Main

Impaired growth in IUGR fetuses with absent or reversed EDV in the umbilical artery is probably secondary to inadequate exchange of nutrients and end products of metabolism at the placenta. Reduced exchange may be due in part to umbilical arterial blood flows that are low, often even when expressed on a per-kg fetal weight basis (1). Low flows may be secondary to high placental resistance caused by an abnormal placental vascular anatomy and/or an abnormally high vasomotor tone. Elevated resistance may be caused by hypovascularity of the IUGR placenta, which has been observed in some studies (2–4) but not all (5, 6). Or, it may be caused by elevated vascular tone secondary to reduced placental nitric oxide synthase levels (7) and/or elevated fetal plasma endothelin (8, 9) and angiotensin II (10) concentrations, which also has been observed in such pregnancies. In any event, maternal-fetal transfer of substances whose transfer is flow-limited (e.g. oxygen) would be reduced by low umbilical arterial blood flows. Indeed, oxygen and glucose extraction from the maternal circulation by the placenta in IUGR pregnancies is impaired to the extent that maternal uterine vein oxygen tensions and glucose concentrations may be abnormally high (11, 12), even though the fetuses themselves may be hypoxic and/or hypoglycemic (13, 14). Fetal hypoxia, in turn, could explain the higher than normal umbilical arterial and venous lactate concentrations sometimes observed in IUGR fetuses (13–15).

On the other hand, low umbilical blood flow cannot explain abnormalities in transfer of other substrates, such as amino acids, that are actively transported from the maternal to fetal circulation by the placenta. Amino acid concentrations are often reduced in plasma of IUGR fetuses, despite maternal concentrations that are as high or higher than those of controls (16, 17). This may be due to reduced amino acid transporter activity, in that transport of the nonmetabolizable neutral amino acid analogs, AIB and methylaminoisobutyric acid (18), is reduced in placental microvillous membrane vesicles isolated from IUGR placentas (19–21).

The objective of the current study was to measure fetoplacental vascular resistance and vascular reactivity, and maternofetal transfer of oxygen, glucose, AIB, and antipyrine in perfused cotyledons from IUGR fetuses with absent or reversed umbilical arterial EDV, compared with preterm and term normally grown controls. We hypothesized that fetoplacental resistance would be elevated and transport reduced in placentas from IUGR pregnancies.

METHODS

Placentas were collected at delivery from term, normally grown fetuses (TN;n = 10), preterm, normally grown fetuses (PN) with normal umbilical arterial EDV (30.8 ± 0.8 wk, n = 6), and preterm IUGR fetuses (<3rd centile) (PI) with absent or reversed umbilical arterial EDV (30.8 ± 0.6 wk, n = 6). All fetuses were morphologically normal. Complete placentas were weighed after the membranes and cord were removed. Clinical data for the three groups are shown in Table 1. Consent was obtained from the mother for the use of placental tissue and access to medical records. The protocol was approved by the University of Toronto Human Ethics Committee. Within 20 min, a peripheral cotyledon was cannulated and the maternal and fetal circulations perfused, by methods described previously (22).

Table 1 Clinical characteristics of mothers, fetuses, and placentas* * TN, term normally grown; PN, preterm normally grown; PI, preterm IUGR;P1, P value for TN versus PN;P2, P value for PN versus PI; A or REDV, absent or reversed end-diastolic velocity.
Full size table

The perfusion medium consisted of Earle's balanced salt solution, with added L-arginine (0.07 g/L; Sigma Chemical Co., St. Louis, MO) and dextran (30 g/L fetal and 7.5 g/L maternal), MW 35,000–50,000 (United States Biochemical, Cleveland, OH). Glucose (0.6 g/L) and heparin (2 USP units/mL; Hepalean, Organon Teknika, Toronto, ON, Canada) were added to the fetal perfusate. Glucose (1 g/L), antipyrine (30 mg/L), and AIB (28 mg/L; Sigma Chemical Co.) were added to the maternal perfusate. Perfusate pH was adjusted to 7.34–7.38 (fetal) and 7.36–7.42 (maternal) with bicarbonate buffer. The maternal perfusate was equilibrated with 95% O2/5% CO2, and the fetal perfusate was equilibrated with 94% N2/6% CO2 (Praxair Products, Mississauga, ON, Canada), except during hypoxia (see below). A miniature temperature probe (Physiotemp, Clifton, NJ) was inserted into the intervillous space, and temperature was maintained between 36°C and 37°C.

