BIOLOGY OF REPRODUCTION 63, 559569 (2000)
Expression, Ontogeny, and Regulation of Hypoxia-Inducible Transcription Factors in
the Human Placenta1
Augustine Rajakumar and Kirk P. Conrad2
Departments of Obstetrics, Gynecology and Reproductive Sciences and of Cell Biology and Physiology, University of
Pittsburgh School of Medicine and Magee-Womens Research Institute, Pittsburgh, Pennsylvania 15213
ABSTRACT
Placental hypoxia likely plays an important role in both nor-
mal placental development and pathology. Yet, the molecular
mechanisms of hypoxia signaling in this organ are virtually un-
explored. Therefore, we investigated the expression of the hyp-
oxia inducible transcription factors (HIF) in normal human pla-
centas spanning the first trimester to term. Several key obser-
vations emerged: 1) HIF-1a and -2a mRNA were present in pla-
centas of all gestational ages but with greater variability during
early pregnancy; 2) overall, HIF-1a mRNA was expressed at a
constant level in all placentas, whereas HIF-2a mRNA increased
significantly with gestational age; 3) both HIF-1a and -2a pro-
tein decreased significantly with gestational age; and 4) HIF-1a
and -2a immunoreactivity were overlapping in cellular distri-
bution being expressed by the syncytiotrophoblast, villous cy-
totrophoblast, and fetoplacental vasculature with both nuclear
and cytoplasmic localization. Next, we studied the regulation of
these transcription factors by oxygen using placental villous ex-
plants in culture from first-trimester and term placentas. The
major findings were 1) HIF-1a and -2a protein, but not mRNA,
was induced by hypoxia in the placental villous explants; 2) HIF-
1a DNA-binding activity was also stimulated by hypoxia; and 3)
glucose transporter-1 mRNA (a known target of HIF) was also
increased by hypoxia in placental villous explants. We suggest
that physiological hypoxia contributes to the increased expres-
sion of HIF-1a and -2a protein in early placentas and that reg-
ulation of these transcription factors by hypoxia in the human
placenta occurs at the level of protein and not mRNA.
developmental biology, gene regulation, placenta, syncytiotro-
phoblast, trophoblast
INTRODUCTION
Successful pregnancy outcome is based on healthy pla-
cental development that involves effective trophoblast an-
choring to and invasion of the uterus (reviewed in [1]). In
the early stages of human placental development, that is,
before the 10th wk of gestation, the blood flow to the in-
tervillous space is low [2]. The only available data on pla-
cental oxygen levels were collected by Rodesch et al. [3],
who used a transcervical polarographic oxygen electrode to
measure the partial pressure of oxygen in the intervillous
space. They obtained values of 17.9 6 6.9 mmHg at Ges-
tational Weeks 810 and of 60.7 6 8.5 mmHg at Weeks
1
This work was supported by National Institutes of Health Grants PO1
HD30367, RO1 HL56410, and Research Career Development Award
KO4 HD01098.
Correspondence: Kirk P. Conrad, Magee-Womens Research Institute,
2
204 Craft Ave., Pittsburgh, PA 15213. FAX: 412 641 1503; e-mail:
rsikpc@mail.magee.edu
Received: 7 February 2000.
First decision: 9 March 2000.
Accepted: 30 March 2000.
Q 2000 by the Society for the Study of Reproduction, Inc.
ISSN: 0006-3363. http://www.biolreprod.org
559
1213the latter being not significantly different from val-
ues measured concurrently in the endometrium. The higher
levels at Weeks 1213 of pregnancy presumably result from
a rise in intervillous blood flow that, at least in part, stems
from trophoblast invasion and physiological remodeling
of uterine spiral arteries from small-resistance caliber ves-
sels to large-capacitance uteroplacental arteries allowing for
unimpeded blood flow into the intervillous space [1]. Thus,
during the first 10 wk or so of development, the human
placenta resides in a relatively hypoxic environment.
The proper and timely proliferation and differentiation
of the villous cytotrophoblast stem cell are crucial to ade-
quate placentation [4]. Generally speaking, proliferation of
villous cytotrophoblast is required for adequate supply of
these cells. They differentiate into the syncytiotrophoblast,
which lines the villous placenta and separates maternal
blood from villous core. This syncytial cell layer performs
numerous functions critical to successful pregnancy out-
come, including gas and nutrient transport and pregnancy
hormone production. Alternatively, some of the villous cy-
totrophoblast stem cells in the so-called anchoring villi that
abut the maternal decidua invade the uterine parenchyma
and spiral arteries as extravillous trophoblasts. Cytotropho-
blast phenotype is influenced by oxygen; for instance, hyp-
oxia stimulates the proliferation of villous cytotrophoblast
cells while impairing their invasion [56]. Successful preg-
nancy outcome also depends on the proper development of
the fetoplacental vasculature in the villous core, which pro-
vides for the delivery of oxygen and nutrients from the
intervillous space to the fetus [4]. Thus, vasculogenesis and
angiogenesis, as well as the entire repertoire of associated
growth factors, are remarkably active during placental de-
velopment. These processes are well known to be critically
regulated by oxygen. Finally, local ischemia/hypoxia has
been implicated in the abnormal cytotrophoblast differen-
tiation and placental development in preeclampsia [710].
The molecular consequences of cellular hypoxia have
been extensively reviewed in recent years (e.g., [11]). A
variety of genes are modulated as an adaptive response to
hypoxia, such as those involved in glucose transport, gly-
colysis, angiogenesis, and erythropoiesis [11]. A prerequi-
site for many hypoxia-mediated responses is the basic-he-
lix-loop-helix (bHLH)-Per-Arnt-Sim (PAS) family member
transcription factors HIF-1a [12] and HIF-2a [13, 14],
which heterodimerize with the aryl hydrocarbon nuclear
translocator protein (ARNT or HIF-1b; 11, 12) and then
bind to the hypoxia response element (HRE; core consensus
sequence59-CGTG-39) in responsive genes. Although the
identity of the oxygen sensor upstream of the HIF tran-
scription factors remains elusive [15], the molecular mech-
anisms of HIF activation by hypoxia are becoming clearer
and include protein stabilization [16, 17], thiol-redox reg-
ulation [18, 19], and interaction with CBP/p300 [19, 20].
Interestingly, there are agents besides hypoxia that have
been shown to activate HIF under normoxic conditions;
560 RAJAKUMAR AND CONRAD
these include insulin and IGF-1 [21, 22] and v-SRC [23],
as well as transition metals, such as cobalt chloride, and
iron-chelating agents [24]. Finally, HIF transcription factors
have been shown to be critical for mouse development.
Indeed, HIF-1a [25], -2a [26], and -1b [27] homozygous
knockout mice all manifested embryonic lethality.
The contribution of HIF to human placental development
is virtually unexplored. In a survey of various tissues in
humans, Weiner et al. [28] demonstrated the expression of
both HIF-1a and -1b in the term human placenta by North-
ern analysis. The presence of HIF-1 DNA binding activity
in isolated cytotrophoblast cells from term placenta was
documented by electrophoretic mobility shift assay [29].
Also, a survey of transcription factor mRNA expression by
isolated trophoblasts from first-trimester human placenta
was recently published and included HIF-1a, -2a, and -1b
[30]. Accordingly, there were three major objectives of the
present work: 1) to determine the relative abundance of
HIF-1a and -2a mRNA and protein in the human placenta
from early gestation to term; 2) to investigate the regulation
of HIF mRNA-, protein-, and DNA-binding activity by ox-
ygen in the human placenta by using villous explant cul-
tures from first-trimester and term placentas; and 3) to ex-
amine the cellular localization of HIF-1a and -2a protein
in the first-trimester and term human placentas. Given the
major role of oxygen in the regulation of normal and ab-
normal placental development and function, the molecular
mechanisms of oxygen transduction by this organ require
further elucidation.
