Compensatory Feto-Placental Upregulation of the Nitric Oxide System during Fetal Growth Restriction

Background Fetal Growth Restriction is often associated with a feto-placental vascular dysfunction conceivably involving endothelial cells. Our study aimed to verify this pathogenic role for feto-placental endothelial cells and, coincidentally, demonstrate any abnormality in the nitric oxide system. Methods Prenatal assessment of feto-placental vascular function was combined with measurement of nitric oxide (in the form of S-nitrosohemoglobin) and its nitrite byproduct, and of the endogenous nitric oxide synthase inhibitor asymmetric dimethylarginine. Umbilical vein endothelial cells were also harvested to determine their gene profile. The study comprised term pregnancies with normal (n = 40) or small-for-gestational-age (n = 20) newborns, small-for-gestational-age preterm pregnancies (n = 15), and bi-chorial, bi-amniotic twin pregnancies with discordant fetal growth (n = 12). Results Umbilical blood nitrite (p<0.001) and S-nitrosohemoglobin (p = 0.02) rose with fetal growth restriction while asymmetric dimethylarginine decreased (p = 0.003). Nitrite rise coincided with an abnormal Doppler profile from umbilical arteries. Fetal growth restriction umbilical vein endothelial cells produced more nitrite and also exhibited reciprocal changes in vasodilator (upwards) and vasoconstrictor (downwards) transcripts. Elevation in blood nitrite and S-nitrosohemoglobin persisted postnatally in the fetal growth restriction offspring. Conclusion Fetal growth restriction is typified by increased nitric oxide production during pregnancy and after birth. This response is viewed as an adaptative event to sustain placental blood flow. However, its occurrence may modify the endothelial phenotype and may ultimately represent an element of risk for cardiovascular disease in adult life.


Introduction
Fetal growth restriction (FGR) complicates up to 8% of pregnancies and carries high perinatal morbidity and mortality. [1] Infants born as small-for-gestational-age (SGA) also present a greater risk for cardiovascular disease during adulthood. [2] No definite explanation is available for the latter occurrence, although a deranged prenatal or perinatal programming of certain mechanisms [3,4,5] may be involved.
An alteration of the feto-placental circulation, reportedly involving the endothelium, plays a key part in the onset and progression of FGR. [6] Among the several agents controlling this vascular district, nitric oxide (NO) appears most important for its known role under normal conditions and as a target for pathological insults, [7] including possibly events that are identified with the development of FGR. 8,9 However, besides being impaired by a variety of stimuli, NO may also have a compensatory function, and several situations in the adult document this possibility. [8].
Our objective was to study NO function in pregnancies complicated by FGR. The approach was comprehensive and included: (i) measurement of NO and its main metabolite, nitrite (NO 2 ), along with the natural NO synthase inhibitor asymmetric dimethylarginine (ADMA); (ii) analysis of Doppler velocimetry in umbilical arteries; and (iii) assessment of gene profile in umbilical vein endothelial cells (HUVEC) collected at the time of delivery. Our ultimate aim was to define the nature of any linkage between NO and FGR and, coincidentally, provide a possible insight into the alleged negative impact of FGR on adult health.

Study Groups
The study comprised three groups. Group 1: pregnancies being scheduled for umbilical Doppler velocimetry at 36 wk and yielding term infants with average (AGA) and small-forgestational age (SGA; ,2.5 kg at term or ,10th percentile body weight) body weight (BW) (n = 60). Group 2: pregnancies with isolated preterm FGR (,10th percentile body weight; mean gestational age 3461 wk, n = 15). Group 3: bi-chorionic, biamniotic twin pregnancies with offspring of discordant growth (mean gestational age 3562 wk, n = 12 pregnancies). All patients included in the study were normotensive. A subset of neonates (n = 45, equally distributed among the groups 1 and 2) were subdivided according to BW in normal (75th-25th centile) (n = 15), at the limit of normality (,25th-10th centile) (n = 15) and in the SGA category (,10th centile, with ,3rd centile being considered a severe SGA, n = 15). Mean 6 SD of BW centiles for each group was as follows: 75th-25th centile: 52.82614.45; ,25th-10th centile: 17.1364.23; ,10th-3rd centile: 8.2461.75; ,3rd centile: 1.6160.51. BW profile for the population is given in Fig. S1. SGA infants were further characterized by the Rohrer's ponderal index (PInd) (weight in grams* 100/length in cm 3 ). Average PInd with appropriate-for-gestational age infants at term is 2.4. [9] PInd decreases with FGR, because abdominal growth lags behind the increase in length (asymmetric growth). Thus, in its low range PInd can be used to characterize further the grade of FGR in newborns. Cord blood (artery and vein) was collected in all cases, while peripheral blood (heel sampling) was obtained from certain newborns at the time of delivery and 24 or 72 h afterwards. Blood being left after routine procedures was used for the latter purpose. Clinical characteristics of groups and demographic details are found in Tables S1, S2, and S3.