The flow rate of the fetal perfusion pump (Masterflex, Cole Parmer Chicago, IL) was adjusted at the start of the stabilization period to achieve a fetal perfusion pressure of 20–30 mm Hg (equivalent to cotyledon perfusion pressure in sheep (23). Flows were then maintained constant for the remainder of the stabilization period and the experiment. Fetal perfusion pressure was measured with an inline pressure transducer (CDXIII; Cobe Laboratories, Lakewood, CO) and was recorded on a chart recorder (Soltec Primeline, Sun Valley, CA). The cotyledon was allowed to stabilize for 45–60 min before beginning the baseline period. Experiments were abandoned if fetal artery pressure was not stable to within 2 mm Hg during this interval, or if, at any point during the study, leakage from the fetal to the maternal side was detected (rate of fetal venous return fell and/or fluid accumulated on the surface of the cotyledon). Leakage in the other direction (i.e. a rise in fetal venous return) was never observed. Mean perfusate flow rates were 0.25 mL min−1 g−1 (fetal) and 0.5 mL min−1 g−1 (maternal), measured with ball-in-tube flowmeters (Gilmont Instruments, Barrington, IL) that were calibrated by timed collection and expressed relative to cotyledon weight determined at the end of the study.

Experiments began with a 30-min baseline period. Perfusion pressure was recorded, and maternal and fetal arterial and venous perfusate samples were collected near the beginning and end of this interval (time-points B1 and B2). Maternal perfusion was then switched to perfusate equilibrated with a gas mixture of 5% CO2/95% N2 to induce a 20-min hypoxic period. Pressures and perfusate samples were collected near the middle and near the end of hypoxia (time-points H1 and H2). Maternal perfusion with the oxygenated perfusate was then resumed, and the cotyledon was allowed to recover for 40 min. This was followed by a second 30-min baseline period with pressure recorded and perfusate samples collected near the beginning and end of this interval (time-points B3 and B4). In most experiments, nitroglycerin (GTN; Sabex, Boucherville, QC, Canada) was then infused for 20 min into the fetal arterial line at an infusion rate of <1% of the fetal arterial flow rate. The final fetal arterial concentration achieved was 10−5 M, a concentration sufficient to induce maximal relaxation (24). Then, after a 20-min recovery period, angiotensin II (0.05 μg; Sigma Chemical Co.) was injected as a bolus into the fetal artery to induce vasoconstriction.

Fetoplacental vascular resistance was calculated as the fetal perfusion pressure divided by the fetal flow rate. Gas tensions and pH were measured with a pH/blood gas analyzer (model 170; Corning Medical, Medfield, MA) within 60 s of sampling. Samples of perfusate were stored frozen (−20°C) until analysis for glucose (YSI model 27; Yellow Springs Instruments, Yellow Springs, OH), lactate, antipyrine, and AIB concentrations. Lactate concentration was measured colorimetrically using a kit based on the conversion of NAD to NADH (Sigma Chemical Co.) on a narrow band spectrophotometer (Shimadzu type UV1600; Mandel Scientific, Toronto, ON, Canada). Antipyrine was also assayed by a colorimetric method (25). AIB concentration was measured with an automated HPLC (model 1050, Hewlett-Packard, ON, Canada) equipped with a fluorescence detector and linear chart recorder. Precolumn derivitization with o-phthaldehyde was used with a mobile phase of 25% acetonitrile in 75% acetate buffer pH 7.2 at a flow rate of 0.6 mL/min. The column was of C18 reversed phase (200 × 4.6 mm with a nominal particle size of 5 μm). Detection was at excitation frequency 365 nm and emission frequency 410 nm.

At the end of the experiment, the cotyledon that had been perfused during the study was fixed by perfusion with 2.5% glutaraldehyde, then dissected out, dry blotted, and the wet weight determined. The entire cotyledon was cut into ∼5-mm cubes that were immersed in fixative in coded vials. Without breaking the code, we randomly selected three cubes from each vial, which were sent to a blinded observer (C.D. Pfarrer) who used scanning electron microscopy to evaluate the villous anatomy according to previously established criteria (2).