MATERIALS AND METHODS
Placental Collection and Processing
Placentas were obtained from women with normal preg-
nancies undergoing elective cesarean section at term and
from women undergoing elective pregnancy terminations at
512 gestational wk under approval of the Institutional In-
ternal Review Board of Magee-Womens Hospital.
For evaluating the expression and ontogeny of the hyp-
oxia-inducible transcription factors (HIF), the placental tis-
sues were snap-frozen or placed into 10% formalin imme-
diately after extraction from the uterus. For evaluating the
regulation of HIF expression in placental tissues, villous
explants were prepared as described by Benyo et al. [31],
with modifications. For term placenta, several cotyledons
were excised at random and rinsed extensively in sterile
saline to remove blood. Decidua and large vessels were
removed from the villous placenta by blunt dissection. The
villous tissue was then finely dissected into 510 mg pieces
while in the bath of sterile saline. The pieces were exten-
sively washed two or three more times before use. First-
trimester villi were prepared in a similar fashion.
Villous Explant Culture
We placed 30 to 50 mg of villous tissue into each well
of a 24-well plate (Becton Dickinson, Franklin Lakes, NJ)
containing 1.0 ml of Medium 199 (Mediatech, Cellgro,
Herndon, VA) supplemented with 10% fetal bovine serum
(FBS; Summit Technology, Ft. Collins, CO) and antibiotics.
Explants were incubated at 378C for a 1224 h preincuba-
tion period on an orbital shaker (60 rpm, Belly Dancer;
Stovall Life Science Inc., Greensboro, NC) under standard
tissue culture conditions of 5% CO2-balance room air (nor-
moxia; 20.94% O2) in a cell culture incubator (Forma Sci-
entific, Marietta, OH). Cultures were gently shaken because
it has been reported in stationary cultures of hepatoma cells
maintained under normoxic conditions that the formation
of unstirred fluid layers leads to reduced pericellular pO2
and consequently, to elevated erythropoietin secretion ([32,
33] and unpublished results). After a medium change and
the addition of any treatments, the plates were again placed
on the orbital shaker at 378C for various times under either
normoxia or reduced O2 (hypoxia; 2% O2-5% CO2-balance
nitrogen) in the cell culture incubators. When the tissues
were treated with cobalt chloride (final concentration of 100
mM), the tissues were incubated under normoxic condi-
tions. In experiments involving actinomycin D (10 mg/ml)
or cyclohexamide (100 mM) to inhibit transcription and
translation, respectively, the villous explants were pretreat-
ed for 2 h and then subjected to either the normoxic or
hypoxic incubation conditions in the continuous presence
of the inhibitor.
Using a Tucker chamber and a radiometer O2 electrode
[34], the pO2 of the hypoxic cell culture incubator was rou-
tinely found to be 1520 torr, comparable to the levels ob-
served in the intervillous space during early gestation. To
achieve a more severely hypoxic environment of 1% O2,
modular incubation chambers (Billups-Rotherberg Inc., Del
Mar, CA) were used. These chambers were filled with a 1%
O2-5% CO2-balance nitrogen gas mixture. The pO2 was
routinely 510 torr, which was well maintained for at least
24 h. These chambers were then placed in the cell culture
incubator for the 378C environment. Cytotoxicity was mon-
itored for each experiment based on the release of lactate
dehydrogenase into the medium and was consistently found
to be 610%, irrespective of the O2 atmosphere, as previ-
ously reported [31].
Hepatoma G2 Cell Culture
Hepatoma G2 (HepG2) cells were obtained from the
American Type Culture Collection (Manassas, VA). These
cells were used as a positive control for HIF expression
[35]. The cells were seeded at a density of 50 000 per
square centimeter and were grown in Eagle minimum es-
sential medium supplemented with nonessential amino ac-
ids (Mediatech, Cellgro, Herndon, VA), sodium pyruvate
(1.0 mM; Gibco-BRL, Grand Island, NY), 10% FBS, and
antibiotics.
Electrophoretic Mobility Shift Assay
Mobility shift assays were carried out using modified
procedures of Semenza et al. [36]. For the preparation of
nuclear extracts, villous explants were rapidly pooled from
four to six wells (about 300 mg total wet weight). The
excess medium was blotted on a sterile 4 3 4 gauze, and
the tissues were transferred into four volumes of buffer A
(10 mM Tris-HCl, pH 7.8; 1.5 mM MgCl2; and 10 mM
KCl containing 0.1% Triton-X 100). Buffers A, C, and D
used in the nuclear extract preparation contained 0.5 mM
dithiothreitol (DTT), 0.4 mM phenymethylsulfonyl fluo-
ride, 2 mg per milliliter each of leupeptin, pepstatin, and
aprotinin, as well as 1 mM sodium vanadate. Care was
taken to process the tissue as quickly as possible, and all
the steps were carried out at 48C. Villous tissue was ho-
mogenized in buffer A using a Tissuemizer (Tekmar, Cin-
cinnati, OH) at setting 60 for 45 sec. Cell lysis was mon-
itored microscopically by staining with hematoxylin. The
homogenate was centrifuged at 850 3 g for 5 min to pellet
the nuclei. The crude nuclear pellet was resuspended in 2.5
volumes of buffer C (20 mM Tris-HCl, pH 7.8; 0.42 M
 HIF-1a, -2a, AND -1b IN THE HUMAN PLACENTA
NaCl; 1.5 mM MgCl2; and 20% glycerol). When preparing
nuclear extracts from HepG2 cells, we substituted 0.42 M
KCl for the NaCl. The nuclear proteins were extracted by
gentle rocking at 48C for 30 min. The extract was centri-
fuged at 21 000 3 g for 30 min and then dialyzed against
200 volumes of buffer D (20 mM Tris-HCl, pH 7.8, 0.1 M
KCl, 0.2 mM EDTA, 20% glycerol) for 4 h. The dialyzed
nuclear extract was subjected to mild sonication with a mi-
crotip for 5 sec (Sonifier-Cell Disruptor; Heat Systems-Ul-
trasonics Inc., Plainview, NY) and then was centrifuged at
21 000 3 g for 15 min. The cleared supernatant was stored
in aliquots at 2708C until further use. Protein concentration
was estimated by the Bradford method using reagents from
Bio-Rad (Hercules, CA).
For the preparation of the probe, oligonucleotide repre-
senting the erythropoietin gene HRE sequence 5 9 -
GCCCTACGTGCTGTCTCA (200 ng) was end labeled us-
ing 50 mCi g32P-ATP (3000 Ci/mM, NEN-Dupont) and 20
units of T4 polynucleotide kinase (New England Biolabs,
Beverly, MA) in a 20 ml volume. The labeled oligonucle-
otide was mixed with twofold excess of cold antisense ol-
igonucleotide, boiled at 1008C for 10 min, and slowly al-
lowed to cool to room temperature (RT). The double-
stranded probe was then subjected to electrophoresis on a
20% polyacrylamide gel in 13 TBE (0.09 M Tris, 0.09 M
boric acid, and 2 mM EDTA), and the wet gel was wrapped
in plastic wrap and exposed for autoradiography. The band
was excised from the gel, and the labeled probe was ex-
tracted in a buffer containing 10 mM Tris-HCl, pH 8.0; 1
mM EDTA; and 1 M NaCl at 378C overnight. After a phe-
nol:chloroform extraction, the probe was precipitated with
100% ethanol in the presence of carrier tRNA (10 mg). The
labeled probe was recovered by centrifugation at 21 000 3
g for 15 min, washed once in 70% ethanol, dried, and dis-
solved in minimal amount of sterile water. The radioactivity
was measured using a Beckman scintillation counter.
