Adaptive responses in uteroplacental metabolism and fetoplacental nutrient shuttling and sensing during placental insufficiency

Abstract

Glucose, lactate, and amino acids are major fetal nutrients. During placental insufficiency-induced intrauterine growth restriction (PI-IUGR), uteroplacental weight-specific oxygen consumption rates are maintained, yet fetal glucose and amino acid supply is decreased and fetal lactate concentrations are increased. We hypothesized that uteroplacental metabolism adapts to PI-IUGR by altering nutrient allocation to maintain oxidative metabolism. Here, we measured nutrient flux rates, with a focus on nutrients shuttled between the placenta and fetus (lactate-pyruvate, glutamine-glutamate, and glycine-serine) in a sheep model of PI-IUGR. PI-IUGR fetuses weighed 40% less and had decreased oxygen, glucose, and amino acid concentrations and increased lactate and pyruvate versus control (CON) fetuses. Uteroplacental weight-specific rates of oxygen, glucose, lactate, and pyruvate uptake were similar. In PI-IUGR, fetal glucose uptake was decreased and pyruvate output was increased. In PI-IUGR placental tissue, pyruvate dehydrogenase (PDH) phosphorylation was decreased and PDH activity was increased. Uteroplacental glutamine output to the fetus and expression of genes regulating glutamine-glutamate metabolism were lower in PI-IUGR. Fetal glycine uptake was lower in PI-IUGR, with no differences in uteroplacental glycine or serine flux. These results suggest increased placental utilization of pyruvate from the fetus, without higher maternal glucose utilization, and lower fetoplacental amino acid shuttling during PI-IUGR. Mechanistically, AMP-activated protein kinase (AMPK) activation was higher and associated with thiobarbituric acid-reactive substances (TBARS) content, a marker of oxidative stress, and PDH activity in the PI-IUGR placenta, supporting a potential link between oxidative stress, AMPK, and pyruvate utilization. These differences in fetoplacental nutrient sensing and shuttling may represent adaptive strategies enabling the placenta to maintain oxidative metabolism.

NEW & NOTEWORTHY These results suggest increased placental utilization of pyruvate from the fetus, without higher maternal glucose uptake, and lower amino acid shuttling in the placental insufficiency-induced intrauterine growth restriction (PI-IUGR) placenta. AMPK activation was associated with oxidative stress and PDH activity, supporting a putative link between oxidative stress, AMPK, and pyruvate utilization. These differences in fetoplacental nutrient sensing and shuttling may represent adaptive strategies enabling the placenta to maintain oxidative metabolism at the expense of fetal growth.

INTRODUCTION

Normal fetal growth and oxidative metabolism are fueled by glucose, lactate, and amino acids delivered across the placenta (1). In response to reduced oxygen and nutrient supply, placental and fetal tissues optimize their allocation and utilization of these substrates to maintain oxidative metabolism (2). During placental insufficiency and fetal growth restriction in humans and sheep models, fetal oxygenation and glucose concentrations are lower and lactate concentrations are higher (3–5). However, the mechanisms governing the exchange of these and other nutrients between the fetus and the placenta remain incompletely understood. Glycolysis produces pyruvate, which can be converted to lactate or oxidized to acetyl CoA. Under normal conditions, fetal circulating pyruvate concentrations are low, ∼1/20th that of lactate on a molar basis, with little net release of pyruvate by the fetus to the placenta (6–10). However, during sustained hypoxemia, fetuses have higher lactate and pyruvate concentrations and greater pyruvate output to the placenta (6, 7, 9, 11, 12). This suggests an accelerated fetoplacental lactate-pyruvate shuttle during hypoxemia whereby the fetus releases pyruvate to the placenta and, in turn, placental tissues preferentially use fetal pyruvate, over maternal glucose, to produce lactate as a fuel for the fetus (6). Other fetoplacental nutrient shuttles have been described in normal pregnancies. The amino acids glutamate and serine have flux from the fetus to the placenta that is matched by a reciprocal net uptake of glutamine and glycine, respectively, resulting from an exchange between the placenta and the fetal liver (13–16). However, little is known about the effect of placental insufficiency-induced intrauterine growth restriction (PI-IUGR), which features hypoxemia, on lactate-pyruvate shuttling or the amino acid shuttles. Alterations in these shuttles may be an important strategy to optimize substrate utilization and maintain oxidative metabolism in placental and fetal tissues (5, 6, 17).

Nutrient sensors coordinate metabolism in response to nutrient, oxygen, and growth factor signals (18–20). AMP-activated protein kinase (AMPK) is activated by low nutrient levels and hypoxemia, whereas AKT and mechanistic target of rapamycin (mTOR) protein complex (mTORC)1 are activated in response to increased nutrients and anabolic hormones (20–22). AMPK activation promotes glycolysis and mitochondrial function to restore cellular ATP levels (21–24) and can activate pyruvate dehydrogenase (PDH) (25, 26), which catalyzes the oxidation of pyruvate to acetyl CoA for entry into the tricarboxylic acid (TCA) cycle. AKT is a major signaling target for insulin action, among other anabolic growth factors (20). In the placenta, mTOR activation functions as a nutrient sensor that parallels nutrient supply and increases anabolic and growth pathways (27, 28). AKT mediates mTOR activation through repression of upstream inhibitors of mTOR protein (20, 22). Furthermore, AMPK can suppress mTOR activity, providing coordinated cross talk between these nutrient sensors. Thus, understanding AMPK, AKT, and mTOR signaling in the PI-IUGR placenta is important given the multiple input signals and downstream metabolic pathways that these nutrient sensors integrate.

In the present study, we leveraged a sheep model to test the hypothesis that uteroplacental metabolism adapts to PI-IUGR by altering the allocation of nutrients between the placenta and fetus, therefore enabling placental tissues to maintain oxygen consumption rates. To test this, we performed in vivo metabolic studies to measure uterine, uteroplacental, and fetal oxygen and nutrient flux rates, complemented with placental tissue molecular and biochemical measurements of regulatory genes, proteins, and enzymes involved in nutrient metabolism and mitochondrial function. We also evaluated AMPK, AKT, and mTOR signaling through analysis of placental cotyledon tissue. This approach allowed us to evaluate the effect of PI-IUGR on nutrients shuttled between the placenta and fetus (lactate-pyruvate, glutamine-glutamate, and glycine-serine) and the mechanisms enabling the placenta to maintain oxidative metabolism.

METHODS

Sheep Model of Placental Insufficiency

All animal procedures were approved by the Institutional Animal Care and Use Committee at the University of Colorado. Pregnant Columbia-Rambouillet ewes supplied from Nebeker Ranch (Lancaster, CA) were studied at the Perinatal Research Center, Aurora, CO, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International. Experimental work was performed and reported according to the Animal Research: Reporting of In Vivo Experiments guidelines (29). PI-IUGR fetuses were created by exposing pregnant ewes to elevated humidity and temperature (40°C for 12 h, 35°C for 12 h) from ∼37 days gestation age (dGA; term = ∼147 dGA) to ∼116 dGA in an environmentally controlled room as previously described (3, 5). Control (CON) fetuses were from pregnant ewes exposed to normal humidity and temperatures daily (25°C) in an environmentally controlled room and pair-fed to the intake of the PI-IUGR ewes. After treatment, all ewes were exposed to normal humidity and temperatures until time of study. All ewes were carrying singleton pregnancies. Feed and water intake logs and medical records were maintained daily. Ewes were housed in individual carts during the duration of experimental procedures.

