normal women and in women with recurrentspontaneous abortions. Eur J Immunogenet 2004;31:275278.

Aldrich CL, Stephenson MD, Karrison T, Odem RR, Branch DW, Scott JR,Schreiber JR, Ober C. HLA-G genotypes and pregnancy outcome in coupleswith unexplained recurrent miscarriage. Mol Hum Reprod 2001;7:11671172.

Apps R, Gardner L, Sharkey AM, Holmes N, Moffett A. A homodimeric complex ofHLA-G on normal trophoblast cells modulates antigen-presenting cells viaLILRB1. Eur J Immunol 2007;37:19241937.

Apps R, Gardner L, Moffett A. A critical look at HLA-G. Trends Immunol 2008a;29:313321.

Apps R, Gardner L, Traherne J, Male V, Moffett A. Natural-killer cell ligands at thematernal-fetal interface: UL-16 binding proteins, MHC class-I chain relatedmolecules, HLA-F and CD48. Hum Reprod 2008b;23:25352548.

Apps R, Murphy SP, Fernando R, Gardner L, Ahad T, Moffett A. Human leucocyteantigen (HLA) expression of primary trophoblast cells and placental cell lines,determined using single antigen beads to characterize allotype specicities ofanti-HLA antibodies. Immunology 2009;127:2639.

Arruvito L, Sanz M, Banham AH, Fainboim L. Expansion of CD4+CD25+andFoxP3+ regulatory T cells during the follicular phase of the menstrual Cycle:implications for human reproduction. J Immunol 2007;178:25722578.

Arruvito L, Sotelo AI, Billordo A, Fainboim L. A physiological role for inducibleFoxP3+ TReg cells: lessons from women with reproductive failure. ClinImmunol 2010;136:432441.


at Hacettepe U



ownloaded from

Ashkar AA, Black GP, Wei Q, He H, Liang L, Head JR, Croy BA. Assessment ofrequirements for IL-15 and IFN regulatory factors in uterine NK celldifferentiation and function during pregnancy. J Immunol 2003;171:29372944.

Athanassakis I, Pais M, Ranella A, Vassiliadis S. Detection of soluble HLA-G levels inmaternal serum can be predictive for a successful pregnancy. Transplant Proc 1999;31:18341837.

Avendano C, Franchi A, Jones E, Oehninger S. Pregnancy-specicbeta-1-glycoprotein 1 and human leukocyte antigen-E mRNA in human sperm:Differential expression in fertile and infertile men and evidence of a possiblefunctional role during early development. Hum Reprod 2009;24:270277.

Berger DS, Hogge WA, Barmada MM, Ferrell RE. Comprehensive analysis of HLA-G: implications for recurrent spontaneous abortion. Reprod Sci 2010;17:331338.

Bhalla A, Stone PR, Liddell HS, Zanderigo A,Chamley LW.Comparison of the expressionof human leukocyte antigen (HLA)-G and HLA-E in women with normal pregnancyand those with recurrent miscarriage. Reproduction 2006;131:583589.

Billington WD. The immunological problem of pregnancy: 50 years with the hope ofprogress. A tribute to Peter Medawar. J Reprod Immunol 2003;60:111.

Borrego F, Kabat J, Sanni TB, Coligan JE. Cellular immunology and immune regulationNK cell CD94/NKG2A inhibitory receptors are internalized and recycleindependently of inhibitory signaling processes. J Immunol 2002;169:61026111.

Boyle LH, Gillingham AK, Munro S, Trowsdale J. Selective export of HLA-F by itscytoplasmic tail. J Immunol 2006;176:64646472.

Bulmer JN, Morrison L, Longfellow M, Ritson A, Pace D. Granulated lymphocytes inhuman endometrium histochemical and immunohistochemical studies. HumReprod 1991;6:791798.

Chen W, Jin W, Hardegen N, Lei KJ, Li L, Marinos N, McGrady G, Wahl SM.Conversion of peripheral CD4+CD25- naive T cells to CD4+CD25+

regulatory T cells by TGF-beta induction of transcription factor Foxp3. J ExpMed 2003;198:18751886.

Chen LJ, Han ZQ, Zhou H, Zou L, Zou P. Inhibition of HLA-G expression via RNAiabolishes resistance of extravillous trophoblast cell line TEV-1 to NK lysis.Placenta 2010;31:519527.

