Pleiotropic actions of IP6K1 mediate hepatic metabolic dysfunction to promote nonalcoholic fatty liver disease and steatohepatitis

Abstract

Objective

Obesity and insulin resistance greatly increase the risk of nonalcoholic fatty liver disease and steatohepatitis (NAFLD/NASH). We have previously discovered that whole-body and adipocyte-specific Ip6k1deletion protects mice from high-fat-diet-induced obesity and insulin resistance due to improved adipocyte thermogenesis and insulin signaling. Here, we aimed to determine the impact of hepatocyte-specific and whole-body Ip6k1 deletion (HKO and Ip6k1-KO or KO) on liver metabolism and NAFLD/NASH.

Methods

Body weight and composition; energy expenditure; glycemic profiles; and serum and liver metabolic, inflammatory, fibrotic and toxicity parameters were assessed in mice fed Western and high-fructose diet (HFrD) (WD: 40% kcal fat, 1.25% cholesterol, no added choline and HFrD: 60% kcal fructose). Mitochondrial oxidative capacity was evaluated in isolated hepatocytes. RNA-Seq was performed in liver samples. Livers from human NASH patients were analyzed by immunoblotting and mass spectrometry.

Results

HKO mice displayed increased hepatocyte mitochondrial oxidative capacity and improved insulin sensitivity but were not resistant to body weight gain. Improved hepatocyte metabolism partially protected HKO mice from NAFLD/NASH. In contrast, enhanced whole-body metabolism and reduced body fat accumulation significantly protected whole-body Ip6k1-KO mice from NAFLD/NASH. Mitochondrial oxidative pathways were upregulated, whereas gluconeogenic and fibrogenic pathways were downregulated in Ip6k1-KO livers. Furthermore, IP6K1 was upregulated in human NASH livers and interacted with the enzyme O-GlcNAcase that reduces protein O-GlcNAcylation. Protein O-GlcNAcylation was found to be reduced in Ip6k1-KO and HKO mouse livers.

Conclusion

Pleiotropic actions of IP6K1 in the liver and other metabolic tissues mediate hepatic metabolic dysfunction and NAFLD/NASH, and thus IP6K1 deletion may be a potential treatment target for this disease.

Graphical abstract

Highlights

• •IP6K1 promotes NAFLD/NASH.
• •IP6K1 reduces mitochondrial oxidative capacity and increases insulin resistance in hepatocytes.
• •IP6K1 enhances protein O-GlcNAcylation in the liver.
• •Whole-body Ip6k1-KO mice are robustly protected from NAFLD/NASH whereas HKO mice display partial improvements.
• •IP6K1 is upregulated in the livers of human NASH and cirrhosis patients.

