of June 19, 2018.This information is current as

ResponsesTranscript 2 Receptor Suppresses B Cell Binding of HLA-G to ITIM-Bearing Ig-like

Pistoia, Edgardo D. Carosella and Nathalie Rouas-FreissAgaugué, Guitta Maki, Elisa Ferretti, Sylvie Bruel, Vito Abderrahim Naji, Catherine Menier, Fabio Morandi, Sophie

ol.1300438http://www.jimmunol.org/content/early/2014/01/22/jimmun

published online 22 January 2014J Immunol 

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2014 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on June 19, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Binding of HLA-G to ITIM-Bearing Ig-like Transcript 2Receptor Suppresses B Cell Responses

Abderrahim Naji,*,† Catherine Menier,*,† Fabio Morandi,‡ Sophie Agaugue,*,†

Guitta Maki,*,† Elisa Ferretti,‡ Sylvie Bruel,*,† Vito Pistoia,‡ Edgardo D. Carosella,*,†,1 and

Nathalie Rouas-Freiss*,†,1

Inhibition of B cells constitutes a rational approach for treating B cell–mediated disorders. We demonstrate in this article that the

engagement of the surface Ig-like transcript 2 (ILT2) inhibitory receptor with its preferential ligand HLA-G is critical to inhibit

B cell functions. Indeed, ILT2–HLA-G interaction impedes both naive and memory B cell functions in vitro and in vivo. Partic-

ularly, HLA-G inhibits B cell proliferation, differentiation, and Ig secretion in both T cell–dependent and –independent models of

B cell activation. HLA-G mediates phenotypic and functional downregulation of CXCR4 and CXCR5 chemokine receptors on

germinal center B cells. In-depth analysis of the molecular mechanisms mediated by ILT2–HLA-G interaction showed a G0/G1 cell

cycle arrest through dephosphorylation of AKT, GSK-3b, c-Raf, and Foxo proteins. Crucially, we provide in vivo evidence that

HLA-G acts as a negative B cell regulator in modulating B cell Ab secretion in a xenograft mouse model. This B cell regulatory

mechanism involving ILT2–HLA-G interaction brings important insight to design future B cell–targeted therapies aimed at

reducing inappropriate immune reaction in allotransplantation and autoimmune diseases. The Journal of Immunology, 2014,

192: 000–000.

Bcells represent an important arm of adaptive immune re-sponses through Ag presentation, as well as productionand secretion of Abs. This is the result of a complex

process of differentiation that starts when B cells encounter Agthrough engagement with the B cell Ag receptor that mediates theinternalization, processing, and presentation of Ags to T cells, animportant requirement for a successful immune response (1, 2).This engagement initiates a signaling cascade leading to B cellactivation (3), which takes place after either T cell–dependent or –independent pathways. B cells subsequently differentiate into plasmacells or memory B cells (4). Failure in this tightly regulated processcauses deleterious effects as exemplified by: 1) the autoimmune

process via autoantibodies, autoantigen presentation to and acti-vation of autoreactive T cells (5); and 2) graft rejection throughthe production of Abs against donor MHC Ags, alloantigen pre-sentation to and priming of alloreactive T cells (6, 7). Inhibition ofB cells constitutes, therefore, a rational approach for treating B cell–mediated disorders. One such treatment modality is rituximab, aB cell–depleting mAb against CD20 currently used in the treat-ment of B cell malignancies (8) and autoimmune disorders (5).Nevertheless, efforts to identify novel B cell–targeted therapies re-main essential (9). To this end, understanding B cell functions andtheir regulatory pathways is of critical importance.Like NK cells, T cells, macrophages, and dendritic cells, B cells

express ITIM-bearing receptors such as the Ig-like transcript 2(ILT2)/LILRB1/CD85j (10). ILT2 binds to the third domain ofHLA class I molecules associated with b2-microglobulin, withhighest affinity for the nonclassical HLA class I molecule HLA-G(11). Notably, the effect of HLA-G interaction with its receptor ILT2depends on its multimerization state because HLA-G multimers bindILT2 with higher affinity and slower dissociation rates than mono-mers (12). ILT2–HLA-G interaction mediates inhibition of NK andCD8+ T cell cytolysis function (13), CD4+ T cell proliferation (14),generation of suppressor T cells (15, 16), and maturation of dendriticcells (17). However, no data are currently available on the role ofILT2–HLA class I interaction in regulating B cell responses.In contrast with classical HLA class I, HLA-G shows low poly-

morphism and is expressed in a limited number of healthy tissuessuch as cytotrophoblast, thymus, cornea, and erythroblasts (18).HLA-G can be expressed as seven different isoforms includingfour membrane-bound (HLA-G1 to -G4) and three soluble (HLA-G5to -G7) proteins through alternative splicing of the primary HLA-Gtranscript. HLA-G expression is tightly regulated by environmentalfactors, such as stress, cytokines, hormones, nutrient deprivation,and hypoxia (19). Therefore, although HLA-G protein expressionis highly restricted in healthy tissues, it can be detected in manytissues under pathologic conditions such as allografts, autoimmuneprocesses, and tumor lesions (18).

*Commissariat a l’Energie Atomique et aux Energies Alternatives, Institut desMaladies Emergentes et des Therapies Innovantes, Service de Recherche en Hemato-Immunologie, Hopital Saint-Louis, 75010 Paris, France; †Universite Paris Diderot,Sorbonne Paris Cite, L’Institut Universitaire d’Hematologie, Hopital Saint-Louis, UniteMixte de Recherche E5, 75010 Paris, France; and ‡Laboratory of Oncology, G. GasliniChildren’s Hospital, 16147 Genoa, Italy

1N.R.-F. and E.D.C. share equal credit for senior authorship.

Received for publication February 25, 2013. Accepted for publication December 7,2013.

This work was supported by the Commissariat a l’Energie Atomique et aux EnergiesAlternatives, the Agence de la Biomedecine, the Italian Ministry of Health, ProgettiStrategici 2006 “Identification of Novel Receptors and/or Ligands Involved in theInteractions between Cells of the Immune System with Hematologic Tumors andMesenchymal Stem Cells” and “Role of Microenvironment in Tumor Progression.Identification of New Targets for the Development of Novel Therapeutic Strategies,”and the Associazione Italiana per la Ricerca sul Cancro.

Address correspondence and reprint requests to Dr. Nathalie Rouas-Freiss, Com-missariat a l’Energie Atomique et aux Energies Alternatives, Service de Recherchesen Hemato-Immunologie, Hopital Saint-Louis, Institut Universitaire d’Hematolo-gie, 1, Avenue Claude Vellefaux, 75010 Paris. E-mail address: Nathalie.Rouas-Freiss@cea.fr

Abbreviations used in this article: Ctrl, control nanoparticle; FDC, follicular den-dritic cell; GC, germinal center; ILT2, Ig-like transcript 2; MI, migration index; PI,propidium iodide; PIR-B, paired Ig-like receptor B; PWM, pokeweed mitogen; RM,repeated-measures; RT, room temperature; SN, supernatant; TFH, follicular helper T;TT, tetanus toxoid.

Copyright� 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1300438

Published January 22, 2014, doi:10.4049/jimmunol.1300438 by guest on June 19, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The biological functions of HLA-G have been originally de-scribed in maternal–fetal tolerance (20), and more recently inallograft acceptance and tumor escape (18). HLA-G is particularlyrelevant in the clinical setting because increased levels of solubleHLA-G have been detected in biological fluids from patients un-dergoing allotransplantation (15, 16), or suffering from inflamma-tory and autoimmune disorders (21) or solid tumors (22). In thesepathological situations, HLA-G levels were associated with betterallograft survival, reduced autoimmune activity, or tumor escapefrom immunosurveillance (18). In an attempt to identify novel B cellregulatory pathways, we have investigated the inhibitory effectsmediated by the interaction between HLA-G and ITIM-bearingIg-like receptor ILT2 in B cell biology.

