This information is current as
of June 17, 2017.
Differential Gene Expression in
Endometrium, Endometrial Lymphocytes,
and Trophoblasts during Successful and
Abortive Embryo Implantation
Chandrakant Tayade, Gordon P. Black, Yuan Fang and B.
Anne Croy
J Immunol 2006; 176:148-156; ;
doi: 10.4049/jimmunol.176.1.148
http://www.jimmunol.org/content/176/1/148
Subscription
Permissions
Email Alerts
This article cites 37 articles, 8 of which you can access for free at:
http://www.jimmunol.org/content/176/1/148.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2006 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
References
The Journal of Immunology
Differential Gene Expression in Endometrium, Endometrial
Lymphocytes, and Trophoblasts during Successful and
Abortive Embryo Implantation1
Chandrakant Tayade,2,3* Gordon P. Black,3* Yuan Fang,* and B. Anne Croy†
T
he maternal-fetal interface is a dynamic site in which fetally derived trophoblast cells interact with maternal tissue that includes immune cells. During early pregnancy in
many species, including pigs, maternal endometrium becomes enriched for cells of the innate immune system, particularly NK cells
(1– 4). Uterine NK (uNK)4 cells have been most extensively studied in humans, mice, and rats, species with invasive hemochorial
placentation accompanied by endometrial decidualization. In these
species, uNK cell recruitment is associated with decidual induction
and does not require a conceptus, although presence of a conceptus
extends the viability of uNK cells. Pigs are a species with noninvasive epitheliochorial placentation. Decidua is not induced and
conceptus attachment is required for uNK cell recruitment. uNK
cells have only been studied over the first 35 of 114 days of porcine gestation with functional activity (target killing) and enrichment (1, 5) reported between gestation days (gd) 12–28. Murine
uNK cells make substantial contributions to maternal arterial
changes by mechanisms that include production of IFN-␥ (6) and
*Department of Biomedical Sciences, University of Guelph, Guelph, Ontario, Canada; and †Department of Anatomy and Cell Biology, Queen’s University, Kingston,
Ontario, Canada
Received for publication July 18, 2005. Accepted for publication October 25, 2005.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
The porcine partial coding sequences presented in this article have been submitted
to GenBank under the following accession numbers: AY616676 (VEGF), AY836553
(HIF-1␣), AY562551 (IFN-␥), AY781397 (FasL), AY781398 (Fas), AY572787
(TNF-␣), AY577818 (IL-1␤), and AY577819 (IL-1R).
2
Address correspondence and reprint requests to Dr. Chandrakant Tayade, Department of Biomedical Sciences, University of Guelph, Guelph, Ontario, Canada
N1G2W1. E-mail address: [email protected]
3
C.T. and G.P.B. contributed equally to this work.
Abbreviations used in this paper: uNK, uterine NK; gd, gestation day; HIF-1␣,
hypoxia-inducible factor 1-␣; VEGF, vascular endothelial growth factor; LCM, laser
capture microdissection; NP, nonpregnant; FasL, Fas ligand; aRNA, antisense RNA.
4
Copyright © 2005 by The American Association of Immunologists, Inc.
synthesis of hypoxia-inducible factor 1-␣ (HIF-1␣)-regulated vascular endothelial growth factor (VEGF) (7, 8). These changes enhance blood supply to developing conceptuses. No functions have
been attributed to porcine uterine lymphocytes.
Commercial pork breeds ovulate 14 –16 ova (9, 10). Fertilized
ova develop as spheres until gd10, then undergo rapid elongation
and attach to noneroded endometrial epithelium between gd11–13
(11). Maternal angiogenesis is initiated at gd15 (12, 13). In commercial pigs, significant conceptus loss (⬃30%) occurs during the
attachment and immediate postattachment intervals (gd15–30). Attachment stage porcine blastocysts are heterogeneous in shape and
in the onset of trophoblastic production of estrogen and IFNs ␦ and
␥ (14). These trophoblastic IFNs are proposed to play important
roles during attachment of conceptuses to the uterus (15). The most
rapidly maturing (elongating), earliest attaching blastocysts appear
to create a hostile environment within the endometrial secretions
for their more slowly developing, genetically normal littermates
and appear to outcompete them for survival. Major events in early
pig pregnancy are summarized in Fig. 1. Cytokines play important
roles during embryo attachment and, in pigs, conceptuses invoke
acute-phase inflammatory responses that involve IL-1␤, TGF-␤,
TNF-␣, and IFN-␥ (16). IL-1␤ promotes rapid trophoblastic elongation (17).
The Chinese Meishan pig is a prolific breed whose litter size
more closely matches its ovulation rate of 14 –16. Reciprocal embryo transfers between Meishan and North American commercial
Large White pigs identified the endometrium as the regulator of
conceptus outcome, because blastocysts from commercial pigs
transferred to Meishan survived, whereas Meishan litters were
smaller when gestated in commercial dams (18). The endometrium
in Meishan implantation sites is more heavily vascularized than
that in commercial breeds (19). To address whether the lymphocytes recruited to early implantation sites in commercial meat pigs
contribute to endometrial vascular development, we studied
changes in gene expression between virgin and peri-implantation
0022-1767/05/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
Prenatal mortality reaching 30% occurs during the first weeks of gestation in commercial swine. Mechanisms for this are unknown
although poor uterine blood supply has been postulated. In other species, vascular endothelial growth factor, hypoxia-inducible
factor 1-␣, and IFN-␥ regulate gestational endometrial angiogenesis. Vascular endothelial growth factor and hypoxia-inducible
factor 1-␣ are also important for placental angiogenesis while trophoblastic expression of Fas ligand is thought to protect conceptuses against immune-mediated pregnancy loss. In this study, we document dynamic, peri-implantation differences in transcription of genes for angiogenesis, cytokine production, and apoptosis regulation in the endometrium, and laser capture microdissected endometrial lymphocytes and trophoblasts associated with healthy or viable but arresting porcine fetuses. In healthy
implantation sites, endometrial gene expression levels differed between anatomic subregions and endometrial lymphocytes showed
much greater transcription of angiogenic genes than trophoblasts. In arresting fetal sites, uterine lymphocytes had no angiogenic
gene transcription and showed rapid elevation in transcription of proinflammatory cytokines Fas and Fas ligand while trophoblasts showed elevated transcription of IFN-␥ and Fas. This model of experimentally accessible spontaneous fetal loss, involving
blocked maternal angiogenesis, should prove valuable for further investigations of peri-implantation failure of normally conceived
and surgically transferred embryos in many species, including the human. The Journal of Immunology, 2006, 176: 148 –156.
