Hepatocyte-specific HuR deficiency enhances WDSW-induced NAFLD

Although two recent studies have reported the hepatic-specific role of HuR modulating lipid metabolism in mouse NAFLD models, both studies used liver-specific HuR knockout mice by cross-breeding a HuRflox/flox mouse with albumin-Cre mice, not hepatocyte-specific knockout mice [20, 22]. To delineate the hepatocyte-specific role of HuR in the NAFLD disease progression, HuRhKO mice were generated by tail-vein injection of HuRflox/flox mice with AAV8-TBGP-Cre recombinase and AAV8-TBGP-GFP was used as a control. HuRhKO and control mice were fed ad libitum a WDSW for 4 weeks. As shown in Additional file 1: Fig.S1a–c, the HuR mRNA and protein levels were significantly reduced in the liver of HuRhKO mice. Immunofluorescence staining further confirmed the deletion of HuR in hepatocytes (Additional file 1: Fig. S1d). As shown in Fig. 1a, b, hepatocyte-specific deletion of HuR exacerbated WDSW-induced hepatic lipid accumulation and liver injury following 4-week feeding as indicated by increased lipid accumulation (H&E and Oil Red O staining), and increased serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels. Although the liver index (ratio of liver/body weight), serum cholesterol, triglycerides, glucose, and alkaline phosphatase (ALP) levels were not significantly altered (Additional file 1: Fig. S1e–g), hepatic levels of triglycerides and cholesterol in HuRhKO mice were much higher than those in control mice (Fig. 1c). However, no significant change was noticed in the serum ALT and AST in HuRhKO and control mice under a normal diet (Additional file 1: Fig. S1h, i), which was consistent with the H&E staining (Additional file 1: Fig. S1j).

Fig. 1

Hepatocyte-specific HuR deficiency enhances WDSW-induced NAFLD. Age and gender-matched HuRhKO and control mice were fed ad libitum a high-fat diet with 42% kcal from fat and containing 0.1% cholesterol plus a high fructose-glucose solution (23.1 g/L d-fructose + 18.9 g/L d-glucose) (Western diet plus sugar water, WDSW) for 4 weeks. a Representative image of the liver and the liver sections stained with H&E and Oil red O staining (scale bar, 100 μm for 10 × magnification). b Serum liver enzyme levels (AST, ALT). c Hepatic triglyceride (TG) and total cholesterol (TC). d Bile acid composition profile in the serum is expressed by % of total bile acids. e Total bile acids (BA), total primary BA, TCA, and total conjugated BA in the serum. Data are expressed as mean ± SEM. Statistical significance: *p < 0.05 vs Control, n = 10–12 (both male and female)

Previous studies have reported that NALFD disease severity is associated with specific changes in circulating bile acids in human NASH patients and mouse NASH models. Serum taurocholic acid (TCA) level is significantly increased in NASH patients [23, 24]. To examine whether the serum bile acid profile was changed in HuRhKO mice after 4-week WDSW feeding, serum bile acid composition and levels were measured using LC–MS/MS [24]. As shown in Fig. 1d, the percentage of TCA in total bile acids was significantly increased from 10% (Control) to 23% (HuRhKO), while βMCA was significantly decreased from 26% (Control) to 15% (HuRhKO) after 4-week WDSW feeding. Total serum bile acid levels were significantly increased in HuRhKO mice; including, total primary, secondary, and conjugated bile acids (Fig. 1e and Additional file 1: Table S2). In addition, serum levels of TCA, TβMCA, TCDCA, TωMCA, and THDCA in HuRhKO mice were much higher than those in control mice (Additional file 1: Table S3).

Hepatocyte-specific HuR deficiency enhances WDSW-induced NAFLD by modulating the global transcriptomic profile

To illustrate the underlying mechanisms by which hepatic HuR deficiency-induced NAFLD disease progression, we performed RNA-seq transcriptome analysis. As shown in Fig. 2a, b, WDSW-feeding induced the upregulation of 192 genes and down-regulation of 160 genes in HuRhKO mice compared to the control mice. Gene Ontology analysis showed that WDSW-feeding significantly impacted the major pathways in biological process (BP), cellular component (CC), and molecular function (MF) related to metabolic processes and immunological responses (Additional file 1: Fig. S2a–c) in HuRhKO mice compared to the control mice, including cholesterol metabolic process, biosynthetic process, inflammatory response, extracellular exosome, lipoprotein, extracellular matrix, oxidoreductase activity, fatty acid-binding, etc. Furthermore, KEGG pathways analysis showed that metabolic pathways, including steroid hormone biosynthesis and inflammatory response, were dysregulated in HuRhKO mice (Additional file 1: Fig. S2d).

