Gene expression of nitric oxide synthase by human umbilical vein endothelial cells: the effect of fetal plasma from pregnancy with umbilical placental vascular disease

Authors


* Professor B. Trudinger, Department of Obstetrics and Gynaecology, University of Sydney at Westmead Hospital, P.O. Box 533 Wentworthville, New South Wales 2145, Australia.

Abstract

Objective To determine whether endothelial cell injury would be produced by factor(s) released into the fetal circulation, manifested by altered messenger RNA expression of nitric oxide synthase.

Design Case–control study.

Setting University teaching hospital.

Samples Fetal plasma was collected from 34 normal pregnancies, 44 pregnancies with umbilical placental vascular disease identified by an abnormal umbilical Doppler and 11 pregnancies with maternal pre-eclampsia but with normal umbilical Doppler studies.

Methods Aliquots from a common culture of human umbilical vein endothelial cells (HUVECs) were incubated with fetal plasma from the members of the three patient groups. The total RNA was extracted from the endothelial cells and mRNA for nitric oxide synthase was measured by reverse transcription and semi-quantitative polymerase chain reaction (RT-PCR). This was standardised by comparison of the amplified inducible nitric oxide synthase (iNOS) or endothelial constitutive nitric oxide synthase (ecNOS) to glyceraldehyde 3-phosphate dehydrogenase (GAPDH).

Main outcome measure Endothelial cell gene expression of iNOS and ecNOS.

Results The mRNA expression of iNOS and ecNOS were significantly higher (P < 0.05) in HUVECs stimulated by fetal plasma from pregnancies with umbilical placental vascular disease [iNOS 1.12 (0.16); ecNOS 1.78 (0.18)] when compared with normal pregnancies [iNOS 0.56 (0.06); ecNOS 1.06 (0.10)]. In the maternal pre-eclampsia group, the NOS expression [iNOS 0.76 (0.11); ecNOS 1.39 (0.26)] did not differ from normal pregnancy. In the vascular disease group, there was no difference in NOS expression between the subgroups with and without maternal pre-eclampsia.

Conclusions Our study demonstrates that umbilical placental vascular disease is associated with a factor(s) in fetal plasma that produces an increase in the expression of iNOS and ecNOS mRNA by endothelial cells. Our findings raise the possibility that the release of factors causing an up-regulation of iNOS and ecNOS in the endothelium in the fetal placenta may occur as part of an inflammatory response of the vascular endothelium to injury.

Introduction

Vascular disease in the placenta may be associated with adverse effects on both the fetus and its mother. Pathological changes are present in both the maternal uteroplacental circulation and the fetal umbilical placental circulation. In the fetus, intrauterine growth restriction (IUGR) may occur. This is a major cause of perinatal mortality and morbidity. In the mother, hypertension and the multisystem manifestations of pre-eclampsia may occur. The clinical features of the syndromes of the fetus and its mother may occur separately or together. While the pathogenic mechanisms of the disease process are not clear, endothelial cell injury and dysfunction in the placental circulations appears to be a pivotal step in the process of pre-eclampsia1,2. To date, studies of endothelial cell function in placental insufficiency have largely focussed on the maternal uteroplacental vascular compartment, which is characterised by reduced perfusion3,4. The fetal circulation has been less studied.

We have recently studied the origins of the vascular pathology in the fetal umbilical placental circulation. This may be identified antenatally by a ‘high resistance’ waveform pattern of the umbilical artery flow velocity waveforms5,6. Thrombosis and vessel obliteration are seen on histological study7,8. Platelet activation and consumption can be demonstrated9,10. We postulated that these changes in the fetal platelet population would follow contact with activated endothelial cells. Incubation of a standard human umbilical vein endothelial cells (HUVEC) culture with 10% fetal plasma caused endothelial cell expression of adhesion molecules of the immunoglobulin gene superfamily indicative of endothelial cell activation11. Apoptosis was increased12. Synthesis of proinflammatory cytokines IL-6 and IL-8 also occurred13. As further evidence these processes were occurring, we showed that fetal plasma from patients with an abnormal umbilical artery Doppler study contained increased levels of IL-6 and IL-814 and the soluble fractions of the adhesion molecules ICAM-1 and PECAM-111. All of these findings pointed to injury of the vascular endothelium in the fetal umbilical placental circulation in cases with high resistance umbilical artery Doppler waveforms.

