Second-generation antisense oligonucleotides against β-catenin protect mice against diet-induced hepatic steatosis and hepatic and peripheral insulin resistance

Abstract

Although mutations in the Wnt/β-catenin signaling pathway are linked with the metabolic syndrome and type 2 diabetes in humans, the mechanism is unclear. High-fat-fed male C57BL/6 mice were treated for 4 wk with a 2′-O-methoxyethyl chimeric antisense oligonucleotide (ASO) to decrease hepatic and adipose expression of β-catenin. β-Catenin mRNA decreased by ≈80% in the liver and by 70% in white adipose tissue relative to control ASO-treated mice. β-Catenin ASO improved hepatic insulin sensitivity and increased insulin-stimulated whole body glucose metabolism, as assessed during hyperinsulinemic-euglycemic clamp in awake mice. β-Catenin ASO altered hepatic lipid composition in high-fat-fed mice. There were reductions in hepatic triglyceride (44%, P < 0.05) and diacylglycerol content (60%, P < 0.01) but a 30% increase in ceramide content (P < 0.001). The altered lipid content was attributed to decreased expression of sn-1,2 diacylglycerol acyltransferase and mitochondrial acyl-CoA:glycerol-sn-3-phosphate acyltransferase and an increase in serine palmitoyl transferase. The decrease in cellular diacyglycerol was associated with a 33% decrease in PKCε activation (P < 0.05) and 64% increase in Akt2 phosphorylation (P < 0.05). In summary, Reducing β-catenin expression decreases expression of enzymes involved in hepatic fatty acid esterification, ameliorates hepatic steatosis and lipid-induced insulin resistance.—Popov, V. B., Jornayvaz, F. R., Akgul, E. O., Kanda, S., Jurczak, M. J., Zhang, D., Abudukadier, A., Majumdar, S. K., Guigni, B., Petersen, K. F., Manchem, V. P., Bhanot, S., Shulman, G. I., Samuel, V. T. Second-generation antisense oligonucleotides against β-catenin protect mice against diet-induced hepatic steatosis and hepatic and peripheral insulin resistance.

Nonalcoholic fatty liver disease (NAFLD) is currently the most common liver disease in the United States with prevalence rates as high as 46% (1). NAFLD is closely associated with the metabolic syndrome and type 2 diabetes (2, 3). The development of insulin resistance in NAFLD has been attributed to the ability of specific lipid metabolites to interfere with insulin signaling. For example, an increase in cellular diacylglycerol (DAG) activates novel PKC isoforms, which impair insulin signaling (4). The regulation of cellular lipid content is under a complex set of controls that balance calorie availability with energetic demands.

The Wnt signaling pathway represents one set of controls. The Wnt molecules are a highly conserved family of secreted, lipid-modified glycoproteins that control multiple processes during embryogenesis (5). In adult tissues, the Wnt pathways are involved in bone turnover, adipogenesis, and liver regeneration (5–7). In addition, aberrant activation of the Wnt/β-catenin signaling drives the progression of numerous malignancies, including cancers of the gastrointestinal system (8). All of these processes can affect metabolic pathways and suggest that Wnt signaling pathways regulate intermediary metabolism.

Recent genomewide association studies have linked the Wnt pathway with metabolic diseases (9–11). Studies in humans and rodents have identified a mutation in LRP6, a coreceptor in the WNT pathway, that predisposes to the metabolic syndrome (9). Human genetic studies found an association between a loss-of-function mutation in the Wnt 10B gene and obesity (10). Similarly, 2 single nucleotide polymorphisms in the LRP5 gene were associated with obesity as well (12, 13). Although these studies clearly show that Wnt signaling is associated with metabolic disorders, the mechanisms by which Wnt proteins affect metabolism is not clear.

Classic Wnt signaling pathways converge on β-catenin (13). Wnt/β-catenin signaling starts with binding of Wnt ligands to Wnt cell surface receptors. This leads to cytoplasmic stabilization of β-catenin and subsequent nuclear translocation and regulation of transcriptional activity. Wnt/β-catenin signaling has a critical role in regulating hepatic function and metabolism. β-Catenin recently has been shown to mediate the hepatic response to starvation (14) and oxidative stress (15). β-Catenin itself may have discrete cellular roles. It exists in the cell within 2 distinct pools: associated with cadherins and the actin cytoskeleton at cellular junctions and cytosolic/nuclear pool, where it participates in Wnt signaling. The cadherin-bound β-catenin is essential for cell adhesion and migration and is not as prone to phosphorylation and degradation as the cytosolic β-catenin (16).