Materials and MethodsCells

PBMCs were isolated from blood of healthy volunteer donors from theFrench Blood Establishment (EFS, Saint-Louis Hospital, Paris, France) bydensity-gradient centrifugation over Ficoll-Histopaque 1077 (Sigma). B cellswere isolated from PBMCs using the Dynal untouched-B cell isolation Kitaccording to the manufacturer’s instructions (Dynal). The purity of CD19+

cells was assessed by flow cytometry analysis and reached 95%. Besides,B cells were purified from human tonsils obtained from children and adultsundergoing routine tonsillectomy (Brest Medical School, Brest, France) aspreviously described (23). 3T6 cell lines either wild type or transfected withthe human CD40L were kindly provided by Dr. Nathalie Guriec (BrestMedical School). For chemokine receptor and migration experiments,mononuclear cells were isolated from tonsil by Ficoll-Histopaque densitygradient and depleted of T lymphocytes by rosetting with neuraminidase-treated sheep erythrocytes. T cell–depleted fractions contained 95% CD19+

cells. B cells purified from tonsil were incubated with CD10 mAb (DAKO)and separated by immunomagnetic beads into CD10+ germinal center(GC) (24) and CD102 non-GC B cells at 4˚C to prevent spontaneousapoptosis of GC B cells. CD10+ GC B cells contained 90% CD38high cells,as assessed by flow cytometry (25). CD10+ GC B cells were cultured 36 hwith or without rCD40L (100 ng/ml; Immunotools). CD102 non-GC B cellswere cultured 36 h with or without rCD40L and goat anti-human Ig Abs(2 mg/ml; Jackson Immunoresearch). Human TFH cells were purified fromhuman tonsil mononuclear cells using anti-ICOS mAb (Santa Cruz) andimmunomagnetic selection with anti-mouse IgG1-coated magnetic beads(Miltenyi). This positive selected cell fraction contained 95% CD4+

CXCR5high TFH cells. The human FDC-like cell line HK is derived fromfollicular dendritic cells (FDCs) of the human tonsil and was kindly pro-vided by Dr. Y.S. Choi (Ochner Clinic Foundation, New Orleans, LA). Thehuman Burkitt lymphoma cell line Raji was obtained from the AmericanType Culture Collection. All cells were cultured in RPMI 1640 mediumcontaining 10% FCS, 2 mM L-glutamine, 1% Fungizone, and gentamicin(Life Technologies). Surgically removed tonsils and normal peripheral bloodsamples were obtained after written informed consent was given; theseconsents were obtained from all patients according to the Helsinki Dec-laration, and the study was approved by our local ethics committee.

HLA-G aggregation onto nanoparticles

rHLA-G1 or rHLA-G5 was used as source of HLA-G proteins, and it isreferred to in this article as HLA-G. rHLA-G1 was produced in the humanlymphoblastoid cell line 721.221.HLA-G1 (kindly provided by Dr. Fran-cesco Puppo, Department of Internal Medicine and Medical Specialties,University of Genoa, Genoa, Italy) by transfection of the 721.221 parentalcell line with human HLA-G1 cDNA. Supernatants (SNs) were collectedfrom the 721.221.HLA-G1 cell line after 72-h culture in RPMI 1640 10%FCS at 37˚C and 5% CO2, and subsequently used as source of HLA-G afterbeing aggregated onto nanoparticles (15, 26). In brief, nanoparticles aremonodispersed magnetic particles with a 300-nm diameter and are coatedwith goat anti-mouse IgG covalently bound to their surface (Bio-Adembeadsgoat anti-mouse; Ademtech, Pessac, France). Nanoparticles were coatedwith anti–HLA-G MEM-G/9 mAb (Exbio, Vestec, Czech Republic). Afterwashing, MEM-G/9–coated particles were incubated overnight at 4˚C withHLA-G–containing SNs. The particles were then washed and resuspended inculture medium before use in experiments. Control particles were incubatedwith a medium negative for HLA-G. rHLA-G5 protein was produced in SF9insect cells infected with HLA-G5– and human b2m-baculovirus, as pre-viously described (15). The eluate containing rHLA-G5 was used as a source

of HLA-G, which was used in experiments after being aggregated ontomagnetic beads. Nanoparticles were incubated for 1 h with anti–HLA-G55A6G7 mAb (Exbio). After wash, 5A6G7-coated beads were incubatedovernight at 4˚C with HLA-G5–containing eluate. The nanoparticles werethen washed and resuspended in culture medium before use as HLA-G. Thecontrol nanoparticles (named Ctrl) were obtained in similar manner using aneluate that does not contain HLA-G proteins. HLA-G capture by nano-particles was assessed by immunoblot analysis, as previously described (15).

Cell activation

Pokeweed mitogen (PWM; Phytolacca americana; Sigma) was used toactivate PBMCs at the concentration of 2 mg/ml. PWM SN is a 0.22-mmfiltered SN obtained from PBMC cultures activated with PWM during 5 d.PWM SN was subsequently used to activate purified B cells at 1/4 finaldilution. Pansorbin (Staphylococcus aureus, Cowan strain I; Calbiochem)was used as a direct B cell activator through binding to Ig and TLR2 at thefinal dilution of 1/2000 (27). Tetanus toxoid (TT) from Pasteur Institute(Paris, France) was used at various concentrations to activate PBMCs. TonsilB cells were activated during 5 d with 75 Gy–irradiated CD40L-transfected3T6 fibroblast monolayer in presence of 50 U/ml IL-2 (Chiron) and 10 ng/mlIL-10 (Peprotech).

HLA-G treatment and sensitization

In most experiments, cells were pretreated for at least 18 h with Ctrl orHLA-G. At the end of this sensitization period, cells were washed and thenactivated and used in experiments. Alternatively, cells were treated with Ctrlor HLA-G simultaneously to the activation step. Nanoparticles were re-moved by magnetic depletion before flow cytometry analysis and in vitromigration assays. Cells were treated using 2.5 3 103 nanoparticles/cell, aspreviously described (26).

Cell proliferation

Pretreated cells with either HLA-G, Ctrl, or medium (Ø) were platedin triplicate wells at 105 cells/well, activated for 1–5 d, and pulsed with[3H]thymidine (1 mCi/well; Amersham, Biosciences). Cells were harvested18 h later, and thymidine incorporation into DNA was quantified on amicrobeta counter (Wallac 1450; Pharmacia).

Blocking experiments

To block HLA-G aggregated onto beads, we used the 87G mAb (Exbio). Inbrief, HLA-Gwere incubated for 2 h in PBS 0.1%BSA containing 50mg/mlazide-free 87G mAb. The nanoparticles were then washed and used forexperiments. To block the ILT2 receptor, we previously incubated cellswith 50% human AB serum for 1 h at 37˚C to ensure a total blockade ofmembrane-bound FcR; then we added 20 mg/ml anti-ILT2 (clone GHI-75;BD Pharmingen) for 2 h at 37˚C.

Flow cytometry

Abs used for flow cytometric analyses were conjugated with either FITC,PE, ECD, or PC5 (Beckman Coulter, BD Pharmingen, Caltag Laboratories,or R&D Systems). In brief, cells were first incubated 30 min at 4˚C in 20%human serum and subsequently labeled with Abs. Irrelevant isotype-matched Abs were systematically used. Cells were analyzed on EPICSXL4 flow cytometer using Expo32 software (Beckman Coulter) or onFACSCalibur (BD) where at least 5 3 103 events were acquired and an-alyzed using the CellQuest software (BD). Results are expressed as per-centage of positive cells or mean of relative fluorescence intensity obtainedas follows: mean fluorescence intensity obtained with specific Ab 4 meanfluorescence intensity obtained with irrelevant isotype-matched Ab.

Immunofluorescence

Cells were pretreated with either HLA-G, Ctrl, or medium (Ø) and thenactivated. After 5 d, cells were harvested from the culture flasks and cyto-spins were prepared using SuperFrost/Plus slides (Merck) and a Cytospin 3(Shandon). For staining, cells were fixed and permeabilized in 90% ethanol,causing disruption of the cell membrane, and then incubated for 30 min withFITC-labeled goat anti-human Ig (Beckman Coulter) recognizing the mainisotypes (IgA, IgD, IgG, IgM) or FITC-labeled isotype-matched controlAb (Beckman Coulter). Indeed, after cell fixation with ethanol, no surfaceIg could be detected while allowing detection of intracytoplasmic Igs.Nuclei were labeled in red with propidium iodide (PI; Sigma). Slides wereanalyzed using a fluorescence microscope (Bio-Rad MRC1024; Bio-Rad).The percentage of cells positive for intracytoplasmic Ig was then deter-mined by counting cells with fluorescent cytoplasm per 100 nuclei countedin a representative area.

2 B CELL REGULATION BY HLA-G/ILT2 INTERACTION

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Chemotaxis

Chemotaxis was investigated using 5-mm pore-size transwell plates (Co-star) as described previously (28). Five hundred thousand cells were dis-pensed in the upper chamber, whereas chemokines or medium alonewas added to the lower chamber. CXCL12 (Immunotools) and CXCL13(Abnova) were tested at 300 ng/ml. Plates were incubated for 2 h at 37˚C.Migrated cells were collected and counted, and migration index (MI) wascalculated as following: (number of migrated cells/number of dispensedcells) 3 100.

