Skeletal muscle amino acid uptake is lower and alanine production is greater in late gestation intrauterine growth-restricted fetal sheep hindlimb

Abstract

In a sheep model of intrauterine growth restriction (IUGR) produced from placental insufficiency, late gestation fetuses had smaller skeletal muscle mass, myofiber area, and slower muscle protein accretion rates compared with normally growing fetuses. We hypothesized that IUGR fetal muscle develops adaptations that divert amino acids (AAs) from protein accretion and activate pathways that conserve substrates for other organs. We placed hindlimb arterial and venous catheters into late gestation IUGR (n = 10) and control (CON, n = 8) fetal sheep and included an external iliac artery flow probe to measure hindlimb AA uptake rates. Arterial and venous plasma samples and biceps femoris muscle were analyzed by mass spectrometry-based metabolomics. IUGR fetuses had greater abundance of metabolites enriched within the alanine, aspartate, and glutamate metabolism pathway compared with CON. Net uptake rates of branched-chain AA (BCAA) were lower by 42%–73%, and muscle ammoniagenic AAs (alanine, glycine, and glutamine) were lower by 107%–158% in IUGR hindlimbs versus CON. AA uptake rates correlated with hindlimb weight; the smallest hindlimbs showed net release of ammoniagenic AAs. Gene expression levels indicated a decrease in BCAA catabolism in IUGR muscle. Plasma purines were lower and plasma uric acid was higher in IUGR versus CON, possibly a reflection of ATP conservation. We conclude that IUGR skeletal muscle has lower BCAA uptake and develops adaptations that divert AAs away from protein accretion into alternative pathways that sustain global energy production and nitrogen disposal in the form of ammoniagenic AAs for metabolism in other organs.

INTRODUCTION

Intrauterine growth restriction (IUGR) fetuses and small-for-gestational-age neonates have less lean muscle mass than their appropriately grown counterparts (41, 60). This smaller muscle mass persists throughout childhood and into adulthood (26, 33). Epidemiological studies link low birth weight and decreased muscle mass to insulin resistance (38, 59, 72), development of the metabolic syndrome and type 2 diabetes (1, 2, 78), and increased risk for adverse cardiovascular events (e.g., stroke and myocardial infarction) later in life (4). Thus, reduced skeletal muscle growth in utero may lead to lasting consequences that adversely affect lifelong metabolic health (3, 12, 17, 47). The mechanisms that lead to lower rates of skeletal muscle growth in IUGR fetuses and increase the risk of later-life metabolic disease remain incompletely understood.

Previously, we have shown in a sheep model of placental insufficiency-induced IUGR that reduced skeletal muscle protein accretion rates were associated with smaller muscle mass in IUGR fetuses when compared with normally growing controls (67). Net total amino acid (AA) uptake rates by the hindlimb were lower in IUGR compared with control fetuses, despite normal circulating plasma concentrations of nearly all AAs (67). Previous studies also have indicated that hindlimb metabolism largely consists of skeletal muscle-specific metabolism rather than bone or other tissues (6, 28, 30, 32). However, the mechanisms responsible for lower hindlimb net (total) AA uptake rates and lower skeletal muscle protein synthesis and accretion rates in the IUGR fetus are unknown.

The overall goal of this study was to identify metabolite differences using a metabolomics approach in fetal hindlimb skeletal muscle from IUGR and control late gestation fetal sheep. Using chronic surgical catheterization of the fetal abdominal aorta and femoral vein, AAs and metabolites were measured in the arterial influx and venous efflux from the hindlimb to determine differences in AA uptake and metabolite profiles that were associated with IUGR. Biceps femoris biopsies were concurrently obtained to measure relative concentrations of skeletal muscle-specific hydrophilic metabolites. We hypothesized that IUGR fetal muscle develops adaptations that divert AAs away from protein accretion into alternative metabolic pathways that conserve energy for the fetus and maintain cellular metabolism for other organs.

MATERIALS AND METHODS

Animal care and IUGR model.

Study protocols were approved by the University of Colorado Anschutz Medical Campus Institutional Animal Care and Use Committee [no. 77617(10)E]. All experiments were performed in accordance with relevant guidelines and regulations from the Guide for the Care and Use of Laboratory Animals. Experimental details are reported according to the Animal in Research: Reporting In Vivo Experiments guidelines (40). The current study used animals from a cohort that was previously published in Rozance et al. (67).

Pregnant Columbia-Rambouillet mixed-breed sheep were randomized and housed in two different environmental chambers. Ewes were either housed in an environmental chamber with elevated ambient temperatures (40°C for 12 h; 35°C for 12 h) and 40% humidity from 38 days gestation (dGA; term = 147 dGA) to 116 dGA, which produces placental insufficiency and IUGR (IUGR group; n = 10) (5, 9), or an environmental chamber with normal ambient temperatures and humidity from 43 dGA to 120 dGA [control (CON) group; n = 8). After environmental treatment, all sheep were housed in normal ambient temperatures and humidity for the remainder of the studies. All sheep were given ad libitum access to water. Maternal feed intake was matched on an absolute basis between sheep in CON and IUGR groups. All fetuses in the study were with the product of singleton pregnancies, with the exception of one fetus in the IUGR group that was a product of a triplet pregnancy that was incidentally found at the time of necropsy. The catheterized triplet fetus (fetal weight: 1,466 g) was included in the analysis because it was not an outlier for any physiological parameters measured within the IUGR group as previously published (67).

Fetal surgical procedures.

Pregnant sheep underwent a surgical procedure at 127 ± 1 dGA for the placement of fetal and maternal catheters according to previously published methods (9, 67, 79). Sheep were fasted for 24 h and thirsted for 12 h before surgery. Anesthesia was induced with an intravenous dose of 0.2 mg/kg of diazepam and 20 mg/kg of ketamine and maintained with 2%–4% isoflurane for the duration of the surgical procedure. The fetal lamb was exposed by maternal laparotomy and hysterotomy. In the “study” hindlimb, a polyvinyl catheter was placed in the pudendoepigastric venous trunk with the tip advanced 1 cm into the common femoral vein (V). The deep circumflex iliac artery and vein and the pudendoepigastric arterial trunk were ligated to minimize collateral circulation to midline structures that might contaminate external iliac blood flow to the skin, bone, and muscle of the hindlimb (67). A transit time ultrasonic blood flow transducer (3 mm, Transonic Systems, Ithaca, NY) was placed around the external iliac artery of the study limb for continuous blood flow measurement. A distal arterial occluder was placed for zero-flow corrections. In the “nonstudy” hindlimb, a catheter was placed in the pedal artery with the tip positioned in the external iliac artery, just before the descending aorta (A). A 0.5-mL biopsy of the biceps femoris muscle was obtained at the time of surgery and frozen in liquid nitrogen for baseline isotope enrichment analysis used in a previous publication (67). Ampicillin (500 mg) was injected into the amniotic fluid before the uterus was sutured closed. Once the maternal linea alba and skin were closed, the maternal femoral artery and vein were catheterized. All the catheters and the flow probe were tunneled subcutaneously to the maternal flank and gathered into a pouch that was sutured to the ewe’s skin. After recovery from anesthesia, ewes received 2 days of postoperative treatment with analgesic (1.1 mg/kg of Banamine im, bid), probios (10 g of Probios po, bid) and a total of 5 days postoperative care before experimentation.