Surgery for Placement of Chronic Vascular Catheters

Surgery was performed at ∼125 days of gestation (∼147-day gestation length) to surgically place indwelling polyvinyl catheters (20 gauge) in the maternal and fetal vasculature (6, 30). Ewes were fasted for 24 h before surgery. A maternal jugular catheter was placed for administration of diazepam (0.2 mg·kg⁻¹) and ketamine (17.5 mg·kg⁻¹), and ewes were then maintained on isoflurane inhalation anesthesia (2–5%) for the remainder of the surgical procedure. At surgery, procaine penicillin G (600,000 U im) and ampicillin (500 mg intra-amniotically) were administered before closing uterine and abdominal incisions. Maternal catheters were placed in the femoral artery and femoral vein via a groin incision. Fetal catheters were placed in the common umbilical vein, fetal artery (advanced into the abdominal aorta), and femoral vein (advanced into the inferior vena cava). A uterine vein catheter was placed and advanced into the common vein draining the pregnant uterus. An additional catheter was placed in the fetal left hepatic vein, as previously described (30), but data obtained from those samples are not presented here. Catheters were filled with 5% heparinized saline and subcutaneously tunneled to the ewe’s flank, exteriorized through the skin, and kept in a plastic pouch sutured to the skin. Flunixin meglumine analgesic (Banamine, 2.2 mg·kg⁻¹ im) and probiotics (Probios, 10 g oral) were administered for 72 h postoperatively to the ewe.

Metabolic Nutrient Uptake Study and Tissue Collection

After a minimum of 5 days of postoperative recovery, metabolic studies were performed (∼132 days) to measure uterine and umbilical blood flow and oxygen and nutrient uptake rates. A 3-mL bolus of ³H₂O was infused into fetal vein catheters, followed by a continuous infusion at 3 mL/h (15 μCi/mL ³H₂O) to measure blood flow as previously described (6). To trace glucose metabolism, a [6,6-²H]glucose stable isotope tracer was infused into the maternal circulation at 12 mL/h (50 mg/mL with 12-mL bolus) and a [U-¹³C]glucose tracer was infused into the fetal circulation at 3 mL/h (30 mg/mL [U-¹³C]glucose with 3-mL bolus). After 90–120 min, blood was simultaneously sampled from the maternal artery, uterine vein, umbilical vein, and fetal artery four times at 20- to 30-min intervals to characterize the steady-state period. Fetal blood was replaced isovolumetrically with heparinized maternal arterial blood (15 mL·h⁻¹) throughout the steady-state period.

Immediately after the metabolic study, ewes were anesthetized with intravenous diazepam (0.2 mg·kg⁻¹) and ketamine (17.5 mg·kg⁻¹) to deliver the fetus via maternal laparotomy and hysterotomy. Subsequently, a lethal dose of sodium pentobarbital (390 mg·mL⁻¹, Fatal Plus; Vortech Pharmaceuticals) was administered intravenously to euthanize the ewe and fetus. For each animal, fetal weight was recorded and organs were dissected, weighed, and snap-frozen in liquid nitrogen. Total uterine weight was measured, followed by dissection and measurement of the uterine membrane, uterine tissue, and placentome weight. Individual placentomes were classified into categories (A, B, C, D) based on gross morphological appearance to determine whether hypoxia affected the proportion of placentomes across categories (31). Representative type A and B placentomes were selected (n = 3 from each animal) to allow for similar comparisons between groups that are independent of gross morphology differences. Each placentome was separated into caruncle or cotyledon sides and snap-frozen. To obtain homogeneous tissue samples for downstream analysis, the cotyledon tissues from each animal were combined and ground in liquid nitrogen.

Biochemical Analyses

In all venous and arterial samples, whole blood Po₂ and O₂ content were measured with the ABL 800 Flex blood gas analyzer (32). Plasma glucose and lactate concentrations were measured with the Yellow Springs Instrument model 2900 Select Biochemistry Analyzer. Pyruvate concentrations were determined in deproteinized whole blood samples (30, 33). Plasma amino acids were measured with a Dionex ICS 5000+ high-pressure ion chromatograph with Pickering PCX Pinnacle 120-4 channel variable wavelength detector for postcolumn derivatization and ultraviolet detection (Thermo Electron North America LLC) (34). Plasma ³H₂O concentrations were measured by liquid scintillation (5, 30).

Glucose Tracer Enrichments

Glucose tracer enrichments [molar percent excess (MPE)] for [U-¹³C]glucose (m + 6) were measured in the fetal artery and umbilical vein plasma samples as previously described (32). For the [6,6-²H]glucose tracer, glucose tracer MPEs (m + 2) were measured in maternal artery, uterine vein, fetal artery, and umbilical vein. Glucose was converted to the aldonitrile peracetate derivative for GC-MS analysis. Glucose [U-¹³C] enrichment was monitored at m/z of 334/328 ratio, and glucose [6,6-²H] enrichment was monitored at m/z of 330/328 ratio. Glucose MPE was calculated as the difference in peak area ratios between unenriched (baseline) and enriched samples.

Calculations

We studied 6 CON and 10 PI-IUGR maternofetal units. Because of an umbilical venous catheter failure, fetal uptake data and uteroplacental uptake data are not available on one PI-IUGR animal.

Uterine and umbilical blood flow were determined by steady-state diffusion of ³H₂O (35). Uterine uptake rates were calculated with the Fick principle, multiplying uterine blood flow by maternal artery-uterine vein difference in substrate concentration. Umbilical (fetal) uptake rates were calculated with the Fick principle, multiplying umbilical blood flow by umbilical vein-fetal artery difference in substrate concentration (35–37). Net uteroplacental uptake rates were calculated as the difference between absolute rates of uterine minus umbilical uptake. Fetal glucose utilization rate using [U-¹³C]glucose MPE and [6,6-²H]glucose MPE was calculated with net tracer uptake rates as previously described (32). Both tracers produced similar rates (r² = 0.90 across all fetuses, P < 0.001), and the average utilization rate measured by both tracers was calculated. Uterine glucose utilization rate was calculated with net uterine tracer uptake for glucose [6,6-²H] and maternal arterial MPE. Uteroplacental glucose utilization rate was calculated as the difference between uterine and umbilical glucose utilization rates, measured with [6,6-²H]glucose.