Biochemistry

Cells were incubated for various times in the presence of HLA-G or Ctrl.Total proteins were then extracted from cell lysates and quantified usingBCA protein assay kit (Pierce). Total protein extracts (30 mg) were sub-jected to SDS-PAGE. Proteins were then resolved on 10% SDS-PAGE,except for mTOR proteins (6% SDS-PAGE), and then transferred ontonitrocellulose membranes. Membranes were subjected to immunoblottingusing Abs to proteins or phospho-proteins from AKT pathway and Foxopathway kits (Cell Signaling). Membranes were subsequently probed witha-tubulin (Sigma). Quantification of blotted proteins was performed bydensitometry of scanned films using the Fluorchem software (Alpha-Innotech).Immunoblotting of HLA-G was performed using 12% SDS-PAGE and theanti–HLA-G 4H84 mAb (15).

Cell cycle

Cells were pretreated with HLA-G, Ctrl, or medium (Ø). Cells were washedand fixed in 70% ethanol in PBS and incubated at 4˚C overnight. Washedcells were incubated in PBS containing 40 mg/ml PI (Sigma) and 100 mg/ml DNase-free RNase A on ice for at least 10 min, as previously described(26). Cell cycle parameters were acquired using LSR flow cytometer andCellQuest software (Becton Dickinson). Cell cycle distribution was de-termined by automatic analysis using the flow cytometry analysis softwareFlowJo. The percentage of cells in each cell cycle phase, that is, G0/G1, S,and G2/M, is provided.

Apoptosis

Apoptosis induction was evaluated using Annexin VFITC/PI kit (BeckmanCoulter) according to manufacturer’s instructions, and stained cells wereanalyzed using EPICS XL4 flow cytometer and Expo 32 software (BeckmanCoulter). Cells treated with 100 ng/ml staurosporine (Sigma) were usedas positive control.

HLA-G–specific ELISA

Concentration of soluble forms of HLA-G was evaluated on SNs from thehuman FDC line HK (kindly provided by Dr. Y.S. Choi) (29), and fromhuman TFH cells that had been cultured for 48 h in RPMI 1640 10% FCS inpresence or absence of 50 ng/ml IL-10 (Boehringer Ingelheim) or 100 pg/ml TGF-b (R&D Systems). HLA-G ELISA was performed using Max-iSorp Nunc-Immuno 96 microwell plates (Nunc) coated overnight at 4˚Cwith mAb MEM-G/9 (Exbio) in 0.001 M PBS, pH 7.4. After three washeswith PBS 0.05% Tween 20, plates were saturated with 200 ml PBS 2%BSA for 30 min at room temperature (RT); 100 ml of samples or standardwas added to each well and incubated at RT for 1 h. Plates were washedthree times and then incubated with 100 ml biotinylated anti-b2m mAbNAMB-1 at RT for 1 h (kindly gifted by Dr. Soldano Ferrone). After threewashes, plates were incubated at RT for 1 h with streptavidin-horse radishperoxidase (GE Healthcare) 1:4000 in PBS 0.1% Tween 20, 0.1% BSA, for1 h at RT. After three additional washes, plates were incubated with theTMB substrate (Sigma) for 5 min at RT. H2SO4 5 M was then added, andoptical densities were measured at 450 nm. The assay’s lowest thresholdwas 2 ng/ml HLA-G. Each sample was tested in duplicate.

Cytokines and Ig ELISA

SNs were harvested and used in ELISA tomeasure levels of cytokines (IL-2,IFN-g, IL-4, and IL-10) or Ig (IgA, IgG, and IgM). IL-2, IFN-g, IL-4, andIL-10 were measured by Th1/Th2 human ELISA Kit (eBioscience), ac-cording to the manufacturer’s instructions. Human Ig present in SN wasmeasured using human IgA, IgG, and IgM ELISA quantitation kits (Bethyl),according to the manufacturer’s instructions. Plasma from mice were har-vested after centrifugation of blood 10 min at 1800 rpm, aliquoted, andstored at 280˚C until ELISAwere performed. Total IgG was quantifiedusing the mouse total IgG EasyTiter ELISA kit (Pierce). Mouse cytokineswere quantified using the Th1/Th2/Th17 Multi-Analyte ELISArray kit(Tebu-Bio).

Slot-blot analysis

A total of 5 3 106 M8-pcDNA cells were washed with cold PBS, and thepellet was lysed and boiled for 5 min. After centrifugation, the SN wasblotted on a 10% SDS-PAGE, and proteins were transferred on Immobilon-P (Millipore, Bedford, MA). Blots were saturated with 5% fat-free milk/PBS/0.2% Tween 20 for 1 h and then incubated with different mouseplasma samples diluted in PBS 1:100 using the mini-Protean II multiscreenapparatus (BioRad). Detection was performed using HRP-conjugated anti-mouse Ab (Sigma-Aldrich) with the ECL kit (Amersham). Incubationswith mouse plasmas and secondary goat anti-mouse Abs were conductedin PBS 0.2% Tween 20, 0.5% BSA.

Mice and immunizations

Female BALB/c mice (6–10 wk of age) were obtained from Charles RiverLaboratories (L’Arbresle, France). Groups of mice were injected s.c. with10 3 106 human melanoma M8-pcDNA (HLA-G2 xenogeneic cells) orM8–HLA-G cells (HLA-G+ xenogeneic cells) (13) resuspended in PBS.When indicated, the human cells were pretreated for 1 h with the anti–HLA-G 87G mAb at 20 mg/ml (Exbio) or the isotypic control (IgG2a;Exbio) before graft. The xenograft area was measured with digital calipersat indicated time points postgraft. Volume was estimated by the followingformula (L 3 W2)/2, with L = length and W = width in millimeters. Allexperimental protocols were approved by the ethics review committee foranimal experimentation of the Saint-Louis Hospital (Paris, France) andfollowed the guiding principles for the care and use of animals approvedby our local committee.

Statistical analysis

Statistical analysis has been performed using GraphPad Prism (GraphPadSoftware). Gaussian distribution of data was tested using Kolmogorov–Smirnov test. Mann–Whitney U test (2 groups) or Kruskal–Wallis test ($3groups) has been performed when Gaussian distribution was not assumed.One-way ANOVA test has been performed for experiment data assumingGaussian distribution and comparing $3 groups at a single time point.Besides, a multiple-comparisons test has been performed to compare sta-tistical significance between groups (Tukey’s multiple-comparisons test).In time-series experiments with data comparing $3 groups, we performed arepeated-measures (RM) two-way ANOVA test; then a multiple-comparisonstest (Tukey’s multiple-comparisons test) has been performed. Significancewas provided with the results assuming p , 0.05 as significant and weremarked with asterisks in the figures.

ResultsHLA-G inhibits T cell–dependent B cell response throughinteraction with ILT2

Among the HLA class I molecules, HLA-G constitutes the pref-erential ligand of ILT2, which is the sole known HLA-G receptorexpressed by CD19+ B cells (Fig. 1A). To elucidate the effects ofILT2 engagement by soluble HLA-G, we first investigated the pro-liferative responses of PBMCs polyclonally activated with PWM,which triggers B cells via a T cell–dependent pathway. Becauseaggregation of HLA-G is critical for optimal interaction with ILTs(15, 17), PBMCs were activated in the absence or presence ofHLA-G aggregated onto nanoparticles (i.e., HLA-G). As control,we used nanoparticles processed in the same manner as for HLA-Gnanoparticles but using an HLA-G–free source (i.e., Ctrl). HLA-Gcapture by nanoparticles was assessed by immunoblot analysis(Fig. 1B, insert). Results show that HLA-G strongly inhibitedPWM-induced PBMC proliferation compared with controls (Fig.1B). In parallel to HLA-G aggregated onto nanoparticles, we usedan rHLA-G protein unbound to magnetic particles and found that,although similar, its effects were less potent than those mediatedby aggregated HLA-G (data not shown). Of note, we have testeddenatured HLA-G protein coated onto nanoparticles that does nolonger inhibit B cell proliferation. Such denatured HLA-G proteinis no longer capable of binding to ILT2 (data not shown).The HLA-G–driven inhibition of PWM-activated PBMC pro-

liferation was dose- and time-dependent (Fig. 1C) and could be

The Journal of Immunology 3

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

reversed by neutralizing ILT2–HLA-G interaction with mAbsagainst HLA-G (anti–HLA-G; Fig. 1D) or ILT2 (a-ILT2; Fig. 1E).The maximum inhibition was reached when HLA-G was added inthe early steps of PWM activation (Fig. 1F) or upon preincubationof PBMCs with HLA-G for 18 h before PWM stimulation (Fig.1G). Besides, stimulating B cells with anti-human Ig and CD40L,as well as CpG, has brought similar results showing that in thoseconditions of stimulation, ILT2–HLA-G interaction strongly in-hibits B cell proliferation (data not shown). Moreover, cytokineassays carried out with PBMCs incubated in the presence of HLA-G showed: 1) reduced levels of IFN-g and, to a lesser extent, IL-2at day 5; and 2) increased IL-10 concentrations earlier at day 2compared with control (Fig. 1H).We then investigated the effects of HLA-G on PWM-induced