Fig. 2

Hepatocyte-specific HuR deficiency enhances WDSW-induced NAFLD by modulating the global transcriptomic profile. Total liver RNA of HuRhKO and control mice fed with WDSW for 4 weeks was processed for transcriptome sequencing (RNA-Seq). Significantly up-or down-regulated genes were determined as fold change ≥ 1.8 and p-value < 0.05. a Volcano plots in HuRhKOvs. control group. Red dots indicate upregulated genes; green dots indicate downregulated genes; black dots indicate not differentially expressed genes. b Hierarchical clustering heatmaps for differentially expressed genes in HuRhKOvs. control group. A Z-score was calculated for the RNA-Seq data to normalize tag counts. Red and green colors indicate up- and down-regulated gene expression, respectively. c Representative heatmap of the key genes involved in fatty acid and lipid metabolism in the liver of HuRhKOvs. control group. d Relative mRNA levels of the key genes involved in fatty acid and lipid metabolism: Acc1, Fasn, Elovl6, Fads1, Fads2, Lxrα, Lxrβ, Pparα, Pparβ, Cpt1α, Pnpla3, and Pnpla2 (Atgl). The mRNA levels were determined by real-time RT–PCR and normalized with HPRT1 as an internal control. Data are expressed as the mean ± SEM. Statistical significance relative to control: * p < 0.05, ** p < 0.01, *** p < 0.001 (n = 10–12, male and female)

Hepatocyte-specific HuR deficiency enhances WDSW-induced dysregulation of fatty acid and lipid metabolism

Dysregulation of lipid metabolism contributes to the development of NAFLD [25]. As shown in Fig. 2c, most of the genes involved in the fatty acid biosynthesis pathway were increased in HuRhKO mice compared to the control mice after 4 weeks of WDSW feeding; including acetyl CoA carboxylase (Acc1), fatty acid synthase (Fasn), elongation of very-long-chain fatty acids member 6 (Elovl6), fatty acid desaturases (Fads1&2), liver X receptor (Lxrα&β), peroxisome proliferator-activated receptor (Pparα&β), carnitine palmitoyltransferase 1 (Cpt1α), patatin-like phospholipase domain containing (Pnpla3), and adipose triglyceride lipase (Atgl, also known as Pnpla2), etc. The mRNA expression levels of key genes involved in hepatic lipid metabolism were further confirmed by real-time RT–PCR. As shown in Fig. 2d, the mRNA levels of Acc1, Fasn, Elovl6, Fads1, Fads2, Lxrα, Lxrβ, Pparα, Pparβ, Cpt1α, Pnpla3, and Atgl were significantly increased in HuRhKO mice compared to the control mice after 4-weeks of WDSW feeding.

Hepatocyte-specific HuR deficiency enhances WDSW-induced inflammation and oxidative stress

Inflammation and stress response are important driving forces in promoting NAFL to NASH progression [26,27,28]. RNA-seq analysis showed that the key genes involved in inflammatory and stress responses, such as F4/80, Cd68, Cd63, C-X-C Motif Chemokine Ligand 1 (Cxcl1), Cxcl10, chemokine ligand 2 (Ccl2), C–C chemokine receptor type 2 (Ccr2), Caspase 3, Activating transcription factor 4 (Atf4), interleukin 1α (IL-1α), were significantly increased in HuRhKO mice fed with a WDSW for 4 weeks (Additional file 1: Fig. S3a). Hepatocyte-specific deletion of HuR promoted WDSW-induced macrophage infiltration to the liver as indicated by the IHC staining of F4/80 antigen; a mature cell surface glycoprotein expressed at high levels on various macrophages (Fig. 3a). Real-time PCR results further showed that the mRNA expression levels of major marker genes of macrophages, inflammatory cytokines and chemokines, such as F4/80, Cd68, Cd63, Integrin alpha M (also known as Cd11b) (Fig. 3b), Cxcl1, Cxcl10, Ccl2, Ccr2, IL-1α, IL-1β, tumor necrosis factor α (Tnfα), and IL-6, were significantly increased (Fig. 3c).