Fetal arterial and venous blood sample measurements.

Ewes were conscious and freestanding in their pens and were allowed free access to food and water. Food and water intakes were measured daily to ensure normal hydration and nutrition (67). Four paired, simultaneously drawn A and V blood samples were obtained over 1 h as the external iliac blood flow rate was recorded to establish steady-state conditions. An isovolemic transfusion of heparinized maternal blood (24 mL) was administered to the fetus during the sampling period to replace fetal blood sampled. The results of this metabolic study, including fetal hindlimb glucose, lactate, and total AA net uptake rates and protein metabolic rates, were previously published (67). Our goal was to evaluate the metabolic profile of the most severe IUGR phenotype. From the previously published data set of 13 IUGR fetuses (67), 3 IUGR fetuses fell within 2 standard deviations of the CON fetal weight mean; therefore, they were removed from this metabolomics study to produce a subset of 10 severely affected IUGR fetuses. For the AA net uptake rate measurements only, 2 of 10 IUGR fetuses were excluded because of flow probe malfunction. From this final subset of 10 IUGR and 8 CON fetuses, fetal arterial blood gas values (Radiometer ABL 800 Flex Blood Gas Analyzers, Copenhagen, Denmark), plasma glucose and lactate concentrations (YSI 2900 Biochemistry Analyzer, Yellow Springs, OH), and plasma insulin (7), IGF-1 (9), cortisol (46), and norepinephrine (48) concentrations are presented here to demonstrate the physiological characteristics of the fetuses included in the present study (Table 1).

Table 1.

Fetal blood and plasma parameters

	CON	IUGR
Fetal weight, g	3,324 ± 137	1,331 ± 132*
Fetal hindlimb weight, g	344 ± 17	157 ± 14*
Biceps femoris weight, g	18.1 ± 0.7	8.7 ± 0.9*
Total fetuses	8	10
Male fetus, %	50	50
Gestational age, days	134 ± 0	134 ± 0
pH	7.36 ± 0.0	7.33 ± 0.1*
$\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, mmHg	50.4 ± 0.7	51.6 ± 0.8
$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, mmHg	20.6 ± 0.7	13.6 ± 0.9*
$\text{S}{\text{a}}_{{\text{O}}_2}$, %	48.3 ± 1.6	22.0 ± 2.7*
O2 content, mmol/L	3.3 ± 0.2	1.4 ± 0.2*
Hematocrit, %	34.1 ± 1.0	32.3 ± 1.3
Hemoglobin, mmol/L	6.9 ± 0.2	6.5 ± 0.3
Glucose, mg/dL	17.9 ± 0.5	11.0 ± 1.0*
Lactate, mmol/L	2.0 ± 0.1	2.5 ± 0.2
Insulin, ng/mL	0.4 ± 0.1	0.1 ± 0.0*
IGF-1, ng/mL	108.2 ± 15.2	27.9 ± 6.0*
Cortisol, ng/mL	20.2 ± 5.3	24.3 ± 6.1
Norepinephrine, pg/mL#	617.1 ± 252.5	5,670 ± 2,474*

This cohort of animals is a subset of animals from a previously published study (67). CON (n = 8) and IUGR (n = 10). CON, control; IUGR, intrauterine growth restriction.

^*P ≤ 0.05, unpaired Student’s t tests. Data are represented as means ± SE;

^#Data were log transformed for analysis.

Fetal net hindlimb AA uptake rates.

Fetal plasma A and V AA concentrations were measured using HPLC as previously described (67). External iliac plasma flow was calculated by multiplying external iliac blood flow by (1-hct). Hindlimb (net) AA uptake rates were calculated by multiplying the mean A-V concentration difference from the four steady-state blood draws by the mean hindlimb plasma flow during the draw period. The essential AAs (histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, and valine) were summed to calculate the total essential AA uptake rate, and the remaining nonessential AAs (alanine, arginine, asparagine, aspartate, cysteine, glutamate, glutamine, glycine, ornithine, proline, serine, taurine, and tyrosine) were summed to calculate the total nonessential AA uptake rate. All hindlimb AA uptake rates were normalized to 100 g of fetal hindlimb weight and compared between CON and IUGR groups.

Fetal skeletal muscle collection.

After conclusion of the metabolic study, ewes received a dose of diazepam (0.2 mg/kg) and ketamine (20 mg/kg) intravenously, and fetuses were delivered via maternal laparotomy and hysterotomy. The biceps femoris muscle was exposed from the study hindlimb of the anesthetized fetus, and a biopsy was obtained and immediately frozen in liquid nitrogen. This sample was used for metabolomics analysis. A bolus dose of pentobarbital sodium (Fatal Plus; Vortech Pharmaceuticals, Dearborn, MI) was administered intravenously to both the mother and the fetus, after which the fetal weight, fetal hindlimb weight, and biceps femoris weight were obtained from the nonstudy hindlimb. No physical differences were observed between the study and the nonstudy hindlimb.

Hydrophilic metabolite extraction.

Frozen biceps femoris muscle samples were extracted in 15 mg/mL ice-cold lysis/extraction buffer (50% methanol, 30% acetonitrile, 20% water). Fetal plasma samples from the fourth paired A and V blood draw were extracted at 1:25 dilution with the same lysis/extraction buffer. Samples were agitated at 4°C for 30 min followed by centrifugation at 10,000 g for 15 min at 4°C. Protein and lipid pellets were discarded, whereas supernatants were injected into an ultra-HPLC system (Vanquish, Thermo, San Jose, CA) and run on a Kinetex C18 column (150 × 2.1 mm internal diameter, 1.7 μm particle size; Phenomenex, Torrance, CA) at 25°C using a 3-min isocratic method (5% Optima acetonitrile, 95% Optima H₂O, 0.1% formic acid) flowing at 250 μL/min (56). The ultra-HPLC system was coupled online with a Q Exactive mass spectrometer (Thermo, Bremen, Germany), scanning in full mass spectrometer mode (2 μscans) at 70,000 resolution in the 60–900 mass-to-charge ratio range, 4 kV spray voltage, 15 sheath gas and 5 auxiliary gas, operated in negative and then positive ion mode (separate runs, 3 min each) (56). Mass spectrometer stability was assessed by determining peak area coefficient of variation for a technical mixture injected every 10 runs.

Metabolite and pathway analysis.

Metabolite assignments were determined using Maven (Princeton, NJ) (16) following conversion of .raw files into .mzXML format through MassMatrix (Cleveland, OH) and assignments confirmed against a subset of 650 light- and heavy-labeled standards (IROATech, Sigma-Aldrich, St. Louis, MO) as described previously (19).