Net transfer and consumption by the cotyledon was determined as previously described in detail (22) from concentrations (C) and flows (Q) measured in maternal (M) and fetal (F) arterial (A) and venous (V) catheters where transfer = [(CFV –CFA) × QF] and consumption = [CMA–CMV) × QM −transfer]. Clearances (= transfer/CMA) of AIB and antipyrine were calculated as previously described (26). Transport and consumption values were expressed per unit cotyledon wet weight. ANOVA (Statview 4.5, Abacus Concepts, Berkeley CA), with repeated measures when appropriate, was used to test for differences between the three groups. When group or group-time interactions were significant, separate ANOVAs were performed to compare PI and PN groups and PN and TN groups. Fisher's exact test was used to test proportion for significant differences between groups. Comparisons between PI and TN groups were not made. Baseline measurements refer to B1, B2, B3, and B4 values. Results are presented as mean ± SE where n is the number of placentas.

RESULTS

The preterm IUGR group did not differ significantly in gestational age at delivery from the preterm normally grown group, whereas birth weights and placental weights were significantly reduced (Table 1). Of interest was the fact that the weight of the perfused cotyledon did not differ significantly between groups (PI = 14 ± 3 g; PN = 19 ± 3 g, TN = 20 ± 2 g), nor did fetal perfusion rate (PI = 4.1 ± 0.4 mL min−1; PN = 4.1 ± 0.6 mL min−1, TN = 4.9 ± 0.3 mL min−1) or fetal arterial perfusion pressure (PI = 22.7 ± 0.4 mm Hg; PN = 23.9 ± 0.8 mm Hg, TN = 24.0 ± 0.6 mm Hg at B1). Groups also did not differ significantly from the preterm normally grown group values for fetal arterial perfusate PO2 (27.2 ± 0.8 mm Hg at B1), or for maternal perfusion rate (13.8 ± 0.4 mL/min) and maternal arterial perfusate PO2, either at baseline (563 ± 23 mm Hg at B1) or during hypoxia (27 ± 2 mm Hg at H1). Random tissue samples obtained from preterm placental cotyledons were examined by a blinded observer and scored according to characteristics indicated in Table 2. Five of six cotyledons in the PN group were assigned a normal morphology, and five of six in the preterm IUGR group were assigned an abnormal morphology by this observer (Table 2). The most striking difference between normal (Fig. 1) and abnormal (Fig. 2) villous anatomy was the marked reduction in terminal villi and terminal villous buds in the preterm IUGR group. Villous buds normally contain the terminal capillary loops (2).

Table 2 SEM evaluation of placental morphology of preterm placentas
Full size table
Figure 1
figure 1

Scanning electron micrographs of terminal placental villi from preterm normally grown fetuses:(a) intermediate villi (IV) branch into multiple villous buds (arrows mark artifacts) (×1000, ID# PN 30 (30 wk);(b) buds (marked by asterisks) at higher magnification (×2500, ID# PN 30). Note the absence of fibrin or wrinkling of syncytiotrophoblast surface. (c) High magnification showing one intermediate villous (IV) and three terminal villous buds covered with microvilli (×5000, ID# PN 18, 34 wk).

Full size image
Figure 2
figure 2

Scanning electron micrographs of terminal placental villi from preterm IUGR fetuses. (a) Intermediate villi (IV) branch into terminal villi (TV) that are typically devoid of villous buds, leaving more room for intervillous space (×1000, ID# PI 25, 31 wk). Villous surface wrinkling (arrows) and “microbuds”(arrowhead) were also common. (b) Also observed were abnormally thin terminal villi (arrow) and few normal-sized terminal villi (TV) (×2500, ID# PI 36, 29 wk). (c) Thin terminal villi sometimes had club-shaped tips (arrows) (×2500, ID# PI 25). (d) An intermediate villous showing “microbuds”(asterisks) but normal microvilli on the syncytiotrophoblast surface (×5000, ID# PI 36).

