Maturation of lipid metabolism in the fetal and newborn sheep heart

Abstract

At birth, the fetus experiences a dramatic change in environment that is accompanied by a shift in myocardial fuel preference from lactate and glucose in fetal life to fatty acid oxidation after birth. We hypothesized that fatty acid metabolic machinery would mature during fetal life in preparation for this extreme metabolic transformation at birth. We quantified the pre- (94-day and 135-day gestation, term ∼147 days) and postnatal (5 ± 4 days postnatal) gene expression and protein levels for fatty acid transporters and enzymes in hearts from a precocial species, the sheep. Gene expression of fatty acid translocase (CD36), acyl-CoA synthetase long-chain 1 (ACSL1), carnitine palmitoyltransferase 1 (CPT1), hydroxy-acyl dehydrogenase (HADH), acetyl-CoA acetyltransferase (ACAT1), isocitrate dehydrogenase (IDH), and glycerol phosphate acyltransferase (GPAT) progressively increased through the perinatal period, whereas several genes [fatty acid transport protein 6 (FATP6), acyl-CoA synthetase long chain 3 (ACSL3), long-chain acyl-CoA dehydrogenase (LCAD), very long-chain acyl-CoA dehydrogenase (VLCAD), pyruvate dehydrogenase kinase (PDK4), phosphatidic acid phosphatase (PAP), and diacylglycerol acyltransferase (DGAT)] were stable in fetal hearts and had high expression after birth. Protein expression of CD36 and ACSL1 progressively increased throughout the perinatal period, whereas protein expression of carnitine palmitoyltransferase 1a (fetal isoform) (CPT1a) decreased and carnitine palmitoyltransferase 1b (adult isoform) (CPT1b) remained constitutively expressed. Using fluorescent-tagged long-chain fatty acids (BODIPY-C₁₂), we demonstrated that fetal (125 ± 1 days gestation) cardiomyocytes produce 59% larger lipid droplets (P < 0.05) compared with newborn (8 ± 1 day) cardiomyocytes. These results provide novel insights into the perinatal maturation of cardiac fatty acid metabolism in a precocial species.

NEW & NOTEWORTHY This study characterized the previously unknown expression patterns of genes that regulate the metabolism of free fatty acids in the perinatal sheep myocardium. This study shows that the prenatal myocardium prepares for the dramatic switch from carbohydrate metabolism to near complete reliance on free fatty acids postnatally. Fetal and neonatal cardiomyocytes also demonstrate differing lipid storage mechanisms where fetal cardiomyocytes form larger lipid droplets compared with newborn cardiomyocytes.

INTRODUCTION

One of the curious features of adult-onset cardiac pathology is the reversion of specific proteins to fetal isoforms that promote aerobic glycolysis (1, 2). Although the primary triggers for this isoform switching are not known, metabolic stress has been suggested to play a key role (2). Although investigators are seeking to determine whether this return to a fetal, glycolytic metabolic profile reflects a maladaption in disease states, few studies are focusing on the normal, ubiquitous metabolic switch that occurs around the time of birth. Understanding the normal ontogeny of the metabolic genes of the maturing myocardium will broaden our understanding of the fuel requirements at different stages of development (3–6) and may shine light on the processes underlying the reversion of adult genes to their immature expression patterns (1).

It is well established in several mammalian species that the fetal heart uses aerobic glycolysis with glucose and lactate (5, 7–10) as primary fuels, whereas the adult heart prefers oxidative phosphorylation using long-chain fatty acids (11–13). Although it is known that the shift from glycolysis and carbohydrate oxidation in immature cardiomyocytes to oxidative phosphorylation in more mature myocardium occurs soon after birth (5, 14), changes in metabolic machinery that underlie this shift are unknown.

The fetus is hypoxemic relative to the adult but has readily available lactate generated by the placenta and low levels of circulating fatty acids, conditions that favor carbohydrate oxidation. The postnatal increase in the energy demand by the heart with increasing systolic load, along with increased oxygen availability and elevated circulating fatty acid concentrations, require a switch to a higher energy yield substrate such as fatty acids. There is an increase in carnitine palmitoyltransferase 1 (CPT1) gene expression during the transition from fetus to newborn (15), but it is not known whether this increase begins before birth or is primarily stimulated by the birth process or postnatal suckling. Because fuel substrates, metabolic demands, and arterial pressure (16) all change within minutes at birth, the machinery required to convert fatty acids found in milk to the generation of ATP presumably must be ready to function at the time of birth in both altricial and precocial species. Relatively little is known about the developmental maturation of additional machinery responsible for augmenting the use of lipids in the postnatal period in a precocial species such as sheep, whose cardiac developmental timeline relative to birth is more similar to humans than rodents and other altricial species.

Under normal conditions, the late-gestation fetal sheep heart does not oxidize fatty acids to an appreciable degree (17), presumably because both oxygen and free fatty acid substrates are low (18). However, when a fetal heart is exposed to infused exogenous palmitate, it can both take up and oxidize the substrate, though not to the same degree as a newborn lamb (17). This demonstrates the capacity of the immature myocardium to process lipids in utero, even though it is not exposed to high concentrations of this substrate under normal conditions. We hypothesized that the expression of genes that underlie fatty acid metabolism occurs gradually rather than suddenly (19, 20) so that by the time of birth, the system is prepared to metabolize free fatty acids acquired from mother’s milk.