Statistical and pathway analyses of metabolites were conducted with MetaboAnalyst 4.0 (15). For each metabolite, outliers were identified by Grubbs’ and Tukey’s test and replaced by the median of the individual group, and zero values were replaced by 50% of the lowest value from both groups. Data were then normalized by median, log transformed, and autoscaled. For the skeletal muscle, the groups analyzed were CON and IUGR. A Student's t test was used to determine significant differences in metabolites between groups (CON and IUGR). For the plasma samples, the groups analyzed were CON-A, CON-V, IUGR-A, and IUGR-V. A two-way ANOVA was conducted to analyze the effects of group (CON and IUGR), vessel (A and V), and their interaction. Tukey’s post hoc test was performed to determine differences among groups. The false discovery rate-adjusted P values [based on the Benjamini-Hochberg false discovery rate ≤10% (76)] were applied to adjust for multiple comparison testing (82). Significance was designated at an adjusted P ≤ 0.05, and marginal significance was designated at an adjusted P ≤ 0.06–0.15, mainly to maximize the number of potentially important metabolites involved in muscle metabolism for generation of pathway analysis.

For both the plasma and skeletal muscle, principal component analysis (PCA) was used as an unsupervised method to allow unbiased identification of patterns of metabolites between samples. In addition, partial least squares-discriminant analysis (PLS-DA), a supervised multivariate statistical method, was performed to identify metabolites that differ between groups to account for unwanted biological variation between animals in principal components when using PCA. Pathway analysis for both the plasma and skeletal muscle samples was performed with the top 30 metabolites with the highest variable importance in projection (VIP) scores regardless of adjusted P values. The homo sapiens Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway library was used. The pathway analysis algorithms used were the hypergeometric test for over representation analysis and relative-betweenness centrality for the pathway topology analysis.

Skeletal muscle gene expression analysis.

Total RNA was extracted and purified from snap-frozen fetal biceps femoris muscle (100 mg) using TRIzol LS (Invitrogen, Grand Island, NY) and RNeasy Mini Kit (Qiagen, Germantown, MD) with RNase-Free DNase Set (Qiagen) according to the manufacturer’s protocols. RNA (2 µg) was converted to complementary DNA (cDNA; 100 ng/μL) using SuperScript III First-Strand Synthesis System (Invitrogen) according to the manufacturer’s protocol. All cDNA samples (3–20 ng) were run in triplicate using FastStart Universal SYBR Green Master (Roche, Pleasanton, CA). The quantitative real-time PCR was performed with a relative standard curve of pooled skeletal muscle cDNA for relative quantification (Lightcycler 480 II; Roche Life Science, Indianapolis, IN) using the following conditions of amplification: 95°C for 5 min; 95°C for 15 s, 60°C for 30 s, 72°C for 30 s (40 cycles); and melting curve from 60°C to 95°C (10, 71). Primers (0.5 μM, final concentration) used were designed for SYBR green assays using either Ovis aries or Bos taurus genome sequence (Table 2) (8, 73, 74). The gene of interest was normalized to the average of three reference genes: β-actin (ACTB), glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and ribosomal protein S15 (RPS15). Expression of reference genes was not different between treatment groups. For each gene, the expression data are presented as a fold change compared with the average of the CON group. The quantitative real-time PCR experiments and analysis were performed according to the Minimum Information for Publication of Quantitative Real-Time PCR Experiments guidelines (11).

Table 2.

Primers used for real-time quantitative PCR assays

Gene Symbol	Gene Name	Forward Primer	Reverse Primer	Accession Number	Species
ACTB	β-Actin	TGCAGAAAGAGATCACTGCC	GACAGCGAGGCAGGATGG	NM_001009784	Ovis aries
ALT1	Alanine aminotransaminase 1	AGCCCTTCACCGAGGTCAT	CACGCCTGCAAGATGCGC	NM_001083740	Bos taurus
AST	Aspartate transaminase	AAAGCTCCCGAGTTCTCCAT	CCCTGATAGGCCGAGTCAA	XM_004015042	Ovis aries
BCAT1	Branched-chain amino acid transaminase 1	CATCCTGGACTTGGCACACA	CAGGCGGTACCTGAACCAAA	NM_001009444	Ovis aries
BCAT2	Branched-chain amino acid transaminase 2	TGTCCTCCGTTTCCACAAGG	AGCTTTACACCGGGAGCATC	XM_027978581	Ovis aries
BCKD	Branched-chain α-keto acid dehydrogenase	CGGCAGGGCCAGATCATC	GCCATAGTTGGTCATGTAGAA	XM_024984691	Bos taurus
BCKDK	Branched-chain ketoacid dehydrogenase kinase	AAAGTGGGTGGACTTTGCCA	GCATCGGGATGAAGGGGAAA	XM_004020920	Ovis aries
GAPDH	Glyceraldehyde-3-phosphate dehydrogenase	TGGAGGGACTTATGACCACTG	TAGAAGCAGGGATGATGTTCT	NM_001190390	Ovis aries
GLS	Glutaminase	CCCAGAAGGCACAGACATGGTTGG	GGGCAGAAGCCACCATTAGCCA	XM_027962793	Ovis aries
GS	Glutamine synthetase	GAAAGCCTGCAGAGACCAAT	GCCATTGGAAGGCCAACCA	NM_001040474	Bos taurus
LDHA	Lactate dehydrogenase A	CATGGCCTGTGCCATCAGTA	GGAAAAGGCTGCCATGTTGG	XM_027959817	Ovis aries
LDHB	Lactate dehydrogenase B	GAGGGAGCGATCCCAAACAA	CAGAATGCTGATGGCACACG	XM_027967824	Ovis aries
PC	Pyruvate carboxylase	GCACAGCATGGGGCTTGGCT	AACTGGGCCAGGTCCCCCAC	XM_027959891	Ovis aries
PDK4	Pyruvate dehydrogenase kinase 4	CCCAGAGGACCAAAAGGCAT	GGGTCAGCTGTACAGGCATC	XM_004007738	Ovis aries
PDH	Pyruvate dehydrogenase	GTTAAGGGGGCTGCTAGGTG	AGCCACTGCGTACTGTGAAA	XM_004021953	Ovis aries
PFK	Phosphofructokinase 1	TGGTGGCTCCATGCTGGGGA	GCAGGGCGTGGATGCTGTGA	XM_004006406	Ovis aries
PK	Pyruvate kinase	ACCACGCAGAGACCATCAAG	GGTCCTTTAGTGTCCAGGGC	XM_012180891	Ovis aries
RPS15	Ribosomal protein S15	ATCATTCTGCCCGAGATGGTG	CGGGCCGGCCATGCTTTACG	XM_015096022	Ovis aries

All primer sequences are from 5′ to 3′. Accession numbers refer to the published gene sequences from the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/gene) from which the primer sequences were designed.

Statistical analysis.