Fig. 3

Hepatocyte-specific HuR deficiency enhances WDSW-induced inflammation and oxidative stress. a Representative images of immunohistochemistry (IHC) staining of F4/80 (scale bar, 100 μm for 10 × and 20 μm for 40 × magnification). Relative mRNA levels of genes involved in inflammation and oxidative stress were determined by real-time RT–PCR and normalized with HPRT1 as an internal control. b Macrophage markers: F4/80, Cd68, Cd63, and Cd11b; c Chemokines: Cxcl1, Cxcl10, Ccl2, and Ccr2; Inflammatory cytokines: IL1α, IL1β, Tnfα, and IL-6; d Neutrophil activation: Ncf4, Cybα, Cybβ, Ncf2, IL2rg, and Vcam1; Stress: Caspase 3 and Atf4. Data are expressed as the mean ± SEM. Statistical significance relative to control: * p < 0.05, ** p < 0.01, *** p < 0.001 (n = 12, male and female)

Neutrophils, the most abundant leukocytes in circulation, play a vital role in innate immunity [29]. However, inappropriate activation of neutrophils can cause tissue damage, which has been involved in different diseases, including various liver diseases [30, 31]. Our recent study reported that in the WDSW-induced NASH mouse model, the expression levels of major genes involved in neutrophil activation were significantly upregulated in the liver [24]. As shown in Additional file 1: Fig. S3b, the key genes involved in neutrophil activation, such as NADPH oxidase 2 [Nox2, also known as neutrophil cytochrome b heavy chain (Cybβ), or p91phox], neutrophil cytosolic factor 2 (Ncf2, also known as p67phox), Ncf4 (also known as p40phox), Cybα (also known as p22phox), IL-2 receptor gamma unit (IL2rg), intercellular adhesion molecule 1 (ICAM1) and vascular cell adhesion molecule 1 (vcam1), were significantly upregulated in the liver of WDSW-fed HuRhKO mice. The upregulation of mRNA levels was further confirmed by real-time RT–PCR (Fig. 3d). Pathway analysis further showed that oxidative phosphorylation in mitochondria and MAPK signaling pathways were significantly dysregulated in WDSW-fed HuRhKO mice compared to control mice (Additional file 1: Fig.S4 and S5).

HuR is a suppressor of lncRNA H19 transcription

We and others have previously reported that aberrant expression of lncRNA H19 is closely associated with hepatic inflammation and liver fibrosis in various liver diseases, including NASH [12,13,14]. To further identify the potential underlying mechanisms by which hepatic deletion of HuR promotes NAFLD progression, we examined H19 expression in the livers of human NASH patients and WDSW-induced NASH mouse models. As shown in Fig. 4a, hepatic H19 mRNA levels were increased more than tenfold in human NASH patients compared to healthy controls. Similarly, in NASH mice fed WDSW for 21 weeks, hepatic H19 mRNA levels were increased more than 30-fold compared to control mice (Fig. 4b). Analysis of the publically available RNAseq data set from a most recent study (GSE143358) showed that H19 is the most significantly upregulated gene in liver-specific HuR knockout mice (Fig. 4c) [20]. In a newly developed preclinical NASH–HCC model with C57/BL6NJ mice, H19 is in the top ten upregulated genes (GSE197884) (Fig. 4c) [32]. We also found that H19 was significantly upregulated in the liver of HuRhKO mice fed with WDSW for 4 weeks (Fig. 4d).