Abbreviations

NAFLD

Nonalcoholic fatty liver disease

NASH

Nonalcoholic steatohepatitis

Ip6k1-KO

whole body Inositol hexakisphosphate kinase-1 knock out

HKO

Hepatocyte specific Ip6k1 knock out

IP6

Inositol hexakisphosphate

5-IP7

5-diphosphoinositol pentakisphosphate or 5-inositol pyrophosphate

AdKO

Adipocyte specific Ip6k1 knockout

WD

Western Diet

HFrD

High fructose diet

LTCDS

Liver Tissue Cell Distribution System

WT

Wild type

Lox

Ip6k1 flox

CLAMS

Comprehensive Lab Animal Monitoring System

NMR

Nuclear Magnetic Resonance

OCR

Oxygen consumption rate

TNP

N2-(m-Trifluoro benzyl), N6-(p-nitrobenzyl)purine

FCCP

Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone

ECAR

Extracellular acidification rate

ARP

Acidic ribosomal protein large P0

GSEA

Gene set enrichment analysis

TMM

The trimmed mean of M-values

LC-MS/MS

Liquid chromatography with tandem mass spectrometry

HDMC

Phepatocellular carcinoma down-regulated mitochondrial carrier protein

DNPME

2,4-dinitrophenol-methyl ether

NAFLD

Nonalcoholic fatty liver disease

NASH

Nonalcoholic steatohepatitis

Ip6k1-KO

whole body Inositol hexakisphosphate kinase-1 knock out

HKO

Hepatocyte specific Ip6k1 knock out

IP6

Inositol hexakisphosphate

5-IP7

5-diphosphoinositol pentakisphosphate or 5-inositol pyrophosphate

AdKO

Adipocyte specific Ip6k1 knockout

WD

Western Diet

HFrD

High fructose diet

LTCDS

Liver Tissue Cell Distribution System

WT

Wild type

Lox

Ip6k1 flox

CLAMS

Comprehensive Lab Animal Monitoring System

NMR

Nuclear Magnetic Resonance

OCR

Oxygen consumption rate

TNP

N2-(m-Trifluoro benzyl), N6-(p-nitrobenzyl)purine

FCCP

Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone

ECAR

Extracellular acidification rate

ARP

Acidic ribosomal protein large P0

GSEA

Gene set enrichment analysis

TMM

The trimmed mean of M-values

LC-MS/MS

Liquid chromatography with tandem mass spectrometry

HDMC

Phepatocellular carcinoma down-regulated mitochondrial carrier protein

DNPME

2,4-dinitrophenol-methyl ether

1. Introduction

A family of three mammalian inositol hexakisphosphate kinases (IP6Ks) converts the biomolecule inositol hexakisphosphate (IP6) to the inositol pyrophosphate (5-IP7) [[1], [2], [3], [4], [5]]. IP6 and 5-IP7 regulate various cellular processes including metabolism [[1], [2], [3], [4]]. Of the IP6K isotypes, the functions of IP6K1 have been the most characterized in vivo [3,6]. IP6K1 promotes insulin secretion from pancreatic β cells [7] but impairs insulin signaling by inhibiting the insulin effector protein kinase Akt [8]. Thus, in a chronic energy-rich state, IP6K1 contributes to hyperinsulinemia and insulin resistance [8]. Moreover, IP6K1 reduces whole-body energy expenditure by inhibiting thermogenesis in adipocytes [9]. IP6K1 diminishes energy oxidation and augments fatty acid biosynthesis in adipocytes by inhibiting the metabolism-enhancing AMP-activated protein kinase (AMPK) [9,10]. IP6K1 also reduces adipose tissue- and serum levels of the insulin sensitizing and metabolically favorable adipokine adiponectin [9,10]. Consequently, whole-body (Ip6k1-KO or KO) and adipocyte-specific Ip6k1-KO mice display improved insulin sensitivity and energy metabolism and are protected from high-fat-diet-induced (HFD: 60% kcal from fat, no added cholesterol) obesity, hyperinsulinemia, and insulin resistance [[8], [9], [10]].

Obesity-induced hyperinsulinemia and insulin resistance greatly increase the risk of nonalcoholic fatty liver disease and steatohepatitis (NAFLD/NASH) [11,12]. In obesity, dysfunctional adipocytes release abnormal amounts of free fatty acids (FFAs), which are taken up by the liver, leading to lipotoxic liver injury [11,12]. This injury and the associated activated macrophages together provide a stimulus for hepatic stellate cell (HSC) activation into a fibrogenic phenotype [[12], [13], [14]]. Moreover, obesity-induced hyperinsulinemia causes pathway-selective insulin resistance, leading to reduced glucose uptake but increased gluconeogenesis and de novo lipogenesis in the hepatocytes [15]. Altered cholesterol and fructose metabolism in hepatocytes also contribute to the development of NAFLD/NASH [11,12,16,17]. Thus, obesity-induced liver disease often begins with benign steatosis that can subsequently progress to steatohepatitis with inflammation and fibrosis and, in some individuals, cirrhosis and hepatocellular carcinoma, with cardiovascular disease being a common comorbidity [11,18]. Although there are no FDA-approved drugs to treat NASH, many current therapies in development focus on improving energy metabolism, insulin sensitivity, inflammation, and fibrosis [11,12,19,20]. A protein or pathway that has pleiotropic beneficial effects on these processes may be an ideal treatment target for NAFLD/NASH.

Despite being the major inositol pyrophosphate-generating enzyme in the hepatocytes [8,9], the role of IP6K1 in hepatic metabolism and in the etiology of NAFLD/NASH remains poorly understood. Therefore, by using a choline-deficient Western diet and a high fructose diet (HFrD) (WD: 40% kcal fat, 1.25% cholesterol and HFrD: 60% kcal fructose) to induce the full spectrum of NAFLD/NASH [[21], [22], [23]] and hepatic metabolic dysfunction in mice [24], we aimed to determine the impact of hepatocyte-specific and whole-body deletion of Ip6k1 on hepatic metabolism and NAFLD/NASH. The expression of IP6K1 in human NASH liver samples was also assessed.

2. Materials and methods

2.1. Human liver tissue samples

Human liver studies were approved by the Saint Louis University Institutional Biosafety Committee. Frozen deidentified human liver samples from normal individuals as well as from NASH and cirrhosis patients were obtained from the Liver Tissue Cell Distribution System (LTCDS) at The University of Minnesota.

2.2. Animal diets and study designs

All animal studies were approved by the Saint Louis University Institutional Animal Care and Use Committee. WT, Ip6k1-KO [8,9], Lox, and HKO mice on C57BL6 (JAX) background were housed at 23 °C under 12-hour light/dark cycles. The mice were fed ad libitum with free access to water and were in good health condition throughout the duration of treatment. Ip6k1-flox (Lox) and albumin-Cre (JAX: 003574) mice were crossed to generate HKO mice following a standard protocol [9]. Age-matched littermates were used in all experiments.

Ten-week-old chow-fed male and female mice were fed the WD for 18 weeks. Whole body energy expenditure measurement (CLAMS) was performed at the 10th week. Insulin, pyruvate, glucagon, and glucose tolerance tests (ITT, PTT, GgTT, and GTT) were performed in WD-fed male mice at the 14th, 15th, 16th, and 17th weeks, respectively, to assess progressive insulin resistance. ITT was performed in WD-fed female mice at the 14th week. Moreover, 10-week-old chow-fed mice were fed a HFrD for 16 weeks. GgTT and ITT were performed at the 10th and 12th weeks, respectively. After the feeding period was over, the animals were fasted for 5h and euthanized. Tissues were collected for further studies.