Ethics Statement
All participants provided written informed consent. Parents provided written approval for the procedures on the newborns, which did not add to the standard routine. The Ethics Committee of the University of Pisa approved protocol and procedures of the study.

Doppler Velocimetric Analysis
Doppler ultrasound analysis was performed on the umbilical arteries with a Technos scanner (Esaote Biomedica, Genova, Italy) being equipped with color flow imaging, pulsed Doppler capability and convex probes (3.5-5 MHz for transabdominal examination). Records were stored for subsequent retrieval, and the same person (F.A.L.S.) was responsible for all the examinations. Patients were studied in recumbent position after a 10-min rest. The result was labeled as abnormal in the case of absent or reversed end-diastolic flow. Similarly abnormal was considered any end-diastolic flow, with a peak-systolic/end-diastolic velocity ratio exceeding the mean value of the group by 2 SD.

Cell Culture
HUVEC from a subset of patients (n = 36, equally distributed among the indicated groups) were isolated at the time of delivery and were grown in culture according to a previous protocol [10] for measurement of endothelial nitric oxide synthase (eNOS) protein and activity. Miniarray analysis was performed on the same population of cells.

Analytical Procedures
NO content of blood was assessed by measuring NO 2 [10] and nitroso-hemoglobin (NOHb). The latter assay was carried out with the X-band electron paramagnetic resonance spectrometer ALEXIS (Bruker Biospin, Karlsruhe, Germany) in a high sensitivity microwave cavity. [11] Whole blood with known concentrations of NO 2 (1-10 mM) and Na 2 S 2 O 4 (20 mM) served as a reference. [12] ADMA was assayed in the same samples with a commercial kit (Pantec, Torino, Italy). eNOS activity of HUVEC was derived from the degree of arginine-to-citrulline conversion, [10] and NO 2 was measured in parallel in the cell incubate.

Western Immunoblotting
HUVEC lysates were separated by SDS-PAGE and then subjected to immunoblotting for eNOS. Briefly, the primary antibody (rabbit anti-human eNOS, 1:1000; Transduction Laboratories, Lexington, KY) was incubated with the blotted membranes overnight at 4uC and blots were subsequently hybridized with a secondary antibody (goat anti-rabbit, 1:4000; Transduction Laboratories). b-Actin (Santa Cruz Biotechnology, Santa Cruz, CA) served as reference. The resulting chemiluminescent signal was recorded with a quantitative digital imaging system capable of recognizing any saturation of the input (Quantity One; BioRad, Hercules, CA). Results are given as the eNOS/b-actin ratio in arbitrary units (control = 1), with conditions ensuring the reading of signals over a linear range.

Miniarray Analysis
Total RNA was extracted from HUVEC with the guanidine thiocianate-phenol-chloroform mixture, and hybridization of an aliquot (4 mg) was performed with Human Endothelial Cell Biology Oligo GEArrays (Superarray Biosciences, Frederick, USA). Singleton pregnancies being scheduled for umbilical Doppler velocimetry at 36 wk and yielding term infants with average (AGA) body weight were studied comparatively with isolated severe FGR singleton pregnancies (n = 6 per group). A total of 114 genes were examined (Table S4), all being selected for their general relevance to endothelial function. Gene expression was measured by chemiluminescence, taking care to ensure the same exposure for each membrane. Quantitative analysis was obtained with a digital system excluding any saturation (Quantity One; BioRad). The signal for each gene was measured against housekeeping genes incorporated into the system, as recommended by the manufacturer. Background noise was also subtracted in the process. A 5-fold change of expression in either direction was taken as the limit for acceptance of data.

Statistics
Data are presented as the mean 6 SD. Comparisons were made using Student's t-test or ANOVA followed by the Tukey test. Regression analysis was performed with the best fitting for a linear or exponential relationship. Differences are considered significant with p,0.05.

