| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
. A model for hematogenous initiation of placental inflammations
* Department of Pediatrics, University of California, Los Angeles School of Medicine; and
Department of Medical Microbiology and Immunology and the University of Alberta Perinatal Research Centre, Edmonton, Canada
Correspondence: Larry J. Guilbert, Department of Medical Microbiology and Immunology and the University of Alberta Perinatal Research Centre, 6-25 HMRC, University of Alberta, Edmonton, Canada T6G 2S2. E-mail: Larry.Guilbert{at}ualberta.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(IFN-)-treated
syncytiotrophoblasts via monocyte LFA-1 for >48 h, during which time
the monocytes induce trophoblast apoptosis and subsequent damage of the
trophoblast layer. Optimal monocyte-mediated syncytiotrophoblast death
requires both lipopolysaccharide (LPS) and IFN- and is inhibited by
either anti-tumor necrosis factor (TNF) antibody or epidermal growth
factor. Syncytiotrophoblast damage is largely limited to culture
surfaces in the vicinity of bound monocytes. These results show that
activated maternal monocytes bound to the placental barrier can induce
focal damage mediated by the inflammatory cytokine TNF-
and suggest
a route for maternal leukocyte infiltration into the fetal
stroma.
Key Words: blood • apoptosis • reproductive immunology • adhesion molecules • cytokines
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Inflammation in the villous placenta (placental villitis) is associated with repeated spontaneous abortions [2 , 3 ] and with pregnancy-associated disorders such as fetal infections [1 , 4 , 5 ], pre-eclampsia [6 , 7 ], and fetal growth restrictions (IUGR [8 ]). The exact mechanisms of initiation of placental villitis are unknown but loss of the protective trophoblast is a common feature [1 ]. In particular, whether the loss of the ST is a cause or an effect of villitis is not known.
Epithelial/endothelial cell loss is associated with apoptosis, which in
turn may be required for subsequent healing [9
]. Villous
trophoblasts normally undergo apoptosis, primarily in the ST layer
[10
]. The frequency of ST apoptosis increases in
pathological situations [11
, 12
] and with
gestational age [10
]. Although it has been suggested
that ST apoptosis initiates in CT during or before fusion
[13
], what regulates either CT or ST apoptosis is
unknown. In vitro, tumor necrosis factor (TNF-)
[14
], but not Fas ligand [15
], stimulates
CT apoptosis. TNF--stimulated CT apoptosis is increased by
interferon- (IFN-) and inhibited by epidermal growth factor (EGF)
[16
] and Bcl-2 expression [17
]. However,
the combination of TNF- and IFN- also stimulate ST apoptosis
[16
]. These two cytokines also stimulate monocyte
adhesion to ST via up-regulation of the immune cell-associated adhesion
molecule ICAM-1 on cultured ST [18
].
These latter observations suggest that aberrant ST apoptosis (leading
to ST loss) is mediated by inflammatory cytokines released by bound
maternal monocytes. However, trophoblasts are thought not to be
susceptible to most forms of cell-mediated killing [19
]
and the villous ST in particular appears to be able to protect itself
against cytotoxic attack by a number of mechanisms
[20
21
22
]. We have therefore asked whether adhering
monocytes can damage trophoblasts with a primary cell co-culture model
of peripheral blood monocytes and villous ST, the latter derived from
CT purified from term placentas. We find that LPS-activated monocytes
remain attached to ST cultures in an IFN-, LFA-1-dependent manner
for >48 h, during which time they induce local, TNF--dependent
apoptosis accompanied by focal disruption of the trophoblast culture.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Trophoblast spot cultures
Spot cultures are discrete circles of cultured trophoblasts that
are separated from each other by unoccupied culture surface. An
approximately 1-cm-diameter circle of fibronectin was created by
pipetting a 40-µL droplet of 50 µg/mL human fibronectin
(Collaborative Biomedical, Becton Dickinson, Bedford, MA) in Iscove’s
modified Dulbecco’s medium (IMDM) onto a fresh tissue culture surface
then incubating the dish or coverslip for 1 h at room temperature
and washing once with IMDM. Thawed and washed cytotrophoblasts (8 x 104) in 40 µL of IMDM (GIBCO, Grand Island, NY)