The Journal of Immunology
149
FIGURE 1. A summary of the important events during early development in commercial meat pigs. Routinely, 25–35% prenatal mortality is observed before
gd30. The dashed line indicates the absence of information in later gestation.
Materials and Methods
Animals and tissue collection
Allogeneic-specific, pathogen-free, line-bred Yorkshire gilts (n ⫽ 18; Arkell Swine Research Station, University of Guelph) were either not mated
or bred twice, 24 h apart, naturally or by artificial insemination at their first
estrus. Reproductive tracts were recovered immediately after abattoir
slaughter at the University of Guelph using protocols approved by the
institutional Animal Care Committee. The gd was estimated from the first
mating day.
Uteri were transported on ice to an RNase decontaminated dissection
area and opened longitudinally along the antimesometrial side. For nonpregnant (NP) uteri, endometrial biopsies were collected from random antimesometrial and mesometrial (side of uterine artery entry) sites. At gd15,
attachment sites were identified under dissecting microscope magnification; later stages were visualized without aid. Within litters, gd21 and 23
conceptuses were not similar and they were grouped as healthy or arresting
by fetal length, weight, and vascularity of the placental membranes (Fig.
2). For each pregnant uterus, three or four healthy implantation sites were
studied as individual samples. When arresting fetal sites were recognized,
each one was collected separately, resulting in at least three additional sites
being studied per dam in each later pregnancy. Conceptuses enclosed in
membranes were peeled from each pregnant uterus and analyzed separately. Gestational sacs were opened to identify fetus and amnion. The
trophoblast was then dissected free from these tissues and placed in RNA
FIGURE 2. A, Uterine lumen surface, exposed by an antimesometrial
incision, from which healthy (B) and
arresting (C) gd23 littermates were
obtained. These fetuses were categorized on the basis of weight, length,
and vascularity of the placental membranes. Normally, three to four arresting fetuses were observed in a litter of 10 –12/gilt and sampled
individually as were equal numbers
of healthy implantation sites (L,
length; W, weight). Endometrial tissue samples (E) were collected from
the attachment sites of both types of
fetuses for RNA extraction or cryosectioning. Approximately 500 LCM
lymphocytes (LY) were collected
from the endometrium. Membraneenclosed fetuses were removed, and
the trophoblast (T) was dissected for
RNA isolation from regions not apposed by other membranes.
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
uteri of commercial pigs using quantitative real-time PCR. Endometrium, laser capture microdissected (LCM) endometrial lymphocytes and trophoblasts were compared. By gd20, arresting fetuses (viable but destined to die (Fig. 2)) could be identified.
Relative gene expression from gd20 was therefore compared between samples from littermates who would have different gestational outcomes. Our results suggest porcine uterine lymphocytes
are more actively involved in angiogenesis and oxygen sensing
than trophoblasts and that the maternal endometrium and its lymphocytes acutely withdraw all vascular support from arresting conceptuses. This occurs simultaneously with a highly localized, cytokine-based immune attack.
150
ALTERED GENE EXPRESSION ASSOCIATED WITH FETAL ARREST
lysis buffer. After removing conceptuses, endometrial biopsies (⬃30 mg)
were collected immediately under the attachment sites (mesometrial), antimesometrial to the attachment or between two healthy attachment sites.
Endometrial biopsies were either placed in 600 ␮l of lysis buffer or embedded in OCT (Thermo Shandon) for cryosectioning.
RNA isolation from embryo attachment sites and trophoblasts
Tissues in RLT buffer (RNeasy mini kit; Qiagen) were disrupted using
Kontes pestles (Fisher Scientific). Total RNA was extracted following
manufacturer’s instructions. Briefly, tissue lysate was centrifuged
(15,000 ⫻ g, 3 min) and the cleared lysate was mixed with 700 ␮l of 70%
ethanol. The reaction mixture was washed over an RNeasy mini column.
The RNA was eluted using 50 ␮l of nuclease-free water and quantified by
a RNA/DNA calculator (Genequant Pro; Promega) and stored at ⫺80°C.
cDNA synthesis, cloning, and sequencing
Amplified aRNA from the LCM-isolated lymphocytes and total RNA from
endometrial biopsies were reverse-transcribed using the First Strand cDNA
synthesis kit (Amersham Biosciences) as per manufacturer’s instructions.
The resulting cDNA was stored at ⫺20°C. PCR-amplified products of
VEGF, HIF-1␣, IFN-␥, TNF-␣, IL-1, IL-1R, Fas ligand (FasL), and Fas
were cloned using the TOPO-TA cloning kit (Invitrogen Life Technologies) as per manufacturer’s instructions. Plasmid DNA was purified by the
Genelute plasmid DNA purification kit (Sigma-Aldrich). Sequencing was
done at the Molecular Biology SuperCentre, University of Guelph. Sequences were analyzed by the BLASTN program of the National Center for
Biotechnology Information portal and deposited to GenBank.
