BIOLOGY OF REPRODUCTION 67, 1211–1217 (2002)
Decidual Lymphocytes of Human Spontaneous Abortions Induce Apoptosis
but Not Necrosis in JEG-3 Extravillous Trophoblast Cells1
Enrique G. Olivares,2,3 Raquel Muñoz,3 Germán Tejerizo,3 Marı́a José Montes,3 Francisca Gómez-Molina,4
and Ana Clara Abadı́a-Molina3
Unidad de Inmunologı́a, Departamento de Bioquı́mica y Biologı́a Molecular,3 Departamento de Obstetricia
y Ginecologı́a,4 Facultad de Medicina, Universidad de Granada, E-18012 Granada, Spain
ABSTRACT
The human decidua contains an unusually high proportion of
lymphocytes, mainly NK and T cells, which are potentially cy-
totoxic to the trophoblast when they are stimulated with certain
cytokines. Given the high incidence of spontaneous abortion in
humans and other species, our working hypothesis is that de-
cidual lymphocytes are involved in immunological mechanisms
that attack the trophoblast and induce abortion when any ges-
tational problem arises. To test this hypothesis, flow cytometry
was used to compare decidual lymphocyte populations in first-
trimester spontaneous abortions and elective terminations of
first-trimester pregnancy. We found significantly higher propor-
tions of decidual lymphocytes that expressed activation markers,
and of T cells (mainly T helper cells) in spontaneous abortions
than in elective terminations of pregnancy. Decidual lympho-
cytes from spontaneous abortion, like decidual lymphocytes
from elective termination of pregnancy and peripheral blood
lymphocytes, were however, unable to lyse the JEG-3 extravil-
lous cytotrophoblast cell line in a 51Cr-release assay. Neverthe-
less, decidual lymphocytes from spontaneous abortion, unlike
decidual lymphocytes from elective termination of pregnancy
and peripheral blood lymphocytes, induced apoptosis in JEG-3
cells as determined by DNA fragment-release assay. Hematox-
ylin and eosin staining showed a significantly higher proportion
of apoptotic JEG-3 cells when these cells were treated with de-
cidual lymphocytes from spontaneous abortion than when JEG-
3 cells were cultured with decidual lymphocytes from elective
termination of pregnancy. The ultrastructural signs of apoptosis
were confirmed by electron microscopy. These data support the
hypothesis that activated decidual lymphocytes participate in
human spontaneous abortion by inducing apoptosis but not ne-
crosis of the trophoblast.
apoptosis, decidua, immunology, pregnancy, trophoblast
INTRODUCTION
The decidua is the maternal tissue in closest contact with
the fetal trophoblast. Immunological interrelations between
the mother and fetus during pregnancy are believed to take
place in this tissue. In normal human decidua or endome-
trium (the nongestating tissue counterpart), an unusually
high proportion of leukocytes, including macrophages and
lymphocytes, are detected throughout pregnancy and men-
strual cycles. In early human decidua most lymphocytes are
CD561CD162 NK cells, and a smaller proportion are T
1
Supported by grants from the Fondo de Investigaciones Sanitarias de la
Seguridad Social (Spanish Ministry of Health).
2
Correspondence: FAX: 34958249015; e-mail: engarcia@ugr.es
Received: 22 June 2001.
First decision: 11 July 2001.
Accepted: 17 May 2002.
Q 2002 by the Society for the Study of Reproduction, Inc.
ISSN: 0006-3363. http://www.biolreprod.org
1211
cells [1]. Thus far, the functions of these decidual lympho-
cytes remain unknown. They may play a role in local de-
fense against infection; some authors propose that they also
exert nonimmune functions involving fetal sustenance [2].
Decidual lymphocytes, however, have been shown to be
potentially cytotoxic for the trophoblast [3, 4]. It is there-
fore highly paradoxical that in normal pregnancy these cells
accumulate in the decidua, where under normal conditions,
cytotoxicity should be overridden in favor of maternal tol-
erance.
Although pregnancy has been considered as an example
of semiallogeneic graft acceptance, this graft is not always
successful because between one-third and one-half of all
human conceptions fail to progress [5]. There is increasing
evidence that the immune system participates in the mech-
anism of fetal elimination during spontaneous abortion [6–
8]. In mice and humans, normal pregnancy is related to the
local and peripheral production of Th2 cytokines [9, 10],
whereas abortion is associated with Th1 cytokine produc-
tion [10, 11] or with a decrease in Th2 cytokines [12].
Recent findings in mice have shown that maternal T cells
tend to reject the semiallogeneic embryo, although under
normal conditions this rejection is inhibited by catabolism
of tryptophan by the conceptus [7], by clonal deletion of
fetal-reactive maternal T cells [13], and by the production
of Th2 cytokines [9]. Moreover, human trophoblast ex-
presses HLA-G, a nonclassical human leukocyte antigen
(HLA) class I molecule that binds the inhibitory receptors
of cytotoxicity. The high expression of these receptors by
normal decidual NK cells [14], and also probably by T cells
[15], leads to the protection of the trophoblast against cell
cytolysis [14].
