Disrupted metabolic flux balance between pyruvate dehydrogenase and pyruvate carboxylase in human fatty liver

Abstract

Hepatic metabolism involving pyruvate carboxylase (PC) and pyruvate dehydrogenase (PDH) may be abnormal in fatty livers. In this study, [¹³C]bicarbonate production from [1-¹³C₁]pyruvate in the liver and glycerol glyceroneogenesis were examined in relation to hepatic fat content using hyperpolarized [1-¹³C₁]pyruvate and oral [U-¹³C₃]glycerol. After an overnight fast, 15 subjects with a range of hepatic fat content received hyperpolarized [1-¹³C₁]pyruvate intravenously to assess its conversion to [1-¹³C₁]lactate and [¹³C]bicarbonate in the liver. They also received oral [U-¹³C₃]glycerol, followed by venous blood sampling to examine glucose and the glycerol backbone of the triglycerides released primarily from the liver. From hyperpolarized [1-¹³C₁]pyruvate, participants with high intrahepatic fat fraction produced higher [1-¹³C₁]lactate and lower [¹³C]bicarbonate than those with low liver fat. The fraction of plasma triglycerides derived from oral [U-¹³C₃]glycerol via the TCA cycle was similar between groups. The fraction of plasma [5,6-¹³C₂]glucose, which reflects PC flux, decreased in subjects with fatty liver. In contrast, the fraction of [4,5-¹³C₂]glucose + [6-¹³C₁]glucose, which can be produced via either PC or PDH, was comparable between groups. The study results suggest a shift in pyruvate metabolism in fatty liver, with a decreased metabolic flux ratio of PC/PDH. The methodology in this study provides insights into fatty liver metabolism of human subjects inaccessible previously and is applicable to advanced liver diseases such as cirrhosis and hepatomas.

Keywords: Fatty liver; Glyceroneogenesis; Hyperpolarized; Pyruvate carboxylase; Pyruvate dehydrogenase.

Figure. 1.. HP [1-¹³C₁]pyruvate and oral [U-¹³C₃]glycerol to assess hepatic metabolism through the TCA cycle, gluconeogenesis, and triglyceride synthesis.

The probes used simultaneously, HP [1-¹³C₁]pyruvate and [U-¹³C₃]glycerol, are highlighted in yellow. Phosphorylated glycerol in the liver may enter the glycolytic pathway, gluconeogenesis, or triglyceride synthesis through fatty acid esterification. Glucose and the glycerol backbone of triglycerides derived directly from [U-¹³C]glycerol retain all three labeled carbons. Alternatively, a fraction of glycerol may be metabolized to pyruvate followed by the TCA cycle where a ¹³C may be lost. After export from the TCA cycle, PEP enters gluconeogenesis and glyceroneogenesis, producing glucose and triglyceride-glycerol, respectively, labeled with one or two ¹³C. The other probe, HP [1-¹³C₁]pyruvate, may enter the TCA cycle via either PDH or PC, releasing ¹³CO₂. ¹³C labeling in these products encodes information about metabolic activities through the TCA cycle in mitochondria. DHAP, dihydroxyacetone phosphate; G3P, glycerol-3-phosphate; GA3P, glycerol-3-phosphate; GK, glycerol kinase; PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase; PEP, phosphoenolpyruvate; PEPCK, phosphoenolpyruvate carboxykinase; TCA, tricarboxylic acid.

Figure. 2.. Study protocol.

After an overnight fast, participants drank water containing [U-¹³C₃]glycerol (50 mg/kg bw) at 11:00 AM followed by a series of venous blood draws for three hours with 30-min intervals. Immediately after the second blood draw, subjects inside an MR scanner received HP [1-¹³C₁]pyruvate intravenously for dynamic ¹³C MRS from the liver.

Figure. 3.. Time-courses of HP ¹³C-metabolites and removal of abdominal lipid profile.

(A) Typical time-accumulated ¹³C spectrum from the liver included the abdominal lipid profile, which could be removed by subtracting the baseline ¹³C scan. (B) HP metabolites appeared ~10 s from the injection of HP [1-¹³C₁]pyruvate and most signals became undetectable after 150 s. The presented data is collected from a healthy subject after overnight fasting (60 yo, white female, BMI = 25.3, fat fraction = 3.8 ± 1.9%). (C) [1-¹³C₁]Aspartate peak, uniquely produced via pyruvate carboxylation, was detected only from few subjects with excellent B₀ homogeneity and polarization.

Figure. 4.. Correlation between intrahepatic fat fraction and HP ¹³C measurements.