Fig. 4

HuR is a suppressor of lncRNA H19 transcription. a Relative mRNA levels of lncRNA H19 in the liver of NASH patients and healthy controls were determined by real-time RT–PCR. Statistical significance relative to healthy control, *p < 0.05 (n = 8–10). b Relative mRNA levels of H19 in the liver of mice fed a WDSW or control diet for 21 weeks. Statistical significance relative to control mice, *p < 0.05 (n = 5). c H19 count from the publically available RNAseq data set (GSE143358) in liver-specific HuR knockout mice; H19 count from the most recent study (GSE197884) in a newly developed preclinical NASH–HCC model with C57/BL6NJ mice. Statistical significance relative to control mice, **p < 0.01 and ***p < 0.001 (n = 3). d Relative mRNA levels of H19 in the liver of HuRhKO and control mice fed with WDSW for 4 weeks. Statistical significance relative to control mice, ***p < 0.001 (n = 10–12). e Effect of HuR on H19 promoter activity. HEK-293 cells were transfected with pGL3-human H19-promoter or pGL3-control vector together with pcDNA3-TAP-human HuR or pcDNA3-TAP vector along with Renilla control vector. After transfection for 48 h, the luciferase activities were measured using a dual-luciferase reporter assay kit. The promoter activity was expressed by relative luciferase activity using the ratio of firefly luciferase activity to Renilla luciferase activity. Statistical significance relative to control, ***p < 0.001 (n = 5). f Representative image of the liver and the liver sections stained with H&E and Oil red O staining (scale bar, 50 µm for 20 × magnification) in H19−/− and WT mice fed with WDSW for 4 weeks

To identify the mechanism underlying HuR deficiency-induced H19 expression, a luciferase reporter assay with a human H19 promoter was performed to examine the impact of HuR on H19 transcriptional activity. As shown in Fig. 4e, overexpression of HuR significantly inhibited H19 promoter activity. To further determine the role of H19 in hepatic steatosis and inflammation in vivo, H19−/− mice and gender and age-matched WT mice were fed a WDSW for 4 weeks. As shown in Fig. 4f, H&E and Oil red O staining indicated that H19−/− mice were protected from WDSW-induced hepatic lipid accumulation.

HuR regulates the expression of SphK2 and S1PR2

Previous studies have reported that SphK2, a key enzyme in sphingolipid catabolism, plays a critical role in regulating hepatic lipid metabolism [8,9,10]. Global SphK2 deficient (SphK2−/−) mice developed overt fatty liver compared to the control mice on a two-week high-fat diet [10]. SphK2 is primarily located in the cell nucleus. As shown in Fig. 5a, the nuclear protein levels of SphK2 were significantly decreased in the livers of WDSW-fed HuRhKO mice compared to control mice. Deletion of H19 increased SphK2 nuclear protein level (Fig. 5b). However, neither hepatocyte HuR deficiency nor deletion of H19 significantly altered the total protein levels of SphK2 in the livers (Additional file 1: Fig. S6a, b). Our previous studies also reported that S1PR2 deficiency significantly reduced cholestasis-induced cholangiocyte proliferation and liver injury [33]. As shown in Fig. 5c, d, S1PR2 mRNA levels were significantly upregulated in the livers of human NASH patients and the WDSW-induced NASH mouse model. Similarly, both mRNA and protein levels of S1PR2 were significantly increased in WDSW-fed HuRhKO mice. To further elucidate the potential role of S1PR2 in HuR deficiency-induced metabolic liver injury, we examined the effect of overexpression of S1PR2 on H19 promoter activity using the luciferase report assay. As shown in Fig. 5e, S1PR2 significantly enhanced H19 promoter activity.

Fig. 5

HuR regulates the expression of SphK2 and S1PR2. a Representative immunoblot images of nuclear SphK2 in the livers of HuRhKO and control mice fed with WDSW for 4 weeks are shown. The relative protein level of SphK2 was calculated using histone H3 as a loading control. Statistical significance relative to control, *p < 0.05 (n = 3) b Representative immunoblot images of nuclear SphK2 in the livers of H19−/− and control mice fed with WDSW for 4 weeks are shown. The relative protein level of SphK2 was calculated using histone H3 as a loading control. Statistical significance relative to control, *p < 0.05 (n = 3). c Relative mRNA levels of S1PR2 in the liver of human NASH patients, WDSW-NASH mice, and HuRhKO mice fed with WDSW for 4 weeks, respectively. Statistical significance relative to control, **p < 0.01. d Representative immunoblot images of S1PR2 in the liver are shown and normalized with β-actin as an internal control of HuRhKO mice fed with WDSW for 4 weeks. Statistical significance relative to control, *p < 0.05 (n = 3). e Effect of S1PR2 on H19 promoter activity. HEK-293 cells were transfected with pGL3-human H19-promoter or pGL3-control vector together with pcDNA3-hS1PR2 or control vector along with Renilla control vector. After transfection for 48 h, the luciferase activities were measured using a dual-luciferase reporter assay kit. The promoter activity was expressed by relative luciferase activity using the ratio of firefly luciferase activity to Renilla luciferase activity. Statistical significance relative to control, ***p < 0.001 (n = 5)

