A specific interferon (IFN)-stimulated response element of the distal HLA-G promoter binds IFN-regulatory factor 1 and mediates enhancement of this nonclassical class I gene by IFN-beta.

Type I interferons display a broad range of immunomodulatory functions. Interferon beta increases gene expression at the transcriptional level through binding of factors to the interferon-stimulated response element (ISRE) within the promoters of interferon-inducible genes, such as HLA class I. Despite mutation of the class I ISRE sequence within the nonclassical HLA-G class I gene promoter, we show that interferon beta enhances both transcription and cell surface expression of HLA-G in trophoblasts and amniotic and thymic epithelial cells that selectively express it in vivo. Deletion and mutagenesis analysis of a putative interferon-regulatory factor (IRF)-1 binding site within the HLA-G promoter show that HLA-G transactivation is mediated through an ISRE sequence 746 base pairs upstream from ATG, which is distinct from the interferon-responsive element described within proximal classical class I gene promoters. Electrophoretic mobility shift analysis and supershift analysis further demonstrate that interferon-responsive transcription factors, including IRF-1, specifically bind to the HLA-G ISRE. Our results provide evidence that IRF-1 binding to a functional ISRE within the HLA-G promoter mediates interferon beta-induced expression of the HLA-G gene. These observations are of general interest considering the implication of HLA-G in mechanisms of immune escape involved in fetal-maternal tolerance and other immune privilege situations.

HLA-G is defined as a nonclassical HLA class I antigen (1,2) that was originally found to be restrictively expressed in the human placenta, where it is thought to play a role in maternal tolerance of the fetal semiallograft. HLA-G expression was further characterized on several placental cell types such as extravillous cytotrophoblasts, amniocytes, or endothelial cells of chorionic vessels and also on subsets of thymic epithelial cells (3). HLA-G expression allows down-regulation of both NK 1 and T lymphocyte cytolytic functions (4) through interaction with killing inhibitory receptors, namely p49/KIR2DL4, ILT2, and ILT4 (5). The capacity of HLA-G leader peptides to stabilize surface expression of nonclassical HLA-E antigens also indirectly contributes to modulate cytolytic activity mediated by the widely expressed CD94/NKG2 receptors (6). HLA-G associates with a wide array of nonamer peptides (7) and binds to the CD8 T-cell coreceptor (8,9). Other immunomodulatory roles, such as its capacity to elicit a T-cell receptor-restricted response in transgenic mice (10), to trigger apoptosis of activated T and NK cells bearing the CD8 molecule (11), to impair NK cell migration (12), and to modulate cytokine (13,14) or the function of B or myelomonocytic cells bearing killing inhibitory receptors, have also been evoked. Tissue-specific patterns of HLA-G expression and their frequent alteration in pathological situations such as pregnancy disorders (15), viral infections (16 -19), tumors (20 -24), or transplantation (25) may modulate the mounting of an efficient immune response and thus provide an additional mechanism to escape immune surveillance.
Interferons are classified in a family of related cytokines, type I (IFN-␣, -␤, -, and -) and type II (IFN-␥) interferons, which mediate diverse functions including antiviral, antiproliferative, and immunomodulatory activities. Type I interferons are produced by many cell types (e.g. macrophages, T cells, keratinocytes, and Langerhans cells) in response to viral or bacterial infection and tumors. They display pleiotropic effects on the immune system, including stimulation of NK cells and macrophage activation, T-cell activation and survival, and upregulation of various genes such as IFN-␥ and major histocompatibility complex class I (26). The cascade of events that yield to IFN type I induction involves activation of Jak/STAT transduction pathways and transactivation of inducible gene promoter through the ISRE regulatory element, a binding site for ISGF3 (IFN-stimulated gene factor 3) or IRFs (27,28). After binding to its receptor, IFN type I induces translocation of activated STAT1 and STAT2 into the nucleus and formation of a heterotrimeric complex containing p48/ISGF3␥ (ISGF3). Among transcription factors belonging to the IRF family, IRF-1 and IRF-2 are secondary IFN response factors that interact mainly with ISRE to activate or repress target promoter activity (29).
