Cotyledon and binucleate cell nitric oxide synthase expression in an ovine model of fetal growth restriction

Henry L. Galan, Timothy R. H. Regnault, Timothy D. Le Cras, R. Weslie Tyson, Russell V. Anthony, Randall B. Wilkening, Steven H. Abman


Heat exposure early in ovine pregnancy results in placental insufficiency and intrauterine growth restriction (PI-IUGR). We hypothesized that heat exposure in this model disrupts placental structure and reduces placental endothelial nitric oxide synthase (eNOS) protein expression. We measured eNOS protein content and performed immunohistochemistry for eNOS in placentas from thermoneutral (TN) and hyperthermic (HT) animals killed at midgestation (90 days). Placental histomorphometry was compared between groups. Compared with the TN controls, the HT group showed reduced delivery weights (457 ± 49 vs. 631 ± 21 g; P < 0.05) and a trend for reduced placentome weights (288 ± 61 vs. 554 ± 122 g; P = 0.09). Cotyledon eNOS protein content was reduced by 50% in the HT group (P < 0.03). eNOS localized similarly to the vascular endothelium and binucleated cells (BNCs) within the trophoblast of both experimental groups. HT cotyledons showed a reduction in the ratio of fetal to maternal stromal tissue (1.36 ± 0.36 vs. 3.59 ± 1.2; P≤ 0.03). We conclude that eNOS protein expression is reduced in this model of PI-IUGR and that eNOS localizes to both vascular endothelium and the BNC. We speculate that disruption of normal vascular development and BNC eNOS production and function leads to abnormal placental vascular tone and blood flow in this model of PI-IUGR.

  • endothelial nitric oxide synthase
  • hyperthermia
  • placenta
  • binucleate cell
  • intrauterine growth restriction

intrauterine growth restriction (IUGR) is a significant complication of pregnancy affecting up to 8% of all pregnancies (6, 28). Fetal complications due to IUGR include intrapartum distress, hypoxia, asphyxia, and intrauterine demise. Neonatal complications include meconium aspiration, metabolic and hematologic disturbances, cognitive dysfunction, and cerebral palsy. In addition to perinatal morbidity and mortality, epidemiological studies suggest that IUGR may contribute to adverse health effects in adulthood including cardiovascular disease and diabetes (4, 27). IUGR fetuses may exhibit symmetric or asymmetric patterns of growth. Although the etiologies of asymmetric IUGR are heterogeneous, the clinical and pathological manifestations may be due to universal structural and functional abnormalities in the placenta.

A number of placental structural abnormalities have been described in IUGR pregnancies, including decreased villous number, diameter, and surface area, as well as decreased arterial number, lumen size, and branching (10, 14, 18, 21, 31). Doppler ultrasound studies of the umbilical artery in both humans and sheep provide evidence for an increase in placental vascular resistance in IUGR pregnancies (9, 38). Increased vascular tone and reactivity may be due, in part, to alterations in vasoactive mediators such as endothelin, prostacyclin, thromboxane, and nitric oxide (NO). Hence, increased placental vascular resistance may reduce umbilical blood flow and nutrient delivery to the fetus.

NO is a free radical molecule with potent vasodilator properties and a short half-life that is synthesized by the enzyme NO synthase (NOS). NOS catalyzes the conversion of l-arginine tol-citrulline with NO as a by-product. NO activates soluble guanylate cyclase to produce cGMP that results in vasorelaxation (23). Three isoforms of NOS have been described: endothelial NOS (eNOS), inducible NOS, and neuronal NOS. Gestational age-related changes in eNOS have been described in the umbilical artery and in a number of fetal tissues (23). eNOS has been localized to syncytiotrophoblast and vascular endothelium in humans and to the vascular endothelium in sheep, baboon, rat, and guinea pig (45). NO is a vasoactive molecule that has been shown to be active in a number of vascular beds including the brain, lung, and kidney (1, 24, 36, 39) as well as the placenta and umbilical vessels (41, 45). Past studies have shown that chronic treatment with an antagonist that inhibits all NOS isoforms causes IUGR and preeclampsia-like syndrome in the pregnant rat (22, 44). However, information on the role of placental NOS and its role in the etiology of IUGR is limited.