The goal of this study was to determine the normal maturation of the machinery responsible for lipid metabolism in the developing sheep heart. Therefore, we measured gene and protein expression for fatty acid transporters and enzymes in myocardium at two fetal ages (94-day and 135-day gestation; term is 145 days) and one newborn age (5 day). We have shown in sheep that cardiomyocytes are mononucleated through the first two-thirds of gestation, after which they gradually become binucleated and unable to divide (21). At birth, some 70%–80% of cardiomyocytes are binucleated. The ages were selected to represent a two-thirds gestation timepoint (94 days), where most cardiomyocytes are mononucleated and actively proliferating, and a late gestation timepoint (135 days), where there is a mix of mononucleated and binucleated cardiomyocytes (22). The postnatal ages were chosen to select for a cardiomyocyte population that is primarily binucleated and more mature. To test whether fetal cardiomyocytes are capable of fatty acid uptake and storage in a lipid droplet, we incubated isolated cardiomyocytes from fetal (125-day gestation) and newborn (8-day) animals with a fluorescent-tagged long-chain fatty acid molecule, BODIPY-C₁₂.

MATERIALS AND METHODS

Animals

Animal work was conducted with approval of the Institutional Animal Care and Use Committee and Oregon Health and Science University (Portland, OR). All sheep were of mixed Western breeds, were healthy, and within normal flock range for body weight at time of collection. The study group included a balance of males and females in each age group. Animals were used for imaging studies only [125 days gestation (dGA) and 8 days postnatal (dPN)] or for gene and protein expression analysis only (94 dGA, 135 dGA, 5 dPN).

Pregnant ewes were euthanized as previously described (23) via intravenous injection of pentobarbital sodium (SomnaSol, 80 mg/kg, Henry Schein Animal Health). Fetuses were accessed through hysterotomy and umbilical vein injected with 10,000 U of heparin (Baxter, IL) followed by 10 mL saturated potassium chloride to arrest the heart in diastole. Fetal weight was recorded, and the heart was extracted, trimmed in a standardized method, and weighed. Hearts were collected in one of two ways: flash frozen in liquid nitrogen (for tissue analysis) or dissociated (for imaging studies). Only control hearts were included in the present study.

Lambs were euthanized in a similar manner. In brief, they were given a 10,000 U heparin injection intravenously, and followed by SomnaSol. Lamb weight was recorded, and the heart was extracted, trimmed in a standardized method, and weighed. Hearts were collected in one of two ways, flash frozen in liquid nitrogen (for tissue analysis) or dissociated (for imaging studies).

For imaging studies, twelve 125 ± 1 dGA fetuses were included. One fetus per ewe was collected from the following: 1 ewe carrying a singleton, 10 ewes carrying twins, and 1 ewe carrying triplets. Fetuses had been surgically instrumented to serve as controls for other studies that did not include effects on the control animals. Eight 8 ± 1 dPN lambs were included.

For tissue analysis, eight (gene studies) or six (protein studies) fetuses at each 94 ± 2 dGA and 135 ± 5 dGA were included and seven (gene studies) or six (protein studies) 5 ± 4 dPN lambs were included. The 94 ± 2 dGA group was composed of fetuses from five ewes carrying twins, where only one twin of each pair was included, and one ewe carrying triplets, where all three fetuses were included. The studies at 94 ± 2 dGA included the same five fetuses from twin pregnancies and only one fetus from the triplet ewe. Eight 135 ± 5 dGA fetuses were included from eight ewes carrying twins; only one twin of each pair was included. Fetuses were uninstrumented control twins of experimental fetuses (not in the present study) or fetuses not subjected to a surgical procedure. Seven (gene studies) or six (protein studies) 5 ± 4 dPN lambs were obtained from four ewes carrying triplets where only one lamb was used per ewe, one ewe carrying twins where both lambs were included, and one lamb where information was not available; this last lamb was omitted for protein studies. Lambs were uninstrumented or served as controls for experimental siblings. All groups included a balance of males and females (Table 1).

Table 1.

Sex distribution and sample size at different age groups

	Fetal			Postnatal
Age	94 ± 2 dGA	125 ± 1 dGA	135 ± 5 dGA	5 ± 4 dPN	8 ± 1 dPN
qPCR	Male (3) Female (5)		Male (4) Female (4)	Male (5) Female (2)
Western blot	Male (3) Female (3)		Male (3) Female (3)	Male (3) Female (3)
Lipid droplet imaging		Male (8) Female (4)			Male (4) Female (4)

dGA, days gestational age (term ∼147 dGA); dPN, days postnatal.

Cardiomyocyte Isolation

Cardiomyocytes were enzymatically dissociated using retrograde perfusion as previously described (24). Left and right ventricular free walls and the intraventricular septum were separated for gentle agitation in high potassium buffer [Kraftbrühe buffer (KB) solution; 74 μM glutamic acid, 30 μM KCl, 30 μM KH₂PO₄, 20 μM taurine, 2 mM MgSO_4, 0.5 mM EGTA, 10 mM HEPES, and 10 mM glucose; pH 7.37] for cardiomyocyte release into solution. After cardiomyocyte isolation was completed, cells were rested in KB solution (room temperature, 30 min). A 140-µL aliquot of left ventricular cells (∼250,000 cells/mL) were taken for live imaging.