Nutrient-oxygen metabolic quotients were calculated for glucose, lactate, pyruvate, and amino acids (38–40). Umbilical substrate:oxygen quotients were calculated by dividing the whole blood umbilical vein-fetal artery difference in substrate concentration by the umbilical vein-fetal artery difference in whole blood O₂ content, multiplied by the number of oxygen molecules required to oxidize one molecule of substrate. Uterine substrate:oxygen quotients were calculated similarly except using maternal artery-uterine vein substrate and oxygen differences. Uteroplacental quotients were estimated with the net uptake of each substrate multiplied by the number of oxygen molecules required and divided by the uteroplacental oxygen utilization rate (41).

Gene Expression

Total RNA was isolated from placental cotyledon tissue and reverse transcribed to cDNA, and qPCR was performed as previously described (42–44). Primers were used as previously reported (6, 45, 46). Relative expression was calculated from standard curves for each gene. The geometric mean of seven reference genes was calculated with ACTB, PPIB, RPL37A, RPL41, RPS15, HPRT1, and 18 s and used to normalize qPCR results. These genes demonstrate stability in expression here and in our studies in sheep placental tissue (6, 45). Data are expressed relative to the mean of the CON group.

Protein Expression

Whole cell lysates were prepared from placental cotyledon tissue, and Western blotting was performed as previously described (6). The antibodies used were as previously reported and validated in ovine samples (5, 6, 42, 47). Briefly, 30 µg of protein was loaded with 4× DTT (1 M) in equal volumes of buffer, separated on 4–12% polyacrylamide gels, and transferred onto nitrocellulose membranes (Bio-Rad). Antibody specificity was verified by the presence of a single band at the expected molecular mass as indicated in figures. Bands for phosphorylated and total forms of a protein were verified to be of similar size based on migration in gels when blot images were aligned. Protein bands were visualized with IR-Dye IgG secondary antibody (LI-COR) and protein expression quantified with Image Studio (LI-COR). All samples were run on the same gel. For phosphorylated proteins, data are expressed as a ratio of phosphorylation to total expression in addition to absolute levels of phosphorylated and total protein expression. Before blocking and antibody incubations, the equality of sample loading was measured with the Total Protein Stain (LI-COR). LDH-A protein expression and oxidative phosphorylation proteins were normalized to the total protein stain quantification specific to each blot. Data are presented as fold change relative to the mean of the CON group.

PDH Enzyme Activity

Enzyme activity of PDH was measured (MAK183, Sigma-Aldrich) in placental cotyledon tissue (40 mg) homogenized in 400 µL of ice-cold PDH Assay Buffer as previously described (32). Protein concentrations were determined with a Pierce BCA Protein Assay (ThermoFisher Scientific), and 10 μg of protein was loaded per reaction in duplicate. The PDH assay reaction was performed at 37°C, and absorbance at 450 nm (A₄₅₀) was measured every 5 min for 30 min. The ΔA₄₅₀ was calculated for 15 min of the linear reaction and is proportional to the NADH concentration produced by PDH enzymatic reaction converting pyruvate into acetyl CoA. Results were normalized to the amount of protein loaded in the reaction.

LDH Enzyme Activity

Placental cotyledon tissue (50 mg) was homogenized in 500 µL of ice-cold CelLytic MT Buffer (Sigma-Aldrich), and LDH enzyme activity was assessed with the LDH Activity Assay (ab102526, Abcam) as previously described (32). Protein concentrations were determined as described above, and 1 μg of protein was loaded per reaction in duplicate. LDH assay reactions were performed at 37°C, and A₄₅₀ was measured every 5 min for 40 min. The ΔA₄₅₀ was calculated for 20 min of the linear reaction between 20 and 40 min time points and is proportional to the NADH concentration produced by the LDH enzymatic reaction converting lactate into pyruvate. Results were normalized to the amount of protein loaded in the reaction.

Thiobarbituric Acid-Reactive Substances

Thiobarbituric acid-reactive substances (TBARS) content was measured colorimetrically (no. 700870, Cayman) in placental cotyledon tissue protein lysate samples prepared as described for Western blotting (43). Results are expressed relative to protein content.

Statistical Analysis

Data were analyzed by unpaired Student’s t test or Mann–Whitney U test, when variances were different between groups using F test for equality of variances (as indicated in table and figure legends), and linear regression and correlation analyses were performed with GraphPad Prism 9.0 (GraphPad Software). The analysis used is indicated in the figures and tables. Data are presented as means ± SE. Statistical differences are declared at P ≤ 0.05, and statistical trends toward significance at P ≤ 0.15 are shown. In a secondary analysis, fetal sex effects were evaluated by a two-way ANOVA with fixed effects of treatment (CON, PI-IUGR) and fetal sex. No significant effects of fetal sex were found, and since our study was not powered to detect sex differences and our prior work demonstrates that fetal sex has little effect on blood flow, oxygen, and amino acid fetal uptake rates in in a larger cohort of CON and PI-IUGR fetal sheep (3), fetal sex was not included in the final analyses and data from female and male fetuses were combined. Since fetuses in the PI-IUGR group have a range in fetal weight, PI-IUGR fetuses weighing >2 kg are depicted as pink data points and PI-IUGR fetuses weighing <2 kg are shown as red data points in all graphs. Since this study was not powered to detect effects based on severity within the IUGR, all PI-IUGR samples are shown together for visualization of possible effects related to growth restriction severity based on color labeling.

RESULTS

Fetal and Placental Growth Characteristics

Male and female fetuses in both groups were studied at similar gestational ages. Animals in the PI-IUGR group had 20% fewer placentomes and 40% lower placental weight (Table 1). PI-IUGR fetuses weighed 40% less than CON fetuses (Fig. 1A). Absolute rates of uterine and umbilical blood flow were 30% and 50% lower, respectively, in PI-IUGR compared with CON animals (Table 1). When normalized to fetal weight, umbilical blood flow was 22% lower in PI-IUGR fetuses.

Table 1.

Uteroplacental and fetal growth and blood flow

	CON	PI-IUGR
Male-to-female ratio†	1:4:1	7:3
Gestational age, days	133.5 ± 1.2	133.6 ± 1.1
Placentome number	82.2 ± 4.7	64.9 ± 4.3*
Placentome total weight, g	340.2 ± 39.1	192.0 ± 29.1*
Fetal weight, kg	3.2 ± 0.1	2.1 ± 0.2*
Uterine blood flow, mL/min	1,667 ± 183.2	1,101 ± 143*
Umbilical blood flow, mL/min	565.3 ± 69.4	292.5 ± 36.8*
Umbilical blood flow, mL/min/kg	172.8 ± 13.8	135.2 ± 9.7*

Values are mean ± SE; N = 6 control (CON), N = 10 placental insufficiency-induced intrauterine growth restriction (PI-IUGR). †Sex was not recorded for 1 fetus. *P < 0.05 in PI-IUGR vs. CON by t test.

Figure 1.

