Content uploaded by Eric Jauniaux
Author content
All content in this area was uploaded by Eric Jauniaux on Feb 09, 2014
Content may be subject to copyright.
ORIGINAL ARTICLE Early pregnancy
Investigation of systemic inflammatory
response in first trimester
pregnancy failure
J. Calleja-Agius1, 4,*, E. Jauniaux1, A.R. Pizzey2, and S. Muttukrishna3
1
Academic Department of Obstetrics and Gynaecology, UCL Institute for Women’s Health, University College London, 86-96, Chenies
Mews, London WC1E 6HX, UK
2
Research Department of Haematology, UCL Cancer Institute, Paul O’Gorman Building 72, Huntley Street,
London WC1E 6BT, UK
3
Anu Research Centre, Department of Obstetrics and Gynaecology, University College Cork, Cork University
Maternity Hospital, Wilton, Cork, Republic of Ireland
*Correspondence address. Tel: +356-21693041; Fax: +356-21319527; E-mail: jean.calleja-agius@um.edu.mt
Submitted on May 1, 2011; resubmitted on October 17, 2011; accepted on October 28, 2011
background: The contribution of local and systemic inflammation to the pathophysiology of sporadic first trimester miscarriages
remains unclear. The objective of this study was to investigate the inflammatory response in the circulation of women presenting with
first trimester miscarriage.
methods: Levels of tumour necrosis factor alpha (TNFa), TNF receptors 1 and 2, interferon gamma (IFNg), interleukin (IL)-6 and IL-10
were assayed using cytometric bead arrays in plasma samples from 29 euploid and 21 aneuploid missed miscarriages, 35 normal pregnant
controls and 31 non-pregnant women (NPW). Whole blood flow cytometry was carried out with samples from 17 euploid and 16 aneuploid
miscarriages, 18 pregnant controls and 13 NPW.
results: The plasma of women with euploid miscarriage contained significantly higher circulating levels of TNFa(P,0.005), IFNg
(P,0.005), IL-6 (P,0.005) and IL-10 (P,0.01) than that of pregnant controls, irrespective of gestational age. Significantly (P,0.05)
higher TNF-R1 levels at 6– 9 weeks, and significantly higher TNFa/IL-6 (P,0.001) and significantly lower TNFa/IL-10 (P,0.001) and
IFNg/IL-10 (P,0.001) ratios at 10– 14 weeks, were also found in euploid miscarriage cases compared with pregnant controls.
TNFa/IL-10 ratio in plasma was significantly (P,0.05) lower in miscarriages with an abnormal karyotype than those with normal karyotype.
Normal pregnant women had a significantly higher plasma level of IFNg(P,0.01) and IFNg/IL-10 ratio (P,0.005), a significantly
(P,0.005) lower TNF-R1 level, and a significant (P,0.05) increase in stimulated TNFain monocytes, compared with NPW.
conclusions: Our data confirm that there is an inflammatory reaction in normal pregnancy compared with the non-pregnant state,
which may be disrupted during miscarriage.
Key words: cytokines / miscarriage / inflammation
Introduction
An inflammatory microenvironment is required for successful implant-
ation and tissue remodeling in the first trimester of pregnancy (Challis
et al., 2009). This inflammatory reaction is characterized by an
up-regulation of cytokines, chemokines and their receptors. An exces-
sive inflammatory reaction has been associated with recurrent miscar-
riage or other pregnancy complications such as pre-eclampsia or
premature labour (Chaouat et al., 2004).
Miscarriage was initially believed to happen exclusively due to
abnormal villous development, as described in anatomical and histo-
logical studies in earlier studies (Bouie et al., 1976). However,
recent evidence is now proving that early pregnancy failure is a
disorder of the placenta, and that the villous changes described in
previous studies are not the cause, but rather the consequence of mis-
carriage (Jauniaux and Burton, 2005). Defective placentation is mainly
characterized by a thinner and fragmented trophoblast shell, reduced
cytotrophoblast invasion of the endometrium and incomplete plugging
of the lumen at the tips of the spiral arteries (Hustin, 1990). These
anatomic changes have been found in about two-thirds of early preg-
nancy failures (Hustin, 1990;Jauniaux et al., 1994). This leads to the
absence of physiological changes in most spiral arteries, with prema-
ture onset of the maternal circulation throughout the entire placenta
(Jauniaux et al., 2006). In vivo and in vitro investigations have shown
that during the first trimester of normal human pregnancy, the pla-
centa limits, rather than facilitates, oxygen supply to the fetus during
&The Author 2011. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: journals.permissions@oup.com
Human Reproduction, Vol.27, No.2 pp. 349– 357, 2012
Advanced Access publication on November 29, 2011 doi:10.1093/humrep/der402
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
organogenesis (Jauniaux et al., 2003a). There is a very well-controlled
phenomenon that has to provide a delicate balance between the fetal
metabolic needs and its placenta, and the potential danger of oxidative
stress. This physiological hypoxia of the first trimester gestational sac
may protect the fetus against the deleterious and teratogenic effects of
highly reactive oxygen free radicals (Jauniaux et al., 2006).
Consequently, in early pregnancy failure, the excessive entry of
maternal blood into the intervillous space has two main effects on
the fetus and the placenta. The first effect is a direct mechanical
effect on the villous tissue which becomes enmeshed inside large inter-
villous blood thrombi. The second effect is widespread and indirect
oxygen-mediated trophoblastic damage and increased apoptosis
(Kokawa et al., 1998;Hempstock et al., 2003;Jauniaux et al.,
2003b). This mechanism possibly occurs in all miscarriages, with
the time at which it occurs in the first trimester depending on the
aetiology (Jauniaux et al., 2006). This is in a way analogous to the
ischaemia-reperfusion injury that occurs in pre-eclampsia (Chappell
et al., 2002).