Analysis of metabolomics and pathway enrichment in CON versus IUGR are described above. For fetal physiological and blood/plasma parameters, individual AA uptake rates, and targeted gene expression analysis, the data are presented as mean values ± SE with consideration for statistical significance at P ≤ 0.05. Unpaired Student’s t tests were used for direct comparisons between CON and IUGR groups, and a Mann-Whitney test was used when variances were unequal (Graph Pad Prism 5). Relationships between various AA uptake rates [branched-chain amino acid (BCAA), alanine, glycine, glutamine] and fetal hindlimb weight and/or oxygen content were determined using simple linear regression analysis. Regression relationships were determined for all fetuses pooled, and P ≤ 0.05 was considered statistically significant. The total number of animals per group was not large enough to have sufficient power to fully evaluate for sex differences; thus, we did not include sex in our statistical model.

RESULTS

Fetal characteristics.

Table 1 shows the physiological measurements of the fetuses included in this study, which is a subset of animals from the previously published study (67). The fetal weights and hindlimb weights were 60% and 54% lighter, respectively, in the IUGR group compared with CON (P ≤ 0.05). Additionally, the fetal biceps femoris muscle weight was 52% less in IUGR compared with CON (P ≤ 0.05). IUGR fetal arterial pH, blood oxygen content, and plasma glucose, insulin, and IGF-1 concentrations were 40%–75% lower than CON (P ≤ 0.05). Fetal plasma norepinephrine concentrations were eightfold higher than CON (P ≤ 0.05) (67).

Metabolomics analysis of IUGR and CON fetal skeletal muscle.

The PCA plots showed separation between the IUGR and CON groups (Supplemental Figure S1A; Supplemental Material is available online at https://doi.org/10.6084/m9.figshare.8009198). Subsequently, the top 30 metabolites were identified according to the VIP scores from the PLS-DA scores plot (R² = 0.92, Q² = 0.71) by comparing IUGR versus CON (Fig. 1 A). From the top 30 metabolites, 8 were higher and 22 were lower in IUGR muscle compared with CON muscle (Fig. 1B). In IUGR skeletal muscle, the relative enrichment of the metabolite 2-hydroxyglutarate was 10-fold higher when compared with the CON (adjusted P = 0.013). The metabolites that were lowest in the IUGR muscle compared with CON were 5-hydroxylysine (40%; adjusted P = 0.009) and arginine (30%; adjusted P = 0.009).

Fig. 1.

Metabolomic profiles of skeletal muscle of intrauterine growth restriction (IUGR) and control (CON) fetuses. A: partial least squares discriminant analysis was used to identify metabolites with changes in abundance that defined separation of samples between the IUGR (n = 10 fetuses) and CON (n = 8 fetuses) groups. B: heat map of the 30 metabolites with the highest variable importance in projection scores. Each square is representative of the mean levels of that metabolite for the individual animal. Row values are normalized for each metabolite, and quantitative changes are color coded from red (high) to blue (low). Student’s t test was conducted to detect differences in fold change. *False discovery rate-adjusted P ≤ 0.05.

Skeletal muscle metabolic pathway analysis was performed using the top 30 VIP scores (Table 3). Pathways that were significantly enriched in IUGR skeletal muscle included arginine and proline metabolism; glycine, serine, and threonine metabolism; aminoacyl-RNA biosynthesis; and alanine, aspartate, and glutamate metabolic pathways (adjusted P ≤ 0.15). Selected significant metabolites (pyruvate, 2-hydroxyglutarate, and aspartate) highlighting the alanine, aspartate, and glutamate metabolism pathway in the IUGR and CON fetal skeletal muscle are shown in Fig. 2A.

Table 3.

Metabolic pathways identified in the arterial and venous plasma, and muscle of IUGR vs. control fetuses

	Total in Pathway	Hits	Adj. P Value	Metabolites Up in IUGR vs. CON	Metabolites Down in IUGR vs. CON
Skeletal muscle
Arginine and proline metabolism	77	8	0.0002	Pyruvate, Spermidine	Arginine, Aspartate, N-Acetylornithine, Homocarnosine, 4-Hydroxyproline, g-Glutamyl-g-aminobutyrate
Glycine, serine, and threonine metabolism	48	5	0.007	Pyruvate	Aspartate, Cystathionine, Ectoine, Dimethylglycine,
Aminoacyl-tRNA biosynthesis	75	6	0.007	Lysine, Leucine, Isoleucine	Arginine, Aspartate, Tyrosine
Alanine, aspartate, and glutamate metabolism	24	3	0.059	Pyruvate, 2-Hydroxyglutarate	Aspartate
Valine, leucine, and isoleucine metabolism	27	3	0.067	Pyruvate, Leucine, Isoleucine
Lysine metabolism	32	3	0.091	Lysine, 2-Hydroxyglutarate	Aspartate
Arterial and venous plasma
Citrate cycle (TCA cycle)	20	3	0.051	Citrate	Pyruvate, Succinate
Taurine and hypotaurine metabolism	20	3	0.051	Taurine, Alanine	Pyruvate
Alanine, aspartate, and glutamate metabolism	24	3	0.058	Alanine	Pyruvate, Succinate
Pentose phosphate pathway	32	3	0.101		Glucose, Ribose, Pyruvate

Targeted pathway analysis was performed using MetaboAnalyst. The hypergeometric test and the relative-betweeness centrality test were used for overrepresentation analysis and pathway topology analysis. P values were adjusted using false discovery rate. IUGR (n = 10), intrauterine growth restriction; CON (n = 8), control.

Fig. 2.

Selected features to highlight the tricarboxylic acid (TCA) cycle; alanine, aspartate, and glutamate metabolism; and other key metabolites. A: metabolites from intrauterine growth restriction (IUGR; n = 10 fetuses) and control (CON; n = 8 fetuses) fetal skeletal muscle. B: features from IUGR (shaded box) and CON (open box) fetal arterial (IUGR-A, CON-A) and venous (IUGR-V, CON-V) plasma samples. Data are from MetaboAnalyst as presented in Figs. 1 and 3 and analyzed using Student’s t test (A) or two-way ANOVA (B). *False discovery rate-adjusted P ≤ 0.1 by Tukey’s post hoc test. Box denotes 25th and 75th percentiles; bars denote minimum and maximum values. A, arterial; V, venous.

Metabolomics analysis of IUGR and CON fetal arterial and venous plasma.

PCA plots showed separation between the IUGR and CON groups but not between vessel groups (Supplemental Figure S1B; https://doi.org/10.6084/m9.figshare.8009198). Using a PLS-DA scores plot (R² = 0.99, Q² = 0.89), four groups (CON-A, CON-V, IUGR-A, IUGR-V) were compared to generate VIP scores for each of the metabolites (Fig. 3A). The top 30 metabolites are shown in Fig. 3B, from which 14 metabolites were higher and 16 metabolites were lower in IUGR fetal plasma compared with CON (adj. P ≤ 0.05, group effect; Fig. 3B). Heat maps of the individual animal metabolite profile were generated for the skeletal muscle (Supplemental Figure S2A; https://doi.org/10.6084/m9.figshare.8009198) and the arterial and venous plasma samples (Supplemental Figure S2B; https://doi.org/10.6084/m9.figshare.8009198). The metabolites that were highest in IUGR fetal plasma compared with CON (all with adjusted P ≤ 0.05) included uric acid (4-fold), 2-aminoadipate (3-fold), 4-pyridoxate (3-fold), dopamine (2-fold), and 5,6-dihydrothymine (2-fold). The metabolites that were lowest in the IUGR fetal plasma compared with CON were ascorbate (80%) and 5-hydroxylysine (70%) (Fig. 3B).