The DNA-binding reaction was carried out in 30 ml vol-
ume containing 5 mg of nuclear proteins from HepG2 cells
or 30 to 40 mg of nuclear proteins from villous explants
and 0.1 mg of denatured calf thymus DNA in 10 mM Tris
HCl, pH 7.5; 50 mM KCl; 50 mM NaCl; 1 mM MgCl2; 1
mM EDTA; 5 mM DTT; and 5% glycerol. After a 5-min
preincubation, 5 3 104 cpm (0.5 to 1 ng) of the probe was
added, and the incubation was carried out for another 15
min at RT. The reaction mixture was loaded onto a 5% non-
denaturing polyacrylamide gel and electrophoresed in 0.3
3 TBE for 3 h at a constant 200 volts at 48C. The gels
were dried and exposed for autoradiography. In the case of
supershift assays, 2 mg of the HIF antibodies or IgG con-
trols were added to the mix after the initial 15 min, and the
incubation was further carried out for another 30 min before
loading onto the gel. The gels were dried on Whatman 3
filter paper (Whatman International Co., Maidstone, UK)
with a cellophane support (Bio-Rad) using a gel dryer
(Fisher Biotech, Pittsburgh, PA) and were exposed to Ko-
dak-Bio-max-AR film (Eastman Kodak, Rochester, NY).
Northern Blot Analysis
Total RNA was isolated from HepG2 cells, placental tis-
sues, and villous explants using RNAwiz (Ambion, Austin,
TX). After a brief wash in ice-cold PBS, HepG2 cells were
directly lysed in RNAwiz (1 ml per 75 cm2 flask). Placental
tissues and villous explants were first pulverized using a
Bessman tissue pulverizer (Spectrum Medical Industries,
Laguna Niguel, CA) and then homogenized in 10 volumes
561
of RNAwiz using a Tissuemizer. After preparation, the
RNA pellet was dissolved in a minimal volume of RNase-
free water or in RNase-free water containing 0.5% SDS.
The amount of total RNA was estimated by spectrophotom-
etry (A260). One optical density was taken to be 40 mg/ml.
Ten to 20 mg of total RNA was separated on formaldehyde
containing 1.5% agarose gels and transferred to nylon
membranes (MSI, Westborough, MA). RNA on the mem-
branes was cross-linked by using a Biotech UV-cross-linker
(Fisher).
For the preparation of the cDNA probes, plasmids pBS/
HIF-1a3.23T7 (EcoRI digest of the complete cDNA pro-
duces three bands of sizes 2063, 1011, and 604 base pairs
[bp]), and plasmid pGEX2T/ARNT-C (HIF-1b, an 800-bp
EcoRI/BamHI fragment) were generously provided by Dr.
Gregg Semenza. Plasmid hEPAS-pcDNA3 (HIF-2a, a
2818-bp BamHI fragment) was a kind gift from Dr. Steven
McKnight. A 539-bp fragment of the human b-actin was
made by reverse-transcription polymerase chain reaction
using a sense primer 59-GTGGGGCGCCCCAGGCACCA
(nt 144163) and an antisense primer 59-TCCTTAATGT-
CACGCACGATTTC (nt 660683). All restriction diges-
tion fragments were purified from gel using the Jetsorb
DNA purification kit (Genomed GmbH, Germany).
cDNA probes were made using 2550 ng of the insert,
50 mCi of a-32P-dCTP (3000 Ci/mM, NEN-Dupont), and 2
U of Klenow polymerase using the Multi-prime DNA-la-
beling kit (Amersham Pharmacia Biotech, Piscataway, NJ).
Unincorporated free 32P was removed using a spin column
(Bio-Spin P30, Bio-Rad).
The membranes were washed once briefly in 23 SSC
and were processed further for hybridization. Prehybridi-
zation was carried out for 12 h in 6 ml of buffer containing
50% de-ionized formamide, 53 Denhardt solution, 53 sa-
line-sodium phosphate-EDTA buffer, 1% SDS, 0.5 mg/ml
of denatured salmon sperm DNA, and 0.1 mg per ml of
tRNA at 428C in a Hybaid oven with roller bottles (250 3
35 mm). Labeled probe in 4 ml of prehybridization buffer
was added to the membrane (2 3 106 cpm per milliliter),
and the hybridization was carried out as usual for 1224 h.
The membranes were washed in 23 SSC-0.1% SDS twice
for 5 min at RT and then were washed twice in the same
solution at 378C. The washing was carried out for another
15 min at 508C with 0.1 3 SSC-0.1% SDS when necessary.
The membranes were exposed to Kodak-Bio-max-AR film
(Eastman Kodak).
The membranes were stripped in 0.1% SDS at 608C for
1530 min and reprobed for another gene of interest or for
the housekeeping gene, b-actin, to correct for interlane
loading variations. b-actin was chosen as a housekeeping
gene because it has been reported to be relatively unaf-
fected by hypoxia [37, 38]. Our results supported this con-
tentionb-actin mRNA (hypoxia)/b-actin mRNA (nor-
moxia) was 0.88 6 0.10 for term placental villous explants
(n 5 four placentas). The signal on the film was quantitated
by densitometry using the Videk Harmony program, ver-
sion 4.03 (Videk Corporation, Rochester, NY).
Western Blot Analysis
Total proteins from placental tissues or villous explants
were extracted using 4 volumes of 13 Laemmli buffer (50
mM Tris HCl, pH 6.8, 2% SDS, 10% glycerol) containing
7 M urea, 5 mM DTT, 0.5 mM PMSF, and 1 mg/ml each
of protease inhibitors, leupeptin, aprotinin, and pepstatin.
Tissues were homogenized with a Tekmar Tissuemizer for
562 RAJAKUMAR AND CONRAD
about 1 min at setting 6. The crude homogenate was cen-
trifuged at 10 600 3 g at 48C, and the supernatant was
subjected to mild sonication (microprobe, setting 4 for 5
sec). The supernatant was then stored in aliquots at 2808C.
Protein estimation was carried out using the Bio-Rad re-
agent, and 150 mg protein was used for the Western anal-
ysis. Nuclear proteins from HepG2 cells (510 mg) or from
villous explants (3050 mg) were also used. All prepara-
tions were finally boiled for 5 min and briefly centrifuged.
The proteins were separated on a SDS containing 6% poly-
acrylamide gel (Mini-Protean II cell system, Bio-Rad) at
100 V for one hour. Total protein samples were loaded in
40 ml volume containing the desired amount of protein and
5% b-mercaptoethanol and 0.005% pyronin (Sigma Chem-
ical Co., St. Louis, MO) and nuclear proteins in 2030 ml
volume containing 2% SDS and 5% b-mercaptoethanol.
The proteins were next transferred to PVDF membranes
(Immobilon; Millipore, Bedford, MA) using a semidry
transfer system (Enprotech; Integrated Transfer Systems,
Natick, MA) at a constant current of 1.0 mA per cm2. De-
tection of proteins was carried out after blocking the mem-
branes with a 1% solution of Blocking Reagent (Boehringer
Mannheim, Indianapolis, IN) for 4 h and incubating with
the primary antibody. An anti-HIF-1a monoclonal antibody
and a rabbit polyclonal anti-HIF-2a were obtained from
Novus Biologicals (Littleton, CO; see Immunohistochem-
istry for further details). For HIF-1a, the antibody was di-
luted 1:200 in TBS (Tris-buffered saline) buffer containing
0.1% Tween-20, and for HIF-2a, the antibody was diluted
1:400 in PBS containing 0.1% Tween-20. The correspond-
ing IgG was substituted for the primary antibody at the
same concentration to generate negative control blots. The
membranes were washed in either TBS-T or PBS-T buffer
three times for 10 min each and then were incubated with
alkaline phosphatase-conjugated secondary antibody (1:
2500 dilution for HIF-1 a and 1:5000 for HIF-2 a and b-
actin) for 1 h. The membranes were washed three times in
either TBS-T or PBS-T for 10 min each. They were further
washed in the appropriate buffer without Tween-20 for 10
min and were allowed to equilibrate in alkaline phosphatase
buffer (100 mM Tris, pH 9.5; 150 mM NaCl) for 5 min.
