TXNIP in liver sinusoidal endothelial cells ameliorates alcohol-associated liver disease via nitric oxide production

Abstract

Dysregulation of liver sinusoidal endothelial cell (LSEC) differentiation and function has been reported in alcohol-associated liver disease (ALD). Impaired nitric oxide (NO) production stimulates LSEC capillarization and dysfunction; however, the mechanism underlying NO production remains unclear. Here, we investigated the role of thioredoxin-interacting protein (TXNIP), an important regulator of redox homeostasis, in endothelial cell NO production and its subsequent effects on ALD progression. We found that hepatic TXNIP expression was upregulated in patients with ALD and in ethanol diet-fed mice with high expression in LSECs. Endothelial cell-specific Txnip deficiency (Txnip^ΔEC) in mice exacerbated alcohol-induced liver injury, inflammation, fibrosis, and hepatocellular carcinoma development. Deletion of Txnip in LSECs led to sinusoidal capillarization, downregulation of NO production, and increased release of proinflammatory cytokines and adhesion molecules, whereas TXNIP overexpression had the opposite effects. Mechanistically, TXNIP interacted with transforming growth factor β-activated kinase 1 (TAK1) and subsequently suppressed the TAK1 pathway. Inhibition of TAK1 activation restored NO production and decreased the levels of proinflammatory cytokines, thereby, blocking liver injury and inflammation in Txnip^ΔEC mice. Our findings indicate that upregulated TXNIP expression in LSECs serves a protective role in ameliorating ALD. Enhancing TXNIP expression could, therefore, be a potential therapeutic approach for ALD.

Keywords: NO; TAK1; TXNIP; alcohol-associated liver disease; eNOS; liver sinusoidal endothelial cells.

Figure 1

TXNIP expression is increased in the liver tissue of patients with ALD and ethanol diet-fed mice. (A) Representative image of immunohistochemical staining of TXNIP in liver sections from healthy controls (n = 13), patients with alcohol-associated steatosis (n = 19), and patients with alcohol-associated fibrosis (n = 18). Original magnification, ×200. The integrated optical density (IOD) of TXNIP-positive areas is shown on the right. (B) Association between TXNIP expression and macro-fat or alcohol consumption in ALD patients with steatosis. Alcohol consumption was defined as follows: Heavy (grade 4) > Moderate (grade 3) > Social (grade 2) > Occasional (grade 1) > None (grade 0). (C) Correlation between TXNIP expression and fibrosis grade or alcohol consumption in ALD patients with fibrosis. (D) Representative immunohistochemical images of TXNIP expression in mice fed a control or 4% ethanol diet for 1 year (1y EtOH, n = 4). Scale bars, 50 μM. The IOD of TXNIP-positive areas is shown on the right. (E) Hepatic expression of Txnip mRNA in mice fed a control or 4% ethanol diet for 1 year (1y EtOH, n = 4). Data are presented as means ± SD. *p < 0.05; **p < 0.01. “ns” stands for “not significant.”

Figure 2

Deficiency of Txnip in LSECs increases ethanol-induced liver diseases. (A) Relative mRNA levels of Txnip. Hepatocytes, KCs, LSECs, and HSCs were isolated from WT mice after 10 day plus gavage (10D+1B EtOH; pair-fed, n = 4; ethanol [EtOH], n = 6). (B) Immunofluorescence staining of WT mice. Note the colocalization of TXNIP and LYVE-1 (arrows). Original magnification, ×400. (C-G) Txnip^fl/fl and Txnip^ΔEC mice were fed a liquid control diet or 4% ethanol diet for 1 year (1y EtOH). Representative gross findings and H&E staining of liver sections (C), incidence and maximum size of tumors (D), serum levels of ALT and AST (E) (Txnip^fl/fl, n = 16; Txnip^ΔEC, n =5), and qRT-PCR analysis (F-G) (pair-fed, n = 3; EtOH, n = 4). Note tumor nodule (arrows) compressing adjacent normal liver tissue in Txnip^ΔEC mice. Scale bars, 200 μM. Data are presented as means ± SD. *p < 0.05; **p < 0.01. “ns” stands for “not significant.”

Figure 3

TXNIP regulates LSEC capillarization, NO production, and proinflammatory function. (A) GSEA of RNA-seq data. RNA samples were collected from LSECs isolated from Txnip^-/- and WT mice (n = 3). The 20 most significantly enriched pathways are shown. (B) Heat maps showing changes in the expression of mRNAs related to LSEC differentiation (n = 3). The p values for the comparisons are indicated in color. (C) Relative mRNA levels of LSECs and capillary EC markers (n = 4). LSECs isolated from Txnip^-/- and WT mice treated with LPS (0.75 μg/ml) for 6 h. (D-F) LSECs isolated from Txnip^-/- and WT mice were incubated with LPS (0.75 μg/ml) for 6 h and subjected to western blotting (D) (n = 4), qRT-PCR analysis (E) (n = 4), and NO assay (F) (n = 3). (G) Heat maps showing changes in the expression of mRNAs related to inflammation (n = 3). The p values for the comparisons are indicated in color. (H) Relative mRNA levels of genes related to inflammation. LSECs isolated from Txnip^-/- and WT mice treated with LPS (0.75 μg/ml) for 6 h. Data are presented as means ± SD. *p < 0.05; **p < 0.01. “ns” stands for “not significant.”