These previous findings suggest that decidual T and NK
lymphocytes are prone to attack the trophoblast, and prob-
ably do so during spontaneous abortion, although their cy-
tocidal activity is down-regulated by different mechanisms
during normal pregnancy [7, 9, 13]. To investigate the par-
ticipation of the immune system in human spontaneous
abortion, we compared the lymphocyte populations of de-
cidua and cytotoxic activity of these lymphocytes against
the trophoblast in human spontaneous abortion and elective
termination of pregnancy (ETP).
MATERIALS AND METHODS
Subjects
Fifty-three samples from first-trimester spontaneous abortion (in wom-
en aged 21–32 yr; gestational age, 8 wk 6 10 days) were collected at the
Departamento de Obstetricia y Ginecologı́a, Hospital Universitario San
Cecilio, Granada. Anembryonic pregnancies or fetal death was confirmed
by ultrasonography. Fifty-five specimens from ETP (age range, 23–35 yr;
gestational age, 8 wk 2 days 6 7 days) were obtained at the Clı́nica El
Sur (Málaga) and Gineclı́nica (Granada). Samples were collected by vag-
inal curettage; in spontaneous abortion, curettage was carried out within
24 h after diagnosis. None of the abortions was pharmacologically in-
1212 OLIVARES ET AL.

TABLE 1. Monoclonal antibodies used.

Immunoglobulin
Antibody Specificity Clone isotype Labeled witha Obtained fromb
Isotype control — — IgG1 FITC or PE Sigma
Isotype control — — IgG2a FITC or PE Sigma
Isotype control — — IgG2b FITC or PE Sigma
Isotype control — — IgM FITC Sigma
OKT3 CD3 OKT3 IgG2a FITC or PE OD
Anti-TCRab TCRab WT31 IgG1 FITC BD
Anti-TCRgd TCRgd 11F2 IgG1 FITC BD
OKT4 CD4 OKT4 IgG2b FITC/PE OD
OKT8 CD8 OKT8 IgG2a FITC OD
Leu-2a CD8 SK1 IgG1 PE BD
OKB20 CD20 NU-B2 IgG2b FITC OD
OK-NK CD16 OK-NK IgM FITC OD
CD56 CD56 NK1-nb11 IgG1 PE CALTAG
OKT26a CD25 OKT26a IgG2a FITC OD
Leu-17 CD38 EBVCS-5 IgG1 PE BD
Leu-23 CD69 HB-7 IgG1 PE BD
OKDR HLA-DR SC-2 IgG2a PE OD
a FITC, Fluorescein isothiocyanate (green); PE, phycoerythrin (red).
b OD, Ortho-Diagnostic System, Raritan, NJ; BD, Beckton-Dickinson, Erembodegem-Aalst, Belgium; CALTAG, San Francisco, CA.

duced. Both groups received the same exclusionary criteria: women re-
ceiving any medication or with infectious, autoimmune, or other systemic
or local diseases, were excluded. Informed consent was obtained from
each patient. Blood progesterone from patients with spontaneous abortion
was quantified with a commercial enzyme immunoassay (Boehringer-
Mannheim, Mannheim, Germany). This research was approved by the eth-
ics committee of the Hospital Universitario San Cecilio, Granada.
Extraction of Decidual Lymphocytes
Samples of decidua from different patients were not mixed in order to
avoid the induction of allogeneic reaction of leukocytes. The method of
extraction has been described elsewhere [16]. Briefly, samples from de-
cidua of spontaneous abortion or ETP were thoroughly washed in PBS.
Decidual fragments were finely minced in a small volume of RPMI 1640
(Sigma, St. Louis, MO) and then pushed through a 53-mm sieve (Gallen-
kamp, Loughborough, U.K.). The resultant cell suspension was washed
with RPMI and layered on an equivalent volume of Lymphoprep (Flow
Laboratories, Hertsfordshire, U.K.) at room temperature, and centrifuged
for 20 min at 600 3 g. The cells were collected from the interface, sus-
pended in RPMI, and washed. The cells were then suspended in complete
culture medium (RPMI 1640, 10% fetal calf serum [FCS], 100 U/ml pen-
icillin, and 50 g/ml gentamicin), and incubated for 2 h at 378C in an
atmosphere of 5% CO2 to allow adherent cells to attach to the plastic. The
supernatant containing decidual lymphocytes was then collected, and the
cells were washed and suspended in PBS for flow cytometric analysis.
Lymphocyte viability was microscopically determined by trypan blue ex-
clusion. Only samples with more than 90% viable lymphocytes were used.
To obtain peripheral blood lymphocytes (PBLs), blood samples were
taken from healthy volunteers. We diluted each sample in the same volume
of 0.25% PBS-EDTA, and centrifuged them on Lymphoprep. Thereafter,
we followed the same steps that we used for decidual lymphocytes.