BODIPY-C₁₂ Fluorescent Fatty Acid Live Imaging

Incorporation of BODIPY FL-C₁₂ (Molecular Probes, Cat. No. D3822), an exogenous long-chain saturated fatty acid, into lipid droplets within cardiomyocytes was studied as previously described (25) with minor adjustments for postnatal cardiomyocytes. The BODIPY-C₁₂ conjugate is a 12-carbon-chain-length saturated fatty acid linked to the fluorophore BODIPY (4,4-difluoro-3a,4a-diaza-s-indacene), but biologically resembles an 18-carbon saturated fatty acid. Solutions of BODIPY FL-C₁₂ (10 µM in DMSO) were prepared 1:250 in fetal or newborn KB solution [KB supplemented with 2 mM glutamine, 200 µM sodium pyruvate, 2 or 1 mM lactate (fetal vs. newborn, respectively), 1 or 5 mM glucose (fetal vs. newborn, respectively)] supplemented with 0.1% fatty acid-free bovine serum albumin (BSA) and incubated for 30 min (37°C, in the dark) to allow fatty acid: BSA conjugation.

Cells were incubated in fetal or newborn KB in 8-well microslides (Ibidi, Cat. No. 80821) with 2 µM of BODIPY-C₁₂ for 60 min (39°C or 37°C, for fetal vs. newborn cells, respectively) and imaged using the ×63 oil lens on the ZEISS LSM 880 with Airyscan. Z-stack images were collected after 60 min. BODIPY was imaged with a 488-nm Argon (intensity 0.6, gain 825, digital gain 1.0). About 90–130 slices (0.2 µm thick) were acquired per frame (18–26 µm thickness, 3–12 frames per animal).

Image Analysis

Z-stacks were processed and analyzed as previously described (25). Images were enhanced using Enhance Local Contrast (CLAHE: block size = 9, histogram = 256, maximum = 4), despeckled, and background subtracted (rolling = 5). Lipid droplet particles were analyzed by filtering for the following parameters: circularity (0.8–1), size [0.0314 (minimal detectable size for LSM880) to 3 (exceeds the maximum nonadipocyte lipid droplet size) (26)], and Auto Threshold (Intermodes dark), and images were converted to a mask for lipid droplets and entire cells. For each animal, 10–30 cells were measured in an average of five frames (range of 3–8 frames), depending on the number of cells. The mean area of individual lipid droplets in the present study was 0.19 µm² with a range of 0.036–1.18 µm². Masks were manually verified to ensure the parameters did not accidentally exclude viable or include nonviable cells.

RNA Isolation and Gene Expression

RNA was isolated as previously described (25) from 40 to 50 µg of left ventricular myocardium using TRIzol, a steel bead, and TissueLyser LT (3 min, 50 Hz; Qiagen, Germantown, MD), and purified using RNeasy Mini Columns (Qiagen). Reverse transcription (1 µg RNA) was conducted to synthesize cDNA (High-Capacity cDNA Reverse Transcription Kit; ABI) and diluted 1:20 before PCR. Quantitative PCR was carried out using SYBR Green in the Stratagene Mx3005. Primers are listed in Table 2. Gene expression was analyzed using the ΔCt method. Genes of interest were normalized to housekeeping gene RPL37a, which was not altered by age.

Table 2.

Primer sequences

Gene ID	Forward	Reverse
RPL37a (reference gene)	ACCAAGAAGGTCGGAATCGT	GGCACCACCAGCTACTGTTT
CD36	CTGGTGGAAATGGTCTTGCT	ATGTGCTGCTGCTTATGGGT
FATP6	TTGGAAATGGAGCACGCAGTGA	CTCCCGACTGATCCAATTTTCCCA
ACSL1	GAGCAGAGGTTCTCAGTGAAGCAA	CGGCTGTCCATCCAGGATTCAATA
ACSL3	GACAGATGCCTTCAAGCTGAAACG	GGAATGGACTCTGCCTCACAGTTT
CPT1	GGATGTTTGAGATGCACGGC	GCCAGCGTCTCCATTCGATA
PDK4	CCTGTGATGGATAATTCCCG	TTGGTTCCTTGCTTGGGATA
IDH	CTGTGTTTGAGACGGCTACAAGGA	CGTAGCTGTGGGATTGGCAATGTT
LCAD	TGAAAGCCGCATTGCCATTGAG	ACTTGGATGGCCCGGTCAATAA
VLCAD	AAGATCCCTGAGTGAAGGCCA	TAGAACCAGGATGGGCAGAAA
HADH	AGAAAACCCCAAGGGTGCTGAT	GCCTCTTGAACAGCTCGTTCTT
ACAT1	CTGGGTGCAGGCTTACCTATTTCT	CATAGGGGACATTGGACATGCTCT
GPAT	GAAGTGGCTGGTGAGTTAAACCCT	CAGTCTGATCATTGCCGGTGAAAC
PAP	AGAATGAAGGGAGACTGGGCAAGA	GCAACCAGAGCTCCTTGAATGAGT
DGAT	AGACACTTCTACAAGCCCATGCTC	AGTGCACTGACCTCATGGAAGA

ACAT1, acetyl-CoA acetyltransferase; ACSL1, acyl-CoA synthetase long chain 1; ACSL3, acyl-CoA synthetase long chain 3; CD36, fatty acid translocase; CPT1, carnitine palmitoyltransferase 1; DGAT, diacylglycerol acyltransferase; FATP6, fatty acid transport protein 6; GPAT, glycerol phosphate acyltransferase; HADH, hydroxy-acyl dehydrogenase; IDH, isocitrate dehydrogenase; LCAD, long chain acyl-CoA dehydrogenase; PAP, phosphatidic acid phosphatase; PDK4, pyruvate dehydrogenase kinase; VLCAD, very long chain acyl-CoA dehydrogenase.