Effect of placental insufficiency-induced intrauterine growth restriction (PI-IUGR) on fetal weight and oxygen uptake. A: fetal weight in control (CON) and PI-IUGR groups. CON fetuses are shown with blue circles. PI-IUGR fetuses weighing >2 kg are represented with pink circles, and fetuses weighing <2 kg are represented with red circles. B: conceptual model of uterine, uteroplacental (UtP), and fetal uptake rate calculations. Created with BioRender.com. C: UtP net oxygen uptake rates in CON and PI-IUGR groups. D: fetal net oxygen uptake rates normalized to fetal weight in CON and PI-IUGR groups. Means ± SE are shown. Data were analyzed by t test. **P < 0.01 in PI-IUGR vs. CON.

Maternal and Fetal Nutrient Concentrations

Maternal uterine arterial and uterine venous blood oxygenation and CO₂ values did not differ between PI-IUGR and CON animals (Table 2). In the PI-IUGR fetus, umbilical venous blood oxygen saturation was lower and blood CO₂ content was higher (Table 2). In the fetal artery, blood oxygen saturation and content were lower in PI-IUGR compared with CON fetuses (Table 2). There were no differences between groups in plasma concentrations of glucose or lactate in the maternal artery and uterine vein, yet blood pyruvate concentrations tended to be lower in both vessels in the PI-IUGR group (Table 2). Plasma glucose concentrations in the umbilical vein and umbilical artery were lower in PI-IUGR than CON fetuses, whereas lactate concentrations were higher in PI-IUGR (Table 2). Pyruvate concentrations were higher in umbilical artery, but not umbilical vein, in PI-IUGR fetuses. Several amino acid concentrations were lower in PI-IUGR compared with CON, including isoleucine, valine, glutamine, and serine, whereas concentrations of taurine and asparagine were higher (Table 2). There were no differences in plasma nonesterified fatty acid concentrations in any vessel between CON and PI-IUGR groups.

Table 2.

Uterine and umbilical blood gas and nutrient concentrations in CON and PI-IUGR groups

	Uterine Artery		Uterine Vein		Umbilical Vein		Umbilical Artery
	CON	PI-IUGR	CON	PI-IUGR	CON	PI-IUGR	CON	PI-IUGR
Blood gas
Po2, mmHg	84.6 ± 5.5	86.5 ± 3.1	53.4 ± 5.5	58.2 ± 4.8#	29.1 ± 5.7	24.7 ± 6.5	18.9 ± 2.9	15.2 ± 3.6#
So2, %	97.3 ± 2.7	96.4 ± 2.1	78.4 ± 5.1	80.1 ± 5.5	74.2 ± 8.9	60.1 ± 14.3*	43.6 ± 8.3	27.9 ± 11.4*
Oxygen, mM	6.10 ± 0.72	6.18 ± 0.40	5.00 ± 0.72	5.14 ± 0.32	4.82 ± 0.20	4.15 ± 0.91#	2.90 ± 0.34	2.01 ± 0.81*
Pco2, mmHg	21.1 ± 0.5	21.9 ± 0.3	22.4 ± 0.4	22.9 ± 0.3	23.2 ± 1.0	24.4 ± 0.7	24.9 ± 1.1	25.8 ± 0.4
CO2, mM	35.1 ± 0.6	35.3 ± 0.5	39.2 ± 0.4	39.1 ± 0.4	43.8 ± 0.6	45.2 ± 0.4*	50.9 ± 1.0	54.4 ± 0.8#
Carbohydrates and lipids
Glucose, mM	3.85 ± 0.31	3.91 ± 0.23	3.61 ± 0.29	3.73 ± 0.21	1.31 ± 0.16	1.08 ± 0.13*	1.08 ± 0.16	0.84 ± 0.11*
Lactate, mM	0.63 ± 0.21	0.57 ± 0.14	0.69 ± 0.22	0.65 ± 0.14	2.51 ± 0.96	4.26 ± 2.30†	2.36 ± 0.93	4.12 ± 2.30†
Pyruvate, mM	0.14 ± 0.05	0.09 ± 0.05#	0.13 ± 0.04	0.09 ± 0.04#	0.12 ± 0.03	0.12 ± 0.05	0.12 ± 0.04	0.17 ± 0.04*
NEFA, µM	261.8 ± 245.7	224.6 ± 89.3	265.0 ± 246.3	220.8 ± 93.5	30.6 ± 11.5	37.3 ± 11.0	35.8 ± 10.0	40.7 ± 11.0
Essential amino acids, µM
Histidine	57.6 ± 5.1	47.4 ± 2.6#	54.6 ± 5.5	45.6 ± 2.2#	49.2 ± 2.5	53.0 ± 3.3	42.9 ± 2.5	45.7 ± 2.5
Isoleucine	120.6 ± 12.3	116.2 ± 7.8	110.2 ± 11.3	106.5 ± 6.8	107.5 ± 7.2	87.5 ± 7.0#	88.5 ± 6.3	70.7 ± 5.5#
Leucine	150.2 ± 14.7	136.5 ± 6.5	138.3 ± 13.5	124.8 ± 5.6	155.5 ± 12.1	133.2 ± 13.3	124.2 ± 10.4	105.2 ± 11.1
Lysine	139.5 ± 6.9	131.7 ± 7.1	133.1 ± 6.9	124.8 ± 7.0	77.9 ± 6.7	93.8 ± 7.3	59.2 ± 5.2	71.6 ± 6.2
Methionine	34.2 ± 3.0	31.3 ± 1.4	31.3 ± 3.1	29.4 ± 1.2	89.0 ± 9.6	81.1 ± 8.0	80.5 ± 8.8	72.7 ± 7.2
Phenylalanine	54.9 ± 5.6	52.5 ± 2.3	52.0 ± 5.6	48.6 ± 1.9	112.3 ± 12.0	112.1 ± 7.0	100.4 ± 12.0	100.2 ± 6.6
Threonine	137.9 ± 24.0	128.5 ± 11.2	132.8 ± 23.8	123.3 ± 10.8	247.7 ± 32.7	236.3 ± 23.0	227.8 ± 32.7	213.1 ± 22.5
Tryptophan	33.6 ± 4.6	36.3 ± 1.9	33.2 ± 4.3	33.0 ± 1.9	37.2 ± 2.8	37.5 ± 1.7	33.9 ± 2.7	33.9 ± 1.7
Valine	217.8 ± 25.0	217.4 ± 10.6	205.9 ± 24.0	204.7 ± 10.6	365.7 ± 22.8	304.1 ± 26.0#	335.5 ± 23.1	275.4 ± 23.7#
Nonessential amino acids, µM
Alanine	135.9 ± 17.9	132.6 ± 7.1	130.7 ± 18.6	127.8 ± 7.1	288.7 ± 33.5	313.4 ± 24.2	262.0 ± 31.7	285.2 ± 22.6
Arginine	174.8 ± 5.1	160.4 ± 11.3	164.5 ± 4.2	150.8 ± 10.5	90.3 ± 3.8	87.2 ± 9.3	68.4 ± 3.6	65.9 ± 8.3
Asparagine	43.7 ± 7.7	39.4 ± 3.1	41.8 ± 7.6	36.8 ± 2.9	45.9 ± 2.3	56.3 ± 3.6#	34.3 ± 2.0	43.3 ± 2.9*
Aspartate	6.9 ± 0.5	7.4 ± 0.6	7.2 ± 0.6	7.7 ± 0.6	23.1 ± 1.9	20.5 ± 2.5	22.2 ± 2.5	20.5 ± 2.4
Cysteine	32.4 ± 2.7	28.8 ± 1.9	32.1 ± 2.9	27.7 ± 1.9	18.1 ± 1.0	14.9 ± 2.2	15.9 ± 1.4	13.1 ± 2.0
Glutamate	56.5 ± 3.6	63.7 ± 4.0	59.0 ± 3.2	66.4 ± 3.5	15.8 ± 1.8	15.7 ± 1.8	48.1 ± 5.4	45.1 ± 5.1
Glutamine	290.2 ± 35.8	274.1 ± 14.4	278.3 ± 35.7	258.9 ± 13.0	480.4 ± 26.8	419.1 ± 25.5#	417.0 ± 27.2	351.0 ± 23.1#
Glycine	383.3 ± 36.2	379.3 ± 19.6	388.5 ± 42.9	381.1 ± 19.8	375.9 ± 43.3	374.7 ± 21.5	345.1 ± 41.6	346.7 ± 20.6
Ornithine	90.7 ± 13.6	76.4 ± 6.8	85.3 ± 13.5	70.6 ± 6.1	63.3 ± 5.5	59.4 ± 9.5	59.1 ± 5.3	55.0 ± 8.6
Proline	99.6 ± 17.0	76.0 ± 4.4#	91.5 ± 16.8	69.2 ± 4.1#	157.4 ± 10.4	186.8 ± 17.5	135.9 ± 9.9	165.2 ± 16.1
Serine	79.2 ± 10.6	72.4 ± 4.3	71.6 ± 9.2	66.1 ± 4.0	677.3 ± 79.8	426.3 ± 68.5*	676.3 ± 80.0	426.7 ± 70.4*
Taurine	74.6 ± 8.0	58.3 ± 6.7#	75.7 ± 7.8	59.9 ± 6.5#	107.9 ± 24.9	162.0 ± 17.9#	101.9 ± 26.7	160.1 ± 17.9#
Tyrosine	64.1 ± 8.5	63.6 ± 5.3	60.5 ± 8.0	57.7 ± 4.1	120.0 ± 11.7	121.4 ± 10.0	107.9 ± 12.4	109.6 ± 9.9

