| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of
* Obstetrics and Gynecology,
Health Research and Policy and
Radiation Oncology, Stanford University Medical Center, California, USA
1Correspondence: Department of Obstetrics and Gynecology, Stanford University Medical Center, 300 Pasteur Dr., Room HH 333, Stanford, CA 94305-5317, USA. E-mail: giudice{at}stanford.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: intrauterine growth restriction • fetal growth • hypoxic response
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. There is increasing evidence that the foundations of lifelong health are built in utero and that long-term health risks of IUGR include hypertension (2)
, dyslipidemia, obesity (3)
, diabetes, precocious adrenarche, and infertility (4)
. Thus, investigation into the genetic and epigenetic response of fetal tissues to hypoxia is paramount in investigating acute and chronic effects of in utero hypoxic exposure and long-term development of disease.
Insulin-like growth factors and their binding proteins are key regulators of fetal growth (5)
. IGF-I is a major growth promoter in the fetus; of six binding proteins, IGFBP-1 sequesters IGF-I and regulates minute-to-minute availability of free IGF-I in the circulation (6
, 7)
. IGFBP-1 in the maternal circulation is elevated in pregnancies complicated by IUGR due to uteroplacental insufficiency (UPI), and IGFBP-1 in human fetal serum is elevated in fetal cord blood in babies with IUGR and UPI. IGFBP-1 levels are inversely correlated with birth weight (8)
and fetal cord pO2 (9)
. IGFBP-1 concentrations in fetal blood collected by cordocentesis in the third trimester are significantly higher in growth-restricted babies than in normal-weight babies (10
, 11)
. In animal models of growth restriction, including uterine artery ligation and maternal hypoxia, elevated maternal and fetal serum IGFBP-1 concentrations are associated with growth restriction (12
13
14)
, as is up-regulation of fetal hepatic IGFBP-1 mRNA expression. Thus, IGFBP-1 regulation in maternal serum and fetal tissues can serve as a marker of hypoxia.
The physiologic response to hypoxia is aimed at maximizing oxygen intake and delivery to vital organs, increasing ventilation, heart rate, cardiac output, blood pressure, growth of capillary beds to supply vital organs, production of red blood cells, hemoglobin levels, and blood volume. Hypoxia can also have sympathoadrenal stimulation and lead to an increase in peripheral resistance. Postnatal hypoxia induces increased respiratory effort and a rise in cardiac performance. Paradoxically, in the fetus acute hypoxia causes inhibition of respiratory movement, bradycardia, hypertension, and a rapid fall in combined ventricular output (15)
. Over the course of 
30 min, heart rate and cardiac output gradually recover (16)
. Thus, the physiologic response to hypoxia differs between prenatal life and postnatal life, suggesting that a fetus can respond to hypoxic insults in a tissue- and time-specific manner.
Many physiologically essential proteins, such as glycolytic enzymes, growth factors, vasoactive peptides and angiogenic peptides, are regulated by the challenge of hypoxia in mammalian tissues and cells in culture (17
18
19)
. Activation of the hypoxia-inducing factor family of transcription factors has been suggested to play an important role in gene regulation during hypoxia (17
, 20)
. In the current study, we have investigated the effects of hypoxia on global gene expression in the developing fetus, using a rat model of in utero hypoxia as well as validation of the model by investigating IGFBP-1 in maternal serum and fetal tissues. The effects of hypoxia on fetal tissue-specific changes in gene expression suggests a mechanism underlying growth inhibition and placental dysfunction in pregnancies accompanied by in utero hypoxia and uteroplacental insufficiency.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Animals