B cell differentiation into Ab-secreting cells. A significantly re-duced number of cells with intracytoplasmic Ig (Ig+ cells), as well

as decreased levels of IgA, IgG, and IgM, were observed in thepresence of HLA-G (Fig. 1I, 1J). Interestingly, both T and B cellsare targets of the HLA-G–mediated inhibition with maximumeffect observed when both cell populations had been pretreatedas assessed by HLA-G–mediated proliferation inhibition (Fig.2A). The secretion of IFN-g and IL-2 (Fig. 2B), as well as of IgA,IgG, and IgM (Fig. 2C), was reduced when T and/or B cells werepretreated with HLA-G. In agreement with data obtained withPBMCs, T cell help proliferation and differentiation of CD19+

B cells isolated from tonsil were inhibited by HLA-G (Fig. 3A,3B).

HLA-G inhibits T cell–dependent Ag-specific memory B cellresponses

To investigate the effects of HLA-G on specific B cell responses to arecall Ag, we stimulated PBMCs with TT. Proliferation of PBMCs

FIGURE 1. HLA-G inhibits T cell–dependent activation of peripheral blood B cells through ILT2 interaction. (A) Expression of known HLA-G receptors

was evaluated by flow cytometry analysis on CD19+ PBMCs. Numbers represent the percentage of cells within the corresponding region; one representative

experiment of four is shown. (B and C) PBMCs were activated by PWM in the absence (Ø) or presence of Ctrl or HLA-G aggregated onto nanoparticles

(HLA-G). Proliferation was evaluated on days 1, 3, and 5. Results are given as mean 6 SEM; significance was evaluated with RM two-way ANOVA test.

(B, inset) Capture of HLA-G by nanoparticles compared with Ctrl was assessed by immunoblot using anti–HLA-G mAb; one representative experiment of

five is shown. (B) Cells (105/well) were treated using 2.5 3 103 nanoparticles/cell, which corresponds to 25 3 107 nanoparticles/well. (D and E) PBMCs

were pretreated with Ctrl or HLA-G for 18 h and then activated with PWM. Neutralizing anti–HLA-G (a-HLA-G) (D) or anti-ILT2 (a-ILT2) (E) mAb was

used to prevent ILT2–HLA-G interactions. Proliferation was evaluated at day 5 (D) and day 3 poststimulation (E). Results are given as mean 6 SEM,

significance was evaluated with one-way ANOVA test. (F) PBMCs were activated by PWM and then HLA-G were added after 24, 48, or 96 h. (G) PBMCs

were first pretreated with HLA-G for 2, 6, or 18 h. Then HLA-G–coated nanoparticles were removed and PBMCs were activated by PWM. (Ø) refers to

cells cultured in medium for 18 h before PWM treatment. (F and G) Proliferation was evaluated at day 5 poststimulation. (B–G) Results are means of

thymidine incorporation corrected for background values (Dcpm) 6 SEM from five independent experiments. (H) IFN-g, IL-2, and IL-10 levels were

measured by ELISA at days 2 and 5 poststimulation in SNs containing PBMCs activated with PWM in the presence of Ctrl or HLA-G or medium alone (Ø).

Results are given as mean 6 SEM (n = 4); significance was evaluated with RM two-way ANOVA test. (I) PBMCs were activated by PWM in the absence

(Ø) or presence of Ctrl or HLA-G aggregated onto nanoparticles (HLA-G). The percentage of intracytoplasmic Ig+ cells was evaluated by immunofluo-

rescence (n = 6). Original magnification 325; one representative experiment is shown. (J) IgA, IgG, and IgM concentrations were measured by ELISA at

day 5 poststimulation of PBMCs activated by PWM in the absence (Ø) or presence of Ctrl or HLA-G aggregated onto nanoparticles (HLA-G). Results are

given as mean 6 SEM (n = 4); significance was evaluated with Kruskal–Wallis test. *p , 0.05.

4 B CELL REGULATION BY HLA-G/ILT2 INTERACTION

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

in response to TT was strongly inhibited by HLA-G, and bothcytokine and Ig secretions were affected by HLA-G with decreasedsecretion of IFN-g, IL-2, IgA, IgG, and IgM (Fig. 3C–F). Decreasednumbers of Ig+ cells were observed among PBMCs after 5-dactivation with TT in the presence of HLA-G (Fig. 3G). Also,reduced expression levels of CD20 and CD69 were observedwithin the CD19+ cells at day 5 after TT stimulation (Fig. 3H, 3I).

HLA-G inhibits T cell–independent B cell responses

Next, we examined the effects of HLA-G on B cells activated byT-independent stimuli using pansorbin or SNs from PWM-acti-vated PBMCs (PWM SN). HLA-G inhibited the proliferation ofPBMCs, as well as of purified B cells, after pansorbin or PWM SNactivation (Fig. 4A). This inhibitory effect was associated with: 1)reduced levels of IFN-g and IL-2 together with enhanced IL-10secretion when purified B cells were stimulated with pansorbin(Fig. 4B); 2) decreased numbers of intracytoplasmic Ig+ cells amongactivated PBMC or purified B cells (Fig. 4C); 3) decreased se-cretion of IgA, IgG, and IgM in culture SNs from purified B cellsstimulated with pansorbin (Fig. 4D); and 4) reduced expressionlevels of CD20, CD69, and CD138 within the CD19+ cells whenpurified B cells were stimulated with pansorbin (Fig. 4E, 4F).

HLA-G downregulates CXCR4 and CXCR5 chemokinereceptor expression on GC B cells and dampens theirchemotaxis to the respective ligands

The chemokine receptors CXCR4 and CXCR5 regulate B celltrafficking within secondary lymphoid follicles (30). Therefore,we examined whether their expression on tonsil CD19+ B cellsubsets was modulated by HLA-G. Results showed that CXCR4expression was significantly downregulated in CD10+ GC B cellstreated with HLA-G compared with control (Fig. 5A, 5C). Con-

versely, CXCR4 expression in CD102CD272 naive and CD102

CD27+ memory B cells was unaffected by HLA-G treatment (Fig.5A, 5C). CXCR5 expression was significantly lower in tonsil GC,naive, and memory B cells treated with HLA-G compared withcontrol (Fig. 5B, 5D). When the same experiments were performedusing peripheral blood naive and memory B lymphocytes, no sig-nificant modulation of CXCR4 or CXCR5 expression by HLA-Gwas observed. GC B cells and non-GC B cells, that is, CD10+ andCD102 B cells, from tonsil were next subjected to in vitro migra-tion assays, using CXCL12 and CXCL13 as chemoattractants.Migration to CXCL12 was significantly downregulated in GCB cells treated with HLA-G, whereas migration of non-GC B cellswas unaffected by HLA-G (Fig. 5E, 5G). Similar results wereobtained using CXCL13 as chemoattractant. MI was significantlylower in GC B cells treated with HLA-G compared with controls,whereas the MI of non-GC B cells was not altered by HLA-G(Fig. 5F, 5H).

HLA-G is secreted by FDCs and TFH cells

We then examined whether the human FDCs cell line HK (29) andfreshly isolated human TFH cells that home in the follicular mi-croenvironment of human tonsil secreted HLA-G in vitro with orwithout known HLA-G release inducers (19). HLA-G was detectedin SNs from the HK cell line in all experimental conditions, whereasTFH cells were found to release HLA-G only upon incubation witheither IL-10 or TGF-b1 (Fig. 5I, 5J).