Full size image

Surprisingly, baseline vascular resistance tended to be lower in the preterm IUGR group (93 ± 6 mm Hg ml−1 min g), although there were no significant differences between the groups (PN = 136 ± 13, TN = 118 ± 7 mm Hg ml−1 min g) (Fig. 3). The three groups did not differ significantly in the change in vascular resistance caused by nitroglycerin or angiotensin II (Fig. 3). The increase in resistance caused by hypoxia was similar in the two preterm groups, but the increase was significantly greater in term compared with preterm normally grown fetuses (p = 0.001) (Fig. 3).

Figure 3
figure 3

Fetoplacental vascular resistance (mean ± SE). x axis labels indicate measurements obtained during first baseline period (B1, B2), at 10 and 20 min of hypoxia (H1, H2, respectively), during second baseline period (B3, B4), at end of 20-min infusion of 10−5 M glyceryl trinitrate (GTN), 20 min after GTN (B5), and peak resistance after a 0.05-μg angiotensin II bolus (AII). PI = preterm IUGR. PN = preterm normally grown. TN = term normally grown. Break in x axis is due to different N between B1 and B4, and B4 and AII.

Full size image

There was no significant difference in baseline glucose consumption between preterm (0.21 ± 0.02 μmol min–1 g−1) and term (TN = 0.23 ± 0.02 μmol min–1 g−1) placentas from normally grown fetuses. However, baseline glucose consumption was 2-fold higher in placentas from preterm IUGR (0.44 ± 0.05 μmol min–1 g−1) than in PN fetuses (p = 0.043) (Fig. 4). Baseline oxygen consumption and lactate production tended to be higher in preterm IUGR placentas (Fig. 4), but differences between groups did not reach significance. During hypoxia, there were no significant differences between groups in the significant decrease in oxygen consumption, and significant increases in glucose consumption and lactate production that were observed at this time (Fig. 4). There were also no significant differences between groups in the significant increase in lactate production and oxygen consumption caused by nitroglycerin infusion, although the increase in lactate production was almost significantly greater in placentas from preterm IUGR than in those from preterm normally grown fetuses (p = 0.054) (Fig. 4). Nitroglycerin did not significantly alter placental glucose consumption.

Figure 4
figure 4

Placental oxygen consumption, glucose consumption, and lactate production (mean ± SE). x axis labels and group labels are as in Figure 3, except G = GTN. Break in x axis is due to different N between B1 and B4, and B4 and G. N were as shown on Figure 3, except for glucose consumption by TN placentas where N = 8 from B1 to B4, and N = 6 from B4 to G.

Full size image

Glucose transfer, antipyrine clearance, and lactate production decreased significantly after the first baseline measurement (B1), then remained essentially constant for the remaining baseline measurements, whereas oxygen transfer significantly decreased after hypoxia (B1 and B2 > B3 and B4;p < 0.0001;Fig. 5) with no significant differences between groups. Baseline measurements of glucose transfer, oxygen transfer, and antipyrine clearance did not differ significantly between groups (Fig. 5). During hypoxia, there were no significant differences between groups with regard to the significant decrease in oxygen transfer, glucose transfer, and antipyrine clearance observed at this time (Fig. 5). Nitroglycerin infusion significantly decreased oxygen transfer by 30%—a result that is similar to previous findings in term placentas (22). This response tended to be greatest in preterm control placentas, although the response did not differ significantly between groups (Fig. 5). Nitroglycerin also significantly decreased antipyrine clearance by ∼35% and, again, the response tended to be greatest in the preterm control group, but differences between groups were not significant (Fig. 5). A small decrease in glucose transfer (∼14%) during nitroglycerin was not significant.

Figure 5
figure 5

Maternofetal oxygen transfer, glucose transfer, and antipyrine clearance (mean ± SE). The x axis labels and group labels are as in Figure 3, except G = GTN. Break in x axis is due to different N between B1 and B4, and B4 and G. N were as shown on Figure 3, except for glucose transfer by TN placentas, where N = 8 from B1 to B4 and N = 6 from B4 to G.

Full size image

Clearance of AIB was measured at four time points in the protocol (B1, H2, B3, and nitroglycerin). Clearance did not change significantly from baseline values [PI = 0.08 ± 0.01 (n = 6), PN = 0.07 ± 0.01 (n = 6), TN = 0.07 ± 0.01 (n = 8) mL min –1 g−1] during hypoxia or nitroglycerin infusion, nor did clearances differ significantly between groups.