All animal experiments were performed in strict accordance with National Institutes of Health guidelines, and animal protocols were approved by the Administrative Panel on Laboratory Animal Care (A-PLAC) at Stanford University. The experimental design (Fig. 1
) involved timed pregnant Sprague-Dawley rats divided into two groups: those exposed to hypoxia (9.5–10.5% O2) and those exposed to 21% O2 for 3 h, 1 day, or 11 days. Animals exposed to hypoxia showed hyperventilation and reduced activity. Consistent with the results of other investigators, maternal food intake was decreased in the hypoxic group, and the reduction of maternal food consumption associated with hypoxia has been demonstrated not to account for the major component of fetal weight reduction in a well-controlled study (21)
. Cesarean sections were performed at day 20. After dissection from extra-embryonic membranes, seven organs from different embryos of the same pregnant rat were dissected and pooled. These included brain, heart, lung, liver, intestine, kidney, and placenta. Pooled organs were minced and frozen in liquid nitrogen immediately. For in situ hybridization of embryos, 9 day timed pregnant Sprague-Dawley rats were divided into two groups: those exposed in the hypoxia chamber (9.5–10.5% O2) and those exposed to 21% O2 for 7 days. Cesarean sections were performed at day 16. After dissection from extra-embryonic membranes, whole embryos from 16 day dams were fixed in 4% paraformaldehyde (Electron Microscopy Science, Ft. Washington, PA, USA) containing 1x PBS at 4°C for 4 to 8 h before being transferred to 0.5 M sucrose in 1x PBS at 4°C for up to 24 h. Fixed embryos were then frozen in O.C.T. compound (Sakura Finetechnical Co., Torrance, CA, USA) in plastic molds, using liquid nitrogen.
|
Western ligand blot for IGFBP-1, a marker of in utero hypoxia
Five microliters human nonpregnancy serum, 1 µL human amniotic fluid (1:10 dilution), 5 µL nonreduced samples of human pregnancy serum, and 5 µL, 10 µL, or 15 µL pregnant rat serum from normoxic and hypoxic groups were diluted in 5x Laemmli gel loading buffer to a final concentration of 50 mM Tris-HCl, pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, and 100 mM dithiothreitol. Samples were then heated to 100°C for 10 min in a water bath before electrophoresis. Proteins were separated by 12% sodium dodecyl sulfate (SDS)-PAGE and electroblotted overnight onto a 0.1 µm nitrocellulose membrane (Schleicher and Schuell, Inc., Keene, NH, USA). Membranes were blocked in Tris-buffered saline (10 mM Tris-HCl, pH 7.4, 154 mM NaCl, 0.05% Tween 20) containing 10% nonfat dry milk (10% TBS-T) for 1 h. Membranes were then washed twice in Tris-buffered saline containing 1% nonfat milk (1% TBS-T) for 15 min. The filter-immobilized proteins were incubated overnight with 5 x 105 cpm each of 125I IGF-I and 125I IGF-II (Diagnostic Systems Laboratories, Inc., Webster, TX, USA) at 4°C. Membranes were subsequently washed twice in 1% TBS-T for 15 min, air dried, and visualized by autoradiography with exposure to Kodak X-Omat AR film (Eastman Kodak Company, Rochester, NY, USA) at –80°C for up to 4 days. Densitometry was performed to compare the difference between hypoxic and normoxic groups.
In situ hybridization
Eight-micron frozen serial sections of embryos were mounted on Superfrost/Plus slides (Fisher Scientific, Pittsburgh, PA, USA), then immersed in fresh 4% paraformaldehyde at 4°C for 15 min and rinsed with 3x PBS before dehydrating in an ascending series (%/vol) of alcohol. Rat IGFBP-1 cDNA was kindly provided by Dr. L. Murphy at the University of Manitoba, Canada. Riboprobes were labeled with 35S-UTP (Amersham, Piscataway, NJ, USA) using Riboprobe Combination System kit (Promega, Madison, WI, USA).
