Personal model-assisted identification of NAD+ and glutathione metabolism as intervention target in NAFLD

Abstract

To elucidate the molecular mechanisms underlying non-alcoholic fatty liver disease (NAFLD), we recruited 86 subjects with varying degrees of hepatic steatosis (HS). We obtained experimental data on lipoprotein fluxes and used these individual measurements as personalized constraints of a hepatocyte genome-scale metabolic model to investigate metabolic differences in liver, taking into account its interactions with other tissues. Our systems level analysis predicted an altered demand for NAD+ and glutathione (GSH) in subjects with high HS Our analysis and metabolomic measurements showed that plasma levels of glycine, serine, and associated metabolites are negatively correlated with HS, suggesting that these GSH metabolism precursors might be limiting. Quantification of the hepatic expression levels of the associated enzymes further pointed to altered de novo GSH synthesis. To assess the effect of GSH and NAD+ repletion on the development of NAFLD, we added precursors for GSH and NAD+ biosynthesis to the Western diet and demonstrated that supplementation prevents HS in mice. In a proof-of-concept human study, we found improved liver function and decreased HS after supplementation with serine (a precursor to glycine) and hereby propose a strategy for NAFLD treatment.

Keywords: NAFLD; glutathione; personalized genome‐scale metabolic modeling; serine.

© 2017 The Authors. Published under the terms of the CC BY 4.0 license.

Figures

Figure 1. Generation of VLDL kinetics and plasma metabolomics data

1. To identify the metabolic changes in response to increased hepatic steatosis (HS), secretion rate of the VLDLs from the liver of subjects was measured and the plasma metabolite levels were detected in subjects with varying degrees of HS.

2. BMI, insulin resistance (HOMA‐IR), plasma TGs, and ALT levels are significantly correlated with the independently measured liver fat.

3. The subjects were categorized into two groups as high (n = 43) and low HS (n = 43). Body mass index (BMI), fasting plasma insulin (FPI), plasma triglycerides (TGs), and plasma alanine aminotransferase (ALT) levels are found to be significantly different between the two groups. Data are presented as means ± SD. *P < 0.05; Student's t‐test.

Figure 2. Personalized modeling of liver in subjects with varying degrees of HS

1. A

   Schematic illustration of how personalized genome‐scale metabolic modeling can be performed accounting the interactions between other tissues and red blood cells for the development of effective therapeutic approaches for non‐alcoholic fatty liver disease (NAFLD). Solid and dashed arrows show the outputs and inputs to the tissues, respectively.

2. B, C

   The correlation between the predicted intracellular fluxes of the liver and hepatic steatosis (HS) is assessed and compared with the (B) apolipoprotein B (apoB) and (C) triglycerides (TG) content in the total VLDL production.

3. D

   The fluxes carried by the reactions catalyzed by NNT, GSR, and GPXs are found to be the most correlated reactions with the HS (Dataset EV3). Green arrow indicates the significant correlation of the flux carried by the reactions and HS.

4. E

   The net fat influx (NFI) is calculated as the differences in the uptake and secretion rates of FAs and its correlation with the intracellular fluxes are assessed.

Figure 3. Correlation of the HS and plasma metabolomics data

1. A, B

   Hepatic steatosis (HS) is measured by the magnetic resonance imaging, and the plasma level of ˜520 metabolites was detected by untargeted metabolomics profiling. (A) The correlation between the HS and the plasma metabolites and (B) the Pearson correlation between the significantly (P < 0.05) correlated metabolites are presented. Red and blue colors represent the positive and negative correlation of the HS and plasma metabolite levels, respectively. Metabolites that can be converted to glycine marked bold.

Figure 4. Identification of significantly changed metabolites in subjects with high HS

The plasma level of ~520 metabolites was detected by untargeted metabolomics profiling and significantly (P < 0.05) changed metabolites are presented using volcano plot. Metabolites that can be converted to glycine marked bold.

Figure 5. Glycine is the limiting substrate in the synthesis of GSH in NAFLD

1. A

   Glycine is found to be the limiting substrate for the synthesis of glutathione (GSH) in subjects with NAFLD. Decreased de novo synthesis of serine has also been reported in subjects with NASH. The decreased plasma level of glycine, serine as well as other associated metabolites betaine and N‐acetylglycine in subjects with high HS is confirmed with the metabolomics study. In order to confirm model‐based predictions, serine was supplemented to the subjects with high HS since serine‐derived carbon can be converted to GSH to satisfy the increased demand for GSH in NAFLD. Blue arrows indicated downregulation whereas the red arrow indicates upregulation of the reactions. Green arrows indicate the significant correlation of the flux carried by the reactions and HS. [c], cytoplasm; [m], mitochondria.

2. B–E

   The mRNA expressions of the (B) nicotinamide nucleotide transhydrogenase (NNT), (C) glutathione reductase (GSR), (D) glutamate–cysteine ligase, catalytic subunit (GCLC), and (E) glutamate–cysteine ligase, modifier subunit (GCLM) are measured in the livers obtained from 12 morbidly obese subjects who had undergone bariatric surgery and seven healthy individuals (mean ± SD). *P < 0.05; Student's t‐test.

Figure 6. Supplementation of NAD+ and GSH precursors prevent NAFLD

Ten mice were given the Western diet supplemented with NR (400 mg/kg/day) and serine (300 mg/kg/day) la gavage and NAC (1 g/l) in the drinking water, and ten mice were only given the Western diet for 14 days.

1. A–E

   Hepatic lipids including (A) triglycerides (TG), (B) cholesterol esters (CE), (C) ceramides (CER), (D) sphingomyelin (SM), (E) phosphatidylethanolamine (PE) (normalized to phosphatidylcholine (PC)) are shown in treated (cocktail supplemented) (n = 10), and control (n = 10) mice (mean ± SEM).

2. F

   Quantification of amino acids from the liver of the same mice before and after supplementation (mean ± SEM).

3. G

   Analysis of the molecular species of TGs extracted from the livers of the mice. Results from the control group are expressed as 100%, and results from the treated group are expressed as % of the control group (mean ± SEM).

4. H

   mRNA expression of the fatty acid synthase (Fasn) in the liver tissue of mice before and after the supplementation. Six subjects received one oral dose of L‐serine (200 mg/kg/day) for 14 days (mean ± SEM).

5. I–L

   The human plasma (I) alanine aminotransferase (ALT), (J) aspartate aminotransferase (AST), (K) alkaline phosphatase (ALP), and (L) TGs levels are presented in each human subject involved in the study before and after the supplementation with serine.

Data information: (A–G) *P < 0.05, **P < 0.01, ***P <0.001; Student's t‐test.Source data are available online for this figure.

Figure 7. Three‐step strategy for the treatment of NAFLD and associated disorders

Based on our results, we postulate a potential treatment strategy for NAFLD patients based on increased oxidation of fat and increased synthesis of GSH. A cocktail can be supplemented to NAFLD patients to boost these metabolic processes and to decrease the hepatic lipid accumulation. NR can be supplemented to boost the oxidation of the fat in the mitochondria by generating NAD+. Serine and NAC can be included in the cocktail to boost the level of GSH which is required for preventing the accumulation of incomplete products of fatty acids oxidation. l‐carnitine can also be added to the cocktail to boost the fatty acid uptake into mitochondria.