Christiansen OB, Mohapeloa HP, Steffensen R, Jersild C. HLA-C and -Bw typingof couples with unexplained recurrent miscarriages. J Reprod Immunol 1997;37:6377.

Colonna M, Navarro F, Bellon T, Llano M, Garcia P, Samaridis J, Angman L, Cella M,Lopez-Botet M. A common inhibitory receptor for major histocompatibilitycomplex class I molecules on human lymphoid and myelomonocytic cells. J ExpMed 1997;186:18091818.

Colonna M, Samaridis J, Cella M, Angman L, Allen L, OCallaghan CA, Dunbar R,Ogg S, Cerundolo V, Rolink A. Cutting edge: hman myelomonocytic cellsexpress an inhibitory receptor for classical and nonclassical MHC class Imolecules. J Immunol 1998;160:30963100.

Cooper MA, Fehniger TA, Turner SC, Chen KS, Ghaheri BA, Ghayur T, Carson WE,Caligiuri MA. Human natural killer cells: a unique innate immunoregulatory rolefor the CD56bright subset. Blood 2001;97:31463151.

Coulam CB. Epidemiology of recurrent spontaneous abortion. Am J Reprod Immunol1991;26:2327.

Desai N, Filipovits J, Goldfarb J. Secretion of soluble HLA-G by day 3 humanembryos associated with higher pregnancy and implantation rates: assay ofculture media using a new ELISA kit. Reprod BioMed Online 2006;13:272277.

Ellis SA, Sargent IL, Redman CWG, McMichael AJ. Evidence for a novel HLA antigenfound on human extravillous trophoblast and a choriocarcinoma cell line.Immunology 1986;59:595601.

Favier B, LeMaoult J, Rouas-Freiss N, Moreau P, Menier C, Carosella ED. Researchon HLA-G: an update. Tissue Antigens 2007;69:207211.

Favier B, LeMaoult J, Lesport E, Carosella ED. ILT2/HLA-G interaction impairsNK-cell functions through the inhibition of the late but not the early events ofthe NK-cell activating synapse. FASEB J 2010;24:689699.

Flores AC, Marcos CY, Paladino N, Arruvito L, Williams F, Middleton D, Fainboim L.KIR receptors and HLA-C in the maintenance of pregnancy. Tissue Antigens 2007;69:112113.

Fuzzi B, Rizzo R, Criscuoli L, Noci I, Melchiorri L, Scarselli B, Bencini E, Menicucci A,Baricordi OR. HLA-G expression in early embryos is a fundamental prerequisitefor the obtainment of pregnancy. Eur J Immunol 2002;32:311315.

Geraghty DE, Wei X, Orr BT, Koller BH. Human leukocyte antigen F: an expressedHLA gene composed of a class I coding sequence linked to a novel transcribedrepetitive element. J Exp Med 1990;171:118.

Geraghty DE, Stockschleader M, Ishitani A, Hansen JA. Polymorphism at the HLA-Elocus predates most HLA-A and -B polymorphism.Hum Immunol 1992;33:174184.

Ghiringhelli F, Menard C, Terme M, Flament C, Taieb J, Chaput N, Puig PE,Novault S, Escudier B, Vivier E et al. CD4+CD25+ regulatory T cells inhibitnatural killer cell functions in a transforming growth factor-beta-dependentmanner. J Exp Med 2005a;202:10751085.

Ghiringhelli F, Puig PE, Roux S, Parcellier A, Schmitt E, Solary E, Kroemer G,Martin F, Chauffert B, Zitvogel L. Tumor cells convert immature myeloiddendritic cells into TGF-beta-secreting cells inducing CD4+CD25+ regulatory Tcell proliferation. J Exp Med 2005b;202:919929.

Gomez-Casado E, Martinez-Laso J, Vargas-Alarcon G, Varela P, Diaz-Campos N,Alvarez M, Alegre R, Arnaiz-Villena A. Description of a new HLA-E (E*01031)allele and its frequency in the Spanish population. Hum Immunol 1997;54:6973.

Gomez-Lozano N, de Pablo R, Puente S, Vilches C. Recognition of HLA-G by theNK cell receptor KIR2DL4 is not essential for human reproduction. Eur JImmunol 2003;33:639644.