Quantitative real-time PCR
LCM and RNA amplification
Real-time PCR (LightCycler; Roche Diagnostics) was used to quantify
expression of target genes relative to ␤-actin in the endometrial biopsies
and in lymphocytes. Each sample was analyzed at least twice and averaged.
Primers, designed using the Primer 3 software program (具http://frodo.wi.
mit.edu/cgi-bin/primer3/primer3_www.cgi典), are given in Table I. The
Quantitect SYBR green I PCR mix kit (Qiagen) was used for the quantification of gene expression. LightCycler reactions were performed in 20 ␮l
of total reaction volume as per manufacturer’s instructions. PCR products
were gel-purified using the Wizard DNA purification system (Promega)
and/or plasmid DNA with specific inserts were quantified and diluted serially to generate standard curves for each gene. The LightCycler program
for each gene was denaturation (94°C, 15 min); PCR amplification and
quantification (95°C, 10 s; 58°C, 5 s; 72°C, 20 s) with the fluorescence
measurement at specific acquisition temperatures for 5 s, repeated for 45
cycles. The melting program was 70 –95°C at the rate of 0.1°C/s with
continuous fluorescence measurement, with the final cooling step at 40°C.
Data were quantified using RelQuant LightCycler analysis software; the
normalized ratio was calculated by the software using the following formula: ((median [target]/median [reference])/(median [targetcal]/median
[referencecal] ⫻ correction factor)) ⫻ multiplication factor.
RelQuant software was used to calculate the normalized ratio using the
coefficient file. The correction and multiplication factors were used to correct the PCR efficiency and variations between the different real-time PCR
runs. Second-derivative maximum analysis, arithmetic baseline adjustment, and polynomial calculation methods were used for quantification.
Baseline curve, melting curve, melting point, crossing point, slope error
(0.1– 0.5), and correlation (r-1) were critically monitored for each round of
analysis. The ratio between the target gene and ␤-actin was used as a level
of mRNA expression.
Statistical analyses
Statistical analyses were performed by the nonparametric Friedman test
using SAS software (SAS 8.2; SAS Institute) for comparison among the
groups. Post hoc analysis for planned comparisons between different paired
groups was done using the Wilcoxon signed rank test. A value of p ⬍ 0.05
Table I. Primers used for real-time quantitative PCR
Porcine Gene
VEGF
HIF-1␣
IFN-␥
FasL
Fas
TNF-␣
IL-1␤
IL-1R
␤-Actin
Primer
Sequence
Product Size
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
ACGACGAAGGTCGGAGTGT
AAATGCTTTCTCCGCTCTGA
TTTAGATTTGGCACAATGAC
AGGGACTCTGGATTTGGTTC
GGTAGCTCTGGAAACTGATG
GGCCCTGGAACATAGTCTGG
GGCCTGTGTCTCCTTGTGAT
TTCCAGAGGGATGGATCTTG
CCACGTGTGAACATGGAGTC
ATCATTGGCACCTCTTCAGC
ACTCACCCCTCCCTCTTTGT
GCTGATGGTGTGAGTGAGGA
CAGCCATGGCCATAGTACCT
CCACGATGACAGACACCATC
CCTTGCTTTGCAGGTAATCA
AGTGGGTGGGTGTACAAAGC
ACGTGGACATCAGGAAGGAC
ACATCTGCTGGAAGGTGGAC
196
228
166
246
229
224
216
213
210
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
Sections from frozen endometrial biopsies were cut (7 ␮m) and stained
with a modified rapid H&E protocol to identify uterine lymphocytes. All
solutions, including the stains, were supplemented with 0.5 U/␮l RNase
inhibitor (Promega). Slides were air-dried (5 min), placed into a slide box
over desiccant and moved to the LCM (Pix Cell IIe; Arcturus). Endometrial
lymphocytes were transferred individually to high-sensitivity LCM caps
using laser pulse settings of 55 mW for 0.7 ms at 7.5 ␮m. LCM was
performed rapidly at room temperature. Five hundred lymphocytes were
captured for each sample.
RNA was extracted using a Picopure RNA isolation kit and Extracsure
assembly in an alignment tray (Arcturus) as per manufacturer’s instructions. Briefly, the cap containing laser-captured cells was aligned in the
Extracsure assembly and filled with 10 ␮l of extraction buffer. After centrifugation to collect cell extract in a microcentrifuge tube, 10 ␮l of 70%
ethanol were added to the cell extract and loaded onto a preconditioned
RNA purification column. The cell extract was purified by the RNA purification column. The RNA was eluted in 11 ␮l of nuclease-free water and
stored at ⫺80°C. RNA amplification was conducted using the MessageAmp II antisense RNA (aRNA) kit (Ambion) as per the manufacturer’s
instructions. The procedure consisted of reverse transcription with oligo(dT) primer bearing a T7 promoter and in vitro transcription of DNA
with T7 RNA polymerase to produce copies of aRNA for each mRNA in
the sample. Briefly, to 11 ␮l of total RNA isolated using the Picopure
extraction kit, 1 ␮l of T7 oligo(dT) primer was added, and the mixture was
incubated (10 min, 70°C). Eight microliters of reverse transcription master
mix were added to each sample and incubation continued (2 h, 42°C).
Second-strand cDNA synthesis followed immediately using a master mix
comprised of 20 ␮l of cDNA sample just obtained, 63 ␮l of nuclease-free
water, 10 ␮l of 10⫻ second-strand buffer, 4 ␮l of dNTP mix, 2 ␮l of DNA
polymerase, and 1 ␮l of RNA H and incubation (2 h, 16°C). The resulting
cDNA was purified using a cDNA filter cartridge as per kit instructions,
eluted in 20 ␮l of nuclease-free water and was used for in vitro transcription to synthesize aRNA (16 ␮l of purified cDNA, 4 ␮l of each T7 ATP,
CTP, GTP, UTP 75 mM solutions, T7 10⫻ reaction buffer and T7 enzyme
mix, 6 h, 37°C). The aRNA was eluted in 100 ␮l of nuclease-free water and
stored at ⫺80°C.