Fig. 3.

Metabolomic profiles of arterial and venous plasma of intrauterine growth restriction (IUGR) and control (CON) fetuses. A: partial least squares-discriminant analysis was used to identify metabolites with changes in abundance that defined separation of samples between the IUGR (n = 10 fetuses) and CON (n = 8 fetuses) groups. B: heat map of the 30 metabolites with the highest variable importance in projection scores. For simplicity, each square is representative of the mean levels of that metabolite per group and vessel. Row values are normalized for each metabolite, and quantitative changes are color coded from red (high) to blue (low). A, arterial plasma; FBP, fructose 1,6-bisphosphate; G3P, sn-glycerol 3-phosphate; NAML, N-Acyl-d-mannosaminolactone; THcHDO, 3D-(3,5/4)-trihydroxycyclohexane-1,2-dione; V, venous plasma. ANOVA was performed on all 4 groups (CON-A, CON-V, IUGR-A, and IUGR-V). Student’s t test was performed separately on the arterial and venous plasma of IUGR versus CON. *False discovery rate-adjusted P ≤ 0.05.

A metabolic pathway analysis was performed on the top 30 VIP metabolites from the fetal plasma (Table 3). Significantly enriched pathways identified were the tricarboxylic acid (TCA) cycle; taurine and hypotaurine metabolism; and alanine, aspartate, and glutamate metabolism (adjusted P ≤ 0.15). Top features in the IUGR and CON fetal arterial and venous plasma highlighting the TCA cycle, alanine, aspartate, and glutamate metabolism and other key metabolites are shown in Fig. 2B.

Individual AA uptake rates across the fetal hindlimb.

The total essential and nonessential net hindlimb AA uptake rates were lower by 63% and 81%, respectively, in IUGR compared with CON (Fig. 4A). Individual net hindlimb essential AA uptake rates for the BCAAs leucine, valine, and isoleucine were 42%, 73%, and 50% lower, respectively, in IUGR compared with CON (P ≤ 0.05). The remaining hindlimb essential AA uptake rates (lysine, threonine, phenylalanine, histidine, and tryptophan) were 54%–110% lower in IUGR compared with CON (P ≤ 0.05) (Fig. 4B). Nonessential net hindlimb AA uptake rates for alanine, glycine, and glutamine were 158%, 107%, and 115% lower, respectively, in IUGR compared with CON (P ≤ 0.05). The remaining nonessential AAs glutamate, serine, arginine, and ornithine were 42%–90% lower in IUGR compared with CON (P ≤ 0.05) (Fig. 4B).

Fig. 4.

Reduced hindlimb uptake of essential and nonessential amino acids in the intrauterine growth restriction (IUGR) fetus. A: total amino acid (AA) uptake rates. B: individual AA uptake rates. Data are represented as means ± SE; IUGR (n = 8) and control (CON; n = 8). Arg, arginine; Asp, aspartate; ASPG, asparagine; Cys, cysteine; Glu, glutamate; His; Lys, lysine; Met, methionine; Orn, ornithine; Phe, phenylalanine; Pro, proline; Ser, serine; Tau, taurine; Thr, threonine; Trp, tryptophan; Tyr, tyrosine. *Statistical significance at P ≤ 0.05 vs. CON, Student’s t test. C: linear relationship between hindlimb weight and alanine (Ala), glutamine (Gln), glycine (Gly), leucine (Leu), isoleucine (Ile), and valine (Val) uptake rates. IUGR fetuses and CON fetuses with R² and P values are shown.

We further investigated the relationship between the BCAA, alanine, glycine, and glutamine uptake rates, hindlimb weight, and fetal arterial blood oxygen content. We found a positive association between BCAA, alanine, glycine, and glutamine net uptake rates with hindlimb weight, with the smallest hindlimbs demonstrating net release of alanine, glycine, and glutamine (Fig. 4C). Fetal arterial oxygen content correlated positively with the uptake rates of BCAAs (leucine, r² = 0.43, P = 0.005; isoleucine, r² = 0.40, P = 0.008; valine, r² = 0.33, P = 0.02), alanine (r² = 0.46, P = 0.004), glycine (r² = 0.51, P = 0.002), and glutamine (r² = 0.42, P = 0.007). Fetuses with the lowest oxygen content showed a release of alanine, glycine, and glutamine from the hindlimb (data not shown).

Quantitative real-time PCR analysis.

To understand the activation or repression of specific metabolic pathways in IUGR muscle as indicated by the pathway analysis, we measured expression of target genes involved in energy metabolism (rate-limiting steps of glycolysis and the TCA cycle) and AA catabolism (Fig. 5). Expressions of phosphofructokinase 1 (PFK), pyruvate kinase (PK), and pyruvate carboxylase (PC) were similar between groups. However, expression of pyruvate dehydrogenase (PDH), which decarboxylates pyruvate to acetyl-CoA, was 39% lower (P ≤ 0.05). Expression of pyruvate dehydrogenase kinase 4 (PDK4), an inhibitor of PDH activity, was 3.3-fold higher in IUGR compared with CON (P ≤ 0.05), as we have demonstrated previously (8). Gene expressions of lactate dehydrogenase A (LDHA) and lactate dehydrogenase B (LDHB) were both 42% lower in the IUGR compared with CON (P ≤ 0.05). Likewise, the expressions of branched-chain amino acid transaminase 1 (BCAT1) and branched-chain amino acid transaminase 2 (BCAT2), which transaminate BCAAs to their ketoacids, were 60% and 26% lower in IUGR compared with CON, respectively (P ≤ 0.05). We did not detect any significant difference in the expression of branched-chain α-keto acid dehydrogenase (BCKD), an enzyme that decarboxylates ketoacids. However, the gene expression of its inhibitor, branched-chain ketoacid dehydrogenase kinase (BCKDK), was lower by 31% (P = 0.07). Additionally, the expressions of both alanine aminotransaminase 1 (ALT1) and aspartate transaminase (AST) were lower by 38%, and glutaminase (GLS) expression was lower by 40% in IUGR compared with CON (P ≤ 0.05).

Fig. 5.

