comparative reactivity of different hla-g monoclonai ......f. lenfant'. r le boutatl>9r>...
TRANSCRIPT
S. FournelX. HueM. Aguerre-GirrC. SolierM. LegrosC. Praud-BrethenouM. MoussaG. ChaouatA. BerrebiA. BensussanF. LenfantP. Le Bouteiller
Comparative reactivity of different HLA-Gmonoclonai antibodies to soiubie HLA-Gmoiecuies
Key words:
Soluble HLA-G: HU-G isoforms; HLA-Gmonoclonal antibodies; ELiSA
Acknowledgments:This work was supported by Institut National deia Sante et de la Recherche Medicaie, byAssociation pour ia Recherche sur le Cancer(9489), by Universite Paui Sabatier and byEtablissement Frangais des Greffes. We wish tothank Dr. F. C. Grunnet for providing us with theSHLA-B7 and mHLA-B7 transfectants, Dr. D.Geraghty, Dr. H. Waldmann and Pr. J.R Revillardfor their gifts of mAbs, Pr. E. Weiss. Dr. S. Maier.Dr. H, Grosse-Wilde and Dr. V. Rebmann forproviding us a caiibrated, standard sHLAG. A.Blascliitz is acknowledged for her helpfui advicein the estabiisment of ELiSA procedure and G.Cassar for the flow cytometry analysis.
Received 37 October 1999, revised,accepted for puOllcation 20 Match 2O00
' Munkseaard 2000TTssue Antigens . ISSN 0001-2815
nssue Antigens 2000; S6; 510-518Printed in Denmark . All rights reserved
MQstract: Different HLA-G monoclonal antibodies (inAbs) were first evalu-
ated for their capability to identify soluble HLA-G (sHLA-G) in ELISA. Three
of them, namely 87G, BFL.1 and MEM-G/9, when used as coating mAbs
together with W6/32 capture mAb, identified p2-mia'oglobulin (|32m)-associ-
ated-sHLA-G but not soluble HLA-B7 (sHLA-B7) in cell culture super-
natants from transfected cells. By comparison, the anti-HLA class 1 mAb 90
did recognize both sHLA-G and sHLA-B7. By using these HLA-G mAbs,
sHLA-G was identified in amniotic fluids as well as in aiiture supernatantsof first trimester and term placental explants but not in cord blond. Intron
4-retaining sHLA-G isoforms were identified in some amniotic fluids by the
use of an intnin 4-spedfic mAb {16G1). Reactivity of these different HLA-
G mAbs was then compared to determine their respective binding sites on
soiubie and membrane-bound HLA-G. Using both ELISA and flow cytome-
try analysis, we showed that they did not compete with each other, which
suggested that they did not recoipiize the same determinants. Finally, we
report that two mAbs directed against the a l domain of HLA class I heavy
chain (mAb 90 and YTH 862) did compete with 87G. therefore demonstrat-ing that this latter mAb recognized an epitope localized on this external
domain of HLA-G.
HLA-G class Ib gene is characterized by a low polymorphism,
unique structural featui-es and a limited tissue distribution (1, 2).
Major structural characteristics of HLA-G are as follows. First, the
promoter region of HLA-G differs in many ways from the other
major histocompatibitity compiex (MHC) class I genes (3,4). Second,
the cytoplasmic domain of HLA-G is shorter than in class Ia mol-
ecules (5). Third, the primary HLA-G transcripts are differentially
spliced, giving rise to six different isoforms (6-9). Four of them
encode membrane-bound products lacking either one or two exter-
nal domains, two others generate soluble proteins. These soluble
forms of HLA-G (sHLA-G). a full-length sHLA-Gl and a shorter
sHLA-G2 form lacking the a2 domain, are encoded by alternatively
spliced transcripts that retain intron 4 which contains a stop codon,
Authors'
S.X. Hue',M.C. Soller'.M. Legros'.C. Praud-Brethenou',M. Mousao'.G. CfiaouBl'.A. BerrebP.A. Bensussan',F. Lenfant'.R Le Boutatl>9r>
'Insmj! Nsilonai de laSante et lie la RecdercheM^icale U395, Hdpital dePurpan, Toulouse. France.
'Biologle de la RelationMatemo Foetale. IMSERM U131, HOpltai AntoineBeclSre. Cfamart, Frarice.
^Service de GyrtecclogleObst^rique. HOpltal UGisue. Toulouse. Fiance.
*Facun6 de M^Oecine deCr#tell, INSERM U4it8,CrSteil, France
Corraapondance to;pn 11 tope Le BnutelllerInsitiul National de la Sant^
et de la Rechetctie
Centm HospKalo.Univefsiiaire Oe Purpan,BP 3026
31024 Toulouse Cedax 03FranceFa»: +33 562 748 386Tel: +33 562 748374
510
Fournel et a l : HLA-6 mAb reactivity
thus precluding the translation of transmembrane and cytoplasmic
domains {7, 10, 11).
