AMP-activated protein kinase activation and NADPH oxidase inhibition by inorganic nitrate and nitrite prevent liver steatosis

Abstract

Advanced age and unhealthy dietary habits contribute to the increasing incidence of obesity and type 2 diabetes. These metabolic disorders, which are often accompanied by oxidative stress and compromised nitric oxide (NO) signaling, increase the risk of adverse cardiovascular complications and development of fatty liver disease. Here, we investigated the therapeutic effects of dietary nitrate, which is found in high levels in green leafy vegetables, on liver steatosis associated with metabolic syndrome. Dietary nitrate fuels a nitrate-nitrite-NO signaling pathway, which prevented many features of metabolic syndrome and liver steatosis that developed in mice fed a high-fat diet, with or without combination with an inhibitor of NOS (l-NAME). These favorable effects of nitrate were absent in germ-free mice, demonstrating the central importance of host microbiota in bioactivation of nitrate. In a human liver cell line (HepG2) and in a validated hepatic 3D model with primary human hepatocyte spheroids, nitrite treatment reduced the degree of metabolically induced steatosis (i.e., high glucose, insulin, and free fatty acids), as well as drug-induced steatosis (i.e., amiodarone). Mechanistically, the salutary metabolic effects of nitrate and nitrite can be ascribed to nitrite-derived formation of NO species and activation of soluble guanylyl cyclase, where xanthine oxidoreductase is proposed to mediate the reduction of nitrite. Boosting this nitrate-nitrite-NO pathway results in attenuation of NADPH oxidase-derived oxidative stress and stimulation of AMP-activated protein kinase and downstream signaling pathways regulating lipogenesis, fatty acid oxidation, and glucose homeostasis. These findings may have implications for novel nutrition-based preventive and therapeutic strategies against liver steatosis associated with metabolic dysfunction.

Keywords: microbiota; nitrate; nitric oxide; nitrite; steatosis.

Fig. 1.

Cardiovascular and metabolic phenotype. (A) After 5 wk of dietary treatment with an HFD in combination with the NOS inhibitor l-NAME, mice were subjected to an i.p. glucose tolerance test (IPGTT) (2 g/kg body weight), and time-course changes in glucose levels were measured. (B) Area under the curve (AUC) calculated from IPGTT data. (C) Intraperitoneal insulin tolerance tests (IPITTs) were performed after 6 wk of treatment (0.75 IU/kg body weight), and glucose levels were followed for 2 h. (D) Inverse AUC calculated from IPITT data. (E) Mean arterial blood pressure was recorded by the noninvasive tail monitoring system at week 6 of treatment. (F) At killing (i.e., 7 wk of treatment), fresh mesenteric arteries from the mice were isolated, and reactivity to increasing concentrations of acetylcholine was measured. Data are presented as the mean ± SEM. n = 6 to 10 mice per group. *, **, ***, **** denote P < 0.05, P < 0.01, P < 0.001, and P < 0.0001, respectively, between indicated groups. An “a” denotes P < 0.05 between Control vs. HFD, “b” denotes P < 0.05 between Control vs. HFD+Nitrate, and “c” denotes P < 0.05 between HFD vs. HFD+Nitrate. Tests were performed by two-way repeated measures (RM) ANOVA (A, C, and F) or one-way ANOVA followed by Holm–Sidak test (B, E, and D). IPGTT, i.p. glucose tolerance test; IPITT, i.p. insulin tolerance test; HFD, high-fat diet+l-NAME.

Fig. 2.

Nitrate, nitrite, and iron-nitrosylated species levels in blood and liver. In tissues obtained after 7 wk of dietary treatment with an HFD in combination with the NOS inhibitor l-NAME, (A) plasma nitrate, (B) plasma nitrite, (C) liver nitrate, (D) liver nitrite, (E) blood iron-nitrosylated-hemoglobin, and (F) liver dinitrosyl iron complexes (DNICs) were measured for the three animal groups. Data are presented as the mean ± SEM. n = 6 to 9 per group. *, **, and *** denote P < 0.05, P < 0.01, and P < 0.001, respectively, between indicated groups tested by Kruskal–Wallis test and Dunn’s test (A–E). HFD, high-fat diet+l-NAME.

Fig. 3.