Tissue sections were first treated with 0.2 M HCl to remove basic proteins, followed by incubation in 2x SSC at 70°C for 30 min. Slides were treated with pronase E to remove protein associated with the mRNA and acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine (pH 8.0) for 10 min. Subsequently, the slides were washed with 1x PBS and dehydrated in an ascending grade (%/vol) of alcohol. Eventually, the tissues sections were incubated overnight at 50°C in a hybridization cocktail containing 50% formamide, 0.06 M NaOH, 10 mM Tris pH 8.0, 5 mM EDTA, 1x Denhardt’s solution, 10% dextran sulfate, 1 mg/mL tRNA, 10 mM DTT, and 105 cpm/µl sense or anti-sense riboprobes. After hybridization, all slides were washed with 5x SSC containing 10 mM DTT at 50°C for 1 h, then dipped into 50% formamide with 2x SSC and 10 mM DDT at 65°C for 30 min. All hybridized slides were rinsed with a solution of 0.5 M NaCl, 10 mM Tris (pH 7.5), and 5 mM EDTA before treatment with RNase A at 37°C for 30 min to digest nonhybridized probes. Slides were later washed with a descending series of SSC (2x, 1x, 0.5x) at room temperature for 15 min each. Finally, slides were washed with 30% ethanol containing 300 mM NH4Ac, then an ascending grade (%/vol) of ethanol, and air dried and coated with NTB2 Kodak autoradiographic emulsion (Eastman Kodak). The NTB-coated slides were stored in the dark for 4 days, then developed in the Kodak D-19 developer and fixer, followed by counterstaining with hematoxylin and 0.125% eosin. After counterstaining, the slides were dehydrated with an ascending grade (%/vol) of ethanol and cleared with xylene before mounting with coverslips using DPX mounting medium (Electron Microscopy Science). Densitometry was used to compare the difference between hypoxic and normoxic groups in fetal tissues.
cRNA preparation
Fifty to 100 mg of each group of pooled organs was homogenized with Polytron PT 1200C homogenizer (Kinematica AG, Littau-Luzern, Switzerland) and used to prepare total RNA. According to the manufacturer’s instruction using TRIzol (Invitrogen Life Technologies, Carlsbad, CA, USA) method, 100 µg of total RNA was cleaned and precipitated using RNeasy Mini Kit (Qiagen, Valencia, CA, USA) to prepare the template for cDNA synthesis. A T7-(dT)24 oligo primer was used to synthesize double-stranded cDNA by the Superscript Choice System (Gibco-Invitrogen), which was subsequently cleaned up by Phase Lock Gels (PLG)-phenol/chloroform extraction and ethanol precipitation. Then ENZO BioArray High Yield RNA Transcript Labeling Kit (T7) (ENZO, Farmingdale, NY, USA) was used to generate biotinylated cRNA. Additional cRNA clean-up was carried out by RNeasy Mini Kit prior to the fragmentation of biotinylated cRNAs with 5x fragmentation buffer (Tris 200 mM, pH 8.1, KOAc 500 mM, MgOAc 150 mM). The chemically fragmented cRNAs were then hybridized on Affymetrix RG_U34A rat chips at the Stanford University School of Medicine Protein and Nucleic Acid (PAN) Facility, screening for 8799 rat genes and ESTs, followed by fluorescence labeling and optical scanning.
Data analysis
Raw data without normalization generated from Affymetrix Microarray Suite 5.0 (MAS 5.0) (Affymetrix, Santa Clara, CA, USA) were analyzed by GeneSpring software 5.1 (Silicon Genetics, Redwood City, CA, USA). Mean values were obtained for gene readouts after normalization to the 50th percentile of the distribution of all measurements in each chip. Per gene normalization was performed using the median value of each gene throughout different chips in the same organ and same experimental conditions. Fold ratios were derived from comparing normalized data between hypoxic and normoxic groups. Genes up- or down-regulated by >1.5-fold by hypoxia were filtered. Nonparametric testing was then applied with P < 0.05 for statistical significance of genes regulated by hypoxia compared with normoxic controls. In addition to GeneSpring, significance analysis of microarrays 1.21 (SAM) software (22)
was used to verify the GeneSpring analysis. The data obtained from optical scanning were analyzed in Affymetrix Microarray Suite 5.0 and normalized by a scaling factor of 500. Normalized data from the same organ and triplicate of the same experimental conditions were imported into a Microsoft Excel spreadsheet. The data were first transformed by cube root. The averages of the cube root of hypoxic or normoxic groups were then calculated before performing the regression analysis. The averages of cube root-transformed data for hypoxic and normoxic groups were used to predict the values in hypoxic and normoxic groups in each experiment. The values derived from regression were then transformed by a power of three, followed by SAM analysis with a cutoff at 1.5 fold. Finally, the results of SAM analysis were compared with the results from GeneSpring analysis. Only genes fulfilling criteria for both analyses were selected for further analysis and validation. Venn diagrams comparing GeneSpring and SAM were generated. A new gene list containing genes shown in GeneSpring analysis and SAM analysis was created in GeneSpring after comparing the results from both analyses. Gene lists in each organ were created based on the average of normalized data of hypoxic vs. normoxic groups with three replicates in three different exposures to hypoxia. Welch ANOVA was applied with P <0.05 for statistical significance. K-means clustering (23)
was used to demonstrate the gene regulatory patterns in each organ. A gene list was created based on each cluster and ontological classification was then used to cluster the genes in each list.