Constitutive expression of HLA-G translated products has been
reprodudbly detected in placenta and thymus. Within the plaeenta,
the following cellular subtypes were found to bind HLA-G mAbs
by immunohistochemistry: invading extravillous cytotrophoblast
(12, 13-15) endothelial cells lining fetal vessels in the chorionic villi
(13), mesenchymal Hofbauer cells (16) and amnion epithelial cells
(17, 18). Soluble forms of HLA-G have been detected in anmiotic
fluids and plasma (18-20) and, in one report, in some syncytiotro-
phoblast (16). HLA-G was also found in some medullary thymic
epithelial cells (21, 22).
Functions of HLA-G are now emerging (1). HLA-G binds nonam-
er peptides (23, 24) and is capable of presenting them to CD8+ T
cells (25, 26). HLA-G is the ligand of few killer cell immunoglobuiin-
like receptors (2, 27) and modulates HLA-E cell surface expression
upon binding to a nonamer derived from the HLA-G leader se-
quence (28).
A number of groups have reported the use of HLA-G monoclonal
antibodies (mAbs) (14,16,18,23,29). Only partial characterization of
these reagents has been made and very few comparative studies have
been carried out to examine their respective reactivities (1). Investi-
gations on their potential use in ELISA for the detection of soluble
forms of HLA-G are also limited (16, 19), and little information is
available about which antigenic determinant they can recognize. By
comparing the reactivity of three HLA-G mAbs in ELISA, we show in
this report that they al! identified p2-microglobulin (p2m)-associated
sHLA-G but not sHLA-B7, that they did not compete with each other
and therefore did not recognize the same epitopes. Additional compe-
tition experiments performed by flow cytometry with mAbs directed
against the a l domain of HI.A class I molecules demonstrate that the
antigenic determinant recognized by the 87G anti-HLA-G mAb is
located on this external domain of HLA-G.
(29), 87G agG2b; gift of Dr. D. Geraghty, Seattle, WA. USA) (23) and
MEM-G/9 agGl. Exbio, Prague, Czech Republic) are HLA-G-specific
mAbs. MEM-G/9 mAb was raised against native, properly folded,
p2m-associated HI.A G. It reacts with HLA-G on the cell surface but
not with denatured molecules, under the conditions of Western blot-
ting. MEM-G/1 (igGl, Exbio, Prague, Czech Republic) reacts with the
denatured HLA-G HC. 16G1 (IgG2a. gift of Dr. D. Geraghty, Seattle,
WA, USA) is specific for a synthetic peptide corresponding to the
amino acid sequence of sHLA-G intron 4 (23). 10.3.6 anti-murine MHC
class n I-A"* (IgG2a) (34) and control isotypic mAbs (Dako, Trappes,
France) were used as negative controls.
Cells
The human gestational choriocarcinoma cell lines JAR and JEG-3 and
the lymphoblastoid cell tines .221 and CIR were obtained from the
American Type Culture Collection (Rockville, MD, USA). JAR-Gl
cells, transfected with the full-length membrane-bound HLA-Gl en-
coding cDNA, and JAR-sGl, transfected with the full-length, intron 4
retaining soluble HLA-Gl (sHLA-Gl) cDNA, have been already de-
scribed (13, 35). 221-sGl cells were similarly transfected with sHLA-
Gl encoding cDNA. ClR-sB7 cell line transfected with sHLA-B7 en-
coding cDNA, and ClR-mB7, transfected with membrane HLA-B7 en-
coding cDNA, were kindly provided by Dr. F. C. Grumet (Stanford,
CA, USA) and have been previously described (36). Cell lines were cul-
tured in RPMI 1640 with Glutamax (Life Technologies. Cergy-Ponto-
ise, France) supplemented with 10% FCS, 1 mM sodium pyruvate, 100
rU/ml penicillin and 100 ng/ml streptomycin (Sigma, St Quentin Fal-
lavier, France). Cultures were maintained in a humid atmosphere at
37°C containing 5% CO2. Cell culture supernatants were collected,
centrifuged at 1200 rpm for 5 min at 4°C, and stored at -20°C until
use. In some cases they were concentrated five times using Macrosep
lOK (10 kDa cutoff) centrifugal concentrator (Paul Filtron Corpor-
ation, Northborough, MA, USA).
Material and methods
Monoclonal antibodies
W6/32 mAb (IgG2a) recognizes a framework determinant expressed
on p2m-assodated HLA class I heavy chain (HC) (30); both human
mAb 90 (IgG 1, gift of Pr. j . P Revillard, Lyon, France) and rat YTH862
(IgG2b, gift of Dr. H. Waldmann, Oxford, UK) recognize a framework
determinant located on the a l domain of HLA class I HC (31); HCA2
mAb (IgGl, gift of Pr. H. Ploegh, Cambridge, MA, USA), recognizes
free HC of HLA-A, -E, -F and -G as well as some selected -B and -C
alleles (32,33); BFL.l agG2a, Coulter Immunotech, Marseille, France)
Amniotic fluids and cord blood sera
Amniotic fluids were obtained by amniotie puncture of healthy
pregnant women (16 to 24 weeks of pregnancy). Cord blood sera
were obtained from term placentas. Samples were centrifuged at
2000 rpm for 5 min, concentrated five times as described above, or
not concentrated, and stored at - 2 0 ^ .