We hypothesized that heat-stress exposure early in ovine pregnancy disrupts placental development and reduces placental eNOS protein expression. To address this hypothesis, we used an established ovine model of placental insufficiency and intrauterine growth restriction (PI-IUGR) induced by hyperthermia (HT) beginning early in gestation. This model is characterized by asymmetric fetal growth and reduced placental mass (8). It is also associated with reduced uterine and umbilical blood flows, reduced transplacental amino acid flux, reduced glucose and oxygen transport capacity, as well as abnormal umbilical arterial and aortic Doppler velocimetry (3, 5,9, 30, 37). To test our hypothesis, eNOS protein expression of placentomes from 90-day-gestation HT animals and thermoneutral (TN) control sheep were compared by Western blot analysis, immunohistochemistry, and histology.


Experimental design and animal care.

This study was approved by the University of Colorado Health Sciences Center Animal Care and Use Committee. Eight mixed-breed (Columbia-Rambouillet) ewes with time-dated singleton pregnancies were used for this study. Beginning at 40 days gestation (term = 147 days), four ewes were placed in a HT environment and four ewes were kept at ambient conditions serving as TN controls. The environmental chamber conditions have been previously described and include a temperature that cycles at 40°C for 12 h during the day and 35°C during the night and humidity that is kept between 30 and 40% (9). All ewes were pair fed and offered water ad libitum. Sheep were euthanized at 85–93 days gestation, and fetal and placentome weights were recorded. The placentome was divided into cotyledon (fetal) and caruncle (maternal) components, which were frozen in liquid nitrogen. Placentomes were also sectioned and placed in 10% formalin and paraffin embedded for histology and immunolocalization studies.

Western blot analysis.

Caruncles and cotyledons were homogenized separately in buffer containing leupeptin, phenylmethylsulfonyl fluoride, betamercaptoethanol, and pepstatin A. Western blot analysis was performed on crude homogenates of whole caruncles and cotyledons from each animal (25 μg of total protein) by use of a monoclonal antibody to eNOS (Transduction Laboratories, Lexington, KY), as previously described (16). Western blots were performed with NuPAGE 4–12% Bis-Tris gradient gels (Novex, San Diego, CA). Densitometry was performed with a UMAX PowerLook II scanner (UMAX DATASystems, Hsinchu, Taiwan) and National Institutes of Health Image software. Western blot analysis was used to compare eNOS protein content between1) caruncles and cotyledons from TN pregnancies,2) cotyledons of increasing weights from TN pregnancies, and3) cotyledons of comparable weights from HT and TN pregnancies (3.86 ± 0.2 vs. 3.72 ± 0.1 g, respectively; P = 0.63)

Histology and immunohistochemistry for eNOS protein.

Midline cross sections of placentomes from each treatment group were fixed in 10% buffered formalin and embedded in paraffin. Paraffin sections 6 μm thick were mounted onto Superfrost Plus slides (Fisher Scientific, Fairlawn, NJ) for staining with hematoxylin and eosin and for immunohistochemistry.

A single pathologist (R. W. Tyson) who was blind to the treatment groups performed the histomorphometry on hematoxylin- and eosin-stained sections. A single placentome of comparable size from each animal was analyzed. Vessel density and vessel diameters were measured for surface vessels of the placentome. Vessel density was performed by point counts using a grid eyepiece and standard magnification (×100). Counts were taken when cross points of the grid touched the wall of the fetal or maternal vessel or fell within the lumen. Using an eyepiece micrometer, we then measured the vessel wall thickness and the vessel outside diameter for vessel wall thickness-to-diameter ratios. This ratio was calculated by a formula previously described by Abman et al. (2) for pulmonary arterioles. The area of fetal villi and maternal stroma were measured on three random ×10 microscopic fields, and the ratio of fetal villi to maternal stroma was compared between treatment groups by using Image-Pro Plus morphometric image analysis software (Media Cybernetics, Silver Spring, MD). Images were captured by a Polaroid digital microscopic camera. Once the images were captured, the areas of fetal tissue were outlined and the total fetal tissue area was calculated. The total area of the image was calculated, and the fetal area was subtracted to give the maternal tissue area. The ratio of fetal to maternal area was then calculated. Figure1 is an example of the computer-captured image of a ×4 histological view of a sheep cotyledon demonstrating the acquisition of the areas of interest outlined in black.