NO2, NOHb and ADMA in Umbilical Blood
Regression analysis showed a significant inverse relationship with exponential course between NO 2 levels in the umbilical vein and growth of newborns (R 2 = 0.809; p,0.001, Fig. 1A). NO 2 elevation was particularly marked with BW below the 10th centile (Fig. 1A). A similar pattern emerged from umbilical artery sampling, but the magnitude of the NO 2 change was smaller and a linear inverse relationship between NO 2 and BW values was evident only from the 25 th centile downwards (R 2 = 0.262; p = 0.018, Fig. 1B).
NOHb was also elevated in umbilical blood of FGR pregnancies, regardless of whether it originated from the artery or the vein ( Figg. 1C and 1D). However, at variance from NO 2 , the inverse relationship linking NOHb with BW was linear with the vein (R 2 = 0.169; p = 0.024, Fig. 1C) and exponential with the artery (R 2 = 0.705; p,0.001, Fig. 1D).
Contrary to NO 2 and NOHb, changes in ADMA levels were less consistent and varied in sign depending on vessel and degree of FGR. As shown in Fig. 1E, findings with the umbilical vein paralleled those for NO-related variables, and a linear inverse relationship linked ADMA content with BW (R 2 = 0.226; p = 0.003). No such relationship was noted instead with the umbilical artery (Fig. 1F), and ADMA was either decreased (,10 th centile) or marginally increased (,25 th -10 th centile).
A separate analysis was carried out with SGA pregnancies of BW below the 10 th centile, where levels of compounds in umbilical vein and artery were correlated with the PInd of infants. For this purpose, a PInd of 2 was taken as the dividing limit between symmetric and asymmetric growth groups (see Methods). As shown in Fig. 2, pregnancies with PInd below 2 had significantly higher amounts of NO 2 and NOHb in both vessels. Conversely, under the same conditions ADMA level was lower only in the umbilical artery.
eNOS Function in HUVEC eNOS protein expression and activity and NO 2 release were significantly higher in HUVEC from the 25th centile BW group downwards (Fig. 3A and 3B). Furthermore, the same pattern was seen when relating these variables to the PInd of newborns with BW below the 10 th centile. Fig. 3C documents this point, even though the subset of pregnancies was distributed over a relatively narrow range compared to that considered in Fig. 2.

Umbilical Artery Doppler Analysis vs. NO Markers in Umbilical Blood and Cells from Pregnancies with Normal or Restricted Fetal Growth
A subset of the patients, including preterm and term pregnancies with FGR, also had a Doppler velocimetric analysis of the umbilical arteries. Limited to this group, NO 2 levels in umbilical vein and artery blood were increased with an abnormal Doppler profile (Fig. 4A), while in contrast NOHb and ADMA levels were decreased ( Fig. 4B and 4C). This finding appeared to be independent of fetal growth status or gestational age. In addition, eNOS protein and activity in HUVEC, with the attendant NO 2 in the incubate, were elevated when dealing with SGA pregnancies with Doppler derangement (Fig. 4D and 4E).

NO 2 and NOHb in Umbilical Blood from Twins with Discordant Intrauterine Growth
Umbilical vein NO 2 did not differ between normal (NT) and small (ST) twin when Doppler analysis was normal (Fig. 5A), while in contrast the compound was elevated in the umbilical artery of the small twin (Fig. 5B). No difference between the two vessels was noted with an abnormal Doppler, and NO 2 levels were consistently higher in the SGA infant of the pair (Fig. 5A and 5B).
NOHb, on the other hand, was elevated in both umbilical vessels of the SGA infant even in the absence of any anomaly in the velocimetric analysis ( Fig. 5C and 5D). However, when SGA was associated with an abnormal Doppler, NOHb abated in the umbilical artery (Fig. 5D) and tended to rise in the vein (Fig. 5C). Blood levels of ADMA in either vessel did not differ, regardless of BW and the state of umbilical artery blood flow (data not shown).

Blood NO 2 and NOHb in Newborns with Normal or Restricted Growth
As shown in Fig. 6 (A, B), newborns below the 25th centile presented a marked increase in blood NO 2 and NOHb at 24 h of age which, in certain cases, was maintained up to the 72-h. Conversely, a modest elevation, only for (NO 2 ), or no elevation at all was seen in the normal offspring. Furthermore, infants with an abnormal Doppler profile tended to have higher levels of blood NO with both term and preterm pregnancies (Fig. 6C). No such correlation was noted in the case of NOHb (Fig. 6D).