Values are means ± SE; N = 6 control (CON), N = 10 placental insufficiency-induced intrauterine growth restriction (PI-IUGR). NEFA, nonesterified fatty acids; So₂, oxygen saturation. *P < 0.05 by t test, #P < 0.15, by t test, and †P < 0.05 with Mann–Whitney test for nonparametric data in PI-IUGR vs. CON within a vessel as indicated.

Uteroplacental and Fetal Oxygen and Nutrient Flux

Blood gas levels and nutrient concentrations were coupled with blood flow measurements across the uterine circulation, represented by maternal artery (input) and uterine vein (output), and the umbilical circulation, represented by the umbilical vein (input) and umbilical artery (output), to calculate net nutrient uterine and umbilical uptake rates (Fig. 1B). The difference between uterine and umbilical uptake rates equals net uteroplacental uptake. To account for differences in placental and fetal mass, all uteroplacental rates are normalized to placental weight and all fetal (net umbilical) rates are normalized to fetal weight. Absolute uterine and fetal uptake rates are provided in Supplemental Table S1 (available at https://doi.org/10.6084/m9.figshare.22056947.v1). Uteroplacental oxygen uptake was maintained in PI-IUGR and similar to CON (Fig. 1C). Fetal oxygen uptake trended lower in PI-IUGR compared with CON (Fig. 1D).

Uteroplacental and fetal net nutrient uptake rates (weight specific) are shown in Fig. 2, A and B, and presented below.

Figure 2.

Effect of placental insufficiency-induced intrauterine growth restriction (PI-IUGR) on fetal and uteroplacental nutrient flux and glucose metabolism rates. A: uteroplacental net uptake rates of major nutrients and selected amino acids normalized to placental weight in control (CON) and PI-IUGR groups. Negative uptake rates indicate net output. NEFA, nonesterified fatty acids. CON fetuses are shown with blue circles. PI-IUGR fetuses weighing >2 kg are represented with pink circles, and fetuses weighing <2 kg are represented with red circles. B: net fetal uptake rates of major nutrients and selected amino acids normalized to fetal weight in CON and PI-IUGR groups. C: uteroplacental glucose utilization rates normalized to placental weight in CON and PI-IUGR groups. D: fetal glucose utilization rates normalized to fetal weight in CON and PI-IUGR groups. E: fetal glucose oxidation rates normalized to fetal weight in CON and PI-IUGR groups. Means ± SE are shown. Data were analyzed by t test. *P < 0.05 and **P < 0.01 in PI-IUGR vs. CON; †P < 0.05 for 1-sided t test vs. 0.

Glucose.

There was no difference in uteroplacental glucose uptake between groups (Fig. 2A). Fetal glucose uptake rates were 20% lower in PI-IUGR compared with CON fetuses (Fig. 2B). In PI-IUGR, uteroplacental glucose utilization, measured with maternal infusion of [6,6-²H]glucose, was maintained (Fig. 2C) and fetal glucose utilization, average of rates measured with [6,6-²H]glucose and fetal infusion of [U-¹³C]glucose, was reduced (Fig. 2D), both utilization rates mirroring net glucose uptake rates. Glucose oxidation tended to be lower in PI-IUGR compared with CON fetuses (Fig. 2E).

Lactate.

There was no difference in uteroplacental lactate output between groups (Fig. 2A). Fetal lactate uptake tended to be lower in PI-IUGR fetuses compared with CON (Fig. 2B), likely reflecting increased endogenous fetal lactate production (48).

Pyruvate.

Uteroplacental pyruvate uptake tended to be higher and was >0 in PI-IUGR compared with CON groups (Fig. 2A). PI-IUGR fetuses had nearly eightfold higher pyruvate output rate compared with CON fetuses (Fig. 2B).

Nonesterified fatty acids.

No detectable uteroplacental nonesterified fatty acid (NEFA) uptake was observed in CON and PI-IUGR groups (P > 0.50 in each group, 1-sided t test) (Fig. 2A). However, fetal uptake of NEFA was lower in PI-IUGR compared with CON fetuses (P = 0.06) and >0 in both groups (Fig. 2B), albeit at relatively low rates compared with the other major nutrients (nanomolar vs. micromolar).

Amino acids.