Gene expressions of enzymes in the fetal skeletal muscle. A: genes involving in the glycolysis and tricarboxylic acid (TCA) cycle pathways. B: branched-chain amino acid catabolism in the intrauterine growth restriction (IUGR; n = 10) versus control (CON; n = 8). *P ≤ 0.05, unpaired Student’s t tests. ALT, alanine aminotransaminase 1; AST, aspartate transaminase; BCAT1, branched-chain amino acid transaminase 1; BCAT2, branched-chain amino acid transaminase 2; BCKD, branched-chain α-keto acid dehydrogenase; BCKDK, branched-chain ketoacid dehydrogenase kinase; GS, glutamine synthetase; GLS, glutaminase; LDHA, lactate dehydrogenase A; LDHB, lactate dehydrogenase B; PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase; PDK4, pyruvate dehydrogenase kinase 4; PFK, phosphofructokinase; PK, pyruvate kinase.

DISCUSSION

Summary of key findings.

Our metabolomics approach demonstrated that the main metabolic (KEGG) pathways impacted in IUGR fetal skeletal muscle were the TCA cycle and AA metabolism. The shared metabolic pathway identified in both IUGR fetal plasma and in skeletal muscle was alanine, aspartate, and glutamate metabolism. The net uptake rates of the BCAAs alanine, glycine, and glutamine were lower in IUGR fetuses compared with CON and strongly correlated with fetal hindlimb weight. IUGR fetuses with the smallest hindlimbs demonstrated net release of alanine, glycine, and glutamine, potentially to shuttle gluconeogenic substrates or nitrogen out of the skeletal muscle and to the liver or other organs. In addition, both the metabolomics analysis and gene expression levels indicated that in the IUGR skeletal muscle, pyruvate may be used as a substrate for alanine production. Other metabolite differences indicated energy deficiency in the IUGR hindlimb, including lower phosphate concentrations in femoral arterial and venous plasma, lower intramuscular concentrations of adenosine, and higher femoral venous plasma concentrations of uric acid. The lower rate of BCAA uptake by the hindlimb is a likely contributor to the lower rate of skeletal muscle protein accretion and growth in the IUGR fetus compared with CON, shown previously (67). Limiting the energy cost of net protein accretion at the expense of skeletal muscle growth allows for energy conservation in the IUGR fetus. This is a positive fetal response to the lower supplies of nutrients and oxygen because of placental insufficiency and likely functions to promote fetal survival. These results support our hypothesis that IUGR fetal skeletal muscle develops adaptations that divert AAs away from those pathways that lead to protein accretion and into selective metabolic pathways that sustain global energy production and nitrogen disposal (Figs. 6 and 7).

Fig. 6.

Pathway diagram of glycolysis, pentose phosphate pathway (PPP), and tricarboxylic acid (TCA) cycle highlighting metabolites that differed between control (CON) and intrauterine growth restriction (IUGR) fetal skeletal muscle and plasma. Selected metabolites are shown from the top 30 variable importance in projection scores that were higher (green), lower (red), or unchanged (black) in IUGR (n = 10) fetal skeletal muscle (M) and/or femoral arterial and venous plasma (AV) when compared with CON (n = 8) fetuses. The gene expression of selected enzymes (italics) in the fetal skeletal muscle also are shown. PPP production of ribose was lower in IUGR muscle and AV, as was adenosine in IUGR muscle for DNA/RNA and ATP production. Despite lower glycolytic intermediates in IUGR AV, pyruvate concentrations were higher in IUGR muscle. LDHA, LDHB, and PDH expressions was lower in IUGR muscle, and alanine concentrations were higher in IUGR AV, supporting alanine release from muscle, as opposed to lactate production or conversion of pyruvate to acetyl-CoA for entry into the TCA cycle. Lysine and its catabolic product α-hydroxyglutarate were higher in IUGR muscle, yet α-ketoglutarate concentrations were unchanged; leucine/isoleucine concentrations also were higher in IUGR muscle, yet BCAT expression was lower, further supporting limited substrate entry into the TCA cycle in IUGR muscle. Finally, intermediates in the arginine synthesis pathway were lower in IUGR muscle. ALT, alanine aminotransaminase 1; BCAT1, branched-chain amino acid transaminase 1; BCAT2, branched-chain amino acid transaminase 2; FBP, fructose 1,6-bisphosphate; G3P, sn-glycerol 3-phosphate; LDHA, lactate dehydrogenase A; LDHB, lactate dehydrogenase B; PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase; PDK4, pyruvate dehydrogenase kinase 4; PFK, phosphofructokinase; PK, pyruvate kinase.

Fig. 7.

Summary of proposed metabolic adaptations in intrauterine growth restriction (IUGR) skeletal muscle that conserve energy utilization and result in the release of ammoniagenic amino acids (AAs; alanine, glutamine, glycine) and uric acid. Metabolites, AA uptake rates, and gene expressions of enzymes (italics) in the IUGR (n = 10) fetal skeletal muscle or plasma that were higher (green), lower (red), or unchanged (black) when compared with control (CON; n = 8) fetuses. Black and gray-shaded arrows and font show proposed increases and decreases, respectively, in metabolic pathway flux. Based on these analyses, we propose that reduced oxidative metabolism (through glycolysis and the TCA cycle) and reduced availability of phosphate and adenosine result in increased entry of ADP into a salvage pathway for ATP generation. AMP and ammonia are then overly produced in IUGR muscle. AMP is further processed to uric acid via AMP deaminase, and excess ammonia is converted and released into the plasma in the forms of glutamine and glycine. Alanine, which is a gluconeogenic substrate and released into the circulation, is produced as a result of a series of transamination reactions that alleviate the load of pyruvate and glutamate. α-KG, α-ketoglutarate; ADA, adenosine deaminase; BCAA, branched-chain amino acid; BCKA, branched-chain α-ketoacids; GCS, γ-glutamylcysteine synthetase; GLS, glutaminase; GS, glutamine synthetase; Pi, orthophosphate; PPi, diphosphate.

Lower AA uptake rates and the release of alanine, glycine, and glutamine from the IUGR hindlimb.

Glucose, oxygen, and AA availability promote protein synthesis and accretion rates in fetal skeletal muscle (6, 24). Previously, we demonstrated lower protein accretion rates in IUGR fetal skeletal muscle compared with normally growing CON fetal sheep (67). When normalized to the weight of the hindlimb, the hindlimb blood flow rates and net glucose and lactate uptake rates were similar between IUGR and CON fetal sheep; however, weight-normalized hindlimb oxygen delivery and consumption rates were 40% and 29% lower, respectively, in IUGR compared with CON (67). The net total hindlimb AA uptake rates were 55% lower in IUGR compared with CON, indicating that the uptake of AAs was reduced relative to glucose in muscle (67). Therefore, lower oxygen delivery to the hindlimb might limit the capacity of skeletal muscle to take up and metabolize AAs into anabolic pathways or the capacity for AA oxidation.