2.3. Glucose, insulin, pyruvate, and glucagon tolerance tests

For GTT or PTT, glucose, or pyruvate (2 g/kg BW, i.p.) was injected in 16-h-fasted animals. For ITT or GgTT, insulin (0.75 U/kg BW, i.p.) or glucagon (15 μg/kg BW, i.p.) was injected in 5-h fasted mice. Blood glucose levels were measured using a glucometer by puncturing the tail veins of mice before and after the indicated time periods of injection [8,9,25].

2.4. Hematoxylin and eosin (H&E) staining and NAFLD feature scoring

Overnight-formalin-fixed tissues were processed and stained. NAFLD features were scored using a semiquantitative scoring system [26]. Steatosis was scored as: 0 (no steatosis or steatosis occupying less than 5% of the hepatic parenchyma), 1 (steatosis occupying 5–33% of the parenchyma), 2 (33–66% of parenchyma), 3 (>66% of parenchyma); hepatocyte enlargement, lobular and portal inflammatory cell infiltration were scored as: 0 (none), 1 (mild, one to two foci/200 X field), 2 (moderate, three to four foci/200 X field), and 3 (severe, more than four foci/200 X field).

2.5. Sirius red staining and analysis of the total collagen content

Paraffin-embedded liver sections were stained with 0.1% Sirius Red in saturated picric acid for 2 h. Slides were then washed in water, dehydrated with ethanol and xylene, and mounted. The degree of collagen accumulation was assessed by morphometric analysis, as described previously [27]. Briefly, about 15–20 nonoverlapping images randomly selected from each liver section were captured at 20× magnification. The same threshold was applied to all images. Sirius Red staining was quantified by digital image analysis using ImageJ software (NIH). The amount of collagen in the Ip6k1-KO or HKO livers was expressed relative to the amount of collagen in the corresponding WT or Lox control group.

2.6. Measurement of TAG, AST, ALT, LDL, and HDL

Serum was prepared immediately after blood collection and was aliquoted and frozen at −80 °C. TAG, AST, ALT, and LDL and HDL cholesterol were measured using commercial kits following the manufacturer's instructions. Hepatic lipid extraction for TAG measurement was performed following the standard protocol [28].

2.7. Mitochondrial oxygen consumption rate (OCR)

Murine primary hepatocytes were isolated as described previously [25] and plated onto 50 μg/ml of collagen (type I) coated 96-well XF plate in complete DMEM (2 × 10⁴ cells/well). After 24 h, vehicle (DMSO: PBS, 1:3) or TNP (1 μM) was added for 2 h. Subsequently, DMEM was replaced with vehicle or TNP containing XF medium (5 mM pyruvate and 2.5 mM glucose) for 1 h. The plate was then placed in a Seahorse Analyzer. After baseline measurements, 3 sequential injections were applied to the wells: oligomycin (complex V inhibitor, 1 μg/ml), FCCP (uncoupler, 1 μM), and antimycin A (complex III inhibitor, 0.8 μM) combined with rotenone (complex I inhibitor, 3 μM). Basal, uncoupled, and maximal OCRs were calculated by averaging of the values from each phase [9]. To compare the OCRs in Lox and HKO hepatocytes, the above assay was performed without TNP.

2.8. Cell culture

Human hepatocyte carcinoma cell line HepG2 (ATCC) and human hepatic stellate cell line LX-2 (provided by Dr. Scott Friedman, Mount Sinai, New York) were cultured in DMEM+1%Pen/Strep+10%FBS.

2.9. HSC activation assay

For quantitative RT-PCR (qRT-PCR) analysis, after 24 h of seeding, the cells (2 × 10⁵ cells/well in a 12-well plate) were serum starved for 24 h and were treated subsequently with TGFβ1−/+ TNP (10 μM) for 24 h.

2.10. RNA isolation and qRT-PCR

RNA isolation and qRT-PCR were conducted as described earlier [27]. Acidic ribosomal protein large P0 (ARP) mRNA was used as control and the comparative threshold cycle method was used to calculate changes in mRNA abundance.

2.11. Proteomic studies

The IP6K1-interactome was identified using our published methods [29]. Briefly, IP6K1 was immunoprecipitated from healthy and NASH human liver lysates. Proteins co-precipitated with IgG-IP6K1 and IgG-control were identified by LC-MS/MS. Proteins that were only co-precipitated with IgG-IP6K1 were considered as IP6K1 interactors.

2.12. Protein O-GlcNAcylation studies

Glucose-induced protein O-GlcNAcylation was determined by injecting glucose (2 g/kg body weight, i.p.) or PBS to 16-h-fasted animals [30]. After 4 h, the mice were euthanized and their livers were flash frozen and stored at −80 °C for immunoblot analysis.

2.13. Immunoblotting studies

Tissue and cells were lysed in a RIPA buffer containing a protease-phosphatase inhibitor tablet. Total protein was quantified using a BCA protein assay kit. Equal amounts of total protein were loaded on 10% SDS-PAGE. For collagen detection, a 7% SDS-PAGE was used [31]. Proteins were detected by immunoblotting following our standard protocol [[8], [9], [10],32]. Densitometric analyses of protein bands were performed using ImageJ software.

2.14. Statistics

Trial/prior experiments were used to determine the sample size with ample statistical power. Animals were excluded if they showed any signs of random sickness. Numbers of mice (n) used in experiments are indicated in the legends. Immunoblots were quantified using ImageJ software (NIH). Data are presented as mean ± SEM within dot plots. Each symbol represents an individual sample. For multiple comparisons, one-way or two-way ANOVA with the Holm-Šidák multiple comparison test was used, and for two independent data sets, two-tailed unpaired Student's t-test was used. The area under the curve and statistical significance were calculated using GraphPad Prism, 8.2.1.

