Introduction
Associated with adverse outcomes spanning fetal to adult life, fetal growth restriction (FGR) constitutes a major obstetrical complication affecting 5-7% of all pregnancies. Perinatal mortality is substantially high in FGR and approximately 1 in 4 of the stillborn fetuses is growth restricted. Perinatal morbidities include asphyxia, preterm delivery, neonatal depression, and a spectrum of metabolic, respiratory, and neurological complications. Long-term risks include continuing growth deficit, cerebral palsy, and neurodevelopmental abnormalities. Beyond the perinatal period and infancy, adverse consequences of FGR extend into adult life. Experimental and epidemiological evidence indicates that chronic intrauterine deprivation may induce epigenetic programming and a thrifty phenotype setting the stage for subsequent development of adult illnesses including hypertension, stroke, ischemic heart disease, type 2 diabetes, and central obesity. Constrained fetal growth thus constitutes a major health care concern.
Compromised fetal supply line has long been proposed as a major underlying mechanism for limiting fetal growth. Consistent with this concept, Doppler ultrasound studies have shown increasing umbilical circulatory impedance associated with progressive fetal decompensation in FGR, eventually leading to absent end-diastolic flow (AEDF) in the umbilical artery. This hemodynamic deterioration is associated with a spectrum of adverse perinatal outcomes including stillbirths and perinatal asphyxia. There is morphological evidence, moreover, correlating umbilical artery AEDF to diminished fetoplacental vascular branching in FGR pregnancies. The molecular mechanisms underlying aberrant fetoplacental angiogenesis in FGR, however, require further elucidation.
Angiogenesis is a complex biological process controlled by agonists and antagonists directly or indirectly promoting or inhibiting angiogenic activity. Normal pregnancy represents a balanced angiogenic state. Current evidence, however, indicates the presence of an antiangiogenic state in several pregnancy complications. Maynard et al first proposed an antiangiogenic state in preeclampsia involving up-regulation of soluble fms-like tyrosine kinase-1 of placental origin in maternal plasma, which opposes proangiogenic vascular endothelial growth factor (VEGF) and placental growth factor (PlGF). Subsequently, the Romero group has provided extensive evidence for the presence of an antiangiogenic state in maternal circulation in several pregnancy disorders including preeclampsia, small-for-gestational-age (SGA) births, fetal death, and preterm labor, and in association with placental massive perivillous fibrin deposition, a condition known to be associated with recurrent miscarriage and stillbirths. Others also have documented an antiangiogenic state in pregnancies resulting from in vitro fertilization, obesity, twin transfusion syndrome, and invasive placentation. Moreover, there is evidence that antiangiogenic agents in maternal serum may serve as potential biomarkers for subsequent recognition of fetal growth compromise and adverse pregnancy outcomes.
These angiogenic mediators in maternal circulation are predominantly of placental origin. Information, however, on the role of specific proangiogenic and antiangiogenic mechanisms operating at the placental level remains limited. Elucidation of these placenta-specific angiogenic mechanisms will not only extend our understanding of the causal pathway for restricted fetal growth but may also lead to the development of biomarkers for FGR that may allow early recognition of FGR and differentiation of constitutionally small fetuses from those that are truly growth restricted. The latter currently remains challenging despite advances in fetal sonography. Such mechanistic understanding may also lead to the emergence of new management strategies for promoting fetal growth.
The purpose of this study, therefore, was to investigate the placental angiogenic mechanisms in FGR pregnancies complicated with fetal hemodynamic compromise as evidenced by umbilical artery AEDF. Specifically, we determined in these pregnancies placental expressions of angiogenic genes utilizing oligonucleotide microarray analysis, and to confirm the significant array findings of differentially expressed messenger RNA (mRNA) transcripts by quantitative real-time polymerase chain reaction (qPCR), Western Blot analysis of protein expression, and immunohistochemistry for localization and quantification in placental tissues.