Chemiluminescent detection was carried out using the
CDP-Star substrate (Boehringer Mannheim) diluted 1:200
in the alkaline phosphatase buffer for 5 min. Membranes
were exposed for different times to Kodak Bio-maxAR
film.
The membranes were stripped with a buffer containing
62.5 mM Tris HCl, pH 6.8; 2% SDS; and 100 mM b-ME
at 508C for 1530 min and reprobed with another antibody
of interest or with a monoclonal anti-b-actin antibody
(Clone AC-15, Sigma; 1:1000 dilution) to correct for load-
ing variations.
Immunohistochemistry
Two anti-HIF-1a monoclonal antibodies were used. One,
a generous gift from Dr. Helen Turley, is clone Hif28b (a
mouse IgG1k) directed against amino acids 329530 of the
molecule [39]. The other designated clone H1a67 (a mouse
IgG2b) was purchased from Novus Biologicals and is di-
rected against amino acids 432528. Dr. Turley also pro-
vided a HIF-2a antibodyclone EP190b (a mouse IgG1k)
directed against amino acids 535631 of the molecule [40].
There was no cross-reactivity of the HIF-1a antibodies with
HIF-2a, and vice versa [39] (see Results).
Immediately after extraction from the uterus, placental
tissues were fixed in 10% formalin for at least 24 h and
then dehydrated and embedded in paraffin. Seven-micron
sections were cut and mounted on Fisher Superfrost/Plus
glass slides. After removal of the paraffin, the sections were
subjected to antigen retrieval by using Antigen Unmasking
Solution (Vector Laboratories Inc., Burlingame, CA) and a
microwave oven. After quenching of endogenous peroxi-
dases using 1.0% hydrogen peroxide in methanol for 3 min
and blocking with 1.5% normal horse serum for 20 min,
the tissues were incubated with the monoclonal antibodies
for 2 h at room temperature. Preliminary studies showed
that the optimal concentration for the H1a67 HIF-1a anti-
body was 10 mg/ml. The other antibodies were provided as
cell culture supernatants at 10 mg/ml, and they were applied
directly to the tissue sections. For negative controls, the
IgG2b was substituted for the H1a67 HIF-1a at the same
concentration. For the other antibodies, RPMI cell culture
medium containing 10% FCS, 0.002% sodium azide, and
10 mg/ml of IgG1k was used. Using the Vectastain Elite
ABC Kit and 3,39-diaminobenzidine (DAB) Peroxidase
Substrate Kit (both from Vector), immunoreactive HIF-1a
and -2a were detected. All of the villous placental tissues
depicted in Figure 6 were run in the same immunohisto-
chemical procedure and developed with the DAB peroxi-
dase substrate for identical times (11.5 min). Tissue sec-
tions were lightly counterstained with Hematoxylin Gill 2
(Sigma). After dehydration in ethanol and xylene solutions,
a coverslip was applied using Cytoseal XYL (Stephens Sci-
entific, Riverdale, NJ).
Data Analysis
The data in all graphs are presented as mean or mean 6
SEM. They were analyzed by one- or two-factor univariate
analysis of variance (ANOVA). When appropriate, post hoc
comparisons between individual group means were made
by Fishers protected least-significant difference tests. A P
value of less than 0.05 was considered to be significant
[41].
RESULTS
We first investigated the expression and ontogeny of the
major hypoxia inducible transcription factors in normal hu-
man placentas of different gestational ages. By Northern
analysis, HIF-1a and -2a mRNA were present in normal
placentas at all stages of gestation with greater variability
in expression during early pregnancy (Fig. 1A). Overall,
however, the HIF-1a mRNA levels did not change signif-
icantly with advancing gestational age when normalized ei-
ther for b-actin (Fig. 1B) or 18S RNA (data not shown;
values of both P 5 NS by ANOVA). In contrast, HIF-2a
mRNA showed a significant increase during gestation,
whether normalized for b-actin (Fig. 1B, P 5 0.001) or
18S RNA (data not shown, P , 0.02). Although b-actin
and HIF-1a mRNA did not change significantly with ges-
tational age when normalized for 18S RNA (P 5 NS by
ANOVA), there was a trend for increased levels of both at
58 gestational wks. These observations indicate that both
HIF-1a and -2a mRNA are constitutively expressed in the
human placenta.
A representative example of the expression and ontog-
eny of HIF-1a and -2a protein by Western analysis is de-
picted in Figure 1C. The graphical summary of the data
from all placentas (some tested on duplicate blots and the
values averaged) is presented in Figure 1D. As a positive
control, total protein from a normal-term villous explant
 HIF-1a, -2a, AND -1b IN THE HUMAN PLACENTA 563

FIG. 1. A) Representative Northern blot depicting the expression and ontogeny of HIF-1a and -2a mRNA in the human placenta. Ten micrograms of
total RNA were electrophoresed and Northern analysis performed (see Materials and Methods). The 4.4-kilobase (kb) HIF-1a (top panel), 5.8-kb HIF-
2a (middle panel), and the 1.8 kb b-actin (bottom panel) mRNA that was used to normalize for loading variability are shown. Each lane contains total
RNA from a different placenta. B) Graphical summary of all data is presented in A (mean 6 SEM). The n indicates the number of different placentas
analyzed for each gestational age. There was a significant effect of gestational age on HIF-2a mRNA expression (P 5 0.001 by ANOVA). * P , 0.05
compared with all other gestational ages. Note that comparison of the abundance of mRNA for HIF-1a and -2a cannot be made because the specific
activity of the probes is unknown. C) Representative Western blot depicting the expression and ontogeny of HIF-1a and -2a protein in the human
placenta. Total protein was extracted in 7 M urea containing Laemmli buffer, electrophoresed on SDS containing 6% polyacrylamide gels, and subjected
to Western analysis as described in Materials and Methods. The 120-kDa HIF-1a, 115-kDa HIF-2a, and 43-kDa b-actin proteins are shown. The total
protein from a normal-term villous explant subjected to 46 h of hypoxia represents a positive control. The exposure times for HIF-1a and -2a,
respectively, were 30 and 3 min. Each lane contains total protein from a different placenta. D) Graphical depiction of all of the data for expression of
HIF-1a and -2a protein in the human placenta (mean 6 SEM). The n refers to the number of different placentas tested at each gestational age. There
was a significant effect of gestational age on both HIF-1a and -2a protein expression (P , 0.01 and , 0.005 by ANOVA, respectively). * P , 0.05
compared with 58 wk; 1 P , 0.05 compared with all other gestational ages. Note that comparison of the abundance of protein for HIF-1a and -2a
cannot be made because the relative efficacy of the antibodies is unknown.

exposed to 2% hypoxia in culture for 46 h was also
probed for HIF-1a and -2a (see below). A significant effect
of gestational age was observed for both HIF-1a and HIF-
2a (P , 0.01 and , 0.005 by ANOVA, respectively). The
protein expression was significantly greater at the earlier
gestational stages.
To determine whether hypoxia regulates HIF expression
in the human placenta and, if so, at the level of mRNA,
protein, or both, we subjected villous explants to standard
tissue culture conditions (normoxia) or to an hypoxic en-
vironment (1 or 2% O2; see Materials and Methods). As
shown in Figure 2, A and B, there was no consistent effect
of hypoxia on either HIF-1a or -2a mRNA expression in
villous explants derived from first-trimester placentas of 5,
7, or 12 gestational wk. Additional placental villous ex-
plants from first-trimester placentas were tested. After 6 h
of incubation in normoxic or hypoxic conditions, normal-
ized HIF-1a mRNA was 2.29 and 2.07, 2.18 and 2.10, and
1.97 and 1.63, respectively, for the villous explants from
7-, 9-, and 11-wk placentas. Normalized HIF-2a in the
same tissues was 1.51 and 1.68, 1.55 and 1.83, and 1.50
and 1.39, respectively. Comparable results were obtained
using term villous explants (Fig. 3, A and B), that is, there
was no influence of hypoxia on either HIF-1a or -2a
mRNA expression (P 5 NS by ANOVA). The HepG2 cells
also failed to show any difference in the mRNA expression
between normoxic and hypoxic conditions (Fig. 3A). Thus,
placental HIF expression does not appear to be regulated
by hypoxia at the level of mRNA.