It has been proposed that there are primary and secondary causes
of oxidative stress leading to early pregnancy failure (Hempstock et al.,
2003). Primary causes are mainly chromosomal abnormalities which
are found in at least 50% of miscarriages and are usually associated
with abnormal trophoblast invasion of the uterine deciduas (Hustin,
1990;Greenwold et al., 2003). Secondary causes may involve mater-
nal leukocytes and other immune factors including cytokines, such as
tumour necrosis factor alpha (TNFa). However, so far, their role in
the trophoblast-decidual interaction in normal and abnormal first
trimester remains unclear (Jauniaux et al., 2006). Studies have
shown that the circulating cytokine levels and the decidual cytokine
profile in the decidua are different in women experiencing recurrent
miscarriages (Jenkins et al., 2000;Makhseed et al., 2000;von Wolff
et al., 2000;Baxter et al., 2001), however little is known about spon-
taneous miscarriages. Major fetal abnormalities leading to fetal demise
in the first trimester could also lead to secondary placental dysfunction
as the placental development becomes increasingly dependent on
fetal synthesis towards the end of the first trimester (Shurtz-Swirski
et al., 1991).
Previous studies in humans (Raghupathy et al., 2000;Rezaei and
Dabbagh, 2002;Raghupathy and Kalinka, 2008)and mice (Doi et al.,
1999;Clark and Croitoru, 2001;Joachim et al., 2003) provide evi-
dence for an increase in Th-1/Th-2 ratio, with a shift towards Th-1
in recurrent miscarriage. In contrast, peripheral blood mononuclear
cells (PBMCs) have been shown to secrete higher levels of Th-1 cyto-
kines and lower levels of Th-2 cytokines, in first trimester pregnant
women with normal outcome compared with women with spontan-
eous miscarriage (Zenclussen et al., 2002) or a history of recurrent
miscarriage who subsequently miscarried (Bates et al., 2002). In
general, Th-1 cytokines which have been measured include interleukin
(IL)-2, TNFaand b, interferon gamma (IFNg) and Th-2 cytokines
include IL-4, IL-5, IL-6 and IL-10. Although similar methods have
been used to estimate circulating and intracellular cytokine levels
(mainly ELISA and flow cytometry of cultured PBMCs or whole
blood) there are conflicting reports in the literature about the state
of systemic inflammation in early pregnancy loss, warranting further
investigation with karyotyping of the products of conception.
Therefore, the objective of this study is to investigate pro- and anti-
inflammatory cytokines and soluble receptor concentrations in
maternal plasma and PBMCs in miscarriage patients of normal karyo-
type in the first trimester, and to evaluate the systemic inflammatory
response in first trimester spontaneous early pregnancy loss compared
with normal pregnancy.
Materials and Methods
Plasma and whole blood samples
Pregnant women attending the Early Pregnancy Unit and diagnosed with a
missed miscarriage on scan were recruited as cases. A missed miscarriage
was diagnosed by ultrasound on the basis of absent fetal heart activity after
5 weeks of gestation or in the presence of an empty gestational sac, before
the expulsion of the fetus or placental tissues. Control pregnant women
were asymptomatic normal pregnant women with an ongoing pregnancy
who came to Antenatal Clinic for booking or to the Early Pregnancy
Unit for a reassurance scan, and were followed up prospectively until
term delivery. Non-pregnant women (NPW) in the luteal phase of their
menstrual cycle were recruited among staff at UCL as a control group
for comparison. All blood samples were collected as part of an ongoing
prospective study, for which the participants gave informed consent. In
the case of women with missed miscarriage, the blood was sampled im-
mediately upon diagnosis. The study was approved by the UCLH Research
Ethics Committee.
All pregnant cases were ,14 weeks’ gestation as calculated from the
first day of the last menstrual period and confirmed by ultrasound meas-
urement of crown-rump length. Inclusion criteria were healthy non-
smoking women with a spontaneous singleton pregnancy, and not on
any medication until the time of blood sampling. In the case of missed
miscarriages, there was an ultrasound confirmation of an absent fetal
heartbeat. Exclusion criteria were patients with a history of recurrent
miscarriage, the presence of a hydatiform mole or a multiple pregnancy,
any congenital uterine anomaly, cervical incompetence, large uterine
fibroids, a known thrombophilia or any other medical condition needing
chronic drug treatment.
The products of conception obtained from patients with missed
miscarriage who underwent elective removal of products of conception,
were karyotyped at a commercial laboratory (TDL, London) using
standard culturing, suspension harvest and G-band analysis methodology
as previously described (Greenwold and Jauniaux, 2002).
Venous blood (10 ml) was collected by sterile venepuncture and 1.5 ml
was aspirated aseptically for whole blood cell culture and analysis, while
the rest was centrifuged within 2 h of collection, and the plasma
supernatant was stored at 2208C until assayed.
Study groups
In total, there were 106 successfully karyotyped placentae. There were 50
placentae with male karyotype, of which 29 were euploid and 21 were tri-
somies (majority Trisomies 16 and 22, but also Trisomies 7, 8, 13, 20 and
21 were found in individual cases). There were 51 placentae with female
karyotype, of which 25 were euploid and 26 were aneuploid (7 monoso-
mies XO and 19 trisomies). There were also five triploidies.
For the purpose of this study, in order to avoid the risk of maternal con-
tamination, the study group included 29 patients with missed miscarriage
of normal male karyotype (14 at 6 –9 and 15 at 10 –14 weeks), 21 patients
with missed miscarriage of aneuploid (trisomic) male karyotype (9 at 6 –9
and 12 at 10–14 weeks) and the control groups included 35 pregnant
women (16 at 6–9 and 19 at 10 –14 weeks) and 31 NPW. All patients
were matched for age, parity, ethnicity and BMI.
350 Calleja-Agius et al.
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
Multiplex bead arrays
TNF receptors 1 and 2, TNFa, IFNg, IL-6 and IL-10 were assayed in
maternal plasma using the BD
TM
Cytometric Bead Array Human Soluble
Protein Flex Sets (BD Biosciences, San Jose, Canada). Cytometric bead
array permitted multiplexed analysis of different cytokines and receptors
on smaller quantities of plasma, compared with conventional ELISA
systems. The experiment was set up according to the manufacturer’s
instructions and as previously published (Chen et al., 1999). Acquisition
of the sample data was performed using the BD FACSArray
TM
bioanalyzer
flow cytometer (BD Biosciences, CA, USA) and data were presented in
graphical and tabular formats using the FCAP Array
TM
Software. The
limits of detection for each assay were 0.7 pg/ml for TNFa, 1.8 pg/ml
for IFNg, 1.6 pg/ml for IL-6, 0.13 pg/ml for IL-10, 5.2 pg/ml for
TNF-R1 and 1.4 pg/ml for TNF-R2. The intra-assay coefficient of variation
was 10.2% for TNFa, 10.8% for IFNg, 6.4% for IL-10, 9.4% for IL-6, 2.6%
for TNF-R1 and 7.1% for TNF-R2. The inter-assay coefficient of variation
was 5% for TNFa, 5% for IFNg, 8% for IL-6, 11% for IL-10, 10.1% for
TNF-R1 and 5.6% for TNF-R2.