Placental supernatants
Supernatants were prepared as previously desaibed (37). Briefly,
first trimeter and term placentas were obtained from healthy preg-
Tissue Anti^ns 2000: 65: 510-518 511
Foume! et a l : HLA-G mAb reactivity
tiant women, and 1.5 g of tissue was washed in HBSS and incubated
in 10 ml of RPMl-10% FCS. At various times, supernatants were
harvested, and stored at -20°C.
Specific soluble HLA-G (sHLA-G) ELISA
96-well microtiter plates {Corning Costar Corporation, Cambridge,
MA. USA) were coated with different mAbs at the indicated concen-
trations in O.IM carbonate buffer pH 9.5 for 1 h at 37''C then over-
night at 4°C. After three wash^ with phosphate-buffered saline
(PBS) containing 0.05% Tween 20 (PBS-Tween), plates were satu-
rated with 100 til of PBS containing 4% BSA overnight at 4T. Cell
culture supernatants (50 |il) were added to each well. After incuba-
tion for 2 h at 37°C, plates were washed three times with PBS-
Tween. Biotinylated mAb (50 \i\) was added to each well and incu-
bated for 1 h at 37X. Plates were washed five times with PBS-
Tween and further incubated with a peroxidase-conjugated avidin
(Dako) for 15 min at 3rC. After five washes with PBS-Tween. 50
nl of ortho-phenyl-diamine peroxidase substrate (Sigma) was added
to each well and incubated for 15 min at room temperature. The
reaction was stopped by addition of 50 nl of 2N H2SO4. The relative
1.5
I
e.5
8
•C1i)-sB70221-561
.AFig. 1. Comparative reactivity of different HLA-G mAbs in ELI-SA. Microwells were coated either with 87G (2 jiK/ml), BFL.l (2 ng/ml).
MEM-G/9 (ascite 1/5000). 16G1 (2 Mg/mi), mAb 90 (ascite 1/100), MEM-G/1(ascite 1/1000). HCA2 (liybridoma supernatant 1/5) or isotypic control Ab
(mixture of IgG2a. IgGl and IgM at 2 jig/ml). Then cell culture super-nalants (non concentrated for 87G, MEM-G/9 and 16G1 mAbs, 5X a>ncen-trated for the other mAbs) were added and bound molecules identified by
using biotinylated W6/32 mAb (10 Mg/nil), as described in Materials andMethods. To assess ELISA specificity, cell culture supernatants of 221 sGl
cells (white bar) were compared to culture supernatants of ClR-sB7 cells(black bar). Results are expressed as means±SD of absorbancy units (492
trni) from four (S7G, MEM-G/9), five (BFL.1, mAb90. MEM-G/1, HCA2. con-trol Ab) or two (16G1) separate experiments. Mean OD obtained with 87Gwas 2.84 ±0^6.
concentrations of sHLA-G or sHLA-B7 were estimated from absorb-
ance measui-ed at 492 nm (duplicate wells) on a MRX-ELISA
microplate reader (Dynatech. Guyancourt, France). A standard
supernatant of sHLA-G/.221 transfectant, produced by Prof. E.
Weiss and Dr. S. Maier (Mtmich, Germany), and calibrated by Drs.
H. Grosse-Wilde and V. Rebmann (Essen, Germany), as previously
described (38), served as reference in standard calibration curves.
Row cytometry competition studies
Cells (0.5x10^) were incubated with blocking mAb diluted in PBS
supplemented with 1% FCS and 0.2% NaN3 for 30 min at AX. After
two washes in the same meditun, 87G mAb (5 [lg/ml in PBS/FCS/
Azide) was added to the cells for an additional 30 min at 4°C. The
cells were then washed and incubated with phycoerythrin-conju-
gated streptavidin (Coulter Immunotech) for 30 min at 4X. After
subsequent washes, cells were analyzed on a Coulter Epics Elite
fiow cytometer (Margency, France) gated to excluded non-viable
cells.
Results
Several HLA-G mAbs identify sHLA-G In cell culture
supernatants of sHLA-G transfectants
To investigate the ability of several HLA-G mAbs to detect sHLA-
G present in cell culture supernatants, different HLA-G-recognizing
mAbs were coated to the microtiter plates and the biotinyIated-W6/
32 anti-HLA class I mAb was used to reveal the bound molecules.
As shown in Fig. 1. four HLA-G mAbs, namely 87G, BFL.l. MEM-
G/9 and 16G1, although with different affinities, specifically recog-
nized sHLA-G in cell culture supernatant from a lymphoblastoid
cell line transfected with sHLA-G-encoding cDNA (.221-sGl). By
comparison, no HLA-G mAb recognized sHLA-B7 in cell culture
supernatant from a lymphoblastoid cell line transfected with sHLA-
HLA-B7-encoding cDNA (ClR-sB7). The different signal intensities
obtained in ELISA with these three HLA-G mAbs are likely to re-
flect different levels of binding affity. mAb 90 which binds to a
common epitope of HLA class I HC (31), including HLA-G (P. Le
Bouteiller, tmpublished data), did recognize both sHLA-G and
SHLA-B7. In contrast, neither MEM-G/1 nor HCA2 mAbs, which
were raised against denatured HLA class I HC. nor isotypic control
Abs, reacted against sHLA-G or sHLA-B7.