In contrast, the expression of both HIF-1a and -2a pro-
tein is regulated by hypoxia in the human placenta. Figure
4, A and B, depicts the Western analyses using total and
nuclear proteins extracted from villous explants, respec-
tively (see Materials and Methods). Proteins of 120 and
115 kDa representing HIF-1a and -2a, respectively, were
enhanced by 2% hypoxia. Term villous explants that were
exposed to 1% hypoxia showed even further increases in
HIF-1a but not in HIF-2a expression, which was appar-
ently already maximal at 2% O2 (Fig. 4A). By reprobing
the same membrane with an antibody directed against b-
actin, uniform loading was documented. Nuclear extracts
from HepG2 cells were used as a positive control (Fig. 4A).
To establish the DNA-binding activity of HIF, electro-
phoretic mobility shift assays (EMSA) were performed. A
564 RAJAKUMAR AND CONRAD
FIG. 2. A) Expression of HIF-1a and -2a mRNA in first-trimester villous
explants subjected to normoxia (N) or hypoxia (H; 2% O2). Ten mg of total
RNA were electrophoresed, and Northern analysis was performed (see
Materials and Methods). The 4.4-kb HIF-1a (top panel), 5.8-kb HIF-2a
(middle panel), and 1.8-kb b-actin (bottom panel) mRNA are shown. B)
Graphical illustration of the data presented in A (mean values). Note the
inclusion of data from an additional first-trimester (7-wk) placenta. There
was no effect of hypoxia on mRNA expression.
distinct HIF complex that moves above a constitutively ex-
pressed complex was observed in the positive control
HepG2 cells and also in the placental villous explants ex-
posed to hypoxia (Fig. 5, AD). The signal in the placental
villous explants was generally less than the HepG2 cells,
even though we used considerably more villous nuclear ex-
tract in each reaction mixture3040 compared with 5
10 mg. We considered that this discrepancy might be due
to an abundance of CREB/ATF that may compete with
HIF-1a for the labeled oligonucleotide. Therefore, we add-
ed cold AP1/CREB consensus oligonucleotide as a com-
petitor. This maneuver somewhat enhanced the DNA-bind-
ing activity of HIF-1a and decreased the binding activity
of the constitutively expressed band (putatively CREB/
ATF). The HIF complex for the HepG2 cells was frequently
a doublet, whereas the HIF complex observed for the nu-
clear extracts from villous explants generally migrated with
the faster-moving lower band of the HepG2 complex. Spec-
ificity of the HIF DNA binding was established by utilizing
FIG. 3. A) Expression of HIF-1a, -2a, and -1b mRNA in term villous
explants subjected to normoxia (N) or hypoxia (H). Twenty micrograms
of total RNA was electrophoresed and analyzed by Northern analysis (see
Materials and Methods). The 4.4-kb HIF-1a (top panel), 5.8-kb HIF-2a
(second panel), 2.6-kb HIF-1b (third panel), and 1.8-kb b-actin (bottom
panel) mRNA are shown. This blot is representative of identical experi-
ments conducted on four placentas. Northern analysis of total RNA from
HepG2 cells exposed to normoxic and hypoxic conditions for 4 h is also
shown. B) Graphical depiction of the data shown in A (mean 6 SEM).
There was no significant effect of hypoxia on HIF mRNA expression by
ANOVA.
antibodies to HIF-1a and -1b in supershift assays (Fig. 5B).
Substitution of the appropriate IgG control for the HIF-1a
and -1b antibodies at the same concentration did not affect
the mobility of the HIF complex.
The amount of HIF DNA-binding activity varied with
the degree of hypoxia (see Fig. 5C). That is, term villous
explants exposed to 1% O2 demonstrated more HIF DNA-
binding activity than did those incubated with 2% O2. Co-
balt chloride duplicated the hypoxia response in villous ex-
plants maintained under normoxic incubation conditions
(Fig. 5C). The effect of preincubating HepG2 cells and term
villous explants with actinomycin D or cyclohexamide is
shown in Figure 5D. Although actinomycin D had little or
no effect on HIF DNA-binding activity, cyclohexamide
completely abolished the binding. Finally, as a negative
control, SP1 consensus oligonucleotide was employed in
the mobility-shift assay to further determine whether the
 HIF-1a, -2a, AND -1b IN THE HUMAN PLACENTA
FIG. 4. A) Expression of HIF-1a and -2a protein in villous explants ex-
posed to normoxia (N) or hypoxia (H). Total protein (150 mg) that was
extracted in 7 M urea containing Laemmli buffer was separated on SDS
containing 6% polyacrylamide gels and subjected to Western analysis
(see Materials and Methods). A representative blot of three experiments
is shown with the 120-kDa HIF-1a, 115-kDa HIF-2a, and 43-kDa b-actin
protein. The last two lanes contain nuclear protein from HepG2 cells (10
mg per lane) that serves as a positive control. H2%, 2% O2; H1%, 1%
O2; otherwise, H alone denotes 2% O2. B) Nuclear proteins were extract-
ed from HepG2 cells (10 mg per lane) and villous explants (50 mg per
lane), as explained in Materials and Methods and subjected to Western
analysis. Nuclear proteins from 7- and 12-wk and term villous explants
that were exposed to normoxia (N) and hypoxia (H) for 6 h were analyzed
along with the nuclear proteins from the HepG2 cells, the latter serving
as a positive control.
hypoxia response was specific. The ubiquitous transcription
factors SP1 and SP3 present in nuclear extracts of HepG2
cells and term villous explants were not affected by hypoxia
(Fig. 5E).
Cells that are exposed to hypoxia turn on a battery of
genes as an adaptive response in many instances via the
HIF pathway, for instance, glucose transporter-1 (Glut-1)
[11]. Therefore, we measured Glut-1 mRNA. The same
membrane depicted in the Northern blot in Figure 3A was
stripped and reprobed for Glut-1 expression. Although the
HepG2 cells exposed to hypoxia showed some increase in
Glut-1 mRNA, the villous explant response was more
marked, and at each time point, Glut-1 mRNA was clearly
more abundant in the hypoxic tissues (Fig. 5F).
Because placental tissue is comprised of several cell
types, we employed immunohistochemistry to discern lo-
calization of HIF-1a and -2a. Additionally, this technique
provides information regarding cytoplasmic and/or nuclear
expression. All results presented in Figure 6 are represen-
tative of at least three placentas studied during the first
trimester and term. Figure 6, af, depict the results for HIF-
1a in first trimester placentas. This transcription factor was
observed in the syncytiotrophoblast, villous cytotropho-
blast, fetoplacental vascular endothelium, and possibly oth-
er cell types in the villous core. Comparable results were
observed for term villous placenta (Fig. 6, gi), except that
at this gestational stage, the villous cytotrophoblast is rare
565
FIG. 5. A) Electrophoretic mobility shift assays representative of identical
experiments conducted on four placentas employing nuclear extracts
from term villous explants and HepG2 cells (positive control) exposed to
normoxia (N) or hypoxia (H). The identical trends were seen in all ex-
periments. A double-stranded oligonucleotide containing the hypoxia re-
sponse element from the erythropoietin gene was end-labeled with 32P-g
ATP. Approximately 50 000 cpm was mixed with 510 mg and 3040 mg,
respectively, of nuclear proteins from the HepG2 cells and term villous
explants. The protein-DNA complex was separated on a 5% nondenatur-
ing gel in 0.33 TBE. The gels were dried and exposed for autoradiography.