Another well-characterized IFN cis-acting regulatory element is the IFN-␥ activation site (GAS), which rather mediates the immediate response of several genes to IFN-␥ (30).
The pattern of HLA-G expression is tightly regulated due to cell-specific transcriptional control, but the regulatory pathways controlling its tissue-specific transactivation remain to be established. HLA-G gene expression can be activated by interleukin 10 (31), glucocorticoids, or stress treatment (32) independently of classical HLA class I genes or by IFNs (33)(34)(35)(36), despite deleterious mutations within the HLA-G promoter of critical cis-acting regulatory elements involved in the constitutive and interferon-inducible transcription of HLA-class I genes (enhancer-A, ISRE, and site ␣) (37)(38)(39).
Our study aimed at identifying the molecular mechanism by which IFN-␤ enhances both transcription and cell surface expression of the nonclassical HLA-G antigen in trophoblast, amnion, and thymus-derived epithelial cells that also express it in vivo. We report that IFN-␤ up-regulates HLA-G transcription through promoter transactivation and binding of IFNresponsive factors, including IRF-1, to a functional ISRE mapped at 746 bp from the ATG, within the HLA-G promoter.

Cells, Tissues, and Primary
Cultures-The JEG-3 choriocarcinoma cell line was purchased from American Type Culture Collection. The LT-TEC2 cell line was derived from an independent proliferative clone of thymic epithelial cells transformed by the SV40LT oncogene (40). Cell lines were respectively maintained in Dulbecco's modified Eagle's medium or RPMI 1640 medium (Life Technologies, Inc.) supplemented with L-glutamine and 10% fetal calf serum.
Normal thymuses were obtained from children (age range, 5 days to 2 years) undergoing cardiac surgery. Primary TEC cultures were established as described previously (41). Thymic fragments were minced into small pieces, washed, and incubated in RPMI 1640 medium and 20% horse serum in 75-cm 2 flasks for 10 -12 days. Adherent TECs were collected and subcultured into Primaria 24-well plates (Becton Dickinson, San Jose, CA).
First trimester trophoblasts were obtained from voluntarily interrupted normal pregnancies at 5-7 weeks of gestation (with local ethic committee approval).
Fetus-surrounding membranes (amniochorion) were extracted from whole term placenta after normal delivery. Chorionic membrane was stripped from the amnion by scraping. Amnion epithelial cells (AECs) were prepared according to Hammer et al. (42). Amnion membrane was minced into small pieces and dissociated by two rounds of 1-h trypsinization in phosphate-buffered saline, 0.05% trypsine (Difco, Detroit, MI), and 40 units/ml DNase and washing. The resulting amnion epithelial cells were grown in Dulbecco's modified Eagle's medium: Ham's F-12 (Life Technologies, Inc.) supplemented with 10% fetal calf serum, glutamine, and antibiotics for 6 -12 days.
Cells and trophoblast explants were stimulated with human recombinant IFN-␤ (1000 units/ml; PeproTech EC, London, United Kingdom) for 2-72 h as further stated or not stimulated.
RNase Protection Assay (RPA)-Single-stranded radiolabeled RNA probes were synthesized using the MAXIscript in vitro transcription kit (Ambion, Austin, TX) with T7 RNA polymerase and [␣-32 P]CTP as described previously (31). The HLA-G template was prepared by PCR amplification of a genomic fragment in the 3Ј-untranslated region with G-1089F forward primer (5Ј-CCCTTTGTGACTTCAAGAAC) and T7 promoter containing reverse primer T7G.1250R (5Ј-GGATCCTAATAC-GACTCACTATAGGGAGGTTATAGCTCAGTGGCCCAC). Cyclophilin standard template was obtained from Ambion. Protection of HLA-G transcripts was carried out using a HybSpeed RPA kit (Ambion). 5 g of total RNA were hybridized with 5 ϫ 10 5 cpm of HLA-G and cyclophilin riboprobes for 10 min in 10 l of hybridization buffer and digested for 30 min with RNase A/T1 mix. Radiolabeled protected fragments were precipitated and separated on a 5% acrylamide (19:1) gel. The gel was dried and exposed to a molecular imager (Bio-Rad) for quantification.