Western Blotting

Protein was isolated as previously described (25). In brief, 20–30 µg of left ventricular myocardium was homogenized in RIPA lysis buffer (MilliporeSigma) with a Complete Mini Protease 227 Inhibitor tablet (MilliporeSigma) and phosphatase inhibitor cocktails I and II (MilliporeSigma). Homogenate was lysed with chilled lysis buffer and a stainless-steel bead (4 min, 50 Hz; TissueLyser LT). Protein was quantified by a standard bicinchoninic acid assay (Pierce BCA Protein Assay Kit) and diluted to 2 µg/µL across all groups. Gels were made using Invitrogen SureCast Stacking Buffer (4% acrylamide) and Resolving Buffer (12% acrylamide). Protein (15 µg) was loaded with 6X dye (10% beta-mercaptoethanol) and separated by SDS-PAGE (90 min, 100 mV, Bio-Rad) with running buffer (24.8 mM Tris, 191.8 mM glycine, 3.5 mM SDS). Protein was transferred to PVDF membranes with transfer buffer (24.8 mM Tris, 191.8 mM glycine) and blocked [5% milk in Tris-buffered saline + 0.01% Tween 20 (TBST), 1 h, room temperature]. Primary antibody incubation (4°C, overnight) was followed by TBST rinses (3 × 10 min) and subsequent secondary incubation (room temperature, 1 h). Primary antibody dilutions and sources are listed in Table 3. For proteins of interest close in molecular weight to housekeeper proteins [acyl-CoA dehydrogenase long chain (ACADL) and the oxidative phosphorylation (OXPHOS) cocktail], blots were stripped with Restore Western Blot Stripping Buffer (Thermo Fisher Scientific) and probed again with either glyceraldehyde 3-phosphate dehydrogenase (GAPDH) or α-tubulin.

Table 3.

Antibodies used

Protein	Dilution	Cat. No.	Vendor	RRID or Reference
CD36	1:1,000	133625	abcam	RRID:AB_2716564
ACSL1	1:1,000	177958	abcam	Jonker and Louey (21)
CPT1a	1:500	12252	Cell Signaling	RRID:AB_2797857
CPT1b	1:1,000	22170	ProteinTech	Ascuitto and Ross-Ascuitto (19)
ACADL	1:1,000	17526	ProteinTech	Giussani et al. (22)
CS	1:1,000	14309	Cell Signaling	RRID:AB_2665545
OXPHOS cocktail containing: NDUFB8, SDHB, UQCRC2, MTCO1, ATP5A	1:500	110413	abcam	RRID:AB_2629281
GAPDH (reference protein)	1:5,000	47339	Novus Biologicals	RRID:AB_10010294
α-Tubulin (reference protein)	1:5,000	2125	Cell Signaling	RRID:AB_2619646
Anti-rabbit HRP (secondary antibody)	1:5,000	7074	Cell Signaling	RRID:AB_2099233
Anti-mouse HRP (secondary antibody)	1:5,000	7076	Cell Signaling	RRID:AB_330924

ACSL1, acyl-CoA synthetase long chain 1; ATP5A, ATP synthase F1 subunit α; CD36, fatty acid translocase; CPT1a, carnitine palmitoyltransferase 1a (fetal isoform); CPT1b, carnitine palmitoyltransferase 1b (fetal isoform); CS, citrate synthase; MTCO1, cytochrome c oxidase subunit 1; NDUFB8, NADH ubiquinone oxidoreductase subunit B8; SDHB, succinate dehydrogenase complex subunit B; UQCRC2, ubiquinol cytochrome b-c1 complex subunit 2.

SuperSignal West Dura chemiluminescence substrate was used to visualize membranes (Thermo Fisher Scientific). GBOX (SynGene) and GeneSys software (version 4.3.7.0) were used to image and analyze the blots. Carnitine palmitoyltransferase 1a (fetal isoform) (CPT1a) blots were developed on film (GeneMate Blue Autoradiography Film, Cat. No. 490001) in a dark room (15 min exposure). Band intensities were expressed as area under the curve and normalized against α-tubulin or GAPDH, which did not have different expression between ages.

Statistics

BODIPY-C₁₂ lipid droplet volume, area, and droplet number were analyzed by unpaired t test. One-way ANOVA, followed by Tukey’s multiple comparison test, was used to test for differences in gene and protein expression. All data were analyzed using GraphPad Prism 6 and are presented as means ± SD unless noted otherwise. P values <0.05 were considered statistically significant.

RESULTS

See Fig. 1 for graphical representation of the fatty acid metabolism pathway under investigation.

Figure 1.