Flow Cytometry
One hundred microliters of a suspension of 1 3 106/ml of decidual
lymphocytes in PBS was incubated with 10 or 20 ml of the appropriate
monoclonal antibody (Table 1) for 30 min at 48C in the dark. Cells were
washed, suspended in 1 ml of PBS, and immediately analyzed in a flow
cytometer (Ortho Cytoron Absolute; Ortho Diagnostic Systems). To iden-
tify dead cells we incubated lymphocytes with propidium iodide (Sigma).
Although the isolated cells were mostly lymphocytes, cells were studied
in the electrically gated lymphocyte cluster. The percentage of cells that
were antibody-positive was calculated by comparing then with the appro-
priate isotype control.
Cell Line
JEG-3, an extravillous trophoblast (EVT) choriocarcinoma cell line,
was maintained in complete culture medium RPMI 1640 containing 5%
FCS, and cells were used as targets when they were in the log phase of
growth.
Culture of Lymphocytes with Interleukin-2
Decidual lymphocytes or PBLs were cultured for 4 days in complete
medium containing 100 U/ml of interleukin-2 (IL-2; Sigma). After this
period of incubation, cells were washed and suspended in complete culture
medium for cytotoxicity studies.
51
Cr-Release Assay
JEG-3 cells were removed from the flask with 0.05% trypsin/0.02%
EDTA, washed, and suspended in complete culture medium. Cells (2 3
104) were added to each well of a 96-well flat-bottomed plate (Becton
Dickinson, San Jose, CA) and left to incubate overnight at 378C in 5%
CO2. The plates were washed twice with medium, and 3 mCi of Na251CrO4
(Amersham, Buckinghamshire, U.K.) in 35 ml of culture medium was
added to each well. After overnight labeling in a moist atmosphere of 5%
CO2 at 378C, the cells were washed three times, and then 100 ml of com-
plete culture medium was added. Effector cells from decidua or PBLs were
added in a volume of 100 ml of culture medium to obtain different effector:
target ratios. After 5 h of incubation at 378C, 100 ml of supernatant was
removed from each well and counted with a gamma counter. Percentage
cytotoxicity was calculated according to the formula
% cytotoxicity 5 [(test cpm 2 spontan. cpm)
4 (max. cpm 2 spontan. cpm)] 3 100.
This and the following experiments were carried out in triplicate.
DNA Fragment Assay
To investigate the induction of apoptosis in target cells we studied
DNA fragmentation as an early sign of this phenomenon. One hundred
microliters of complete medium containing 2 3 104 JEG-3 cells was added
to each well of a 96-well flat-bottomed plate. Then, 0.5 mCi of [3H]thy-
midine ([3H]TdR; Amersham) was added per well, and the plates were
incubated overnight at 378C in 5% CO2. The supernantants were discarded
and 100 ml of complete culture medium was added. Different amounts of
effector cells from decidua or PBLs were added in a volume of 100 ml of
culture medium to obtain different effector:target ratios, as shown in the
figures. After 5 h of incubation at 378C in 5% CO2, the plates were cen-
trifuged at 400 3 g for 10 min, and supernatants were replaced with 200
ml of hypotonic lysing buffer (10 mM Tris, 1 mM EDTA pH 7.5) con-
taining 0.2% Triton X-100. Twenty-five microliters of supernatant was
carefully removed from each well and counted in a b-scintillation counter.
Wells with target cells and no effectors were used to determine sponta-
neous release (spontan. cpm). A solution of 0.1% SDS was used as a
positive control for DNA fragmentation. To determine maximal cpm (max.
cpm), wells with target cells and no effectors were harvested onto glass
 DECIDUAL LYMPHOCYTES INDUCE APOPTOSIS IN THE TROPHOBLAST 1213

fiber filters using a cell harvester (Skatron Instruments AS, Lier, Norway)
and counted in a b-scintillation counter. The results were expressed ac-
cording to the formula
% fragmented DNA 5 [(test cpm 2 spntan. cpm)
4 (max cpm 2 spontan. cpm)] 3 100.
Our [3H]TdR-release assay was based on the methods described of
Matzinger [17] and Duke et al. [18]. As in the Matzinger test [17], we
carried out the assay in a microtiter plate format; however, as in classical
DNA fragment assays [18], radioactivity was measured directly in the
supernatant after centrifugation. The direct measurement of DNA frag-
ments makes the assay more sensitive.

Light Microscopy
Fifty thousand JEG-3 cells were cultured on a Lab-Tek chamber slide
system (Nalge Nunc International, Naperville, IL) and allowed to attach;
the slide was then washed with PBS, and 1.5 3 106 decidual lymphocytes
in 500 ml of culture medium was added. After 5 h of incubation at 378C
in 5% CO2, the slide was washed with PBS, fixed with methanol, stained
with hematoxylin and eosin, and then mounted under coverslips with di-
butyl phthalate xylene mountant for microscopy. Each preparation was
examined at a magnification of 10003 oil immersion with a 1003 objec-
tive lens and a 103 eyepiece. JEG-3 cells were much larger than lym-
phocytes, therefore, the two types of cell were easily distinguished. Apo-
ptotic JEG-3 cells were identified by the condensation of nuclear hetero-
chromatin into a crescent apposed to the nuclear membrane, and later into
a single body or multiple-dense bodies. Apoptotic and nonapoptotic JEG-
3 cells were counted and the results were expressed as percentages of
apoptotic JEG-3 cells.