Whole blood flow cytometry for cellular
expression and inflammatory response
Whole blood flow cytometry performed in a subgroup of patients with
miscarriage with normal male karyotype (n¼17: n¼7 at 6 –9 weeks
and n¼10 at 10–14 weeks), miscarriages with aneuploid male karyotype
(n¼16: n¼6 at 6– 9 weeks and n¼10 at 10–14 weeks), pregnant con-
trols (n¼18: n¼8 at 6–9 weeks and n¼10 at 10 –14 weeks) and NPW
(n¼13) of the same cohort of patients. The patients in the whole blood
flow cytometry subgroup were matched for age, parity, ethnicity and BMI.
Optimal experimental conditions were selected based on initial valid-
ation experiments with a range of doses of (0.1–40 ng/ml) of lipopolysac-
charide (LPS) and incubation times (2–20 h). All experiments were carried
out as previously described by our group (Calleja-Agius et al., 2011) using
40 ng/ml of LPS as an inflammatory stimulus with an incubation period of
12 h. In brief, dual antibody labelling was carried out with monocyte (CD
14) specific and cytokine or receptor-specific mouse anti-human anti-
bodies (AbD Serotec, Oxford, UK). Using aseptic techniques, 1.5 ml of
whole blood from each individual patient was mixed with 1.5 ml of
Dulbecco phosphate-buffered saline without Ca
++
,Mg
++
and the final
solution divided into aliquots. Ten microlitres of monensin sodium solution
(a Golgi apparatus inhibitor used to stop cytokine release by the cell;
Sigma-Aldrich, St. Louis, USA) and 10 ml of LPS 40 ng/ml solution were
added to one aliquot (stimulated), while only 10 ml of monensin sodium
solution (Sigma-Aldrich) without LPS was added to the other aliquot.
Both aliquots were placed in a water bath at 378C, shaking at 20 revs/
min for 12 h incubation. Each aliquot was further divided depending on
the antibody combination, and Fix and Perm
w
Cell Permeabilization
Reagents (Caltag
TM
Invitrogen, UK) were added following the manufac-
turer’s instructions. Each sample was read by the flowcytometer Dako
CyAn Advanced Digital Processing, operated through the Summit 4.31
Software. The results were then saved in fetal calf serum 2.0 file format
for post-acquisition spectral compensation and data analysis. The cell
populations were classified as monocytes, granulocytes and lymphocytes
by their granularity and size, and further gating was done on CD-14
positive cells, which represent monocytes.
Statistical analysis
Data were log transformed to approximate normal distribution, and
two-way analysis of variance was carried out using category ( pregnant
controls versus miscarriage of normal karyotype) and gestational age
subgroups (6–9 and 10 –14 weeks). All patients were matched for age,
BMI, parity and ethnicity. Median and interquartile ranges were calculated
for each patient category within each gestational group. Comparison of
cytokine levels in plasma between patient categories in each gestational
group were tested for significance using an unpaired t-test of the log-
transformed data. This analysis was repeated for intracellular cytokine
levels in monocytes, both at baseline and at stimulated levels using
40 ng/ml LPS, including the % rise for each cytokine upon stimulation. In
this case, significance testing was carried out using Mann –Whitney
U-test, since the number of samples was limited. Descriptive statistics
and significance testing was carried out using SPSS v.17 for Windows
(SPSS Inc., Chicago, IL, USA). A P≤0.05 was considered statistically
significant.
Results
Demographics
There were no statistical differences between the groups regarding
maternal age, parity, BMI and ethnicity. There were no significant
differences between the mean gestation durations in each patient
subgroup within both studies.
Miscarriages of normal karyotype
compared with pregnant controls
at 6–9 weeks’ gestation
Differences in cytokine levels and ratios were standardized to the
comparator group, and expressed as percentages of the levels or
ratios in the comparator group.
Maternal plasma concentrations
Patients with miscarriage of normal karyotype had significantly higher
circulating levels of TNFa(337.1%, P¼0.005), TNF-R1 (34.1%,
P,0.05), IFNg(299.5%, P,0.005), IL-6 (195%, P,0.005) and
IL-10 (2584.6%, P¼0.01) in the 6 –9 gestational group compared
with gestation-matched normal pregnant women (Fig. 1). There was
no statistical difference in the cytokine ratios between the two
patient categories at this gestational stage.
Cellular expression
At 6–9 weeks, there was a significant (24.6%, P,0.05) increase in
unstimulated IL-6 levels in monocytes in miscarriages of normal
karyotype compared with pregnant controls. Otherwise, there was
no significant difference in any of the baseline or stimulated cytokine
levels or receptor levels.
Miscarriages of normal karyotype
compared with pregnant controls
at 10–14 weeks’ gestation
Maternal plasma concentrations
Patients with miscarriage of normal karyotype had significantly higher
circulating levels of TNFa(615.7%, P,0.001), IFNg(375.6%,
P,0.001), IL-6 (115%, P,0.001) and IL-10 (3330.8%, P,0.001)
in the 10 –14 weeks gestational groups compared with gestation-
matched normal pregnant women (Fig. 1). In this gestational group,
there was also a significantly higher ratio of TNFa/IL-6 (197.7%,
Cytokines in early pregnancy failure 351
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
P,0.001), and significantly lower TNFa/IL-10 (279.6%,
P¼0.001) and IFNg/IL-10 (286.1%, P,0.001) ratios, in miscar-
riages compared with normal pregnancy (Table I).