Hepatocyte-specific HuR deficiency enhances WDSW-induced dysregulation of bile acid homeostasis

Bile acids are critical in regulating hepatic lipid, glucose, and energy metabolism as they are important signaling molecules [24, 34, 35]. LC–MS/MS analysis of serum bile acid levels indicated that hepatocyte-specific HuR deficiency aggravated WDSW-induced disruption of bile acid homeostasis (Fig. 1d, e, Additional file 1: Table S2 and S3). Real-time PCR analysis showed that the expression levels of two rate-limiting enzymes in the bile acid synthesis pathway, cholesterol 7 alpha-hydroxylase (Cyp7α1) and cholesterol 27 alpha-hydroxylase (Cyp27α1), were significantly decreased. In contrast, the expression levels of small heterodimer partner (Shp) and Na + -taurocholate cotransporting polypeptide (Ntcp) were significantly upregulated in the livers of WDSW-fed HuRhKO (Fig. 6a). To further determine the impact of hepatocyte-specific deletion of HuR on enterohepatic circulation, the bile acid composition and levels in the liver, intestinal ileum, and cecal contents were measured using LC–MS/MS. As shown in Fig. 6b, the percentage of TCA in total hepatic bile acids was increased from 27% (Control) to 40% (HuRhKO), while βMCA was decreased from 39% (Control) to 24% (HuRhKO). Although the total liver bile acids showed no significant changes between the WDSW-fed HuRhKO mice and control mice, the TCA level, the ratio of total primary conjugated bile acids to total primary unconjugated bile acids, the ratio of total conjugated bile acids to total unconjugated bile acids, and the ratio of total secondary conjugated bile acids to total secondary unconjugated bile acids were increased in HuRhKO mice (Fig. 6c, Additional file 1: Table S4 and S5). Moreover, it should be noted that TCA has been shown to be a potent activator of S1PR2 [36]. In the intestinal ileum, as shown in Additional file 1: Fig. S7a, the percentage of TCA in total bile acids was increased from 27% (Control) to 35% (HuRhKO). Hepatocyte HuR deficiency had no significant effect on intestinal total bile acids, the ratio of total primary conjugated bile acids to total primary unconjugated bile acids, and the ratio of total conjugated bile acids to total unconjugated bile acids in the intestine (Additional file 1: Fig. S7b and Additional file 1: Table S6, S7). However, hepatocyte HuR deficiency significantly reduced bile acid levels, including total bile acids, total unconjugated bile acids, total primary bile acids, and total secondary bile acids in the cecal contents (Fig. 6d, Additional file 1: Fig. S7c, d, and Additional file 1: Tables S8, S9). Under normal diet feeding conditions, the deletion of HuR in hepatocytes significantly changed bile acid levels and composition in the serum. As shown in Additional file 1: Fig. S8a, the percentages of TCA and TβMCA in total bile acids increased from 14 to 63% and 3% to 20%, respectively, in normal diet fed HuRhko mice. While the percentage of βMCA decreased from 25 to 1%. The total bile acid level and ratios of total primary bile acid to total bile acid and total primary bile acid to total secondary bile acids were also increased in the serum of normal diet-fed HuRhko mice (Additional file 1: Tables S10, S11). There were no significant changes in hepatic bile acid level and composition in normal diet-fed HuRhko mice (Additional file 1: Fig. S8b, Tables S12, S13). However, the total bile acid levels in the intestine and cecal contents were reduced, although the bile acid composition was not significantly changed in normal diet-fed HuRhko mice (Additional file 1: Tables S14–17; Fig. S9).