Transmission Electron Microscopy

For transmission electron microscopy, we cultured decidual lympho-
cytes with JEG-3 cells on a slide as we did for light microscopy. The
preparation was fixed in 1% glutaraldehyde in 0.1 M sodium cacodylate/
0.1% sucrose buffer. Cells were washed in the same buffer and postfixed
in 2% osmium tetroxide, dehydrated through a graded acetone series, and
embedded in Epon 812. Thin sections were cut on a Reichert-Jung Ultracut
E ultramicrotome (Vienna, Austria), stained with lead citrate, and visual-
ized and photographed in a transmission electron microscope (10C; Carl
Zeiss, Inc., Oberkochen, Germany). Apoptotic cells were identified mainly
by the nuclear alterations described in the previous section of light mi-
croscopy, and by the formation of apoptotic bodies. We also confirmed
the integrity of the plasma membrane.
Statistical Analysis
Proportions of decidual lymphocytes and percentages of apoptotic cells
were compared with the Student t-test. A P value of , 0.05 was consid-
ered statistically significant.
RESULTS
Lymphocyte Populations of Decidua from First-Trimester
ETP and Spontaneous Abortion
Figure 1 shows the comparison between the lymphocyte
populations from spontaneous abortion deciduas and pre-
sumed normal deciduas (i.e., ETP). No significant differ-
ences were found between the two types of decidua in the
population of CD201 B cells or in the NK populations of
CD561, CD161, CD561CD161, CD561CD162 and
CD56 1 CD8 1 lymphocytes. However, the number of
CD31, TCRab1, and CD31TCRab1 T cells were sig-
nificantly higher in spontaneous abortion. This increase was
determined mainly by T helper cells, as CD41 lympho-
cytes were also significantly elevated in spontaneous abor-
tion compared with those in ETP. CD81 lymphocytes
(CD31CD81 or CD81TCRab) also appeared to contrib-
ute to the increase in decidual T lymphocytes in sponta-
neous abortion, although the differences in the proportions
of these populations were not statistically significant. We
found no differences in the populations of Tgd lympho-
FIG. 1. Comparison of the proportions of decidual lymphocytes from
elective termination of pregnancy (white; n 5 25) and spontaneous abor-
tion (black; n 5 25) by flow cytometry analysis. Results are shown as
means 6 SD. *P , 0.05; **P , 0.01; ***P , 0.005; ****P , 0.001;
*****P , 0.00005.
cytes. Decidual lymphocytes from spontaneous abortion
seemed to be activated because the percentages of lympho-
cytes expressing the activation markers CD25, CD38, and
CD69 were elevated in spontaneous abortion. Populations
of activated T cells (CD31CD251 and CD41CD691) and
probably activated NK cells (CD56 1 CD25 1 and
CD561HLA-DR1) were significantly higher in spontane-
ous abortion. A small population of CD561TCRab1 cells,
which may correspond to natural T (NT) cells [19] was also
significantly elevated in spontaneous abortion in compari-
son with ETP.
We did not observe significant differences between em-
bryonic and anembryonic pregnancies in the lymphocyte
populations of spontaneous abortion deciduas (not shown),
suggesting that the presence or absence of embryo is not a
determining factor in the variations in decidual lymphocyte
proportions. Although serum levels of progesterone were
significantly lower in patients that had spontaneous abortion
(10.3 6 10.4 ng/ml) than in ETP (26.7 6 6.2 ng/ml) (P ,
0.01), we did not find any correlation between these levels
and the proportions of spontaneous abortion decidual lym-
phocytes (not shown).
1214 OLIVARES ET AL.
FIG. 2. Two representative experiments
of seven 51Cr-release assays of the lytic ac-
tivity against JEG-3 cells of decidual lym-
phocytes from spontaneous abortion
(v—v), elective termination of pregnancy
(m—m), elective termination of pregnancy
stimulated with IL-2 (n—n), peripheral
blood lymphocytes (m—m), and peripher-
al blood lymphocytes stimulated with IL-2
(□—□). Results are shown as means 6
SD.
Cytotoxicity of Decidual Lymphocytes from ETP
and Spontaneous Abortion Against Extravillous
Cytotrophoblast JEG-3 Cells
Neither decidual lymphocytes from ETP, spontaneous
abortion, nor PBLs showed spontaneous cytotoxicity
against JEG-3 cells in the 51Cr-release assay (which iden-
tifies necrosis). These target cells were, however, lysed
when the lymphocytes were previously stimulated with IL-
2 (Fig. 2). Decidual lymphocytes from spontaneous abor-
tion, but not PBLs or decidual lymphocytes from ETP,
spontaneously induced apoptosis in JEG-3 cells, as deter-
mined by the DNA fragment assay (Fig. 3).