Cellular expression
At 10–14 weeks, there was a significant decrease (256.4%, P,0.05)
in the level of stimulated IFNgin miscarriages of normal karyotype
Figure 1 Median and interquartile range for all cytokines and
receptors (pg/ml) measured in plasma in both gestational subgroups,
in pregnant controls and miscarriages of normal karyotype. Signifi-
cance testing was carried out using unpaired T-test to compare preg-
nant controls and euploid miscarriages. At 6 – 9-week gestation:
n¼16 pregnant controls, n¼14 euploid miscarriages; and at 10 –
14-week gestation: n¼19 pregnant controls, n¼15 euploid miscar-
riages. Y-axis is cytokine/receptor (pg/ml).
Figure 1 Continued
352 Calleja-Agius et al.
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
compared with pregnant controls. There was no significant difference
in any of the other baseline or stimulated cytokine levels or receptor
levels.
Comparison of euploid with aneuploid
missed miscarriages
Maternal plasma concentrations
The levels of TNFa, TNF-R1, TNF-R2, IFNg, IL-6 and IL-10 were not
significantly different between miscarriages with normal and abnormal
karyotype, irrespective of gestational stage. TNFa/IL-10 ratio in the
plasma was significantly (P,0.05) lower in miscarriages with an
abnormal karyotype compared with those with normal karyotype in
both gestational groups (Fig. 2).
Cellular expression
There was no significant difference between intracellular levels of
TNFa, TNF-R1, TNF-R2, IFNg, IL-6 and IL-10 in the monocytes of
both groups.
Comparison of normal pregnant state
with luteal phase of the non-pregnant state
Maternal plasma concentrations
Normal pregnant women in the first trimester (6 –14 weeks) had
significantly higher (163%, P¼0.005) plasma levels of IFNg, and
significantly lower (250.5%, P¼0.001) TNF-R1 than NPW. There
was also a significantly higher (117%, P¼0.001) IFNg/IL-10 ratio in
normal pregnant controls than NPW. There was no significant
difference in TNFa, TNF-R2, IL-6 or IL-10 levels, or in ratios of
TNFa/IL-10, TNFa/IL-6 or IFNg/IL-6 (Fig. 3and Table II).
Cellular expression
In monocytes, there was a significant increase (61%; P,0.05) in
stimulated TNFain pregnant women compared with NPW. Upon
........................................................................................
Table I Median and interquartile range for all cytokine
ratios measured in plasma in the 10– 14 week
gestational subgroup, in pregnant controls (n519)
and miscarriages (n515).
Cytokine,
ratio
Pregnant
controls,
10–14 weeks
Miscarriage,
10–14 weeks
P-value
(unpaired
T-test),
comparison
between
controls
versus
miscarriages
at 10–14
weeks
TNFa/IL-6 0.44 (0.44, 0.54) 1.31 (0.58, 1.46) ,0.001
TNFa/IL-10 5.39 (5.39, 5.64) 1.10 (0.85, 1.31) 0.001
IFNg/IL-6 1.13 (1.13, 2.08) 2.48 (1.79, 3.56) 0.08
IFNg/IL-10 13.85 (13.85, 23.5) 1.92 (1.41, 3.54) ,0.001
There were no significant differences for the cytokine ratios in the 6–9 week
gestational subgroup.
Figure 2 Median and interquartile range for TNFa/IL-10 ratio
measured in plasma in both gestational subgroups, in euploid
(n¼29) and aneuploid (n¼21) male miscarriages. Significance
testing was carried out using Mann –Whitney. There was a significant-
ly (P,0.05) lower TNFa/IL-10 ratio in the plasma of miscarriages of
abnormal karyotype compared with those with normal karyotype.
Figure 3 Median and interquartile range for all cytokines and
receptors (pg/ml) measured in plasma in pregnant controls
between 6–14-week gestation (n¼35) and NPW (n¼31). Signifi-
cance testing was carried out using unpaired t-test. *P,0.05,
**P,0.01.
........................................................................................
Table II Median and interquartile range for all cytokine
ratios measured in plasma in pregnant controls between
6–14-week gestation (n535) and NPW (n531).
Cytokine
ratio
Pregnant controls
(6–14 weeks)
Non-pregnant
women
P-value
Median
(interquartile
range)
Median
(interquartile
range)
TNFa/IL10 3.48 (1.36, 5.61) 4.74 (3.04, 5.38) 0.09
TNFa/IL6 0.11 (0.10, 0.44) 0.14 (0.11, 0.44) 0.8
IFNg/IL10 32.00 (13.85, 98.46) 14.74 (2.24, 18.92) 0.001
IFNg/IL6 0.97 (1.06, 8.00) 0.44 (1.13, 2.05) 0.19
Significance testing was carried out using unpaired t-test.
NPW, non-pregnant women.
Cytokines in early pregnancy failure 353
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
stimulation with 40 ng/ml LPS, there was a median rise of 537.9%
from baseline in TNFain pregnant controls, when compared with a
median rise of 235.3% in NPW. Other differences were not significant
although IFNglevels decreased less in pregnant controls when
compared with NPW, upon LPS stimulation (0.3 versus 26.7%), and
there was only a median rise of 5.5% in IL-10 upon LPS stimulation
in samples from pregnant women, when compared with a median
of 97.7% in NPW (Table III).
Discussion
This study demonstrates that there is an inflammatory reaction in
normal pregnancy compared with the non-pregnant state, which
may be altered in spontaneous miscarriage. Our findings also show
that there are higher plasma and intracellular levels of some
pro-inflammatory cytokines in pregnant compared with NPW, con-
firming that normal pregnancy is an inflammatory state as described
previously (Redman and Sargent, 2010). In cases of spontaneous mis-
carriage with normal karyotype, plasma cytokine levels were generally
higher than in pregnant controls.
However, a limitation of this study concerns reliability of the
statistical significance of the results due to multiple testing in the
flow cytometry study, which is exacerbated by the relatively small
sample sizes. The low sample sizes are because of the difficulty in
getting chromosomally normal miscarriage samples.
In normal pregnancy, there is an increase in systemic inflammation
(Challis et al., 2009), which explains the higher plasma levels of
IFNgand IFNg/IL-10 in normal pregnant women compared with
NPW in circulation, and the higher rise in TNFa(and possibly
lower IL-10) upon stimulation of monocytes. This is in contrast
with the findings of another study carried out on cultured PBMCs
(Bates et al., 2002), where the levels of the pro-inflammatory cytokine,
IFNg, were decreased while IL-10 levels were higher in pregnant con-
trols compared with NPW. However, in our study, all NPW had their
blood samples taken in the luteal phase, unlike in the latter study.