Fig. 6

Hepatocyte-specific HuR deficiency enhances WDSW-induced dysregulation of bile acid homeostasis. a Relative mRNA levels of Cyp7a1, Cyp27a1, Shp, and Ntcp are shown and normalized with HPRT1 as an internal control in the liver of HuRhKO vs. Control mice fed with WDSW for 4 weeks. b BA composition profile in the liver is expressed by % of total BA. c TCA, the ratio of total primary conjugated BA to total primary unconjugated BA, the ratio of total conjugated BA to total unconjugated BA, and the ratio of total secondary conjugated BA to total secondary unconjugated BA in the liver. d Total BA, total unconjugated BA, total primary BA, and total secondary BA in the cecal contents. Data are expressed as the mean ± SEM. Statistical significance relative to control: * p < 0.05, ** p < 0.01, n = 8–12

Hepatocyte-specific HuR deficiency enhances WDSW-induced hepatic fibrosis

To further examine the impact of hepatic HuR in WDSW-induced NAFLD disease progression, both control and HuRhKO mice were fed ad libitum a WDSW for 12 weeks to induce NASH and early fibrosis. As shown in Additional file 1: Fig. S10a, HuR protein levels were significantly downregulated in the livers of HuRhKO mice. As expected, H&E staining showed that HuRhKO mice fed WDSW for 12 weeks exacerbated intra-acinar (lobular) inflammation, hepatocellular ballooning, and macrovesicular steatosis (Fig. 7a). Similar to the 4-week WDSW feeding study in HuRhKO mice (Fig. 4d), the hepatic H19 expression level was significantly upregulated in 12-week WDSW-feeding HuRhKO mice (Fig. 7b). The key genes involved in the fatty acid synthesis and lipid metabolism were also increased in HuRhKO mice compared to the control mice, including sterol regulatory element-binding protein 2 (Srebp2), Pparα, elovl fatty acid elongase 7 (Elovl7), 3-hydroxy-3-methylglutaryl coenzyme-A (HMG-CoA) reductase (HMG-CoAR), PPARG Coactivator 1 Alpha (Pgc1α), lipoprotein lipase (Lpl), fibroblast growth factor 21 (Fgf21), fatty acid desaturase (Fads2), and Cyclin D1 (Additional file 1: Fig. S10b). As shown in Fig. 7c, HuRhKO mice fed WDSW for 12 weeks resulted in enhanced macrophage infiltration to the liver, as indicated by IHC staining of F4/80. In addition, the mRNA levels of F4/80, Cd11b, Cd63, Cd68, Ccl2, Ccr2, Tnfα, Cd14, Tlr4, ceramide kinase (Cerk), Caspase 1, and cyclooxygenase 2 (Cox-2) were significantly increased in the liver of HuRhKO mice (Additional file 1: Fig. S11a). Furthermore, as shown in Additional file 1: Fig. S11b, the expression levels of key genes involved in neutrophil activation were also significantly upregulated in the liver of HuRhKO mice fed WDSW for 12 weeks, including Nox2, neutrophil cytosolic factor 1 (Ncf1), Ncf2, Ncf4, Cybα, Il2rg, Elastin, and Selectin. These findings were consistent with the results in HuRhKO mice fed WDSW for 4 weeks, indicating hepatocyte-specific HuR deficiency enhances WDSW-induced dysregulation of hepatic lipid metabolism and activation of inflammatory response.

Fig. 7

Hepatocyte-specific HuR deficiency aggravates WDSW-induced hepatic fibrosis. HuRhKO and control mice were fed ad libitum a WDSW for 12 weeks. a Representative H&E staining images of the liver sections (scale bar, 20 μm for 40× magnification). b Relative mRNA levels of H19 in the liver of HuRhKO and control mice. c Representative images of the liver sections stained with F4/80, Picro Sirius Red staining, and IHC staining of CK19 (scale bar, 20 μm for 40 × magnification). d Relative mRNA levels of genes involved in hepatic fibrosis were determined by real-time RT–PCR and normalized with HPRT1 as an internal control. Relative mRNA levels of Ck19, α-Sma, Tgfβ, Lox12, Sox4, Sox9, Ctgf, Mmp2, Mmp7, Sctr, Postn, and S1pr2 are shown. Data are expressed as the mean ± SEM. Statistical significance relative to control: *p < 0.05, **p < 0.01, ***p < 0.001, n = 10–12.