Fig. 1.

Example of the Image Pro Plus morphometric image demonstrating how fetal (F) and maternal (M) tissues are outlined. Areas (mm2) are reported on a separate display screen (not shown).

Immunohistochemistry for eNOS was performed on 6-μm-thick sections of paraffin-embedded whole placentomes. Slides were dewaxed with 100% xylene. Slide preparation and antigen retrieval were performed as previously described by Le Cras et al (16). Slides were washed in PBS, and sections were blocked for 45 min with Super Block (Sky Tek, Logan, Utah) diluted 1:10 (vol/vol) in 1× PBS. Slides were then incubated for 2 h with a rabbit polyclonal primary antibody against eNOS (Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:2,000 or with an IgG1-negative control (Jackson Laboratories, West Grove, PA). Sections were then washed in 1× PBS. Sections were then incubated for 45 min with a biotin-labeled anti-mouse (for IgG1) or anti-rabbit (for eNOS) secondary antibody at a dilution of 1:10,000. Slides were washed in 1× PBS. Slides were incubated in streptavidin-biotin-horseradish peroxidase solution and developed with diaminobenzidine (DAB) using the Vectastain ABC and DAB kit (Vector Laboratories, Burlingame, CA). The NiCl enhancement-DAB color development reaction was stopped by washing with water. Slides were dehydrated in 70% ethanol-30% water, 95% ethanol-5% water, 100% ethanol, and 100% xylene and then coverslipped.

Statistical analysis.

Gestational age, cotyledon number, fetal and placental weights, and Western blot densitometry values between treatment groups were compared by Wilcoxon test by use of SigmaStat 2.0 (SPSS, Chicago, IL). Histomorphometry between HT and NT groups was compared by unpairedt-tests. Data are presented as means ± SE, and aP value of <0.05 was considered significant.


Table 1 shows gestational ages and fetal and placental weights at death for each study group. Both groups were delivered at similar gestational ages. Fetal weight was significantly reduced in the HT group (P < 0.05). A trend for a decrease in placental weight was noted in the HT group.

View this table:
Table 1.

Gestational age and fetal and placental weights for each study group

Figure 2 shows a representative Western blot for eNOS in cotyledon and caruncle pairs for the TN group. These findings show that eNOS protein is predominantly located in the cotyledon component of the placentome. The cotyledon has 14-fold higher eNOS protein concentration compared with caruncle. Figure3 depicts a representative Western blot and corresponding linear regression plot of cotyledons of different weights (25 μg protein load for each) from TN controls. Linear regression analysis shows a strong correlation in TN controls (r 2 = 0.8) between eNOS protein and cotyledon size. Figure 4 shows a representative Western blot with corresponding histogram of the densitometry demonstrating a twofold decrease in eNOS protein content in the cotyledons of the HT pregnancies compared with the TN group (P < 0.05). Figure 5shows immunolocalization of eNOS protein to chorionic binucleate cells within the fetal trophoblast and to the endothelium of the blood vessels within the fetal villi.

Fig. 2.

Representative Western blot demonstrating a 14-fold greater endothelial nitric oxide synthase (eNOS) protein concentration in the cotyledons (n = 4) compared with caruncles (n = 4) from an animal in the thermoneutral (TN) group (P < 0.001).

Fig. 3.

Representative Western blot showing a positive correlation between eNOS protein concentration and cotyledon mass (R 2 = 0.84).

Fig. 4.

Representative Western blot depicting a 2-fold reduction in the cotyledons from the animals in the hyperthermic (HT) group (n = 4) compared with TN controls (n = 4; P < 0.05).