Moreover, we have compared cytokine and soluble receptor levels
both in plasma and intracellularly in monocytes. In contrast to previous
investigations where experiments were done on cultured PBMCs
(Bates et al., 2002;Sacks et al., 2003), in this study we used a whole-
blood cell culture system which has been successfully used by other
groups (Colin et al., 2004;Zahran et al., 2006;van Nieuwenhoven
et al., 2008), together with a Golgi apparatus inhibitor used to stop
cytokine release by the cell, in order to closely reflect the in vivo
conditions. Whole blood analysis is more physiological, when
compared with studies involving cultured PBMCs, because in the
latter, the purification step itself can alter cytokines and other
markers of inflammation. The TNFa/IL-10 ratio was not significantly
different between the pregnant and NPW, however this may be
attributed to the low detectable levels of TNFa.
Both pro-inflammatory (TNFa, IFNg) and anti-inflammatory (IL-6,
IL-10) cytokines were significantly elevated in the plasma of patients
presenting with miscarriages of normal karyotype when compared
with pregnant controls, irrespective of gestation. Since the Th-1/
Th-2 ratio gives a better picture of the overall immune response,
rather than the individual cytokine levels, the cytokine ratios were cal-
culated. The significantly higher TNFa/IL-6 ratio in miscarriages of
normal karyotype compared with normal pregnancy is consistent
with a shift towards Th-1 type of immune response with an increase
in pro-inflammatory cytokines, which has already been shown previ-
ously by others investigating miscarriages sampled after diagnosis
(Makhseed et al., 1999,2000;Raghupathy et al., 2000). However, in
the late first trimester (10–14 weeks), the TNFa/IL-10 and IFNg/
IL-10 ratios were lower in euploid miscarriages compared with
normal pregnancy. This contradicts the hypothesis of a shift in the
.............................................................................................................................................................................................
Table III Median and interquartile range for all cytokines and receptors measured in monocytes, in pg/ml, in pregnant
controls (n518) and NPW (n513).
Cytokine/receptor Pregnant controls (6– 14 weeks) Non-pregnant women P-value
Median (interquartile range) Median (interquartile range)
TNFR-1 median 2.53 (1.63, 3.81) 1.29 (1.00, 3.67) 0.29
TNFR-2 median 15.57 (12.33, 20.04) 9.73 (9.39, 12.99) 0.12
TNFa(0 LPS) 1.53 (0.65, 5.73) 1.92 (1.20, 2.29) 0.69
TNFa(40 LPS) 8.50 (7.14, 20.48) 5.27 (3.85, 9.19) 0.03
% rise in TNFa537.91 (207.82, 1209.81) 235.36 (109.65, 491.71) 0.09
IFNg(0 LPS) 80.29 (69.15, 95.47) 88.21 (85.91, 94.89) 0.07
IFNg(40 LPS) 81.00 (72.62, 93.11) 64.13 (53.28, 89.26) 0.12
% rise in IFNg20.29 (21.44, 18.54) 226.69 (237.07, 21.91) 0.32
IL-10 (0 LPS) 1.57 (1.34, 3.01) 2.13 (1.54, 3.77) 0.60
IL-10 (40 LPS) 1.79 (0.97, 4.41) 3.35 (1.77, 8.77) 0.34
% rise in IL10 5.45 (230.49, 93.64) 97.71 (24.51, 258.61) 0.25
IL-6 (0 LPS) 1.94 (1.65, 3.65) 2.30 (1.60, 3.62) 0.68
IL-6 (40 LPS) 6.37 (3.64, 9.37) 5.10 (3.51, 6.06) 0.54
% rise in IL-6 97.58 (37.41, 224.85) 74.42 (220.67, 135.00) 0.94
Significance testing was carried out using unpaired t-test. Cytokines were measured at basal level and upon stimulation with 40 ng/ml LPS. The % rise from basal to stimulated was
calculated as [(stimulated-basal)/basal] multiplied by 100. TNF-R1 and TNF-R2 were measured at basal level only.
354 Calleja-Agius et al.
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
Th-1/Th-2 ratios. We hypothesize that there could be changes in
oxidative stress in the feto –placental unit happening at around 9
weeks gestation (Jauniaux et al., 2000;Jauniaux et al., 2001), which
could modulate the cytokine balance both in normal pregnancies
and in miscarriages. This can explain our findings that in the early
first trimester (6 –9 weeks), there were no changes in the Th-1/
Th-2 ratios, even though there was a rise in the plasma levels of
both pro- and anti-inflammatory cytokines in women with euploid
miscarriages compared with normal pregnancy. With the premature
entry of maternal oxygenated blood in late first trimester miscarriages
leading to an increase in free radicals in the placenta (Jauniaux et al.,
2000), there may be a compensatory rise in the anti-inflammatory
cytokine IL-10 in the maternal circulation, possibly to dampen the
rise in pro-inflammatory cytokines in the placenta. Our preliminary
results of measuring the placental cytokine levels in normal first trimes-
ter and miscarriages show increased TNFaand decreased receptors
in villous tissue of miscarriages.
Our study contributes to this literature by confirming that the
miscarriages studied were of normal karyotype. Other studies have
recruited women with recurrent miscarriage, but did not specify the
karyotype of the products of conception of the index miscarriage.
Karyotype of the fetus which is eventually miscarried has been
shown to have an effect on the cytokine levels present in the maternal
circulation of recurrent miscarriages (Yamada et al., 2004). However,
in this study, we investigated this difference in spontaneous miscar-
riages. A comparison was done between male euploid and aneuploid
miscarriages. In this case, the TNFa/IL-10 ratio in plasma was signifi-
cantly lower in miscarriages with an abnormal karyotype compared
with those with normal karyotype. However since the TNFa/IL-10
ratio was already lower in euploid miscarriages compared with
normal pregnancy, it would appear that the difference in immune func-
tion in miscarriage is actually exaggerated in aneuploid miscarriages.