Reporter Constructs-The 1.4-kb fragment of the HLA-G promoter was prepared by PCR amplification of a JEG-3 genomic DNA region (nucleotide Ϫ1438 to the ATG) using primers sense Ϫ1438F and antisense Ϫ14R as described previously (36). PCR products were cloned into the pGL3-basic promoterless vector (Promega), upstream of the firefly luciferase reporter gene. The 500-bp HLA-G promoter fragment/pGL3 vector construct (Ϫ500) was then generated by digesting the former vector (Ϫ1400) using SacI restriction enzyme. A mutated ISRE site was introduced within the Ϫ1400 construct by PCR amplification using complementary mutated oligonucleotides sense 5Ј-CACAGATACATT-GTCTGGGAACTTAGAACTTAAGAGCTTTGTGAGTCC and antisense 5Ј-GGACTCACAAA-GCTCTTAAGTTCTAAGTTCCCAGACAATGTAT-CTGTG (Ϫ1400mut), carrying the same mutation as in EMSA (see Fig.  3). Nucleotide sequences of constructs were controlled by sequencing.
Transient Transfection and Luciferase Activity Assays-Cells were seeded in 24-well plates 16 -20 h before transfection, at 60% confluence. Cells were transfected by 1 g of the HLA-G promoter/pGL3 firefly luciferase reporter construct and 10 ng of pRL-TK Renilla luciferase vector (Promega) as an internal control for transfection using the Exgen 500 reagent (Euromedex, Souffelweyersheim, France) according to the manufacturer's instructions. On next day, cells were fed with culture medium alone or supplemented with IFN-␤. Firefly and Renilla luciferase activities of precleared cell extracts (10 l) were measured 20 h after induction, after sequential injection of the firefly and Renilla luciferase substrates (50 l) provided in the Dual Luciferase Reporter assay system (Promega) using a luminometer (EG and G Berthold).
Firefly luciferase activity values (relative light units/s), were either measured in duplicate or divided by the Renilla luciferase activity values to correct for transfection efficiency and expressed as a mean of at least three independent experiments.

IFN-␤ Enhances HLA-G and Classical HLA Class I Cell
Surface Expression in Trophoblast, Amnion, and Thymus-derived Cells-Primary cultured cell types expressing HLA-G in vivo were chosen as a model to study HLA-G modulation by IFN-␤. Human AECs and TECs were derived from term placenta amnions and surgically removed pediatric thymuses. These primary cells maintain HLA-G expression ex vivo, although a down-regulation of the level of cell surface HLA-G antigens occurring throughout the culture led to a weak steadystate level of HLA-G cell surface antigens at the time of analysis (45). A trophoblast-derived choriocarcinoma cell line (JEG-3) and a thymic epithelial-derived cell line (LT-TEC2) were also included in the study. LT-TEC2, which is derived by SV40 transformation of TEC primary culture, shares several features with primary cells (40).
Cells were left untreated or subjected to a 48-h IFN-␤ treatment and analyzed in flow cytometry experiments. HLA-G as well as classical HLA class I surface expression was significantly induced on amniotic and thymic primary cultures as assessed by the enhancement of the mean of fluorescence intensity (mfi) after IFN-␤ stimulation (Fig. 1A). IFN-␤-mediated induction was confirmed using several primary cultures derived from distinct individuals (Fig. 1B). Lower levels of induction were also observed in the trophoblast-derived JEG-3 cell line (Fig. 1B). Fold levels of HLA-G antigen surface induction varied from 3 to 1.5 times, depending on the cell lineage (Fig.  1B), and were higher in primary cultures expressing low (TECs) or intermediate (AECs) levels of steady-state HLA-G antigens due to a decreased HLA-G expression resulting from an 8 -15-day culture. We also show that, like primary TECs, IFN-␤-treated LT-TEC2 cells retain the capacity to up-regulate HLA-G as well as major histocompatibility complex class I antigens (data not shown).
