Impaired cardiac energy metabolism in embryos lacking adrenergic stimulation

Abstract

As development proceeds from the embryonic to fetal stages, cardiac energy demands increase substantially, and oxidative phosphorylation of ADP to ATP in mitochondria becomes vital. Relatively little, however, is known about the signaling mechanisms regulating the transition from anaerobic to aerobic metabolism that occurs during the embryonic period. The main objective of this study was to test the hypothesis that adrenergic hormones provide critical stimulation of energy metabolism during embryonic/fetal development. We examined ATP and ADP concentrations in mouse embryos lacking adrenergic hormones due to targeted disruption of the essential dopamine β-hydroxylase (Dbh) gene. Embryonic ATP concentrations decreased dramatically, whereas ADP concentrations rose such that the ATP/ADP ratio in the adrenergic-deficient group was nearly 50-fold less than that found in littermate controls by embryonic day 11.5. We also found that cardiac extracellular acidification and oxygen consumption rates were significantly decreased, and mitochondria were significantly larger and more branched in adrenergic-deficient hearts. Notably, however, the mitochondria were intact with well-formed cristae, and there was no significant difference observed in mitochondrial membrane potential. Maternal administration of the adrenergic receptor agonists isoproterenol or l-phenylephrine significantly ameliorated the decreases in ATP observed in Dbh^−/− embryos, suggesting that α- and β-adrenergic receptors were effective modulators of ATP concentrations in mouse embryos in vivo. These data demonstrate that adrenergic hormones stimulate cardiac energy metabolism during a critical period of embryonic development.

the adrenergic hormones epinephrine (EPI) and norepinephrine (NE) are key mediators of stress responses and sympathetic nervous system activities in adult mammals. NE in particular is also essential for embryonic development. Targeted disruption of the gene dopamine β-hydroxylase (Dbh), which codes for the enzyme that converts dopamine into NE, led to a loss of NE and EPI and embryonic lethality due to heart failure in mice (54). In contrast, disruption of the subsequent enzymatic step catalyzed by phenylethanolamine n-methyltransferase (Pnmt) led to the loss of EPI without concomitant developmental phenotypes (21). Thus, although EPI may contribute to adrenergic activity in the embryo, NE is clearly of critical importance for heart development.

How NE influences heart development in utero is not fully understood. In Dbh^−/− embryos, signs of cardiac distress begin to appear on embryonic day (E)10.5, and ∼50% of the Dbh^−/− embryos die by E11.5 (54). Remarkably, the heart appears to develop and function normally up to this point but then deteriorates rapidly into heart failure within 24 h of the first signs, which include sluggish cardiac contractions, arrhythmia, and asynchrony, as observed via echocardiography in utero (43). Recent work has shown that Dbh^−/− hearts display significantly delayed conduction speeds across the atrioventricular junction relative to age-matched littermate controls (1). These results suggest important adrenergic influences on the development of cardiac structure and function but do not fully explain how NE affects them.

At these early embryonic stages of development, there is no sympathetic innervation of the heart, and the adrenal glands have not yet formed. Instead, NE is produced in the embryonic heart itself as well as outside of the heart in primordial sympathetic ganglia and brainstem neurons (20–22). With respect to heart function, NE appears to be acting primarily through β-adrenergic receptors since Dbh^−/− embryos can be rescued by providing the β-agonist isoproterenol in the maternal drinking water, whereas the α-agonist l-phenylephrine is partially effective at rescuing the heart failure and lethality (55). These results established that NE stimulates cardiovascular function during early embryonic development, but what are the important physiological targets regulated by adrenergic stimulation in the embryonic heart?

To gain insight about adrenergic actions in cardiac development, we recently performed a genome-wide expression screen of Dbh^−/− and Dbh^+/+ embryonic hearts. A key finding from this screen demonstrated that the largest category of differentially expressed genes (∼31% of total) were those involved in metabolism (43). In adult mammals, adrenergic hormones are known to have profound and widespread influences on metabolism. In the liver, for example, β-adrenergic stimulation inhibits glycolysis, promotes gluconeogenesis, and stimulates breakdown of glycogen (6, 50). In cardiac and skeletal muscle, glycolysis and oxidative phosphorylation (OXPHOS) are increased in response to β-adrenergic stimulation to produce more available energy during stress. Free fatty acids are released from adipose tissue, and these serve as the primary fuel source for cardiac metabolism in adult mammals, but prior to birth the heart principally uses carbohydrates (8, 10, 17, 23, 26, 42, 48). The shift from carbohydrate to lipid metabolism in the heart occurs mainly at birth and is often referred to as the “fetal shift” in cardiac metabolism (19, 36, 37). Surprisingly little, however, is known regarding adrenergic influences on cardiac metabolism during the embryonic period.

There is compelling evidence indicating that aerobic metabolism in the mitochondria becomes increasingly important as the heart transitions from the embryonic to the fetal stages of development. For example, genetic mutations that decrease or disrupt OXPHOS often result in heart failure and embryonic lethality (29, 34, 35). Furthermore, mutations that disrupt mitochondrial structure and function have also interfered with cardiomyocyte differentiation and development (11, 12, 18, 58), and many of these have ultimately succumbed to heart failure and embryonic lethality. Thus, there appears to be an “embryonic shift” from primarily anaerobic to aerobic metabolic mechanisms in the heart during the embryonic-fetal transition period of development (2).

We hypothesized that adrenergic hormones play a critical role in facilitating the metabolic shift toward aerobic oxidative phosphorylation in cardiac mitochondria during embryonic development. Here, we tested this hypothesis by examining metabolic profiles of ATP, ADP, and ATP/ADP ratios in Dbh^−/− and control embryos. In parallel, we also examined other metabolic indices, including cardiac oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) as estimators of aerobic and anaerobic metabolism, respectively. In addition, we performed detailed ultrastructural analysis of mitochondria in adrenergic-deficient and control hearts.