Schematic of fatty acid uptake, activation, oxidation, ATP production, and esterification, highlighting the genes and proteins under investigation. Stars indicate the gene was upregulated before birth. FA, fatty acid; FATP6, fatty acid transport protein 6; CD36, fatty acid translocase; ACSL1/3, acyl-CoA synthetase long chain; PDK4, pyruvate dehydrogenase kinase; PD, pyruvate dehydrogenase; CPT1, carnitine palmitoyltransferase 1; VLCAD/LCAD, very long/long chain acyl-CoA dehydrogenase; HADH, hydroxyacyl-coenzyme A dehydrogenase; ACAT, acetyl-CoA acetyltransferase; CS, citrate synthase; IDH, isocitrate dehydrogenase; SDHB, succinate dehydrogenase complex subunit B; GPAT, glycerol phosphate acyltransferase; PAP, phosphatidic acid phosphatase; DGAT, diacylglycerol acyltransferase. Adapted from Lopaschuk et al. (27).

BODIPY-C₁₂ Lipid Droplet Formation

Individual BODIPY-C₁₂ lipid droplets were significantly larger in fetal cardiomyocytes compared with newborn cardiomyocytes (Fig. 2A). There was no difference in the number of droplets relative to cell area (Fig. 2B).

Figure 2.

Lipid droplet accumulation in fetal (white: 125 ± 1 dGA, n = 12) and newborn (black: 8 ± 1 dPN, n = 8) cardiomyocytes. A: average lipid droplet area (µm²). B: number of lipid droplets normalized to cell area. C: representative images of BODIPY-C₁₂ (green) lipid droplet accumulation in fetal and newborn cardiomyocytes. Data are presented as means ± SD. *P < 0.05. dGA, days gestational age; dPN, days postnatal.

Sarcolemmal Fatty Acid Transporter Expression

Fatty acid translocase (CD36) and fatty acid transport protein 6 (FATP6) are responsible for fatty acid transport into the cardiomyocyte (28, 29). The mRNA levels for CD36 showed progressive increases with advancing age (Fig. 3A), and this was also reflected in the protein expression (Fig. 3C). In contrast, FATP6 mRNA levels were low throughout gestation but were high after birth (Fig. 3B).

Figure 3.

Developmental myocardial gene (relative to RPL37a) and protein expression (relative to α-tubulin) of sarcolemmal fatty acid transporters fatty acid translocase (CD36) (A and C) and fatty acid transport protein 6 (FATP6) (B) in left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 8 genes, n = 6 proteins), 135 ± 5 dGA (half black: n = 8 genes, n = 6 proteins), and 5 ± 4 dPN (white: n = 7 genes, n = 6 proteins) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. dGA, days gestational age; dPN, days postnatal.

Fatty Acid Acylation Enzyme Expression

Acyl-CoA synthetase long chain (ACSL) is responsible for the acylation and activation of fatty acids upon entry to the cell (30, 31). ACSL isoform 1 (ACSL1) had stepwise increases in gene expression with age (Fig. 4A), and protein levels followed the gene expression (Fig. 4C). In contrast, acyl-CoA synthetase long chain 3 (ACSL3) gene levels were low in the fetal heart and high postnatally (Fig. 4B).

Figure 4.

Developmental myocardial gene (relative to RPL37a) and protein expression (relative to GAPDH) of fatty acid acylation enzymes acyl-CoA synthetase long chain 1 (ACSL1) (A and C) and acyl-CoA synthetase long chain 3 (ACSL3) (B) in left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 8 genes, n = 6 proteins), 135 ± 5 dGA (half black: n = 8 genes, n = 6 proteins), and 5 ± 4 dPN (white: n = 7 genes, n = 6 proteins) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. dGA, days gestational age; dPN, days postnatal.

Mitochondrial Fatty Acid Transporter Expression

Carnitine palmitoyltransferase 1 (CPT1) shuttles fatty acids across the outer membrane of the mitochondrion and is thus the rate-limiting enzyme for fatty acid utilization by the mitochondrion (32). It is found in three isoforms designated as the liver isoform (CPT1a or CPT1-L), which is expressed in the fetal heart, the muscle isoform [carnitine palmitoyltransferase 1b (adult isoform) (CPT1b) or CPT1-M], which is most highly expressed in the adult heart, and the brain isoform (CPT1c). Cardiomyocytes express both CPT1a and CPT1b isoforms (33) and in rat cardiomyocytes the shift from CPT1a (fetal) to CPT1b (adult) occurs at weaning (34). Although the mRNA levels for all CPT1 isoforms increase with age (Fig. 5A), this pattern is not mirrored in protein expression of the isoforms. Protein levels of CPT1a (the fetal isoform) decrease from 94 to 135 days gestation and remain low in the newborn myocardium (Fig. 5B). Adult isoform (CPT1b) protein expression appears to be constitutively expressed, consistent with the findings of Bartelds et al.(15) (Fig. 5C).

Figure 5.

Developmental myocardial gene (relative to RPL37a) and protein expression (relative to GAPDH) of mitochondrial fatty acid transporters (CPT1) carnitine palmitoyltransferase 1 (A), carnitine palmitoyltransferase 1a (fetal isoform) (CPT1a) (B), and carnitine palmitoyltransferase 1b (adult isoform) (CPT1b) (C) in left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 8 genes, n = 6 proteins), 135 ± 5 dGA (half black: n = 8 genes, n = 6 proteins), and 5 ± 4 dPN (white: n = 7 genes, n = 6 proteins) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. CPT1a protein ladder not present due to the use of film exposure (15 min), which did not transfer protein ladder. dGA, days gestational age; dPN, days postnatal.