In the current study, we further evaluated the net hindlimb uptake rates of individual AAs to determine whether the lower total AA uptake rate included specific AAs or groups of AAs (essential vs. nonessential). The net uptake rates of nearly all of the AAs were lower in the IUGR than CON hindlimb. Specifically, the uptake rates of essential BCAAs (leucine, valine, and isoleucine) were lower in IUGR compared with CON. Human and rat studies have shown that BCAAs, especially leucine, independently stimulate muscle protein synthesis (28, 32). In addition, fetal sheep studies have shown that leucine uptake and oxidation rates are lower during hypoxic states, resulting in decreased protein accretion and use of leucine for energy production compared with normoxic controls (53). The decrease in BCAA uptake into IUGR fetal muscle also could be due to the limited capacity of the skeletal muscle for oxidative metabolism (8). In support of this concept, we found that the hindlimb uptake rates of BCAA were lower in the IUGR group and positively correlated with fetal arterial blood oxygen content. Furthermore, BCAA uptake by the skeletal muscle is directly related to circulating concentrations of insulin, a key stimulator of fetal growth (28, 29). Thus, the reduced protein synthesis rates, accretion rates, and overall slower growth of IUGR skeletal muscle, could be due to the lower skeletal muscle BCAA uptake rates.

Despite the lower essential BCAA uptake rates by the IUGR hindlimb, essential AA concentrations (lysine, leucine/isoleucine) were higher in the IUGR muscle compared with CON. This pattern of metabolites indicates a reduction in the intramuscular utilization of essential AAs for protein accretion and/or catabolic processes (84). We have previously shown that in the IUGR fetal hindlimb, there were lower protein synthesis and accretion rates compared with CON (67). In addition, lower gene expressions of BCAT1 and BCAT2 suggest a decrease in transamination of BCAAs to α-ketoacids, the first step in BCAA catabolism. Interestingly, lysine is one of the least oxidized AAs in the fetus (42, 44, 50), potentially resulting in accumulation within the muscle if it is not used for protein synthesis. However, catabolic intermediates in lysine metabolism were higher in both the IUGR muscle (2-hydroxyglutarate, 10-fold) and plasma (2-aminoadipic, 3-fold) compared with CON. Previous work in human primary cells (endothelial cells, smooth muscle cells, and lung fibroblasts) showed that hypoxic stress increases cellular 2-hydroxyglutarate concentrations (58). The significance and/or consequence of this striking increase in 2-hydroxyglutarate in the fetal IUGR muscle is yet to be determined.

Hindlimb uptake rates of alanine, glutamine, and glycine were lower in IUGR and strongly correlated with fetal hindlimb weight and arterial oxygen content. In fact, the IUGR fetuses with the lowest hindlimb weight had negative uptake rates of alanine, glutamine, and glycine, indicating net release of these AAs by the hindlimb. In human conditions of elevated ammonia, ammoniagenic AAs (alanine, glutamine, and glycine) are generated in the tissue to maintain nitrogen balance by carrying ammonia molecules to the liver where they can be converted to urea and excreted by the kidney (27, 68). In IUGR fetal skeletal muscle, several processes may be responsible for a net release of these three AAs. Ammonia is generated from transamination of BCAAs to their corresponding ketoacids. Consistent with this, one of the metabolic pathways that was enriched in both the skeletal muscle and plasma was the alanine, aspartate, and glutamate pathway. In skeletal muscle, alanine and glutamate are synthesized from BCAAs through the transfer of an amine group to pyruvate, oxaloacetate, or α-ketoglutarate, with release of their α-ketoacid products (Fig. 7). In adults, this process is enhanced especially during times of stress, including sepsis, starvation, burn injury, and trauma by increased activity of muscle BCAA aminotransferase (BCAT) (34, 35). Alternatively, in an environment of hypoxia and limited ATP availability as in the IUGR fetus, ammonia is generated from the recycling of ATP via ADP and adenylate kinase. This process generates not only ATP but also AMP, which is further processed to uric acid via AMP deaminase (75), thus resulting in higher levels of plasma uric acid and ammoniagenic AAs (excess tissue ammonia) in the IUGR fetus (Fig. 7).

Glutamine is further synthesized through the formation of an amide bond between ammonia and glutamate (31). Glutamine acts as a nitrogen shuttle among organs to deliver nitrogen to rapidly dividing cells or to eliminate ammonia from muscle (18). Consistent with this cycle, lower expression of GLS, an enzyme that converts glutamine to glutamate, and maintained expression of GS (glutamine synthase) indicates that the synthesis of glutamine was favored over glutamate in IUGR muscle. IUGR fetal skeletal muscle, therefore, preferentially forms alanine, glycine, and glutamine from the limited supply of BCAAs and releases them into the circulation, possibly to the liver for gluconeogenesis (39, 57) to sustain glucose availability to vital organs (77), or for disposal of nitrogen to the liver for urea production (31) (Fig. 7).

Our findings, however, differ from the classically described BCAA-alanine and glutamine cycle in adults, which is mediated by increased BCAT activity and decreased BCKD activity. We observed lower expression of BCAT1 and 2 with maintained expression of BCKD. There are several possible explanations for this difference. First, acute metabolic stressors described in the adult result in the stimulation of proteolysis to make BCAAs available to other organs (34, 35). Conversely, in placental insufficiency, the fetus adapts to a chronic and progressive decline in nutrient and oxygen availability by slowing protein synthesis as opposed to increasing protein breakdown (67). Second, lower BCAT1 and 2 expression in IUGR fetal muscle might be reflective of the overall reduction of BCAA uptake by the IUGR fetal hindlimb, another adaptation to long-standing reduction in AA supply from the placenta (13, 49, 64). Finally, lower BCAT1 and 2 expression might reflect developmentally regulated differences in AA metabolism between the fetus, when growth of the muscle is expendable, and the adult, when muscle is fully mature.

We also found higher concentrations of adrenaline and dopamine in IUGR fetal plasma, which is consistent with previous reports of increased catecholamines in IUGR and hypoxic fetuses (20, 37, 67, 70). Higher levels of catecholamines in the IUGR fetuses have been shown to suppress insulin secretion and stimulate hepatic gluconeogenesis (43, 85), which also supports interorgan communication between the IUGR fetal muscle and the liver.

Higher levels of pyruvate in the IUGR skeletal muscle.

Despite lower glycolytic intermediates, pyruvate concentration was higher in IUGR muscle compared with CON (Fig. 6). Pyruvate is the end-product of glycolysis and can enter multiple metabolic pathways, including conversion to lactate, alanine, or entry into the TCA cycle. IUGR fetuses have lower whole body fractional rates of glucose oxidation (8, 45). Increased gene expression of PDK4 and reduced gene expression of PDH suggest decreased PDH activation as a mechanism for limited glucose oxidation (8). However, despite these differences in gene expression, PDH activity is increased in IUGR skeletal muscle (61). Further studies are needed to measure pyruvate flux in the IUGR muscle and understand its role linking glycolysis and TCA cycle activity. Lower expression of LDHA and LDHB in IUGR muscle indicates limited interconversion between pyruvate and lactate, consistent with normal LDH activity in IUGR muscle (61). Gene expression of ALT, the transaminase that catalyzes the reversible reaction between pyruvate and alanine, also was lower in IUGR fetal skeletal muscle compared with CON. However, higher concentrations of circulating alanine in the venous pool, along with evidence of release of alanine from the smallest IUGR hindlimbs, supports increased glucose to alanine flux in the muscle via pyruvate and likely downregulation of ALT to prohibit conversion of alanine back to pyruvate (14, 57).