METHODS

Mice.

All procedures and handling of mice were conducted in accordance with and approved by the University of Central Florida Institutional Animal Care and Use Committees. The Dbh mouse strain was kindly provided by Dr. Richard Palmiter (University of Washington, Seattle, WA) (54) and maintained as described previously (1, 43). Timed pregnancies were determined by the presence of a vaginal plug (denoted as E0.5) and further confirmed by high-resolution ultrasound (Vevo 2100 instrument with 40 MHz transducer; Visualsonics) at E8.5. For the dosing of maternal mice, drinking water was supplemented with isoproterenol (ISO; 0.02 mg/ml), l-phenylephrine (0.02 mg/ml), or timolol (1 mg/ml) beginning at E8.5, as described previously (49, 55). Vitamin C was added to the control and drug-containing drinking water bottles at a concentration of 2 mg/ml to help minimize oxidation of the drugs (55).

Embryonic tissue collections.

All embryos used in this study appeared healthy and viable at the time of isolation, as judged by their size (crown-rump lengths), color, texture, morphology, and overall appearance. Microscopic examination confirmed a beating heart and bright red blood coursing through the embryonic circulation. Unhealthy and dead embryos were discarded. From the living specimens collected, no apparent differences were observed between adrenergic-deficient and control embryos based on these gross examinations at the time of isolation. Upon isolation, the heads were removed and used for genotyping. For RNA and biochemical assays, hearts or trunks were flash-frozen in liquid nitrogen and stored at −80°C.

Reagents.

Chemical reagents were purchased from Sigma-Aldrich (St. Louis, MO), except where noted otherwise. Electron microscopy-grade reagents for transmission electron microscopy (TEM) were purchased from Electron Microscopy Sciences (Hatfield, PA). Cell culture reagents were purchased from Invitrogen (Carlsbad, CA), except for fetal bovine serum, which was obtained from Hyclone Laboratories (Logan, UT).

ATP and ADP measurements.

Briefly, embryonic tissue was homogenized in 6% trichloroacetic acid (TCA) for 1 min and then centrifuged at 6,000 g for 5 min at 4°C. The supernatant was then removed, and TCA was neutralized with tris-acetate, as described previously (16). ATP measurements were performed with ATPlite Bioluminescence Assay (Perkin-Elmer) as instructed by the manufacturer's protocol. ATP/ADP measurements were performed using ApoSENSOR ADP/ATP Ratio Bioluminescence Assay Kit (BioVision). Standard curves were generated with known concentrations of ATP and ADP. Luminescence was detected in an Envision Multilabel plate reader (Perkin-Elmer), and ATP measurements were normalized to total protein concentrations.

Lactate measurements.

Embryos were homogenized in 8% perchloric acid for 1 min and then centrifuged at 6,000 g for 4 min. Absorbance readings at 340 nm before and after the addition of l-lactate dehydrogenase were performed as described (5). The supernatants for lactate measurements were combined with nicotinamide adenine dinucleotide (NAD) solution (2.5 M NAD, 0.2 M glycine buffer, and 100 μl of ≥500 U/mg protein l-lactate dehydrogenase from bovine heart) in a 96-well plate. Increase in absorbance at 340 nm was compared with a standard curve of known lactate concentrations. Lactate measurements were done in triplicate and normalized to protein.

Glucose measurements.

Flash-frozen embryos were homogenized for 1 min and then centrifuged at 12,000 g for 1 min. The supernatant was removed, and 2 μl was used to measure glucose concentrations using a WaveSense Presto blood glucose-monitoring system (AgaMatrix) (15). Samples were compared with a standard curve of known d-glucose concentrations and normalized to total protein concentrations. Measurements were done in triplicate.

Glycogen measurements.

Flash-frozen embryos were homogenized for 30 s in 1 M potassium hydroxide solution and then incubated at 37°C for 1 h. Samples were then boiled for 5 min and centrifuged at 13,000 g for 5 min at room temperature (27). The supernatant was removed, and glycogen was measured using a Glycogen Assay Kit (Sigma-Aldrich) as per the manufacturer's protocol. Background glucose levels were subtracted from glycogen measurements, and samples were normalized to total protein concentrations. Measurements were performed in duplicate.

OCR and ECAR measurements.

E10.5 and E11.5 mouse hearts were isolated under aseptic conditions and cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Hyclone Laboratories) that had been charcoal-stripped to remove catecholamine and steroid hormones (41). The medium was additionally supplemented with penicillin G (100,000 U/l) and streptomycin (100 mg/l). Hearts were cultured in a Seahorse Biosciences XF24 Islet Capture Microplate with mesh grids placed on top of the specimen to prevent it from floating and from probe interference. Basal OCR and ECAR were measured simultaneously at 10-min intervals over a period of 2 h using a Seahorse XF^e Biosciences system. Rotenone (5 μM) and antimycin A (20 μM) were added simultaneously (61) to block mitochondrial electron transport.

Gene expression.