We further examined the sHLA-G specificity of 87G, MEM-G/9
and BFL.l mAbs. Whereas these mAbs identified sHLA-G in .221-
sGl culture supernatants, none of them detected any soluble protein
512 nssue Antf ens 2000: Sft;
Foumel et al : HLA-G mAb reactivity
in cell culture supernatants from 2l\, CIR or ClR-sB7 cells (Fig. 2).
Using 87G as coating mAb, sHLA-G was also detected in culttire
supernatants of JAR transfected with a cDNA encoding sHLA-Gl
(JAR-sGl), whereas no soluble protein was identified in the culture
medium either from tintransfected JAR cells or from JAR cells frans-
fected with a cDNA encoding membrane HLA-Gl QAR-Gl) (Fig. 2).
Using serial concentrations of a calibrated sHLA-G reference (20),
we established standard curves for 87G and MEM-G/9 HLA-G
mAbs (Fig. 3). Both curves revealed linearity for sHLA-G concen-
trations in the range of 30 to 260 ng/ml.
We then used the ELISA described above (coating mAb : 87G)
to investigate the presence of sHLA-G in few biological fiuids and
culttu'e media from placental explants. As shown in Fig. 4A, sHLA-
G was identified in amniotic fitiids of healthy pregnant women (19-
24 weeks of pregnancy) but not in cord blood sera. Similar results
were obtained with MEM-G/9 coating mAb (data not shown). It
should be noted that cord blood sera always gave a quite high
background in ELISA after coating the plates with isotypic IgG
control, which was never the case for amniotic fluids or the different
cell culture supernatants (these latter control values have been sub-
sfracted from the values shown in Fig. 4). In contrast, by using
mAb 90 (anti HLA-A, -B, -C, -G) or B1.23. 2 (anti-HLA-B and -C but
not HLA-G) coating mAbs (29) and W6/32 detection mAb, we have
observed that the cord blood sera did contain sHLA class Ia mol-
ecules (data not shown). sHLA-(; was also detected in culture super-
natants from first trimester or term placenta explants (Fig. 4B).
Based on ELISA standard curves (see Fig. 3), we estimated that the
concenfrations of sHLA-G in amniotic fluids were in the range of
35 to 130 ng/ml, depending on the samples, and in the range of 30
to 60 ng/ml for placental supernatants. Tbe use of 87G and MEM-
G/9 mAbs did not allow to disaiminate between sHLA-G isoforms
possibly derived from shedding or proteolytic cleavage of mem-
brane molecules and those retaining intron 4. Therefore we used a
conformation independent mAb 16G1 which reacts uniquely with
Fig. 2. Specific detection of sHLA-G with 87G, IVIEMG/9 andBFL.l coating inAbs. (A), 87G (1 ^g/ml), (B), MEM-G/9 (1/5000) and (C),
BFL.1 (2 |ig/ml) were coated on microtiter plates and biotinylated W6/32mAb (10 |ig/nil) was added to revealed the bound sHLA-G molecules (see
Materials and methods). To assess the specificity of the EUSA, cell culturesupernatants from 221-sGl and JAR-sGl cells were used as positive control
and those from JAR. JAR-Gl (which express membrane HLA-Gl), 221, CIRand C1R-SB7 cells were used as negative control. All culture supernatants
were used non concentrated, except JAR-sGl which was concentrated 5times. Results are expressed as means±SD of absorbancy units (492 nm) on
ten separate experiments for 87G. seven for BFL.1 and five for MEM-G/9mAb.
nssue Anti^ns 2000: 55: 510-518 513
Fournel et al : HLA-G mAb reactivity
87G r = 0.9738
UIU
492
oO
0.5
0
1
0.8
0.6
0.4
0.2
MEM-G /9 = 0.999
0 50 100 150 200 250 300
Concentration of sHLA-G1 (ng/ml )
Fig. 3. ELISA standard curves for sHLA-G using 87G or MEM-G/9 mAbs. Varying concentrations of calibrated standard supernatant of
sHLA-Gl/,221 transfectant (see Material and methods) were coated into
miooplates. Binding of either 87G (2 ng/ml) or MEM-G/9 (1/5000) mAbs
followed by biotinylated W6/32 was measured by ELISA as described in
Material and nfiethods. Results are expressed as means of duplicates.