supplemented with 10% fetal bovine serum (FBS; GIBCO) were then plated
onto the fibronectin-defined spots. After a 4-h incubation at 37°C
the nonadherent cells and debris were removed by gentle washing with
pre-warmed IMDM and the adherent cells refed as a combined culture with
1 mL 10% FBS/IMDM containing 10 ng/mL EGF (Prepro-Tech, Rocky
Hill, NJ) plus 50 µg/mL gentamicin. The medium was changed after 5
days, 200 U/mL IFN- (Collaborative Biomedical Products, Bedford, MA)
added, and the culture continued for 24 h to up-regulate ICAM-1 as
previously described [18
]. The spot cultures were again
washed, the surface around the spot rapidly dried, and a 40-µL bubble
containing medium alone or medium with LPS- (0.5 µg/mL, Sigma catalog
no. L8274) activated monocytes added (2 x 104 human
serum-treated or 1.2 x 105 untreated monocytes per
spot, protocols below). After 2 h of culture, unbound monocytes
were removed from the coverslips by nine washes in phosphate-buffered
saline (PBS) as previously described [18
] and from 35-mm
dishes by a single 2-mL wash with warm medium, which was added to the
center of the dish and removed simultaneously from the edge by
aspiration. The washed cultures were then refed with 2 mL of medium
containing IFN-, EGF, LPS, or anti-TNF- antibody (10 µg/mL,
clone 195, Boehringer-Mannheim, Laval, Quebec) such that all culture
spots shared a common supernatant. After 24 or 48 h of culture,
the medium was removed, the cultures washed with 2 mL of warm medium as
described above, and 1 mL of 2% (w/v) glutaraldehyde in IMDM (pH 7.4)
slowly flowed over the spots from the center of the tissue culture dish
to cross-link adhering monocytes (fixation time was 1 h at room
temperature).
Peripheral blood monocytes
Monocytes (90–95% pure by CD11c immunohistochemistry) were
prepared from peripheral blood leukocytes by selective adherence as
previously described [18
]. Where specified, monocytes
were incubated 15 h with human serum before activation with LPS.
All monocytes were activated by incubation with 0.5 µg/mL LPS for
2 h before harvesting for co-culture with trophoblasts. In some
experiments monocytes were pre-incubated 1 h at 4°C with mouse
anti-human LFA-1 antibody (anti-CD11a, clone 25.3, 10 µg/mL,
Immunotech, Westbrook, ME) or with control mouse IgG1 antibody (10
µg/mL, Dako, Dimension Labs, Mississauga, Ontario), then washed with
PBS before transfer to trophoblasts for the binding reaction described
above.
Immunohistochemistry
Immunohistochemistry of glutaraldehyde-fixed cells was carried
out as described by the manufacturer (Zymed Laboratories, South
Francisco, CA) and as described previously [18
].
Monocytes were identified with mouse anti-human CD45 (clone B-A11,
Biosource International, Camarillo, CA, 2.5 µg/mL). All preparations
of trophoblasts were assessed for mesenchymal cell contamination by
staining for vimentin as previously described [23
].
TNF- bioassay
The biologically active TNF content of culture supernatants was
determined by the L929-8 bioassay as previously described
[25
]. Half-maximal detection was 5 pg/mL and the lower
limit of detection 0.5 pg/mL.
Detection of DNA nicking by TUNEL
Trophoblasts were washed once with PBS, fixed with
acetone/methanol (1:1) for 10 min at room temperature, washed three
times with PBS, then subjected to TUNEL as previously described
[14
].
Statistical analysis
Where noted, differences between experimental and control data
was evaluated by Student’s t test with Microsoft Excel.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-treated ST, unbound monocytes removed, and adhering monocytes
cross-linked to the trophoblasts with glutaraldehyde 24 h later.
Monocyte adhesion persisted at least 48 h after initiation of
co-culture in the presence of LPS and IFN-, and that which persisted
at 24 h was 82.0 ± 1.8% LFA-1 dependent (as determined by
inhibition of binding by preincubation with LFA-1 antibody) in pooled
data from five coverslips in two independent experiments (data not
shown).
Adhering monocytes disrupts the underlying syncytiotrophoblast
culture
Co-incubation of LPS-activated monocytes with
IFN--treated ST for 24 h leads to the appearance of clusters of
pyknotic nuclei in the vicinity of adhering monocytes, suggesting
trophoblast apoptosis (data not shown). ST apoptosis was confirmed by
TUNEL histochemistry, which measures DNA nicking [26
].