RNA was isolated from flash-frozen embryonic hearts using TRIzol reagent and converted to cDNA using a High Capacity cDNA Reverse Transcription Kit (Invitrogen). Real-time PCR was performed using SYBR Green Fast reagent in an AB7500 machine (Applied Biosystems). Genes of interest were normalized to the housekeeping gene β-actin. Forward and reverse primers were as follows: Pgc-1α, 5′-TATGGAGTGACATAGAGTGTGCT-3′ and 5′-CCACTTCAATCCACCCAGAAAG-3′, Primer Bank ID 6679433a1; Tfam, 5′-GAGCGTGCTAAAAGCACTGG-3′ and 5′-CCACAGGGCTGCAATTTTCC-3′ (33); Sirt1, 5′-TGTGAAGTTACTGCAGGAGTGTAAA-3′ and 5′-GCATAGATACCGTCTCTTGATCTGA-3′(33); β-actin, 5′-CATCACTATTGGCAACGAGC-3′ and 5′-ACGCAGCTCAGTAACAGTCC-3′ (24).

mtDNA quantification.

mtDNA content was measured as described previously (45). Briefly, total DNA was extracted using TRIzol from flash-frozen embryonic hearts. Quantitative PCR of DNA was performed using SYBR Green Fast reagent in an AB7500 machine (Applied Biosystems). The nuclear gene hexokinase 2, with forward and reverse primers 5′-GCCAGCCTCTCCTGATGT-3′ and 5′-GGGAACACAAAAGACCTCTTCTGG-3′, was used to compare the mitochondrial-encoded 16s ribosomal RNA gene with forward and reverse primers 5′-CCGAAGGGAAAGATGAAAGA-3′ and 5′-TCGTTTGGTTTCGGGGTTTC-3′, respectively. Measurements were performed in duplicate.

Transmission electron microscopy.

Briefly, isolated embryonic hearts were placed in Karnovsky's fixative (2% paraformaldehyde, 2.5% glutaraldehyde, and 0.1 M sodium cacodylate, pH 7.3) for 1 h. Samples were then rinsed in 0.1 M sodium cacodylate with 3 μM CaCl₂ (pH 7.3) and postfixed in 1% osmium tetroxide buffered in sodium cacodylate (pH 7.3), 0.8% potassium ferrocyanide, and 3 μM CaCl₂ solution. Samples were stained en bloc with 2% uranyl acetate and dehydrated in graded ethanol solutions. Samples were embedded in Durcupan, and sections were cut at 80 nm using a Leica UTC Ultramicrotome and diamond knife and then mounted on copper grids. Thin sections were stained with routine TEM double stain: 4 min in 4% uranyl acetate and then 4 min in Reynolds Lead Citrate. Thin sections were examined using an FEI 268D TEM at 50 Kv, and images were recorded using an AMT XR-60 digital camera. Mitochondrial morphometric and glycogen analyses were performed with ImageJ software (National Institutes of Health).

Oil red O staining.

Oil Red O (5 mg/ml in isopropanol) stock stain was diluted to the working solution (60% stock in distilled water) and used to stain for lipid droplets, as described previously (47). Embryos were immediately placed in 2% fresh paraformaldehyde overnight and moved to 30% sucrose and 0.02% sodium azide at 4°C for storage. Frozen sections were made on a Leica CM1850 cryostat 12 μm thick at −20°C. Briefly, the slides were air-dried, fixed in formalin, and washed with running tap water for 1 min. They were then rinsed with 60% isopropanol and stained with freshly prepared Oil Red O working solution for 15 min. The staining solution was then removed, and the samples were rinsed with 60% isopropanol, rinsed further with distilled water, and mounted using Vectashield mounting medium (Vector Laboratories). Digital micrographs were obtained using a Leica DM2000 microscope, and images were analyzed for lipid droplets using ImageJ software (National Institutes of Health).

Free fatty acid quantifications.

Fatty acids were measured using the Free Fatty Acid Quantification Colorimetric/Fluorometric Kit (Biovision) according to the manufacturer's instructions. Briefly, flash-frozen embryos were homogenized for 30 s and then centrifuged at 12,000 g for 1 min. One-hundred microliters of 1% Triton X-100 in chloroform was added to the supernatant. Samples were centrifuged at 12,000 g for 10 min, and the lower phase was transferred and allowed to air-dry; pellets were then resuspended in fatty acid buffer. Fluorescence readings were performed using an Envision Multilabel plate reader (Perkin-Elmer) (excitation: 535 nm; emission: 590 nm). Known concentrations of palmitic acid were used to generate a standard curve for this assay, and all samples were measured in duplicate.

JC-1 dye.

Myocytes were isolated and cultured on coverglass (for microscopy) or 48-well plates (for flow cytometry), as described previously (38). Cardiomyocytes were cultured for 48 h before staining with JC-1 dye (5 μg/ml; Invitrogen) for 20 min. The dye was washed with PBS, and samples were either viewed using a Perkin-Elmer Spinning Disk confocal microscope or quantified using a BDFacs Canto flow cytometer (BD Biosciences). Gates were set with unstained cells and FCCP-treated controls (50 μM) and quantified emission filters appropriate for Alexa Fluor 488 nm and R-phycoerythrin. Data were analyzed using FCSExpress software (DeNovo).

Statistics.

Data are expressed as means ± SE. Student's t-tests were performed to compare means between adrenergic-competent and adrenergic-deficient groups, with P < 0.05 required to reject the null hypothesis. One-way analysis of variance (ANOVA) was performed for multiple comparisions, with Bonferroni post hoc testing for comparison between individual groups.

RESULTS

ATP is depleted in adrenergic-deficient embryos.

To determine whether adrenergic deficiency affects embryonic metabolism, we measured ATP and ADP concentrations in adrenergic-competent (Dbh^+/− and Dbh^+/+) and -deficient (Dbh^−/−) embryos.

Throughout this and previous studies, we observed no significant difference in Dbh^+/+ and Dbh^+/− embryos for any assays employed, and these genotypes are phenotypically indistinguishable. Thus we have combined Dbh^+/+ and Dbh^+/− mice into a single group that we will hereafter refer to as “adrenergic competent”. Dbh^−/− mice fail to produce NE or EPI, and most will succumb to heart failure and embryonic lethality between E10.5 and E15.5 (54) unless rescued by maternal supplementation of alternative catecholamine substrates (54) or β-adrenergic agonists such as ISO (55).