3. Results

3.1. Hepatocyte-specific Ip6k1 deletion (HKO) protected mice from diet-induced hyperglycemia, hepatic steatosis, and fibrosis

Young (2-month-old), chow-fed HKO mice displayed substantially reduced IP6K1 protein in the liver, leaving the other metabolic tissues unaffected (Figure 1A). The residual level of IP6K1 in the HKO mouse liver was due to IP6K1's expression in non-parenchymal hepatic cells. Ip6k2 or Ip6k3 was not upregulated in HKO mouse livers, ruling out the possibility of compensation of Ip6k1 loss by these isotypes (Fig. S1A).

Figure 1.

Hepatocyte-specific Ip6k1 deletion (HKO) protected mice from Western diet-induced hyperglycemia, hepatic steatosis, and fibrosis. A. Ip6k1 was deleted in the liver but not in the EWAT, IWAT, or skeletal muscle of HKO mice. Data obtained from young, chow-fed mice. B. HKO mice displayed improved glucose disposal in an insulin tolerance test (ITT). C. Glucose-induced peak blood glucose levels were reduced, and its disposal was increased in HKO mice (glucose tolerance test, GTT). D. HKO mice displayed reduced peak blood glucose levels and increased disposal of glucose in a pyruvate tolerance test (PTT). E. Glucagon injection increased blood glucose levels to a lesser extent in HKO mice (glucagon tolerance test, GgTT). F. Liver weight was reduced in HKO mice. G–H. Steatosis and TAG were reduced in HKO mouse livers. Representative H&E staining images of 8 mice (n = 8). I–J. Serum TAG, AST, and ALT levels were reduced in HKO mouse livers. K–L. HKO mice displayed reduced fibrosis compared with Lox mice. Images are examples from different mice. For comparison of densitometry, the mean value of Lox was set as 1 (n = 10 per cohort). Number of mice (n) used in each experiment are presented as individual datapoints. WD-fed male mice were used unless otherwise indicated. Data are presented as mean ± SEM within dot plots. For multiple comparisons, two-way ANOVA with the Holm-Šidák multiple comparison test was used, and for two independent data sets, two-tailed unpaired Student's t-test was used. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

Metabolic profiles of chow-fed Lox and HKO mice were similar (data not shown). However, HKO mice were protected from diet-induced metabolic dysfunction and fibrosis. Insulin- or glucose-mediated glucose disposal (insulin or glucose tolerance test, ITT or GTT) was improved in WD-fed HKO mice (Figure 1B–C and Figs. S1B–S1C). Pyruvate and glucagon increase blood glucose levels partly by enhancing hepatic gluconeogenesis, evidenced by an increase in the blood glucose following their injection (pyruvate or glucagon tolerance test, PTT or GgTT). WD-fed HKO mice displayed significant decrease in pyruvate- or glucagon-induced rise in blood glucose levels (Figure 1D–E and Figs. S1D–S1E). Moreover, WD-fed HKO mouse livers exhibited reduced hepatomegaly, steatosis, and triglyceride (TAG) accumulation (Figure 1F–H) and serum levels of TAG and hepatotoxicity markers such as aspartate and alanine aminotransferases (AST and ALT) (Figure 1I–J). Chronic hepatocellular metabolic dysfunction provides a stimulus for HSC activation into a fibrogenic phenotype [[12], [13], [14]]. Sirius Red staining of collagen fibers showed that WD-induced hepatic fibrosis was reduced (∼40–50%) in HKO mouse livers (Figure 1K–L). Expression of Acta2 (gene for the alpha smooth muscle actin, α-SMA, a marker of HSC activation) was significantly reduced and collagen type 1 alpha 1 and alpha 2 (Col1a1 and Col1a2) were moderately diminished, whereas the expression levels of Col3a1, the metallopeptidase inhibitor 1 (Timp1), and matrix metalloproteinase 2 (Mmp2) were unaltered in WD-fed HKO mouse livers (Fig. S1F). Expression of the inflammatory receptor TLR4 and activating phosphorylation of the downstream kinase JNK were similar in Lox and HKO mice (Fig. S1G). Moreover, HFrD feeding developed insulin resistance and mild hepatic steatosis in Lox but to a lesser extent in HKO mice (Fig. S1H-S1N). This diet did not cause fibrosis in Lox or HKO mice. Thus, hepatocyte-IP6K1 modulates liver metabolism and the development of NAFLD/NASH.

3.2. Hepatocyte-specific Ip6k1 deletion (HKO) increased mitochondrial oxidative capacity of hepatocytes but did not alter body weight or energy expenditure in mice

Increased uncoupled mitochondrial oxidative capacity in adipocytes reduces fat accumulation in Ip6k1-KO mice [9,10]. Similarly, WD-fed HKO mouse livers displayed increased levels of hepatic proteins that mediate mitochondrial oxidative capacity including the hepatic isotype of the mitochondrial uncoupling protein UCP2 (Figure 2A–C). Moreover, mitochondrial OCR was increased in HKO hepatocytes in vitro due to augmented mitochondrial uncoupling (ATP-coupled OCR was unaltered) (Figure 2D–F). Glycolysis (extracellular acidification rate) was also increased in HKO hepatocytes (data not shown). Inhibition of the IP6K pathway by the inhibitor TNP [32,33] displayed similar effects in HepG2 and WT (Lox) primary hepatocytes (Fig. S2A–S2D). However, localized improvement in hepatocyte energy expenditure was not adequate to alter body weight or whole-body energy expenditure as WD-fed HKO mice did not exhibit alterations in these parameters (Figure 2G–I and Figs. S2E–S2G). We observed similar effects in HFD-fed HKO mice (data not shown).