Quantification of Apoptotic JEG-3 Cells Treated
with Decidual Lymphocytes from ETP
and Spontaneous Abortion
JEG-3 cells cultured with decidual lymphocytes from ei-
ther ETP or spontaneous abortion were stained with he-
matoxylin and eosin, and the proportions of apoptotic JEG-
3 cells were determined by light microscopy. The propor-
tions of apoptotic cells were significantly higher in the
preparations of JEG-3 cells cultured with decidual lympho-
cytes from spontaneous abortion than with those from ETP
(P 5 1.86 3 1025) (Fig. 4).
Ultrastructure of JEG-3 Cells Treated with Decidual
Lymphocytes from ETP and Spontaneous Abortion
JEG-3 cells cultured with decidual lymphocytes from
ETP or spontaneous abortion were observed with electron
microscopy. Most JEG-3 cells showed a normal morphol-
ogy after culture with decidual lymphocytes from ETP (Fig.
5A). However, when JEG-3 cells were cultured with decid-
ual lymphocytes from spontaneous abortion, we detected
lymphocytes that sent membrane prolongations to or that
FIG. 3. Two representative experiments
of seven DNA fragment assays of activity
against JEG-3 cells of decidual lympho-
cytes from spontaneous abortion (v—v),
elective termination of pregnancy (m—m),
and peripheral blood lymphocytes
(m—m). Results are shown as means 6
SD.
interacted with JEG-3 cells (Fig. 5, B–E). Some of the JEG-
3 cells exhibited clear signs of apoptosis, although their
plasma membrane remained intact (Fig. 5, D–F): conden-
sation and shrinkage of nuclear material, chromatin aggre-
gation, pyknosis of the nuclei (Fig. 5, E and F), compacted
cytoplasm containing numerous vacuoles (Fig. 5, D–F), cell
blebbing, and cell fragmentation with the formation of ap-
optotic bodies (Fig. 5E). No signs of necrosis were evi-
denced.
DISCUSSION
The high rate of pregnancy loss in humans [5] suggests
the existence of selective mechanisms that stop gestation
when conditions are less than optimal. Experimental evi-
dence [6–8, 10–12] has shown that the maternal immune
system is involved in the elimination of the fetus. Our re-
sults, in which we show that populations of T lymphocytes
are increased in spontaneous abortion decidua (Fig. 1), and
that spontaneous abortion decidual lymphocytes induced
apoptosis in the EVT cell line JEG-3 (Figs. 3–5), also sup-
port this view.
Although decidual lymphocytes constitutively express
activation antigens [4, 20], the significant increase in the
proportions of decidual lymphocytes expressing these an-
tigens (CD25, CD38, and CD69) in spontaneous abortion
(Fig. 1) further supports that a greater degree of immune
activation occurs in this situation [21–23]. In mice and hu-
mans, normal pregnancy is related to the local and periph-
eral production of Th2 cytokines [9, 10], whereas abortion
is associated with Th1 cytokine production [10, 11] or with
a decrease in Th2 cytokines [12]. It is therefore probable
that the increase in CD41 cells detected by us in sponta-
neous abortion decidua corresponds to Th1 cells. A small
population of cells with a peculiar phenotype
(CD561TCRab1) was also significantly higher in spon-
 DECIDUAL LYMPHOCYTES INDUCE APOPTOSIS IN THE TROPHOBLAST
taneous abortion in comparison with ETP (Fig. 1). This
phenotype may correspond to that of CD31CD561 natural
T (NT) cells, detected in human liver, which produce main-
ly Th1 cytokines [19]. In this connection, decidual
CD561TCRab1 cells may also be a source of Th1 cyto-
kines in human spontaneous abortion. It is interesting that
decidual NTs, which secrete Th1 cytokines, have been
shown to be involved in murine abortion [24].
An increase in Th1 cytokines may directly induce cy-
totoxicity or activate cytotoxic cells, which may lead to
destruction of the trophoblast [6]. Although recent results
have shown that human decidual lymphocytes may play a
role in placental detachment during parturition [25], these
lymphocytes from normal pregnancies were unable to spon-
taneously lyse the trophoblast in vitro [4, 16]. Nevertheless,
they became lytic for the trophoblast after stimulation with
IL-2 [3, 4] (Fig. 2). This suggested that decidual lympho-
cytes were potentially cytotoxic, although in normal preg-
nancies the preponderance of Th2 cytokines probably
blocks Th1 cells, and hence inhibits cytotoxicity against the
trophoblast [6, 9]. On the contrary, in spontaneous abortion
we expected Th1-activated decidual lymphocytes to lyse
trophoblast cells. Nevertheless, we found that neither the
decidual lymphocytes of spontaneous abortion nor those of
ETP or PBLs could lyse the JEG-3 EVT cell line in the
51Cr-release assay. On the other hand, decidual lympho-
cytes of spontaneous abortion, unlike decidual lymphocytes
of ETP or PBLs, spontaneously induced DNA fragment
release in JEG-3 (Fig. 3). In this connection, using histo-
logical methods, we confirmed the incidence of a signifi-
cantly higher proportion of JEG-3 cells showing morpho-
logical signs of apoptosis, but not of necrosis, when these
cells were cultured with decidual lymphocytes from spon-
taneous abortion (Figs. 4 and 5).