ATP concentrations were significantly decreased in adrenergic-deficient embryos as early as E10.5 and declined precipitously by E11.5 (Fig. 1A). Conversely, ADP concentrations increased over the same time period in the Dbh^−/− group relative to controls (Fig. 1B). Notably, both ATP and ADP concentrations were unchanged at E9.5 relative to adrenergic-competent controls (Fig. 1, A and B). Steady-state ATP, ADP, and ATP/ADP values, generated from known ATP and ADP standard curve concentrations, are shown in Table 1 for adrenergic-competent and deficient embryos.

Fig. 1.

ATP, ADP, and ATP/ADP measurements in adrenergic-deficient embryos compared with controls. A–C: ATP (A), ADP (B), and ATP/ADP (C) ratio measurements on embryonic days (E)9.5, E10.5, and E11.5 in adrenergic-competent (black bars) and adrenergic-deficient littermates (open bars). D: ATP measurements on E10.5 and E11.5 in adrenergic-deficient embryos with and without isoproterenol (ISO; 0.02 mg/ml in maternal drinking water). E: ATP measurements on E11.5 in adrenergic-competent embryos with (n = 16) and without (n = 14) timolol (1 mg/ml in maternal drinking water). F: ATP measurements on E11.5 in adrenergic-competent (black bars) and adrenergic-deficient embryos (open bars) with and without l-phenylephrine (0.02 mg/ml in maternal drinking water). Numerical values within the bars refer to the no. (n) of samples analyzed. For this experiment, 1-way ANOVA was performed with Bonferroni's multiple comparison test used to evaluate differences between individual groups. All other comparisons in this figure (A–E) were analyzed using Student's t-test. Data are shown as fold change compared with littermate controls. Fold change calculations were generated from steady-state ATP and ADP concentrations (see also Table 1). *P < 0.05; **P < 0.005; ***P < 0.001.

Table 1.

ATP, ADP, and ATP/ADP ratio values of adrenergic-deficient and control embryos

	[ATP], nmol/mg protein	[ADP], nmol/mg protein	ATP/ADP Ratio (AU)	No. of Samples
E9.5
Competent	40.6 ± 10.7	19.8 ± 6.6	3.9 ± 1.2	12
Deficient	32.4 ± 8.3	10.8 ± 2.9	3.8 ± 1.2	6
E10.5
Competent	26.8 ± 3.1	9.1 ± 1.9	5.1 ± 1.0	13
Deficient	19.5 ± 7.33	9.9 ± 4.1	2.6 ± 0.8	5
E11.5
Competent	7.6 ± 0.5	0.8 ± 0.06	9.4 ± 1.3	8
Deficient	1.0 ± 0.4***	4.3 ± 1.3**	0.2 ± 0.03***	4

Data are represented as means ± SE.

AU, arbitrary units; E9.5, E10.5, and E11.5, embryonic days 9.5, 10.5, and 11.5, respectively.

^**P < 0.01;

^***P < 0.001.

ATP/ADP ratios were virtually identical in adrenergic-competent and -deficient embryos at E9.5 but then diverged dramatically over the next 2 days. In control embryos, the ATP/ADP ratio increased steadily over this time period, but in Dbh^−/− embryos, this ratio dropped by 50% at E10.5 and by >95% (∼48-fold reduction) at E11.5 (Fig. 1C). It is important to note that all embryos collected for these analyses appeared healthy and viable, as described in methods. Despite their outward appearance, however, Dbh^−/− embryos exhibited an ATP/ADP energy deficit beginning around E10.5 that rapidly became much more severe by E11.5.

To determine whether the observed energy depletion was due to the absence of β-adrenergic stimulation, we provided ISO in the maternal drinking water to rescue the phenotype. When the experiment was repeated under these conditions, the ATP deficits disappeared (Fig. 1D). We observed a 1.8-fold increase in ATP concentrations at E10.5 and a 4.5-fold increase in E11.5 Dbh^−/− ATP concentrations after ISO treatment compared with those observed without the addition of ISO. These results demonstrate that β-adrenergic stimulation was effective at preventing the energy loss resulting from adrenergic deficiency in developing mouse embryos.

Conversely, we applied the nonselective β-adrenergic receptor antagonist timolol (1 mg/ml) in the maternal drinking water to determine whether blockade of β-receptors would influence ATP concentrations in control (adrenergic-competent) embryos. E11.5 embryos from dams that received timolol contained significantly less (∼22% reduction, P < 0.01; n = 16 for no drug controls, n = 14 for timolol-treated group) ATP compared with age-matched controls that did not receive timolol (Fig. 1E). Taken together, these results suggest that adrenergic stimulation of β-receptors is important for maintaining embryonic ATP concentrations.

However, we noticed that although timolol was clearly effective at lowering ATP levels relative to controls during this period of development (Fig. 1E), there was a greater decrease in ATP concentrations in the Dbh^−/− age-matched embryos (compare Fig. 1, A and E, for E11.5). There are a number of possible explanations for this finding, including a potential role for α-adrenergic receptors in regulating embryonic energy metabolism. To test this, we applied the α-adrenergic receptor agonist l-phenylephrine in the maternal drinking water beginning on E8.5 and again collected embryos on E11.5 to assess ATP concentrations. Remarkably, adrenergic-deficient embryos obtained from dams that received l-phenylephrine had significantly higher concentrations of ATP than adrenergic-deficient embryos from dams that did not receive any drug treatment, as assessed by one-way ANOVA (Fig. 1F). The recovery was not complete, but nevertheless achieved 50–60% of the ATP levels observed in controls. These results strongly suggest that embryonic ATP concentrations are positively regulated through α- as well as β-adrenergic receptor stimulation.

Influence of adrenergic hormones on embryonic carbohydrate and lipid metabolism.