Total amino acid uptake by the uteroplacenta (Fig. 2A) was not >0 in the CON group (P = 0.90, 1-sided t test) but was >0 in the PI-IUGR group (P < 0.01, 1-sided t test). Total amino acid net uptake was lower in PI-IUGR compared with CON fetuses (Fig. 2B). We next looked at the uptake rates of glutamine, glutamate, glycine, and serine, which are known to be actively shuttled between the fetus and placenta (13–16, 49). There were no differences in net uteroplacental uptake of glutamate and serine or release of glutamine and glycine (Fig. 2A). The PI-IUGR compared with CON fetus had decreased net uptake of glycine, with no differences in glutamine, glutamate, or serine uptake (Fig. 2B).

Nutrient-oxygen metabolic quotients.

We next looked at the nutrient:oxygen quotients, which are a flow-independent means to assess net nutrient uptake per carbon atom relative to oxygen uptake. There were no differences in individual or cumulative sum of uteroplacental metabolic quotients, although the glucose quotient tended to be lower and amino acid quotient tended to be higher in PI-IUGR (Fig. 3A). The cumulative sum of quotients was not >1 in either group. This suggests that the carbon atoms from these substrates are used to fuel oxidative metabolism, leaving little additional carbon substrate available for biosynthesis or tissue growth. On the fetal side, the individual glucose, amino acid, and pyruvate quotients were lower in PI-IUGR (Fig. 3B). The cumulative sum of umbilical metabolic quotients was >1 in the CON group and higher compared with PI-IUGR.

Figure 3.

Uteroplacental and fetal nutrient metabolic quotients. A: individual nutrient-oxygen metabolic quotients for uteroplacental (UtP) glucose, pyruvate, amino acids, and lactate and the cumulative (total) sum in control (CON) and placental insufficiency-induced intrauterine growth restriction (PI-IUGR) groups. CON fetuses are shown with blue circles. PI-IUGR fetuses weighing >2 kg are represented with pink circles, and fetuses weighing <2 kg are represented with red circles. B: individual umbilical nutrient-oxygen metabolic quotients for glucose, lactate, amino acids, and pyruvate and the cumulative (total) sum in CON and PI-IUGR groups. Means ± SE are shown. Data were analyzed by t test. *P < 0.05 and **P < 0.01 in PI-IUGR vs. CON; ‡P < 0.05 for 1-sided t test of the cumulative sum vs. 1.0.

Uteroplacental Tissue Responses

To understand the molecular signaling pathways that may underlie the in vivo nutrient flux data, we measured the following pathway-specific targets in uteroplacental tissue (cotyledonary) in CON and PI-IUGR groups.

Nutrient utilization.

There was no change in expression of the genes for the glucose transporters GLUT1 and GLUT4 or glycolytic enzymes, including PFK1, PKLR, PKM1, and PKM2, between CON and PI-IUGR groups (Fig. 4A). Gene expression of the mitochondrial pyruvate transporters (MPC1, MPC2), pyruvate dehydrogenase (PDH), and PDH kinases (PDK1, PDK2, and PDK4) were not different between groups. Gene expression of LDHA tended to be higher, but there were no differences in gene expression of LDHA, LDHB, or lactate transporters MCT1 and MCT4 (Fig. 4A). There was no difference in abundance of phosphorylated PDH or total PDH protein, yet the ratio of phosphorylated inactive PDH to total PDH protein was lower in PI-IUGR compared with CON placentas, favoring increased activity (Fig. 4, B and C). Indeed, PDH enzymatic activity was higher (Fig. 4D). There were no differences in LDH protein expression or LDH enzymatic activity (Fig. 4, C and D). PI-IUGR uteroplacental tissues had decreased expression of glutaminase (GLS), glutamate dehydrogenase (GLUD1), and glutamine ligase (GLUL, also known as glutamine synthetase) (Fig. 4E).

Figure 4.

Metabolic adaptations in placental tissue. A: expression of genes involved in glucose uptake and utilization, pyruvate oxidation, and lactate metabolism was measured in control (CON) and placental insufficiency-induced intrauterine growth restriction (PI-IUGR) placental tissue. CON fetuses are shown with blue circles. PI-IUGR fetuses weighing >2 kg are represented with pink circles, and fetuses weighing <2 kg are represented with red circles. B: protein expression was measured by Western blotting in CON and PI-IUGR placental tissue lysates and quantified. A representative blot of each protein is shown, and protein molecular masses (kilodaltons) are indicated on right. C: protein abundance of phosphorylated (ph, S293) and total pyruvate dehydrogenase (PDH) and the ratio of phosphorylated to total PDH (left) and protein abundance of LDH (right) in placental tissue in CON and PI-IUGR groups. Actin abundance is shown, demonstrating equality of loading. D, left: PDH activity measured in placental tissue. Right: LDH activity measured in placental tissue. E: expression of genes for glutamine-glutamate metabolism measured in placental tissue of CON and PI-IUGR groups. Means ± SE are shown. Data were analyzed by t test. *P < 0.05 in PI-IUGR vs. CON.

Nutrient sensing and signaling.

We next looked at AMPK, AKT, and mTOR signaling to determine the effect of PI-IUGR on these nutrient sensors. The phosphorylation of AMPK protein (T172) was increased in PI-IUGR compared with CON placental tissue, with no change in total abundance of AMPK, resulting in an increased phosphorylated-to-total AMPK ratio, representing activation (Fig. 5, A and B). Phosphorylation of AKT (S473) was also higher in PI-IUGR placenta. Whereas there was no difference in the ratio of phosphorylated to total AKT protein, the phosphorylation of GSK3β (S9), a target of AKT signaling, and ratio of phosphorylated to total GSK3β was increased (Fig. 5, A and C). The abundance of phosphorylated mTOR (S2448) protein was increased. There was no difference in total mTOR protein abundance or the ratio of phosphorylated to total protein (Fig. 5, A and D). The phosphorylation of S6 kinase (S6K, T421, S424) and ribosomal S6 (S6, S235, S236) protein, downstream targets of mTOR activation, were increased in PI-IUGR compared with CON placentas. There was no change in total abundance of S6K or S6, yet the ratio of phosphorylated to total S6 protein was higher in PI-IUGR. The phosphorylation of 4EBP1 (T37, T46), another target of mTOR signaling, tended to be higher in PI-IUGR compared with CON placentas, and there was no change in total abundance of 4EBP1 or the ratio (Fig. 5D).

Figure 5.

Effect of placental insufficiency-induced intrauterine growth restriction (PI-IUGR) on nutrient sensing and signaling pathways in the placenta. Protein abundance was measured by Western blotting on control (CON) and PI-IUGR placental tissues. A: representative Western blot images are shown, and protein molecular masses (kilodaltons) are indicated on right for phosphorylated (ph) and total protein abundance of AMP-activated protein kinase (AMPK) (T172), AKT (S473), GSK3β (S9), mechanistic target of rapamycin (mTOR) (S2448), p70S6K (T421, S424), S6 (S235,6), and 4EBP1 (T37, T46). B: protein abundance of phosphorylated, total, and phosphorylated:total AMPK. CON fetuses are shown with blue circles. PI-IUGR fetuses weighing >2 kg are represented with pink circles, and fetuses weighing <2 kg are represented with red circles. C: protein abundance of phosphorylated, total, and phosphorylated:total AKT and GSK3β. Band shown for GSK3α was not included. D: protein abundance of phosphorylated, total, and phosphorylated:total mTOR, S6K, S6, and 4EBP1. Means ± SE are shown. Data were analyzed by t test. *P < 0.05, **P < 0.01, and ***P < 0.001 in PI-IUGR vs. CON.