DISCUSSION

Contrary to our hypothesis, baseline fetoplacental vascular resistance in the perfused placental cotyledons of severely IUGR pregnancies with absent or reversed umbilical arterial EDV was not higher than those obtained from normally grown preterm controls. These data are in agreement with the normal perfusion rates and pressures previously reported for perfused cotyledons from IUGR pregnancies with absent umbilical arterial EDV (27). We recognize that cotyledons were not randomly selected for perfusion. However, evaluation by a blinded observer showed that five of six cotyledons that had been perfused in the study in the PI group showed the same abnormalities in villous structure previously described for random samples obtained from entire IUGR placentas (2). The observed reduction in the number of terminal villi and villous buds suggests that the cotyledons chosen for perfusion probably also had the reduced vascularity often observed in IUGR placentas (2–4).

In the current study, the anatomic abnormalities observed in the cotyledons of severely IUGR pregnancies with absent or reversed umbilical arterial EDV were not associated with an elevation in fetoplacental vascular resistance when expressed per g cotyledon wet weight and measured in vitro. We believe that inaccuracies associated with the measurement of fetoplacental vascular resistance in vitro and/or possible intergroup differences in wet-to-dry-weight ratios cannot entirely explain this result. At the perfusion pressure (24 mm Hg) used in our study, the fetoplacental flow rate per g cotyledon weight (0.25 mL min−1 g−1) would be equivalent to a flow rate of 45 mL min−1 kg−1 birth weight in the term normally grown group. This flow rate is ∼60% less than that of human fetuses at term measured in vivo by Doppler ultrasound (110 mL min−1 kg−1) (28). Whether lower flows in vitro were due to elevated vascular resistance in vitro and/or to our use of an underestimated perfusion pressure is not known. Nevertheless, very large increases in resistance in the order of 300–500% are required to reduce umbilical arterial EDV to zero, based on fetal sheep experiments (29) and theoretical studies (30). Thus, it is highly unlikely that errors associated with the in vitro perfusion technique could entirely account for resistance in the preterm IUGR group being, on average, the lowest of the three groups.

The lack of elevated cotyledon resistance in the preterm IUGR group may result from the fact that hypovascularity is due primarily to reduced capillarization of IUGR placentas (2, 6) and—because most resistance resides at the arteriolar rather than capillary level in systemic circulations—as a consequence, reduced capillarization may have been insufficient to elevate resistance. It is noteworthy, however, that not all studies find significant vascular abnormalities in IUGR placentae (5, 6), whereas others find abnormalities in IUGR placentas that are independent of the umbilical arterial EDV (31). Indeed, evidence of such independence was observed for one placenta in each of the preterm groups in the current study. In any event, results suggest that abnormal EDV in the umbilical artery in vivo in IUGR is not caused solely by elevated vascular resistance secondary to abnormal placental anatomy.

There is considerable experimental evidence to support the association between absent or reversed EDV in the umbilical artery and elevated vascular resistance in the placental microcirculation (29, 32) or venous outflow tract of the placenta (i.e. the umbilical vein, ductus venosus, and/or hepatic circulation) (33). Thus, there appear to be two possible explanations for the results of the current study: (1) that absent or reversed EDV in the umbilical artery is caused by elevated resistance in the venous outflow tract or (2) that increased resistance in vivo is caused by a vascular stimulus that is absent in our in vitro system. The first possibility, that elevated resistance resides within the venous outflow tract, deserves further study. The second possibility is also interesting. It suggests that the cause of the elevated resistance is not anatomic, and therefore, if the cause were known, the elevated resistance might be reversed by the appropriate therapeutic intervention. Possible causes for such an elevation in resistance include placental hypoxia (22, 34) elevated maternal intervillous pressure (35), or elevated circulating endothelin (8) or angiotensin II (10) concentrations.