ST damage in progress was assessed by the incidence of clusters (>3
nuclei) of apoptotic nuclei. Already completed disruption was assessed
by the number of holes (>50 µm) in the ST monolayer. Figure 1A
shows a typical cluster of apoptotic nuclei (as assessed by TUNEL
analysis) in the vicinity of a number of trophoblast-bound monocytes.
After loss of the apoptotic nuclei and associated cells, this structure
is typical of a hole in the ST culture. ST cultures without added
monocytes morphologically resembled the ST monolayers to the left of
the apoptotic cluster in Figure 1A and contained an average of 1.6 ± 0.1 holes and 7.0 ± 1.0 apoptotic clusters per 1-cm-diameter
culture spot while monocyte-ST co-cultures showed an approximate
fivefold increase in both apoptotic clusters and holes per culture
(Fig. 1B) . At both 24 (Fig. 1)
and 48 h (data not shown), the
number of apoptotic clusters and trophoblast culture holes were
proportional. In the remainder of this article, the number of holes in
the ST culture is the measure of trophoblast disruption. The number of
monocytes adhering to ST monolayers was always larger than the number
of holes forming and was consistent within a single experiment using
the same preparations of primary trophoblasts and monocytes. However,
both the number of monocytes adhering and the holes forming
individually varied with different preparations of trophoblasts and
monocytes (e.g., compare Figs. 1
2
3
).
|
|
|
are required for optimal monocyte-mediated
trophoblast disruption-treated ST, unbound monocytes
removed, and either medium (10% FBS in IMDM) or medium containing
IFN-, LPS, EGF, anti-TNF- antibody, or combinations of these were
added to the washed cultures. The numbers of monocytes and monolayer
holes per spot culture were assessed after 24 h of co-culture
(Fig. 2
). Both IFN- and LPS alone promoted monocyte adhesion and
trophoblast death. However, the combination of LPS and IFN- were
required for both maximal adhesion and hole formation.
TNF antibody and EGF inhibit monocyte-mediated trophoblast death
Because TNF- induces villous ST apoptosis [16
],
we next asked whether TNF- mediated the cytotoxic effects of
adhering monocytes. Neutralizing anti-TNF antibody approximately halved
the number of monocytes adhering but almost completely eliminated hole
formation (Fig. 2)
. TNF--induced ST apoptosis is inhibited by
co-incubation with EGF [16
]. We therefore asked whether
EGF would also inhibit monocyte-mediated ST apoptosis. We found that
EGF did not decrease monocyte adhesion stimulated by IFN- and LPS
but eliminated hole formation. Thus, monocyte-induced ST disruption is
mediated by TNF-, and monocytes can adhere to trophoblasts without
subsequent damage.
Trophoblast damage induced by adhering monocytes is a local effect
The above observation that monocyte-dependent ST disruption could
be inhibited by antibody suggests that TNF- is acting in the
supernatant and not at the surface of tightly bound monocytes. In the
experiments depicted in Figure 2
TNF- could be detected at low
levels in the culture supernatant (<200 pg/mL at 24 h). Although
a single dose of 200 pg/mL is too low to induce ST loss (even after
96 h of culture [16
]), low steady-state levels of
TNF- could possibly induce trophoblast disruption. To assess this
possibility, we compared ST hole formation in spot cultures containing
different numbers of adhering monocytes. Four ST culture spots were
established in a 35-mm culture dish and individual 40 µL bubble
cultures established with monocytes added to only two (see Fig. 3A
). Nonadherent monocytes were removed after a 2-h adhesion period,
fresh medium containing IFN- and LPS added to cover all the spots,
and the cultures continued for 48 h. During the 48-h culture, some
monocytes migrated from the spots to which monocytes were originally
seeded (monocyte spot, Fig. 3B
) to neighboring control spots; however,
80% remained on the original (monocyte) spots. When assessed for
disruption (by the number of holes per spot), control spots had 30
holes and monocyte spots had >250. Spots cultured in dishes to which
no monocytes were added to any spot (without monocytes, Fig. 3B
) had
<10 holes per spot. Thus, the low steady-state levels of TNF-
present in the bulk culture supernatant minimally induces trophoblast
damage. However, the observation that the extent of damage is roughly
proportional to the number of monocytes per 1-cm spot suggests elevated
steady-state accumulation in the vicinity of adhering monocytes.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