Materials and Methods
Study design
In this prospective study, angiogenic gene expression was analyzed on placental samples collected from pregnancies complicated with FGR and umbilical artery Doppler AEDF (study group, n = 7), and from uncomplicated pregnancies (control group, n = 7). FGR was defined as an ultrasound-estimated fetal weight <10th percentile for gestational age following the American Congress of Obstetricians and Gynecologists (ACOG) guidelines. We followed these guidelines to restrict the terms “fetal growth restriction” to the fetus and “small for gestational age” to the neonate. Pregnant mothers receiving prenatal care were approached for consent and participation in the study. The inclusion criteria for the study group were: singleton pregnancies, gestational age ascertained according to the ACOG guidelines, gestational age >36 weeks, ultrasound biometric diagnosis of FGR, AEDF in the umbilical artery Doppler, and delivery by cesarean delivery. This mode of delivery was chosen to minimize possible placental oxidative stress from labor and delivery as demonstrated by others. The exclusion criteria were labor; pregnancy complications other than FGR such as preeclampsia, multiple gestation, fetal aneuploidy, fetal malformations, prolonged rupture of membranes, and chorioamnionitis; placental pathology such as abruption, placenta previa, and placental accreta; and maternal diseases such as infection, diabetes, and chronic hypertension. The control group included uncomplicated pregnancies, matched by gestational age with the study group, and delivered by elective cesarean delivery before labor indicated by previous cesarean delivery or breech presentation. The birthweight centile was determined utilizing the US national reference values as reported by Oken et al. The institutional review boards of Truman Medical Center/University of Missouri–Kansas City School of Medicine and Winthrop University Hospital approved the study. Informed consent was obtained from each patient according to the institutional review board protocol.
Placental sample acquisition
Within 10 minutes after the placental delivery, placental tissue samples were collected using a sterile scalpel. Three samples each measuring approximately 2 × 2 × 2 cm were removed from a mediobasal location as defined by Wyatt et al. Each sample was then divided into 3 identical pieces, rinsed thoroughly in sterile phosphate buffer solution, and snap-frozen in liquid nitrogen. The samples were stored in a biorepository according to the institutional regulations.
Angiogenesis gene array analysis
Two angiogenesis-oriented arrays were used for gene expression analysis. Angiogenic SuperArray HS-009 contained 96 genes and 16 internal controls and angiogenic SuperArray OHS-024 contained 114 genes and 14 internal controls (SuperArray Biosciences Corp, Frederick, MD).
The total RNA was extracted from placental tissues using Trizol (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. RNA samples (3 μg) were then reverse transcribed and labeled with biotin-16-dUTP (Roche, Nutley, NJ) by polymerase chain reaction (PCR) using the AmpoLabeling-LPR kit (SuperArray Biosciences Corp). Then, probes were hybridized to either GEArray HS-009 or GEArray OHS-024. The hybridization signals were generated by GEArray chemiluminescent detection kit (SuperArray Biosciences Corp) and evaluated by the Kodak Image Station 4000R and the ChemiDoc XRS (BioRad Laboratories, Hercules, CA), respectively. The relative expression levels of angiogenesis gene were normalized to a panel of housekeeping genes of the corresponding array and cross-linked using gene symbols.
qPCR analysis
qPCR analysis was used to confirm the significant angiogenic findings of the differentially expressed mRNA transcripts (SuperArray Biosciences Corp). Total RNA (1.5 μg/sample) was reverse transcribed using the First Strand cDNA synthesis kit according to manufacturer’s instructions (Roche). The primers were designed using the Primer3 software (MIT) for sense and antisense and purchased from Origene Technologies (Rockville, MD). The sequences of forward and reverse primers were: human GAPDH (GenBank accession no. NM_002046 ) 5′ CTCTCTGCTCCTCCTGTTCGAC 3′ and 5′ TGAGCGATGTGGCTCGGCT 3′ and human neuropilin (NRP)-1 (GenBank accession no. NM_001024628 ) 5′ ACG ATG AAT GTG GCG ATA CT 3′ and 5′ AGT GCA TTC AAG GCT GTT GG 3′. Real-time qPCR was performed by using Fast SYBR Green on a StepOne Plus real-time PCR system (Applied Biosystems, Grand Island, NY). Relative expression values were calculated using the 2-delta delta Ct method and were normalized against reference gene GAPDH. In these calculations we accounted for the PCR efficiency of the individual PCR reactions, calculated on the basis of linear regression as described elsewhere. The specificity of amplification was confirmed by evaluation of the melting curve.