Figure 2.

Hepatocyte-specific Ip6k1 (HKO) deletion increased mitochondrial oxidative capacity of hepatocytes but did not alter body weight or energy expenditure in mice. A–C. Hepatic levels of the mitochondrial oxidative phosphorylation proteins (NDUFB8, SDHB, UQCRC2, MTCO1, and ATP5A for complexes I–V, respectively, and UCP2) were augmented in HKO mice. D–E. Hepatocytes isolated from chow-fed HKO mice displayed increased OCR in vitro due to mitochondrial uncoupling (n = 5 experimental replicates). F. UCP2 protein level was increased in HKO hepatocytes used in Figure 2D. G–H. Body weight of Lox and HKO mice were similar at the beginning and end of WD diet feeding and intraabdominal adipose tissue stores in WD-fed mice appeared similar. I. Whole body energy expenditure (EE) was similar in Lox and HKO mice. Number of mice (n) used in each experiment are presented as individual datapoints. WD-fed male mice were used unless otherwise indicated. Data are presented as mean ± SEM within dot plots. Hepatocytes were isolated from young, chow-fed male mice. Two-tailed unpaired Student's t-test. ∗P < 0.05, ∗∗P < 0.01.

3.3. Whole-body deletion of Ip6k1 (KO) reduced body weight and robustly protected mice from western diet-induced hyperglycemia, hepatic steatosis, liver injury, and fibrosis

Based on our previous findings with regard to adipocyte deletion of Ip6k1 [9] and current discoveries with regard to hepatocyte deletion of Ip6k1, we hypothesized that the improved metabolic phenotype caused by adipocytes lacking Ip6k1 should protect whole-body Ip6k1-KO mice from NAFLD/NASH to a greater extent than hepatocyte deletion alone. WD moderately increased body weight (∼5 g) in WT but not in whole-body Ip6k1-KO male mice (Figure 3A). Accordingly, WD-fed Ip6k1-KO mice accumulated less fat (Figure 3B–C and Fig. S3A). WD-fed Ip6k1-KO mice also displayed improved glucose disposal following glucose or insulin injections (Fig. S3B–S3E). The livers of Ip6k1-KO mice did not have the pale and enlarged appearance of WT mice due to reduced TAG accumulation (Figure 3D–G and S3F). Histologic NAFLD feature scoring revealed that micro- and macro-steatosis and hepatocyte hypertrophy were reduced in Ip6k1-KO mice (Figure 3H). Levels of the fatty acid biosynthetic enzymes acetyl-CoA carboxylase (ACC) and fatty acid synthase (FAS) were reduced in the KO mouse livers (Figure 3I and Fig. S3G). Cholesteryl esters but not free cholesterol levels were significantly reduced in the KO mouse livers (Fig. S3H–S3I). Moreover, Ip6k1-KO livers had fewer inflammatory foci, reduced expression of the inflammatory receptor TLR4, and reduced activating phosphorylation of the downstream protein kinases JNK and NFκB (Figure. 3H-last panel, 3J and Fig. S3J–S3K). Serum levels of TAG, low-density lipoprotein (LDL), but not high-density lipoprotein (HDL), cholesterol, AST, and ALT were also lower in WD-fed Ip6k1-KO male mice (Figures. 3K–L and S3L–S3M). Consequently, WD-induced hepatic fibrosis was substantially (∼70–80%) reduced in KO mouse livers (Figure 3M−N). NAFLD/NASH parameters (insulin resistance, inflammation, hepatic steatosis, and fibrosis) were improved to a greater extent in Ip6k1-KO mice than in HKO mice compared with those observed in the corresponding control mice. WD-fed Ip6k1-KO female mice also gained less body and fat weight (Fig. S3N–S3Q). The livers of female WT but not Ip6k1-KO mice appeared pale (Fig. S3Q). Liver weight and fat accumulation were marginally less in WD-fed Ip6k1-KO females (Figs. S3R-S3S). Yet, KO females exhibited significantly reduced serum TAG and ALT levels (Figs. S3T-S3U). WD feeding (for 18 weeks) did not induce fibrosis in WT or Ip6k1-KO female mice. Thus, due to pleiotropic effects of IP6K1 on various metabolic tissues [6], whole-body Ip6k1 deletion robustly protected mice from WD-induced weight gain, insulin resistance, metabolic dysfunction, hepatic steatosis, and fibrosis.

Figure 3.