Nitric oxide is known to play important functional roles in a variety of physiological systems. Within the vasculature, nitric oxide induces vasodilatation, prevents neutrophil/platelet adhesion to endothelial cells, inhibits platelet aggregation, controls smooth muscle cell proliferation and migration, maintains an endothelial cell barrier function and regulates programmed cell death (apoptosis)15. Nitric oxide is synthesised from l-arginine by the action of nitric oxide synthase, an enzyme existing in three isoforms: neuronal nitric oxide synthase (nNOS, NOSI), inducible nitric oxide synthase (iNOS, NOSII) and endothelial constitutive nitric oxide synthase (ecNOS or NOSIII). The isoforms of nitric oxide synthase have been cloned and sequenced. Many factors can influence and regulate the expression of this enzyme16. Nitric oxide can influence many aspects of the inflammatory cascade and increased levels of the isoform of nitric oxide synthase have been demonstrated in inflammation17. There is debate about whether this is beneficial or harmful.

In this study, we tested the hypothesis that fetal placental vascular disease is associated with the presence of factor(s) in the fetal circulation that affect endothelial cell function and in particular the expression of nitric oxide synthase. Endothelial cell nitric oxide synthase activity was determined by measuring the expression of messenger RNA to the two isoforms of nitric oxide synthase (iNOS and ecNOS). The effect of fetal plasma from normal and umbilical placental insufficiency pregnancy on HUVECs was determined.

Methods

We collected fetal blood at delivery and investigated its effect on HUVECs in culture.

We studied the actions of the fetal plasma from three groups of pregnant women. Fetal plasma was collected from 34 normal pregnancies and 44 fetuses with an abnormal umbilical artery Doppler study indicating umbilical placental vascular disease. This group was subdivided according to the absence (n= 23) or the presence (n= 21) of associated maternal pre-eclampsia. Our third group was plasma collected from 11 fetuses in which the pregnancies were complicated by maternal pre-eclampsia but umbilical artery Doppler studies were normal. In the normal group, there were no identifiable medical or obstetric diseases. All normal pregnancies delivered at term (>37 weeks) by spontaneous vaginal delivery or elective caesarean section (for reasons not associated with fetal compromise) and the birthweight was above the 10th centile. The umbilical placental vascular disease group was identified by an abnormal umbilical artery Doppler study (greater than 95th centile using our reported normal range6) performed within one to six days prior to delivery. Maternal pre-eclampsia was defined as a blood pressure ≥140/90 mmHg on at least two occasions at least 6 hours apart occurring after the 20th week of gestation accompanied by proteinuria (>300 mg/24 hours). Fetal blood was collected from the umbilical vein at delivery. Plasma was prepared from blood anticoagulated with citrate to avoid the confounding effects of cellular products released into serum during blood coagulation. The plasma was separated within 1 hour and stored at −70°C until assayed.

HUVECs were isolated from the umbilical cord of fetuses from normal pregnancies and vaginal delivery. We used 0.1% collagenase type 1 (Worthington Biochemical, New Jersey, USA) for digestion to separate the endothelial cells. The cells were cultured in Dulbecco's modified eagle medium (DMEM, Gibco BRL, Maryland, USA) containing 25 mM Hepes (Sigma, Missouri, USA), 100 u/mL Penicillin G and 100 μg/mL streptomycin sulphate (CSL Biosciences, Melbourne, Australia), 10 u/mL heparin (DBL, Melbourne, Australia), 20% fetal calf serum (Gibco BRL) and 20 μg/mL endothelial cell growth factor (Boehringer Mannheim, Mannheim, Germany) within 1% gelatin (Bio-Rad Laboratories, California, USA) coated tissue culture flasks (25 cm2). Every 48–72 hours the medium was replaced until the cells reached confluence and were detached with trypsin/EDTA (CSL Biosciences). Cells from different umbilical veins were mixed, centrifuged and frozen in liquid nitrogen tank until required. The same mixture of cell lines was used to test all of the fetal plasma samples.