disrupts trophoblast monolayers by inducing apoptosis
[16
, 18
], we hypothesized that the focal
loss of the placental ST barrier is caused by release of TNF- from
adhering monocytes. We find that activated monocytes adhere to cultured
ST via trophoblast-expressed ICAM-1 and monocyte LFA-1 and remain
viable and bound long enough (>48 h) to induce apoptosis (clusters of
DNA-nicked nuclei visualized by TUNEL) and focal loss of syncytialized
trophoblasts (holes in the ST monolayer). Optimal maintenance of
monocyte adhesion to the trophoblast culture requires the continuous
presence of LPS and IFN-, thus appearing to require continuous
monocyte LFA-1 activation by LPS [27
] and trophoblast
ICAM-1 expression (by IFN- [18
]). Optimal
monocyte-induced disruption of the trophoblast monolayer requires
IFN- and LPS in the culture supernatant. Disruption, but not
monocyte adhesion, is completely inhibited by neutralizing anti-TNF
antibody. Thus, placental ST disruption can occur via apoptosis induced
by ICAM-1/LFA-1-bound monocytes that release TNF-.
Because monocytes activated by LPS release TNF-, monocyte-dependent
trophoblast damage could be due to accumulation of TNF- in the
culture supernatant during the 24- to 48-h co-culture. However,
monocyte-dependent ST disruption does not reflect the accumulation of
TNF- in the extended supernatant but is localized in that the extent
of damage (incidence of holes) in trophoblast culture spots reflected
the number of monocytes present on separate culture spots that shared a
common supernatant. Nonetheless, the observation that anti-TNF antibody
inhibits monocyte-induced trophoblast disruption suggests that the
interaction between TNF--producing monocyte and TNF-receptors on the
ST was not completely private (juxtacrine) and may involve short-range
diffusion of secreted TNF- (paracrine). The localized nature of
monocyte-mediated ST damage may partly explain why placental villitis
is a focal inflammation and suggest that it is initiated by
concentrated adhesion of activated maternal leukocytes to a relatively
small area on the protective trophoblast.
Placental ST apoptosis appears to be a normal component of placental
aging [10
] but its incidence is increased in disorders
such as IUGR [11
, 12
] or pre-eclampsia that
are associated with increased placental villitis [6
,
7
]. Pre-eclampsia has been suggested to be a disease of
Th1-type immunity [28
] that includes expression of the
quintessential TH1 regulatory cytokine IFN- and increased production
of TNF- [29
]. Placental ST expression of ICAM-1
becomes significant after only 6 h of exposure to IFN- or
TNF- [18
]; thus, the findings in this study suggest
that even transient expression of these cytokines in peripheral blood
during a cell-mediated (TH-1 regulated) immune response could lead to
monocyte adhesion followed by loss of the ST and initiation of
villitis. Such elevated levels of TNF- and IFN- are present in
the peripheral circulation for days after immune challenge by
infections or during tissue rejection responses
[30
31
32
].
We find that adhering monocytes optimally kill underlying ST when both
IFN- and LPS are in the culture medium. Monocytes must be activated
to both bind (activation of LFA-1) and produce TNF-. The requirement
for LPS may reflect a more general requirement for monocyte activation
because bacterial LPS is only one of several microbial products or
infections that activate macrophages [33
34
35
]. The
requirement for IFN- may reflect its stimulation of monocyte TNF-
production [36
], and/or its ability to enhance
TNF--stimulated ST apoptosis [16
]. Serum IFN- is
present for days after an appropriate immune challenge
[32
]. Monocytes circulating during cellular immune
responses may also produce both TNF- and IFN-
[37
38
39
40
] or, alternatively, both TNF- and IFN-
could derive from maternal T cells, which could also bind to
trophoblast ICAM-1 via activated LFA-1 [41
,
42
]. Finally, although stimulation conditions are
unknown, IFN- could come from the villous trophoblast itself
[43
] or from placental macrophages in the villous stroma
[44
]. Thus, it is not unlikely that the conditions for
cell-mediated placental trophoblast killing exist in the placenta and
that the same in vivo conditions (a maternal cell-mediated
immune response) that promotes trophoblast expression of ICAM-1 may
simultaneously activate leukocyte LFA-1, stimulate leukocyte TNF-
production, and place sufficient IFN- in maternal circulation to
allow optimal monocyte-induced trophoblast damage.