Whole-body deletion of Ip6k1 (KO) reduced body weight and robustly protected mice from Western diet-induced hyperglycemia, hepatic steatosis, liver injury, and fibrosis. A–D.Ip6k1-KO males displayed reduced body and fat weight and had darker, normal-appearing livers compared with the pale enlarged livers of WT mice. NMR analysis showed a reduction in percent body fat but unaltered lean and fluid masses in the knockouts. E–G.Ip6k1-KO male mice displayed reduced weight and fat accumulation in the liver. H. Macro- and micro-steatosis, hepatocyte hypertrophy, and inflammatory foci were reduced in Ip6k1-KO mice. NAFLD features were quantified from the H&E slides (n = 4 mice per cohort). I. The lipogenic ACC and FAS proteins were reduced in Ip6k1-KO livers. J. TLR4 (normalized to GAPDH in Fig. 3I) and stimulatory phosphorylation of JNK and NFκB were reduced in Ip6k1-KO mouse livers. K–L. Serum TAG and LDL were reduced in Ip6k1-KO mice. M–N.Ip6k1-KO mouse livers were robustly protected from fibrosis. For comparison, the mean value of WT was set as 1 (n = 6 per cohort). Number of mice (n) used in each experiment are presented as individual datapoints. WD-fed male mice were evaluated. Data are presented as mean ± SEM within dot plots. For multiple comparisons, two-way ANOVA with the Holm-Šidák multiple comparison test was used, and for two independent data sets, two-tailed unpaired Student's t-test was used. ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

3.4. IP6K1 regulates various target proteins and pathways that modulate liver metabolism and fibrosis

IP6K1 regulates various targets and pathways in vivo [3,6]. To identify the pathways that were altered in WD-fed Ip6k1-KO liver, RNA-Seq studies were performed. Coupled and uncoupled mitochondrial oxidative pathways were upregulated, whereas the gluconeogenic pathway was downregulated, which conform with our observation of improved metabolic phenotypes in HKO mice (Table S1, blue and orange denote upregulation and downregulation, respectively; Figs. S4A–B). IP6K1 reduces adipocyte mitochondrial oxidative capacity via direct inhibition of AMPK [9,10]. Activating phosphorylation of AMPK was increased in WD-fed Ip6k1-KO and HKO mouse livers (Figs. S4C–S4D).

RNA-Seq studies also showed that the hepatic transcriptomic profiles of a fibrotic response such as collagen synthesis, HSC activation, integrin signaling and extracellular matrix (ECM) production [13,34] were downregulated in WD-fed Ip6k1-KO mouse livers (Table S1 and Fig. S4E). Accordingly, markers of fibrosis and HSC activation such as Col1a1, α-SMA, and desmin were reduced at the protein level in KO mouse livers (Figure. 4A and S4F). The integrin receptors activate multiple proteins including the focal adhesion kinase (FAK) to induce the expression of ECM proteins and TGFβ1-induced HSC activation in fibrosis [35,36]. The integrin β1 protein was downregulated and the activating phosphorylation of FAK was decreased in WD-fed Ip6k1-KO mouse livers (Figure 4B and S4G–S4H). Based on these results, we checked whether IP6K1 regulates HSC activation. Ip6k1 and Ip6k2 were comparably expressed in the human HSC cell line LX-2 under basal and TGFβ1-treated conditions (Fig. S4I), whereas Ip6k3 was undetectable (data not shown). Interestingly, siRNA-mediated Ip6k1 depletion reduced the expression of α-SMA in TGFβ1-treated LX-2 cells (Figure 4C). Ip6k2 depletion did not produce any such effect (data not shown). Moreover, TNP treatment reduced the expression of Acta2, the gene for α-SMA (Figure 4D). TNP did not alter Col1a1 expression under basal conditions but significantly impaired TGFβ1-induced expression of this gene (Figure 4E).

Figure 4.

IP6K1 regulates various target proteins and pathways that modulate liver metabolism and fibrosis. A. Col1a1, α-SMA, and desmin were reduced in Ip6k1-KO mouse livers. For ImageJ analysis of Col1a1, all three Col1a1 bands were compared between WT and Ip6k1-KO samples. B. Integrin β1 and stimulatory phosphorylation of FAK (Y397) were reduced in Ip6k1-KO mouse livers. C.Ip6k1 depletion (siRNA) reduced α-SMA protein levels in TGFβ1-treated LX-2 cells. D. TNP treatment reduced mRNA expression of Acta2 in TGFβ1-treated LX-2 cells. E. TNP treatment impaired TGFβ1-induced mRNA expression of Col1a1 in LX-2 cells. F–G. IP6K1 but not IP6K2 was upregulated in the livers of human NASH and alcoholic cirrhosis patients. α-SMA was used as a marker of profibrogenic stellate cell activation in NASH and cirrhosis. ImageJ quantification was used to compare the expression of proteins in normal vs NASH and cirrhosis livers. H–I. Levels of protein O-GlcNAcylation were increased in human NASH and cirrhosis livers. For quantification, ImageJ analysis was performed on the most prominent O-GlcNAcylated bands (indicated by arrows) using β-actin as a control. J–M. Protein O-GlcNAcylation was reduced in WD-fed Ip6k1-KO and HKO mice compared with control mouse livers. For quantification, ImageJ analysis was performed on two O-GlcNAcylated bands that were the most prominent (indicated by arrows). N. Glucose-induced protein O-GlcNAcylation was reduced in the chow-fed HKO mouse liver. O. Model showing pleiotropic mechanisms by which IP6K1 promotes the development of NAFLD/NASH. Chronic unhealthy diets stimulate IP6K1-mediated inactivation of AMPK and activation of protein O-GlcNAcylation and integrin signaling in the liver. Consequently, mitochondrial oxidative capacity and insulin signaling are impaired, whereas HSCs are activated to a fibrogenic phenotype. IP6K1 also facilitates adipose tissue dysfunction, which causes inappropriate fatty acid release with subsequent uptake in the liver contributing to lipotoxic injury. The net effect is a phenotype of NASH and liver fibrosis. Solid and dotted red arrows denote current and previous discoveries from our group. Back arrows represent information obtained from the literature. Each experiment is presented as individual datapoints. Data are presented as mean ± SEM in dot plots. For multiple comparisons, two-way ANOVA with the Holm-Šidák multiple comparison test was used, and for two independent data sets, two-tailed unpaired Student's t-test was used. ∗P < 0.05, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