Fatty Acid β-Oxidation Expression

The first step in fatty acid β-oxidation is catalyzed by long-chain or very long-chain fatty acyl-CoA dehydrogenase, LCAD and VLCAD, respectively. LCAD and VLCAD gene expression levels are low in mid and late fetal life and increase more than twofold postnatally (Fig. 6, A and B). Interestingly, LCAD protein expression did not follow the gene expression levels but rather increased toward term and was not different postnatally (Fig. 6E). Downstream β-oxidation pathway enzyme hydroxyacyl CoA dehydrogenase (HADH) displayed a stepwise increase in expression throughout development (Fig. 6C), whereas acetyl-CoA acetyltransferase (ACAT1) gene expression increased with advancing gestation but wasn’t higher after birth (Fig. 6D).

Figure 6.

Developmental myocardial gene (relative to RPL37a) and protein expression (relative to GAPDH) of fatty acid β-oxidation enzymes in long chain acyl-CoA dehydrogenase (LCAD) (A and E), very long chain acyl-CoA dehydrogenase (VLCAD) (B), hydroxy-acyl dehydrogenase (HADH) (C), and acetyl-CoA acetyltransferase (ACAT1) (D) left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 8 genes, n = 6 proteins), 135 ± 5 dGA (half black: n = 8 genes, n = 6 proteins), and 5 ± 4 dPN (white: n = 7 genes, n = 6 proteins) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. dGA, days gestational age; dPN, days postnatal.

Glycolysis/β-Oxidation Regulator Expression

The relative contributions to oxidative metabolism from glycolysis versus fatty acid β-oxidation are heavily regulated by pyruvate dehydrogenase kinase (PDK4). This protein deactivates the final step of glycolysis via phosphorylation and subsequent deactivation of pyruvate dehydrogenase, which would normally yield acetyl-CoA. Inhibition of acetyl-CoA production permits increased carbon flux through the β-oxidation of fatty acids, which also yields acetyl-CoA. Gene expression of PDK4 was low throughout gestation and was higher in the newborn myocardium (Fig. 7).

Figure 7.

Developmental myocardial gene (relative to RPL37a) of glycolysis/β-oxidation regulator pyruvate dehydrogenase kinase 4 (PDK4) in left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 8 genes), 135 ± 5 dGA (half black: n = 8 genes), and 5 ± 4 dPN (white: n = 7 genes) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. dGA, days gestational age; dPN, days postnatal.

Tricarboxylic Acid Cycle Expression

Isocitrate dehydrogenase (IDH) produces the first reducing equivalent (NADH) for supplying electrons to complex I of the electron transport chain. Expression of the IDH gene increases toward term and into postnatal life (Fig. 8A). The first step in the tricarboxylic acid cycle (TCA) cycle joins oxaloacetate and acetyl-CoA and is catalyzed by citrate synthase (CS). CS protein expression was not different between ages (Fig. 8B).

Figure 8.

Developmental myocardial gene (relative to RPL37a) and protein expression (relative to GAPDH) of TCA cycle enzymes isocitrate dehydrogenase (IDH) (A) and citrate synthase (CS) (B) in left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 8 genes, n = 6 proteins), 135 ± 5 dGA (half black: n = 8 genes, n = 6 proteins), and 5 ± 4 dPN (white: n = 7 genes, n = 6 proteins) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. The CS protein band was confirmed to be the correct size in a separate blot; protein ladder did not show up for this particular blot. dGA, days gestational age; dPN, days postnatal.

Electron Transport Chain Expression

Complex I subunit NADH ubiquinone oxidoreductase subunit B8 (NDUFB8) protein expression was low in fetuses and had significantly higher expression in the newborn (Fig. 9A). Electron transport chain subunit protein expression from complex II [succinate dehydrogenase complex subunit B (SDHB)], complex III [ubiquinol cytochrome b-c1 complex subunit 2 (UQCRC2)], and complex IV [cytochrome c oxidase subunit 1 (MTCO1)] was not different between ages (Fig. 9, B–D). ATP synthase (complex V) subunit ATP synthase F1 subunit α (ATP5A) protein levels had a similar pattern to NDUFB8, with low levels in the fetal ages and higher expression in newborn myocardium (Fig. 9E).

Figure 9.

Developmental myocardial protein expression (relative to GAPDH) of electron transport chain subunits NADH ubiquinone oxidoreductase subunit B8 (NDUFB8) (A), succinate dehydrogenase complex subunit B, (SDHB) (B) ubiquinol cytochrome b-c1 complex subunit (UQCRC2) (C) , cytochrome c oxidase subunit 1 (MTCO1) (D), and ATP synthase F1 subunit α(ATP5A) (E) in left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 6 proteins), 135 ± 5 dGA (half black: n = 6 proteins), and 5 ± 4 dPN (white: n = 6 proteins) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. ATP5A, ATP synthase F1 subunit α; dGA, days gestational age; dPN, days postnatal.

Fatty Acid Esterification Pathway Expression

Glycerol phosphate acyltransferase (GPAT) catalyzes the joining of a glycerol-3-phosphate with fatty acyl-CoA to form lysophosphatidic acid, the rate-limiting step in de novo glycerolipid synthesis (35). GPAT gene expression had a stepwise increase with age (Fig. 10A). Phosphatidic acid phosphatase (PAP) catalyzes the production of diacylglycerol (DAG) by dephosphorylation of phosphatidate. Diacylglycerol acyl transferase (DGAT) adds the final fatty acyl-CoA to DAG to produce triacylglycerol. Both PAP and DGAT had low gene expression in both 94 and 135 days fetal myocardium with higher expression in newborns (Fig. 10, B and C).