the intron 4 retaining soluble form of HLA-G (23). We compared its
reactivity with 87G and MEM-G/9 mAbs by testing eleven other
amniotic fiuid samples in ELISA. Results are shown in Fig. 5. Very
similar results were obtained with 87G and MEM-G/9 mAb: both
mAbs detected the same variations of the sHLA-G level of express-
ion, according to amniotic fluid samples. BFL.1 HLA-G mAb also
gave very similar results (data not shown). When 16G1 mAb was
used, the most positive amniotic fluids (samples 8 and 13) were the
same as those detected by the other HLA-G mAbs. Sample 17 ap-
peared more positive with 16G1 mAb than with the other HLA-G
mAbs whereas a slight reactivity was detected in samples 16 and
18. These data demonstrated that at least part of sHI A-G present
in amniotic fluids was intron 4 retaining sHl.A-G is<}form. 16G1
mAb also reacted with placental cell supernatants (Solier et al.,
manuscript in preparation),
Different HLA-G mAbs do not bind to the same epitopes on
sHLA-GI
As demonstrated above, 87G, BFL-1, MEM-G/9, mAb 90 and W6/
32 mAbs recognize sHLA-G in culture media from different sHLA-
G transfectants. To compare the reactivities of these various HLA-
G mAbs in ELISA, we used different combinations of coating and
detection mAbs. Microwells were coated either with 87G, BFL.1,
1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9
AmnJotic fluids Cord blood
24h 48h 12h 24h 48h
•' trim, placentas Term placentas
Fig. 4. Detection of sHLA-G in amnioticfluid, cord blood and cell culture super-natants of first trimester and term placen-ta! explants. A) Seven samples of amni-otic fluids (numbers 1-7). concentrated 5times, and 9 samples ol c<ird blood sera(numbers 1-9), non-concentrated, wereevaluated usii^ sHLA-G-specific ELISA(coating mAb: 87G; detection mAb: bi-otinylated W6/32 mAb). Values are expressedas mean absorbancy units (492 nm) of three sep-
arate EL1SA±SD after substraction of thevalues obtained with the isotypic a)ntrol coatedAbs. Numbers 4 and 9 of cord blocid sjiniples:
OD=0. B) Culture supernatants of first tri-mester or term placental explants wereharvested at indicated times and thenevaluated non-concenlrated using sHLA-G-specific ELISA. Values are expressed as meanabsorbancy units (492 nm) ±SD of eight first tri-mester and fifteen term placental explants. 87G
mAb was coated at 1 pg/ml and biotinylated W6/32 added at 10 Mg'nil. 221-sGl and ClR-sB7 cul-
ture supernatants were used as positive andnegative controls, respectively.
514 TIssw Antigens 2000: BS: 510-51B
Fournel et al : HLA-G mAb reactivity
Fig. 5. Detection of sHLA-G in amnioticfluid: comparative study using three differ-ent HLA-G mAbs. Eleven Tion-concentratedsamples of amniotic fluid (numbers 8-18) were
evaluated using sHLA-G-spedfic ELISA. CoatingmAbs were as follows: 87G (2 g/mI), MEM-G/9 (ascite 1/5(XK)), or 16G1 (2 ig/ml). Detection
mAb: biotinylated W6/32 mAb. Values are ex-pressed as mean absorbancy units (492 nm) of
duplicates ±SD. Values obtained with the iso-typic control coated Abs were all negative. 221-
sGl and C1R-SB7 culture supernatants wereused as positive and negative controls, respec-
tively. OD obtained with 87G and MEM-G/9 on221-sGl were 2.66 and 1.14, respectively.
IB I I 12 13 14 15 16 17 18
Hmniotic fluids
MEM-G/9. W6/32 mAbs or with mAb 90, then the bound sHLA-G
molecules were revealed either by biotinylated-W6/32, -87G or
-WL.] mAbs. As shown in Fig. 6. all the combinations using HLA-
G mAbs or W6/32, albeit with different affinities, allow the detection
of sHLA-G in culture supernatants of 221-sHLA-G, suggesting that
these various mAbs recognize different epitopes on sHLA-G mol-
ecules. However, combinations using 87G, MEM-G/9 or, to a lesser
extent, BFL.l coating mAbs, followed by biotinylated W6/32 detec-
tion mAb, gave the strongest specific signals. In all these combi-
nations, SHLA-B7 is never recognized (Fig. 6, left panel). Interest-
1.5
0.5
Coating mAb
i
C1R-SB7 n221-sG1
JDetection mAb Biotinylated W6/32 BiotinylatedB7G BlotinyiatedBFL.1
Fig. 6. Comparative ELISA using different combinations of coat- the bound sHLA-G molecules (see Materials and methods). Non-concen-ing and detection mAbs. 87G (1 ng/ml), BFL.1 (2 ng/m!), MEM-G/9 (asdte
1/5O(X)), mAb 90 (ascite 1/100), W6/32 (2 ng/ml) and isotype control Ab(IgG2a-(-IgGl -HigM; 2ng/nil) were coated on microwells and biotinylated W6/
), 87G (5 ng/ml) or BFL.l (5 Mg/ml) mAbs were added to revealed
trated cell culture supernatants from 221-sG] cells (white bar) were used as
positive control and those from ClR-sB7 cells (black bar) were used asnegativa control. Results are expressed as means±SD of absorbancy units(492 nm) on three separate experiments.
T7ssue Anti^ns 2000; 65: 510-518 515
Fournel et al : HLA-G mAb reactivity
ingly, when mAb 90 was used as coating mAb and biotinylated 87G
as detection mAb, no protein was detected (Fig. 6, medium panel),
suggesting that 87G mAb recognized the same or a near epitope on
the a l domain of sHLA-G.