Hepatic cell injury and inflammation are the major driving forces of hepatic fibrosis, which is closely associated with mortality in NASH patients [37]. To determine the impact of hepatic HuR on the progression of NASH fibrosis, we performed Picro Sirus Red staining, IHC staining of CK-19 and real-time PCR analysis. As shown in Fig. 7c, 12-week WDSW-feeding induced early fibrosis in HuRhKO mice but much less in control mice. CK-19 staining indicated that hepatocyte-specific HuR deficiency significantly exacerbated WDSW-induced cholangiocyte proliferation. The mRNA expression levels of fibrotic genes were significantly upregulated in 12-week WDSW-feeding HuRhKO mice, including Ck19, smooth muscle actin (α-Sma), transforming growth factor-beta 1 (Tgfβ1), lysyl oxidase-like 2 (Loxl2), SRY (sex-determining region Y)-box 4&9 (Sox4&9), connective tissue growth factor (Ctgf), matrix metallopeptidase 2 and 7 (Mmp2&7), secretin receptor (Sctr), periostin (Postn), and S1pr2, indicating hepatocyte-specific HuR deficiency aggravates WDSW-induced hepatic fibrosis (Fig. 7d).

H19 downregulation alleviates WDSW-induced NAFLD in HuRhKO mice

Based on our results and published RNAseq data, H19 upregulation may represent a major cellular mechanism underlying WDSW-induced NAFLD in HuRhKO mice. To verify the role of H19 in WDSW-induced NAFLD in HuRhKO mice, HuRhKO mice were injected with a recombinant adenovirus encoding an H19 shRNA or control adenovirus (GFP) while being fed ad libitum a WDSW for 4 weeks. As shown in Fig. 8a, the H19 levels in the liver were significantly decreased after injection of adenovirus of H19 shRNA compared to injection of control adenovirus. Down-regulation of H19 reversed the WDSW-induced decrease of nuclear SphK2 protein (Fig. 8b). H&E staining showed that downregulation of H19 in HuRhKO mice reduced WDSW-induced intra-acinar (lobular) inflammation, hepatocellular ballooning, and macrovesicular steatosis (Fig. 8c). The Picro-Sirius Red staining also showed that downregulation of H19 in HuRhKO mice reduced WDSW-induced early fibrosis (Fig. 8d). Together, these results suggest that the upregulation of H19, at least partially, contributes to WDSW-induced NAFLD development in HuRhKO mice.

Fig. 8

H19 downregulation alleviates WDSW-induced NAFLD in HuRhKO mice. HuRhKO mice were injected with a recombinant adenovirus encoding an H19 shRNA or control adenovirus (GFP) while being fed ad libitum a WDSW for 4 weeks. a Relative mRNA levels of H19 in livers injected with H19 shRNA adenovirus compared control adenovirus. Data are expressed as the mean ± SEM. Statistical significance relative to control: *p < 0.05, n = 5. b Representative immunoblot images of nuclear SphK2 in the liver are shown. Relative protein levels of nuclear SphK2 were normalized using histone H3. Statistical significance relative to control: *p < 0.05, n = 5. c Representative H&E staining images of the liver sections (scale bar, 20 μm for 40 × magnification). d Representative Picro Sirius Red staining images of the liver sections (scale bar, 20 μm for 40 × magnification). e Schematic diagram of the potential mechanism of HuR in attenuating WDSW-induced NAFLD. The current study demonstrated that HuR functions as an important regulator of hepatic lipid metabolism, enterohepatic bile acid homeostasis, inflammation, and fibrosis by suppressing H19 expression and modulating the SphK2 nuclear protein level. Bile acid-induced activation of S1PR2 may also contribute to NASH fibrosis. Hepatocyte-specific modulation of HuR expression and its downstream target, H19, may be used to develop potential therapeutics for NAFLD