Oil-Red O staining and NADPH oxidase activity in liver. (A) Representative images of liver sections from the three mice groups at the moment of killing stained with Oil Red O to visualize neutral lipids and Mayer´s hematoxylin to visualize the cell nuclei and tissue morphology. (Scale bars: 1 µm.) Insets show four times magnified details of the images to highlight the lipid-staining morphology. (B) Quantification of neutral lipids in liver sections using ImageJ software. (C) NADPH oxidase-derived superoxide and hydrogen peroxide production was measured by Amplex red and expressed as relative light units (RLUs), in the liver of the three groups of mice at killing. (D and E) p67phox and NOX2 (gp91phox) protein levels in the liver were determined by Western blot. Densitometric quantification is presented as p67phox/vinculin and NOX2/vinculin. (F) Correlation between Oil Red O staining and NADPH oxidase activity in the liver. Data are presented as the mean ± SEM. n = 5 to 9 per group. * and ** denote P < 0.05 and P < 0.01, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test (B, C, and E) or Pearson correlation test (D). HFD, high-fat diet+l-NAME.

Fig. 4.

Oil-Red O staining in livers of germ-free mice. (A) Representative images of liver sections from the three groups of germ-free mice at killing, stained with Oil Red O to visualize neutral lipids and Mayer´s hematoxylin to visualize the cell nuclei and tissue morphology. (Scale bars: 5 µm.) (B) Quantification of neutral lipids in liver sections was done using ImageJ software. Data are presented as the mean ± SEM. n = 5 to 6 per group. * and ** denote P < 0.05 and P < 0.01, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test. HFD, high-fat diet+l-NAME.

Fig. 5.

Liver AMPK and Akt expression. After 7 wk of dietary treatment, mice were killed, and livers from the animals were individually homogenized. The protein levels of p-AMPK/AMPK (A) and p-Akt/Akt (B) were measured by Western blot, and densitometric quantification is presented as the ratio of p-AMPK/AMPK or p-Akt/Akt. Data are presented as the mean ± SEM. n = 9 to 10 (A and B) per group. * and ** denote P < 0.05 and P < 0.01, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test. HFD, high-fat diet+l-NAME.

Fig. 6.

Expression of metabolic regulatory proteins in the liver. After 7 wk of dietary treatment, mice were killed, and livers from the animals were individually homogenized. The mRNA expression and protein levels of (A and B) SREBP1c, (C and D) ACC, (E and F) PGC1α, and (G and H) ACADM were measure in the three groups of animals. mRNA expression levels were determined by real-time PCR and protein levels by Western blot. Densitometric quantification is presented as SREBP1c/vinculin, p-ACC/vinculin, PGC1α/vinculin, and ACADM/vinculin. Data are presented as the mean ± SEM. n = 6 to 8 (A, C, E, and G) and n = 7 to 8 (B), n = 8 to 10 (D), n = 5 to 6 (F), n = 7 to 8 (H) per group. * and ** denote P < 0.05 and P < 0.01, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test. ACADM, medium chain acyl-CoA dehydrogenase; ACC, acetyl-CoA carboxylase; HFD, high-fat diet+l-NAME; p-ACC, phospho-ACC; PGC1α, peroxisome proliferator-activated receptor γ coactivator 1 alpha; SREBP1c, transcription factor sterol regulatory element-binding protein 1c.

Fig. 7.

Oil Red O staining and NADPH oxidase activity in HepG2 cells. Cells were treated with control, steatosis, or steatosis + nitrite media for 24 h. (A) Representative images of HepG2 cells from the three groups stained with Oil Red O to visualize neutral lipids. (Magnification: 40×.) (B) Semiquantification of neutral lipids in the HepG2 cells measuring absorbance at 520 nm. Each dot represents an individual value. (C) NADPH oxidase-derived superoxide production in the HepG2 cells, and expressed as relative light units (RLUs). Each dot represents an individual value. To assess the role of oxidative stress in the development of steatosis, the cells were simultaneously incubated for 24 h with the superoxide scavenger tempol (D and E) or the NADPH oxidase inhibitor GLX481304 (F and G). Then, neutral lipids accumulation was measured by Oil Red O (D and F), and NADPH oxidase-derived superoxide production was measured with a lucigenin-dependent chemiluminescence signal (E and G). Data are presented as the mean ± SEM. n = 6 (A), n = 11 to 17 (B), n = 26 to 50 (C), n = 6 to 8 (D and E), n = 4 to 16 (F and G) per group. *, **, ***, and **** denote P < 0.05, P < 0.01, P < 0.001, and P < 0.0001, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test. Control, 5.5 mM glucose; Steatosis, 25 mM glucose, 10 nM insulin, and 240 µM FFA (steatosis mix); Steatosis + nitrite, steatosis mix + 10 µM sodium nitrite; Steatosis + tempol, steatosis mix + 100 µM tempol; Steatosis + GLX = steatosis mix + 2 to 50 µM GLX481304.