These data show that decidual lymphocytes of sponta-
neous abortion induce apoptosis, but not necrosis, in the
EVT. Our results are consistent with those of Kokawa et
al. [26], who also detected (by molecular biochemical tech-
niques) an increase in low molecular weight DNA frag-
ments in the trophoblast of spontaneous abortion in com-
parison with ETP. Kokawa et al. [26] observed a limited
but detectable cleavage of DNA in the trophoblast of nor-
mal pregnancies. We also detected a proportion of JEG-3
cells with signs of apoptosis (2.96 6 1.73) when these cells
were cultured with decidual lymphocytes from ETP (Fig.
4). Nevertheless, we observed no activity in the DNA frag-
ment assay when decidual lymphocytes of ETP were used
as effectors against JEG-3 cells (Fig. 3). This apparent dis-
crepancy may be attributable to the fact that the incidence
of spontaneous [3H]TdR release by JEG-3 cells during this
assay might have interfered with the detection of the DNA
fragments induced by the decidual lymphocytes from ETP.
Several authors [27, 28] who used hematoxylin and eo-
sin staining have also reported apoptosis in trophoblast
from normal pregnancies. Our percentages for ETP were
much higher than those reported by Smith et al. [27], and
higher than but close to those of Chan et al. [28]. The
differences are probably because we used an in vitro system
with an EVT cell line, whereas those authors determined
apoptosis in sections of placenta. Nevertheless, other au-
thors who used the in situ DNA ligation method reported
proportions higher than ours in first-trimester normal pla-
centa [29]. The incidence of apoptosis during normal preg-
nancy suggests that apoptosis is a physiological mechanism
by which excessive trophoblast expansion is controlled by
decidual lymphocytes during normal pregnancies [29]. Our
1215
FIG. 4. Percentages of apoptotic JEG-3 cells after culture with decidual
lymphocytes from elective termination of pregnancy (ETP; n 5 14) or with
decidual lymphocytes from spontaneous abortion (SAB; n 5 14). The re-
sults were determined by hematoxylin and eosin staining and light mi-
croscopy and were expressed as means 6 SD.
results and those of Kokawa et al. [26] suggest that during
spontaneous abortion, this physiological mechanism of con-
trol is more intensely activated, leading to a more extensive
destruction of the trophoblast by apoptosis, which results
in the end of pregnancy. The mechanism by which apopto-
sis is induced in the trophoblast remains to be determined.
Th1 cytokines such as tumor necrosis factor are directly
cytotoxic to the trophoblast [30]; decidual lymphocytes
such as NK and CD81 cytotoxic T lymphocytes (CTLs)
may also be activated by Th1 cytokines. In fact, the small
CD561 CD251 and CD561 HLA-DR1 subpopulations
that are significantly higher in spontaneous abortion decid-
ua in comparison with ETP (Fig. 1) may correspond to
activated NK cells, which may also contribute to tropho-
blast elimination. Furthermore, the expression of killer in-
hibitory receptors (KIRs) by decidual NK cells, which in
normal pregnancies is high [14], is reduced in human spon-
taneous abortion. This probably leads to the disinhibition
of cell cytotoxicity against the trophoblast [22].
Most spontaneous abortions occur because of chromo-
somal abnormalities in the embryo. The mechanism that
links this (or any) nonimmunological cause to the immune
response of the decidua against the trophoblast remains to
be elucidated. We have found that the presence of an em-
bryo does not appear to be an important factor in the im-
mune response because in our study there was no signifi-
cant difference in the proportions of decidual lymphocytes
between embryonic and anembryonic spontaneous abortion
(not shown). Progesterone up-regulates the production of
Th2 cytokines by lymphocytes [31], inhibits the production
of Th1 cytokines [32], and appears to down-regulate CTL
activity in the uterus [33]. In our cases of spontaneous abor-
tion, we observed a significant decrease in progesterone
concentration in peripheral blood from patients who had
undergone spontaneous abortion. The reduced secretion of
progesterone may trigger a Th1 response, which in turn,
would activate cytotoxicity against the trophoblast. Never-
theless, we found no correlation between the serum con-
centrations of progesterone and any of the proportions of
decidual lymphocytes in spontaneous abortion (not shown).
It is probable that local rather than peripheral concentra-
tions of progesterone correlate better with these propor-
tions.
The Th1-Th2 balance appears to be a physiological
1216 OLIVARES ET AL.

FIG. 5. Ultrastructural changes in JEG-3

cells after interaction with decidual lym-
phocytes from elective termination of
pregnancy (A, n 5 5) or with decidual
lymphocytes from spontaneous abortion
(B–F, n 5 5). A JEG-3 cell with a normal
morphology (A). Decidual lymphocytes
sending out membrane prolongations (B)
or interacting with JEG-3 cells (C and D).