Fig. 5.

A and B: localization of eNOS to the binucleated cells in the trophectoderm at ×40 and ×100 magnification, respectively. C: IgG negative control. D: eNOS localization to the vascular endothelium within the fetal villi.

Histomorphometry showed that there were no differences between HT and TN groups for either fetal surface vessel counts or wall thickness (Table 2). However, there was a reduction in the ratio of fetal to maternal stroma in the HT group compared with TN controls (P < 0.03). Figure6 shows the striking histological differences in fetal to maternal stromal content between the HT and TN groups. The histology of the HT cotyledons was similar to that of a term cotyledon that is shown for comparison.

View this table:
Table 2.

Placentome histomorphometry

Fig. 6.

A and B: histological sections of cotyledons from 90-day postconception TN and HT animals, respectively. Note the marked disparity in the F to M components between the groups.C: histological section of a term TN cotyledon for comparison. Striking similarities are seen between the HT and the term TN sections.


We found that HT animals showed a twofold reduction in cotyledon eNOS protein content. We also found in control animals that eNOS protein content was 14-fold greater in the cotyledon compared with the caruncle. The strong positive correlation of eNOS protein with cotyledon weight is important from a methodological standpoint. This finding directed us to compare cotyledons of similar sizes between the TN and HT groups, thus avoiding the potential bias of comparing different-sized cotyledons. Alternatively, by comparing cotyledons of similar sizes, we may have masked larger differences in eNOS protein content because the HT animals tended to have placentas of smaller weight. Previously shown to localize to vascular endothelium, we found that eNOS protein also localized specifically to the BNCs in the trophectoderm of the cotyledon. No differences in fetal surface vessel number or wall thickness were detected; however, there was a 2.5-fold decrease in the ratio of fetal villous stroma to maternal villous stroma in the HT cotyledons relative to TN cotyledons. This decrease in the fetal villous stroma to maternal villous stroma in the HT cotyledon was strikingly similar to that normally seen at term. This suggests an accelerated maturation of the placenta in the HT animals, a finding that has been previously described in the human IUGR placenta (11). As in previous studies using this model of IUGR, we found a significant decrease in fetal weight in the HT group. Interestingly, this decrease in fetal weight was not accompanied by a significant reduction in placental weight. However, it has been previously shown that placental weight in the near-term ovine fetus is dependent on the duration of heat stress (9). Thus it is likely that, with time, differences in placental weight would have become apparent.

A reduction in eNOS protein in the HT cotyledons provides a potential explanation for the decrease in absolute umbilical blood flow (37, 30) and increase in placental resistance detected by umbilical artery Doppler (9) previously shown in this PI-IUGR model. Reduced eNOS protein would likely result in reduced NO production and increased basal vascular tone and vascular resistance.

Past studies have provided evidence supporting a role for NO in regulating tone in the fetoplacental circulation in early gestation. The NO-cGMP cascade appears to be most active in the first trimester human placenta. Peak NOS activity in human placental homogenates occurs in the first trimester and decreases to term (29). Izumi et al. (12) further showed in human umbilical artery smooth muscle bath studies that the amount of NO and the sensitivity of the smooth muscle to NO decreased with advancing gestation. NOS activity has been shown to rise early in the tissues of a variety of organs in fetal guinea pigs (40). These investigators suggested that NOS activity levels may be related to estrogen levels because the increased NOS activity mirrored the rapid rise and plateau of estrogen. Using Doppler ultrasound, Lees et al. (19) showed that umbilical artery vascular impedance was highest in early pregnancy, which was paradoxical to the higher levels of NOS activity and cGMP levels. It may also be that levels of vasoconstrictors may be high, which may offset the NO-cGMP cascade. Lees et al. speculated that perhaps shear stress in these high-resistance vascular beds was responsible for the increase of NOS activity and increased cGMP levels. Chronic hypoxia has been shown to upregulate eNOS in the rat lung, and although eNOS levels were high, Sato et al. (33) showed that NO production was low under hypoxic conditions. It is possible that NOS activity is higher in early pregnancy because of the relative hypoxic environment of the villi noted in early pregnancy.