Finally, we assessed the relevance of IP6K1 in human NASH. IP6K1, but not IP6K2, was markedly upregulated in the livers of human NASH and cirrhosis patients, indicating that IP6K1 may play a role in the development and progression of liver fibrosis in humans as it does in mice (Figure 4F–G). LC-MS/MS studies of immunoprecipitated IP6K1 from human livers identified the enzyme O-GlcNAcase (OGA) as an IP6K1 interacting protein in NASH (Fig. S4J). OGT (O-GlcNAc transferase) and OGA regulate protein O-GlcNAcylation by adding and removing the β-D-N-acetylglucosamine (O-GlcNAc) moiety to and from target proteins [37,38]. Protein O-GlcNAcylation reduces insulin sensitivity, whereas it stimulates lipogenesis [39]. Increased protein O-GlcNAcylation was found in the livers of human NASH and cirrhosis patients (Figure 4H–I). Moreover, this modification was reduced in WD-fed Ip6k1-KO and HKO mouse livers and HFrD-fed HKO mouse livers compared with control mice (Figures 4J–M and S4K–S4L). Glucose-induced acute protein O-GlcNAcylation [30] was also prevented in chow-fed HKO mouse livers (Figure 4N). These results suggest that IP6K1 promotes net protein O-GlcNAcylation in the liver.

4. Discussion

Our study demonstrates that hepatocyte-specific Ip6k1 deletion improves hepatic metabolism, which ameliorates diet-induced hepatic insulin resistance and steatosis in mice. In the liver, IP6K1 regulates various pathways that modulate energy metabolism, insulin sensitivity, and fibrogenesis. Moreover, adipocyte-specific Ip6k1 deletion enhances thermogenic energy expenditure and reduces β-adrenergic receptor-induced lipolysis, protecting mice from HFD-induced obesity, insulin resistance, and liver fat accumulation [9,10,29]. Furthermore, IP6K1 promotes insulin secretion from pancreatic β cells [7,40], which augments hyperinsulinemia [8], a major contributor to hepatic de novo lipogenesis. Thus, the adverse metabolic effects of adipose tissue-derived FFA and hyperinsulinemia are presumably similar in the livers of Lox and HKO mice. Consequently, HKO mice displayed less protection against NAFLD/NASH compared with whole-body Ip6k1-KO mice. Based on our previous [[8], [9], [10]] and current studies, we propose a model by which pleiotropic actions of IP6K1 promote the development of NAFLD/NASH (Figure 4O). The albumin-cre deleted Ip6k1 in hepatocytes and bipotent liver epithelial progenitor cells. To determine the impact of Ip6k1 deletion specifically on mature hepatocytes, an adeno-associated virus (AAV8-TBG-cre) system should be used in future studies.

IP6K1-regulated other processes may also contribute to NAFLD/NASH development. For example, IP6K1-generated 5-IP7 maintains the cellular polyphosphate levels and pyrophosphorylate protein targets [41,42] that mediate chemotaxis and phagosomal motility of immune cells including macrophages [43,44]. Hence, IP6K1 may also play a role in Kupffer cell-mediated inflammation in NAFLD/NASH. The limited information available suggests the potential significance of other IP6K isotypes in liver and metabolic diseases. Like Ip6k1, Ip6k2 is upregulated in human hepatocellular carcinoma [45]. Ip6k3 expression is limited to skeletal muscle and heart, and its whole body deletion improves skeletal muscle metabolism, protecting mice from age-induced fat gain and insulin resistance [46].

IP6K1-modulated pathways in metabolic tissues may have therapeutic relevance in NAFLD/NASH. The indirectly acting AMPK-activating drug metformin is first-line therapy for type-2 diabetes (T2DM), and newer direct AMPK-activating drugs are in the clinical development process for the treatment of obesity, T2DM, and NASH [47,48]. Pharmacological activation of whole-body-, and liver-targeted mitochondrial uncoupling is also being evaluated as a therapeutic approach to treat obesity, T2DM, and NAFLD/NASH [49]. Targeting integrin/FAK signaling is a potential strategy to treat fibrotic diseases including NASH [50]. Increased protein O-GlcNAcylation blocks insulin signaling and promotes lipogenesis [39,51]. IP6K1 inhibition may ameliorate metabolic dysfunction and NAFLD/NASH by altering these pathways. Although the mechanisms by which IP6K1 regulates insulin signaling, AMPK, mitochondrial uncoupling, and FAK have been studied [6,[8], [9], [10],52], how it stimulates protein O-GlcNAcylation is currently unknown.