In order to avoid the risk of maternal contamination, only male
euploid miscarried placentae were included in this study. However,
one limitation of including only miscarriages of normal male karyotype
is the introduction of a bias whereby the maternal immune response is
altered by the H-Y antigen present in male fetuses. It has been shown
that in patients with recurrent miscarriages, male fetuses are at an
increased risk of being miscarried due to the increased maternal
immune response against H-Y antigens. Anti-H-Y immunity may
partly explain the increased inflammatory response found during mis-
carriage of male fetuses of normal karyotype (Christiansen et al., 2010;
Nielsen et al., 2010a;Nielsen et al., 2010b). In view of this, a subanay-
sis was carried out to compare cytokine levels between women
miscarrying a euploid male fetus, and those controls who eventually
delivered a healthy male baby at term. There was a trend for higher
circulating levels for the pro-inflammatory cytokines (TNFaand
IFNg) in the miscarriage group, but it did not achieve statistical
significance, probably due to the small sample size.
In contrast to our findings in the maternal plasma, the cytokine
levels in monocytes of euploid miscarriages showed significantly
lower levels of the pro-inflammatory cytokine, IFNgupon LPS stimu-
lation, and higher basal IL-6 levels, when compared with pregnant
controls. In a smaller study with n¼5, other investigators found no
shift in Th-1/Th-2 cytokine ratio in PBMCs of pregnant women
prior to ending up having a spontaneous miscarriage (Marzi et al.,
1996). However, similar to our findings, others have shown that
there is a rise in anti-inflammatory, rather than pro-inflammatory cyto-
kines in PBMCs of pregnant women who subsequently miscarry (Bates
et al., 2002). Although, in contrast to our study the patients recruited
in the latter study were recurrent and not spontaneous miscarriagers,
and the karyotype of the products of conception was also unknown,
the trend in the cytokine shift intracellularly in monocytes is similar.
This is inspite of the fact that in our study, the blood was sampled
upon confirmation of the missed miscarriage, while in the latter
study, the blood was sampled prior to the actual miscarriage. This
may suggest that the changes in cytokine levels may be a cause
rather than a consequence of the actual miscarriage.
Cytokines such as TNFado not exist in a stored form, but are
synthesized upon activation of cells derived from the monocyte/
macrophage lineage (Aggarwal et al., 1985b). LPS stimulation could
lead up to a 100-fold increase in the level of cytokine secretion
(Beutler et al., 1986). The production of IFNgin decidual mono-
nuclear cells has recently been shown to be up-regulated by IL-2
and IL-12 by increasing the susceptibility of these monocytes to LPS,
with TNFabeing independent of such a mechanism (Negishi et al.,
2011). In our study, in pregnant women, the rise in pro-inflammatory
TNFalevels upon LPS stimulation was higher compared with non-
pregnant controls, while the increase in stimulated IFNgwas higher
than in euploid miscarriages. IFNgmay play a beneficial role in early
pregnancy. Uterine natural killer cells have been shown to depend
on IFNgduring normal trophoblast invasion in mice (Ashkar and
Croy, 2001), and this cytokine has been shown to be essential in
endometrial angiogenesis in pigs (Tayade et al., 2007).
TNFareceptors, TNF-R1 and TNF-R2, are shed from the cell
surface and exist as soluble receptors in circulation neutralizing the
effect of TNFa(Aggarwal et al., 1985a). These receptor levels have
been shown to be decreased in recurrent miscarriage (Chernyshov
et al., 2005). Early in the first trimester, TNF-R1 levels were higher
in the plasma of miscarriages with normal karyotype compared with
pregnant controls, who in turn had lower levels than NPW. The
high TNF-R1 levels may explain the overall low levels of TNFa
present in the plasma in our study since the assay measures
free TNFa.
The classification of IL-6 into Th-1 or Th-2 cytokine remains
controversial (Chaouat et al., 2007). In the present study, we have
analyzed our data classifying IL-6 as a Th-2 cytokine, as there is litera-
ture showing that IL-6 has anti-inflammatory properties in pregnancy
as explained above. We found that TNFa/IL-6 ratios were significantly
higher in miscarriages than in pregnant controls, supporting IL-6 as a
Th2 cytokine. However, in studies on premature labour (Genc and
Ford, 2010;Wei et al., 2010) and pre-eclampsia (Vural et al., 2010),
IL-6 has been classified as a pro-inflammatory cytokine. IL-6 has
been detected in coelomic and amniotic fluid of early human preg-
nancy (Jauniaux et al., 1996) and has been shown to be involved in
the promotion of Th-2 differentiation and inhibition of Th-1 polariza-
tion (Dubinsky et al., 2008). During the first trimester, IL-6 could be
involved in the generation of new vessels in the placental villous
tissue, and tissue remodeling associated with placentation, thus
making it a Th-2 cytokine at this gestation (Jauniaux et al., 1996).
Our group has further shown, using the same methodology as in
this study, that women with threatened miscarriage who end up mis-
carrying have a shift towards a Th-1 type of immune response in their
circulation, when compared with women with threatened miscarriage
Cytokines in early pregnancy failure 355
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
who have a live birth with no adverse pregnancy outcome
(Calleja-Agius et al., 2011). This further confirms that the shift
towards a Th-1 type of immune response in plasma of miscarriages
of normal karyotype can contribute to prediction of early fetal demise.
This current study sheds new light on the role of cytokines in human
pregnancy and pregnancy failure.
Authors’ roles
J.C.A. wrote the first draft of the paper and carried out sample collec-
tion and the actual experiments, together with the statistical analysis.
E.R.J. contributed to the study design, interpretation of the data, and
to the writing of the manuscript. A.R.P. contributed to the experimen-
tal design and data analysis. S.M. contributed to the experimental
design, data analysis, presentation and interpretation of data.
Funding
This work was supported by UCLH special trustees fund for Early
Pregnancy Research.
Conflict of interest
All authors declare no conflict of interest.
References
Aggarwal BB, Eessalu TE, Hass PE. Characterization of receptors for
human tumour necrosis factor and their regulation by gamma-
interferon. Nature 1985a;6047:665– 667.
Aggarwal BB, Kohr WJ, Hass PE, Moffat B, Spencer SA, Henzel WJ,
Bringman TS, Nedwin GE, Goeddel DV, Harkins RN. Human tumor
necrosis factor. Production, purification, and characterization. J Biol
Chem 1985b;4:2345–2354.