Confluence of endothelial cells was identified by the typical cobblestone appearance. We confirmed that it was endothelial cells in culture by positive immunostaining for monoclonal antibodies against von Willebrand Factor (Silenies, Melbourne, Australia) and thrombomodulin (Serotec, Oxford, UK), and the presence of uptaking Del-Acetylated low-density lipoprotein (Dil-Ac-LDL, Biomedical Technologies, New Jersey, USA). The positive cells were always greater than 97%. Splits from the same HUVECs were thawed and plated in 25 cm2 flasks and grown in culture medium as passage two to reach a confluent monolayer. The cells were made quiescent in DMEM medium without serum for 24 hours before experimental stimulation. Fetal plasma was heparinised (100 u/mL) to prevent clotting of the diluted plasma and added to the cultured cells at a final concentration of 10% fetal plasma for 24 hours. All experiments were performed with double flasks. Stimulated cells were dispersed with trypsin/EDTA and harvested the pellets of cells.

Total RNA was extracted from the stimulated endothelial cells using the RNA zolB Kit (Tel-Test, Texas, USA). The integrity of RNA was confirmed by the presence of intact 18s and 28s bands on 1% agarose gel. The RNA was shown to be free of protein and DNA. The optical density ratio at 260/280 nm was always >1.8 and each RNA concentration was determined by using Spectrophotometer (Beckman Instrument, California, USA) with the Nucleic Soft pac. First-strand cDNA synthesis was performed by reverse transcriptase and oligo (dT)12–18 to prime the reaction (Superscript Preamplification System, Gibco BRL). Specifically, 2 μg of total RNA template sample was transferred to a RNase-free microcentrifuge tube to perform the RT reaction as described by the instruction manual. Control RNA and no RT control had been done at the same time.

The expression of iNOS and ecNOS mRNA in HUVEC exposed to the fetal plasma was assessed using a modification of the semi-quantitative PCR. Preliminary experiments have been done to optimise the amount of cDNA and to assess the PCR optimum assay conditions for quantity. A 5 μL sample of each cDNA was used to perform an iNOS PCR reaction. For the ecNOS reaction, 2.5 μL was used. These were added to a 25 μL volume containing 1.25 units of thermostable DNA polymersae (Advanced Biotechnologies, Surrey, UK) according to the basic protocol for PCR. The primers were designed from the sequence of gene bank. The human iNOS primers were sense, 5′-GGA GCC AGC TCT GCA TTA TC-3′ and antisense, 5′-GTG CAC TCA GCA GCA AGT TC-3′; human ecNOS primers sense, 5′-ACA TCC TGA GGA CGG AGC TGG CT-3′; and antisense, 5′-TGC GTA TGC GGC TTG TCA CCT CC-3′. The amplification profile of iNOS was with denature 95°C for 1 minute, anneal 56°C for 1 minute, extend 72°C for 1 minute and total of 36 cycles. The ecNOS reaction was amplified for 32 cycles at 94°C for 1 min, 68°C for 1 minute, 72° for 1 minute. A housekeeping gene human glyceraldelyde 3-phosphate dehydrogenase (GAPDH) was used to control PCR assays. All PCR system conditions were optimised with respect to the linear range of the amplification reaction and the PCR reactions were duplicated. The iNOS or ecNOS and GAPDH product were then respectively loaded into 2% agarose gel. Separation of DNA fragments was achieved by electrophoresis with an ethidium bromide staining. The gel was photographed with Polaroid Type 665 film over ultraviolet light. The bands on the negative film were scanned and analysed by Personal Densitometer SI using ImageQuant software (Molecular Dynamics, California, USA). The signal intensity for iNOS and ecNOS was counted and expressed relative to GAPDH.

As a positive control for the iNOS RT-PCR assay we prepared RNA from human skeletal muscle18. For the ecNOS, positive control RNA prepared from normal pregnant placental tissue was used19. Genomic DNA contamination was checked by carrying samples through RNA isolation and a first-strand cDNA synthesis reaction. Water was to serve as a control for cDNA contamination.

The differences in clinical data and the levels of mRNA expression of iNOS and ecNOS between the test groups were assessed using one-way analysis of variance with Dunnett's post hoc comparisons. Fisher's exact test was used for comparison between categorical variables. Pearson correlation was used for comparison of the mRNA expression levels of iNOS with ecNOS. Two-way analysis of variance was used to evaluate the difference between iNOS and ecNOS within the study groups controlled by smoking status, labour status and gestational age. A P value of less than 0.05 was considered statistically significant.