Gonen-Gross T, Achdout H, Arnon TI, Gazit R, Stern N, Horejsi V,Goldman-Wohl D, Yagel S, Mandelboim O. The CD85J/leukocyte inhibitoryreceptor-1 distinguishes between conformed and beta-2-microglobulin-freeHLA-G molecules. J Immunol 2005;175:48664874.

Gregori S, Tomasoni D, Pacciani V, Scirpoli M, Battaglia M, Magnani CF, Hauben E,Roncarolo MG. Differentiation of type 1T regulatory cells (Tr1) by tolerogenicDC-10 requires the IL-10-dependent ILT4/HLA-G pathway. Blood 2010;116:935944.

Grimsley C, Ober C. Population genetic studies of HLA-E. Evidence for selection.Hum Immunol 1997;52:3340.

Gros F, Cabillic F, Toutirais O, Le Maux A, Sebti Y, Amiot L. Soluble HLA-Gmolecules impair natural killer/dendritic cell crosstalk via inhibition of dendriticcells. Eur J Immunol 2008;38:742749.

Groux H, OGarra A, Bigler M, Rouleau M, Antonenko S, De Vries JE, GraziaRoncarolo M. A CD4+ T cell subset inhibits antigen-specic T cell responsesand prevents colitis. Nature 1997;389:737742.

Guillaudeux T, Gomez E, Onno M, Drenou B, Segretain D, Alberti S, Lejeune H,Fauchet R, Jegou B, Le Bouteiller P. Expression of HLA class I genes in meioticand post-meiotic human spermatogenic cells. Biol Reprod 2006;55:99110.

Harrison GA, Humphrey KE, Jakobsen IB, Cooper DW. A 14 bp deletionpolymorphism in the HLA-G gene. Hum Mol Genet 1993;2:2200.

Heikkinen J, Mottonen M, Alanen A, Lassila O. Phenotypic characterization ofregulatory T cells in the human decidua. Clin Exp Immunol 2004;136:373378.

Heinrichs H, Orr HT. HLA Non-A, -B, -C Class I genes, their structure andexpression. Immunol Res 1990;9:265274.

Hiby SE, Regan L, Lo W, Farrell L, Carrington M, Moffett A. Association of maternalkiller-cell immunoglobulin-like receptors and parental HLA-C genotypes withrecurrent miscarriage. Hum Reprod 2008;23:972976.

Higuma-Myojo S, Sasaki Y, Miyazaki S, Sakai M, Siozaki A, Miwa N, Saito S. Cytokineprole of natural killer cells in early human pregnancy. Am J Reprod Immunol 2005;54:2129.

Hoare HL, Sullivan LC, Clements CS, Ely LK, Beddoe T, Henderson KN, Lin J,Reid HH, Brooks AG, Rossjohn J. Subtle changes in peptide conformationprofoundly affect recognition of the non-classical MHC class I molecule HLA-Eby the CD94-NKG2 natural killer cell receptors. J Mol Biol 2008;377:12971303.

Hunt JS, Jadhav L, Chu W, Geraghty DE, Ober C. Soluble HLA-G circulates inmaternal blood during pregnancy. Am J Obstet Gynecol 2000;183:682688.

Hunt JS, Pace JL, Morales PJ, Ober C. Immunogenicity of the soluble isoforms ofHLA-G. Mol Hum Reprod 2003;9:729735.

Hunt JS, Petroff MG, McIntire RH, Ober C. HLA-G and immune tolerance inpregnancy. FASEB J 2005;19:681693.

Hviid TV, Hylenius S, Hoegh AM, Kruse C, ChristiansenOB. HLA-G polymorphisms incouples with recurrent spontaneous abortions. Tissue Antigens 2002;60:122132.

Hviid TVF, Hylenius S, Rorbye C, Nielsen LG. HLA-G allelic variants are associatedwith differences in the HLA-G mRNA isoform prole and HLA-G mRNA levels.Immunogenetics 2003;55:6379.

Hviid TVF, Hylenius S, Lindhard A, Christiansen OB. Association between humanleukocyte antigen-G genotype and success of in vitro fertilization and pregnancyoutcome. Tissue Antigens 2004a;64:6669.

Hviid TV, Rizzo R, Christiansen OB, Melchiorri L, Lindhard A, Baricordi OR. HLA-Gand IL-10 in serum in relation to HLA-G genotype and polymorphisms.Immunogenetics 2004b;56:135141.