Protein O-GlcNAcylation has been shown to differentially influence hepatic metabolism and fibrosis [39,51]. Heterozygous OGA-deleted mice (homozygous OGA-KO mice do not survive) develop metabolic dysfunction. Moreover, adipocyte-OGT deletion reduces protein O-GlcNAcylation and ameliorates HFD-induced obesity, insulin resistance, and liver fat accumulation in mice [53,54]. OGT is upregulated in the livers of human NAFLD-associated hepatocellular carcinoma (HCC) and its pharmacological inhibition reduces HCC in a mouse xenograft model [55]. Polyphenolic compounds silibinin and curcumin reduce NAFLD/NASH and protein O-GlcNAcylation in mouse models [56,57]. In contrast, decreased protein O-GlcNAcylation in human cirrhotic livers and the protective effects of hepatocyte-OGT on alcohol-induced cirrhosis in mice have also been demonstrated [58]. Overall, both activation and disruption of protein O-GlcNAcylation have been shown to stimulate HSC activation [[59], [60], [61]]. Perhaps, the regulated modulation of O-GlcNAcylation of specific proteins at a particular stage of obesity, insulin resistance, and NASH may have therapeutic importance rather than its constitutive alteration.

Enhancing mitochondrial uncoupling may induce ATP depletion-mediated cellular injury [62]. Moreover, obesity-related fatty liver is not improved by UCP2 deletion in genetically obese (ob/ob) mice [63]. Encouragingly, pharmacological activation of liver uncoupling by treating mice with DNPME (2,4-dinitrophenol-methyl ether) ameliorates NAFLD without reducing hepatocyte ATP levels [64]. HKO hepatocytes showed unaltered coupled (ATP-dependent) OCR (Figure 2E). In Ip6k1-KO embryonic fibroblasts, ATP levels are higher due to increased glycolysis [65]. Thus, regulated increases in hepatocyte uncoupling without ATP depletion may have beneficial effects in NAFLD.

In most cases, IP6K1-mediated regulation of its metabolic targets can be attributed to the catalytic activity of IP6K1 [3,6]. Thus, pharmacological inhibition of IP6K1 by the pan-IP6K inhibitor TNP [33] reduces intracellular 5-IP7 [3] levels and ameliorates HFD-induced obesity, hyperinsulinemia, insulin resistance, and liver injury in mice by increasing thermogenic energy expenditure and insulin sensitivity [9,32]. TNP enhances mitochondrial oxidative capacity and insulin sensitivity and reduces fatty acid biosynthesis in adipocytes. It also reduces insulin secretion from β cells and FAK activation in mouse embryonic fibroblasts [9,10,32,33,52,66]. Our current study showed that TNP enhances mitochondrial oxidative capacity of hepatocytes and reduces HSC activation. Unfortunately, TNP is not a drug-like compound due to structural and other shortcomings and thus can only be used as a pharmacological tool [3,32]. Therefore, efforts are ongoing to develop drug-like IP6K inhibitor compounds [[67], [68], [69]]. A newly developed IP6K inhibitor has been shown to alleviate kidney fibrosis in a rat model of chronic kidney disease by increasing ATP and reducing hyperphosphatemia [70]. Hopefully, the improved IP6K inhibitor compounds will be evaluated to assess their potential as treatment targets to reverse NASH and its associated fibrosis.

5. Conclusions

Pleiotropic actions of IP6K1 in metabolic tissues mediate hepatic metabolic dysfunction leading to NAFLD/NASH. Targeting only the liver is insufficient as a treatment for NAFLD/NASH because it cannot overcome the influences of dysfunctional adipose tissue and perhaps other tissue types throughout the body (e.g., pancreas). Systemic inhibition of IP6K1 or IP6Ks in general could ameliorate the full spectrum of obesity, T2DM, and NAFLD/NASH.

Author contributions

A.C. conceived the project and designed the experiments; A.C. and M.C. generated HKO mice; M.C. and S.M. maintained mouse colonies; S.M. performed the majority of the experiments; M.C., B.U, K.M, D.P., E.M., J.H. C.G., and S.G. also performed the experiments; A.C. and S.M. analyzed the majority of the data; B.U., B.N-T., and D.C. analyzed the histology data; J.Z. analyzed RNA-Seq data; D.F. analyzed the lipidomic data; A.C. and S.M. prepared the final figures; A.C. wrote the first draft; B.N-T. edited the draft; B.U and S.M. also helped in the editing. All authors provided input and approved the manuscript prior to submission.

Conflict of interest

B.N-T. has been an advisor or consultant for Alimentiv, Allergan, Allysta, Alnylam, Amgen, Arrowhead, Axcella, Boehringer Ingelheim, BMS, Coherus, Cymabay, Enanta, Fortress, Genfit, Gilead, High Tide, HistoIndex, Innovo, Intercept, Ionis, LG Chem, Lipocine, Madrigal, Medimmune, Merck, Mirum, NGM, NovoNordisk, Novus Therapeutics, pH-Pharma, Sagimet, Target RWE, 89Bio; he has stock options in HepGene and received institutional research grants from Allergan, BMS.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Multimedia component

2

Fig S1.

Fig S2.

Fig S3.

Fig S4.