Protein extraction and Western blot
The protein was extracted from placental samples and its concentration was measured by the BCA protein assay method per manufacturer’s protocol (Pierce, Rockford, IL). Each sample (50 μg) was separated in 10% sodium dodecyl sulfate-polyacrylamide gel. The protein was transferred onto nitrocellulose membrane. The membrane was blocked with blocking serum (Thermoscientific, Waltham, MA) followed by incubation with mouse antihuman NRP-1 antibody (1:1000) (Abcam, Cambridge, MA) overnight at 4°C. After washing with Tris-NaCl-tween 20 buffer, the membrane was incubated with antimouse IgG (1:15000) (Abcam) at room temperature for 1 hour. Immunoreactivity was detected using an enhanced chemiluminescence Western blotting system (Thermoscientific). Qualitative analysis was performed and expressed in relation to β-actin.
Immunohistochemical localization and quantification
The placental tissues were fixed in 4% buffered formaldehyde solution, dehydrated, and embedded in paraffin. The 4-μm thickness sections were transferred onto poly-l-lysin-coated slides, deparaffinized, rehydrated, and immunostained with NRP-1 antibody (1:200) using the Vectastain Immpress Reagent Kit (Vector Laboratories, Burlingame, CA). Two investigators blinded to the sample source independently graded immunostain intensity. Slides were first examined at ×4 magnification to identify NRP-1 immunopositive regions in placental sections. In each section, 5 different areas with 8-12 villi per area were selected at random and were evaluated microscopically with a ×40 objective magnification. All sections were scored in a semiquantitative fashion as described by Hsu et al, which considered both the intensity and percentage of cells staining. Intensities were classified as 0 (no staining), +1 (weak staining), +2 (moderate staining), and +3 (very strong staining).
Statistics
Angiogenic data (SuperArray Biosciences Corp) were normalized to corresponding internal controls and resulting gene expression values were cross-referenced between arrays. The Student t test was applied to 67 genes that were common to both SuperArrays and genes with P < .05 considered significant. The final candidate selection was done by filtering for the Bonferroni correction for multiple comparisons ( P < .0007). The values from qPCR were adjusted to GAPDH, and the values from Western blot were adjusted for β-actin, then mean and SD values were determined. The immunostain intensity met the distributional assumptions of a parametric statistical test. Independent groups t test was used to compare the control and FGR groups for these values. Differences were considered significant when P was <.05.
Materials and Methods
Study design
In this prospective study, angiogenic gene expression was analyzed on placental samples collected from pregnancies complicated with FGR and umbilical artery Doppler AEDF (study group, n = 7), and from uncomplicated pregnancies (control group, n = 7). FGR was defined as an ultrasound-estimated fetal weight <10th percentile for gestational age following the American Congress of Obstetricians and Gynecologists (ACOG) guidelines. We followed these guidelines to restrict the terms “fetal growth restriction” to the fetus and “small for gestational age” to the neonate. Pregnant mothers receiving prenatal care were approached for consent and participation in the study. The inclusion criteria for the study group were: singleton pregnancies, gestational age ascertained according to the ACOG guidelines, gestational age >36 weeks, ultrasound biometric diagnosis of FGR, AEDF in the umbilical artery Doppler, and delivery by cesarean delivery. This mode of delivery was chosen to minimize possible placental oxidative stress from labor and delivery as demonstrated by others. The exclusion criteria were labor; pregnancy complications other than FGR such as preeclampsia, multiple gestation, fetal aneuploidy, fetal malformations, prolonged rupture of membranes, and chorioamnionitis; placental pathology such as abruption, placenta previa, and placental accreta; and maternal diseases such as infection, diabetes, and chronic hypertension. The control group included uncomplicated pregnancies, matched by gestational age with the study group, and delivered by elective cesarean delivery before labor indicated by previous cesarean delivery or breech presentation. The birthweight centile was determined utilizing the US national reference values as reported by Oken et al. The institutional review boards of Truman Medical Center/University of Missouri–Kansas City School of Medicine and Winthrop University Hospital approved the study. Informed consent was obtained from each patient according to the institutional review board protocol.