Representative ¹³C spectra from study participants with (A) lean liver (42 yo, white female, BMI = 31.4, fat fraction = 1.6 ± 3.3%) and (B) fatty liver (40 yo, white female, BMI = 32.9, fat fraction = 14.6 ± 8.2%). (C) Lactate signal was positively correlated with intrahepatic fat fraction (R² = 0.334, P = 0.024). Participants with fatty liver (high FF) showed higher lactate as compared to those with lean liver (low FF, P = 0.002). (D) Bicarbonate level showed an opposite trend (R² = 0.280, P = 0.062). Participants with high liver fat had lower bicarbonate (P = 0.003). (E) Alanine did not correlate with fat content (R² = 0.181, P = 0.114). Alanine was lower in participants with high fat content (P = 0.040). (F) The metabolite ratio of bicarbonate to lactate showed a stronger, negative correlation with liver fat content (R² = 0.340, P = 0.040), differentiating participants with high liver fat from the others (P = 0.002). (G) Pearson correlation coefficients between BMI, liver fat, fasting insulin, fasting glucose, HOMA-IR, and HP measurements. * and ** indicate P < 0.05 and P < 0.01, respectively. FF, fat fraction; TP, total ¹³C products.

Figure. 5.. Labeling patterns in TG-glycerol and glucose after [U-¹³C₃]glycerol metabolism through the TCA cycle.

(A) If [U-¹³C₃]glycerol is metabolized via PC-PEPCK, [U-¹³C₃]PEP and [2,3-¹³C₂]PEP are generated. The [2,3-¹³C₂]PEP yields TG-[1,2-¹³C₂]glycerol, TG-[2,3-¹³C₂]glycerol, or [5,6-¹³C₂]glucose. (B) When [U-¹³C₃]glycerol is metabolized via PC-TCA-PEPCK, [1,2-¹³C₂]PEP and [3-¹³C₁]PEP are generated. The [1,2-¹³C₂]PEP yields TG-[1,2-¹³C₂]glycerol, TG-[2,3-¹³C₂]glycerol, or [4,5-¹³C₂]glucose while [3-¹³C₁]PEP leads to TG-[1-¹³C₁]glycerol, TG-[3-¹³C₁]glycerol, or [6-¹³C₁]glucose. (C) If [U-¹³C₃]glycerol is metabolized via PDH-TCA-PEPCK, the same labeling patterns are expected as the PC-TCA-PEPCK path. Open circles represent ¹²C, and filled black circles represent ¹³C. aKG, α-ketoglutarate; OAA, oxaloacetate; PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase; PEP, phosphoenolpyruvate; PEPCK, phosphoenolpyruvate carboxykinase; TCA, tricarboxylic acid.

Figure. 6.. Evolution of ¹³C labeling in TG-glycerol from oral [U-¹³C₃]glycerol in fatty vs. lean livers.

Blue spectra and dots indicate data from participants with low liver fat. Red indicates high liver fat data. (A) Representative ¹³C NMR spectra of the backbones of TGs from blood after administration of ¹³C glycerol. (B) Total TG and (C) ¹³C-labeled TGs were higher in subjects with high liver fat compared to subjects with low liver fat. (D) The fraction of TG-[¹³C₂]glycerol and (E) the mass of TG containing [¹³C₂]glycerol were comparable between the groups. (F) Fractional enrichment of TG-[¹³C]glycerol was generally lower in the high liver fat group (P = 0.064 – 0.518 per time-point) and (G) the time-averaged fractional enrichment had a negative correlation with intrahepatic fat fraction (P = 0.033). * indicates P < 0.05.

Figure. 7.. Evolution of ¹³C labeling in plasma glucose after oral [U-¹³C₃]glycerol in fatty vs. lean livers.

A spectrum and dots in blue indicate data from participants with low liver fat. Red indicates high liver fat data. (A) The relation between multiplets and the isotopomers in glucose. Throughout the experiments, (B) total glucose and (C) ¹³C-labeled glucose were higher in subjects with high liver fat. (D, E) The concentrations of glucose derived from the indirect pathways (e.g., [5,6-¹³C₂]glucose, [4,5-¹³C₂]glucose, and [6-¹³C₁]glucose) were comparable between the groups. (F) The ratio of [5,6-¹³C₂]/([4,5,6-¹³C₃]+[U-¹³C₆]) in glucose was lower in fatty liver. (G) The ratio of ([4,5-¹³C₂]+[6-¹³C₁])/([4,5,6-¹³C₃]+[U-¹³C₆]) in glucose was comparable between the groups. (H) The ratio of [5,6-¹³C₂]/[4,5-¹³C₂] in glucose ratio was lower in fatty liver, implying a decreased PC-to-PDH flux balance. * indicates P < 0.05.