87G mAb recognizes a determinant iocaiized on the a l
domain of HLA-G
We next compared the epitope distribution on membrane-bound
HLA-G by flow cytometry analysis. Trophoblast-derived JEG-3
cells, which are known to constitutively express HLA-G (7, 33),
have been incubated firstly, with various mAbs which recognize
HLA-G (blocking mAbs) and, secondly, with biotinylated 87G
mAb. As shown in Fig. 7, pre-incubation of JEG-3 with 87G
mAb almost abolished the binding of biotinylated 87G. as com-
pared to a preincubation with 10.3.6 control mAb or with HCA2
mAb (this latter mAb having been previously shown not to bind
to the cell surface of JEG-3: data not shown). BFL.l, MEM-G/9
and W6/32 mAbs did not prevent the binding of biotinylated
87G, demonstrating that these mAbs did not recognize the same
epitope as 87G mAb on membrane-bound HLA-G. In contrast,
mAb 90 and YTH862, which both recognize a framework deter-
minant on the a l domain of HLA class I HC (31), inhibited the
binding of 87G mAb. Altogether, these results and those men-
tioned above (ELISA) strongly suggest that 87G mAb did recog-
nize an epitope mapped on the a l domain of membrane-bound
and sHLA-G.
4,00
3.00
2.00
1.00
BiocklngmAb :
Detection mAb : Biotinylated87G
Fig. 7. Competition assays of various HLA-G mAbs : flow cytome-try analysis. JEG-3 celts were pre-incubated either with 87G (5 ng/ml),
10.3.6 (negative control, undiluted hybridoma supernatant), HCA2 (non-
diluted hybridoma supernatant), BFL.1 (5 ng/ml), MEM-G/9 (a.scite, 1/1000),W6/32 (5 ng/ml), mAb 90 (ascite, 1/100) or YTH862 (non-diluted hybridomasupernatant). Then, after intensive washes, cells were incubated with bi-
otinylated 87G mAb (2.5 fig/ml) followed by phycoerythrin-conjugatedstreptavidin. Binding of S7G mAb was evaluated by flow cytometry, as
described in Material and methods. Results are expressed as mean fluor-escence intensity (MFO ±SD on three separated experiments.
Discussion
In this report, we have shown that three HLA-G mAbs. namely,
87G, BFL.1 and MEM-G/9 did recognize sHLA-G molecule, in it;
p2-m-assodated form. They could not recognize HLA-G free HC for
the following reasons: firstly, the detection mAb W6/32 used in ELI-
SA only recognizes a conformational monomorphic determinant ex-
pressed on p2m-associated HLA class I HC (30); secondly, none of
the MEM-G/1 and HCA2 mAbs, which both react with denatured
free HLA-G HC, did boimd proteins from sHLA-G containing super-
natants, showing that such free HC were not present in the cell
culture supernatants used; finally, we have previously shown that
both 87G and BFL.l were able to recognize affinity purified p2m-
assodated-HLA-G in ELISA (35). These results confirm prervious
studies performed by other groups: Fujii et al. (7) mentioned that
87G detected sHLA-G in ELISA and Puppo et al. (19) recently iden-
tified sHLA-G in JEG-3 culture supernatant, as well as in some
biological fluids, by using W6/32 and biotinylated BFL.1 as captui¥
and detection mAbs respectively, after immunodepletion of HLA
class Ia and H1.A-E molecules with mAb TP25.99. To our knowl-
edge, no data has been reported yet with MEM-G/9 mAb.
None of these HLA-G mAbs did recognize sHLA-B7 present in
cultxire supernatant of transfected cells, although it carmot be tot-
ally excluded that they also reacted with some unfrequent sHLA
class Ia alleies. However, this is unlikely for the following reasons:
firstly, they did not detect any soluble protein in all cord bkxxl
samples tested, although we have shown by using mAb 90 or
Bl.23.2 coating mAbs that these cord sera did contain some sHLA-
class Ia molecules, as already reported by other groups (19); second-
ly, both 87G (23) and BFL.1 (29) were described not to bind to pe-
ripheral blood mononuclear cells of various donors expressing vari-
ous HLA class la allelea The three HLA-G mAbs used in this study
clearly recognized intron 4 retaining sHLA-Gl isoform, since only
this species was produced by JAR and 221 transfectants. In con-
trast, they were unlikely to detect the intron 4 retaining sHLA-G2
shorter form lacking the o2 domain, since W6/32 (used as coating
or detection mAb) was shown to recognize an epitope in the a2
domain (31), absent in sHLA-G2 (7).
sHLA-Gl has been detected in amniotic fluid, which is in accord-
ance with several previous reports (18, 20). It was shown that at
least part of the sHLA-G in amniotic fluid derived from HLA-G
cleaved from the membrane (18). The reactivity of 16G1 mAb
specifically raised against intron 4 amino acid that we observed in
some amniotic fiuid samples demonstrated that, in addition to shed-
ded or cleaved sHLA-G, intron 4 retaining sHLA-G isoforms were
also present.
516 nssue Antigens 2000: 66: 510-518
Fournel et al : HLA-G mAb reactivity
We did not find any sHLA-G in cord blood, which is in contrast
with other findings (19). However, knowing that human serum does
contain p2m-free class I HC (39) and that the HLA-G mAbs used in
this study could not detect such a form, it cannot be excluded that
p2m-free sHLA-G HC were present in cord blood.