RT-PCR
Total RNA was collected from each group of pooled organs using the TRIzol method. One hundred micrograms of total RNA was cleaned and precipitated using RNeasy Mini Kit (Qiagen) to prepare the template for reverse transcription. RNA integrity was verified by formaldehyde agarose gel electrophoresis with ethidium bromide staining. The reverse transcription (RT) reaction was carried out using Omniscript kit (Qiagen). Each RT reaction contained 2 µg of total RNA, 2 µL 1x buffer RT, 0.5 mM dNTPs, 1 µM T7-(dT)24 oligo primer, and 4 units of Omniscript reverse transcriptase. This 20 µL reaction was immediately incubated at 37°C for 1 h. Two microliters of RT products were subsequently used as templates for PCR containing 5 µL 1x PCR buffer, 0.2 mM dNTPs, 0.5 µM specific primer pairs, 2.5 units of HotStar Taq polymerase using the Eppendorf Mastercycler Gradient (Eppendorf, Hamburg, Germany). After denaturation at 95°C for 15 min, the 50 µL samples were subjected to 28–38 cycles of amplification consisting of denaturation at 94°C for 1 min, annealing at 51.3–67°C for 1 min, and extension at 72°C for 1 min, followed by a final 10 min of extension at 72°C after completion of cycles. In each reaction, 18S rRNA was used as a reaction control. Primer sets were synthesized at the PAN facility in the School of Medicine of Stanford University. The PCR products were then subject to 2% agarose gel electrophoresis and stained with ethidium bromide.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A) from 11 ± 1.18 to 8.33 ± 1.18 (31.5%) (P<0.05), suggesting that hypoxia can lead to pregnancy failure. Fetal weights in the normoxic and hypoxic groups were 4.16 ± 0.02 g and 2.86 ± 0.26 g, respectively (31.25%) (P<0.05) (Fig. 2B
). Placental weight was reduced from 0.7 ± 0.03 g in the normoxic group to 0.49 ± 0.02 g in the hypoxic group (30%) (P<0.05) (Fig. 2C
). Pregnant rats exposed to hypoxic conditions for 11 days were found to have 2.5 ± 0.5 placental resorption sites; placental resorption was not observed in the normoxic group (P<0.05) (Fig. 2D
).
|
IGFBP-1 has been demonstrated to be increased in the maternal circulation under hypoxic conditions as well as in the fetal circulation and liver with in utero hypoxia and IUGR (14
, 24)
. To validate our model for maternal and in utero fetal hypoxia, IGFBP-1 levels in maternal serum and IGFBP-1 mRNA expression in fetal organs were investigated. A significant 4.39-fold increase in IGFBP-1 levels in maternal serum was found under hypoxic vs. normoxic conditions (Fig. 3
A). In situ hybridization using embryos from pregnant rats exposed to hypoxia from day 9 to day 16 revealed robust induction of IGFBP-1 expression in heart and liver from hypoxia-treated animals (Fig. 3B
). The densitometry of the pictures taken from slides showed that IGFBP-1 in fetal heart and liver was induced by hypoxia by 18.31- and 1.68-fold, respectively. Expression of IGFBP-1 in brain, lung, and kidney did not exhibit significant differences (data not shown).