Carbohydrate metabolism is the principal source of energy for the developing heart prior to birth, although lipid utilization by the heart also begins during the embryonic period as the heart gains oxidative metabolic capability (3, 17, 23, 48, 59). To determine whether key carbohydrate metabolites were altered in adrenergic-deficient embryos, we measured glucose, glycogen, and lactate concentrations in embryos isolated on E10.5 and E11.5 (Table 2). No significant differences were observed in the concentrations of these carbohydrate metabolites at these ages. However, lactate concentrations appeared slightly elevated on average in adrenergic-deficient embryos on both E10.5 and E11.5, but these differences were also not statistically significant. Glycogen concentrations trended lower in the deficient group, but the results were variable and again not found to be significantly different. The biochemical results for glycogen were corroborated by image analysis of glycogen granules from TEM micrographs, which showed a similar, insignificant downward trend in glycogen granules on E11.5 for the adrenergic-deficient group (Table 2). We also measured free fatty acid concentrations and lipid droplets but found no significant differences in these either, although there was a downward trend in the deficient group for lipid droplets, which were located predominantly in the liver at these stages of development. Despite these trends, steady-state levels of free fatty acids as well as glucose, glycogen, and lactate were relatively unchanged (no significant differences) in adrenergic-deficient embryos compared with age-matched littermate controls on E10.5 and E11.5.

Table 2.

Carbohydrate and lipid metabolite concentrations

	E10.5			E11.5
	Competent	Deficient	P value	Competent	Deficient	P value
Glucose, nmol/mg protein	313 ± 17 (n = 10)	260 ± 26 (n = 9)	0.09	436 ± 27 (n = 8)	421 ± 44 (n = 7)	0.77
Glycogen, nmol/mg protein	5.8 ± 1.4 (n = 4)	7.5 ± 1.5 (n = 4)	0.45	9.9 ± 6.5 (n = 4)	4.3 ± 2.2 (n = 3)	0.51
Glycogen granules, %/mm2	5.0 ± 0.6 (n = 32)	6.2 ± 0.9 (n = 42)	0.30	4.8 ± 1.0 (n = 18)	3.3 ± 0.7 (n = 23)	0.20
Lactate, nmol/mg protein	250 ± 50 (n = 6)	422 ± 107 (n = 6)	0.17	293 ± 37 (n = 6)	416 ± 78 (n = 6)	0.18
Lipid (liver) droplets/mm2	5,015 ± 677 (n = 10)	3,528 ± 416 (n = 14)	0.06	5,866 ± 598 (n = 16)	3,988 ± 869 (n = 16)	0.09
Free fatty acids, nmol/mg protein	1.6 ± 0.2 (n = 10)	1.5 ± 0.2 (n = 10)	0.86	1.5 ± 0.1 (n = 8)	1.5 ± 0.3 (n = 7)	0.92

Data are presented as means ± SE.

When we examined the rate of glycolysis, however, significant differences were observed. To perform these measurements, we isolated and cultured whole beating hearts from E10.5 to E11.5 Dbh^−/− and littermate controls and recorded the ECAR at various intervals over a 40-min period. On E10.5, there was little difference in ECAR from adrenergic-deficient and control samples over the entire 40 min (Fig. 2A). On E11.5, however, adrenergic-deficient and control ECARs were similar initially but began to decline in the adrenergic-deficient group after about 10 min and continued to decline further over the next 30 min such that it was less than one-half the initial rate by 40 min (Fig. 2B). This decline was prevented by the addition of ISO in the maternal drinking water (Fig. 2C). These results suggest that glycolytic rate was compromised in adrenergic-deficient hearts by E11.5.

Fig. 2.

Effects of adrenergic deficiency on extracellular acidification rate (ECAR). A and B: ECAR in E10.5 and E11.5 adrenergic-competent (●) and -deficient (□) isolated beating hearts; values are represented as %E10.5 controls. For E10.5 adrenergic-competent hearts, n = 15; adrenergic-deficient hearts, n = 9. For E11.5 adrenergic-competent hearts, n = 28; adrenergic-deficient hearts, n = 8. C: ECAR fold change at 40 min between adrenergic-competent (black bars) and -deficient (open bars) E11.5 isolated hearts with and without ISO in maternal drinking water. Means ± SE at the final (40 min) time point are shown. *P < 0.05.

OCR decreased due to the absence of β-adrenergic stimulation.

Mitochondria play an increasingly important role in the heart at these early embryonic stages of development. OCR measurements from isolated embryonic hearts can be used to estimate mitochondrial respiration. Prior to OCR measurements, beating rates were taken in isolated embryonic hearts to verify the viability of tissue collected. Similar beating rates were observed in the adrenergic-competent and -deficient samples at both ages (Table 3). To confirm that the OCR was from mitochondrial function, we administered rotenone (5 μM) and antimycin A (20 μM) to control hearts to block complexes I and III, respectively. This treatment significantly decreased the overall OCR by 50–60% (Fig. 3, A and B), indicating that at least half of the observed OCR in E10.5–E11.5 mouse hearts was dependent on mitochondrial respiration. The remaining portion likely reflects other oxidation reactions within the cells, as has been commonly observed in other systems (9, 28, 60).

Table 3.

Beating rates of isolated embryonic hearts

	Competent, beats/min	Deficient, beats/min	P Value
E10.5	88 ± 11 (n = 10)	99 ± 21 (n = 4)	0.62
E11.5	51 ± 4 (n = 39)	47 ± 8 (n = 8)	0.59

Data are presented as means ± SE.

Fig. 3.