HLA-G induces G0/G1 cell cycle arrest and acts as a negativeB cell regulator through AKT signaling pathway

To address the precise mechanisms underlying HLA-G–mediatedinhibition, we analyzed B cells for cell cycle progression andapoptosis. Because activated B cells proliferated slowly, the Raji

FIGURE 2. Both T and B cells are targets of the HLA-G–mediated inhibition. (A–C) Cocultures were performed with purified T and B cells isolated from

the same donor and were individually sensitized or not with HLA-G before stimulation with PWM. Experiments were: 1) Ctrl-treated T cells cocultured

with Ctrl-treated B cells [T(Ctrl) + B(Ctrl)]; 2) Ctrl-treated T cells cocultured with B cells pretreated with HLA-G [T(Ctrl) + B(HLA-G)]; 3) T cells

pretreated with HLA-G cocultured with Ctrl-treated B cells [T(HLA-G) + B(Ctrl)]; and 4) T cells pretreated with HLA-G cocultured with B cells pretreated

with HLA-G [T(HLA-G) + B(HLA-G)]. (A) Cell proliferation was evaluated at day 5 poststimulation with PWM. Results are expressed as mean values

(Dcpm) 6 SEM (n = 5). Comparisons between (2), (3), and (4) show that T and B cells pretreated with HLA-G have a synergistic on cell proliferation

inhibitory effect. Significance was evaluated with one-way ANOVA test. (B) IFN-g, IL-2, and IL-10 concentrations were measured by ELISA at day 5

poststimulation (n = 3). Significance was evaluated with Kruskal–Wallis test. Results show that INF-g and IL-2 are reduced significantly with a synergistic

effect when both T and B cells were pretreated with HLA-G (C) IgA, IgG and IgM concentrations were measured by ELISA at day 5 poststimulation (n =

3). Significance was evaluated with Kruskal–Wallis test. Results show no significance between (3) and (4) conditions. *p , 0.05.

The Journal of Immunology 5

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

cell line was used as a B cell model as previously described (31).Indeed, some features of Raji cells may resemble the state ofactivated B cells, and proliferation of Raji cells was inhibited byHLA-G in a dose- and time-dependent manner similar to normalB cells (32). HLA-G was found to alter cell cycle progression ofRaji B cells (Fig. 6A, 6B), without any detectable apoptosis byusing Annexin V/PI or DIOC6 staining (Fig. 6C). Supporting thesedata, additional experiments allowed us to exclude decreased cellviability because of the presence of HLA-G: 1) in our preliminarytests, we did not find any difference in cell viability with trypanblue staining whether B cells were activated in the presence orabsence of HLA-G; 2) in blocking experiments using anti–HLA-Gor anti–ILT2 mAb in the presence of HLA-G, PBMC proliferation

could be restored, demonstrating that cells were still viable afterHLA-G treatment (Fig. 1D, 1E); and 3) no inhibition of cell pro-liferation was observed when HLA-G was added in the later stepsof PWM activation (96 h), showing in this study again that HLA-Gtreatment did not alter cell viability (Fig. 1F).To identify the intracellular signaling cascades involved in

ILT2/HLA-G–mediated inhibition, we analyzed the AKT signal-ing pathway, which is frequently dysregulated in B cell disorders(3). HLA-G decreased the PWM SN–induced phosphorylation ofAKT at Ser473 in primary B cells after 12-h activation (Fig. 6D)without modifying AKT protein levels (data not shown). An essentialfunction of AKT signaling is to regulate cell cycle progressionthrough phosphorylation and inactivation of GSK-3b, c-Raf, and

FIGURE 3. HLA-G inhibits tonsil B cell– and T cell–dependent specific memory B cell responses. (A and B) Tonsil B cells were incubated or not with

HLA-G for 18 h and then stimulated by coculture with gamma-irradiated 3T6 cells expressing CD40L in presence of IL-2 and IL-10. (A) Cell proliferation

was evaluated at day 5 poststimulation. Results are expressed as means of thymidine incorporation corrected for background values (Dcpm)6 SEM (n = 4).

Significance was evaluated with Kruskal–Wallis test. (B) IgG concentrations were determined by ELISA at day 5 poststimulation (n = 3). Results are given

as means 6 SEM, and significance was evaluated with Kruskal–Wallis test. (C) PBMCs from healthy donors were activated by increasing doses of TT or

(D–I) by 10 mg/ml TT. Before activation, PBMCs were either pretreated with medium alone (Ø), Ctrl, or HLA-G. (C) Cell proliferation was evaluated at day

5 poststimulation or (D) at various times after TT stimulation. Results are expressed as means of thymidine incorporation corrected for background values

(Dcpm) 6 SEM (n = 3); significance was evaluated with RM two-way ANOVA test. (E) IFN-g, IL-2, and IL-10 levels were measured by ELISA at day 5

poststimulation. Results are given as mean6 SEM (n = 3); significance was evaluated with Kruskal–Wallis test. (F) IgA, IgG, and IgM concentrations were

measured by ELISA at day 5 poststimulation. Results are given as mean 6 SEM (n = 3); significance was evaluated with Kruskal–Wallis test. (G) Per-

centage of intracytoplasmic Ig+ cells was evaluated by immunofluorescence (n = 4). Original magnification 340; one representative experiment is shown.

(H) PBMCs were treated with Ctrl or HLA-G and stimulated with TT. At day 5 poststimulation, CD20, CD38, CD69, and CD138 cell surface expression

were analyzed by flow cytometry according to CD19 expression (n = 3). One representative dot plot is shown. (I) PBMCs were treated with Ctrl or HLA-G

and stimulated with TT, at day 5 poststimulation; means 6 SEM of the percentage of CD19+ cells expressing CD20, CD38, CD69, and CD138 markers are

given. Significance was evaluated with a Mann–Whitney U test. *p , 0.05.

6 B CELL REGULATION BY HLA-G/ILT2 INTERACTION

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Foxo proteins (33, 34). In this study, we found that HLA-G de-creased the phosphorylation of these downstream signaling pro-teins GSK-3b (Fig. 6D), c-Raf (Fig. 6D), and Foxo1, Foxo3a, andFoxo4 (Fig. 6D).

HLA-G promotes a Th2 response and inhibits B cell Absecretion, therefore inducing in vivo immune tolerance ina murine xenograft model

To provide in vivo evidence of the inhibitory effect mediated byHLA-G on B cells, we finally developed a xenograft model in mice(Fig. 7A). Although there is no murine homolog of HLA-G, thisstudy was made possible by the fact that human HLA-G interactswith the murine receptor paired Ig-like receptor B (PIR-B), thehomolog of human ILTs (35, 36). In this regard, tetramers ofHLA-G were found to prolong skin allograft survival in mice byinducing a tolerogenic environment (37). In agreement with these

data, we found that HLA-G+ xenograft displayed extended sur-vival and growth compared with HLA-G2 counterpart, and thatblocking HLA-G function by a specific Ab inhibited xenograftgrowth in immunocompetent mice (Fig. 7B). We then analyzedthe Ab response from mice injected with HLA-G2 or HLA-G+

xenogeneic cells. Although the total amount of plasma IgG wassimilar between mice groups (Fig. 7C), we observed a dramaticdecrease of anti-xenograft Abs present in the plasma of HLA-G+

xenograft-bearing mice compared with HLA-G2 counterpart atday 27 postgraft (Fig. 7D). To correlate with the slot-blot data (Fig.7D), HLA-G+ versus HLA-G2 xenograft growth over time was asfollows: 270 versus 40 mm3 at day 7, 70 versus 0 mm3 at day 13,0 versus 0 mm3 at days 20 and 27. These results demonstrate thatAg-specific B cell response is drastically affected in HLA-G+

xenograft-bearing mice. Ab data were strengthened by flow cytom-etry experiments showing that the level of mouse IgG specific for

FIGURE 4. HLA-G inhibits both T cell–dependent and –independent activation of peripheral blood B cells. (A–F) PBMCs or purified B cells from

healthy donors were activated by pansorbin or PWM SN. Cells were pretreated with Ctrl, HLA-G, or not (Ø). (A) Cell proliferation was evaluated at day 5

poststimulation. Results are expressed as mean values (Dcpm) 6 SEM (n = 5). Significance was evaluated with one-way ANOVA test. (B) IFN-g, IL-2, and

IL-10 concentrations were measured by ELISA at days 2 and 5 poststimulation of purified B cells with pansorbin. Results are given as mean 6 SEM (n = 3);

significance was evaluated with RM two-way ANOVA test. (C) The percentage of cells positive for intracytoplasmic Ig (Ig+ cells) was evaluated by

immunofluorescence (n = 4). One representative experiment is shown. Original magnification 340. (D) IgA, IgG, and IgM concentrations were measured

by ELISA at day 5 poststimulation of purified B cells with pansorbin (n = 3). Results are given as mean 6 SEM; significance was evaluated with Kruskal–

Wallis test. (E) Purified B cells stimulated with pansorbin were analyzed by flow cytometry within the CD19+ cells for CD20, CD38, CD69, and CD138

cell-surface expression; one representative experiment is shown of three. (F) Means 6 SEM of the percentage of CD19+ cells expressing CD20, CD38,

CD69, and CD138 markers are given (n = 3). Significance was evaluated with Kruskal–Wallis test. *p , 0.05.