The exact role of Wnt/β-catenin in conditions of calorie surplus and insulin resistance has not been defined yet. Liver-specific conditional deletion of β-catenin lowered fasting glucose and improved glucose tolerance in high-fat-fed (HFF) mice, possibly by decreasing FoxO nuclear translocation and expression of gluconeogenic enzymes. However, this was paradoxically associated with increased hepatic lipid content (14). Thus, the underlying mechanism whereby β-catenin impacts hepatic lipid metabolism and hepatic insulin action remains unresolved.

To address these questions, we decreased hepatic and adipose tissue β-catenin expression with second-generation 2′-O-methoxyethyl chimeric antisense oligonucleotides (ASO) in HFF adult C57BL/6 mice. ASOs decrease expression of a specific target predominantly in the liver and white adipose tissue (WAT) but not in other metabolically active tissues (brain, pancreas, skeletal, and cardiac muscle), and thus are an effective tool to study a selective decrease of β-catenin expression (17–20). In contrast to knockout models, ASOs can lower gene expression in adult animals, avoiding any compensatory developmental changes that may arise as a consequence of a genetic deletion. Moreover, ASOs allow validation of potential novel drug targets. We quantified the changes in body composition, whole body energy metabolism, and tissue lipid content to evaluate the phenotype induced by β-catenin ASO in conditions of calorie surplus (HFF mice). We performed hyperinsulinemic-euglycemic clamps to quantify basal and insulin-stimulated glucose metabolism in the liver, muscle, and adipose tissue and evaluated changes in insulin signaling.

MATERIALS AND METHODS

Animals

Male C57BL/6J mice were obtained at 6 wk of age (The Jackson Laboratory, Bar Harbor, ME, USA). The mice were housed at Yale University School of Medicine and maintained in accordance with the Institutional Animal Care and Use Committee guidelines and were individually housed under controlled temperature (23°C) and lighting (12:12-h light-dark cycle, lights on at 7 am), with free access to water and food. After 1 wk of acclimatization, the mice received a high-fat diet (HFD, 55% fat by calories; TD 93075; Harlan Teklad, Madison, WI, USA) for 4–6 wk and were studied at 12–14 wk of age. β-Catenin or control ASO (Isis Pharmaceuticals, Carlsbad, CA, USA) was injected intraperitoneally at a dose of 50 mg/kg per week for 4–6 wk. Mice underwent placement of a right jugular catheter about 10 d before studies. They recovered their presurgical weights by 5 to 7 d after the operation. The Institutional Animal Care and Use Committee of Yale University approved all procedures.

Metabolic cage studies

Fat and lean body mass were assessed by ¹H magnetic resonance spectroscopy (Bruker BioSpin, Billerica, MA, USA). A comprehensive animal metabolic monitoring system (CLAMS; Columbus Instruments, Columbus, OH, USA) was used to measure O₂ consumption, CO₂ production, energy expenditure, activity, and food and water consumption.

Intraperitoneal glucose tolerance test

After an overnight withdrawal of food, mice were weighed and received a bolus intraperitoneal injection of glucose (1 g/kg). Blood samples were obtained from a tail vein at baseline (0) and 15, 30, 45, 60, 90, and 120 min after a glucose challenge for determination of plasma glucose and insulin concentrations.

Hyperinsulinemic-euglycemic clamp

Hyperinsulinemic-euglycemic clamps were performed as previously described (21, 22) and more details available in the supplementary material.

Hepatic lipid metabolites assay

Liver triglyceride (TG) content was determined by using a TG assay kit (Genzyme Diagnostics P.E.I. Inc., Charlottetown, PE, Canada) and a method adapted from Storlien et al. (23). Extraction, purification, and assessment of long-chain fatty acyl-CoA (LCCoA), ceramide, and DAGs from liver by liquid chromatography-mass spectrometry/mass spectrometry have been described previously (24).

Liver lipid measurements

Tissue TGs were extracted using the method of Bligh and Dyer (25) and measured using a DCL TG reagent (Diagnostic Chemicals, Charlottetown, PE, Canada). For DAG extraction, livers were homogenized in a buffer solution (20 mM Tris-HCl, 1 mM EDTA, 0.25 mM EGTA, 250 mM sucrose, 2 mM PMSF) containing a protease inhibitor mixture (Roche, Basel, Switzerland), and samples were centrifuged at 100,000 g for 1 h. The supernatants containing the cytosolic fraction were collected. DAG levels were then measured by liquid chromatography-mass spectrometry/mass spectrometry, as previously described (24, 26). Total cytosolic DAG content is expressed as the sum of individual species. Ceramide and LCCoA were measured as previously described (24). All lipid measurements were done from tissues of animals after a 12 h withdrawal from food.