Discussion
Our study shows activation of the NO system in pregnancies with FGR, specifically when placental vascular resistance is impaired. Elevation of NO 2 and NOHb is ascribed to enhanced eNOS activity, within and outside the placenta. This adaptive response presents distinctive features in HUVEC where appro-priate changes are seen in genes controlling blood flow, angiogenesis, extracellular matrix remodeling and inflammation. Furthermore, a NO enhancement is also evident after birth. Hence, findings suggest the existence of a re-programming of the feto-placental vascular system, and of the fetus as a whole, in the face of adverse intrauterine conditions. Expectedly, its immediate purpose is to preserve an adequate placental blood flow. An open question, however, is whether this event translates into a conditioning process for SGA newborns and, as a corollary, into an element of risk later in life. A possibility in this respect is that, in line with evidence from other sites [13,14], an extended activation of the NO system is followed by down-regulation, specifically of vascular eNOS. The relative hyperoxia of the neonate would favor any such occurrence through depletion of tetrahydrobiopterin (BH 4 ) secondary to oxygen radical formation. [15] Recent data in the rat are also consistent with the idea of an altered hemodynamic control in the adult resulting from FGR-linked NO dysfunction. [16] If confirmed, this concept could open the way to new therapeutic strategies, based on BH 4 supplementation and/or free radical scavenging, to improve the postnatal adaptation of SGA infants.
The source of NO 2 in the feto-placental circulation of FGR pregnancies remains to be determined. Pregnancies complicated by FGR show increased eNOS expression in syncytiotrophoblast cells. [17] Our finding of greater eNOS and iNOS expression in HUVEC confirms the importance of any up-regulation of NOS transcript(s) in the adaptation of the feto-placental vascular district to FGR. One should mention, however, that a reduction of eNOS activity has also been reported, [18] hence highlighting how this enzyme may be conditioned by a host of factors with opposing effects. A related question is whether agents that inhibit eNOS, such as ADMA, [19] are causally linked to FGR, in analogy to observations made with pre-eclampsia. [20] We found that, in the case of FGR, ADMA levels are decreased in the umbilical artery, while the opposite occurs in the vein. This finding accords with the idea of the fetus being able to reset its NO system upwards, and this event, along with a decreased production of ADMA, is conceivably crucial in preserving placental vascular flow. Any such resetting also applies to pregnancies with a severely growthrestricted fetus, where the decreasing ponderal index coincides with higher amounts of NO 2 and NOHb in the umbilical vein. Indeed, this particular relationship is also evident after birth, since infants belonging to the low-centile groups (,25 th centile) present a more active NO system compared to those within a normal range (75 th -25 th centile).
Significantly, we identified erythrocyte-based NOHb as a potential source for the elevated NO in FGR, even in conditions of altered feto-placental resistance. Methodologically, the measurement of NOHb has been introduced as a suitable way to monitor circulating NO in animals and humans. [12] Our report highlights how this novel technique can be exploited to assess NO function during pregnancy, with the attendant possibility of better defining clinical conditions such as an altered feto-placental vascular resistance or FGR. The existence, in fact, of a good matching between NO 2 /NOHb levels in the feto-placental Figure 7. Gene profile in HUVEC from SGA vs. AGA pregnancies. Comparison of gene expression at 75 th -25 th (black squares) vs. ,3 rd (white squares) BW centile. Figure reports transcripts being affected by fetal growth restriction that relate to hemodynamic control (A), angiogenesis (B), extracellular matrix turnover (C) and inflammatory/adhesion process (D). Values are expressed as the intensity ratio between test and reference genes (for details, see Methods). Note that the total gene cohort under examination is given in Table S4. *p,0.05; **p,0.01; ***p,0.001 for ,3 rd vs. 75 th -25 th BW centile. doi:10.1371/journal.pone.0045294.g007 circulation and indices of vascular resistance/flow provides the basis for any such approach.
NO function in FGR is sustained, in part, by an enhanced eNOS activity in feto-placental endothelial cells. Our analysis suggests the occurrence of a reset in endothelial function, being directed to the maintenance of blood flow in umbilical vessels. A qualifying event in this process is the reciprocal variation in eNOS (upwards) and angiotensin-converting enzyme/angiotensin receptor (downwards) expression. Coincidentally, HUVEC from SGA pregnancies manifest a gene expression pattern that would be unfavorable for angiogenesis and vascular remodeling. Some of these gene products -VEGF, VCAM-1 and ICAM-1-have previously been implicated in FGR or pre-eclampsia and have also been assigned predictive value for these conditions. [21,22,23] The overall picture emerging from these data is that feto-placental endothelial cells play a complex role in the pathophysiology of FGR, possibly by contributing to processes that limit the development of the placenta and its responsiveness to hypoxia.
Demonstration of a differential relationship between BW and umbilical blood levels of NO 2 and NOHb in twins with discordant growth further supports a prime role of the feto-placental vasculature, and the fetus as a whole, in the natural course of FGR. Maternal factors are, in fact, common to the two situations and cannot account for the observed differences.

Conclusion
In conclusion, we find that FGR pregnancies present an adjustment in the NO system being characterized by elevation of NO 2 and NOHb and reduction of ADMA. Endothelial cells of the feto-placental circulation are part of this response through a phenotypic change that attests to growth restriction and hemodynamic dysfunction. Assessment of NO function in the clinical setting could become important for the prevention and manage-ment of FGR. Furthermore, if future studies were to confirm that changes in the NO system in SGA infants are linked to increased susceptibility to cardiovascular disorders later in life, new strategies may ensue for identification of individuals at risk and introduction of preventive measures.