|
Genes commonly up- or down-regulated
To assure the reliability of the data analysis of hypoxia-regulated genes in fetal tissues, two independent data analysis algorithms (GeneSpring 5.1 and SAM 1.21) were used. Raw data generated from Affymetrix MAS 5.0 were imported to GeneSpring 5.1 software in a tab-delimited format, followed by per chip and per gene normalization, as described in Material and Methods. Separate gene lists were created for induced genes and repressed genes in each organ after filtering for 1.5-fold change and nonparametric testing. The genes regulated by hypoxia in selected fetal organs and placenta are mostly either acute responses at 3 h or chronic responses at 11 days, except in the lung, which contains more regulated genes at 1 day.
The same batch of .cel files used for GeneSpring analysis was also used in the SAM analysis. Triplicates including normoxic and hypoxic groups of the same organ with the same exposure to hypoxia were imported into Microsoft Excel after normalization in MAS 5.0, and an unpaired two-class comparison was performed after transformation of the readouts from normoxic and hypoxic experimental groups. Genes with significant changes were selected with the cutoff of a 1.5-fold difference.
SAM analysis is more sensitive than GeneSpring analysis, whereas GeneSpring is more stringent. Although SAM analysis detected more genes than GeneSpring analysis, 89.16% of the genes detected by the GeneSpring analysis were also included in the SAM analysis. Gene lists were created accordingly. Genes common to both algorithms were selected for further analyses by examining individual gene lists to identify genes potentially involved in hypoxic responsiveness or by a global approach including k-means clustering and ontological classification (vide infra) to recognize coregulated genes and genes with related functions or pathway signatures (25)
.
Gene regulation
K-means clustering analysis was performed on the log-transformed datasets to investigate the dynamic changes in genes and gene families using the ANOVA test, assuming equal variances (Fig. 4
). Set 1 genes were composed of genes up-regulated 1.5-fold at 3 h that exhibited profound down-regulation at 1 day, returning to baseline at 11 days. Set 2 genes showed a pattern similar to set 1 at 3 h and 11 days, although these genes demonstrated marked up-regulation at 1 day. Genes in set 3 exhibited varied regulation before their marked decline at 11 days. Set 4 genes in all organs except intestine were induced or suppressed at 3 h, then fluctuated at 1 day, and were strongly up-regulated at 11 days of exposure to hypoxia compared with normoxic controls at the same gestational age. Set 4 genes in intestine and set 5 genes in lung, liver, and kidney were up-regulated at 3 h and fluctuated at 1 day before leveling off at 11 days. Set 5 genes in heart and set 6 genes in lung were markedly down-regulated at 3 h, then accumulated after 1 day exposure to hypoxia before being down-regulated again at 11 days. The number of regulatory patterns varied in different organs. Patterns 1, 2, and 3 were shared by all organs.
|
Although all organs shared three similar regulatory patterns, the genes included in a given pattern set were tissue-specific. For example, brain and lung shared only one common gene, Cdkn1a, in pattern 2. Each pattern of k-means clustering was then subjected to ontological classification. Ontological analysis revealed that glycolytic enzymes were mainly up-regulated by hypoxia after 3 h or 1 day exposure, with hexokinase, phosphofructokinase, and aldolase A being up-regulated in most tissues (Table 1
). Genes related to ion transport, calcium homeostasis (Table 2
) and the cytoskeleton were consistently up-regulated regardless of tissue type and duration of exposure to hypoxia. In contrast, genes related to G-protein signaling, growth (Table 3
), different groups of ion transport, and cytoskeletal genes were consistently down-regulated regardless of tissue type and duration of exposure to hypoxia. Several inflammation-related genes were regulated with long-term (11 days) exposure in placenta (Table 4
). Although the same functional category was observed in different organs and/or time points, members included in each gene category were tissue-specific. For example, the category "enzyme" has been included in pattern 1 in brain (Ugcg, Ubiquinone-1ß, Bleomycin hydrolase, HKII, Kynu, SI) and lung (Idi1, Sult-n, Dlgap1, Hri), but genes included in this category varied between these two tissues. The same phenomenon can be found in any category among different organs and regulatory patterns. All gene expression data, cluster groups, and functional categories of the dynamically regulated genes have been placed on our microarray database web server (http://171.65.6.67/microarrayandhypoxia/welcome.htm).