Immunoblotting analysis

Immunoblotting analysis was performed as previously described (22). The following primary antibodies were used: phopho-AKT2^ser473 (Signalway Antibody, College Park, MD, USA), AKT2 (Cell Signaling Technology, Danvers, MA, USA), PKCε (BD Transduction Laboratories, Lexington, KY, USA), β-catenin (Santa Cruz Biotechnology, Santa Cruz, CA, USA), GAPDH (Santa Cruz Biotechnology), histone (Cell Signaling Technology). Immune complexes were detected using an enhanced luminol chemiluminescence system (ECL; Thermo Scientific, Waltham, MA, USA) and subjected to photographic films. Signals on the immunoblot were quantified by optical densitometry (ImageJ, National Institutes of Health, Bethesda, MD, USA).

Total RNA preparation and real-time quantitative PCR analysis

Total RNA was extracted from frozen livers or WAT using RNeasy 96-kit (Qiagen, Germantown, MD, USA), then 1 μg of RNA was reverse transcribed into cDNA with the use of the Quantitect RT kit (Qiagen) as per the manufacturer’s protocol. The abundance of transcripts was assessed by real-time PCR on a 7500 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) with a SYBR Green detection system. Duplicate samples were assayed for both the gene of interest and GAPDH and data were normalized for the efficiency of amplification, as determined by a standard curve included on each run.

Statistical analysis

Data are expressed as the means ± sem. Results were assessed using Student’s t test or 1-way ANOVA (GraphPad, GraphPad Software, La Jolla, CA, USA; Prism 6) followed by Tukey’s multiple comparison test, as appropriate. Values of P < 0.05 were considered significant.

RESULTS

β-Catenin ASO decreased mRNA expression of β-catenin in the liver and WAT

To assess the physiologic role of β-catenin on lipid-induced hepatic steatosis and insulin resistance, we fed male C57BL/6 mice a 55% HFD for 4 to 6 wk while administering either a control ASO without homology to any mRNA target or an ASO against β-catenin. Compared with control ASO-treated mice, β-catenin ASO decreased β-catenin mRNA expression in the liver and WAT (Fig. 1A) mice by ∼70–80% (P < 0.005). Levels of liver β-catenin protein were 25% lower with the knockdown in a whole cell homogenate (Fig. 1B, C). We considered whether subcellular compartmentation of β-catenin may account for the discordance between the effective decrease in mRNA expression and the modest decrease in whole cell protein expression. When we immunostained liver tissue for β-catenin, we noted that β-catenin was predominantly located near the hepatocytes’ membranes and that this membrane staining appeared unaltered by our ASO treatment despite the decrease in β-catenin mRNA expression (Supplemental Fig. 1). We then quantified β-catenin protein expression in discrete subcellular fractions: membrane, cytosol, and nuclear fractions (Fig. 1B, C). We noted significantly decreased levels in the cytosolic β-catenin fraction with the ASO knockdown (Fig. 1B, C), as well as the nuclear and whole cell fraction but not in the membrane fraction. Downstream targets of Wnt signaling such as cyclin D1 had lower mRNA expression (Fig. 1A), as well as lower protein levels (Fig. 1B, C) when measured in a whole cell homogenate. We assessed β-catenin protein expression in WAT whole cell lysates but were unable to detect a significant difference (Supplemental Fig. 3).

Figure 1.

β-Catenin ASO is effective in decreasing expression of β-catenin in the liver and WAT. A) β-Catenin and cyclin D1 mRNA expression by quantitative RT-PCR. B) Immunoblots for β-catenin and cyclin D1 in liver tissue of wild-type HFD mice treated with either β-catenin ASO or control ASO for 4 wks. GAPDH was used as a loading control for cytosol and whole cell protein extracts, histone as a loading control for nuclear, and Na-K ATPase as a membrane loading control (n = 8 per group). C) Densitometry of blots. Protein ratios, normalized to corresponding loading controls, were used to quantify fold change relative to control ASO. P value calculated by 2-way ANOVA. *P < 0.05, **P < 0.005 compared with overnight-unfed control ASO-treated mice.