We speculate that hypoxia-inducible factor (HIF) signaling might play a role in how the IUGR fetal muscle responds to chronically lower substrate and oxygen delivery from the placenta. It has been proposed that HIF signaling is activated within mature muscle exposed to hypoxia (HIF-2α) and promotes the upregulation of glucose transporter 1 and 4, insulin-like growth factor binding protein-1, and the glycolytic enzymes (36, 63), all of which might explain the higher levels of pyruvate in IUGR fetal skeletal muscle. Likewise, it has been reported that immature skeletal muscle cells are more dependent on glucose and anaerobic metabolism versus mature cells (83). The potential activation of signaling pathways, such as HIF, which might link placental insufficiency and chronic fetal hypoxemia with muscle metabolic adaptations to maintain glycolysis and/or pyruvate metabolism in the context of lower plasma glucose concentrations requires further study.

Lower rates of arginine and proline metabolism.

The most significant metabolic pathway identified by KEGG analysis in the IUGR skeletal muscle was arginine and proline metabolism. In the IUGR muscle, both arginine and proline concentrations were lower compared with CON. In late gestation fetal sheep, the level of arginine concentration in the skeletal muscle is positively associated with muscle weight (69). In neonatal piglets, arginine supplementation enhances protein synthesis in the skeletal muscle (25) similar to the action of BCAAs. Therefore, lower concentrations of arginine observed in the IUGR fetal muscle might contribute to lower skeletal muscle growth rates compared with CON. As the precursor to nitric oxide, arginine serves as a nitrate reservoir in muscle (54, 62, 81). Additionally, metabolites involved in redox homeostasis were lower in the IUGR muscle, specifically dimethylglycine, cystathionine, and γ-glutamylcysteine (precursor to glutathione) (51, 52, 65). In the IUGR plasma, there were lower levels of ascorbate and Cys-Gly but higher levels of dehydroascorbate. Future studies are needed to investigate whether lower concentrations of arginine result in lower nitric oxide and impaired vasodilation of the microcirculation and whether an oxidative stress may contribute to reduced oxygen and nutrient delivery to skeletal muscle in the IUGR fetus.

Several metabolites identified within the arginine and proline metabolic pathway are also involved (directly or indirectly) in the urea cycle (e.g., spermidine, arginine, aspartate, acetylornithine, homocarnosine, and γ-glutamyl-GABA). In addition, metabolites acetylcitrulline, homocitrulline, and N-carbamyl-l-glutamate are related to urea cycle metabolism and were lower in IUGR muscle compared with CON (Fig. 6). The urea cycle primarily occurs in the liver and, to a lesser extent, the kidney and intestine. The skeletal muscle is not known to have complete urea cycle function, but enzymes in the cycle that are present in the muscle could contribute to metabolism of nitrogen-containing molecules as they relate to nitrogen balance, oxidative stress, and generation of intermediate AAs. These findings, and urea cycle enzymes as they relate to fetal skeletal muscle metabolism, warrant further investigation.

Evidence for lower energy status in IUGR muscle.

Several metabolites in the IUGR plasma and muscle indicate lower energy status, consistent with our previously observed reduction in hindlimb oxygen consumption rate (67). First, orthophosphate and diphosphate concentrations were lower in both the arterial and femoral venous plasma of IUGR fetuses, which could result in lower ATP production by the muscle. Second, the concentration of ribose was lower in both the IUGR plasma and skeletal muscle compared with CON (Fig. 6). Third, the purine adenosine was lower in the IUGR muscle, and purines are also used for the generation/synthesis of ATP (Fig. 7).

Low levels of ribose and purines (particularly adenosine) in the IUGR muscle could indicate an increased consumption of ribose sugars and purines as global energy currency, at the expense of local skeletal muscle energy metabolism. Consistent with this hypothesis, we observed a fourfold increase in uric acid in IUGR arterial and femoral venous plasma compared with CON. Uric acid is a catabolite of purines (e.g., adenosine and guanine) and is elevated, particularly in the fetus and neonate, in situations of extreme energy deficiency, including inborn errors of fat metabolism (e.g., medium-chain acyl-CoA dehydrogenase deficiency) and carbohydrate metabolism (e.g., glycogen storage diseases and hereditary fructose intolerance) (66). Glucose is metabolized through the pentose phosphate pathway to generate ribose 5-phosphate, NADPH, and erythrose 4-phosphate (80). Ribose phosphate is used in the synthesis of nucleotides and nucleic acids. It is possible that there is reduced pentose phosphate pathway flux in IUGR muscle in favor of glycolysis for pyruvate production, and consequently, decreased intramuscular synthesis of nucleic acids to support myoblast replication (71, 84). In addition, both purine and pyrimidine (e.g., adenosine and cytosine, respectively) levels were lower in the IUGR muscle compared with CON, suggesting that DNA and RNA contents might be lower as a result of slower rates of replication, transcription, and cellular growth. These results indicate that the IUGR muscle is deficient in both energy and growth and develops adaptations to provide substrates (such as alanine, glutamine, and uric acid) to be used as energy substrates globally, and to sustain cellular metabolism and support growth and essential metabolism of other organs, such as the brain or heart. Additionally, future work should focus on determining whether fetal sex plays a role in the adaptation of fetal skeletal muscle energy metabolism to an IUGR environment (22, 23).

Perspectives and Significance

Metabolomics and pathway analysis identified global metabolic changes unique to IUGR fetal hindlimb skeletal muscle. We conclude that IUGR skeletal muscle develops adaptations in an environment of low oxygen and nutrient availability to prioritize alanine and glutamine production to support the growth and metabolism of other organs in the fetus in lieu of protein accretion and growth. The consequent effects on skeletal muscle metabolism result in elevations in uric acid and other ammoniagenic AAs (e.g., glycine). Although this may help to sustain global energy needs, the consequences of reduced protein accretion in the IUGR fetal muscle can be seen in the phenotypes of reduced muscle mass and reduced myofiber size. The most concerning aspect of this phenotype in IUGR fetuses is that their small muscle mass and myofiber number will limit their postnatal hypertrophic muscle growth (26, 33, 41). In addition, the restricted growth of these fetuses not only affects the mass of the skeletal muscle, but also has a serious metabolic cost with more likelihood of developing adult metabolic diseases and cardiovascular events (e.g., diabetes, obesity, coronary disease, etc.) (21, 55). Determining changes in key metabolic pathways in the IUGR fetal skeletal muscle has potential to provide an experimental basis for potential targets for hormonal, nutrient, and oxygen supplements to increase IUGR muscle growth.

GRANTS

This research was supported by the National Institutes of Health Grants R01-HD-079404 (to L. D. Brown), R01-DK-088139 (to P. J. Rozance), and R01-DK-108910 (to S. R. Wesolowski) and by The Center for Women’s Health Research at the University of Colorado School of Medicine (to L. D. Brown).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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