Placental sample acquisition
Within 10 minutes after the placental delivery, placental tissue samples were collected using a sterile scalpel. Three samples each measuring approximately 2 × 2 × 2 cm were removed from a mediobasal location as defined by Wyatt et al. Each sample was then divided into 3 identical pieces, rinsed thoroughly in sterile phosphate buffer solution, and snap-frozen in liquid nitrogen. The samples were stored in a biorepository according to the institutional regulations.
Angiogenesis gene array analysis
Two angiogenesis-oriented arrays were used for gene expression analysis. Angiogenic SuperArray HS-009 contained 96 genes and 16 internal controls and angiogenic SuperArray OHS-024 contained 114 genes and 14 internal controls (SuperArray Biosciences Corp, Frederick, MD).
The total RNA was extracted from placental tissues using Trizol (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. RNA samples (3 μg) were then reverse transcribed and labeled with biotin-16-dUTP (Roche, Nutley, NJ) by polymerase chain reaction (PCR) using the AmpoLabeling-LPR kit (SuperArray Biosciences Corp). Then, probes were hybridized to either GEArray HS-009 or GEArray OHS-024. The hybridization signals were generated by GEArray chemiluminescent detection kit (SuperArray Biosciences Corp) and evaluated by the Kodak Image Station 4000R and the ChemiDoc XRS (BioRad Laboratories, Hercules, CA), respectively. The relative expression levels of angiogenesis gene were normalized to a panel of housekeeping genes of the corresponding array and cross-linked using gene symbols.
qPCR analysis
qPCR analysis was used to confirm the significant angiogenic findings of the differentially expressed mRNA transcripts (SuperArray Biosciences Corp). Total RNA (1.5 μg/sample) was reverse transcribed using the First Strand cDNA synthesis kit according to manufacturer’s instructions (Roche). The primers were designed using the Primer3 software (MIT) for sense and antisense and purchased from Origene Technologies (Rockville, MD). The sequences of forward and reverse primers were: human GAPDH (GenBank accession no. NM_002046 ) 5′ CTCTCTGCTCCTCCTGTTCGAC 3′ and 5′ TGAGCGATGTGGCTCGGCT 3′ and human neuropilin (NRP)-1 (GenBank accession no. NM_001024628 ) 5′ ACG ATG AAT GTG GCG ATA CT 3′ and 5′ AGT GCA TTC AAG GCT GTT GG 3′. Real-time qPCR was performed by using Fast SYBR Green on a StepOne Plus real-time PCR system (Applied Biosystems, Grand Island, NY). Relative expression values were calculated using the 2-delta delta Ct method and were normalized against reference gene GAPDH. In these calculations we accounted for the PCR efficiency of the individual PCR reactions, calculated on the basis of linear regression as described elsewhere. The specificity of amplification was confirmed by evaluation of the melting curve.
Protein extraction and Western blot
The protein was extracted from placental samples and its concentration was measured by the BCA protein assay method per manufacturer’s protocol (Pierce, Rockford, IL). Each sample (50 μg) was separated in 10% sodium dodecyl sulfate-polyacrylamide gel. The protein was transferred onto nitrocellulose membrane. The membrane was blocked with blocking serum (Thermoscientific, Waltham, MA) followed by incubation with mouse antihuman NRP-1 antibody (1:1000) (Abcam, Cambridge, MA) overnight at 4°C. After washing with Tris-NaCl-tween 20 buffer, the membrane was incubated with antimouse IgG (1:15000) (Abcam) at room temperature for 1 hour. Immunoreactivity was detected using an enhanced chemiluminescence Western blotting system (Thermoscientific). Qualitative analysis was performed and expressed in relation to β-actin.