Knockdown of β-catenin expression with ASO did not alter body weight or whole body energy balance

Total body weight, weight gain, fat mass, and lean body mass were similar between the control ASO group and the β-catenin ASO group after 4 wk of HFD, as assessed by ¹H magnetic resonance spectroscopy (Table 1). There were no detectable changes in whole body energy balance. Specifically, food intake, locomotor activity, and energy expenditure, as assessed by indirect calorimetry, were not altered by treatment with β-catenin ASO.

TABLE 1.

Physiologic parameters

Physiologic parameter	Control ASO	β-Catenin ASO	P value
Final weight (g)	25.1 ± 0.5	25.7 ± 0.6	0.3
Weight gain (g)	4.2 ± 0.5	4.3 ± 0.6	0.9
Fat (%)	8.89 ± 0.7	7.85 ± 0.64	0.3
Muscle (%)	18.35 ± 0.4	19.15 ± 0.5	0.2
Activity (counts/hr)	201 ± 18	190 ± 15	0.65
Energy expenditure [kcal/(kg-h)]	18.4 ± 0.6	17.66 ± 0.4	0.31
VO2 (ml/kg/h)	3819 ± 121	3669 ± 84	0.33
VCO2 (ml/kg/h)	3042 ± 84	2917 ± 64	0.3
RER	0.80 ± 0.01	0.79 ± 0	0.74
Serum TGs (mg/dl)	118.7 ± 12.4	74.16 ± 7.2	0.03*
Serum cholesterol (mg/dl)	124.6 ± 5.3	103.8 ± 4.8	0.03*
Serum NEFA (mEq/L)	1.1 ± 0.1	0.9 ± 0.1	0.03*
Serum adiponectin (ng/ml)	35.69 ± 2.4	47.73 ± 2.6	0.006**
Fasting insulin (μU/ml)	10.1 ± 0.9	9.6 ± 1	0.05
Fasting glucose (mg/dl)	112 ± 4.2	115 ± 4.5	0.5
ALT (U/L)	22.9 ± 3.1	30.3 ± 5.4	0.2

Treatment with β-catenin ASO does not affect weight and body composition as assessed by ¹H magnetic resonance spectroscopy and does not alter whole body energetics after 4 wk of HFD and ASO treatment (control ASO or β-catenin ASO); however, serum TGs, cholesterol, and free fatty acids are lower in the β-catenin knockdowns (n = 8 per group). Data are expressed as means ± sem. VO₂, O₂ consumption; VCO₂, CO₂ consumption; RER, respiratory exchange ratio. P value calculated by 2-way ANOVA.* P < 0.05, **P < 0.005 compared with overnight-unfed control ASO-treated mice fed HFD.

β-Catenin ASO was effective in improving serum lipid profile in fat-fed mice

There was no difference in fasting plasma glucose and plasma insulin concentrations between controls and ASO-treated mice. β-Catenin ASO treatment improved plasma lipid profile in HFF mice: plasma cholesterol concentration was reduced by 17% (P < 0.05), TG concentration by 37% (P < 0.05), and nonesterified fatty acid (NEFA) concentration by 20% (P < 0.05) (Table 1). In addition, plasma adiponectin was significantly higher in the β-catenin ASO group (P < 0.05).

β-Catenin knockdown prevented lipid-induced hepatic and peripheral insulin resistance

To assess the effect of β-catenin knockdown on glucose homeostasis, we measured whole body glucose metabolism in HFF mice by intraperitoneal glucose tolerance test (IPGTT). The excursion in plasma glucose concentration following IPGTT suggested a modest improvement (Fig. 2A) with a 13% reduction in the area under the glucose excursion in the β-catenin ASO relative to the control ASO-treated mice (P = 0.03). Although the plasma insulin concentrations were not significantly different between the groups (Fig. 2B), the values for the β-catenin ASO-treated mice tended to be lower. Taken together, these data suggest that β-catenin ASO may improve insulin sensitivity in HFF mice.

Figure 2.

Treatment with β-catenin ASO improves whole-body glucose tolerance. HFD β-catenin ASO mice have lower glucose (A) and insulin (B) levels compared with WT mice fed a HFD during IPGTT. Area under the curve for glucose was 13% lower for the β-catenin ASO group compared with the control ASO group (P = 0.05), and 12% lower when evaluating insulin excursions (P = 0.16). IPGTT was performed after ovenight food withdrawal as detailed; mice had been fed HFD for 4 wk prior to the experiment (n = 7–8 per group). *P < 0.05.