|
|
|
|
RT-PCR
The reliability of the microarray quantitative data was independently confirmed by RT-PCR using the same total RNA samples as in the microarray experiments. Genes common to GeneSpring and SAM analyses in each gene list, either up-regulated or down-regulated, were randomly selected for validation. Specific primer sets were designed accordingly to each selected gene for RT-PCR. The sequences and product sizes of these primer sets are shown in Table 5
. As shown in Fig. 5
, regulation of genes determined by RT-PCR was consistent with the microarray analyses, including tissue-specific expression of some genes. For instance, Cyp7a1 is suppressed by hypoxia in lung after 3 h of exposure to hypoxia, whereas it is induced in heart after 1 day exposure. The regulation of a gene can be time-specific in the same organ. For example, Cdk2 was induced by hypoxia in brain after 1 day of exposure whereas the expression of this enzyme is suppressed after 11 days of exposure.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 27)
. The role of oxygen in fetal growth is underscored by observations that IUGR is seen frequently in pregnancies complicated by decreased maternal oxygen-carrying capacity (e.g., maternal cyanotic heart disease and anemia) as well as localized hypoxia due to poor uterine perfusion (28)
. The role of oxygen in fetal growth is further underscored by observations of significant fetal growth restriction and increased fetal and neonatal morbidity and mortality in pregnancies accompanied by fetal anemia and hypoxia (e.g., Rh isoimmunization). In the current study, we have found that chronic, but not acute, exposure to hypoxia results in significant fetal growth restriction, consistent with clinical observations in humans (29
, 30)
and models of in utero hypoxia (14)
and uterine artery ligation (31)
. In addition, chronic hypoxic stress results in decreased litter size and pregnancy failure (Fig. 2)
. The impediment in maintaining pregnancy is unlikely due to failure in implantation, because the dams were exposed to hypoxia beginning on day 9 of gestation. Whether the reductions in litter size and fetal weight are caused by impaired placental development remains to be determined.
IGFBPs regulate minute-to-minute bioavailability of IGFs in the circulation and modulate IGFs actions (32)
. IGFBP-1 is an inhibitory factor for fetal growth and expressed mainly in the fetal liver. In humans and animals, IGFBP-1 levels in maternal and fetal circulation correlate negatively with fetal size, especially in fetuses with profound IUGR (8
, 33
, 34)
. Although liver expression of the rat IGFBP-1 gene has been characterized between day 16 in utero and 16 days postnatal (35)
, in the current study fetal IGFBP-1 mRNA expression in various organs was determined when pregnant rats were exposed to hypoxic conditions from day 9 to day 16. Hypoxia increased IGFBP-1 expression in fetal heart and liver by 18.31-fold and 1.68-fold, respectively. The impressive induction of IGFBP-1 in fetal heart is a novel finding and raises the possibility of its role in limiting IGF-1 action or perhaps direct action (36)
in cardiac myocytes to accommodate the cardiac changes that accompany adaptation to hypoxia in utero. We have demonstrated 4.39-fold increase in IGFBP-1 levels in maternal serum at day 20 after 11 days of exposure to hypoxia compared with normoxic controls, validating the hypoxic environment of the dams and fetuses in the experimental paradigm used in the gene expression studies. IGFBP-1 expression was up-regulated by hypoxia in heart and liver in the in situ hybridization experiments, although it was not included in the gene lists in the microarray study. Data from the SAM and GeneSpring analyses in heart and liver demonstrate up-regulation of IGFBP-1 in the SAM analysis, but it was excluded in the GeneSpring analysis due to marginal P values. Thus, it was excluded in the gene list showing common genes from both analyses.