Ashkar AA, Croy BA. Functions of uterine natural killer cells are mediated
by interferon gamma production during murine pregnancy. Semin
Immunol 2001;4:235–241.
Bates MD, Quenby S, Takakuwa K, Johnson PM, Vince GS. Aberrant
cytokine production by peripheral blood mononuclear cells in
recurrent pregnancy loss? Hum Reprod 2002;9:2439–2444.
Baxter N, Sumiya M, Cheng S, Erlich H, Regan L, Simons A,
Summerfield JA. Recurrent miscarriage and variant alleles of mannose
binding lectin, tumour necrosis factor and lymphotoxin alpha genes.
Clin Exp Immunol 2001;3:529–534.
Beutler B, Krochin N, Milsark IW, Luedke C, Cerami A. Control of
cachectin (tumor necrosis factor) synthesis: mechanisms of endotoxin
resistance. Science 1986;4753:977– 980.
Bouie J, Philippe E, Giroud A, Boue A. Phenotypic expression of lethal
chromosomal anomalies in human abortuses. Teratology 1976;1:3– 19.
Calleja-Agius J, Muttukrishna S, Pizzey AR, Jauniaux E. Pro- and
antiinflammatory cytokines in threatened miscarriages. Am J Obstet
Gynecol 2011.
Challis JR, Lockwood CJ, Myatt L, Norman JE, Strauss JF III, Petraglia F.
Inflammation and pregnancy. Reprod Sci 2009;2:206– 215.
Chaouat G, Ledee-Bataille N, Dubanchet S, Zourbas S, Sandra O, Martal J.
TH1/TH2 paradigm in pregnancy: paradigm lost? Cytokines in
pregnancy/early abortion: reexamining the TH1/TH2 paradigm. Int
Arch Allergy Immunol 2004;2:93 –119.
Chaouat G, Dubanchet S, Ledee N. Cytokines: important for
implantation? J Assist Reprod Genet 2007;11:491–505.
Chappell LC, Seed PT, Briley A, Kelly FJ, Hunt BJ, Charnock-Jones DS,
Mallet AI, Poston L. A longitudinal study of biochemical variables
in women at risk of preeclampsia. Am J Obstet Gynecol 2002;
1:127–136.
Chen R, Lowe L, Wilson JD, Crowther E, Tzeggai K, Bishop JE, Varro R.
Simultaneous quantification of six human cytokines in a single sample
using microparticle-based flow cytometric technology. Clin Chem
1999;9:1693–1694.
Chernyshov VP, Vodyanik MA, Pisareva SP. Lack of soluble TNF-receptors
in women with recurrent spontaneous abortion and possibility for its
correction. Am J Reprod Immunol 2005;5:284–291.
Christiansen OB, Steffensen R, Nielsen HS. The impact of anti-HY
responses on outcome in current and subsequent pregnancies of
patients with recurrent pregnancy losses. J Reprod Immunol 2010;
1:9–14.
Clark DA, Croitoru K. TH1/TH2,3 imbalance due to cytokine-producing
NK, gammadelta T and NK-gammadelta T cells in murine pregnancy
decidua in success or failure of pregnancy. Am J Reprod Immunol 2001;
5:257–265.
Colin IM, Isaac J, Dupret P, Ledant T, D’Hautcourt JL. Functional
lymphocyte subset assessment of the Th1/Th2 profile in patients with
autoimmune thyroiditis by flowcytometric analysis of peripheral
lymphocytes. J Biol Regul Homeost Agents 2004;1:72–76.
Doi TS, Marino MW, Takahashi T, Yoshida T, Sakakura T, Old LJ,
Obata Y. Absence of tumor necrosis factor rescues RelA-deficient
mice from embryonic lethality. Proc Natl Acad Sci USA 1999;
6:2994–2999.
Dubinsky V, Junovich G, Gentile T, Gutierrez G. IL-6 as a regulatory factor
of the humoral response during pregnancy. Am J Reprod Immunol 2008;
3:197–203.
Genc MR, Ford CE. The clinical use of inflammatory markers during
pregnancy. Curr Opin Obstet Gynecol 2010;2:116– 121.
Greenwold N, Jauniaux E. Collection of villous tissue under ultrasound
guidance to improve the cytogenetic study of early pregnancy failure.
Hum Reprod 2002;2:452–456.
Greenwold N, Jauniaux E, Gulbis B, Hempstock J, Gervy C, Burton GJ.
Relationship among maternal serum endocrinology, placental
karyotype, and intervillous circulation in early pregnancy failure. Fertil
Steril 2003;6:1373–1379.
Hempstock J, Jauniaux E, Greenwold N, Burton GJ. The contribution of
placental oxidative stress to early pregnancy failure. Hum Pathol 2003;
12:1265–1275.
Hustin J. Histological study of the materno-embryonic interface in
spontaneous abortion. Placenta 1990;11:477–486.
Jauniaux E, Burton GJ. Pathophysiology of histological changes in early
pregnancy loss. Placenta 2005;2–3:114– 123.
Jauniaux E, Zaidi J, Jurkovic D, Campbell S, Hustin J. Comparison of colour
Doppler features and pathological findings in complicated early
pregnancy. Hum Reprod 1994;12:2432– 2437.
Jauniaux E, Gulbis B, Schandene L, Collette J, Hustin J. Distribution of
interleukin-6 in maternal and embryonic tissues during the first
trimester. Mol Hum Reprod 1996;4:239–243.
Jauniaux E, Watson AL, Hempstock J, Bao YP, Skepper JN, Burton GJ.
Onset of maternal arterial blood flow and placental oxidative stress.
A possible factor in human early pregnancy failure. Am J Pathol 2000;
6:2111–2122.
Jauniaux E, Watson A, Burton G. Evaluation of respiratory gases and
acid-base gradients in human fetal fluids and uteroplacental tissue
between 7 and 16 weeks’ gestation. Am J Obstet Gynecol 2001;
5:998–1003.
356 Calleja-Agius et al.
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from
Jauniaux E, Gulbis B, Burton GJ. The human first trimester gestational sac
limits rather than facilitates oxygen transfer to the foetus–a review.