To further quantify the tissue-specific changes in insulin action, we performed hyperinsulinemic-euglycemic clamp experiments in awake HFF mice, combined with radiolabeled glucose, to quantify the impact of β-catenin ASO treatment on insulin-stimulated glucose metabolism in liver, muscle, and adipose tissue (Fig. 3). As before, fasting plasma glucose and insulin concentrations were similar in both groups (Table 1). Basal rates of endogenous glucose production were also similar (Fig. 3A). Under hyperinsulinemic-euglycemic conditions, β-catenin ASO treatment increased whole body insulin responsiveness; the glucose infusion rate required to maintain euglycemia was nearly 2-fold higher in the β-catenin ASO-treated group [control ASO: 16 ± 2.2 vs. β-catenin ASO: 30.9 ± 1.7 mg/(kg-min), P < 0.0005, Fig. 3B, C]. The increased requirement of exogenous glucose infusion was largely attributed to improved suppression of endogenous glucose production in the β-catenin ASO-treated mice (Fig. 3D). β-Catenin ASO also increased (P < 0.05) insulin-stimulated peripheral glucose uptake by 30% (Fig. 3D). The tissue-specific contributions to insulin-stimulated glucose uptake were further resolved by measuring 2-deoxyglucose uptake in key insulin responsive tissues: WAT, brown adipose tissue, gastrocnemius, quadriceps, and the heart (Supplemental Figs. 2 and 3F). Insulin suppressed plasma nonesterified fatty acids concentrations to a similar degree (control ASO: 0.61 ± 0.05 vs. β-catenin ASO: 0.48 ± 0.03 mM, P = 0.08) with an equal suppression of serum fatty acids (control ASO: 40 ± 5% vs. β-catenin ASO: 43 ± 6% mM, P = 0.44). The increase in insulin-stimulated whole body glucose uptake was associated with improvements in insulin-stimulated uptake of glucose in WAT (Fig. 3F) and, surprisingly, the heart (Supplemental Fig. 2). However, β-catenin protein expression in the heart was not affected (Supplemental Fig. 3).

Figure 3.

β-Catenin ASO-treated mice are protected from diet-induced hepatic and peripheral insulin resistance, as quantified by a hyperinsulinemic-euglycemic clamp. A) Basal glucose levels in overnight-unfed mice treated with control or β-catenin ASO. B) Increased glucose infusion rates in the β-catenin ASO-treated mice compared with the controls during the clamp. C) Average of the glucose infusion rates required to maintain euglycemia during the clamp. D) Increased insulin-stimulated peripheral glucose uptake. E) Insulin-stimulated suppression of endogenous glucose production during the clamp. F) Increased WAT glucose uptake in the β-catenin group with the clamp. n = 6–8 per group; the β-catenin group was injected with β-catenin ASO; the control group received injections of control ASO as detailed in Materials and Methods. Both groups were fed HFD for 4 wk prior to the clamp, while receiving the injections of AS. P value calculated by 1-way ANOVA. *P < 0.05 vs. control ASO; **P < 0.01 vs. control ASO.

β-Catenin ASO improved hepatic steatosis

We measured hepatic lipid content to further investigate the mechanisms by which β-catenin ASO improved hepatic insulin sensitivity. Liver histology was assessed in hematoxylin and eosin-stained liver sections by a pathologist blinded to the treatment groups. The control group had increased macro- and microvesicular steatosis compared with the β-catenin ASO group (Supplemental Fig. 4). However, there was no evidence of ballooning degeneration or fibrosis in either group. Serum aspartate aminotransferase/alanine aminotransferase levels were similar between groups (Table 1).

The development of insulin resistance in fat-fed animals has been associated with the accumulation of specific ectopic lipid species (26). The improvement in hepatic insulin sensitivity with β-catenin ASO was associated with a 44% reduction in hepatic TG: β-catenin, 14 ± 0.8; control, 22 ± 1.7 mg TG/g liver tissue (P < 0.05) (Fig. 4A). We also noted a 60% reduction in DAG in a plasma membrane fraction (P < 0.01) (Fig. 4C) in the β-catenin ASO-treated mice. DAG in a cytosolic fraction (which includes lipid droplets) was not affected. β-Catenin ASO did not alter LCCoA levels (Fig. 4B). Ceramides have also been implicated in the pathogenesis of insulin resistance (27). In contrast to the reduction in hepatic DAG, hepatic ceramide content was increased by 30% in the β-catenin ASO-treated animals (P < 0.001) (Fig. 4E).

Figure 4.