Figure 4

Endothelial TXNIP overexpression blocks LSEC capillarization and increases NO production. (A) Western blot analysis of TXNIP in control (Ctrl) and TXNIP-overexpressing (TXNIP OE) TMNK-1 cells. (B) Relative mRNA levels of LSECs and capillary EC markers (n = 4 biological replicates). Ctrl and TXNIP-OE TMNK-1 cells were incubated with saline or LPS (0.75 μg/ml) for 24 h. (C-D) Ctrl and TXNIP-OE TMNK-1 cells were incubated with saline or LPS (0.75 μg/ml) for 24 h and subjected to western blotting (C) and NO assay (D) (n = 3 biological replicates). Data are presented as means ± SD. *p < 0.05; **p < 0.01.

Figure 5

TXNIP inhibits TAK1/JNK signaling in LSECs. (A) Volcano map showing DEGs in LSECs isolated from Txnip^-/- and WT mice (n = 3). (B) The top-10 most significantly enriched pathways contributing to TXNIP function based on the KEGG enrichment analysis. (C) Protein expression of TAK1 and JNK. LSECs isolated from Txnip^-/- and WT mice were incubated with LPS (0.75 μg/ml) for 30 min (n = 4). (D) Western blot analysis of the liver tissue (pair-fed, n = 3; ethanol [EtOH], n = 4). Txnip^fl/fl and Txnip^ΔEC mice were fed a control or 4% ethanol diet for 1 year (1y EtOH). (E) Western blot analysis (n = 4 biological replicates). Control (Ctrl) and TXNIP-overexpressing (TXNIP OE) TMNK-1 cells were incubated with LPS (0.75 μg/ml) for 30 min. (F) Immunofluorescence staining of TXNIP and TAK1 in TMNK-1 cells. Original magnification, ×400. Scale bars, 20 μM. (G) Representative coimmunoprecipitation analysis of TXNIP and TAK1 in TMNK-1 cells. Data are presented as means ± SD. *p < 0.05; **p < 0.01. “ns” stands for “not significant.”

Figure 6

TAK1 inhibition restores eNOS levels and ameliorates ethanol-induced liver injury, oxidative stress, and inflammation caused by Txnip deficiency in LSECs. (A-C) LSECs isolated from Txnip^-/- and WT mice were incubated with NG25 (500 nM) for 3 h, incubated with LPS (0.75 μg/ml) for 0.5 or 6 h, and subjected to western blot analysis (A), qRT-PCR analysis (B and C), and NO assay (B) (n = 4). (D-G) Txnip^fl/fl and Txnip^ΔEC mice were fed a 5% ethanol diet with NG25 (5 mg/kg/day) for 4 weeks (4w EtOH) and subjected to H&E staining (D), blood chemistry analysis (E), oxidative stress analysis (F), and qRT-PCR analysis (G) (n = 5). Scale bars, 50 μM. Data are presented as means ± SD. *p < 0.05; **p < 0.01. “ns” stands for “not significant.”

Figure 7

Positive correlation between hepatic TXNIP expression and LSEC capillarization in patients with ALD. (A) Representative image of immunohistochemical staining for LYVE-1 and CD31 in liver sections from healthy controls (n = 13) and in patients with alcohol-associated fibrosis (n = 18). Scale bars, 50 μM. (B) LYVE-1- and CD31-positive areas. (C) Negative correlation of hepatic LYVE-1 expression and fibrosis grade or alcohol consumption. (D) Positive association of hepatic CD31 expression and fibrosis grade or alcohol consumption. Alcohol consumption was defined as follows: Heavy (grade 4) > Moderate (grade 3) > Social (grade 2) > Occasional (grade 1) > None (grade 0). (E) Immunofluorescence staining of liver sections. Note the colocalization of TXNIP and LYVE-1 or CD31 (arrows). Original magnification, ×400. (F) Immunohistochemical images of TXNIP expression. Note the TXNIP-positive LSECs (arrows) in the human liver tissue. Scale bars, 30 μM. (G) Correlation of TXNIP expression and LYVE-1- or CD31-positive area in healthy controls and ALD patients. Data are presented as means ± SD. *p < 0.05; **p < 0.01. “ns” stands for “not significant.”