IFN-␤ Enhances the Levels of HLA-G Transcripts in Trophoblast, Amnion, and Thymus-derived Cells-To further characterize mechanisms involved in IFN-␤-mediated HLA-G surface induction, we investigated the effect of this cytokine on HLA-G transcript levels by RT-PCR (Fig. 2, A and B) and RNase protection assay (Fig. 2B). Pan HLA-G primers specific for exon 2-and 3Ј-untranslated region-containing transcripts were used to amplify all HLA-G isoforms ( Fig. 2A, top panel), whereas HLA-G primer sets located in exon 3 and intron 4 were used to specifically detect HLA-G5 soluble isoforms ( Fig. 2A, middle  panel). We show that all HLA-G transcripts, including those encoding membrane-bound and soluble HLA-G5 isoforms, are significantly enhanced after IFN-␤ treatment in both primary cultured amniotic and thymic cells. Similar RT-PCR results were reproducibly obtained using primary TEC and AEC cultures derived from several individuals ( Fig. 2A). IFN-␤-mediated enhancement of HLA-G messengers was also observed in treated first trimester trophoblast explants (Fig. 2B). Concomitant protection of HLA-G-and cyclophilin-specific transcripts from RNase degradation, measured by a RNase protection assay, allowed us to further quantify the level of IFN-␤ induction to a 3.5-fold enhancement of HLA-G messages in trophoblast explants (Fig. 2B, right panel).
IFN-␤ Transactivation of the HLA-G Promoter Is Mediated through an Upstream ISRE-HLA-G gene transcription is upregulated in response to IFN-␤, despite the disruption of enhancer A/ISRE/site ␣-regulatory sequences known to be the target of IFN response factors within the classical HLA class I gene promoter. We thus investigated whether cis-regulatory elements mediating HLA-G transcriptional activation by IFN-␤ could be identified within a 1.4-kb region of the HLA-G promoter.
Computer search and manual sequence analysis based on sequence homologies led us to identify two putative ISRE sites within the HLA-G promoter. The upstream one, located at Ϫ746 bp from the ATG, beside a GAS element (46), is highly homologous to the consensus ISRE (Fig. 3). The other one, lying downstream at Ϫ380 bp, is less conserved.
To assess the functionality of these sequences, we analyzed HLA-G promoter activity in JEG-3 and LT-TEC2 cells. Fragments of the 1.4-kb HLA-G promoter were subcloned upstream of the firefly luciferase reporter gene into the promoterless pGL3 basic vector. Extracts from transfected cells subjected or not to a 20-h IFN-␤ treatment were analyzed for luciferase activity. To discriminate between both ISRE putative sites, the activity of the 1.4-kb HLA-G promoter fragment construct (Ϫ1400) containing both ISREs was compared with that of a shorter construct (Ϫ500) with the upstream ISRE deleted and retaining solely the downstream ISRE (Fig. 4).
We show that the activity of the Ϫ1400 construct exhibits a 3-fold enhancement upon IFN-␤ induction in TECs, demonstrating that the 1.4-kb promoter region is responsive to IFN-␤. Enhancement of promoter activity was much weaker but reproducible (1.4 Ϯ 0.4) in JEG-3 choriocarcinoma cells, in accordance with the very weak cytokine-mediated transcriptional activation of HLA-G (33,36,47) and classical HLA class I (43) described previously in this cell line. In contrast, promoter activity of a deleted fragment (Ϫ500) lacking the upstream putative site was not affected by IFN treatment in both cell types (Fig. 4), showing that deletion of the upstream ISRE impaired transcriptional activation of the 1.4-kb HLA-G promoter fragment. Functionality of the upstream ISRE was further analyzed by directed mutagenesis. No enhancement of HLA-G promoter activity was observed using a construct in which the upstream ISRE site is mutated within the 1.4-kb promoter fragment (Fig. 4), suggesting that deleterious muta-tions of this site preventing the binding of transcription factors impair IFN-␤ transactivation of the 1.4-kb HLA-G promoter fragment.