The Journal of Immunology
was considered significant. Data are presented as box plots showing median and quartiles which were prepared using Statgraphics Centurion
software.
Results
Cloning and analysis of partial coding sequences
VEGF, HIF-1␣, IFN-␥, FasL, and Fas detected in porcine uterine
lymphocytes and TNF-␣, IL-1␤, IL-1R detected in porcine endometrium were cloned in TOPO-TA cloning vectors. Basic local
alignment search tool analysis revealed 98 –100% sequence homology with the published nucleotide sequences.
Gene expression in endometrial biopsies
and interconceptus attachment sites. For these genes, transcription
in the mesometrial interconceptus endometrium was higher than in
the antimesometrial endometrium. HIF-1␣ was most abundantly
expressed in antimesometrial tissue ( p ⬍ 0.05) at gd19, compared
with mesometrial and interconceptus tissues.
Time-course analysis with living and dying fetuses. To define
more precisely the control of mesometrial angiogenesis immediately postattachment, a time-course study was undertaken from
gd15–23 with three or more pregnancies per time point. From
gd20, arresting fetuses could be identified in each litter and tissues
associated with each of these fetuses were processed separately
from tissues associated with apparently healthy, viable conceptuses. In endometrium from attachment sites with healthy conceptuses, both VEGF (Fig. 4A) and HIF-1␣ (Fig. 4B) were expressed
and expression progressively increased to gd23. No VEGF and
minimal HIF-1␣ were found at attachment sites of arrested fetuses.
In healthy attachment sites, endometrial transcription of IFN-␥ was
highest at gd15. For the other gd studied, levels were only slightly
above those in virgin uteri (Fig. 4C) and no significant differences
were found. Endometrium associated with arresting fetuses had
much higher transcription of IFN-␥ than endometrium associated
FIGURE 3. Transcription of VEGF (A), HIF-1␣ (B), IFN-␥ (C), and TNF-␣ (D) in endometrial subregions at gd19. For each region, samples were
collected relative to multiple implant sites in three gilts. Three or more NP gilts were also studied. Samples were dissected from the mesometrial (entry
site of maternal blood vessels), antimesometrial (opposite to mesometrial side), and mesometrial interconceptus endometrium. Gene expression is relative
to ␤-actin and is expressed as a normalized ratio that differs between anatomical subregion. Significantly higher expression for VEGF, TNF-␣, and IFN-␥
was found mesometrially (ⴱ, p ⬍ 0.05, ⴱⴱ, 0.01), compared with antimesometrially or in interconceptus sites. HIF-1␣ expression was significantly higher
antimesometrially, compared with mesometrially and between conceptus attachment sites (p ⬍ 0.05). M, Mesometrial side; AM, antimesometrial side; IC,
interconceptus. Data are presented by box plot showing median and quartiles.
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
Virgin vs gd19. To determine whether genes promoting angiogenesis are expressed uniformly in pregnant uterus, endometrial
biopsies of NP (n ⫽ 3) and gd19 (n ⫽ 3) gilts were studied for
expression of VEGF, HIF-1␣, IFN-␥, and TNF-␣. All four genes
were detected in NP endometrium, and their transcription was altered by pregnancy (Fig. 3). In pregnant uteri, the change in gene
expression varied by anatomic location. For VEGF, IFN-␥, and
TNF-␣, transcription was significantly higher ( p ⬍ 0.05, p ⬍ 0.01)
in the mesometrial endometrium compared with antimesometrial
151
152
ALTERED GENE EXPRESSION ASSOCIATED WITH FETAL ARREST
with viable littermates ( p ⬍ 0.05) at both gd21 and 23 and IFN-␥
transcription increased ( p ⬍ 0.01) in these sites between gd21
and gd23.
To determine whether changes in proinflammatory cytokines
other than IFN-␥ correlated with fetal arrest, TNF-␣, IL-1␤, and
IL-1R were quantified in endometrium at gd21 and 23. All three
genes were highly up-regulated in endometrium from arresting
conceptus attachment sites (gd21, data not shown; gd23 (Fig. 4F)).
IL-1␤ was the most elevated transcript followed by TNF-␣, IFN-␥,
and IL-1R. In the attachment sites of healthy conceptuses, endometrial transcription levels for these genes were significantly lower
( p ⬍ 0.01, p ⬍ 0.05).
FasL, a molecule associated with apoptosis, was induced in the
endometrium by pregnancy. Endometria associated with arresting
fetuses had significantly higher levels ( p ⬍ 0.05) of FasL expression than endometria associated with healthy fetuses (Fig. 4D).
Expression of Fas in endometria from healthy implantation sites
was variable but always lower than that in endometria from sites
with arresting fetuses ( p ⬍ 0.05; Fig. 4E). At gd23, Fas expression
in arresting attachment sites was significantly higher than in
healthy littermate sites ( p ⬍ 0.01).