Because the HIF complex observed in villous explant tissue was weaker,
110 ng of a mixture of double-stranded AP1 and CREB was added to
the binding reaction, which moderately enhanced the intensity of the HIF
complex by reducing the CREB/ATF complex. All experiments were there-
fore carried out in the presence of these oligonucleotides (see Materials
and Methods for additional details). NS, Nonspecific band; C, constitu-
tively expressed band; HIF, mobility shift caused by HIF-1a and HIF-1b
heterodimer. B) Supershift assays using HIF-specific antibodies. Electro-
phoretic mobility shift assays were performed using nuclear extracts from
HepG2 cells (positive control) and from 7-wk and term villous explants.
After an initial 15-min incubation, 2 mg of HIF-1a or -1b (data not shown)
antibody was added to the reaction mixture, and the incubation was con-
tinued for another 3060 min on ice. Mouse IgG2b and rabbit IgG were
substituted for the HIF-1a and -1b antibodies, respectively, as negative
controls. The complexes were separated on gel and analyzed by autora-
diography (see Materials and Methods for further details). C) Electropho-
retic mobility shift assays of nuclear extracts prepared from term villous
explants exposed for 6 h to normoxia (N), hypoxia at 2% O2 (H2%),
hypoxia at 1% O2 (H1%), or CoCl2. D) Electrophoretic mobility shift as-
says of nuclear extracts prepared from HepG2 cells and term villous ex-
plants treated with actinomycin D (ACT) or cyclohexamide (CHX) prior
to and during subsequent incubation in normoxic (N) or hypoxic (H) con-
ditions (see Materials and Methods for additional details). E) Electropho-
retic mobility shift assays showing the presence of the ubiquitous tran-
scription factors SP1 and SP3 in the nuclear extracts of HepG2 cells and
term villous explants. The levels of these transcription factors were not
regulated by hypoxia. F) Northern analysis depicting the expression of the
2.8-kb glucose transporter-1 mRNA in HepG2 cells and term villous ex-
plants exposed to normoxic (N) and hypoxic (H) incubation conditions.
The blot from Figure 3A was stripped and reprobed for the Glut-1 ex-
pression shown herefor b-actin, see Figure 3A.
566 RAJAKUMAR AND CONRAD

FIG. 6. Localization of immunoreactive HIF-1a and HIF-2a in the first-trimester and term villous placenta (brown staining). ac) Serial sections of a
7-wk placenta for immunolocalization of HIF-1a using clones H1a67 monoclonal IgG2b and Hif28b monoclonal IgG1k antibodies in a and c, respectively.
Substitution of the primary antibody with mouse IgG2b or IgG1k as a negative control, respectively, is shown in the upper and lower panels of b. df)
Serial sections of a 9.5-wk placenta for immunolocalization of HIF-1a again using the two clones of HIF-1a antibodies in (d) and (f), respectively. The
negative controls using IgG2b and IgG1k are depicted in the upper and lower panels of e. gi) Serial sections of a term placenta for immunolocalization
of HIF-1a, again using the clones H1a67 and Hif28b monoclonal antibodies in g and i, respectively. The negative controls using IgG2b and IgG1k are
shown in the upper and lower panels of h, respectively. jl) Tissue sections of a 9.5-wk (j) and normal-term placenta (l) for immunolocalization of HIF-
2a using the clone EP190b monoclonal IgG1k antibody. Substitution of the primary antibody with mouse IgG1k is shown in the upper and lower panels
of k for the 9.5-wk and term placenta, respectively. The syncytiotrophoblast is denoted by the thin arrow, the villous cytotrophoblast by the arrow head,
and fetoplacental blood vessels by the thick arrow. Original magnification: 3200.
 HIF-1a, -2a, AND -1b IN THE HUMAN PLACENTA
and not easily identified. Significantly, the data were similar
for the two different HIF-1a monoclonal antibodies em-
ployed, although the immunostaining using clone H1a67
was generally more intense (see figure legend and Materi-
als and Methods for further details). Figure 6, jl, portrays
the findings for HIF-2a in the first trimester (Fig. 6j and
the top of Fig. 6k), and term placentas (the bottom of Fig.
6k and Fig. 6l). The localization of HIF-2a was generally
overlapping with HIF-1a. Both nuclear and cytoplasmic lo-
calization are evident for the two transcription factors.
DISCUSSION
Despite the close link between placental hypoxia and
normal placental development and various placental pa-
thologies, relatively little is known about the molecular
mechanisms of oxygen transduction by this organ. Accord-
ingly, our objectives were to investigate the ontogeny of
expression of the hypoxia-inducible transcription factors in
the human placenta, as well as their regulation by oxygen
using villous explant cultures.
A major finding was that HIF-2a mRNA increased sig-
nificantly with gestational age, whereas HIF-1a remained
unchanged as assessed by Northern analysis (Fig. 1, A and
B). An increase in HIF-2a mRNA per cell, an increase in
the number of villous cells expressing the message, or both
could account for the phenomenon. Because HIF-2 a
mRNA may be expressed preferentially in endothelial cells
[13, 14], the development and maturation of the fetopla-
cental vasculature in the villous core with advancing ges-
tational age [4] might explain the finding. Ultimately, in
situ hybridization is needed to resolve this issue.
Another major finding was that the abundance of both
HIF-1a and -2a protein in the human placenta significantly
decreased with gestational age as determined by Western
analysis (Fig. 1, C and D). In fact, there was an inverse
relationship between HIF-2a mRNA and protein expres-
sion, with the former significantly rising and the latter de-
clining with gestational age (compare Fig. 1, B and D).
Because HIF-1a and -2a protein, but not mRNA, are in-
duced in the placenta by hypoxia (see below), the finding
of increased protein expression in the first trimester is con-
sistent with the low intervillous blood flow and physiolog-
ical hypoxia that has been reported during the early stages
of placental development [2, 3]. Interestingly, despite the
establishment of intervillous blood flow after the first tri-
mester, both HIF-1a and -2a proteins were still detected in
second and third trimester placentas. On the one hand, per-
sistent protein expression may be partly a consequence of
the relatively low PO2 measured in the intervillous space
even after the establishment of intervillous blood flow60
torr (vs. arterial PO2 of 75100 torr)which was compa-
rable to the endometrial PO2 [3]. On the other hand, various
growth factors including IGF-I and EGF [21, 22] may con-
tribute to this paradoxical persistence of HIF-1a and -2a
protein expression after the first trimester and possibly to
the elevated levels observed during the first trimester. In-
deed, the expression of IGF-I and -II, as well as of the IGF-
binding proteins by the placenta has been described [42].
Our findings are perhaps reminiscent of those recently re-
ported by Iyer and colleagues that indicate that mouse em-
bryonic stem cells showed relatively high levels of HIF-1a
expression under nonhypoxic conditions [25]. Furthermore,
the presence of HIF-1a and -2a protein in the human pla-
centa may explain the constitutive expression of erythro-
poietin by this organ [43].
In order to evaluate the regulation of HIF-1a and -2a in
567
the human placenta by oxygen, we investigated cultured
villous explants subjected to a 20%, 2%, or 1% oxygen
atmosphere as previously described [31]. A major obser-
vation was that HIF-1a, -1b, and -2a were not regulated
by oxygen at the level of mRNA in either first trimester
(Fig. 2, A and B) or term (Fig. 3, A and B) villous explants,
as assessed by Northern analysis. Nor was regulation of the
HIF message by hypoxia observed in the HepG2 cells (Fig.
3A). Although there may have been some diminution of the
hypoxia-induced HIF-1a DNA-binding activity by actino-
mycin D for the HepG2 cells as determined by EMSA, no
such decrease was observed for the term villous explants
(Fig. 5D). Thus, the studies with actinomycin D and HIF-
1a DNA binding activity are consistent with the Northern
analysis, at least for placental villous explants. Although
some investigators have reported hypoxia modulation of
HIF mRNA, others have not, and the phenomenon may
depend on the tissue or cells being investigated, on incu-
bation conditions, and on the degree of hypoxia ([44], and
see citations in [11], as well as present results).