Immunohistochemical localization and quantification
The placental tissues were fixed in 4% buffered formaldehyde solution, dehydrated, and embedded in paraffin. The 4-μm thickness sections were transferred onto poly-l-lysin-coated slides, deparaffinized, rehydrated, and immunostained with NRP-1 antibody (1:200) using the Vectastain Immpress Reagent Kit (Vector Laboratories, Burlingame, CA). Two investigators blinded to the sample source independently graded immunostain intensity. Slides were first examined at ×4 magnification to identify NRP-1 immunopositive regions in placental sections. In each section, 5 different areas with 8-12 villi per area were selected at random and were evaluated microscopically with a ×40 objective magnification. All sections were scored in a semiquantitative fashion as described by Hsu et al, which considered both the intensity and percentage of cells staining. Intensities were classified as 0 (no staining), +1 (weak staining), +2 (moderate staining), and +3 (very strong staining).
Statistics
Angiogenic data (SuperArray Biosciences Corp) were normalized to corresponding internal controls and resulting gene expression values were cross-referenced between arrays. The Student t test was applied to 67 genes that were common to both SuperArrays and genes with P < .05 considered significant. The final candidate selection was done by filtering for the Bonferroni correction for multiple comparisons ( P < .0007). The values from qPCR were adjusted to GAPDH, and the values from Western blot were adjusted for β-actin, then mean and SD values were determined. The immunostain intensity met the distributional assumptions of a parametric statistical test. Independent groups t test was used to compare the control and FGR groups for these values. Differences were considered significant when P was <.05.
Results
Clinical characteristics
The basic clinical characteristics of the control and the FGR groups are presented in the Table . There were no differences between the groups regarding maternal age and gestational age. The population was racially diverse and the sample size did not permit any analysis of the impact of race. Notably, this was not an objective of this study. The birthweights and the placental weights were significantly lower in the FGR group than those in the control group. The birthweights were <5th centile for the study group and >20th centile for the control group. There were no fetal deaths in this population.
| Parameter | Control (n = 7) | FGR (n = 7) | Significance |
|---|---|---|---|
| Maternal age, y | 26.57 ± 2.60 | 25.71 ± 1.85 | NS |
| Gestational age, wk | 39 ± 0.85 | 37.46 ± 0.86 | NS |
| Birthweight, g | 3181.86 ± 234.31 | 2377.86 ± 72.88 | P < .01 |
| Placental weights, g | 563.66 ± 43.17 | 217.60 ± 54.58 | P < .001 |
Down-regulation of placental NRP-1 mRNA expression in FGR
To investigate that placenta-specific angiogenic mechanisms may underlie restricted fetal growth, we first determined expressions of genes in angiogenic pathway using 2 angiogenesis-oriented arrays (SuperArray Biosciences Corp and Qiagen, Valencia, CA). Several angiogenesis-related genes were differentially expressed in placentas from FGR pregnancies. A representative microarray photograph depicts expression of angiogenic genes in control and FGR in Figure 1 , A. Densitometric analysis showed significant reduction in NRP-1 in FGR placentas ( Figure 1 , B). A heat map depicting the results obtained in control and FGR placentas is shown in Figure 1 , C. The scatter plot shown in Figure 1 , D, demonstrates underexpression and overexpression of genes in FGR placentas. Although several genes in the study group were down-regulated by >30% of those in the control group, only NRP-1 was inhibited by >2-fold ( Figure 2 ). When the gene expression profiles of the 2 microarray platforms, SuperArray HS-009 and SuperArrayOHS-024 (SuperArray Biosciences Corp), were cross-referenced and filtered by Student t test, only 3 genes were significantly underexpressed, of which only NRP-1 remained significant ( P < .0007) when further filtering was performed by Bonferroni correction for repeated measures ( Figure 3 ).