Endometrial lymphocytes
To define the contribution of endometrial lymphocytes to angiogenesis during early pregnancy, a time-course analysis of transcription of VEGF, HIF-1␣, and IFN-␥ by pools of 500 LCMcaptured mesometrial lymphocytes was conducted. VEGF
expression progressively increased from NP to gd23 in lymphocytes associated with healthy fetuses (Fig. 5A; p ⬍ 0.01). In lymphocytes obtained from endometria associated with arresting fetuses, no VEGF was detected at gd21 and only marginal
expression was detected at gd23. In healthy implantation sites, the
level of VEGF transcription in lymphocytes was much higher than
in total endometrial biopsies, suggesting that lymphocytes are a
major source of this angiogenic product. A progressive increase
was also detected in transcription of HIF-1␣ in lymphocytes from
healthy attachment sites to gd23 (Fig. 5B; p ⬍ 0.05). No HIF-1␣
transcription was detected in lymphocytes from attachment sites
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
FIGURE 4. Expression of VEGF (A), HIF-1␣ (B), IFN-␥ (C), FasL (D), and Fas (E) in endometrial biopsies; n ⫽ 3 gilts for each gd time point studied.
For time-course analysis, three to four implantation sites from each gilt were evaluated. At gd21 and 23, a minimum of three or more healthy and arresting
implantation sites were examined for gene expression. VEGF and HIF-1␣ expression increased progressively to gd 23, compared with gd15 and 19 (ⴱ, p ⬍
0.05) with significantly higher VEGF expression in healthy gd21 attachment sites (ⴱⴱ, p ⬍ 0.01). Unexpectedly, no VEGF and significantly lower HIF-1␣
expression were detected in attachment sites of arresting conceptuses at gd21, compared with healthy sites (p ⬍ 0.05). Both FasL and Fas were expressed
in NP uterus but at lower levels. Elevated levels of FasL/Fas were detected at the attachment sites of arrested conceptuses (ⴱ, p ⬍ 0.05; ⴱⴱ, 0.01) at both
gd21 and 23, compared with healthy attachment sites. At gd15, significantly elevated Fas was found, compared with gd19 and gd21 healthy attachment
sites (p ⬍ 0.05). Gene expression was normalized and expressed as a ratio of target to the housekeeping gene, ␤-actin. F, Expression of proinflammatory
cytokines at the endometrial attachment sites of gd23-arresting fetuses. Significantly elevated IFN-␥, TNF-␣, IL-1␤, and IL-1R were observed at arresting
sites, compared with healthy sites (ⴱ, p ⬍ 0.05, ⴱⴱ, 0.01). ND, Not detected; H, healthy; A, arresting.
The Journal of Immunology
153
containing arresting fetuses. IFN-␥ transcription was similar in
lymphocytes from NP uteri and from healthy implantation sites but
significantly higher in lymphocytes from attachment sites with arresting fetuses at gd21 (Fig. 5C; p ⬍ 0.05), then declined at gd23
but remained significantly higher ( p ⬍ 0.05) than in healthy attachment site lymphocytes. In continuing studies at gd 30, we have
found that the arresting embryos have, for the most part, been
successfully eliminated. We therefore speculate that the implant
site crisis started before we could see gross differences at gd21 and
that it was resolving by gd23 when lower IFN-␥ transcription was
found in endometrial lymphocytes.
FasL and Fas (Fig. 5, D and E) were transcribed by endometrial
lymphocytes from NP animals and the transcripts were significantly elevated after gd15 ( p ⬍ 0.01, p ⬍ 0.05). In healthy attachment sites, lymphocyte transcription of both genes peaked at
gd21 and then declined. Both genes were more abundantly transcribed in lymphocytes associated with implantation sites containing arresting fetuses and, again, levels at gd21 exceeded those
at gd23.
from healthy fetuses but, at both time points, trophoblasts had
fewer transcripts than did maternal lymphocytes from the same
implantation site (Fig. 6A). VEGF transcription was significantly
lower in trophoblasts from arresting compared with healthy fetuses
( p ⬍ 0.05) and declined from gd21 to 23. HIF-1a transcription
decreased in trophoblasts between gd21 and 23 whether or not the
fetus was arresting ( p ⬍ 0.05; Fig. 6B). Trophoblastic IFNs are
produced abundantly in pigs from gd12. At gd21, there was no
noticeable difference in IFN-␥ expression by trophoblasts from the
two types of littermates (Fig. 6C), but by gd23, IFN-␥ expression
had increased ⬃10-fold in trophoblasts of arresting fetuses ( p ⬍
0.05). Transcription of FasL (Fig. 6D) and Fas (Fig. 6E) changed
dynamically in the trophoblasts of healthy gd21 and 23 fetuses
with the decline of FasL and elevation of Fas ( p ⬍ 0.05). Low
FasL transcription and high Fas transcription were found in trophoblasts from arresting fetuses. The elevated expression of both
IFN-␥ and Fas in gd23 trophoblasts indicates that the decline of
expression of other genes in trophoblasts from arresting fetuses
was not yet due to loss of trophoblast cell viability.
Trophoblasts
Trophoblasts were dissected from normal and arresting fetuses and
examined for gene expression at gd21 and gd23. VEGF transcription increased ( p ⬍ 0.01) between gd21 and 23 in trophoblasts
Discussion
Angiogenesis is a carefully regulated process that occurs in both
maternal and fetal implantation site tissues. The pig is an excellent
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
FIGURE 5. Transcription profiles for endometrial lymphocytes isolated by LCM suggest that lymphocytes contribute to and regulate angiogenesis
through transcription of VEGF (A), HIF-1␣ (B), and IFN-␥ (C). Marginal VEGF, no HIF-1␣, and unusually high IFN-␥ transcription (p ⬍ 0.05) were found
in lymphocytes obtained from attachment sites containing arresting conceptuses, compared with the lymphocytes from healthy attachment sites. FasL/Fas
(D and E) transcription was greater in lymphocytes associated with arresting than viable conceptuses (p ⬍ 0.05). Gene expression is relative to ␤-actin.
Animal numbers and implantation sites studied are as given in the legend to Fig. 4. ND, Not detected.