This study failed to demonstrate the expected impairment in maternal to fetal transfer of oxygen, glucose, and AIB in preterm IUGR cotyledons. Because maternal and fetal flow rates were the same in all three groups in vitro, the similarity in antipyrine clearance under baseline conditions suggests that cotyledons in all three groups had similar functional diffusing capacity per g cotyledon weight. This result was surprising, given the obvious anatomic differences we observed in villous anatomy in the IUGR cotyledons (e.g.Fig. 2). Nevertheless, this result suggests that the activities of the neutral amino acid transporters for AIB (18) and the glucose transporter (likely primarily GLUT1 (36) are indistinguishable from normal in IUGR placental tissue when expressed on a per weight basis and studied under in vitro conditions. Normal AIB transport in the current study contrasts with reduced amino acid concentrations in plasma of IUGR fetuses in vivo (16, 17) and with reduced AIB and methylaminoisobutyric acid transport in microvillous membrane vesicles isolated from IUGR placentas studied in vitro (19–21). The reason for the discrepancy in results is not known. In vivo, net transport of glucose and amino acids may be lower in IUGR because of the smaller size of the placenta. Furthermore, the activity of the transporters might be influenced by circulating substances (e.g. hormones or competitive inhibitors) which were not present in the simple perfusion medium used in our in vitro study.

Despite constant flows, antipyrine transfer and oxygen transfer decreased significantly during both hypoxic vasoconstriction and nitroglycerin-induced vasodilation in all three groups. Transfer was presumably reduced during hypoxia by the reduction in vessel surface area caused by vasoconstriction. Nitroglycerin, however, presumably increased surface area by dilating vessels. However, the indiscriminate vasodilation caused by nitroglycerin may have disrupted normal maternal-fetal flow matching, thereby reducing transfer as suggested previously (22). This observation may have clinical significance, because nitroglycerin has been proposed as a therapy for IUGR (37). Although this treatment may well “improve” umbilical arterial Doppler waveforms by transiently decreasing placental resistance, the vasodilation of fetal vessels that were constricted because they were poorly perfused on the maternal side may actually reduce placental transfer and, therefore, be harmful rather than helpful to the fetus.

Glucose consumption by preterm IUGR placentas was twice that of preterm controls in the current study. The placenta is a major consumer of the glucose taken up from the maternal circulation (26), so that an elevated placental consumption of glucose could contribute to the greater difference in glucose concentration between the maternal and fetal circulation in IUGR fetuses with abnormal umbilical arterial EDV (12). However, in the current study, glucose transfer to the fetal side was not diminished. Instead, increased placental glucose consumption was sustained by increased uptake of glucose from the maternal side. Possibly, glucose transport capacity is so high that doubling placental glucose consumption has an insignificant effect on placental glucose transfer. This suggestion is consistent with normal glucose transporter protein levels in IUGR placentas (36, 38).

In the current study, glucose and oxygen consumption per g remained constant with development from preterm to term in normally grown placentas. This result is in accord with results from large fragments of placental tissue incubated in vitro, but it contrasts with results obtained from small fragments, in which placental glucose and oxygen consumption decreased with advancing gestation (39). As shown previously in the perfused human cotyledon at term, hypoxia increased fetoplacental vascular resistance (34), glucose consumption, and lactate production, and it decreased oxygen consumption (40). The current study further shows that the vasoconstrictive response to hypoxia increased, whereas the metabolic effects remained unchanged with increasing gestation.

In summary, fetoplacental resistance per g cotyledon weight was not elevated in cotyledons from preterm IUGR fetuses with absent or reversed EDV in the umbilical artery relative to preterm controls, even though anticipated differences in villous anatomy were observed. These data suggest that reduced umbilical arterial EDV in such fetuses is not due to an elevation in placental microvascular resistance secondary to an anatomic abnormality. Instead, reduced umbilical arterial EDV may be caused by an elevated resistance in the umbilical venous outflow tract, or by elevated placental resistance caused by a vascular stimulus that is absent in vitro. Net maternofetal transfer of oxygen and glucose, and clearance of aminoisobutyric acid and antipyrine per g cotyledon weight, were also normal in IUGR placentas. However, IUGR placentas were not entirely normal in vitro. Glucose consumption was doubled in preterm IUGR placentas relative to preterm controls. These data suggest that reduced substrate delivery in IUGR fetuses is due to the reduced weight of the IUGR placenta, abnormal fetal and/or maternal flow rates, abnormal fetal and/or placental metabolism, or to the influence of substances not present in our simple perfusion medium.