Fig. 8.

Assessment of superoxide production in HepG2 cells by EPR. Cells were treated with control, steatosis mixture, or steatosis + nitrite for 48 h. (A and B) Quantification of second-peak amplitudes of DMPO/^•OH adducts from EPR spectra in the membrane fraction of HepG2 cells. Nitrite treatment prevented steatosis-induced increase of O₂^•− production (A), similar to that achieved with the NOX2/NOX4 inhibitor GLX481304 (B). Coincubation with SOD (400 U/mL) completely blunted EPR signals, confirming that DMPO/^•OH adducts quantified are from O₂^•− (DMPO/^•OOH) and not from the hydroxyl (^•OH) radical. (C) Representative EPR spectrums from A and B experiments. Data are presented as the mean ± SEM. n = 4 to 11 per group. * and **** denote P < 0.05 and P < 0.0001, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test.

Fig. 9.

Regulation of AMPK and NADPH oxidase activity in HepG2 cells. Cells were treated with control, steatosis, or steatosis + nitrite media for 24 h. The protein levels of p-AMPK/AMPK were measured by Western blot. Densitometric quantification is presented as the ratio p-AMPK/AMPK (A). Cells were treated with control, steatosis, steatosis + nitrite, steatosis + compound C (an AMPK inhibitor), or steatosis + compound C + nitrite media for 24 h, and NADPH oxidase-derived superoxide formation was measured with lucigenin-dependent chemiluminescence signal in the HepG2 cells, and expressed as relative light units (RLUs) (B). Using similar conditions, the effects of an NO donor (DETA-NONOate) and a cGMP analog (8-pCPT-cGMP) on p-AMPK/AMPK ratio were assessed (C) during steatosis. Data are presented as the mean ± SEM. n = 4 (A) and n = 5 to 60 (B) per group. *, **, and **** denote P < 0.05, P < 0.01, and P < 0.0001, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test. Control, 5.5 mM glucose; Steatosis, 25 mM glucose, 10 nM insulin, and 240 µM FFA (steatosis mix); Steatosis + nitrite, steatosis mix+10 µM sodium nitrite; Steatosis + Compound C, steatosis + 20 µM Compound C; Steatosis + Compound C + nitrite, steatosis mix + 20 µM Compound C + 10 µM sodium nitrite.

Fig. 10.

Mechanisms contributing to nitrite-mediated protection against steatosis in HepG2 cells. Cells were treated with control, steatosis, steatosis + nitrite, steatosis + DETA-NONOate (a slow releasing NO-donor), steatosis + cGMP analog (8-pCPT-cGMP), steatosis + Febuxostat (a specific inhibitor of XOR), or steatosis + ODQ (an inhibitor of soluble guanylyl cyclase) media for 24 h. (A) Accumulation of neutral lipids was measured by Oil Red O, and (B) NADPH oxidase-derived superoxide formation was measured with lucigenin-dependent chemiluminescence signal, and expressed as relative light units (RLUs). Data are presented as the mean ± SEM. n = 5 to 32 (A) and n = 5 to 60 (B) per group. *, **, ***, and **** denote P < 0.05, P < 0.01, P < 0.001, and P < 0.0001, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test. Control, 5.5 mM glucose; Steatosis, 25 mM glucose, 10 nM insulin, and 240 µM FFA (steatosis mix); Steatosis + nitrite, steatosis mix + 10 µM sodium nitrite; Steatosis + DETA-NONOate, steatosis mix+ 5 µM DETA-NONOate; Steatosis + cGMP analog, steatosis mix + 10 µM cGMP analog; Steatosis + Febuxostat, steatosis mix + 50 nM Febuxostat; Steatosis + ODQ, steatosis mix + 10 µM ODQ.

Fig. 11.