Figure 10.

Developmental myocardial gene (relative to RPL37a) of fatty acid esterification enzymes glycerol phosphate acyltransferase (GPAT) (A), phosphatidic acid phosphatase(B) (PAP), and diacylglycerol acyltransferase (DGAT) (C) in left ventricular (LV) tissue of 94 ± 2 dGA (black: n = 8 genes), 135 ± 5 dGA (half black: n = 8 genes), and 5 ± 4 dPN (white: n = 7 genes) myocardium. Data are presented as means ± SD, nonshared letters indicate statistically significant differences, P < 0.05. dGA, days gestational age; dPN, days postnatal.

DISCUSSION

Circulating fatty acids sharply increase with suckling in newborns. Given this abrupt change in nutritive input, it stands to reason that the systems required to transport, oxidize, and esterify fatty acids must be maturing prenatally so they are primed to function at birth. The aim of this study was to determine the maturation of the lipid metabolism system in the perinatal sheep heart. We hypothesized that components of this system are in place before birth in sheep, even though some investigators have suggested that the metabolic transition occurs suddenly after birth in other animal models.

The physiology regulating lipid droplet morphology and number is an active area of study (36). In adults, intracellular lipid droplet size is indicative of metabolic status, with larger and more abundant droplets being observed in heart failure, diabetes, and metabolic disease (37). When exposed to BODIPY-C₁₂, we showed that lipid droplet size was significantly larger in fetal cardiomyocytes compared with newborn, suggestive of suppressed fatty acid oxidation. Larger lipid droplets in fetal cardiomyocytes suggest a propensity for storage of exogenous lipids rather than β-oxidation as supported by the low fetal gene expression of LCAD and VLCAD, the first step of β-oxidation. It is possible that fatty acyl-CoA substrate would then accumulate in the mitochondrial matrix rather than effectively cycle through β-oxidation. Low expression of LCAD and VLCAD might account for the differences in fatty acid oxidation in fetuses infused with palmitate compared with newborn lambs (17). The ability for uptake and oxidation in palmitate-infused fetuses suggests that components of the transport pathways are mature enough to allow entry of fatty acids into the mitochondrial matrix. Furthermore, developmental upregulation of CPT1 as well as downstream mitochondrial β-oxidation enzymes HADH and ACAT1 and TCA cycle enzyme IDH explains why the fetal myocardium is able to oxidize exogenously supplied fatty acids.

Critical components of the lipid metabolism pathway including a sarcolemmal fatty acid transporter (CD36), cytosolic fatty acid activation enzyme (ACSL1), mitochondrial transporter (CPT1), β-oxidation enzymes (HADH and ACAT1), TCA cycle enzyme (IDH), and esterification enzyme (GPAT) are upregulated before birth. These transporters and enzymes presumably enable the fetal heart to oxidize and esterify fatty acids to some degree in preparation for postnatal life and to be functional by birth. Interestingly, however, not all genes of interest showed this prenatal maturation: FATP6, ACSL3, LCAD, VLCAD, PDK4, NDUFB8, ATP5A, PAP, and DGAT had low expression in mid and late fetal life and higher expression in the postnatal period. PDK4 is critical in the shift from glycolysis to β-oxidation by regulating the flow of acetyl-CoA through modulating pyruvate dehydrogenase (38). It is not clear whether this upregulation is initiated in the final days before birth, is stimulated by the immediate changes in cardiac function and oxygenation at birth, or in response to the initiation of enteral feeding and high circulating lipid levels.

Most gene mRNAs and proteins measured in this study were in concordance. Reliable, specific antibodies were not available for many of the proteins of interest. Thus, concordance was not able to be determined for all genes studied. However, gene expression for CPT1 (nonspecific to isoform a or b) increased with age, whereas CPT1a protein expression dropped off during fetal life. CPT1b, the adult/muscle isoform, was notably similar between ages. Bartelds et al. (15) reported CPT1b (CPT1-M) protein expression during development in sheep myocardium and concluded that it was constitutively expressed over the latter half of gestation. CPT1a and b activities were not measured in this study. Thus, we cannot conclude that constitutive protein expression levels are indicative of constitutive function. CPT1a has been shown to be regulated by epigenetic mechanisms in the context of a high-fat diet (39), which would be encountered with suckling, this could explain the higher CPT1 expression postnatally. In addition, long-chain acyl-CoA dehydrogenase (LCAD) gene expression did not match the appearance of protein. These discrepancies could be due to differential regulation at the translational level. There appear to be species differences, so it is also possible that LCAD is not the primary acyl-CoA dehydrogenase (ACD) involved in energy production of the sheep heart. In humans, VLCAD is the primary acyl-CoA dehydrogenase involved in fatty acid oxidation and LCAD is thought to be of little consequence for energy production. Sheep may rely primarily on VLCAD for cardiac fatty acid oxidation similar to humans (40), whereas mouse models have suggested that both LCAD and VLCAD are active in murine cardiomyocytes (40–42).

Fatty acid metabolism gene expression increases during fetal life in human hearts between 10–12 wk and 16–18 wk (43), suggesting that the commencement of fatty acid metabolic maturation may occur early in gestation. CD36, ACSL1, and GPAT gene upregulation would presumably augment fatty acid transport into the cell, acylation, and esterification within the fetal cardiomyocyte, respectively. Fatty acids acetylated by ACSL1 are preferentially destined for storage in a lipid droplet (44).