154
ALTERED GENE EXPRESSION ASSOCIATED WITH FETAL ARREST
animal for study of implantation site angiogenesis and its maternal-fetal cross-regulation because maternal and fetal tissues are
clearly distinguishable and can be studied independently as pure
cell types. Pig placentation is simply apposed epitheliochorial; the
uterine epithelium is not breached by trophoblasts (20). The endometrium was not uniform in its expression of VEGF, the gene
we selected as a prototypic marker for angiogenesis. The highest
VEGF expression occurred mesometrially, where major branches
of the uterine and ovarian arteries supply the uterus. Species with
epitheliochorial placentation use growth of subendothelial capillary networks to expand maternal blood supply to the placenta
rather than spiral arterial modification. Nonetheless, porcine endometrial lymphocytes share angiogenic potential with uNK cells
from hemochorial implantation sites of women and mice (7, 21).
When compared with entire endometrium, transcription of VEGF
was much higher in lymphocytes than in the complex tissue, suggesting lymphocytes may be major stimuli for endometrial angiogenesis during gestation. VEGF transcription in lymphocytes also
exceeded that in trophoblasts, the interface tissue more commonly
studied.
Lymphocytes in normal porcine implantation sites transcribe
IFN-␥, as reported for uNK cells in species with hemochorial pla-
centation (5–7). Previously, IFN-␥ in porcine implantation sites
was solely attributed to trophoblasts. Our relative quantification
indicated that lymphocytes, not trophoblasts, are likely the major
producers of this cytokine in early porcine implantation sites.
TNF-␣ transcription was not found in the lymphocytes from normal implantation sites but low, differentially distributed levels occurred in healthy implantation site endometrium (Fig. 3D).
uNK cells are enriched 2- to 3-fold in pig mesometrial endometrium early after trophoblast attachment (1), but because we
dissected all lymphocytes, IFN-␥ production cannot be attributed
exclusively to uNK cells. Similarly, contributions of lymphocyte
subsets other than uNK cells to angiogenesis and to oxygen sensing cannot be established until reagents become available for rapid
discrimination of porcine uterine lymphocyte subsets and permit
reassessment of our data using lymphocyte subset analysis. We
also cannot assume that the lymphocyte subset mixtures we studied on different gd or between living and arresting littermates were
similar. The transcriptional changes we documented may reflect
either activation of cells already at attachment sites, transcripts in
newly recruited cells or a mixture of both. It is anticipated that
many genes, in addition to those examined, are altered in healthy
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
FIGURE 6. Expression of VEGF (A), HIF-1␣ (B), IFN-␥ (C), FasL (D), and Fas (E) in trophoblasts. Elevated VEGF transcription was found in gd21
and 23 trophoblasts of healthy conceptuses, compared with gd21 and 23 arresting ones, whereas HIF-1␣ transcription was significantly high at gd21 (p ⬍
0.05). Significantly lower VEGF and HIF-1␣ expression were found in trophoblasts from arresting conceptuses (ⴱ, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.01). Unusually high
levels of IFN-␥ expression were induced by gd23 (p ⬍ 0.05) in arresting conceptuses, compared with healthy littermates and compared with gd21 healthy
and arresting conceptuses. FasL expression declined between gd21 and 23 for both healthy and arresting conceptuses but less transcription occurred in the
latter but significantly elevated in healthy sites (p ⬍ 0.05), compared with arresting. Fas expression increased from gd21–23 for both healthy and arresting
conceptuses but was higher in the latter (p ⬍ 0.05). Gene expression is relative to ␤-actin.
The Journal of Immunology
way not only to promote agriculture but also to give insight into
potential pathways for human embryo loss following conception or
assisted reproductive techniques.
Acknowledgments
We thank Dr. D. Bienzle for access to her LCM and Lightcycler facilities
and Dr. H. Engelhardt for helpful guidance during the course of these
studies.
Disclosures
The authors have no financial conflict of interest.
References
1. Engelhardt, H., B. A. Croy, and G. J. King. 2002. Conceptus influences the
distribution of uterine leukocytes during early porcine pregnancy. Biol. Reprod.
66: 1875–1880.
2. Moffett-King, A. 2002. Natural killer cells and pregnancy. Nat. Rev. Immunol. 2:
656 – 663.
3. Peel, S. 1989. Granulated metrial gland cells. Adv. Anat. Embryol. Cell Biol. 115:
1–112.
4. Kammerer, U., M. Schoppet, A. D. McLellan, M. Kapp, H. I. Huppertz,
E. Kampgen, and J. Dietl. 2000. Human decidua contains potent immunostimulatory CD83⫹ dendritic cells. Am. J. Pathol. 157: 159 –169.
5. Croy, B. A., A. Waterfield, W. Wood, and G. J. King. 1988. Normal murine and
porcine embryos recruit NK cells to the uterus. Cell. Immunol. 115: 471– 480.
6. Ashkar, A. A., J. P. Di Santo, and B. A. Croy. 2000. Interferon ␥ contributes to
initiation of uterine vascular modification, decidual integrity, and uterine natural
killer cell maturation during normal murine pregnancy. J. Exp. Med. 192:
259 –270.
7. Wang, C., T. Tanaka, H. Nakamura, N. Umesaki, K. Hirai, O. Ishiko, S. Ogita,
and K. Kaneda. 2003. Granulated metrial gland cells in the murine uterus: localization, kinetics, and the functional role in angiogenesis during pregnancy.
Microsc. Res. Tech. 60: 420 – 429.
8. Li, X. F., D. S. Charnock-Jones, E. Zhang, S. Hiby, S. Malik, K. Day, D. Licence,
J. Bowen, M. L. Gardner, A. King, et al. 2001. Angiogenic growth factor messenger ribonucleic acids in uterine natural killer cells. J. Clin. Endocrinol. Metab.
86: 1823–1834.