This study was conducted with the approval of the Western Sydney Area Health Service Ethics Committee.

Results

The clinical data are summarised in Table 1. Not surprisingly, the group with umbilical placental vascular disease delivered earlier and the infant birthweight was reduced. The centile infant birthweight for gestational age was also lower in this group. The placenta was smaller. More pregnancies required caesarean delivery.

Table 1.  Clinical characteristics. Values are expressed as mean (SEM) or n.
 Normal pregnancy (n= 34)Umbilical placental vascular disease (n= 44)Pre-eclampsia normal Doppler (n= 11)
  1. *P < 0.05, umbilical placental vascular disease group or maternal pre-eclampsia only group versus normal pregnancy group.

  2. **P < 0.001, umbilical placental vascular disease group or maternal pre-eclampsia only group versus normal pregnancy group.

Maternal age (years)31.53 (0.87)29.32 (0.97)32.82 (2.08)
Gestational age (week)39.56 (0.20)32.82 (0.63)**36.59 (1.09)*
Infant birthweight (g)3431.41 (105.38)1587.61 (124.08)**2752.27 (238.62)*
Centile weight (%)51.84 (5.11)11.50 (2.42)**33.44 (8.94)
Placenta weight (g)666.00 (26.54)380.93 (32.22)**613.00 (55.31)
No. of ≤5th centile023**2
No. of vaginal delivery4 44
No. of smoking8205

Both iNOS mRNA and ecNOS mRNA were amplified using RT-PCR. The endothelial cells were incubated with fetal plasma from our three study groups and RNA was then isolated from the cells. The densitometry results of mRNA expression for iNOS and ecNOS controlling by GAPDH are shown in Fig. 1. There were significant increases in both the iNOS mRNA expression [1.12 (0.16)] and the ecNOS mRNA expression [1.78 (0.18)] in pregnancy with umbilical placental vascular disease when compared with normal pregnancy [iNOS 0.56 (0.06); ecNOS 1.06 (0.10)]. However, the maternal pre-eclampsia with normal umbilical Doppler group [iNOS 0.76 (0.11); ecNOS 1.39 (0.26)] did not differ significantly from the normal pregnancy group. In the umbilical placental vascular disease whole group, there were no differences in the expressions of both iNOS and ecNOS mRNA when the subgroups with maternal pre-eclampsia present [iNOS 1.12 (0.20); ecNOS 1.58 (0.27)] and absent [iNOS 1.13 (0.24); ecNOS 1.96 (0.25)] were compared (Fig. 2).

Figure 1.

The mRNA expression of iNOS and ecNOS in HUVEC induced by fetal plasma from three study groups. The expression of iNOS mRNA and ecNOS mRNA was significantly higher in the group with umbilical placental vascular disease than those in the group with normal pregnancy (*P < 0.05). The expression of iNOS mRNA and ecNOS mRNA in the group of maternal pre-eclampsia with normal umbilical artery Doppler does not differ to the normal group (P > 0.05). Bars indicate mean (SEM).

Figure 2.

In the umbilical placental vascular disease group, the whole group to the subgroup of absence of maternal pre-eclampsia and subgroup of presence of pre-eclampsia were compared. There were no different expressions of both iNOS mRNA and ecNOS mRNA by HUVEC (P > 0.05). Bars indicate mean (SEM).

There was a significant positive correlation between iNOS mRNA and ecNOS mRNA expression (γ= 0.27; P < 0.05) in each of our groups. Two-way analysis of variance was carried out to determine the effect of smoking, mode of delivery and gestational age on iNOS mRNA and ecNOS mRNA expression within the study groups. There were no significant interactions between smoking, route of delivery and gestational age within the study groups for both iNOS and ecNOS expression. In this study, 65 patients were also included in our study of the endothelial cell adhesion molecule expression11. Two-way analysis of variance showed no difference in the three study groups for the common and uncommon patients.

Discussion

Our study has demonstrated that exposure of umbilical vein endothelial cells to fetal plasma from pregnancies with umbilical placental vascular disease increases the expression of both iNOS mRNA and ecNOS mRNA. However, there was no significant increase in mRNA expression of the two isoforms of nitric oxide synthase when maternal pre-eclampsia was present in the absence of Doppler evidence of vascular disease in the fetal umbilical placental circulation. As a secondary result, we found that nitric oxide synthase activity in the umbilical placental vascular disease group was independent of whether maternal pre-eclampsia was present or absent.