Before this study, the temporal regulation of genes that underlie the metabolic machinery required to both oxidize and esterify fatty acids had not been determined for large mammals in the perinatal period. In addition, it was unclear whether the esterification and subsequent storage of exogenously supplied fatty acids were equally functional in fetal and newborn hearts or whether the esterification pathway undergoes a gradual maturational process. Since the ACSL1 gene and protein expression increases with age, along with GPAT gene expression, while PAP and DGAT are only higher in the newborn, some but not all of the enzymes required for fatty acid esterification are maturing before birth. It stands to reason that newborn myocardium would have the highest need for esterification enzymes due to drastic increases in circulating lipids at the time of the first feed. The prenatal maturation of these elements could enable storage of fatty acids in immature cardiomyocytes to some degree, especially if the mitochondrial elements needed to metabolize the lipids to energy are not yet mature, but the implications of exposure to high concentrations of exogenous lipids coupled with this immature esterification system remain unclear.

Although the number of fetuses in the uterus has not been shown to influence gene and protein expression, it is nonetheless important to note that the samples included in this study came from a mixture of singleton, twin, and triplet pregnancies. In addition, the present study was not powered to identify sex differences; no differences were found for gene expression, protein expression, or lipid droplet size.

These observations are important to consider in the context of intravenous lipid support in premature infants. The gradual maturation of fatty acid metabolism machinery described here agrees with a previous study showing that the fetal heart is capable of fatty acid oxidation (17, 45), though not to the same degree as newborns. We also observe herein that fetal cardiomyocytes store fatty acids in larger lipid droplets compared with newborns, which may be indicative of suboptimal lipid processing as seen in failing adult hearts. Lipid exposure in utero has been shown to cause metabolic stress and impaired cardiac function along with increased lipid peroxidation in rat offspring (46). This increased lipid peroxidation, indicative of increased oxidative stress, could be the mechanism underlying lipid-induced cardiotoxicity. The near-term fetal heart does appear to have some antioxidative mechanisms in place (47), but the degree to which they are able to mitigate higher levels of oxidative stress from premature exposure to lipid remains to be determined. Although several studies have investigated the impact of hypoxia on antioxidant responses in fetal sheep (22, 48), the role of lipid exposure on the cardiac antioxidant response has not been studied. Future studies should measure antioxidant activity and lipid peroxidation in lipid-infused fetuses or a model of preterm birth to determine the viability of this hypothesis under varying conditions of oxygenation.

An additional clinical concern is the possibility that intravenous lipid nutrition protocols could be disadvantageous. A randomized controlled trial of short-term exposure to exogenous lipids in premature infants led to a reduction in myocardial left ventricular peak systolic apical circumferential strain at age 23–28 (49), suggesting impaired myocardial function in these young adults. The infants in this randomized controlled trial were born at 28 wk gestational age (∼70% developed) and the fetuses used for the imaging studies in the present study are ∼85% developed (125 dGA, term ≈147 dGA), so presumably the ability of premature babies to handle exogenous lipids is on par or less able than those cardiomyocytes used in our present study. Exposing immature cardiomyocytes to lipid resulting in larger lipid droplet stores may be integral to the mechanism leading to impaired myocardial contractile function in young adulthood.

Future studies should attempt to determine which specific perinatal changes (circulatory, respiratory, nutritive, hormonal, or other) are most highly associated with upregulation of fatty acid metabolism genes at different parts in the pathway. Our data show that a number of genes have a stepwise increase during gestation, but for most, the expression in the newborn lamb is higher than that in the fetus. It is possible that birth, or suckling, is the triggering event for this considerable increase, but upregulation by simple increased availability of free fatty acids in milk cannot be understated.

These findings increase our understanding of prenatal metabolic abilities, an important consideration in devising lipid nutrition strategies in preterm infants. Fetal cardiomyocytes store exogenous long-chain fatty acids in larger droplets relative to newborn, suggesting a different strategy for fatty acid storage and utilization in fetal life compared with postnatal life. The immature heart may be less able to properly esterify and package fatty acids, resulting in a haphazard storage mechanism in the form of larger lipid droplets that may have long-term consequences for preterm or vulnerable infants receiving parenteral lipids. These observations offer novel insights into the metabolic maturation of fatty acid metabolism in cardiomyocytes in the perinatal period. Future studies should compare changes in the relative expression and function of the metabolic machinery required for carbohydrate metabolism in relation to the changes in lipid machinery that we have reported here. Characterization of this normal metabolic transition aids in our comprehension of the reversal of this process seen in failing adult hearts that revert back to a fetal metabolic profile.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This research was supported by the National Heart, Lung and Blood Institute (Grant R01 HL146997 to K.L.T.). R.R.D. was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development National Research Service Award Fellowship (Grant F30 HD096812) and by the American Heart Association Grant 19PRE34380190. K.L.T. was supported by National Institutes of Health Grant P01 HD034430-20.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

R.R.D. conceived and designed research; R.R.D. and S.L. performed experiments; R.R.D. and K.L.T. analyzed data; R.R.D., S.L., and K.L.T. interpreted results of experiments; R.R.D. prepared figures; R.R.D. drafted manuscript; R.R.D., S.L., and K.L.T. edited and revised manuscript; S.L. and K.L.T. approved final version of manuscript.