Hypoxia triggers a series of systemic, cellular, and metabolic responses that allow tissues to adapt to the damaging effects of the lack of oxygen (37
, 38)
. However, hypoxia can also induce gene expression that is detrimental to an organism (39)
. These responses can be acute or have more long-term effects. Central to these effects can be direct regulation of a single gene or the regulation of a number of genes. Despite great advances in research in gene regulation by hypoxia in adult mammals and cell lines, molecular mechanisms underlying fetal development in response to hypoxia remain poorly understood. In this study, we have shown that the number of gene regulatory patterns varies among different organs (Fig. 4)
. Although similar regulatory patterns of sets 1, 2, and 3 exist in all seven organs investigated, the majority of genes included in the same pattern in different organs are distinct. Kinases, cytoskeleton, molecules associated with calcium homeostasis, and some ion transporters are up-regulated by hypoxia in a tissue-specific manner. Many enzymes involved in glycolysis have been identified as hypoxia-responsive genes (19)
. Although an organism responds to hypoxia by introducing anaerobic glycolysis as the predominant pathway for cellular ATP production, enhanced lactate production with its consequential acidosis potentially limits the source of ATP despite compensated glucose supply via increased glucose transport. In our results, glycolytic enzymes are stimulated mainly by hypoxia in the early stages of hypoxic exposure (3 h, 1 day). This suggests that increased glycolysis is an acute response in the fetus exposed to hypoxia, likely compensating for energy deprivation.
In contrast to glycolytic enzymes, genes related to the cytoskeleton, ion transport, G-protein-dependent signaling, and growth are down-regulated by hypoxia in all tissues examined. In spite of some growth-related genes up-regulated at a given time, genes related to growth are consistently inhibited in all organs, suggesting that the eventual fate of an animal under hypoxic conditions is the restriction of growth.
During gestation, calcium transport in the placenta represents the primary site of regulating fetal calcium homeostasis (40)
. but little is known about the exact mechanism of transport. Evidence implicates that some developmentally expressed cytosolic calcium binding proteins have a crucial role in modulating or shuttling cytosolic calcium, since they are bestowed with a high affinity for calcium (41)
. An elevation in intracellular calcium, often acting via the calcium receptor protein calmodulin, is a ubiquitous signal that regulates a number of critical cellular functions, including activation of transcription factors, DNA synthesis and repair, cell cycle regulation, nuclear envelop breakdown, and apoptosis (42)
. Calcium homeostasis can be modulated through diverse molecules, including channels, transporters, binding proteins, and hormones. Neuronal brain damage induced by hypoxia has been suggested to be mediated by excessive calcium influx through excitatory amino acid receptors (43
44
45)
and voltage-gated calcium channels. Glutamate receptors can act as either calcium channels or transducers, signaling through a G-protein dependent pathway. In the latter, glutamate receptors are coupled to phospholipase C and signal via its downstream cascade (46)
. According to the results herein, several glutamate receptors are regulated by hypoxia in a tissue-specific manner. In addition to glutamate receptors, we also found a variety of calcium homeostasis-regulating molecules to be up-regulated in response to hypoxia in a tissue-specific manner. These molecules include calcium channels (ROB3, Cacna1d, LOC24621, Cacna1c, Cacna1 s, Cacnb4, Cacna1b), calcium transporters (Slc24a2, PMCA3, Atp2b1), 5 HT receptors (Htr3a, Htr5b, 5Ht-2, Htr1b, Htr4), neuropeptide receptors (Npy5r), calcium binding proteins (Calb1, Chp, Parvalbumin, S-100 
, S-100 ß), IP3 receptors (InsP3R, Itpr1, Itpr2), calcium-decreasing factor (preprocaldecrin), and hormones (VIP, calcitonin, parathyroid hormone). Together, these results suggest that calcium homeostasis plays an important role during organogenesis under hypoxic conditions.