The Journal of Immunology 7

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

xenografted cells was quantitatively reduced in HLA-G+ xenograftmice compared with HLA-G2 (data not shown).Finally, we analyzed the presence of Th1 versus Th2 cytokines in

mice plasma. Although IFN-g and IL-2 levels were decreased (Fig.7E–H), IL-4 and IL-10 were detected at high levels (Fig. 7G, 7H)in mice injected with HLA-G+ cells compared with HLA-G2

cells. This cytokine environment in mice injected with HLA-G+

cells is strongly in favor of a Th2 immune response.

DiscussionImmune regulation is achieved by the integration of positive signalsdelivered by Ag receptors, costimulatory/adhesion molecules, andcytokine receptors, with negative signals provided by inhibitoryreceptors. Identifying these receptor–ligand interactions is im-portant for understanding and elucidating mechanisms that pre-vent leukocyte overactivation that could potentially result in tissuedamage/autoimmunity. These molecules also potentially representtargets for therapies directed at modulating immune responsive-ness.Although effects of the ILT2–HLA-G receptor–ligand interac-

tion have been described extensively for myeloid and T cells, the

only effect of ILT2 ligation on B cells shown previously was usingan anti-ILT2 mAb (clone HP-F1) (10). In that study, pretreatmentof an EBV-transformed B cell line with anti-ILT2 mAb inhibitedBCR-induced Ca2+ flux. In that same study, the authors reportedthat anti-ILT2 mAb had a minimal effect on primary B cells, ar-guing that a higher order magnitude of ILT2 cross-linking may benecessary to have an effect on primary cells. Our study is the onlyone to date that demonstrates an effect of ILT2 via HLA-G onvarious primary B cell responses.We demonstrate that HLA-G aggregated onto nanoparticles in-

hibits both T cell–dependent and –independent naive and memoryB cell responses through ILT2 interaction. This inhibition targetsB cell proliferation, as well as B cell differentiation, into Ab-secreting cells. The rationale of HLA-G aggregation onto nano-particles lies in the fact that it mimics the active polymeric structurefound in vivo (11). Indeed, both crystallography and biochemicalstudies have shown that ILT2 binds mostly cell-surface multi-mers in vitro and in vivo, and that multimers signal more effi-ciently than monomers (12, 38). Moreover, HLA-G aggregationhas shown its effectiveness in previous functional studies (15, 26,37).

FIGURE 5. HLA-G affects CXCR4 and CXCR5 expression on tonsil GC B cells leading to migration inhibition. (A and C) CXCR4 and (B, D) CXCR5

expression was evaluated on purified CD10+ (GC) and CD102 (non-GC) B cells, gating on CD27+ B cells and CD272 B cells. Cells were stimulated and

cultured with Ctrl or HLA-G before being analyzed. (A and B) Results are expressed as mean of relative fluorescence intensity, n $ 4 experiments;

significance was evaluated with Mann–Whitney U test. (C and D) Cell expression of CXCR4 and CXCR5 was assessed within the CD10+ or CD102 subsets

by flow cytometry. One representative experiment is shown for at least four experiments. (E–H) Purified CD10+ B cells (GC) and CD102 B cells (non-GC)

were stimulated and cultured with Ctrl or HLA-G before being subjected to in vitro migration assay using as chemoattractant (E, G) CXCL12 or (F, H)

CXCL13. Results are expressed as MI. Results are given as means 6 SEM (n = 6). Significance was evaluated with Mann–Whitney U test. (I and J)

Quantification of HLA-G in SNs from HK cell line and TFH cells cultured for 48 h with medium alone (Ø), 50 ng/ml IL-10, or 100 pg/ml TGF-b. Histogram

shows means 6 SEM (n = 3). *p , 0.05.

8 B CELL REGULATION BY HLA-G/ILT2 INTERACTION

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Furthermore, HLA-G downregulates IL-2 and IFN-g, whereasit increases IL-10 production by activated PBMCs and purifiedB cells, suggesting the occurrence of synergistic inhibitory activityof HLA-G and IL-10 that are interdependent in some experimentalsystems (39). Also, we found that HLA-G favors a Th2 cytokineprofile at the expenses of Th1 cytokines in vitro and in vivo.A striking finding was the ability of HLA-G to inhibit B cell Ab

production as shown by reduced Ig secretion irrespective of theB cell activation protocols used. Such an HLA-G inhibitory role onB cell response is in accordance with clinical observations showingreduced HLA alloantibody levels in kidney transplant patients (40)or inhibition of the humoral response in heart transplant patientswho express HLA-G (41).We show that HLA-G dampens expression of CXCR4 and

CXCR5 in GC B cells. Accordingly, the latter cells exhibited a dra-matically reduced chemotaxis to CXCL12 andCXCL13. CXCR5wassignificantly downregulated by HLA-G in naive and memory B cellsfrom tonsil, but migration of these cell fractions to CXCL13 wasunaffected. Taken together, our findings suggest that HLA-G par-ticipates in the control of B cell trafficking in the GC. This hypothesisis supported by our findings showing the secretion of HLA-G by theFDC HK cell line and purified tonsil TFH cells in vitro.HLA-G reduces B cell proliferation by inducing a G0/G1 cell

cycle arrest. We defined signaling events leading to HLA-G in-

hibitory effects through decreased phosphorylation of AKT,GSK-3b, c-Raf, and Foxo proteins. All these effects convergeto activate inhibitors or to inhibit activators of cell survival andproliferation. Presumably, SHP-1 is responsible for HLA-G–mediateddephosphorylation of AKT, GSK3-b, and Foxo proteins. Indeed,SHP-1 has been identified as the phosphatase associated with theITIM of ILT-2 on B cells, whereas no association was detectedwith SHP-2 and SHIP (10, 42). Altogether, these data show thatHLA-G plays a critical role in regulating B cell fate decision, suchas cell proliferation and differentiation.Disruption of the delicate balance between activating and in-

hibitory signals regulating normal B cell activation can lead to theproduction of autoantibodies and autoimmunity (5), or alloanti-bodies responsible for allograft rejection (7). In these clinical set-tings, HLA-G is particularly relevant because increased levels ofHLA-G have been described in plasma from solid organ transplantpatients with better graft acceptance (15, 39), and in plasma andcerebrospinal fluid from multiple sclerosis patients with reduceddisease activity (43). Notably, overexpression of ILT2 has beenreported in the course of the same pathologies (44). Our studybrings a new highlight on HLA-G–induced tolerance operating inthese patients that is likely achieved by acting on the B cell re-sponse. Indeed, we show in this article the in vivo inhibitory roleof HLA-G on B cells through the use of a xenograft model. This

FIGURE 6. HLA-G induces G0/G1 cell cycle arrest and inhibits B cells via AKT pathway. (A) Raji B cells were pretreated with Ctrl or HLA-G for 18 h

or medium (Ø), and analyzed after 24 h. Cell cycle was determined by PI staining. The fraction of cells in cycle was normalized to the fraction in S phase.

One representative experiment of six is provided as histograms and dot plot. (B) Percentage of cells in G0/G1, G2/M or S phase are provided as means 6SEM (n = 6). Significance was evaluated with one-way ANOVA test. (C) B cells were pretreated with Ctrl or HLA-G for 18 h or medium (Ø), and apoptosis

induction was assessed at 24 and 48 h by Annexin V (AV) and PI staining using FACS analysis. Staurosporine was used as positive control. One rep-

resentative experiment of six is shown. (D) Purified B cells were sensitized with Ctrl or HLA-G for 18 h and then stimulated with PWM SN for 12 h.

Phosphorylation of AKTSer473, GSK-3b, c-Raf, and Foxo1, Foxo3a, and Foxo4 was detected by immunoblot analysis with phosphospecific Abs as indi-

cated. a-Tubulin was used to assess that same amounts of proteins were run in polyacrylamide gels. One representative experiment of three is shown. Green

arrows indicate decreased phosphorylation. Numbers indicate the relative protein quantifications as mean 6 SEM (n = 3) using Ctrl values as baseline.