The above arguments also suggest that the placental trophoblast barrier
could be damaged whenever pregnant women mount a cell-mediated immune
response. Indeed, placental villitis is not uncommon [1
]
and its incidence and severity are increased in primary maternal
infections of cytomegalovirus (CMV) [4
] and
Toxoplasma gondii [5
], infections that elicit
maternal cellular immunity. However, elevated levels of IFN- and
TNF- in the placental intervillous space does not necessarily
translate into placental damage, as is observed during responses to
malaria parasites during multiparous pregnancies [45
].
There are a number of pregnancy-related mechanisms that may reduce the
potential for maternal cell-mediated immune damage to the villous
placenta. Antigen-specific IFN- and TNF- production by lymph node
and spleen cells that are induced by foot pad Leishmania
major infection is suppressed during pregnancy
[46
]. The placenta also provides a biological buffer
against even muted increases in circulating TNF- because it is a
major source of soluble TNF p55 receptors during pregnancy
[47
, 48
]. The placenta also produces agents
that suppress maternal cell-mediated immunity. Progesterone, produced
by villous ST [49
], promotes TH2 cytokine production and
reduces TH1 production [Xu, Guilbert, and Mosmann, unpublished
results; 50] and suppresses TNF- production from mononuclear
phagocytes [51, reviewed in 52]. Prostaglandin E2
(PGE2), also a product of the villous ST
[53
54
55
], selectively inhibits cytokine production from
TH1 cells [56
] and inhibits IL-12 production from
dendritic cells [57
, 58
], thereby
inhibiting TH1 development. Finally, we show in this article that the
ability of monocyte-released TNF- to kill underlying ST is
completely reversed by EGF, which together with TGF-, activate the
EGF receptor [reviewed in ref. 59
] and are produced by the villous
placenta [60
, 61
]. Thus, trophoblast damage
induced by maternal cell-mediated immunity may only occur if placental
protective mechanisms fail. The conditions leading to their failure are
under investigation.
|
ACKNOWLEDGEMENTS |
|---|
Received May 16, 2000; revised July 3, 2000; accepted July 5, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
/IFN--induced villous trophoblast apoptosis Placenta 20,423-430[Medline]
by placental blood mononuclear cells is associated with protection in an area with high transmission of malaria J. Infect. Dis. 179,1218-1225[Medline]
production in activated mouse macrophages by progesterone J. Immunol. 160,5098-5104This article has been cited by other articles:
|
|
 |
|
  W Zammiti, N Mtiraoui, H Khairi, J-C Gris, W Y Almawi, and T Mahjoub Associations between tumor necrosis factor-{alpha} and lymphotoxin-{alpha} polymorphisms and idiopathic recurrent miscarriage Reproduction, March 1, 2008; 135(3): 397 - 403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Blois, M. Tometten, J. Kandil, E. Hagen, B. F. Klapp, R. A. Margni, and P. C. Arck Intercellular Adhesion Molecule-1/LFA-1 Cross Talk Is a Proximate Mediator Capable of Disrupting Immune Integration and Tolerance Mechanism at the Feto-Maternal Interface in Murine Pregnancies J. Immunol., February 15, 2005; 174(4): 1820 - 1829. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Berman, G. Girardi, and J. E. Salmon TNF-{alpha} Is a Critical Effector and a Target for Therapy in Antiphospholipid Antibody-Induced Pregnancy Loss J. Immunol., January 1, 2005; 174(1): 485 - 490. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Chan, M.F. Stinski, and L.J. Guilbert Human Cytomegalovirus-Induced Upregulation of Intercellular Cell Adhesion Molecule-1 on Villous Syncytiotrophoblasts Biol Reprod, September 1, 2004; 71(3): 797 - 803. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Chan, D. G. Hemmings, A. D. Yurochko, and L. J. Guilbert Human Cytomegalovirus-Caused Damage to Placental Trophoblasts Mediated by Immediate-Early Gene-Induced Tumor Necrosis Factor-{alpha} Am. J. Pathol., October 1, 2002; 161(4): 1371 - 1381. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