JEG-3 cells showing typical apoptotic mor-
phology: apoptotic bodies (arrows in E)
and condensed chromatin with an intact
plasma membrane (E and F). Original
magnifications: 33400 (A), 35000 (B),
36300 (C and D), 34000 (E), and 34800
(F).

mechanism that may lead either to Th2 cytokine production
and successful pregnancy or Th1 cytokine production and
spontaneous abortion [6]. It seems that decidual lympho-
cytes are always prone to kill the trophoblast, although their
activity is blocked under normal conditions [7, 9, 10, 12–
15]. Maternal immune activity against the trophoblast may
be an efficient biological mechanism to eliminate a fetus
when any gestational problem arises; this would select only
normal fetuses to develop to term. Conception rates in
mammals are high enough to allow for such a mechanism
of natural selection.
ACKNOWLEDGMENTS
We are grateful to Dr. S. Jordán and Dr. C. Sánchez from the Clı́nica
el Sur (Málaga), and Dr. A. Stolzenburg from Gineclı́nica (Granada), and
to Dr. A. Caño from Servicio de Obstetricia y Ginecologı́a del Hospital
Universitario San Cecilio de Granada for providing us with decidual spec-
imens. We thank K. Shashok for improving the manuscript, Dr. Pepi León
for her skillful help with the figures, and Dr. Antonio Rios for his help
with the electron microscopy images.
REFERENCES
1. Bulmer JN. Immune cells in decidua. In: Kurpisz M, Fernandez N
(eds.), Immunology of Human Reproduction. Oxford: Bios Scientific
Publishers Ltd.; 1995: 313–334.
2. Guimond MJ, Wang B, Croy BA. Engraftment of bone marrow from
severe combined immunodeficient (SCID) mice reverses the repro-
ductive deficits in natural killer cell-deficient tge26 mice. J Exp Med
1998; 187:217–223.
3. King A, Loke YW. Human trophoblast and JEG choriocarcinoma cells
are sensitive to lysis by IL-2-stimulated decidual NK cells. Cell Im-
munol 1990; 122:435–448.
 DECIDUAL LYMPHOCYTES INDUCE APOPTOSIS IN THE TROPHOBLAST
4. Abadı́a-Molina AC, Ruiz C, Montes, MJ, King A, Loke YW, Garcı́a-
Olivares E. Immune phenotype and cytotoxic activity of lymphocytes
from human term decidua against trophoblast. J Reprod Immunol
1996; 31:109–123.
5. Wilcox AJ, Weinberg CR, O’Connor JF, Baird DD, Schlatherer JP,
Canfield RE, Armstrong EG, Nisula BC. Incidence of early loss of
pregnancy. New Engl J Med 1988; 319:189–194.
6. Raghupathy R. Th1-type immunity is incompatible with successful
pregnancy. Immunol Today 1997; 18:478–482.
7. Munn DH, Zhou M, Attwood JT, Bondarev I, Conway SJ, Marshall
B, Brown C, Mellor AL. Prevention of allogeneic fetal rejection by
tryptophan catabolism. Science 1998; 281:1191–1193.
8. Baker JM, Bamford AI, Antezak DF. Modulation of allospecific CTL
responses during pregnancy in equids: an immunological barrier to
interspecies matings? J Immunol 1999; 162:4496–4501.
9. Lin H, Mosmann TR, Guilbert L, Tuntipopipat S, Wegmann TG. Syn-
thesis of T helper 2-type cytokines at the maternal-fetal interface. J
Immunol 1993; 151:4562–4573.
10. Marzi M, Vigano A, Trabattoni D, Villa ML, Salvaggio A, Clerici E,
Clerici M. Characterization of type 1 and type 2 cytokine production
profile in physiologic and pathologic human pregnancy. Clin Exp Im-
munol 1996; 106:127–133.
11. Tangri S, Wegmann TG, Lin H, Raghupathy R. Maternal anti-placen-
tal reactivity in natural, immunologically-mediated fetal resorptions. J
Immunol 1994; 152:4903–4912.
12. Piccinni MP, Beloni L, Livi C, Maggi E, Scarselli G, Romagnani G.
Defective production of both leukemia inhibitory factor and type 2 T-
helper cytokines by decidual T cells in unexplained recurrent abor-
tions. Nat Med 1998; 4:1020–1024.
13. Jiang SP, Vacchio MS. Multiple mechanisms of peripheral T cell tol-
erance to the fetal ‘‘allograft.’’ J Immunol 1998; 160:3086–3090.
14. Ponte M, Cantoni C, Biassoni R, Tradori-Cappai A, Bentivoglio G,
Vitale C, Bertone S, Moretta A, Moretta L, Mingari MC. Inhibitory
receptors sensing HLA-G1 molecules in pregnancy: decidua-associ-
ated natural killer cells express LIR-1 and CD94/NKG2A and acquire
p49, an HLA-G1-specific receptor. Proc Natl Acad Sci U S A 1999;
96:5674–5679.