The localization of eNOS protein to the BNC is of particular interest because it is a trophoblast cell that also produces progesterone and ovine placental lactogen. The BNC migrates from the fetal side of the trophectoderm to the maternal syncytium to deliver these hormones into the maternal bloodstream. In a study comparing eNOS and inducible NOS isoforms in the placenta of various species, Zarlingo et al. (45) found that eNOS localized to both syncytiotrophoblasts and vascular endothelium in the human placenta, but only to the vascular endothelium of sheep and other species. They studied these animals late in gestation, and it is possible that eNOS expression was undetectable since NOS levels are decreased in late gestation (29). In our study, immunohistochemistry was performed on placentomes at midgestation, which is a time in gestation when eNOS expression is higher (29). In all ruminant placentas, the BNC is found in the trophectodermal epithelium and are found in fairly constant proportions across gestation (13, 42,43). The BNC has been shown to migrate across the fetomaternal junction (43). BNCs appear to have two primary functions: they form the maternal-fetal syncytium needed for implantation and placentomal growth, and they produce and deliver protein and steroid hormones (43).

The localization of eNOS to the BNC together with other known features of the BNC allows us to draw comparisons with the trophoblast of the human placenta. The BNC is the only migratory cell in the ovine placenta moving across the fetal-maternal interface to the maternal uterine epithelium forming the fetal-maternal syncytium necessary for implantation. Because they are the only migratory cells in the ovine placenta, BNCs are analogous to invasive cytotrophoblasts in the human placenta. The human cytotrophoblast differentiates to become cell columns that invade the uterine epithelium to anchor the fetus and establish blood flow to the placenta and fetus (46). The human cytotrophoblast also produces the syncytiotrophoblast that produces proteins and steroid hormones (46). Similarly, the BNC produces protein and steroid hormones. The BNC and the human syncytiotrophoblast also share eNOS localization (45). Given this information, we speculate that the BNC may be the ovine placental correlate to the cyto- and syncytiotrophoblast of the human placenta because it has both hormonal and migratory functions. We further speculate that the BNC develops a vascular phenotype much like the human cytotrophoblast, which has been shown to express receptors and other markers of endothelial cells (46). The localization of eNOS to the BNC in our study is consistent with this speculation and rational from a teleological viewpoint.

Because the NO-cGMP cascade is most active in early gestation, this may provide some insight into the mechanisms of reduced placental weight and IUGR in this model. The effects of heat treatment on placental growth early in gestation may be critical because this is a period of rapid vascular growth in the placenta. Although we found no difference in the number or wall thickness of the surface vessels of the placentomes, it remains necessary to quantitate vasculature changes within the fetal villi. Preliminary data on fetal villous vascular casts from our lab suggest a significant disturbance in angiogenesis of HT placentas (T. R. H. Regnault, R. V. Anthony, and R. B. Wilkening, unpublished data). Disruption of placental growth and probably vascular growth might result in altered of eNOS protein expression. The reduction in eNOS protein may be due to the direct effect of heat on endothelial cells and BNC or to an overall reduction in the number of endothelial cells and BNC in proportion to other cells. Regardless of the mechanism, a reduction in eNOS will probably affect the NO-cGMP cascade and alter basal vascular tone and possibly vessel growth. NOS activity and cGMP levels shown to be highest early in human pregnancy may be critical in the formation of new villous vessels (19). This may contribute to the reduction in umbilical blood flow and increase in placental vascular resistance thus affecting nutrient transport to the placenta and the fetus destined to become IUGR. The relationship between reduced eNOS protein and IUGR would be consistent with studies showing that administration of a nonspecific inhibitor of NOS into the drinking water of pregnant rats in early gestation produces an asymmetrically growth restricted fetus (22,32, 42).