9. Anderson, L. H., L. K Christenson, R. K. Christenson, and S. P. Ford. 1993.
Investigations into the control of litter size in swine. II. Comparisons of morphological and functional embryonic diversity between Chinese and American
breeds. J. Anim. Sci. 71: 1566 –1571.
10. Youngs, C. R., L. K. Christenson, and S. P. Ford. 1994. Investigations into the
control of litter size in swine. III. A reciprocal embryo transfer study of early
conceptus development. J. Anim. Sci. 72: 725–731.
11. Keys, J. L., and G. J. King. 1990. Microscopic examination of porcine conceptusmaternal interface between days 10 and 19 of pregnancy. Am. J. Anat. 188:
221–238.
12. Dantzer, V., and R. Leiser. 1994. Initial vascularisation in the pig placenta. I.
Demonstration of nonglandular areas by histology and corrosion casts. Anat. Rec.
238: 177–190.
13. Winther, H., A. Ahmed, and V. Dantzer. 1999. Immunohistochemical localization of vascular endothelial growth factor (VEGF) and its two specific receptors,
Flt-1 and KDR, in the porcine placenta and non-pregnant uterus. Placenta 20:
35– 43.
14. La Bonnardiere, C., F. Martinat-Botte, M. Terqui, F. Lefevre, K. Zouari,
J. Martal, and F. W. Bazer. 1991. Production of two species of interferon by
Large White and Meishan pig conceptuses during the peri-attachment period.
Reprod. Fertil. 91: 469 – 478.
15. Cencic, A., M. Guillomot, S. Koren, and C. La Bonnardiere. 2003. Trophoblastic
interferons: do they modulate uterine cellular markers at the time of conceptus
attachment in the pig? Placenta 24: 862– 869.
16. Geisert, R. D., and, J. V. Yelich. 1997. Regulation of conceptus development and
attachment in pigs. J. Reprod. Fertil. 52(Suppl.): 133–149.
17. Ross, J. W., Malayer, Jr., J. W. Ritchey, and R. D. Geisert. 2003. Characterization
of the interleukin-1␤ system during porcine trophoblastic elongation and early
placental attachment. Biol. Reprod. 69: 1251–1259.
18. Wilson, M. E., N. J. Biensen, C. R. Youngs, and S. P. Ford. 1998. Development
of Meishan and Yorkshire littermate conceptuses in either a Meishan or Yorkshire uterine environment to day 90 of gestation and to term. Biol. Reprod. 58:
905–910.
19. Biensen, N. J., M. E. Wilson, and S. P. Ford. 1998. The impact of either a
Meishan or Yorkshire uterus on Meishan or Yorkshire fetal and placental development to days 70, 90, and 110 of gestation. J. Anim. Sci. 76: 2169 –2176.
20. Guillomot, M., G. E. Flechon, and F. Leroy. 1993. Blastocyst development and
implantation. In Reproduction in Mammals and Man. C. Thbault,
M. C. Levasseur, and R. H. F. Hunter, eds. Ellipses, Paris, pp. 387– 410.
21. Charnock-Jones, D. S., A. M. Sharkey, J. Rajput-Williams, D. Burch,
J. P. Schofield, S. A. Fountain, C. A. Boocock, and S. K. Smith. 1993. Identification and localization of alternately spliced mRNAs for vascular endothelial
growth factor in human uterus and estrogen regulation in endometrial carcinoma
cell lines. Biol. Reprod. 48: 1120 –1128.
22. Cheung, C. Y. 1997. Vascular endothelial growth factor: possible role in fetal
development and placental function. J. Soc. Gynecol. Investig. 4: 169 –177.
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017
implant sites and that novel as well as classical mechanisms could
contribute to implantation site angiogenesis.
Hypoxia is a potent stimulus for VEGF production and is
thought to be essential for development of both embryonic and
placental vasculature in early human pregnancy (22, 23). HIF-1␣,
a prime regulator of oxygen homeostasis, binds to the hypoxia
response element in the VEGF promoter (24) and is reported to
regulate peri-implantation angiogenesis in humans (25), mice (26),
and sheep (27). Our finding of HIF-1␣ expression in a species with
epitheliochorial placentation suggests that transcription is promoted by attachment and growth of conceptuses (Fig. 4C). Induction was, however, variable (Fig. 3B), and there was significant
transcription of HIF-1␣ in the antimesometrial endometrium.
The pig provides a strong experimental model for investigations
of peri-implantation spontaneous loss of apparently normal conceptuses. The fetuses we classified as arresting were not dead.
Indeed, their trophoblasts showed elevated transcription of some
genes, indicative of a response to stress or environmental change.
A highly localized regulatory step effectively blocked maternal
transcription of VEGF and HIF-1␣. The most probable source for
a highly localized regulator, not affecting the entire litter, is trophoblasts within the individual implantation site. A number of danger signals have been described that are recognized by immune
cells (28). Most of these evoke dendritic or phagocytic cell recognition of dying cells (29, 30). Danger signals include natural or
endogenous adjuvants such as stress and heat shock proteins, fragments of fibronectin, hyularonic acid, free DNA, CpG oligonucleotides, and uric acid (31, 32). Elevated circulating fetal DNA and
trophoblast membrane fragments are associated with human fetal
stress (33, 34). TLRs bind some danger molecules and may directly activate lymphocytes. Endometrial immunity could also become activated via the complement cascade (35, 36), changes in
soluble forms of trophoblast transplantation Ags (37), or induction
by trophoblast of adjuvant cytokines such as IL-12 (38). The finding of elevated IFN-␥ gene expression in trophoblasts from arresting fetuses strongly supports the last mechanism. A trophoblastderived natural adjuvant signal could induce the endometrial
cytokine storm documented with gd21 and 23 fetal retardation. A
key question is what cell types and tissues are targets of this cytokine
aggression? We suggest the target is neither trophoblast nor fetus.