Placenta 2003a;24(Suppl A):S86–S93.
Jauniaux E, Hempstock J, Greenwold N, Burton GJ. Trophoblastic
oxidative stress in relation to temporal and regional differences in
maternal placental blood flow in normal and abnormal early
pregnancies. Am J Pathol 2003b;1:115–125.
Jauniaux E, Poston L, Burton GJ. Placental-related diseases of pregnancy:
Involvement of oxidative stress and implications in human evolution.
Hum Reprod Update 2006;6:747–755.
Jenkins C, Wilson R, Roberts J, Miller H, McKillop JH, Walker JJ.
Antioxidants: their role in pregnancy and miscarriage. Antioxid Redox
Signal 2000;3:623–628.
Joachim R, Zenclussen AC, Polgar B, Douglas AJ, Fest S, Knackstedt M,
Klapp BF, Arck PC. The progesterone derivative dydrogesterone
abrogates murine stress-triggered abortion by inducing a Th2 biased
local immune response. Steroids 2003;10 –13:931 – 940.
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;1:21– 26.
Makhseed M, Raghupathy R, Azizieh F, Al Azemi MM, Hassan NA,
Bandar A. Mitogen-induced cytokine responses of maternal peripheral
blood lymphocytes indicate a differential Th-type bias in normal
pregnancy and pregnancy failure. Am J Reprod Immunol 1999;5:273–281.
Makhseed M, Raghupathy R, Azizieh F, Farhat R, Hassan N, Bandar A.
Circulating cytokines and CD30 in normal human pregnancy and
recurrent spontaneous abortions. Hum Reprod 2000;9:2011– 2017.
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 Immunol 1996;
1:127–133.
Negishi M, Izumi Y, Aleemuzzaman S, Inaba N, Hayakawa S.
Lipopolysaccharide (LPS)-induced Interferon (IFN)-gamma Production
by Decidual Mononuclear Cells (DMNC) is Interleukin (IL)-2 and
IL-12 Dependent. Am J Reprod Immunol 2011;65:20 –27.
Nielsen HS, Steffensen R, Lund M, Egestad L, Mortensen LH,
Andersen AM, Lidegaard O, Christiansen OB. Frequency and
impact of obstetric complications prior and subsequent to
unexplained secondary recurrent miscarriage. Hum Reprod 2010a;6:
1543–1552.
Nielsen HS, Wu F, Aghai Z, Steffensen R, van Halteren AG, Spierings E,
Christiansen OB, Miklos D, Goulmy E. H-Y antibody titers are
increased in unexplained secondary recurrent miscarriage patients and
associated with low male : female ratio in subsequent live births. Hum
Reprod 2010b;11:2745– 2752.
Raghupathy R, Kalinka J. Cytokine imbalance in pregnancy complications
and its modulation. Front Biosci 2008;13:985 –994.
Raghupathy R, Makhseed M, Azizieh F, Omu A, Gupta M, Farhat R.
Cytokine production by maternal lymphocytes during normal human
pregnancy and in unexplained recurrent spontaneous abortion. Hum
Reprod 2000;3:713– 718.
Redman CW, Sargent IL. Immunology of pre-eclampsia. Am J Reprod
Immunol 2010;6:534–543.
Rezaei A, Dabbagh A. T-helper (1) cytokines increase during early
pregnancy in women with a history of recurrent spontaneous
abortion. Med Sci Monit 2002;8:CR607–CR610.
Sacks GP, Redman CW, Sargent IL. Monocytes are primed to produce the
Th1 type cytokine IL-12 in normal human pregnancy: an intracellular
flow cytometric analysis of peripheral blood mononuclear cells. Clin
Exp Immunol 2003;3:490–497.
Shurtz-Swirski R, Simon RJ, Cohen Y, Barnea ER. Human embryo
modulates placental function in the first trimester; effects of neural
tissues upon chorionic gonadotropin and progesterone secretion.
Placenta 1991;5:521–531.
Tayade C, Fang Y, Croy BA. A review of gene expression in porcine
endometrial lymphocytes, endothelium and trophoblast during
pregnancy success and failure. J Reprod Dev 2007;3:455 – 463.
van Nieuwenhoven AL, Moes H, Heineman MJ, Santema J, Faas MM.
Cytokine production by monocytes, NK cells, and lymphocytes is
different in preeclamptic patients as compared with normal pregnant
women. Hypertens Pregnancy 2008;3:207–224.
von Wolff M, Thaler CJ, Strowitzki T, Broome J, Stolz W, Tabibzadeh S.
Regulated expression of cytokines in human endometrium throughout
the menstrual cycle: dysregulation in habitual abortion. Mol Hum
Reprod 2000;7:627– 634.
Vural P, Degirmencioglu S, Saral NY, Demirkan A, Akgul C, Yildirim G,
Issever H, Eroglu H. Tumor necrosis factor alpha, interleukin-6 and
interleukin-10 polymorphisms in preeclampsia. J Obstet Gynaecol Res
2010;1:64–71.
Wei SQ, Fraser W, Luo ZC. Inflammatory cytokines and spontaneous
preterm birth in asymptomatic women: a systematic review. Obstet
Gynecol 2010;2:393–401.
Yamada H, Morikawa M, Furuta I, Kato EH, Shimada S, Sata F, Kishi R,
Minakami H. Circulating cytokines during early pregnancy in women
with recurrent spontaneous abortion: decreased TNF-alpha levels in
abortion with normal chromosome karyotype. Hokkaido Igaku Zasshi
2004;3:237–241.
Zahran WA, Ghonaim MM, Koura BA, El-Banna H, Ali SM, El-Sheikh N. A
study of the role of IL-12 in pulmonary tuberculosis using the whole
blood flowcytometry technique. Egypt J Immunol 2006;1:53–65.
Zenclussen AC, Fest S, Busse P, Joachim R, Klapp BF, Arck PC.
Questioning the Th1/Th2 paradigm in reproduction: peripheral levels
of IL-12 are down-regulated in miscarriage patients. Am J Reprod
Immunol 2002;4:245–251.
Cytokines in early pregnancy failure 357
by guest on February 20, 2013http://humrep.oxfordjournals.org/Downloaded from