Affirmation of nitrite-mediated NO production in HepG2 cells by EPR. Changes in first-peak amplitudes from EPR spectrums, together with representative spectrums, for NO production using Fe(DETC)₂ as spin trap. HepG2 cells were incubated with vehicle or Febuxostat (50 nM) for 20 min and then treated with nitrite ¹⁴N (500 µM) or nitrite ¹⁵N (500 µM) for 1 h in the presence of 0.5 mM colloid Fe(DETC)₂ for NO detection. Shown are quantified EPR data for nitrite ¹⁴N (A and B) and nitrite ¹⁵N (C and D) together with representative spectrums for the different experiments. Data are presented as the mean ± SEM. n = 3 to 4 per group. ** and **** denote P < 0.01 and P < 0.0001, respectively, between indicated groups, tested by Kruskal–Wallis test and Dunn’s test.

Fig. 12.

Reversibility of metabolic steatosis in primary human hepatocyte spheroids. Spheroids were exposed to higher concentration of FFA (320 µM) in the medium for 1 wk to induce steatosis. Thereafter, the metabolic steatotic stimuli were removed, and the spheroids were treated with or without increasing concentrations of nitrite (0.2 to 100 µM) for 7 d. The spheroids were stained, and the degree of steatosis (i.e., lipid content) was evaluated by the AdipoRed fluorescence technique. Nitrite improved reversibility of metabolic-induced steatosis in a dose-dependent way (A and B), being significant compared with the steatosis group at concentrations ≥0.6 µM. Representative images are shown for the three groups (nitrite, 10 µM) (staining with Nile Red staining) (A). Simultaneous treatment with febuxostat completely inhibited the favorable effects of nitrite on lipid accumulation (C). Similar to that observed in vivo and in HepG2 cells, steatosis was associated with increased NADPH oxidase activity, measured as changes in relative light units (RLUs) (D) and increased expression of its regulatory subunit p67phox (E), which were both attenuated by nitrite (10 µM) treatment. Data are presented as the mean ± SEM. n = 6 to 16 (B), n = 8 to 12 (C), n = 8 (D), n = 3 to 5 (E) per group. All data were obtained from one donor. *, **, and *** denote P < 0.05, P < 0.01, and P < 0.001, respectively, between indicated groups. Tests performed by one-way ANOVA followed by Holm–Sidak test. (Magnification: 10×.)

Fig. 13.

Drug-induced steatosis in primary human hepatocyte spheroids. Spheroids were exposed to amiodarone (5 µM) in the medium for 14 d to induce steatosis with or without increasing concentrations of nitrite (1 to 100 µM). The spheroids were stained, and the degree of steatosis (i.e., lipid content) was evaluated by the AdipoRed fluorescence technique. Nitrite significantly improved drug-induced steatosis (A and B) at concentrations ≥1 µM. Representative images are shown for the three groups (nitrite, 10 µM) (A). Data are presented as the mean ± SEM. n = 8 to 20 per group. ** and **** denote P < 0.01 and P < 0.0001, respectively, between indicated groups. Tests were performed by one-way ANOVA followed by Holm–Sidak test. (Magnification: 10×.)

Fig. 14.

Metabolic signaling pathways in the liver affected by dietary nitrate. Shown is a simplified scheme illustrating some of the key pathways affected by dietary nitrate to attenuate diet-induced liver steatosis. A high-fat diet (HFD) leads to down-regulation of AMPK activity in at least two ways. First, the surplus of energy will down-regulate AMPK via direct mechanisms involving changes in cellular energy status (AMP/ADP to ATP ratio). Moreover, an HFD leads to activation of NADPH oxidase (NOX) activity, and the resulting oxidative stress can inhibit AMPK activation. Dietary nitrate can counteract this process via formation of nitrite, nitric oxide (NO), and other bioactive nitrogen oxide species. Bioactivation of nitrate depends on commensal bacteria and nitrite reductase activity [e.g., xanthine oxidoreductase (XOR)], respectively. NO serves to down-regulate NOX activity, at least in part via reduction of p67phox expression, and can also activate AMPK in a cGMP-dependent way. The reversal of the HFD-induced AMPK depression by nitrate ultimately affects a number of downstream pathways (i.e., inhibition of SREBP1c, stimulation of ACC phosphorylation, and by restoration of ACADM and PGC1α) to reduce lipid accumulation and steatosis.