Both types of analysis, performed in ELISA and fiow cytometry,
came to the same conclusion that 87G mAb recognizes an epitope
localized on the a l domain of HLA-G HC. A similar conclusion was
drawn from previous investigations (6,40). On the contrary, absence
of competition between 87G, BFL.1 and MEM-G/9 suggest that the
antigenic determinants recognized by these HLA G mAbs are dif-
ferent.
Such HLA-G mAbs should be useful tools to investigate the
possible immunoregulatory functions sHLA-G might exert in nor-
mal and pathological conditions (41. 42). Further epitope mapping
study of these HLA G reagents together with cristallography of
HLA-G molecule may help in more critical interpretation of such
future functional studies.
References
1. Le Bouteiller P, Blaschitz A. Thefunctionality of HLA-G is emer^ng.Immunol Rev 1999: 167; 233-44.
2. I^ Bouteiller P. Solier C. Proll J. Aguerre-Girr M, Fournel S, Lenfant F. PlacentalHLA-G protein expression in vivo: whereand what for. Hum Reprod Update 1999: 5:223-33.
3. Le Bouteiller P. HLA class I chromosomalregion, genes, and products: facts andquestions. Crit Rev Immunol 1994: 14: 89-129.
4. Gobin S, van den Elsen P. The regulationof HLA class I expression: is HLA-G theodd one out? Sem Cane Biol 1999: 9: 55-9.
5. Geraghty DE. Roller BH. Orr HT A humanmajor histocompatibility complex class Igene that encodes a protein with ashortened cytoplasmic segment. Proc NatlAcad SdUSA 1987: 84: 9145-9.
6. Ishitani A, Geraghty DE. Alternativesplicing of HLA-G transcripts yieldsproteins with primary structuresresembling both class I and class IIantigens. Proc Natl Acad SdUSA 1992:89: 3^7-51.
7. Fujii T. Ishitani A. Geraghty DE. A solubleform of the HLA-G antigen is encoded bya messenger ribonucleic acid containingintron 4. J Immunol 1994: 153: 5516-24.
8. Kirszenbaum M. Moreau P, Gluckman E,Dausset J, Carosella E. An alternativelyspliced form of HLA-G mRNA in humantropboblasts and evidence for the presenceof HLA-G transcript in adult lymphocytes.Proc Natl Acad SdUSA 1994: 91: 4209-13.
9. Hiby S. King A. Sharkey A, Loke Y.Molecular studies of trophoblast HLA-G:polymorphism, isoforms. imprinting andexpression in preimplantation embryo.Tissue Ant^ens 1999: 53:1-13.
10. Moreau P, Carosella E, Teyssier M et al.Soluble HLA-G molecule. An alternativelyspliced HLA-G mRNA form candidate toencode it in peripheral blood mononuciearcells and human trophoblasts. HumImmunol 1995: 43: 231-6.
11. Amiot L, Onno M, Drenou B et al.Distribution of HLA-G alternative mRNAsincluding soluble forms in normallymphocytes and in lymphoid cell-derivedletjkemia. Eur f Immunogenet 1996: 23:311-20.
12. McMaster MT, Librach CL, Zhou Y et al.Human placental HLA-G expression isrestricted to differentiatedcytotrophoblasts. y/w»i«Ho/1995: 154:3771-8.
13. Blaschitz A. Lenfant F. Mallet V et al.Endothelial cells in chorionic fetal vesselsof first trimester placenta express HLA-G.FMrf Immunol 1997: 27: 3380-8.
14. Loke YW, King A, Burrows T et al.Evaluation of trophoblast HLA-G antigenwith a specific monoclonal antibody. TissueAntigens 1997: 50: 135-46.
15. Hunt JS. HLA and the maternal-fetalrelationship. Austin: RG Landes Company.1996.
16. Chu W, Fant ME, Geraghty DE, Hunt jS.Soluble HLA-G in human placentas:synthesis in trophoblasts and interferon-gamma-activated macrophages but notplacenta! fibroblasts. Hum Immunol 1998:59: 435^2.
17. Hammer A. Hutter H. Blaschitz A et al.Amnion epithelial cells, in contrast totrophoblast cells, express all classical HLAclass 1 molecules together with HLA-G.Am f Reprod Immunol 1997: 37: 161-71.
18. McMaster M, Zhou Y. Shorter S et al. HLA-G isoforms produced by placentalcytotrophoblasts and found in amnioticfluid are due to unusual glycosylation.f Immunol 1998: 160: 5922-8.
19. Puppo F. Costa M, Contini P et al.Determination of soluble HLA-G and HLA-A, B, and -C molecules in pregnancy.Transplant Proc 1999: 31:1841-3.
20. Rebmann V, Pfeiffer K. Passler M et al.Detection of soluble HLA-G molecules inplasma and anmiotic fluid. Tissue Antigens1999: 53: 14-22.
21. Crisa L. McMaster MT. Ishii JK. Fisher SJ,Salomon DR. Identification of a thymicepithelial cell subset sharing expression ofthe class Ib HLA-G molecule with fetaltrophoblasts. yEr/> Med 1997: 186: 289-98.