The release of factor(s) causing up-regulation of endothelial cell nitric oxide synthase mRNA expression is a component of the inflammatory response. It may be beneficial or harmful. Nitric oxide may influence many components of the inflammatory cascade including adhesion molecule expression, leucocyte endothelial interaction and infiltration of activated leucocytes. We believe that endothelial cell dysfunction is pivotal in the pathogenesis of umbilical placental vascular disease. In the placental circulation, nitric oxide is a potent vasodilator and may serve to regulate umbilical blood flow and vascular tone in the fetal placental circulation20,21. In the introduction to this study, we presented the results of other studies we have carried out which provide evidence to support our concept that inflammation occurs as a result of injury to the vascular endothelium in umbilical placental vascular disease. Fetal platelet activation and consumption are likely triggered by contact with activated endothelium. Plasma from fetuses who are victims of placental vascular disease contains increased levels of proinflammatory cytokines IL-6 and IL-8 and the soluble fractions of ICAM-1 and PECAM-1. Endothelial cells in culture exposed to this plasma express increased amounts of the genes responsible for synthesis of these substances. Most recently, we have directly extracted endothelial cells from the umbilical placental villi without culture and passage and demonstrated increased gene expression. This is compelling evidence for a local inflammatory response with endothelial cell activation and a proinflammatory cytokine response.

The nitric oxide system has been studied in pre-eclampsia and in fetal growth restriction. In pre-eclampsia, it has been suggested to be the underlying explanation for the vascular endothelial dysfunction widespread through the mother and typical of this disease in its most severe form22. Other studies have suggested that nitric oxide plays an important role in maintaining the fetoplacental circulation. It contributes to both maintenance of basal vascular tone and attenuates the action of vasoconstrictors. The pharmacological actions of nitric oxide in the human fetal placental vasculature in vitro have been studied23. Lyall et al.20 reported that a higher concentration of total combined serum nitrite and nitrate was present in umbilical venous serum from pre-eclampsia. Higher plasma total nitrites were shown in IUGR21. Davidge et al.2 measured nitrite production by endothelial cells after exposure to fetal plasma from pre-eclampsia pregnancies and found increased levels compared with normal pregnancies. These results are not in conflict with our study. Our study differed from these studies in design. We classified our pregnancies by the presence of umbilical placental vascular disease and studied the fetal compartment. Our study group was fetuses who were victims of umbilical placental insufficiency with low birthweight, low centile weight, low placental weight and earlier gestation at delivery. In contrast, in pre-eclampsia with no umbilical artery Doppler flow velocity waveform evidence of vascular disease, we showed no alteration of mRNA expression of nitric oxide synthase. Immunostaining of endothelial cell nitric oxide synthases, which are found in the endothelium of the terminal villous capillary and stem villous vessels, was greater in both IUGR and pre-eclampsia cases24. These findings are consistent with our results. Nitric oxide production may improve blood flow in the placental circulation and limit platelet adhesion and aggregation25. It may be a response to a decrease in fetal oxygenation. In another report, there was no difference in the plasma level of the nitric oxide second messenger cGMP in normal and pre-eclamptic pregnancies26. Although the findings concerning nitric oxide synthase activity in normal pregnancy and pre-eclampsia are sometimes conflicting, there is no evidence of a decreased nitric oxide production in pre-eclampsia27. A satisfactory hypothesis linking the changes in the fetal placental circulation with the maternal uteroplacental circulation and pre-eclampsia remains to be established.

Conclusion

Our results provide evidence that the expression of the two isoforms of iNOS and ecNOS mRNA by HUVECs is elevated in response to exposure to fetal plasma from pregnancies with umbilical placental vascular disease. This finding is independent of the presence or absence of maternal pre-eclampsia. These findings support our view that umbilical placental vascular disease may be a primary pathology and not a consequence of the maternal syndrome of pre-eclampsia. The change in nitric oxide synthase may be a component of injury to the vascular endothelium and the associated inflammatory response in the fetal placenta in this disease. The mechanisms causing the abnormal increase in nitric oxide synthase expression and its relevance to the pathophysiologic mechanisms of umbilical placental vascular disease remain to be established.

Ancillary