Mitochondrial oxidative capacity and oxidative stress.

The abundance of proteins representing the mitochondrial oxidative phosphorylation complexes was measured (Fig. 6, A and B). The relative protein abundance of NDUFB8, representing oxidative phosphorylation complex I, was lower in PI-IUGR compared with CON uteroplacental tissue. There were no differences in the abundance of proteins representing complexes II (SDHB), III (UQCRC2), IV (COXII), or V (ATP5A). Expression of SDHB was increased in PI-IUGR compared with CON, and expressions of COX41 and COX42 genes were similar between groups (Fig. 6C). Uteroplacental tissue content of TBARS, a marker of oxidative stress, tended to be higher (Fig. 6D) and was correlated with phosphorylated AMPK protein abundance, supporting a relationship between oxidative stress and AMPK activation in the PI-IUGR placenta (Fig. 6E). Furthermore, PDH activity was associated with phosphorylated AMPK protein abundance (Fig. 6E), supporting a link between oxidative stress, AMPK activation, and pyruvate utilization via PDH (Fig. 6E).

Figure 6.

Effects on mitochondrial capacity and oxidative stress. Protein abundance was measured by Western blotting in control (CON) and placental insufficiency-induced intrauterine growth restriction (PI-IUGR) placental tissues. A: representative Western blot image of oxidative phosphorylation complexes I–V; protein molecular masses (kilodaltons) are indicated on right. B: protein abundance of oxidative phosphorylation complexes I–V in CON and PI-IUGR groups. CON fetuses are shown with blue circles. PI-IUGR fetuses weighing >2 kg are represented with pink circles, and fetuses weighing <2 kg are represented with red circles. C: expression of genes for mitochondrial function, SDHB, COX41, and COX42, in CON and PI-IUGR groups. D: thiobarbituric acid-reactive substances (TBARS) content measured in placental tissue. Means ± SE are shown. Data were analyzed by t test. *P < 0.05 in PI-IUGR vs. CON. E: associations between placental AMP-activated protein kinase (AMPK) activation (ph-AMPK protein abundance), TBARS content, and pyruvate dehydrogenase (PDH) activity.

DISCUSSION

Our results demonstrate shifts in placental and fetal nutrient shuttling in association with metabolic responses in the placenta during placental insufficiency. First, PI-IUGR pregnancies have increased fetal to placental pyruvate flux and placental PDH activation, supporting increased placental capacity to oxidize pyruvate. Second, uteroplacental rates of oxygen utilization and glucose uptake are maintained in PI-IUGR and similar to CON when normalized to placental weight, consistent with prior studies (17, 50, 51). Third, placental expression of genes regulating glutamine-glutamate metabolism is lower in PI-IUGR, supporting downregulation of amino acid shuttling between the placenta and fetus. Fourth, we find an unexpected activation of AMPK with higher AKT and mTOR signaling in the PI-IUGR placenta, suggesting a dissociation between the regulation of these nutrient sensors. Taken together, we speculate that PI-IUGR activates AMPK and PDH in the placenta, which, in part, mediates the shifts in uteroplacental substrate metabolism and nutrient shuttling with the fetus that enable the placenta to maintain weight-specific oxygen consumption (see summary in Fig. 7).

Figure 7.

Summary of proposed metabolic adaptations that may enable the placental insufficiency-induced intrauterine growth restriction (PI-IUGR) placenta to maintain weight-specific oxygen consumption (Vo₂). 1) During placental insufficiency, AMP-activated protein kinase (AMPK) is activated, which in turn activates pyruvate dehydrogenase (PDH) and facilitates increased pyruvate utilization (red). 2) The placenta utilizes pyruvate from the increased fetal flux (green). Pyruvate that is not oxidized via PDH is converted to lactate and delivered to the fetus, creating an accelerated lactate-pyruvate shuttle. 3) This increased preference for pyruvate from the fetus may spare glucose for delivery to the fetus (blue). 4) Glutamine (GLN)-glutamate (GLU) metabolism is lower in the placenta (orange), potentially sparing glutamine for use by the fetus. Thick colored arrows indicate active or increased flux in pathway; thin gray arrows indicate decreased flux. Created with BioRender.com.

PI-IUGR uteroplacental tissues have increased dephosphorylated (active) PDH protein and increased PDH activity, demonstrating increased capacity for pyruvate oxidation to acetyl CoA. We also find increased fetal pyruvate concentrations and net pyruvate output from the PI-IUGR fetal to uteroplacental compartment that is matched by higher uteroplacental net pyruvate uptake. This occurs without an increase in uteroplacental glucose uptake (measured with Fick principle), glucose utilization (measured with isotope tracing), or expression of genes regulating glucose transporters and glycolysis. Thus, during PI-IUGR, increased pyruvate flux from the fetus may fuel oxidation in the placenta, in addition to pyruvate normally produced from glycolysis. Lactate can also be made from pyruvate. However, we did not detect increased placental to fetal lactate flux during PI-IUGR, and placental expression of LDHA, the gene responsible to converting pyruvate to lactate, only tended to be higher, whereas protein abundance and activity of LDH were not different. Collectively, this suggests 1) accelerated fetal to placental pyruvate flux and 2) that within the placenta pyruvate may be preferentially oxidized, with the remainder reduced to lactate and delivered as a fuel to the fetus (step 2, Fig. 7). Future studies using metabolic tracers are needed to establish lactate-pyruvate shuttling given that both the fetus and placenta simultaneously produce and utilize lactate (52). Additional studies measuring placental mitochondrial respiration are also needed to determine the relative preference for pyruvate as an oxidative substrate and to understand the effects of lower oxidative phosphorylation complex I abundance and higher expression of SDHB gene, which encodes for SDHB protein, a component of oxidative phosphorylation complex II, whose protein expression was maintained.

Fetal glucose utilization and oxidation rates are lower in PI-IUGR, as shown here and previously (48, 53). Lower glucose oxidation likely contributes to increased lactate and pyruvate concentrations and increased pyruvate output in the PI-IUGR fetus. More specifically, fetal pyruvate output during PI-IUGR likely occurs in the fetal liver, as prior studies show that the liver releases pyruvate (30). We also find increased fetal pyruvate output and accelerated lactate-pyruvate shuttling between the placenta and fetus during sustained late-gestation hypoxemia in sheep (6). Thus, conditions associated with hypoxemia, including placental insufficiency and IUGR, increase fetal pyruvate output (6, 11, 12). Ongoing studies in models of hypoxemia and PI-IUGR are underway to determine whether the fetal liver is responsible for increased pyruvate output (33).