Rather, we hypothesize that the immune storm is directed toward the
maternal vasculature and its purpose is to eliminate/destroy maternal
support for a dangerous, stressed, or about-to-fail conceptus by initiation of death in active endothelial tip cells and in those uNK cells
providing the strongest support for angiogenesis.
Support for endometrial targets also comes from our examination of FasL/Fas expression. Trophoblasts dynamically express
both molecules differentially between viable and arresting conceptuses. Thus, there is no absolute requirement for death signals of
maternal origin. FasL and Fas are also elevated in the endometrium
and in endometrial lymphocytes associated with arresting, compared with healthy, conceptuses. This combined elevation would
be expected to protect some but not all lymphocytes and some but
not all endometrial cells from death signals.
Further studies will be needed to resolve when heterogeneity for
survival becomes established between porcine littermates and
whether mechanisms defined in this model of peri-implantation
pregnancy failure are applicable to species with other placental
types, particularly humans. We have documented for the first time
the major and relative contributions of endometrial lymphocytes to
both angiogenesis and oxygen sensing in normal implantation sites
and the complete cessation of maternal angiogenesis in sites with
arresting conceptuses. This report has also defined an experimental
pregnancy model that can be interrogated in a precise and detailed
155
156
ALTERED GENE EXPRESSION ASSOCIATED WITH FETAL ARREST
23. Qian, D., H. Y. Lin, H. M. Wang, X. Zhang, D. L. Liu, Q. L. Li, and C. Zhu.
2004. Normoxic induction of the hypoxic-inducible factor-1␣ by interleukin-1␤
involves the extracellular signal-regulated kinase 1/2 pathway in normal human
cytotrophoblast cells. Biol. Reprod. 70: 1822–1827.
24. Semenza, G. L. 2000. HIF-1: using two hands to flip the angiogenic switch.
Cancer Metastasis Rev. 19: 59 – 65.
25. De Marco, C. S., and I. Caniggia. 2002. Mechanisms of oxygen sensing in human
trophoblast cells. Placenta 23: S58 –S68.
26. Daikoku, T., H. Matsumoto, R. A. Gupta, S. K. Das, M. Gassmann,
R. N. DuBois, and S. K. Dey. 2003. Expression of hypoxia-inducible factors in
the peri-implantation mouse uterus is regulated in a cell-specific and ovarian
steroid hormone-dependent manner: evidence for differential function of HIFs
during early pregnancy. J. Biol. Chem. 278: 7683–7691
27. Nau, P. N., T. Van Natta, J. C. Ralphe, C. J. Teneyck, K. A. Bedell,
C. A. Caldarone, J. L. Segar, and T. D. Scholz. 2002. Metabolic adaptation of the
fetal and postnatal ovine heart: regulatory role of hypoxia-inducible factors and
nuclear respiratory factor-1. Pediatr. Res. 52: 269 –278.
28. Matzinger, P. 2002. An innate sense of danger. Ann. NY Acad. Sci. 961: 341–342.
29. Inaba, K., S. Turley, F. Yamaide, T. Iyoda, K. Mahnke, M. Inaba, M. Pack,
M. Subklewe, B. Sauter, D. Sheff, et al. 1998. Efficient presentation of phagocytosed cellular fragments on the major histocompatibility complex class II products of dendritic cells. J. Exp. Med. 188: 2163–2173.
30. Albert, M. L., B. Sauter, and N. Bhardwaj. 1998. Dendritic cells acquire antigen
from apoptotic cells and induce class I-restricted CTLs. Nature 392: 86 – 89.
31. Shi, Y., J. E. Evans, and K. L. Rock. 2003. Molecular identification of a danger
signal that alerts the immune system to dying cells. Nature 425: 516 –521.
32. Rock, K. L., A. Hearn, C. J. Chen, and Y. Shi. 2005. Natural endogenous adjuvants. Springer Semin. Immunopathol. 26: 231–246.
33. Zhong, X. Y., W. Holzgreve, and S. Hahn. 2005. The levels of circulatory cell
free fetal DNA in maternal plasma are elevated prior to the onset of preeclampsia.
Hypertens. Pregnancy 21: 77– 83.
34. Sargent, I. L., S. J. Germain, G. P. Sacks, S. Kumar, and C. W. Redman. 2003.
Trophoblast deportation and the maternal inflammatory response in pre-eclampsia. J. Reprod. Immunol. 59: 153–160.
35. Mao, D., X. Wu, C. Deppong, L. D. Friend, G. Dolecki, D.M. Nelson, and
H. Molina. 2003. Negligible role of antibodies and C5 in pregnancy loss associated exclusively with C3-dependent mechanisms through complement alternative pathway. Immunity 19: 813– 822.
36. Hill, J. A., G. C. Melling, and P. M. Johnson. 1995. Immunohistochemical studies
of human uteroplacental tissues from first-trimester spontaneous abortion.
Am. J. Obstet. Gynecol. 173: 90 –96.
37. Le Friec, G., B. Laupeze, O. Fardel, Y. Sebti, C. Pangault, V. Guilloux,
A. Beauplet, R. Fauchet, and L. Amiot. 2003. Soluble HLA-G inhibits human
dendritic cell-triggered allogeneic T-cell proliferation without altering dendritic
differentiation and maturation processes. Hum. Immunol. 64: 752–761.
38. Chattergoon, M. A., V. Saulino, J. P. Shames, J. Stein, L. J. Montaner, and
D. B. Weiner. 2004. Co-immunization with plasmid IL-12 generates a strong
T-cell memory response in mice. Vaccine 22: 1744 –1750.
Downloaded from http://www.jimmunol.org/ by guest on June 17, 2017