15. Fournel S, Aguerre-Guirr M, Huc X, Lenfant F, Alam A, Toubert A,
Bensussan A, Le Boutillier P. Soluble HLA-G1 triggers CD95-CD95
ligand-mediated apoptosis in activated CD81 cells by interacting with
CD8. J Immunol 2000; 164:6100–6104.
16. King A, Birkby C, Loke YW. Early human decidual cells exhibit NK
activity against K562 line but not against first trimester trophoblast.
Cell Immunol 1989; 118:337–344.
17. Matzinger P. The JAM test: a simple assay for DNA fragmentation
and cell death. J Immunol Methods 1991; 145:185–192.
18. Duke RC, Chervenak R, Cohen JJ. Endogenous endonuclease-induced
DNA fragmentation: an early event in cell-mediated cytolysis. Proc
Natl Acad Sci U S A 1983; 80:6361–6365.
19. Doherty DG, Norris S, Madrigal-Estebas L, McEntee G, Traynor O,
Hegarty JE, O’Farrelly C. The human liver contains multiple popu-
1217
lations of NK cells, T cells and CD31CD561 natural T cells with
distinct cytotoxic activities and Th1, Th2, and Th0 cytokine secretion
patterns. J Immunol 1999; 163:2314–2321.
20. Saito S, Nishikawa K, Morii T, Narita N, Enomoto M, Ichijo M. Ex-
pression of activation antigens CD69, HLA-DR, interleukin-2 recep-
tor-alpha (IL-2Ralfa) and IL-2Rbeta on T cells of human decidua at
an early stage of pregnancy. Immunology 1992; 75:710–712.
21. Maruyama T, Makino T, Sugi T, Matsubayashi TH, Ozawa N, Nozawa
S. Flow-cytometric analysis of immune cell populations in human
decidua from various types of first-trimester pregnancy. Hum Immu-
nol 1992; 34:212–218.
22. Chao KH, Wu MY, Chen CD, Yang JH, Yang YS, Ho HN. The ex-
pression of killer cell inhibitory receptors on natural killer cells and
activation status of CD41 and CD81 T cells in the decidua of normal
and abnormal early pregnancies. Hum Immunol 1999; 60:791–797.
23. Vassiliadou N, Searle RF, Bulmer JN. Elevated expression of activa-
tion molecules by decidual lymphocytes in women suffering sponta-
neous early pregnancy loss. Hum Reprod 1999; 14:1194–1200.
24. Ito K, Karasawa M, Kawano T, Akasaka T, Koseki H, Akutsu Y,
Kondo E, Sekiya S, Sekikawa K, Harada M, Yamashita M, Nakayama
T, Taniguchi M. Involvement of decidual Va14 NKT cells in abortion.
Proc Natl Acad Sci U S A 2000; 97:740–744.
25. Abadı́a-Molina AC, Ruiz C, King A, Loke, YW, Olivares EG. Lym-
phocytes of human term decidua decrease cell adhesion to a plastic
substrate. Hum Reprod 1997; 12:2393–2398.
26. Kokawa K, Shikone T, Nakano R. Apoptosis in human chorionic villi
and decidua during normal embryonic development and spontaneous
abortion in the first trimester. Placenta 1998; 19:21–26.
27. Smith SC, Baker PN, Symonds EM. Placental apoptosis in normal
human pregnancy. Am J Obstet Gynecol 1997; 177:57–65.
28. Chan CC, Lao TT, Cheung AN. Apoptosis in human placenta. Am J
Obstet Gynecol 1998; 179:1377–1378.
29. Al-Lamki RS, Skepper JN, Loke YW, King A, Burton GJ. Apoptosis
in the human placental bed and its discrimination from necrosis using
the in-situ DNA ligation technique. Hum Reprod 1998; 12:3511–
3519.
30. Knofler M, Mosl B, Bauer S, Griesinger G, Husslein P. TNF-alpha/
TNFRI in primary and immortalized first trimester cytotrophoblast.
Placenta 2000; 21:525–535.
31. Piccinni MP, Giudizi MG, Biagiotti R, Beloni L, Giannarini L, Sam-
pognaro S, Parronchi P, Manetti R, Annunziato F, Livi C, Romagnani
S, Maggi E. Progesterone favours the development of human T helper
cells producing Th2-type cytokines and promotes both IL-4 produc-
tion and membrane CD30 expression in established Th1 cell clones.
J Immunol 1995; 155:128–33.
32. Choi BC, Polgar K, Xiao L, Hill JA. Progesterone inhibits in-vitro
embryotoxic Th1 cytokine production to trophoblast in women with
recurrent pregnancy loss. Hum Reprod 2000; 15:46–59.
33. White HD, Crassi KM, Givan AL, Stern JE, González JL, Memoli
VA, Green WT, Wira CR. CD31 CD81 CTL activity within the hu-
man female reproductive tract. Influence of stage of the menstrual
cycle and menopause. J Immunol 1997; 158:3017–3027.