Another interesting finding is the observation of up-regulation in placenta of a substantial number of genes associated with the inflammatory response, primarily after 11 days of hypoxic exposure. In preliminary RT-PCR studies (data not shown) we have found that COX-2 and thromboxane A2 synthase, involved in the eicosanoid pathway, are up-regulated by hypoxia in placenta after 3 h and 1 day exposure. An imbalance in the eicosanoid pathway between synthesis of thromboxane A2 and PGI2 has been postulated to be important during the development of pre-eclampsia (47
, 48)
, a disorder associated with placental hypoxia. Clinically, pre-eclampsia is a leading cause of maternal morbidity, IUGR, and fetal mortality (49)
. Deficient trophoblast invasion during placental development is believed to be a cause of pre-eclampsia (50)
, and complex interactions between the endocrine and immune systems are believed to contribute to its pathogenesis. The placenta plays a key role in the development of pre-eclampsia as the symptoms disappear soon after birth or termination of pregnancy, when the placenta is no longer present. Proinflammatory molecules, including cytokines, have been revealed to be responsible for the general maternal inflammatory response observed in pre-eclampsia (51)
. These cytokines are also known to induce lipid peroxidation and free radical production. Lipid peroxides and the product of free radical-NO reaction, peroxynitrite, can selectively inhibit prostacyclin synthesis while the production of thromboxane A2 is further stimulated. Taken together, the results in the current study suggest that hypoxia may cause placental dysfunction and may subsequently induce pre-eclampsia (in humans) through an inflammatory response, leading to IUGR or cessation of fetal development. Whether the induction of inflammatory genes is a direct effect of hypoxia on the placenta or an indirect effect via other embryonic tissues, such as the yolk sac or embryo per se, warrants further investigation since 11 day hypoxia exposure began at the gestational day 9, when the placenta is not yet fully developed.
In summary, global gene profiling of fetal tissues in a rat model of in utero hypoxia has revealed several compelling mechanisms for the hypoxic response during fetal development. We have demonstrated for the first time that an increase in glycolysis is an acute response to hypoxia and that calcium homeostasis is dysregulated in a global fashion; in addition, regulation of individual genes in response to hypoxia during organogenesis is tissue-specific. We have also shown the up-regulation of inflammation-related genes in the placenta in response to hypoxia. Taken together, we hypothesize that hypoxia-induced IUGR is mediated by a dysfunctional placenta and by tissue-specific pathways, including metabolism, growth, and calcium homeostasis.
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 18, 2003. Accepted for publication May 5, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Ream, A. M. Ray, R. Chandra, and D. M. Chikaraishi Early fetal hypoxia leads to growth restriction and myocardial thinning Am J Physiol Regulatory Integrative Comp Physiol, August 1, 2008; 295(2): R583 - R595. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. C. Wendler, S. Amatya, C. McClaskey, S. Ghatpande, B. B. Fredholm, and S. A. Rivkees A1 adenosine receptors play an essential role in protecting the embryo against hypoxia PNAS, June 5, 2007; 104(23): 9697 - 9702. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. P. Gheorghe, S. Mohan, K. C. Oberg, and L. D. Longo Gene Expression Patterns in the Hypoxic Murine Placenta: A Role in Epigenesis? Reproductive Sciences, April 1, 2007; 14(3): 223 - 233. [Abstract] [PDF]
|
|
|
|
 |
|
 K. Shankar, M. Hidestrand, X. Liu, R. Xiao, C. M. Skinner, F. A. Simmen, T. M. Badger, and M. J. J. Ronis Physiologic and Genomic Analyses of Nutrition-Ethanol Interactions During Gestation: Implications for Fetal Ethanol Toxicity Experimental Biology and Medicine, September 1, 2006; 231(8): 1379 - 1397. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kajimura, K. Aida, and C. Duan Understanding Hypoxia-Induced Gene Expression in Early Development: In Vitro and In Vivo Analysis of Hypoxia-Inducible Factor 1-Regulated Zebra Fish Insulin-Like Growth Factor Binding Protein 1 Gene Expression Mol. Cell. Biol., February 1, 2006; 26(3): 1142 - 1155. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