Effects of adrenergic deficiency on oxygen consumption rate (OCR). A: OCR in E10.5 (●) and E11.5 (■) adrenergic-competent hearts treated with rotenone and antimycin A (dashed line). B: OCR in E10.5 (n = 8) and E11.5 (n = 10) adrenergic-competent hearts before (black bars) and after (gray bars) rotenone and antimycin A treatment. In both A and B, values are represented as %E10.5 basal OCR. C and D: OCR in E10.5 and E11.5 adrenergic-competent (●) and -deficient (□) hearts; values are represented as %E10.5 controls. For E10.5 adrenergic-competent hearts, n = 15; adrenergic-deficient hearts, n = 9. For E11.5 adrenergic-competent hearts, n = 20; adrenergic-deficient hearts, n = 9. E and F: OCR fold change at 15 min between adrenergic-competent (black bars) and -deficient (open bars) E10.5 and E11.5 isolated hearts with and without ISO. Data are represented as means ± SE. *P < 0.05; ***P < 0.001.

We found that adrenergic-deficient hearts had lower rates of oxygen consumption on average compared with littermate controls (Fig. 3, C and D). Nevertheless, OCR increased similarly in adrenergic-deficient and -competent hearts between E10.5 and E11.5 (compare Fig. 3, C and D), but the adrenergic-deficient OCR continued to lag significantly below the control group through E11.5. As indicated above, the hearts appeared to be similar and were beating spontaneously at comparably slow but steady rates in both adrenergic-deficient and -competent hearts during the ex vivo culture period (Table 3). Thus, despite the lack of any clear differences in outward appearance or behavior, adrenergic-deficient hearts consumed oxygen at significantly lower rates than control hearts. To verify that this effect was due to the absence of adrenergic hormones, we provided ISO in the maternal drinking water, which restored OCR to control levels in adrenergic-deficient hearts (Fig. 3, E and F). However, acute administration of ISO to the isolated heart in culture did not result in any significant changes in OCR over a 1-h recording period (not shown). These results suggest that longer-term treatment with ISO in vivo is needed to effectively prevent the decline in OCR observed in adrenergic-deficient hearts.

Mitochondrial biogenesis not affected by the loss of adrenergic hormones.

In theory, the lowered OCR and ATP/ADP ratio could be due to fewer mitochondria resulting from compromised mitochondrial biogenesis in adrenergic-deficient hearts. To test this hypothesis, we measured the expression of key mitochondrial biogenesis genes in isolated hearts, but no significant alterations in mRNA for Pgc-1α (P = 0.58), Tfam (P = 0.47), or Sirt1 (P = 0.65) were found in E10.5 isolated hearts when normalized to the housekeeping gene β-actin (Fig. 4A). Similar results were found on E11.5. We also measured mtDNA content present in adrenergic-deficient and control hearts, but no significant differences were observed (Fig. 4B). There was an apparent increase in mtDNA on E10.5, which could be related to the similar upward trend seen in Pgc-1α from this group (compare Fig. 4, A and B), but neither of these differences were significant, and both trends had disappeared by E11.5. These results suggest that mitochondrial biogenesis is likely not a limiting factor in adrenergic-deficient embryos at these early stages.

Fig. 4.

Mitochondrial biogenesis gene expression and mtDNA content in adrenergic-deficient vs. control hearts. A: fold change of key mitochondrial biogenesis genes between adrenergic-competent and -deficient mRNA from E10.5 (n = 5) and E11.5 (n = 6) hearts, as measured by quantitative RT-PCR analysis. B: mtDNA content fold change between E10.5 and E11.5 adrenergic-competent (black bars) and -deficient (open bars) hearts. Numerical values within the bars refer to the no. (n) of samples analyzed.

Adrenergic-deficient myocytes have intact mitochondrial membranes.

The decreased OCR in adrenergic-deficient hearts suggests that mitochondrial function may be impaired. To properly function, mitochondria must maintain membrane potentials sufficient to drive the proton gradients necessary for OXPHOS in their inner membrane space. The fluorescent dye JC-1 can distinguish between mitochondria with intact membranes and those with compromised membrane potentials via differential fluorescence emissions (46). For example, red fluorescence (∼590 nm) detects aggregates and is indicative of healthy intact membranes, whereas green fluorescence (∼529 nm) detects monomers associated with perturbed mitochondrial potentials. We employed JC-1 staining with flow cytometry in combination with laser-scanning confocal fluorescence microscopy to assess mitochondrial membrane potential integrity in cardiomyocytes isolated from E10.5 and E11.5 adrenergic-competent and -deficient hearts. Flow cytometry for red (aggregates) and green (monomers) showed no difference between adrenergic-deficient and control samples at E10.5 (Fig. 5A) or E11.5 (Fig. 5C). Quantification of the stained primary cardiomyocyte culture showed a red-to-green ratio of 3.4 ± 0.7 in the E10.5 deficient samples (n = 9) and 3.1 ± 0.6 in controls (n = 9) (Fig. 5B). Similar results were observed in the E11.5 primary cardiomyocyte cultures, with a red-to-green ratio of 3.4 ± 1.3 in the deficient group (n = 4) and 2.5 ± 0.8 in the control group (n = 4) (Fig. 5D). Representative mitochondrial staining with JC-1 is shown in Fig. 5E for adrenergic-competent and -deficient myocytes isolated from E11.5 hearts. Despite the fact that there were no significant differences in the ratio of red/green staining in these specimens, the mitochondrial staining pattern in the adrenergic-deficient group appeared to be less densely clustered and less well organized within myocytes than those typically found in age-matched adrenergic-competent littermates, indicating that there may be structural abnormalities in mitochondria from adrenergic-deficient embryos. At this resolution, however, it was difficult to determine whether there were truly structural anomalies in the mitochondria, so we employed TEM for these evaluations, as described below.

Fig. 5.

Analysis of mitochondrial membrane potentials in adrenergic-deficient and control myocytes using flow cytometry and fluorescence microscopy with JC-1 dye. A and C: representative scatter plots of flow cytometry with red and green fluorescence from JC-1 dye in adrenergic-deficient and adrenergic-competent E10.5 and E11.5 embryonic primary cardiomyocytes. Adrenergic-competent samples (red) and adrenergic-deficient samples (blue) in histogram overlay. B and D: ratio of red/green fluorescence in E10.5 and E11.5 adrenergic-competent (open bars) and -deficient (black bars) samples. Numerical values within the bars refer to the no. (n) of samples analyzed. E: representative scanning laser confocal microscopy of JC-1 dye in E11.5 primary cardiomyocytes. Scale bar, 10 μm.

Mitochondrial morphology altered in adrenergic-deficient hearts.

To obtain a detailed view and assessment of mitochondrial structure in adrenergic-deficient and control embryonic hearts, we analyzed cardiac tissue specimens using TEM. Our results show that mitochondria within E10.5 adrenergic-deficient myocardium were enlarged and appeared swollen relative to adrenergic-competent controls (Fig. 6, A and B). High-magnification images show abnormal mitochondrial shapes in the adrenergic-deficient hearts (Fig. 6C). Drawn-to-scale tracings of mitochondrial shapes from control and adrenergic-deficient hearts are shown for ease of comparison in Fig. 6D on E10.5. Similar results were observed in E11.5 adrenergic-deficient hearts, demonstrating swollen and elongated mitochondria compared with those in adrenergic-competent samples (Fig. 6, E–H). Multiple mitochondria had branches and bulges in adrenergic-deficient hearts at E11.5 (Fig. 6H). Adrenergic-deficient myocyte samples had fewer mitochondria per micrograph compared with adrenergic-competent hearts on E11.5, a trend that was evident on E10.5, although the numbers were more similar to control values at that age (Table 4). Mitochondrial length was significantly increased on E10.5 (P < 0.05) and on E11.5 (P < 0.0001) by nearly 30% in the adrenergic-deficient samples vs. controls (Table 4). Mitochondrial surface area was also significantly increased on E10.5 (P < 0.05) and on E11.5 (P < 0.001) in adrenergic-deficient myocytes compared with controls (Table 4). These results clearly show that mitochondrial structure was significantly altered in adrenergic-deficient embryos. On average, the mitochondria were longer, had greater surface area, and displayed more abnormal shapes such as branch points, curvatures, and budding bulges in their membranes compared with their control counterparts.

Fig. 6.

Ultrastructural analysis of mitochondria in adrenergic-deficient and control hearts following evaluation of transmission electron microscopy images. Mitochondrial morphology in adrenergic-competent and -deficient E10.5 (compare A and B) and E11.5 (compare E and F) myocytes. C and G: branched (arrowheads) and swollen mitochondria in adrenergic-deficient samples. D and H: representative drawn-to-scale tracings of abnormally shaped mitochondria in adrenergic-deficient samples relative to adrenergic-competent control tracings at each age. m, Mitochondria; n, nucleus. Scale bar, 500 nm.

Table 4.

Quantification of mitochondrial nos. and morphology from TEM

	E10.5 Adrenergic Competent	E10.5 Adrenergic Deficient	E11.5 Adrenergic Competent	E11.5 Adrenergic Deficient
No./17.5 μm2	11.7 ± 0.5 (n = 58)	10.7 ± 0.4 (n = 103)	13.9 ± 1.1 (n = 38)	11.0 ± 0.7 (n = 38)*
Length, nm	718.8 ± 13.8 (n = 669)	772.4 ± 14.4 (n = 1,106)	637.5 ± 17.4 (n = 529)	888.5 ± 27.9 (n = 418)***
Surface area, μm2	257.5 ± 11.7 (n = 153)	304.7 ± 19.0 (n = 158)*	204.8 ± 10.6 (n = 166)	362.8 ± 28.7 (n = 127)***

Data are expressed as means ± SE.

TEM, transmission electron microscopy.

^*P < 0.05;

^***P < 0.001.

DISCUSSION

Anaerobic glycolysis is the predominant mode of metabolism of the heart during early embryonic development, but aerobic mitochondrial metabolism becomes an increasingly important and essential mode of ATP production at the late embryonic to early fetal stages of development (7, 17, 23, 48). For example, classic studies showed that isolated embryonic rat hearts utilized glycolytic mechanisms through E11; however, glycolysis was not sufficient to maintain maximal heart rates on E12 or E13 (17). Furthermore, energy metabolism was unaffected by the presence of oxygen at E11 in the rat but was significantly and progressively elevated by oxygen in E12 and E13 hearts. This embryonic shift in metabolic capability between E11 and E12 in the rat corresponds roughly to the end of the organogenesis period of embryonic development and the beginning of fetal development. In the mouse, the equivalent stages of development are approximately E9.5–E10.5 (53). It is during this late embryonic period when metabolic deficiencies first become apparent in adrenergic-deficient (Dbh^−/−) mouse embryos.

On E9.5, there were no discernible differences in ATP or ADP concentrations in adrenergic-deficient and control embryos, but ATP/ADP ratios began to decline on E10.5 and much more dramatically so on E11.5 in adrenergic-deficient embryos. This was mediated in part through β-adrenergic receptors since administration of the β-agonist isoproterenol was able to “rescue” the loss of ATP observed in Dbh^−/− embryos. Furthermore, administration of the β-antagonist timolol induced significant decreases in ATP concentrations in adrenergic-competent control embryos. These results support a role for β-adrenergic receptor stimulation in regulation of ATP.

There also appears to be a role for α-adrenergic receptor involvement since the α-agonist phenylephrine was also able to “rescue” ATP concentrations in adrenergic-deficient embryos. In both rescue experiments, ATP depletion was partially rescued by either α- or β-agonists, suggesting that both pathways likely contribute to adrenergic influence of ATP regulation in the developing embryo. It should be noted, however, that further work is needed to fully elucidate the molecular signaling pathways leading from receptor stimulation to regulation of ATP homeostasis. One must be cautious interpreting pharmacological data since partial effects could be attributed to a number of different factors, including drug dosage, secondary or other indirect effects, drug distribution/penetrance in the embryo due to placental barrier, and other factors that may not be readily apparent. These caveats notwithstanding, the results of our pharmacological experiments are consistent with those observed in the genetic knockout experiments described in this study. In both cases, the results showed that adrenergic stimulation is needed to maintain embryonic ATP concentrations during the critical early stages of development.

ATP is generated by glycolysis and mitochondrial oxidative phosphorylation. Our results suggest that the decline in ATP/ADP cannot be fully ascribed to glycolysis since no significant differences were observed in key glycolytic metabolites, including glycogen, glucose, and lactate on E10.5 or E11.5. In addition, ECAR was virtually identical on E10.5 in adrenergic-deficient and control embryos. By E11.5, however, ECAR was selectively decreased in the adrenergic-deficient group, indicating that glycolytic rate may have been compromised. The decline was not apparent immediately but developed over 15–30 min and was prevented completely by supplying the β-agonist isoproterenol in the maternal drinking water. These results suggest that adrenergic stimulation is needed to maintain glycolytic rates after E10.5 in mouse embryos.

As we have shown, OCR was also significantly decreased in adrenergic-deficient hearts on E10.5 and E11.5 relative to littermate controls. These results suggest that aerobic metabolism was likely compromised in these embryos. It appeared that basal OCR in adrenergic-deficient hearts lagged about 1 day behind adrenergic-competent hearts and was only about 50% of control levels on E10.5 and E11.5. These effects were extinguished by supplying isoproterenol in the maternal drinking water, thereby indicating that the absence of β-adrenergic stimulation was responsible for the declining OCRs in Dbh^−/− embryos. These results suggest that oxidative metabolism may have been impaired in adrenergic-deficient embryos.

Since oxidative phosphorylation occurs primarily in mitochondria, we examined mitochondrial ultrastructure directly in the embryonic heart using TEM and found that it was significantly altered in adrenergic-deficient embryos. TEM analyses showed that mitochondria were enlarged and more frequently displayed branching or budding membranes in adrenergic-deficient hearts on E10.5 and E11.5. Despite the abnormal mitochondrial morphology in these hearts, the membranes appeared to be intact with well-formed cristae. Furthermore, no significant alterations in red/green fluorescence ratios were observed following application of the mitochondrial membrane potential-sensitive dye JC-1 in adrenergic-deficient hearts compared with controls. These findings suggest that although the mitochondria were larger and abnormally shaped, the observed structural changes may represent compensatory mechanisms to increase ATP production in energy-starved Dbh^−/− hearts (25, 39). Indeed, enlarged mitochondria have been associated with enhanced metabolic output, whereas smaller fragmented mitochondria have generally been associated with decreased metabolic output (31, 44, 52). In the present study, however, we observed enlarged mitochondria with apparent decreased metabolic output. This phenomenon of elongated mitochondria with diminished function has been demonstrated previously in other models of mitochondrial dysfunction (4, 11, 14, 40).

We tested the hypothesis that Dbh^−/− embryos may be starved for substrate metabolites, but little change was observed in key carbohydrate or lipid substrates. This result was not entirely unexpected since glucose is known to pass freely from maternal blood supply to the embryo (51). These results suggest that metabolic substrates were not depleted in adrenergic-deficient embryos.

The pathophysiology of adult heart failure is commonly characterized by mitochondrial dysfunction, which is indicated by decreased energy levels (30, 56, 57). In the Dbh^−/− embryos, however, signs of energy starvation were seen before any outward precursor of heart failure. Indeed, metabolic deficiencies were observed in Dbh^−/− embryos that were otherwise phenotypically indistinguishable from adrenergic-competent controls at the time of isolation in terms of size, morphology, color, texture, and cardiac beating activity. These results suggest that metabolic defects likely contribute to the subsequent heart failure and embryonic lethality in this model. The depletion of ATP concentrations could be prevented largely by maternal administration of either an α-adrenergic agonist (l-phenylephrine) or a β-adrenergic agonist (ISO), which suggests that both of the major adrenergic signal transduction pathways play important roles to ensure adequate supplies of ATP available for embryonic/fetal growth and development (55).

In summary, we have shown that adrenergic hormones fulfill a critical developmental role by providing the growing embryo with sufficient chemical energy in the form of ATP to enable successful transition from the embryonic to fetal stages. Furthermore, our results demonstrate that adrenergic hormones are necessary to maintain sufficient ATP/ADP, ECAR, and OCR during late periods of embryonic development in preparation for the transition to the fetal period. The discovery of these influential regulatory connections between adrenergic hormones and embryonic energy metabolism opens new paths for the study of cardiovascular development that could give rise to novel therapeutic targets and strategies for treating congenital heart defects (13, 32) as well as adult forms of heart disease.

GRANTS

This work was supported by a grant to S. N. Ebert from the National Heart, Lung, and Blood Institute (R01-HL-78716) and funds from the College of Medicine at the University of Central Florida.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

C.N.B. and S.N.E. conception and design of research; C.N.B., S.A.G., G.F.P., and J.N.P. performed experiments; C.N.B., S.A.G., G.F.P., J.N.P., and S.N.E. analyzed data; C.N.B., J.N.P., and S.N.E. interpreted results of experiments; C.N.B., S.A.G., and J.N.P. prepared figures; C.N.B. drafted manuscript; C.N.B., S.A.G., G.F.P., J.N.P., and S.N.E. edited and revised manuscript; C.N.B., S.A.G., G.F.P., J.N.P., and S.N.E. approved final version of manuscript.