In contrast, of major significance was that both HIF-1a
and -2a proteins were potently induced by hypoxia in first
trimester and term villous explants, as well as in HepG2
cells (Fig. 4, A and B). Presumably, this event was sec-
ondary to protein stabilization [14, 1618] and increased
translation (e.g., [11, 24]). Our experiments with cycloh-
examide and HIF-1a DNA-binding activity were supportive
of the latter mechanism (Fig. 5D, see below), but because
the translation of all proteins was inhibited by the drug, this
methodological approach was nonspecific and did not ex-
clude an important contribution of protein stabilization. In
addition, more HIF-1a protein was expressed in response
to 1.0% rather than 2.0% hypoxia in term villous explants.
The HIF-1a DNA binding activity on EMSA was also more
intense in term villous explants subjected to the 1.0% hyp-
oxic incubation (Fig. 5C). This inverse relationship between
the degree of hypoxia and the production of HIF-1a protein
and DNA-binding activity, at least down to 0.5% O2, has
been previously described [45]. In contrast, no further in-
crease was observed for the HIF-2a protein, which is also
consistent with previous work suggesting that HIF-2a may
be more active than HIF-1a under less severely hypoxic
conditions [39].
Using EMSA, we documented the corresponding DNA
binding activity of HIF-1a for first-trimester and term vil-
lous explants, as well as for the HepG2 cells (Fig. 5). A
constitutive band was also present that may represent
CREB/ATF DNA-binding activity [46]. The authenticity of
the HIF-1a DNA-binding activity was verified by super-
shifts with antibodies specific for HIF-1a and -1b (Fig. 5B).
The HIF complex was a doublet for the HepG2 cells,
whereas the DNA-binding activity of the villous explants
was a single band that aligned with the lower, faster mo-
bility band of the doublet observed for the HepG2 cells
(best portrayed in Fig. 5D). Maxwell et al. showed that the
upper, slower-mobility band of the HIF/DNA complex was
supershifted by a von Hippel-Lindau (VHL) protein anti-
body [47]. Thus, an interaction between VHL protein and
the HIF/DNA complex may not occur in the human pla-
centa during hypoxic conditions comparable to the treat-
ment with cobalt chloride ([47] and Fig. 5C). Further in-
vestigation is needed to substantiate this possibility.
We were unable to identify HIF-2a-binding activity on
EMSA despite the presence of mRNA and protein in the
villous explants, the latter of which was increased by hyp-
oxia as described above. Indeed, the HIF complex that we
568 RAJAKUMAR AND CONRAD
observed was consistently supershifted by the HIF-1a an-
tibody, and there were no other candidate bands. Clearly,
supershift studies using the HIF-2a antibodies are needed.
To our knowledge, there are no published reports identi-
fying endogenous HIF-2a DNA-binding activity by EMSA.
It is possible that the binding conditions are different, and
further work is needed to clarify this important issue.
Because placental villi are composed of many cell types,
we used immunohistochemistry to determine cellular lo-
calization of HIF-1a and -2a (Fig. 6). The technique also
provided insight into nuclear and/or cytoplasmic localiza-
tion. The major findings for the first trimester and term
placentas were 1) the transcription factors showed overlap-
ping expression in the syncytiotrophoblast, villous cytotro-
phoblast, fetoplacental vasculature, and possibly in other
villous core cells and 2) they demonstrated both nuclear
and cytoplasmic localization. The conclusions on HIF-1 a
may be particularly reliable because comparable results
were obtained with two different HIF-1a monoclonal an-
tibodies.
The induction of glucose transporter-1 by hypoxia has
been previously shown to be mediated by HIF-1 [11]. In-
deed, the activation of HIF by hypoxia in the villous ex-
plants was associated with the induction of Glut-1 mRNA
(Fig. 5F). Clearly, further investigation is required to estab-
lish additional links between the hypoxia-inducible tran-
scription factors and normal placental development, as well
as placental pathologies.
ACKNOWLEDGMENTS
We thank Dr. Helen Turley and her colleagues for their magnanimous
contribution of HIF-1a and -2a antibodies, and we also thank Dr. Gregg
Semenza and Dr. Steven McKnight for their generous gifts of the HIF-1a
and HIF-2a cDNA probes, respectively. The authors also thank Rob Smith
for his assistance in the procurement of placentas and Sue Davis for out-
standing clerical support.
REFERENCES
1. Khong TY, Robertson WB. Spiral artery disease. In: Coulam CG,
Faulk WP, McIntyre JA (eds.), Immunological Obstetrics. New York:
W.W. Norton & Company; 1992: 492501.
2. Carter AM. When is the maternal placental circulation established in
man? Placenta 1997; 18:8387.
3. Rodesch F, Simon P, Donner C, Jauniaux E. Oxygen measurements in
endometrial and trophoblastic tissues during early pregnancy. Obstet
Gynecol 1992; 80:283285.
4. Benirschke K, Kaufmann P. Pathology of the Human Placenta, 3rd
ed. New York: Springer-Verlag; 1995.
5. Genbacev O, Joslin R, Damsky CH, Polliotti BM, Fisher SJ. Hypoxia
alters early gestation human cytotrophoblast differentiation/invasion
in vitro and models the placental defects that occur in preeclampsia.
J Clin Invest 1996; 97:540550.
6. Genbacev O, Zhou Y, Ludlow JW, Fisher SJ. Regulation of human
placental development by oxygen tension. Science 1997; 277:1669
1672.
7. Lindheimer MD, Roberts JM, Cunningham FG. Chesleys Hyperten-
sive Disorders in Pregnancy, 2nd ed. Stamford, CT: Appleton &
Lange; 1999.
8. Conrad KP, Benyo DF. Placental cytokines and the pathogenesis of
preeclampsia. Am J Reprod Immunol 1997; 37:240249.
9. Lim K-H, Zhou Y, Janatpour M, McMaster M, Bass K, Chun S-H,
Fisher SJ. Human cytotrophoblast differentiation/invasion in abnormal
pre-eclampsia. Am J Pathol 1997; 151:18091818.
10. Caniggia I, Grisaru-Gravnosky S, Kuliszewsky M, Post M, Lye SJ.
Inhibition of TGF-b3 restores the invasive capability of extravillous
trophoblasts in preeclamptic pregnancies. J Clin Invest 1999; 103:
16411650.
11. Wenger RH, Gassmann M. HIF-1 and the molecular response to hyp-
oxia in mammals. In: Storey KB (ed.), Environmental Stress and Gene
Regulation. Oxford, UK: BIOS Scientific Publishers, Ltd.; 1999: 25
45.
12. Wang GL, Jiang B-H, Rue EA, Semenza GL. Hypoxia-inducible fac-
tor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular
O2 tension. Proc Natl Acad Sci U S A 1995; 92:55105514.
13. Tian H, McKnight SL, Russell DW. Endothelial PAS domain protein
1 (EPAS1), a transcription factor selectively expressed in endothelial
cells. Genes Dev 1997; 11:7282.
14. Ema M, Taya S, Yokotani N, Sogawa K, Matsuda Y, Fjuii-Kuriyama
Y. A novel bHLH-PAS factor with close sequence similarity to hyp-
oxia-inducible factor 1a regulates the VEGF expression and is poten-
tially involved in lung and vascular development. Proc Natl Acad Sci
U S A 1997; 94:42734278.
15. Zhu H, Bunn HF. Oxygen sensing and signaling: impact on the reg-
ulation of physiologically important genes. Respir Physiol 1999; 115:
239247.
16. Huang LE, Gu J, Schau M, Bunn HF. Regulation of hypoxia-inducible
factor 1a is mediated by an O2-dependent degradation domain via the
ubiquitin-proteasome pathway. Proc Natl Acad Sci U S A 1998; 95:
79877992.