IRF-1 Binds to the Functional HLA-G ISRE in Response to IFN-␤ Treatment-EMSAs were carried out to assess direct binding of nuclear factors to the functional ISRE (Ϫ746) after IFN-␤ induction. Nuclear protein extracts were incubated with a radiolabeled probe designated HLA-G ISRE/GAS, which spans nucleotide Ϫ757 to Ϫ727 of the HLA-G promoter sequence and contains the functional ISRE linked to a GAS sequence (Fig. 3). Kinetics of binding of nuclear factors to the HLA-G ISRE/GAS probe was first investigated using nuclear extracts from JEG-3 cells stimulated by IFN-␤ for different time periods (Fig. 5A). Two specific IFN-␤-induced complexes, C1 and C2, were detected in untreated extracts, but their intensity increases until 24 h of IFN-␤ treatment. The stronger C1 complex is constituted of several complexes of close mobility as assessed in EMSA that have undergone a longer gel migration (see LT-TEC2 and trophoblasts EMSA (Fig. 5A)). Part of the binding activities contained within C1 complexes was not fully abolished by an excess of cold homologous competitor (ISRE/GAS), which could be due to the altered binding activity of p48 complexes reported in trophoblast cells (47). The binding pattern of these IFN-induced factors analyzed using trophoblast, AECs, and LT-TEC2 cells extracts was quite similar to that observed in JEG-3 cells, except that the ISRE/GAS probe fully competed for formation of all C1-forming complexes in LT-TEC2 extracts (Fig. 5A).
To evaluate the specificity of binding of factors to the HLA-G ISRE/GAS probe and to discriminate between both IFN-responsive elements, competition experiments were conducted using wild type and mutated HLA-G ISRE and GAS sequences as well as functional ISRE and GAS sites respectively derived from HLA-B7 and IRF-1 IFN-inducible genes. Addition of an excess of cold oligonucleotides either homologous to the HLA-G ISRE/GAS or carrying a mutated GAS sequence (ISRE/GASm) competed binding of C1 and C2 to the probe, as assessed by the disappearance of these complexes. In contrast, an excess of cold competitor carrying a mutated ISRE motif and a wild type GAS motif (ISREmut/GAS) did not alter the pattern of binding of these complexes. In addition, an ISRE oligonucleotide derived from the HLA-B7 promoter sequence competed for binding of the C1-and C2-specific complexes in trophoblast extracts, unlike a functional GAS site derived from the IRF-1 promoter (data not shown). These competition experiments confirm that C1 and C2 IFN-␤-induced complexes specifically bind to the   3. Localization of putative ISRE and GAS interferon-responsive sites within the HLA-G promoter sequence. Arrows indicate the orientation of ISRE and GAS (boxed areas), with regard to the consensus ISRE and GAS described in the literature (GAS cons. and ISRE cons.). Filled squares identify residues mutated to disrupt the ISRE site of the sequence of the mutated HLA-G promoter/pGL3 construct used in the luciferase activity assays and the ISRE and GAS of the mutated HLA-G oligonucleotides (ISREmut/GAS, ISRE/GASmut) used in EMSA (coding strand only).