Metabolic reprogramming describes how cells adapt their substrate metabolism when oxygen availability is low (54, 55). In human studies, it has been suggested that hypoxemia increases glucose consumption (glycolysis) and decreases mitochondrial oxygen consumption via mechanisms resembling metabolic reprogramming in the placenta to explain the reduction in glucose supply yet preservation of oxygen supply to the fetus (2, 55, 56). If there is metabolic reprogramming in the placenta, we would expect decreased uteroplacental oxygen uptake, increased uteroplacental glucose uptake, increased lactate release to the fetus and mother, and decreased glucose supply to the fetus. However, placental glucose utilization and oxygen consumption are maintained in sheep models of PI-IUGR and late-gestation hypoxemia (6, 17, 40, 51). Thus, the present data only support decreased fetal glucose supply and potentially higher placental lactate production during PI-IUGR (steps 2 and 3, Fig. 7) or hypoxemia in humans and sheep models (4, 6, 17, 40, 51). Since pyruvate is central to glucose and lactate metabolism and our data do not support increased placental glucose utilization, we propose that increased placental pyruvate utilization during PI-IUGR and hypoxemia resembles and explains features of metabolic reprogramming that have been previously attributed to increased placental glucose utilization.

Prior studies report lower transplacental amino acid concentration ratios, lower fetal amino acid uptake, including glutamate and glycine, and less fetal glutamate output during PI-IUGR (3, 40, 57). In the PI-IUGR group, our data show lower uteroplacental glutamine output to the fetus, and both fetal glutamine uptake and glutamate output tended to be lower, consistent with results in a larger cohort showing significant reductions in glutamine uptake and glutamate output (3). We also find decreased uteroplacental expression of genes regulating glutamine and glutamate metabolism (GLS, GLUD1, and GLUL). Placental GDH (encoded by GLUD1 gene) activity is also decreased in human IUGR (58). These results suggest lower placental metabolism of glutamine and glutamate, perhaps as a mechanism to allow the fetus to utilize these nutrients (step 4, Fig. 7). Serine-to-glycine shuttling is important for the metabolism and transfer of intermediates in one-carbon metabolism between the fetal liver and placenta (59, 60). Fetal glycine uptake from the placenta is lower and uteroplacental glycine output tended lower in the PI-IUGR, with no differences in serine uptake rates. Together, these results support downregulation of amino acid shuttling between the placenta and fetal liver, the fetal organ responsible for these amino acid exchanges (13–16).

We find increased AMPK activation and AKT and mTOR signaling in the PI-IUGR placenta. The fetal-facing side of the PI-IUGR placenta is perfused with umbilical blood containing low nutrients and an endocrine milieu with low insulin and IGF-1 that would be expected to activate AMPK (21, 22, 28). We recognize that active AKT and mTOR signaling is unexpected given the low nutrient milieu and as AMPK activation normally suppresses mTOR (28). However, some studies in placental samples from IUGR pregnancies in animal models and human subjects support AKT and mTOR activation during growth restriction and dissociation from AMPK-mediated suppression (19, 61). Placentas from human high-altitude pregnancies have greater phosphorylation of AMPK and 4EBP1 proteins, supporting AMPK activation and mTOR signaling (62). The abundance of mTOR protein is higher in human IUGR placentas (63). These data suggest that AMPK-mediated suppression of mTOR may be dissociated and that AKT-mediated activation of mTOR signaling is intact in the placenta of pregnancies complicated with nutrient and oxygen stress (61). Prior studies in this sheep model of PI-IUGR report higher mTOR phosphorylation and no change in phosphorylation of the downstream targets S6K, S6, or 4EBP1 in placental tissue; however, AMPK activation was not measured and other in vivo systemic factors were not reported, which could explain the differences compared with our results (64). In addition, during sustained hypoxemia in sheep, neither activation of AMPK activation nor suppression of mTOR signaling is observed (6). Thus, activation of AMPK during PI-IUGR may be regulated by nutrient and oxidative stress rather than hypoxemia. Indeed, phosphorylation of AMPK is associated with increased oxidative stress (i.e., TBARS). In addition, human high-altitude placentas demonstrate oxidative stress-induced metabolites (65). We speculate that AMPK may promote PDH activation as shown in other studies (25, 26) and as supported by the positive association between PDH activity and AMPK phosphorylation, providing a mechanism for increased pyruvate utilization (step 1, Fig. 7).

In both CON and PI-IUGR groups, the sum of uteroplacental metabolic quotients is close to 1.0, demonstrating a match between carbon substrates and oxidative metabolism. This is also consistent with the function of the placenta in late gestation, after placental mass is established, whereby substrates are used for oxidative metabolism and nutrient exchange rather than growth. It should be noted we find no net uptake of nonesterified fatty acids across the uterine or uteroplacental circulation, consistent with prior work (66). Further studies are needed to understand the impact of putatively lower nonesterified fatty acid uptake in the PI-IUGR fetus. We recognize that other substrates, including glycerol, volatile fatty acids, phospholipids, ketones, and acylcarnitines, may be important; however, additional studies are needed to quantitatively measure these substrates.

In summary, our results highlight the importance of bidirectional nutrient shuttles between the placenta and fetus. We speculate that these fetal-placental nutrient shuttles help sustain placental oxidative metabolism and maintain nutrient supply to the fetus during PI-IUGR (summarized in Fig. 7). Since fetal growth restriction develops despite activation of these shuttles and relatively maintained weight-specific uteroplacental and fetal oxygen consumption rates, additional studies will be needed to test whether these shuttles and metabolic effects are indeed beneficial in enabling the fetus and placenta to defend their rates of oxidative metabolism or, alternatively, if these effects are contributing to the development of fetal growth restriction. Further studies are also needed to address sex differences and the severity of growth restriction, both putative sources of variation that this study was not powered to detect.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL MATERIAL

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.22056947.v1.

GRANTS

This work was supported by National Institutes of Health Grant R01-DK108910 (to S.R.W.). Additional support was provided by the Ludeman Center for Women’s Health Research and by R01-HD093701 (to P.J.R.), R01-HD079404 (to L.D.B.), R01 HL138181 (to L.G.M. and C.G.J.), and R21-HD102628 (to R.A.L.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

H.M.K. and S.R.W. conceived and designed research; D.W., R.B.W., and S.R.W. performed experiments; H.M.K., D.W., and S.R.W. analyzed data; H.M.K., R.A.L., C.G.J., L.G.M., R.B.W., P.J.R., L.D.B., and S.R.W. interpreted results of experiments; H.M.K. and S.R.W. prepared figures; H.M.K. and S.R.W. drafted manuscript; H.M.K., R.A.L., C.G.J., L.G.M., P.J.R., L.D.B., and S.R.W. edited and revised manuscript; H.M.K., D.W., R.A.L., C.G.J., L.G.M., R.B.W., P.J.R., L.D.B., and S.R.W. approved final version of manuscript.