17. Salceda S, Caro J. Hypoxia-inducible factor 1a (HIF-1a) protein is
rapidly degraded by the ubiquitin-proteasome system under normoxic
conditions. J Biol Chem 1997; 272:2264222647.
18. Huang LE, Arany Z, Livingston DM, Bunn HF. Activation of hypoxia-
inducible transcription factor depends primarily upon redox-sensitive
stabilization of its a subunit. J Biol Chem 1996; 271:3225332259.
19. Ema M, Hirota K, Mimura J, Abe H, Yodoi J, Sogawa K, Poellinger
L, Fujii-Kuriyama Y. Molecular mechanisms of transcription activa-
tion by HLF and HIF1a in response to hypoxia: their stabilization and
redox signal-induced interaction with CBP/p300. EMBO J 1999; 18:
19051914.
20. Arany Z, Huang LE, Eckner R, Bhattacharya S, Jiang C, Goldberg
MA, Bunn HF, Livingston DM. An essential role for p300/CBP in the
cellular response to hypoxia. Proc Natl Acad Sci U S A 1996; 93:
1296912973.
21. Zelzer E, Levy Y, Kahana C, Shilo B-Z, Rubinstein M, Cohen B.
Insulin induces transcription of target genes through the hypoxia-in-
ducible factor HIF-1a/ARNT. EMBO J 1998; 17:50855094.
22. Feldser D, Agani F, Iyer NV, Pak B, Ferreira G, Semenza GL. Recip-
rocal positive regulation of hypoxia-inducible factor 1a and insulin-
like growth factor 2. Cancer Res 1999; 59:39153918.
23. Jiang B-H, Agani F, Passaniti A, Semenza GL. V-SRC induces ex-
pression of hypoxia-inducible factor 1 (HIF-1) and transcription of
genes encoding vascular endothelial growth factor and enolase 1: in-
volvement of HIF-1 in tumor progression. Cancer Res 1997; 57:5328
5335.
24. Wang GL, Semenza GL. Desferrioxamine induces erythropoietin gene
expression and hypoxia-inducible factor 1 DNA-binding activity: im-
plications for models of hypoxia signal transduction. Blood 1993; 82:
36103615.
25. Iyer NV, Kotch LE, Agani F, Leung SW, Laughner E, Wenger RH,
Gassmann M, Gearhart JD, Lawler AM, Yu AY, Semenza GL. Cel-
lular and developmental control of O2 homeostasis by hypoxia-induc-
ible factor 1a. Genes Dev 1998; 12:149162.
26. Tian H, Hammer RE, Matsumoto AM, Russell DW, McKnight SL.
The hypoxia-responsive transcription factor EPAS1 is essential for
catecholamine homeostasis and protection against heart failure during
embryonic development. Genes Dev 1998; 12:33203324.
27. Maltepe E, Schmidt JV, Baunoch D, Bradfield CA, Simon MC. Ab-
normal angiogenesis and responses to glucose and oxygen deprivation
in mice lacking the protein ARNT. Nature 1997; 386:403407.
28. Wiener CM, Booth G, Semenza GL. In vivo expression of mRNAs
encoding hypoxia-inducible factor 1. Biochem Biophys Res Commun
1996; 225:485488.
29. Seligman SP, Nishiwaki T, Kadner SS, Dancis J, Finlay TH. Hypoxia
stimulates ecNOS mRNA expression by differentiated human tropho-
blasts. Ann NY Acad Sci 1997; 828:180187.
30. Janatpour MJ, Utset MF, Cross JC, Rossant J, Dong J, Israel MA,
Fisher SJ. A repertoire of differentially expressed transcription factors
that offers insight into mechanisms of human cytotrophoblast differ-
entiation. Dev Genet 1999; 25:146157.
31. Benyo DF, Miles TM, Conrad KP. Hypoxia stimulates cytokine pro-
duction by villous explant from the human placenta. J Clin Endocrinol
Metab 1997; 82:15821588.
32. Wolff M, Fandrey J, Jelkmann W. Microelectrode measurements of
pericellular PO2 in erythropoietin-producing human hepatoma cell
cultures. Am J Physiol 1993; 265:C1266-C1270.
 HIF-1a, -2a, AND -1b IN THE HUMAN PLACENTA
33. Metzen E, Wolff M, Fandrey J, Jelkmann W. Pericellular PO2 and O2
consumption in monolayer cell cultures. Respir Physiol 1995; 100:
101106.
34. Gilson GJ, Mosher MD, Conrad KP. Systemic hemodynamics and
oxygen transport during pregnancy in chronically-instrumented, con-
scious rats. Am J Physiol 1992; 263:H1911-H1918.
35. Maxwell PH, Pugh CW, Ratcliffe PJ. Inducible operation of the eryth-
ropoietin 39 enhancer in multiple cell lines: evidence for a widespread
oxygen-sensing mechanism. Proc Natl Acad Sci U S A 1993; 90:
24232427.
36. Semenza GL, Wang GL. A nuclear factor induced by hypoxia via de
novo protein synthesis binds to the human erythropoietin gene en-
hancer at a site required for transcriptional activation. Mol Cell Biol
1992; 12:54475454.
37. Graven KK, Troxler RF, Kornfeld H, Panchenko MV, Farber HW.
Regulation of endothelial cell glyceraldehyde-3-phosphate dehydro-
genase expression by hypoxia. J Biol Chem 1994; 269:2444624453.
38. Zhong H, Simons JW. Direct comparison of GAPDH, b-actin, cyclo-
philin, and 28S rRNA as internal standard for quantifying RNA levels
under hypoxia. Biochem Biophys Res Commun 1999; 259:523526.
39. Wiesener MS, Turley H, Allen WE, William C, Eckardt K-U, Talks
KL, Wood SM, Gatter KC, Harris AL, Pugh CW, Ratcliffe PH, Max-
well PH. Induction of endothelial PAS domain protein-1 by hypoxia:
characterization and comparison with hypoxia-inducible factor-1a.
Blood 1998; 92:22602268.
40. Talks KL, Turley H, Gatter KC, Maxwell PH, Pugh CW, Radcliffe PJ,
569
Harris AL. The distribution of the hypoxia inducible factors HIF-1a
and HIF-2a in normal and pathological tissues. Am J Pathol 2000;
(in press).
41. Zar JH. Biostatistical Analysis, 2nd ed. Englewood Cliffs, NJ: Pren-
tice Hall; 1984.
42. Han VKM, Bassett N, Walton J, Challis JRG. The expression of in-
sulin-like growth factor (IGF) and IGF-binding protein (IGFBP) genes
in the human placenta and membranes: evidence for IGF-IGFBP in-
teractions at the feto-maternal interface. J Clin Endocrinol Metab
1996; 81:26802693.
43. Conrad KP, Benyo DF, Westerhausen-Larson A, Miles TM. Expression
of erythropoietin by the human placenta. FASEB J 1996;10:760766.
44. Yu AY, Frid MG, Shimoda LA, Wiener CM, Stenmark K, Semenza
GL. Temporal, spatial, and oxygen-regulated expression of hypoxia-
inducible factor-1 in the lung. Am J Physiol 1998; 275:L818L826.
45. Jiang B-H, Semenza GL, Bauer C, Marti HH. Hypoxia-inducible fac-
tor 1 levels vary exponentially over a physiologically relevant range
of O2 tension. Am J Physiol 1996; 271:C1172C1180.
46. Kvietikova I, Wenger RH, Marti HH, Gassmann M. The transcription
factors ATF-1 and CREB-1 bind constituitively to the hypoxia-induc-
ible factor-1 (HIF-1) DNA recognition site. Nucleic Acids Res 1995;
23:45424550.
47. Maxwell PH, Wiesener MS, Chang G-W, Clifford SC, Vaux EC,
Cockman ME, Wykoff CC, Pugh CW, Maher ER, Radcliffe PJ. The
tumour suppressor protein VHL targets hypoxia-inducible factors for
oxygen-dependent proteolysis. Nature 1999; 399:271275.