ISRE motif, whereas the GAS motif or surrounding nucleotides within the HLA-G ISRE/GAS probe are unlikely to interfere or cooperate in the binding of these factors to the ISRE of the HLA-G promoter. To identify the nuclear proteins that bind to the HLA-G ISRE, supershift experiments were conducted. JEG-3 and LT-TEC2 nuclear extracts were incubated with antibodies specific for DNA-binding factors involved in transactivation of IFN-␤-inducible genes before the addition of the HLA-G ISRE/GAS probe. We thus identified the presence of IRF-1 within the C1 complex interacting specifically with the HLA-G ISRE (Fig. 5B). As expected from published data, IRF-1 binds transiently to the ISRE in response to IFN-␤ treatment  5. Binding of nuclear proteins to the HLA-G ISRE in response to IFN-␤. Representative EMSA performed using nuclear extracts from JEG-3, trophoblasts, AECs, or LT-TEC2 cells stimulated with IFN-␤ for different time periods and the HLA-G ISRE/GAS probe (Fig. 3). A, unlabeled homologous (ISRE/GAS), ISRE-mutated (ISREm/GAS), GAS-mutated (ISRE/GASm), or irrelevant (IR) competitors were used in 100 -200-fold molar excess. C1 and C2 complexes bound specifically to the HLA-G ISRE/GAS probe are indicated by an arrow. B, supershift experiments. IRF-1 rabbit polyclonal antibodies supershift (*) an IRF-1-containing complex bound to the HLA-G ISRE/GAS probe (C1). A rabbit polyclonal serum serves as nonspecific control (IR). because IRF-1 shifted complexes can be detected 2 h (Fig. 5B, top panel) or 4 h (Fig. 5B, bottom panel) after induction, whereas they are not observed within untreated extracts or after a 24-h treatment ( Fig. 5B; data not shown). DISCUSSION A previous study reported the ability of IFN-␤ to enhance levels of HLA-G surface antigens, but these data remained preliminary because they were only reported on transfected mouse fibroblasts. Although IFN-␤ was identified as a potent inductor of HLA-G transcription in other cell types (33)(34)(35), the regulatory pathways yielding to this transactivation were not elucidated. Indeed, little is known about sequences and factors that participate in transactivation of the HLA-G promoter, in which most of the conserved regulatory DNA elements involved in the constitutive and cytokine-induced expression of the HLA class I genes are disrupted and nonfunctional.
Our findings represent the first evidence of HLA-G modulation and overall IFN-␤-mediated up-regulation of HLA class I cell surface expression in amnion and thymic epithelial cells expressing HLA-G. We also show that IFN-␤-mediated enhancement of the level of HLA-G surface antigens was comparable to that of classical HLA class I surface antigens and was correlated to an overall enhancement of the levels of HLA-G transcripts. The extent of up-regulation of HLA-G transcript levels we observed in IFN-␤-treated primary cells, cell lines, and trophoblast tissue was in agreement with the 2-2.5-fold enhancement reported previously and may not be attributed to increased stability of mRNA but rather to stimulation of the HLA-G transcription rate (35). Our results confirm that IFN-␤ is a potent inducer of HLA-G transcription, which may explain the effect observed at the surface of HLA-G-expressing cells, although a posttranscriptional effect of IFN-␤ on HLA-G enhancement cannot be excluded.
We further elucidated part of the regulatory mechanisms involved in IFN-␤ induction of HLA-G transcription. We have identified a functional ISRE at position Ϫ746 bp from the ATG within the distal HLA-G promoter, and we demonstrate that this site is necessary to confer IFN-␤ transactivation of the 1.4-kb HLA-G gene promoter fragment. This ISRE does not appear to be conserved among other classical HLA class I promoters and could represent a locus-specific pathway of IFN-␤-mediated induction of the nonclassical HLA-G class I antigen. We also demonstrate that the GAS element, lying downstream of the ISRE, does not play a role in IFN-␤ transactivation of HLA-G.
Previous studies on the transactivating effect of the ISRE element upon IFN stimulation point out that both enhancer A and site ␣ are crucial for maximal induction of HLA class I genes through the ISRE (37,39). The nonclassical HLA-E also contains a specific IFN-␥ response region constituted of two adjacent cis-acting regulatory elements, both of which are required to mediate the full response to IFN-␥ (48). No adjacent cooperative sequence seems to be required for full induction by IFN-␤ through the functional ISRE within the HLA-G promoter, which identifies the particular nature of IFN-␤-mediated HLA-G induction. The level of IFN-␤-mediated induction of HLA-G surface antigens was quite well correlated to the level of HLA-G promoter transactivation through the Ϫ746 ISRE and was similar to that reported using a transfected HLA-G 6.0-kb fragment containing the 1.4-kb promoter fragment and 3Ј sequences (35). Nevertheless, we cannot exclude the presence of other functional IFN-responsive sites within the HLA-G gene.