Significance was evaluated with Mann–Whitney U test. AKT protein does not vary with Ctrl or HLA-G. *p , 0.05.

The Journal of Immunology 9

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

approach was made achievable because human HLA-G can bindand mediate an inhibitory signal via the murine receptor PIR-B, thehomolog of human ILTs (36, 37). The substantial difference be-tween the species (ILT-2 and PIR-B) is the absence of PIR-B ex-pression on murine T and NK cells, whereas ILT-2 is expressed onboth human subsets. On B cells, both ILT-2 and PIR-B are ex-pressed and have been shown to bind to SHP-1 (10, 35). Thus, thedifferences in effect of HLA-G between human and murine B cellimmunity might be because of the fact that HLA-G targets T andNK cells and might affect indirectly B cell functions in human.Our data provided in this work show that when both T and B cellsare sensitized by HLA-G, the effect on proliferation inhibition ismore profound compared with controls. Furthermore, our resultsshow a decreased Ab secretion directed against the xenoantigens inHLA-G+ xenograft-bearing mice compared with their HLA-G2

counterpart. In agreement with the results obtained with humanPBMCs, HLA-G+ xenograft was associated with a shift toward aTh2 cytokine plasma pattern. Indeed, although Th1 cytokines, that is,IL-2 and IFN-g, were detected at high levels in HLA-G2 xenograft-bearing mice, they were quasi-absent in HLA-G+ xenograft-bearingmice. Collectively, these in vivo data led us to hypothesize that atfirst steps, the immune response is specific for the xenograft, and inthis regard the immune protection exerted by HLA-G may targetxenograft-specific immune cells. However, afterward, suppressive

mechanisms involving immunoregulatory cells induced by HLA-Gmay take over by creating a broader immunosuppression towardxenograft-specific and nonspecific cells. Thus, both global and graft-specific immunosuppression may account for the HLA-G–mediatedimmune protection of the xenograft.In conclusion, our findings describe a novel regulatory pathway

of B cells through ITIM-bearing receptor signaling after HLA-Ginteraction and delineate potential targets for therapies of B cellhyperactivity disorders.

AcknowledgmentsWe thank Dr. Nathalie Guriec for providing tonsillar B cells, James Vigneron

and Sylvaine Labaume for technical assistance, Carmen Botella and

Priscila Vianna for contribution to this study, Dr. Marika Pla and Martine

Chopin for animal care facilities, andMichael Yagello and Niclas Setterblad

from the Imagery Department (Institut Universitaire d’Hematologie/Institut

Federatif de Recherche 105, Paris, France) The Imagery Department was

supported by the Conseil Regional d’Ile de France, the Canceropole Ile de

France, and the Ministere de la Recherche. We thank Dr. Lizzia Raffaghello

(Laboratory of Oncology, G. Gaslini Children’s Hospital) for immunohis-

tochemical staining and Dr. Soldano Ferrone (University of Pittsburgh

Cancer Institute, Pittsburgh, PA) for providing NAMB-1 mAb.

DisclosuresThe authors have no financial conflicts of interest.

G

ED

A B C

F

H

FIGURE 7. HLA-G promotes a Th2 response whereas inhibiting B cell Ab secretion and inducing immune tolerance in vivo in a murine xenograft model.

(A) Schematic representation of the experimental procedures. (B) BALB/c mice were injected s.c. with 10 3 106 M8-pcDNA (HLA-G2) or M8-HLA-G

(HLA-G+) xenogeneic cells. M8–HLA-G cells were pretreated with the 87G Ab (a-HLA-G) or the isotype control before s.c. injection into mice. En-

graftment was monitored at indicated time points. Data represent mean 6 SD from one representative experiment with three mice in each group. Sig-

nificance was evaluated with Kruskal–Wallis test. (C) Plasmas were collected day 27 after xenogeneic M8 cell injection. Total IgG was quantified by ELISA

compared with a standard curve established with standard mouse IgG. Results are expressed as the mean titer 6 SD of groups of three plasmas. (D)

Presence of specific Abs against xenoantigens in the plasmas of the different groups of mice was analyzed at day 27 or in kinetic at days 7, 13, 20, and 27 by

slot blot. The numbers indicated on the right correspond to the molecular mass (in kDa) of protein markers. Data from two representative experiments are

shown. (E–H) Plasma concentrations of cytokines from Ctrl mice or mice injected with HLA-G+ xenogeneic cells were quantified by ELISA at day 7

postinoculation. Results are given as means (n = 3). Significance was evaluated with Kruskal–Wallis test. *p , 0.05.

10 B CELL REGULATION BY HLA-G/ILT2 INTERACTION

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

References1. Harwood, N. E., and F. D. Batista. 2008. New insights into the early molecular

events underlying B cell activation. Immunity 28: 609–619.2. Mitchison, N. A. 2004. T-cell-B-cell cooperation. Nat. Rev. Immunol. 4: 308–

312.3. McCubrey, J. A., L. S. Steelman, S. L. Abrams, F. E. Bertrand, D. E. Ludwig,

J. Basecke, M. Libra, F. Stivala, M. Milella, A. Tafuri, et al. 2008. Targetingsurvival cascades induced by activation of Ras/Raf/MEK/ERK, PI3K/PTEN/Akt/mTOR and Jak/STAT pathways for effective leukemia therapy. Leukemia22: 708–722.

4. Rawlings, D. J., M. A. Schwartz, S. W. Jackson, and A. Meyer-Bahlburg. 2012.Integration of B cell responses through Toll-like receptors and antigen receptors.Nat. Rev. Immunol. 12: 282–294.

5. Dalakas, M. C. 2008. Invited article: inhibition of B cell functions: implicationsfor neurology. Neurology 70: 2252–2260.

6. Redfield, R. R., III, E. Rodriguez, R. Parsons, K. Vivek, M. M. Mustafa,H. Noorchashm, and A. Naji. 2011. Essential role for B cells in transplantationtolerance. Curr. Opin. Immunol. 23: 685–691.

7. Liu, C., H. Noorchashm, J. A. Sutter, M. Naji, E. L. Prak, J. Boyer, T. Green,M. R. Rickels, J. E. Tomaszewski, B. Koeberlein, et al. 2007. B lymphocyte-directed immunotherapy promotes long-term islet allograft survival in nonhumanprimates. Nat. Med. 13: 1295–1298.

8. Dighiero, G., and T. J. Hamblin. 2008. Chronic lymphocytic leukaemia. Lancet371: 1017–1029.

9. Kurosaki, T. 2008. Paradox of B cell-targeted therapies. J. Clin. Invest. 118:3260–3263.

10. Colonna, M., F. Navarro, T. Bellon, M. Llano, P. Garcıa, J. Samaridis,L. Angman, M. Cella, and M. Lopez-Botet. 1997. A common inhibitory receptorfor major histocompatibility complex class I molecules on human lymphoid andmyelomonocytic cells. J. Exp. Med. 186: 1809–1818.

11. Shiroishi, M., K. Kuroki, L. Rasubala, K. Tsumoto, I. Kumagai, E. Kurimoto,K. Kato, D. Kohda, and K. Maenaka. 2006. Structural basis for recognition of thenonclassical MHC molecule HLA-G by the leukocyte Ig-like receptor B2(LILRB2/LIR2/ILT4/CD85d). Proc. Natl. Acad. Sci. USA 103: 16412–16417.

12. Shiroishi, M., K. Kuroki, T. Ose, L. Rasubala, I. Shiratori, H. Arase, K. Tsumoto,I. Kumagai, D. Kohda, and K. Maenaka. 2006. Efficient leukocyte Ig-like re-ceptor signaling and crystal structure of disulfide-linked HLA-G dimer. J. Biol.Chem. 281: 10439–10447.

13. Riteau, B., N. Rouas-Freiss, C. Menier, P. Paul, J. Dausset, and E. D. Carosella.2001. HLA-G2, -G3, and -G4 isoforms expressed as nonmature cell surfaceglycoproteins inhibit NK and antigen-specific CTL cytolysis. J. Immunol. 166:5018–5026.

14. Lila, N., N. Rouas-Freiss, J. Dausset, A. Carpentier, and E. D. Carosella. 2001.Soluble HLA-G protein secreted by allo-specific CD4+ T cells suppresses theallo-proliferative response: a CD4+ T cell regulatory mechanism. Proc. Natl.Acad. Sci. USA 98: 12150–12155.