106 Dahl and Hviid

at Hacettepe U



ownloaded from

Ishitani A, Geraghty DE. Alternative splicing of HLA-G transcripts yields protein withstructures resembling both class I and II antigens. Proc Natl Acad Sci USA 1992;89:39473951.

Ishitani A, Sageshima N, Lee N, Dorofeeva N, Hatake K, Marquardt H,Geraghty DE. Protein expression and peptide binding suggest unique andinteracting functional roles for HLA-E, F, and G in maternal-placental immunerecognition. J Immunol 2003;171:13761384.

Jacobs R, Hintzen G, Kemper A, Beul K, Kempf S, Behrens G, Sykora KW,Schmidt RE. CD56bright cells differ in their KIR repertoire and cytotoxicfeatures from CD56dim NK cells. Eur J Immunol 2001;31:31213127.

Jasper MJ, Tremellen KP, Robertson SA. Primary unexplained infertility is associatedwith reduced expression of the T regulatory cell transcription factor FoxP3 inendometrial tissue. Mol Hum Reprod 2006;12:301308.

Kanai T, Fujii T, Keicho N, Tokunaga K, Yamashita T, Hyodo H, Miki A, Unno N,Kozuma S, Taketani Y. Polymorphism of human leukocyte antigen-E gene inthe Japanese population with or without recurrent abortion. Am J ReprodImmunol 2001;45:168173.

King A, Boocock C, Sharkey A, Gardner L, Beretta A, Siccardi AG, Loke YW.Evidence for the expression of HLA-C class I mRNA and protein by humanrst trimester trophoblast. J immunol 1996;16:20682076.

King A, Allan DSJ, Bowen M, Powis SJ, Joseph S, Verma S, Hiby SE, McMichael AJ,Loke YW, Braud VM. HLA-E is expressed on trophoblast and interacts withCD94/NKG2 receptors on decidual NK cells. Eur J Immunol 2000;30:16231631.

Klentzeris LD, Li TC, Dockery P, Cooke ID. The endometrial biopsy as a predictivefactor of pregnancy rate in women with unexplained infertility. Eur J Obstet GynecolReprod Biol 1992;45:119124.

Kunishima S, Nagae M, Mizuno S, Kamiya T, Ozawa K. A new polymorphism in theHLA-F gene 67Ala(GCC) to Ala(GCG). Immunogenetics 1999;49:147148.

Kusumi M, Yamashita T, Fujii T, Nagamatsu T, Kozuma S, Taketani Y. Expressionpatterns of lectin-like natural killer receptors, inhibitory CD94/NKG2A, andactivating CD94/NKG2C on decidual CD56bright natural killer cells differ fromthose on peripheral CD56dim natural killer cells. J Reprod Immunol 2006;70:3342.

LaBonte ML, Letvin NL. Variable NKG2 expression in the peripheral bloodlymphocytes of rhesus monkeys. Clin Exp Immunol 2004;138:205212.

Lachapelle MH, Miron P, Hemmings R, Roy DC. Endometrial T B and NK cells inpatients with recurrent spontaneous abortion. Altered prole and pregnancyoutcome. J Immunol 1996;156:40274034.

Langat DK, Platt JS, Tawk O, Fazleabas AT, Hunt JS. Differential expression ofhuman leukocyte antigen-G (HLA-G) messenger RNAs and proteins in normalhuman prostate and prostatic adenocarcinoma. J Reprod Immunol 2006;71:7586.

Lanier LL, Corliss BC, Wu J, Leong C, Phillips JH. Immunoreceptor DAP12 bearing atyrosine-based activation motif is involved in activating NK cells. Nature 1998;391:703707.

Larsen MH, Bzorek M, Pass MB, Larsen LG, Nielsen MW, Svendsen SG, Lindhard A,Hviid TV. Human Leukocyte Antigen-G in the male reproductive system and inseminal plasma. Mol Hum Reprod, 2011; PMID 21813635

Le Discorde M, Le Danff C, Moreau P, Rouas-Freiss N, Carosella ED.HLA-G*0105N null allele encodes functional HLA-G isoforms

2 dahl 2012 hla early pregnancy loss.pdf

Documents

pregnancy hlae

dividedinto hla class

pregnancy hlag polymorphism

basic role of hla

f function signicance

pregnancy ambiguity

g future studies

different proteins