22. Mallet V, Blaschitz A. Crisa L et al. HLA-Gin human thymus: subpopulation ofmedullary epithelial but not CD83-fdendritic cells expresses HLA-G as amembrane bound and soluble protein. IntImmunol \°^: 11:889-98.
23. Ue N. Malacko AR, Ishitani A et al. Themembrane-bound and soluble forms ofHLA-G bind identical sets of endogenouspeptides but differ with respect to TAPassociation. Immunity 1995: 3: 591-600.
24. Dieht M, Munz C. Keilholz W et al.Nonclassical HLA-G molecules are classicalpeptide presenters. Curr Biot 1996: 6: 305-14.
Ttesue Antigens 2000: 55: 510-518 517
Fournel et al ; HLA-G mAb reactivity
25. Horuzsko A, Antoniou J, Tomlinson P,
Portik-Uobos V, Mellor AL. HLA-Gfunctions as a restriction element and a
transplantation antigen in mice. Int
Immunol 1997: 9: 645-53.
26. Horuzsko A, Portik-Dobos V, Hansen KA,
Markowitz RB, Hetman SW. Mellor AL.Induction of HLA-G-specific human CD8+
T cell lines by stimulation across apolymorphism of HLA-G. Transplant I'roc
1999: 31: 1860-3.
27. Lanier LL. Natural killer cells fertile withreceptors for HLA-G? Proc Natl Acad Sci
USA 1999:96:5343-5.
28. Vales-Gomez M. Reyburn H, Erskine R.
Lopez-Botet M, Strominger J. Kinetics andpeptide dependency of the binding of the
inhibitory NK receptor CD94/NKG2A andthe activating receptor CD94/NKG2-C to
HLA-E. EMBOJ 1999: 18: 4250-60.
29. Bensussan A, Mansur IG. Mallet V et al.
Detection of membrane-bound HLA-G
translated products with a specific
monwlonal antib<xly. Proc Natl Acad Sd
USA 1995:92:10292-6.
30. Barnstable CJ. Bodmer WF, Brown G et al.
I'roduction of monoclonal antibodies to
group A erythrocytes, HLA and otherhuman cell surface antigens-new tools for
genetic analysis. CeS 1978: 14:9-20.
31. Genestier L. Paillot R. Bonnefoy-Berard N
et al. Fas-independent apoptosis ofactivated T cells induced by antibodies to
the HI.A class I alphal domain. Bbod 1997:90: 3629-39.
32. Seitz C, Uchanska-Ziegler B, Zank A,
Ziegler A. The monoclonal antibody HCA2recognises a broadly shared epitope onselected classical as well as several non-
classical HLA class I molecules. Mol
Immunol 1998: 35: 819-27.
33. Sernee MF, Ploegh HL. Schust DJ. Why
certain antibodies cross-react with HLA-Aand HLA-G: epitope mapping of two
common MHC class I reagents. Mol
Immunol 1998: 35: 177-88.
34. Boucraut J, Hakem R, Gauthier A, FauchetR, Le Bouteiller P. Transfected trophoblast-
darived human ceils can express a singleHI.A class I allelic product Tissue
Antigens 1991:37:84-9.
35. Fournel S, Aguerre-Girr M, Campan A.
Salauze L, Berrebi A, ljine Y, Lenfant F. Le
Bouteiller P. Soluble HLA-G: purification
from eucaryotic transfected cells and
detection by a specific ELISA. AmJReprod
Immunol 1999: 42: 22-9.
36. Grumet FC, Krishnaswamy S, See-Tho K,
Filvaroff E, Hiraki DD. Soluble form of an
HLA-B7 class I antigen specifically
suppresses humoral alloimmunlzation. Hum
Immunol 1994: 40: 228-34.
37. Menu E. Kaplan L, Andreu G, Denver L,
Chaouat G. Immunoactive products ofhuman placenta. I. An immunoregulatoryfactor obtained from explant cultures of
human placenta inhibits CTL generationand cytotoxic effector activity. C^ll Immunol
1989: 119:341-52.
38. Westhoff U, Grosse-Wiide H. Soluble H1,Aclass I and class II concentrations in Factor
VIIT and PCC preparations. Vta: Sanguinis
1995: 68: 7;i-6.
39. Puppo F, Bignardi D, O)ntini P et al. p2m-free HI-A class I heavy chain levels in sera
of healthy individuals. Uck of associationwith P2m-associated HL.A class I heavy
chain levels and HL.^ phenotype. Tissue
Antigens 1999: 53: 253-62,
40. Mallet V, PrOll J, Solier C et al. The full
length HLA-Gl and no other alternativeform of HLA G is expressed at the <x\\
surface of transfected cells. Hum Immunol
2000: 6 1 : 212-24.
41. Puppij F. Indiveri F, Scudeietti M, FerroneS. Soluble 1-ll.A antigens: new roies and
uses. Immunol Today 1997: 18:154^.
42. McDonald J. Adamashvili I. Soluble HI.A:
a review of the literature. Hum Immunol
1998: 59: 387-103.
518 nssue Antigens 2000: 55: 510-518