Our finding of reduced placental eNOS concentration in sheep PI-IUGR pregnancies at 90 days gestation contrasts with other studies demonstrating an increase in eNOS or NO production in the fetoplacental circulation of IUGR fetuses (20, 25). Lyall et al. (20) determined NO concentrations by measuring indexes of NO production (nitrites and nitrates) in human umbilical venous blood from IUGR and normal pregnancies. They found that pregnancies complicated by IUGR had significantly higher nitrite and nitrate levels in the umbilical venous plasma. They concluded that the increase in NO indexes could represent a compensatory effort to improve placental blood flow. Myatt et al. (25) used a semiquantitative immunohistochemical technique to show that pregnancies complicated by IUGR or by both preeclampsia and IUGR have increased basal distribution of eNOS in syncytiotrophoblast as well as intense terminal villous capillary endothelial immunostaining. Although several possibilities exist to explain the differences between these studies, one likely explanation is the difference in relative gestational age at the time of tissue collection. If our samples had been collected later in gestation, as were the earlier reports (20, 25), we may have seen comparable increases in eNOS production. In support of this possibility is the demonstration of placental hypoxia at 135 days gestation in our sheep model of PI-IUGR (37). Greater placental hypoxia could lead to enhanced vascular endothelial growth factor production (34), which could then lead to enhanced eNOS production (15). Hypoxia alone has been shown to increase eNOS production in other vascular beds as well. Thus alterations in eNOS expression in IUGR placenta may very well change as a function of the developmental stage of the placenta, necessitating observations at multiple time points during gestation. Therefore, our findings could still be consistent with those of Lyall et al. and Myatt et al. The different findings between our study and that of Lyall et al. and Myatt et al. are important and help direct us toward future studies in our model of IUGR.

Although the heat stress IUGR animals have been shown to be hypoxic at term, one of the limitations of this study is that we did not know the oxygen tension of these fetuses at midgestation. This may be important because, although little is known about the effects of hypoxia on eNOS regulation in the placenta, there are reports suggesting that hypoxia affects eNOS expression in the pulmonary vascular bed. Endothelial NOS gene and protein expression in small resistance vessels of the lung were induced with chronic hypoxia (17), and inhibition of NO production induces vasoconstriction in rat lungs (26). In contrast, Fike et al. (7) have shown that chronic hypoxia decreases both NO production and eNOS synthase in newborn piglet lungs. In addition, NO appears to play an active role in regulating basal pulmonary vascular tone as depicted by an increase in pulmonary vascular resistance with NOS inhibition (35). Because hypoxia can affect eNOS and NO production, it remains necessary to determine the oxygen tensions at midgestation in our IUGR model so that the mechanism of reduced eNOS protein can be better defined.

In summary, heat stress reduces fetal and placental growth, disrupts normal placental histology, and reduces the eNOS protein expression in the cotyledon. We speculate that the placental vascular development will mirror the abnormal villous histology and that this, together with the reduced eNOS production, accounts for the abnormal hemodynamics seen in this model. It remains uncertain whether the reduction in eNOS protein is a result of abnormal vascular development or whether a reduction in eNOS protein contributes to abnormal vascular development. The localization of eNOS to the BNC introduces another variable that could be affected by heat and could impact placental morphology and vascular development. The availability of an animal model that produces an asymmetric pattern of growth restriction by the application of a known environmental stress provides a means of investigating the timing and duration of the environmental stress in terms of fetal and placental growth. It also allows for the investigation of pathophysiological and molecular mechanisms of placental damage induced by thermal stress.


We appreciate the insight of Dr. Frederick C. Battaglia and the technical support provided by Neil Markham, David Hood, Peter Orchard, Karen Trembler, Pat Lenhardt, Sarah Williams, and Kate Fasth.


  • This work was supported by the American Association of Obstetricians and Gynecologists Foundation and Burroughs-Wellcome Fund, the National Institutes of Health NHLBI SCOR Grant HL-57144, National Institutes of Health Program Project Grant PO1HD20761, and March of Dimes Grant no. 6-F497-0174.

  • Address for reprint requests and other correspondence: H. L. Galan, Dept. of Obstetrics and Gynecology, Univ. of Colorado Health Sciences Center, 4200 E. 9th Ave., Campus Box B-198, Denver, CO 80262.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract