Experimental chronic fetal hyperglucagonemia stimulates hepatic pathways for amino acid catabolism and gluconeogenesis

Abstract

Glucagon activates amino acid catabolism and gluconeogenesis in adults. Elevated glucagon concentrations in the fetus occur in pregnancy complications such as fetal growth restriction and hypoxia, yet the impact of chronic fetal hyperglucagonemia is unknown. Using chronically catheterized pregnant sheep, glucose tracers, and liver tissue biopsies, we investigated the effects of nine days of glucagon infusion at 5 or 50 ng·kg⁻¹·min⁻¹ in late-gestation fetal sheep that increased plasma glucagon concentrations by 800%. Glucagon-infused fetuses were euglycemic and exhibited lower plasma and hepatic amino acid concentrations. They also had increased hepatic mRNA expression of amino acid catabolism genes, including ARG2, GLS2, BCAT1, BCAT2, GLUL, HAL, UROC1, and PPARGC1A. Metabolite profiling in liver tissue revealed enrichment of pathways associated with amino acid degradation, elevated tri- and diphosphate nucleotides, and changes in fatty acid metabolites, supporting enhanced hepatic energy metabolism from amino acid oxidation. Hepatic glycogen content was reduced in glucagon-infused fetuses and the gluconeogenic genes PCK1 and G6PC1 were increased, though fetal glucose production was not detected. These findings demonstrate that in the fetal liver, chronic hyperglucagonemia activates amino acid catabolic pathways, indicating a physiological role for glucagon in regulating fetal amino acid homeostasis. These findings have implications for understanding fetal hepatic adaptations during chronic fetal hyperglucagonemia that can occur in the setting of fetal growth restriction or hypoxia.

INTRODUCTION

Glucagon, a hormone produced by pancreatic α-cells, has established roles in adults, such as stimulating hepatic gluconeogenesis and amino acid (AA) catabolism (1). In the adult liver, glucagon promotes AA catabolism as part of a homeostatic feedback loop in which elevated AA concentrations stimulate glucagon release by the pancreatic α-cells. Elevated glucagon concentrations promote AA catabolism in the liver, lowering AA concentrations back to normal (2). In contrast to the adult, the role of glucagon in the fetus has been less well studied. Elevated glucagon concentrations in the fetus are clinically observed in various complications of pregnancy, including fetal growth restriction (FGR), placental insufficiency, and fetal hypoxia (3, 4). Despite these observations, direct evidence linking elevated glucagon in the fetus with specific physiologic consequences is limited.

Sheep models provide critical insights into the metabolic and developmental roles of glucagon during pregnancy. In a seminal study, Devaskar and colleagues demonstrated that glucagon infused into the fetus at a rate of 5 ng/kg/min did not stimulate fetal glucose production yet did stimulate glucose production in adult sheep at the same dose (5). Fetal glucose production could be stimulated by using a higher glucagon infusion rate of 50 ng/kg/min, demonstrating a relative resistance compared to adult sheep. However, the study by Devaskar, Ganguli (5) focused only on glucose production and did not measure fetal AA concentrations or the impact on fetal hepatic AA catabolic or gluconeogenic pathways.

Our group demonstrated that an acute two hour glucagon infusion into fetuses at a rate of approximately 40 ng/kg/min lowered fetal AA concentrations and induced hepatic mRNA expression of several genes responsible for AA catabolism (6). Arginase-1 (ARG1), histidine ammonia-lyase (HAL), and glycine decarboxylase (GLDC) mRNA were all higher in hyperglucagonemic fetuses. With respect to genes that regulate gluconeogenesis, glucagon increased mRNA expression of PCK2 (phosphoenolpyruvate carboxykinase-2 [PEPCK 2]) and G6PC1 (glucose-6-phosphatase) and decreased hepatic glycogen content. Moreover, mRNA expression of PPARGC1A (peroxisome proliferator-activated receptor gamma coactivator-1 alpha), a coactivator which stimulates gluconeogenic and AA catabolic genes, was increased. These transcriptional responses were replicated in isolated fetal sheep hepatocytes exposed to glucagon for 24 hours in vitro. Further, exogenous glucagon stimulated PCK1 (PEPCK-1), ARG2 (arginase-2), GLS1 (glutaminase-1), GLS2 (glutaminase-2), and BCAT2 (branched-chain amino acid transferase-2) in primary fetal hepatocytes, supporting widespread effects on AA catabolic and gluconeogenic pathways.

In addition to acute experimental fetal hyperglucagonemia, our group has tested longer nine-day glucagon infusions into the fetus at both 5 and 50 ng/kg/min to better understand the impact of sustained elevated glucagon concentrations on fetal metabolism (7, 8). In this cohort, fetal plasma AA concentrations decreased due to inhibition of umbilical (net fetal) AA uptake from the placenta, rather than increased fetal AA utilization (8). Further, glucose supply from the placenta and fetal plasma glucose concentrations are normal, while insulin concentrations were decreased (8). Lower insulin and higher glucagon concentrations create a systemic hormonal milieu predicted to activate gluconeogenesis and shift metabolism from anabolism to catabolism in the fetal liver. In the current study, our objective was to test the hypothesis that fetuses with experimentally elevated glucagon for nine days would stimulate hepatic gluconeogenic and AA catabolic pathways. We also performed glucose kinetic studies to test if glucose production was active.

MATERIALS AND METHODS:

Surgical instrumentation of fetus and ewe and experimental design

All experiments were conducted at the Perinatal Research Center, protocol 00470 approved by the University of Colorado School of Medicine Institutional Animal Care and Use Committee, in accordance with the ARRIVE 2.0 guidelines (9). This Center is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. Studies were conducted in Columbia-Rambouillet sheep with singleton pregnancies. Characteristics of these animals, the surgery performed to place indwelling fetal and maternal vascular catheters, and the details of their 8-10 day continuous experimental vehicle or glucagon infusion have already been reported (7, 8).

In the original cohort we reported results from 10 CON animals (vehicle infusion, 0.5% bovine serum albumin in saline [BSA]), 7 animals that received glucagon diluted in 0.5% BSA at a rate of 5 ng/kg/min (GCG-5), and 5 animals that received glucagon diluted in 0.5% BSA at a rate of 50 ng/kg/min (GCG-50) (8). In the current study, we report results only from animals that had a fetal liver biopsy performed while experimental vehicle or glucagon infusions were running (n=7 CON, n=7 GCG-5, and n=4 GCG-50). Due to catheter failure we only report glucose tracer results from a subset of these animals (n=5 CON, n=5 GCG-5, and n=3 GCG-50). For purposes of summary and comparative statistics we have combined the two GCG groups. For graphical presentations we have used different symbols for GCG-5 and GCG-50 animals to distinguish between the two groups. The glucagon doses were selected in accordance with Devaskar et al. (5) and elevate fetal glucagon concentrations to a similar degree as fetal hyperaminoacidemia (10) and hypercatecholinemia(11).

Fetal Metabolic study and tissue collection

After a baseline set of blood draws on the second to last day of the chronic infusion, fetuses were given a primed continuous infusion of [U-¹⁴C]glucose (70 μCi bolus followed by 70 μCi/hr; Cambridge Isotope Laboratories) via a fetal inferior vena cava catheter to measure fetal glucose metabolism (12). After 90 minutes, a primed continuous infusion of ³H₂O (30 μCi followed by 30 μCi/hr; PerkinElmer Life Sciences) was added via a fetal inferior vena cava catheter to measure umbilical blood flow via the steady state transplacental diffusion technique (8, 13). Simultaneously, a primed continuous infusion of L-[1-¹³C]leucine was provided to measure fetal leucine metabolism. Results of the L-[1-¹³C]leucine study are detailed in a separate publication and will not be reported in this manuscript (8). After isotopic steady state was achieved (180 minutes for leucine tracer, 90 minutes for tritiated water and glucose tracers), blood was simultaneously sampled from the fetal lower abdominal artery (representing blood in the umbilical artery) and the umbilical vein catheters (4.5 mL from each) every 15 minutes for a total of 4 draws for analysis of glucose, blood gases, lactate, AAs, and specific activity of [U-¹⁴C]glucose, ¹⁴CO₂, and ³H₂O. During this period, fetuses received an isovolumic transfusion of maternal blood to replace fetal blood sampled and prevent hypovolemia and acute anemia (8). Immediately following the fourth steady state blood draw, fetuses were subjected to a hyperglycemic clamp study to measure fetal insulin secretion (7). The results of the insulin secretion study have been published in a separate manuscript (7).

Glucagon or control infusions were continued overnight following the tracer study, at which time a necropsy was performed. Sheep received diazepam (0.2 mg·kg⁻¹) and ketamine (20 mg·kg⁻¹) intravenously and fetuses were delivered via maternal laparotomy and hysterotomy. The fetal liver was exposed, and a biopsy was taken from the left lobe and immediately frozen in liquid nitrogen before transfer to −80°C. Following this, the mother and fetus were killed with pentobarbital sodium (390 mg·mL⁻¹ intravenously; 12 and 2 ml, respectively; Bortech Pharmaceuticals, Dearborn, MI). Fetal and fetal organ weights and fetal sex were determined and recorded.

Biochemical analysis of blood samples and study calculations

All biochemical assays and calculations have been previously described (6, 8, 12, 14-16). Whole arterial and umbilical venous blood was collected in EDTA-coated syringes and centrifuged at 14,000 g for 3 min at 4 °C to isolate plasma for immediate measurement of glucose concentrations using a Yellow Springs Instrument 2700 (Yellow Springs Instruments, Yellow Springs, OH). Aliquots of plasma were frozen at −80 °C for measurement of AAs and hormones. Plasma glucagon, insulin, IGF-1, and cortisol concentrations were measured using enzyme-linked immunosorbent assays (ALPCO Immunoassays 48-GLUHU-E01, 80-INSOV-E01, 22-IGFHU-E01, and 11-CORHU-E01-SLV, respectively). Plasma norepinephrine and AA concentrations were measured using high-performance liquid chromatography (HPLC).

Measurement of U-¹⁴C-glucose, ¹⁴CO₂, and ³H₂O was performed as previously described (6, 8, 12, 14-16). The radioactivity of glucose was measured by hemolyzing whole blood, deproteinizing the sample, and incubating the supernatant with glucose oxidase to convert glucose into gluconic acid and then passing the samples through anion exchange columns. Samples were then resuspended, mixed with scintillation fluid, and analyzed with a liquid scintillation counter (Packard Tri-Carb 460 C)(14). Measurement of ¹⁴CO₂ was accomplished by injecting whole blood through a rubber stopper into a 1 mL Nunc CryoTube affixed with epoxy inside a 20 mL glass scintillation vial. Hydrochloric acid was injected into the 1 mL vial to release the CO₂ which was trapped in 0.5 mL of Solvable at the base of the scintillation vial. The rubber stopper was removed, scintillation fluid was added, and radioactivity was determined by analysis with a liquid scintillation counter (Packard Tri-Carb 460 C) (15). ³H₂O was measured by solubilizing plasma in Soluene-350 and scintillation fluid followed by analysis with a liquid scintillation counter (Packard Tri-Carb 460 C) (8).

Biochemical characterization of the fetal liver

Glycogen content

Glycogen content of the fetal liver was measured from pulverized tissue as previously described and is reported as mg of glycogen per gm of liver tissue (16).

Protein expression

Protein lysates were prepared from fetal liver and cotyledonary tissue, and Western immunoblotting was performed as previously described (17). Briefly, 30μg of protein was electrophoresed through a 4-12% NuPAGE Bis-Tris gel (Invitrogen; Waltham, MA) and transferred to a 0.20μm pore nitrocellulose membrane. To visualize PEPCK and Glucose-6-Phosphatase, blots were incubated in 5% BSA in Tris-Bis Solution + 1% Tween with rabbit anti-PCK1/PEPCK (1 μg/μL, Abcam Cat# 28455) or rabbit anti-Glucose-6-Phosphatase (1 μg/μL, Abcam cat# 83690) overnight at 4°C. After washing, the membranes were incubated in 5% Non-Fat Dry Milk in Tris-Bis Solution + 1% Tween with a 1:5000 dilution of goat anti-rabbit (Li-COR #926-32211) for 2 h at room temperature. Equality of sample loading was measured using the Revert 700 Total Protein Stain (Li-COR #926-11011). Protein bands were visualized on Odyssey Fc (LiCOR) at 800 nM for 10 min and quantified with Image Studio 6.0 (LiCOR). Results were quantified and expressed as a ratio of PEPCK or Glucose-6-Phosphatase signal to total protein.

mRNA expression

RNA was extracted from approximately 200 mg of pulverized tissue as previously described (6). RNA concentration was measured using a Nanodrop spectrophotometer. Total RNA (2 μg) was reverse transcribed using the SuperScript III First-Strand Synthesis SuperMix (Invitrogen 18080400), and cDNA was diluted 1:10 with sterile water. Quantitative PCR assays for α-actin-2 (ACTA2), arginase-1 (ARG1), arginase-2 (ARG2), beta-2-microglobulin (B2M), branched chain amino acid transaminase-1 (BCAT1), branched chain amino acid transaminase-1 (BCAT2), glucose-6-phosphatase (G6PC1), glycine decarboxylase (GLDC), glutaminase-1 (GLS1), glutaminase-2 (GLS2), glutamate-ammonia ligase (GLUL), histidine ammonia-lyase (HAL), insulin-like growth factor-1 (IGF1), phosphoenolpyruvate carboxykinase-1 (PCK1), phosphoenolpyruvate carboxykinase-2 (PCK2), peroxisome proliferator activated receptor gamma coactivator-1 alpha (PPARGC1A), ribosomal protein S15 (RPS15), tyrosine aminotransferase (TAT), and urocanate hydratase-1 (UROC1) were performed utilizing primers developed for sheep sequences as previously described (6). A melt-curve analysis was examined to ensure a single peak and correct amplification. Samples were analyzed in triplicate, and the standard curve method of relative quantification was used (18). Genes of interest were normalized to the mean of three reference genes (RPS15, ACTA2, and B2M), and fold-change relative to control samples was calculated. Primers are provided in Table 1.

Table 1.

qPCR Primers

Gene	Forward	Reverse
ACTA2	GAAGATCCTGACCGAGCGAG	TCGTTCTCAAAGTCCAGGGC
ARG1	AGCCTTGTCCTGGGCGGAGA	AGATCTGGGTGGACCCTGGCA
ARG2	GCTCCAGCCACAGGAACCCC	TCCAGGGCCGAGAGCAACCC
B2M	CTTGGTCCTTCTCGGGCTG	ATCTTCTGGCGGGTGTCTTG
BCAT1	CATCCTGGACTTGGCACACA	CAGGCGGTACCTGAACCAAA
BCAT2	TGTCCTCCGTTTCCACAAGG	AGCTTTACACCGGGAGCATC
G6PC1	GGATTCTGGATCGTGCAACT	ATCCAATGGCGAAACTGAAC
GLDC	TTTCCAGCCTAACAGCGGAG	AGGATTGGTCCCATGTGCAG
GLS1	CCCAGAAGGCACAGACATGGTTGG	GGGCAGAAGCCACCATTAGCCA
GLS2	CTGGTGCCATTGTTGTGAGC	ATGTGGCATTGCTGAAACCC
GLUL	TGGGCACCCCTTTGGTTGGC	GCCTTGTCCGCTCCCACACC
HAL	GATGCATACACCTTGCGCTG	GGCCAAATAGTCCAGGGCTT
IGF1	GAGACCCTCTGCGGGGCTGA	CTGCTCGAGCCGTACCCCGT
PCK1	TGTCCGAGGAGGATTTTGAG	ATGCCAATCTTGGACAGAGG
PCK2	GCCTGTGCTTCAGGCCCTGG	TGCATGGCCACTGGCACACC
PPARGC1A	CACCAAACCCACAGAGAACC	GTGACTCTGGGGTCAGAGGA
RPS15	ATCATTCTGCCCGAGATGGTG	CGGGCCGGCCATGCTTTACG
TAT	CCTGCCGACAGATCCTGAAG	TCCTCCCGACTGGATAAGCA
UROC1	ACGGGCAGGTGTTCAGCAA	CCATTGGTGATGACCAGCC

ACTA2, α-actin-2; ARG1, arginase-1; ARG2, arginase-2; B2M, beta-2-microglobulin; BCAT1, branched-chain amino acid transaminase-1; BCAT2, branched-chain amino acid transaminase-2; G6PC1, glucose-6-phosphatase catalytic-subunit; GLDC, glycine decarboxylase; GLS1, glutaminase-1; GLS2, glutaminase-2; GLUL, glutamate-ammonia ligase; HAL, histidine ammonia-lyase; IGF1, insulin-like growth factor-1; PCK1, phosphoenolpyruvate carboxykinase 1, cytosolic; PCK2, phosphoenolpyruvate carboxykinase 2, mitochondrial; PPARGC1A, PPARG coactivator-1α; RPS15, ribosomal protein S15; TAT, tyrosine aminotransferase; UROC1, urocanate hydratase-1.

Amino acid concentrations

Free hepatic AA concentrations were measured from pulverized tissue with HPLC as previously described and is reported as μmole per gm of liver tissue (19).

Metabolomics analysis

Fetal liver tissue samples (25 mg) were subjected to targeted metabolomic profiling at the University of Colorado Metabolomic Core as previously described (6). Multivariate principal component analysis was performed using partial least squares discriminant analysis (PLS-DA) with CON and GCG groups. 186 metabolites were detected, and the CON and GCG groups were separated by PLS-DA analysis. The metabolites with the variable importance in projection (VIP) scores (rank order) with FDR values ≤ 0.05 based on one-way ANOVA analysis were identified and entered into a Pearson-Ward unsupervised heatmap generator in MetaboAnalyst 6.0. The lists of significant VIP metabolites were also included in an enrichment analysis through MetaboAnalyst 6.0, with a hypergeometric test, utilizing KEGG IDs to identify the compounds, and were mapped to the SMPDB database. LogP values ≤ 1.30 were used as a threshold for identifying significantly enriched pathways, and enrichment scores were calculated as a ratio of total metabolite hits to expected hits. To interrogate differences between GCG-5 and GCG-50 groups, a similar analysis was performed which was restricted to the GCG animals and focused on differences between these two groups (GCG-5 and GCG-50).

Statistical analysis

Statistical analysis was performed using GraphPad Prism (V.10.1). Results were expressed as mean ± Standard Deviation (SD). For the main analysis we combined the two GCG groups and compared them to the CON group with a Student’s t-test or the Mann-Whitney test for parametric and non-parametric data, respectively. In a post hoc analysis, we compared the GCG-5 and GCG-50 groups using the same approach. P-values <0.05 were accepted as significant. Statistical methods and thresholds for metabolomic analysis were described above.

RESULTS

Fetal and maternal characteristics and nutrient and hormone concentrations

Characteristics of animals included in this study have been reported (7, 8). Data from animals included in this manuscript are summarized in Table 2, including the GCG-5 and GCG-50 groups separately. At the end of the experimental infusions, GCG fetuses had plasma arterial glucagon concentrations that were 800% higher than CON (P<0.0001) and insulin concentrations that were 55% lower than CON (P=0.0012). Fetal cortisol and norepinephrine concentrations were similar in the two groups. While fetal IGF-1 concentrations were not different, hepatic IGF1 mRNA was 45% lower in GCG compared to CON. Fetal plasma arterial glucose concentrations were similar between groups, but total AA concentrations were 46% lower in the GCG group compared to CON (P<0.0001). Despite similar gestational ages at time of collection, GCG fetuses were 11% lighter than CON fetuses, though this is not statistically significant in this cohort of GCG animals (P=0.0729).

Table 2.

Fetal Characteristics and Nutrient and Hormone Concentrations

	Control (n=7)			GCG (All, n=11)				GCG-5 (n=7)			GCG-50 (n=4)
	Mean		SD	Mean		SD	P Value	Mean		SD	Mean		SD	P Value
Gestational Age	134	±	1	135	±	1	0.1391	135	±	1	136	±	0	0.1274
Fetal
% female	33			58			0.2563	43			100			0.0300
Weight (gm)	3273	±	437	2897	±	383	0.0729	2971	±	287	2769	±	538	0.4294
Liver	89.6	±	14.7	68.2	±	17.1	0.0134	73.8	±	14.1	60.4	±	19.4	0.0654
Glucose (mmol/L)	1.1	±	0.2	0.9	±	0.3	0.3608	0.9	±	0.2	1.0	±	0.5	0.7879*
Summed Amino Acids (μmol/L)	4331	±	233	2341	±	748	<0.0001 *	2665	±	539	1774	±	783	0.0508
Glucagon (pg/mL)	38.8	±	24.0	348.4	±	362.2	<0.0001 *	194.3	±	39.7	618.2	±	530.7	0.5273*
Insulin (ng/mL)	0.28	±	0.11	0.13	±	0.06	0.0012	0.10	±	0.05	0.17	±	0.04	0.0217
Cortisol (ng/mL)	19.2	±	15.1	34.3	±	24.8	0.1989	39.5	±	23.1	25.0	±	28.4	0.3776
Norepinephrine (pg/mL)	469	±	186	518	±	316	0.7171	532	±	311	493	±	372	0.8530
IGF-1 (ng/mL)	119	±	24	93	±	42	0.1846	88	±	34	101	±	58	0.6424
Hepatic IGF1 mRNA (ratio to CON)	1.00	±	0.34	0.55	±	0.27	0.0072	0.51	±	0.31	0.63	±	0.19	0.49595

P-Values represent comparisons of all CON vs GCG animals or between GCG-5 and GCG-50 animals using unpaired Student's t-test or the Mann-Whitney test (*).

Hepatic glucose metabolism

Fetal liver glycogen was 55% lower in GCG compared to CON (Figure 1A). Hepatic mRNA content of PCK1 and G6PC1 were five-fold and two-fold higher, respectively, in GCG fetal livers compared to CON (Figure 1B). Hepatic protein expression of phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase (G6Pase) was similar between groups (Figure 1C).

Figure 1. Regulators of Hepatic Glucose Metabolism.

Fetal liver glycogen content (A), PCK1, PCK2, and G6PC1 mRNA content (B), and phophoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase (G6Pase) protein content (C) are shown for control animals (CON, white circles), GCG-5 animals (black diamonds), and GCG-50 animals (gray squares) along with mean and SD bars. CON animals were compared to the combined GCG-5 and GCG-50 groups with Student’s t-test or the Mann-Whitney test (PCK1, G6PC1) and exact P-Values are provided above the horizontal bar. GCG-5 animals were compared to GCG-50 with a Student’s t-test and the * to the right of the vertical bar indicates P<0.05. If there was not a statistically significant difference, no vertical line or * is present.

Fetal glucose metabolism

Rates of fetal glucose utilization and oxidation were lower in the GCG groups compared to CON, though net fetal glucose uptake from the placenta was not different between groups (Figure 2A-C). When normalizing these rates to fetal body weight, only the fetal glucose oxidation rate remained lower in the GCG group relative to CON (Figure 2E-G). Whether expressed relative to body weight or not, glucose production by the fetus was not different between groups (Figure 2D,H). The fetal glucose oxidation fraction (umbilical to uterine uptake ratio) also was not different between groups (56 ± 21% CON vs 60 ±17% GCG, P=0.6991).

Figure 2. Fetal Glucose Metabolic Rates.

Fetal umbilical glucose uptake, glucose utilization, glucose oxidation, and glucose production rates, both absolute (A-D) and relative to fetal body weight (E-H) are shown for control animals (CON, white circles), GCG-5 animals (black diamonds), and GCG-50 animals (gray squares) along with mean and SD bars. CON animals were compared to the combined GCG-5 and GCG-50 groups with Student’s t-test or the Mann-Whitney test (D,H) and exact P-Values are provided above the horizontal bar. GCG-5 animals were compared to GCG-50 with a Student’s t-test but there were no statistically significant differences (P>0.05).

Hepatic amino acid metabolism

Given the impact of the glucagon infusion on fetal AA concentrations we measured fetal liver mRNA content of several genes which regulate AA catabolism (Figure 3). ARG2, GLS2, BCAT1, BCAT2, GLUL, HAL, UROC1, and PPARGC1A mRNA were all higher in the GCG group compared to CON (P<0.05). Concentrations of several AA were significantly lower in GCG compared to CON (Table 3). Those that reached statistical significance included alanine, citrulline, glutamate, glutamine, isoleucine, ornithine, threonine, and valine (P<0.05). No AAs were significantly higher in the GCG group compared to CON.

Figure 3. Hepatic mRNA expression of amino acid catabolism related genes.

Fetal liver mRNA content for genes replated to AA metabolism are shown for control animals (CON, white circles), GCG-5 animals (black diamonds), and GCG-50 animals (gray squares) along with mean and SD bars. CON animals were compared to the combined GCG-5 and GCG-50 groups with Student’s t-test or the Mann-Whitney test (GLS2, BCAT2, UROC1, TAT, PPARGC1A) and exact P-Values are provided above the horizontal bar. GCG-5 animals were compared to GCG-50 with a Student’s t-test and the * to the right of the vertical bar indicates P<0.05. If there was not a statistically significant difference, no vertical line or * is present.

Table 3.

Fetal Liver Amino Acid Concentrations

	Control (n=7)			GCG (All, n=11)				GCG-5 (n=7)			GCG-50 (n=4)
Amio Acid (μmol/gm)	Mean		SD	Mean		SD	P Value	Mean		SD	Mean		SD	P Value
ALA	3.03	±	0.66	2.00	±	1.61	0.0346 *	1.82	±	0.93	2.31	±	2.59	0.6485
ARG	0.80	±	0.14	1.00	±	0.22	0.0503	0.93	±	0.18	1.13	±	0.24	0.1497
ASP	1.86	±	0.52	2.11	±	0.51	0.3238	2.20	±	0.63	1.95	±	0.10	0.2303
ASPG	0.16	±	0.03	0.16	±	0.02	0.6119	0.17	±	0.02	0.15	±	0.02	0.2708
CIT	0.23	±	0.06	0.12	±	0.06	0.0019	0.14	±	0.06	0.07	±	0.05	0.0954
CYST	0.13	±	0.05	0.09	±	0.09	0.2917	0.10	±	0.07	0.07	±	0.13	0.6785
GLN	1.60	±	0.15	0.95	±	0.46	0.0059 *	0.97	±	0.27	0.91	±	0.75	0.2303*
GLU	3.73	±	0.67	2.77	±	0.78	0.0159	2.98	±	0.79	2.39	±	0.71	0.2519
GLY	2.94	±	0.76	3.00	±	0.57	0.8487	3.31	±	0.43	2.45	±	0.31	0.0071
HIS	0.26	±	0.04	0.25	±	0.07	0.5388	0.23	±	0.05	0.28	±	0.10	0.2553
ILEU	0.08	±	0.02	0.05	±	0.03	0.0306	0.05	±	0.04	0.04	±	0.03	0.5012
LEU	0.14	±	0.03	0.15	±	0.04	0.7379	0.14	±	0.05	0.15	±	0.03	0.6905
LYS	0.27	±	0.10	0.31	±	0.17	0.5561	0.25	±	0.06	0.43	±	0.25	0.1030*
MET	0.06	±	0.03	0.04	±	0.03	0.1709	0.04	±	0.03	0.04	±	0.03	0.9687
ORN	0.59	±	0.14	0.33	±	0.12	0.0006	0.34	±	0.10	0.30	±	0.17	0.6180
PHE	0.07	±	0.05	0.03	±	0.05	0.1072	0.04	±	0.06	0.01	±	0.03	0.5417
SER	5.03	±	0.95	5.83	±	2.77	0.4789*	6.66	±	2.90	4.37	±	2.06	0.2027
TAU	4.50	±	1.26	5.17	±	1.45	0.3341	4.87	±	1.58	5.68	±	1.19	0.4023
THR	3.15	±	0.65	2.03	±	0.88	0.0107	2.30	±	0.63	1.55	±	1.15	0.1841
TYR	0.07	±	0.04	0.04	±	0.07	0.3143	0.04	±	0.07	0.05	±	0.10	0.8125
VAL	0.38	±	0.06	0.26	±	0.08	0.0051	0.28	±	0.09	0.23	±	0.06	0.3412

P-Values represent comparisons of all CON vs GCG animals or between GCG-5 and GCG-50 animals using unpaired Student's t-test or the Mann-Whitney test (*).

Metabolite analysis of the fetal liver

Experimental hyperglucagonemia significantly impacted fetal hepatic nucleotides (Figure 4). For the nucleotides of adenosine, guanine and cytosine, the tri- and diphosphates were significantly higher in GCG livers compared to CON. For cytosine and guanine, the monophosphate was significantly lower in GCG livers compared to CON. There were no significant differences for the uracil nucleotides. Furthermore, for all triphosphate nucleotides and three of the four diphosphate nucleotides (ADP, GDP, and CDP) the GCG-5 levels were significantly higher than the GCG-50 levels. For the monophosphate nucleotides the GCG-5 levels were lower than CON for three of the four nucleotides (AMP, CMP, and UMP).

Figure 4. Fetal Liver Nucleotide Content.

Fetal liver nucleotide content for adenosin (A-C), guanine (D-F), cytosine (G-I), and uracil (J-L) are shown for control animals (CON, white circles), GCG-5 animals (black diamonds), and GCG-50 animals (gray squares) along with mean and SD bars. CON animals were compared to the combined GCG-5 and GCG-50 groups with the Mann-Whitney test or Student’s t-test (F,K,L) and exact P-Values are provided above the horizontal bar. GCG-5 animals were compared to GCG-50 with a Student’s t-test and the * to the right of the vertical bar indicates P<0.05. If there was not a statistically significant difference, no vertical line or * is present.

Experimental hyperglucagonemia depleted several metabolic pathway intermediates including citrate, a TCA cycle intermediate, and D-Ribulose 5-phosphate, an intermediate of the pentose phosphate pathway (Figure 5). In GCG-50 livers, phosphate levels were higher than in GCG-5 and CON but the levels of the oxidative stress marker, glutathione disulfide, were greater in GCG-5. Experimental hyperglucagonemia also impacted metabolites involved in hepatic fatty acid metabolism, reducing concentrations of oleic acid (octadecanoic acids), docosohexaenoic acid (DHA), (8Z-11Z-14Z)-Icosatrienoic acid (an omega-6-fatty acid [20:3 n-6]), and (7Z-10Z-13Z-16Z-19Z)-Docosapentaenoic acid (an omega-3 fatty acid [22:5 n-3]). Glucagon infusion increased concentrations of the short-chain fatty acid pentanoate (valerate), and both short chain and medium chain acetyl-carnitine. Experimental hyperglucagonemia elevated levels of N-methylethanolamine phosphate, a phospholipid biosynthesis precursor. It also reduced concentrations of pantethenine, a component of coenzyme A, allantoate which is involved in nitrogen metabolism, and dimethylglycine, a biproduct of choline/glycine breakdown in the GCG-50 group in comparison to CON. Concentrations of picolinic acid, which is a product of tryptophan breakdown were elevated in the GCG-5 vs CON and GCG-50 groups. Despite differential responses in some metabolites to the two different glucagon infusion rates, as demonstrated in Figure 5, glucagon exposure of either dose led to enrichment of many pathways (Figure 6) including many associated with AA catabolism, the urea cycle, mitochondrial electron transport chain, and lower lipid biosynthesis.

Figure 5. Heat Map of Top VIP Fetal Liver Metabolites.

Metabolites in control (CON) and glucagon treated fetal livers (GCG) identified as top variable importance (VIP) metabolites and organized by functional category in a heat map. All VIP have FDR values ≤ 0.05 and are associated with changes in energy metabolism and redox reactions, lipids and fatty acids, and metabolic enzymes and co-factors. Nucleotides and AAs are not shown as they are presented elsewhere.

Figure 6. Top Enriched Pathways in the Fetal Liver.

Bubble plot of enriched pathways identified by inputting all significant VIP metabolites as determined by an FDR value ≤ 0.05. A threshold of Logp values ≤ 1.30 was used to identify significantly enriched pathways, and enrichment scores were calculated as a ratio of total hits to expected hits.

DISCUSSION

In this study, we demonstrated that experimental fetal hyperglucagonemia for nine days at the end of gestation depleted hepatic glycogen and increased hepatic mRNA expression of PCK1 and G6PC1, two key enzymes responsible for gluconeogenesis. Moreover, experimental hyperglucagonemia also increased hepatic mRNA expression of several genes involved in AA catabolism, including ARG2, GLS2, BCAT1, BCAT2, GLUL, HAL, UROC1, and PPARGC1A and resulted in lower fetal plasma AA concentrations. Additionally, we observed significant increases in hepatic tri- and diphosphate nucleotide concentrations for adenosine, guanine, and cytosine, in glucagon-infused fetuses compared to controls. This is consistent with increased hepatic energy production from AA catabolism and oxidation. These findings further support a role for glucagon as a regulator of fetal AA homeostasis, and may have implications for the metabolic adaptations observed in fetuses exposed to chronic intrauterine stressors such as hypoxemia, placental insufficiency, and fetal growth restriction (3, 4).

Directly supporting our hypothesis, nine days of experimental hyperglucagonemia robustly stimulates a robust hepatic AA catabolic transcriptional program, underscoring the metabolic responsiveness of the fetal liver to glucagon signaling. This sustained transcriptional response appears to become more pronounced over time, as our prior study of a two-hour glucagon infusion into the fetus demonstrated less widespread increases in mRNA expression of genes regulating AA catabolism (6). Furthermore, in that acute study, we observed reductions in plasma AA concentrations, but the magnitude of the decreases was less pronounced than in the current study. Notably, the transcriptional response after nine days of glucagon infusion was more consistent with the broader set of gene expression changes observed in isolated fetal sheep hepatocytes exposed high doses of glucagon for 24 hours in vitro (6). Together, these findings indicate that the fetal liver retains significant sensitivity to glucagon’s actions on AA metabolism during sustained exposure.

Our prior manuscript employing these same animals reported the results from our leucine tracer kinetics and demonstrated that whole-fetus leucine utilization and oxidation rates were lower following nine days of experimental hyperglucagonaemia (8). While seemingly contradictory to our increased mRNA expression of hepatic AA catabolic genes, it is important to remember that the majority of leucine oxidation occurs in the muscle and not the liver (20). Furthermore, these leucine kinetic measurements reflect the totality of fetal metabolism and are not specific to the liver. Additionally, the reduced rates of leucine utilization and oxidation may have been driven by the lower plasma and intracellular AA concentrations observed at the end of the glucagon infusion (8, 21).

To more directly assess hepatic metabolism in these fetuses, we analyzed liver tissue metabolites in the current study and found that despite lower plasma and hepatic AA concentrations, three hepatic tri- and diphosphate nucleotides (ATP, GTP, CTP, ADB, GDP, and CDP) were higher in livers of glucagon infused fetuses. Consistent with this, we found lower hepatic concentrations of two out of the four monophosphate nucleotides (GMP and CMP). This nucleotide pattern is expected for livers actively catabolizing and oxidizing AAs to generate high-energy nucleotides (22). This pattern is further supported by enrichment for many AA catabolizing pathways and the urea cycle, the depletion of metabolic pathway intermediates, and lower concentrations of hepatic AA and increased AA catabolites. These metabolite changes highlight that while systemic AA utilization may decrease with the glucagon infusion, the fetal liver maintains significant capacity for AA catabolism and energy generation in response to prolonged glucagon signaling.

Although nine days of experimental fetal hyperglucagonemia produced clear effects when compared to controls, we also observed differences between the GCG-5 and GCG-50 groups. Notably, fetal plasma total AA concentrations were 33% lower in the GCG-50 relative to the GCG5 group, though this failed to reach statistical significance (P=0.0508). Despite this, intrahepatic AA concentrations were largely not different between the groups. Another important set of differences between these groups are the mRNA levels for several genes involved in AA catabolism and in the hepatic metabolites. While the GCG-50 group generally demonstrated higher mRNA levels of ARG2, BCAT2, and GLDC, the hepatic tri- and diphosphate nucleotide concentrations were lower in the GCG-50 group compared to the GCG-5 group. The reasons for these divergent dose-dependent effects remain unclear. One possibility is that the higher mRNA expression observed in the GCG-50 group may have driven greater AA catabolic and oxidative activity earlier during the infusion period, leading to transiently higher rates of AA metabolism. However, by the time tissue was collected after nine days of glucagon exposure, these catabolic pathways may have been downregulated at the level of enzyme activity, despite persistent elevations in gene transcription. Nonetheless, the present study was not specifically powered to detect differences between the two glucagon infusion rates.

Nine days of experimental hyperglucagonemia also had an important impact on fetal hepatic glucose metabolism. Hepatic glycogen content was significantly decreased in glucagon-infused fetuses, suggesting that sustained glucagon exposure promotes glycogenolysis, though this may be transient. In our study testing the impact of a two-hour glucagon infusion we also found lower hepatic glycogen concentrations, but in those experimental conditions fetal glucose concentrations increased with the glucagon infusion (6). Interestingly, while PEPCK and G6Pase protein levels were not increased, we observed significant upregulation of the mRNA transcripts PCK1 and G6PC1 in glucagon-infused fetuses compared to controls. It is notable that PEPCK plays a role not only in gluconeogenesis but also in AA metabolism (23). Thus, in addition to its role in glucose metabolism, the increased expression of PCK1 mRNA in the current study may support enhanced AA turnover. Despite these liver specific changes following experimental glucagon infusions, there was no evidence of net fetal glucose production, as measured by glucose tracer kinetics, consistent with previous observations of relative glucagon resistance in the fetal liver for glucose production (5). Additionally, hepatic protein expression of PEPCK and G6Pase were unchanged between glucagon-infused and control fetuses.

One possible explanation for the apparent resistance to glucagon for stimulating endogenous fetal glucose production are low fetal plasma AA concentrations. In conditions where fetal glucose production is activated in late gestation, the fetal liver increases uptake of AA to support glucose synthesis (24). Conversely, in the current study, chronic glucagon infusion resulted in significantly lower fetal plasma AA concentrations, potentially limiting substrate availability for gluconeogenesis. Moreover, previous work has shown that exogenous AA can stimulate hepatic glucose production (25), further underscoring the role of AAs as critical regulators of fetal glucose metabolism. Thus, it is possible that in both the current chronic infusion study and in earlier studies investigating acute glucagon stimulation, the absence of a measurable gluconeogenic response reflects insufficient AA substrate availability rather than true hepatic insensitivity to glucagon. Another possible explanation for the lack of endogenous fetal glucose production at the end of a nine-day experimental glucagon infusion are the normal fetal glucose concentrations. We have previously demonstrated that hypoglycemia is sufficient to stimulate endogenous glucose production in normally growing fetuses (24). We have also shown that acutely correcting fetal hypoglycemia in a model of placental insufficiency and fetal growth restriction turns off the endogenous fetal glucose production normally seen in this model (Limesand et al., 2007).

Another interesting finding was the significant reduction in hepatic IGF1 mRNA expression following nine days of experimental hyperglucagonemia. This aligns with reduced fetal growth and lower fetal body weights, though fetal arterial plasma IGF-1 concentrations were not statistically lower in these animals (8). Given that AA can stimulate hepatic IGF1 mRNA expression (26, 27), the lower plasma AA concentrations observed in the current study may have contributed to the reduction in hepatic IGF1 mRNA transcript levels.

Our liver metabolite analysis revealed that the glucagon infusion also may decrease fetal hepatic fatty acid biosynthesis or increase fatty acid oxidation, as evidenced by lower concentrations of various omega-3 and omega-6 fatty acids including DHA and elevated levels of acetyl-carnitines. Glucagon has been previously implicated in the regulation of hepatic lipid oxidation in the adult, especially by regulating fatty acid transport into the mitochondria through the rate limiting enzyme carnitine palmitoyl transferase (1, 28). In the fasted wild-type mouse, hepatic β-oxidative genes including carnitine palmitoyltransferase-1 are upregulated in response to fasting, but this response in not observed in glucagon receptor null mice (29). While our study indicates glucagon may be involved in fetal fatty acid oxidation, the direct mechanisms and implications for this regulation in the fetus remain to be elucidated.

In summary, nine days of experimental hyperglucagonemia in the late-gestation fetal sheep activates hepatic AA catabolic pathways highlighting the metabolic responsiveness of the fetal liver to glucagon. Our findings extend previous acute infusion studies by demonstrating that chronic glucagon exposure leads to sustained and more widespread transcriptional activation of AA metabolism genes, increases in hepatic tri- and diphosphate nucleotide pools indicative of increased AA oxidation, and depletion of hepatic glycogen reserves. The lack of a corresponding increase in endogenous fetal glucose production, despite elevated expression of gluconeogenic mRNAs, underscores a functional resistance of the fetal liver to glucagon-induced gluconeogenesis, which may reflect limited substrate availability due to lower fetal plasma and intrahepatic AA concentrations or a sufficient amount of circulating glucose. Collectively, these data demonstrate that in the previously normally growing fetus a predominant metabolic role of glucagon is the regulation of AA concentrations. This is further supported by previous studies demonstrating the inhibitory role that chronic fetal hyperglucagonemia has on placental AA uptake from the parental circulation and transfer to the fetal circulation (8). We recognize that complications of pregnancy resulting in fetal growth restriction cause changes in many other factors, not just fetal glucagon concentrations. Thus, fetal glucagon infusions do not fully recapitulate any of these pathologies. Further investigation is warranted to determine how pathways interrogated in the current study are regulated by glucagon under those pathologic conditions.