HLA-G expression can be induced by IFN-␣ and IFN-␥ in transfected mouse fibroblasts (35). Whether the functional ISRE is able to mediate transcriptional activation of HLA-G upon both type I IFNs remains to be verified. Interestingly, we have previously reported a lack of activation of the 1.4-kb HLA-G promoter fragment in IFN-␥-treated JEG-3 or in IFN-␥-treated Tera-2 cells (36, 37) that displayed a 3-fold enhancement of HLA-G promoter activation upon treatment with IFN-␤ (data not shown). This could suggest that the ISRE mediating the IFN-␤ response within the HLA-G promoter is not involved in IFN-␥ transactivation of HLA-G or requires additional cooperative regulatory elements. These results point out that differential and overlapping mechanisms, including several regulatory pathways and cooperative interactions between cellular factors involved in the specific response to IFNs, regulate IFN-␣, -␤, and-␥ transactivation of HLA-G and other HLA class I genes.
We further provide evidence that IRF-1 binds to the functional ISRE within the HLA-G promoter in response to IFN-␤ and may thus represent a key factor in the regulation of the HLA-G gene in trophoblast, amnion, and thymus-derived cells. Accordingly, IRF-1 is described as the principal transcription factor binding to the ISRE within the major histocompatibility complex class I promoter (49). Supershift analysis failed to identify other known ISRE-binding proteins such as IRF-2, p48, or STAT1␣. An altered activity of the p48/ISGF3␥ subunit of the ISGF3 factor has been reported in trophoblast cells (47). STAT1-containing complexes binding to HLA class I ISREs were not detected in IFN-␥-treated HeLa cells, unlike such GAS-binding complexes (38), suggesting that ISRE-protein complexes would be barely detectable in supershift experiments.
In addition to playing essential roles in IFN responses, IRF-1 proteins are involved in other regulatory processes, i.e. cell cycle regulation, tumor suppression, oncogenic activities, apoptosis, and development and function of immune effector cells (50,51). Interleukin 2 and interleukin 12 have been shown to directly induce IRF-1 gene expression in human T and NK cells (52). Whether these cytokines that promote cell-mediated immune response could also enhance HLA-G expression through binding of IRF-1 to ISRE should be further investigated in immunocompetent cells.
Type I interferons are highly involved in both the innate and specific host protective response, such as the T-cell IFN-␥ response to viral infection (53)(54)(55). Facing this, viruses have evolved numerous strategies to subvert host defenses, including modulation of major histocompatibility complex class I expression or mimicking host cell genes such as IRF-1 or ILT2 (56). Alteration of HLA-G or locus-specific class I gene expression may also interfere in the balance between escape from cytotoxic T lymphocytes and maintenance of protection from NK cells (16 -19, 57). Given that IFN-␤ is produced in virally infected trophoblast and amnion (58,59), specific pathways of HLA-G modulation could permit infected cells to block NK and T-cell responses, as seen during maternal-fetal transmission of human cytomegalovirus (60) or human immunodeficiency virus. Alternatively, lysis of infected material could be triggered by CD94/NKG2C receptors through HLA-E surface expression or by HLA-G-restricted T-cell Receptor and thus prevent pathogen spreading from the placental cells to the fetus. IFN-␤mediated HLA-G up-regulation in TECs could also affect selection of the thymic repertoire.
Type I interferons are used in the treatment of several human pathologies such as infectious diseases, multiple sclerosis, or tumors. Whether the nonclassical HLA-G molecule is upregulated during IFN treatment or plays a role in favoring or impairing immune responses to such therapeutic approaches also remains an interesting issue. Indeed, analysis of aberrant expression of HLA-G combined with loss of classical HLA mol-ecules on melanoma cells could have important practical implications for the selection of patients likely not to benefit from interferon ␣ therapy (61).
Our results point out that IFN-␤-mediated enhancement of HLA-G expression may occur as a way to refine modulation of immune responses during pregnancy, thymic involution, or IFN-␤ treatment of autoimmune diseases and tumor progression.