15. Le Rond, S., C. Azema, I. Krawice-Radanne, A. Durrbach, C. Guettier,E. D. Carosella, and N. Rouas-Freiss. 2006. Evidence to support the role ofHLA-G5 in allograft acceptance through induction of immunosuppressive/regulatory T cells. J. Immunol. 176: 3266–3276.

16. Naji, A., A. Durrbach, E. D. Carosella, and N. Rouas-Freiss. 2007. Soluble HLA-G and HLA-G1 expressing antigen-presenting cells inhibit T-cell alloprolifera-tion through ILT-2/ILT-4/FasL-mediated pathways. Hum. Immunol. 68: 233–239.

17. Liang, S., V. Ristich, H. Arase, J. Dausset, E. D. Carosella, and A. Horuzsko.2008. Modulation of dendritic cell differentiation by HLA-G and ILT4 requiresthe IL-6—STAT3 signaling pathway. Proc. Natl. Acad. Sci. USA 105: 8357–8362.

18. Carosella, E. D., P. Moreau, J. Lemaoult, and N. Rouas-Freiss. 2008. HLA-G:from biology to clinical benefits. Trends Immunol. 29: 125–132.

19. Moreau, P., S. Flajollet, and E. D. Carosella. 2009. Non-classical transcriptionalregulation of HLA-G: an update. J. Cell. Mol. Med. 13(9B): 2973–2989.

20. Rouas-Freiss, N., R. M. Goncalves, C. Menier, J. Dausset, and E. D. Carosella.1997. Direct evidence to support the role of HLA-G in protecting the fetus frommaternal uterine natural killer cytolysis. Proc. Natl. Acad. Sci. USA 94: 11520–11525.

21. Carosella, E. D., P. Moreau, S. Aractingi, and N. Rouas-Freiss. 2001. HLA-G:a shield against inflammatory aggression. Trends Immunol. 22: 553–555.

22. Morandi, F., I. Levreri, P. Bocca, B. Galleni, L. Raffaghello, S. Ferrone,I. Prigione, and V. Pistoia. 2007. Human neuroblastoma cells trigger an immu-nosuppressive program in monocytes by stimulating soluble HLA-G release.Cancer Res. 67: 6433–6441.

23. Jinquan, T., H. H. Jacobi, C. Jing, A. Millner, E. Sten, L. Hviid, L. Anting,L. P. Ryder, C. Glue, P. S. Skov, et al. 2003. CCR3 expression induced by IL-2

and IL-4 functioning as a death receptor for B cells. J. Immunol. 171: 1722–1731.

24. Liu, Y. J., D. E. Joshua, G. T. Williams, C. A. Smith, J. Gordon, andI. C. MacLennan. 1989. Mechanism of antigen-driven selection in germinalcentres. Nature 342: 929–931.

25. Lebecque, S., O. de Bouteiller, C. Arpin, J. Banchereau, and Y. J. Liu. 1997.Germinal center founder cells display propensity for apoptosis before onset ofsomatic mutation. J. Exp. Med. 185: 563–571.

26. Menier, C., C. Guillard, B. Cassinat, E. D. Carosella, and N. Rouas-Freiss. 2008.HLA-G turns off erythropoietin receptor signaling through JAK2 and JAK2V617F dephosphorylation: clinical relevance in polycythemia vera. Leukemia22: 578–584.

27. Bekeredjian-Ding, I., S. Inamura, T. Giese, H. Moll, S. Endres, A. Sing,U. Zahringer, and G. Hartmann. 2007. Staphylococcus aureus protein A triggersT cell-independent B cell proliferation by sensitizing B cells for TLR2 ligands. J.Immunol. 178: 2803–2812.

28. Corcione, A., L. Ottonello, G. Tortolina, P. Facchetti, I. Airoldi, R. Guglielmino,P. Dadati, M. Truini, S. Sozzani, F. Dallegri, and V. Pistoia. 2000. Stromal cell-derived factor-1 as a chemoattractant for follicular center lymphoma B cells. J.Natl. Cancer Inst. 92: 628–635.

29. Kim, H. S., X. Zhang, and Y. S. Choi. 1994. Activation and proliferation offollicular dendritic cell-like cells by activated T lymphocytes. J. Immunol. 153:2951–2961.

30. Hauser, A. E., T. Junt, T. R. Mempel, M. W. Sneddon, S. H. Kleinstein,S. E. Henrickson, U. H. von Andrian, M. J. Shlomchik, and A. M. Haberman.2007. Definition of germinal-center B cell migration in vivo reveals predominantintrazonal circulation patterns. Immunity 26: 655–667.

31. Ghebrehiwet, B., L. Silvestri, and C. McDevitt. 1984. Identification of the Rajicell membrane-derived C1q inhibitor as a receptor for human C1q. Purificationand immunochemical characterization. J. Exp. Med. 160: 1375–1389.

32. Naji, A., C. Menier, G. Maki, E. D. Carosella, and N. Rouas-Freiss. 2012.Neoplastic B-cell growth is impaired by HLA-G/ILT2 interaction. Leukemia 26:1889–1892.

33. Cross, D. A., D. R. Alessi, P. Cohen, M. Andjelkovich, and B. A. Hemmings.1995. Inhibition of glycogen synthase kinase-3 by insulin mediated by proteinkinase B. Nature 378: 785–789.

34. Greer, E. L., and A. Brunet. 2005. FOXO transcription factors at the interfacebetween longevity and tumor suppression. Oncogene 24: 7410–7425.

35. Blery, M., H. Kubagawa, C. C. Chen, F. Vely, M. D. Cooper, and E. Vivier. 1998.The paired Ig-like receptor PIR-B is an inhibitory receptor that recruits theprotein-tyrosine phosphatase SHP-1. Proc. Natl. Acad. Sci. USA 95: 2446–2451.

36. Ristich, V., S. Liang, W. Zhang, J. Wu, and A. Horuzsko. 2005. Tolerization ofdendritic cells by HLA-G. Eur. J. Immunol. 35: 1133–1142.

37. Liang, S., B. Baibakov, and A. Horuzsko. 2002. HLA-G inhibits the functions ofmurine dendritic cells via the PIR-B immune inhibitory receptor. Eur. J.Immunol. 32: 2418–2426.

38. Gonen-Gross, T., H. Achdout, R. Gazit, J. Hanna, S. Mizrahi, G. Markel,D. Goldman-Wohl, S. Yagel, V. Horejsı, O. Levy, et al. 2003. Complexes ofHLA-G protein on the cell surface are important for leukocyte Ig-like receptor-1function. J. Immunol. 171: 1343–1351.

39. Naji, A., S. Le Rond, A. Durrbach, I. Krawice-Radanne, C. Creput, M. Daouya,J. Caumartin, J. LeMaoult, E. D. Carosella, and N. Rouas-Freiss. 2007. CD3+CD4low and CD3+CD8low are induced by HLA-G: novel human peripheralblood suppressor T-cell subsets involved in transplant acceptance. Blood 110:3936–3948.

40. Qiu, J., P. I. Terasaki, J. Miller, K. Mizutani, J. Cai, and E. D. Carosella. 2006.Soluble HLA-G expression and renal graft acceptance. Am. J. Transplant. 6:2152–2156.

41. Sheshgiri, R., V. Rao, A. Mociornita, H. J. Ross, and D. H. Delgado. 2010.Association between HLA-G expression and C4d staining in cardiac transplan-tation. Transplantation 89: 480–481.

42. Tamir, I., J. M. Dal Porto, and J. C. Cambier. 2000. Cytoplasmic protein tyrosinephosphatases SHP-1 and SHP-2: regulators of B cell signal transduction. Curr.Opin. Immunol. 12: 307–315.

43. Feger, U., E. Tolosa, Y. H. Huang, A. Waschbisch, T. Biedermann, A. Melms,and H. Wiendl. 2007. HLA-G expression defines a novel regulatory T-cell subsetpresent in human peripheral blood and sites of inflammation. Blood 110: 568–577.

44. Wiendl, H., U. Feger, M. Mittelbronn, C. Jack, B. Schreiner, C. Stadelmann,J. Antel, W. Brueck, R. Meyermann, A. Bar-Or, B. C. Kieseier, and M. Weller.2005. Expression of the immune-tolerogenic major histocompatibility moleculeHLA-G in multiple sclerosis: implications for CNS immunity. Brain 128: 2689–2704.

The Journal of Immunology 11

by guest on June 19, 2018http://w

ww

.jimm

unol.org/D

ownloaded from