Knockdown of β-catenin expression with ASO in HFF mice is associated with decreased hepatic TGs (A) and DAGs (C, D). It did not alter LCCoA levels (B) and, paradoxically, ceramides (E) were increased in the liver. Expression levels of key lipogenic genes (mtGPAT and DGAT2) involved in lipid re-esterification were decreased with β-catenin knockdown (F) compared with controls. The expression of SPT mRNA, the rate-controlling enzyme of ceramide synthesis was increased (F). Lipid levels and mRNA expression of β-catenin ASO-treated HFF mice is compared with wild-type HFF mice treated with control ASO (n = 8 per group). *P < 0.05, **P < 0.005. P value calculated by 1-way ANOVA.

β-Catenin ASO decreased expression of key genes for TG synthesis

We measured the expression of enzymes and molecules involved in lipogenesis, lipid trafficking, and β-oxidation. There was an ∼50% (P < 0.05) reduction in the mRNA expression of both sn-1,2 diacylglycerol acyltransferase (DGAT2) and mitochondrial acyl-CoA:glycerol-sn-3-phosphate acyltransferase (mtGPAT) in the β-catenin ASO group compared with the control group (Fig. 4F). In comparison, there were no changes in the expression of enzymes regulating β-oxidation, de novo lipogenesis and lipid export (Table 2). Protein levels of mtGPAT were also significantly decreased in the liver (Supplemental Fig. 5). The reduction in mtGPAT may explain the discordance between LCCoA and DAG and TG concentration, as LCCoA levels were similar between the controls and the knockdowns, whereas both DAGs and TGs were significantly lower in the β-catenin ASO-treated mice. In addition, the expression of serine palmitoyl transferase (SPT) was increased and consistent with the observed increase in hepatic ceramide (Fig. 4F). Thus, in HFF mice, loss of β-catenin leads to marked improvement in hepatic insulin sensitivity and reduction in hepatic DAG content associated with a decrease in mtGPAT and DGAT2.

TABLE 2.

Relevant metabolism gene expression in liver and adipose tissue

Tissue, function, and gene	Control ASO	β-Catenin ASO	P value
Liver
Lipogenesis
SREBP1c	0.93 ± 0.19	0.98 ± 0.09	0.8
ACC2	0.94 ± 0.18	0.77 ± 0.14	0.5
FAS	1.2 ± 0.26	1.58 ± 0.48	0.5
SCD1	1.72 ± 0.53	1.56 ± 0.32	0.8
β-oxidation
PPARα	1.05 ± 0.15	0.64 ± 0.05	0.03*
CPT1	1.13 ± 0.23	0.47 ± 0.09	0.03*
Lipid import/export
CD36	1.2 ± 0.2	2.00 ± 0.54	0.15
MTP	1.22 ± 0.23	0.91 ± 0.17	0.3
Gluconeogenesis
PEPCK	1.4 ± 0.19	0.76 ± 0.2	0.04*
G6P	0.99 ± 0.13	0.66 ± 0.15	0.1
Lipolysis
ATGL	0.97 ± 0.04	1.56 ± 0.18	0.005**
β-Catenin target	1.94 ± 0.04	0.79 ± 0.02	0.02*
cyclin D1
Adipose tissue
Hydrolysis
ATGL	1.25 ± 0.27	0.95 ± 0.4	0.4
HSL	1.3 ± 0.28	0.89 ± 0.30	0.2
Adipocyte differentiation
PPARγ	1.48 ± 0.26	0.73 ± 0.2	0.04*
C/EBPα	1.07 ± 0.18	0.99 ± 0.15	0.8

β-Catenin ASO treatment does not alter lipogenesis or hepatic β-oxidation; however, the expression of enzymes involved in glycerol re-esterification (mtGPAT, DGAT2) is significantly reduced (Fig. 4F). Data are expressed as means ± sem. ACC, acetyl CoA-carboxylase; ATGL, adipocyte triglyceride TG lipase; C/EBPα, CCAAT/enhancer binding protein α; CPT, carnitine palmitoyl transferase; DGAT, acyl-CoA:diacylglycerol acyltransferase; FAS, fatty acid synthase; G6P, glucose 6-phosphatase; HSL, hormone sensitive lipase; MTP, microsomal triglyceride TG transfer protein; PEPCK, phosphoenolpyruvate carboxykinase; SPT, serine palmitoyl transferase; PPAR, peroxisome proliferator activated receptor; SCD, stearoyl-CoA desaturase-1; mtGPAT, mitochondrial acyl-CoA:glycerol-sn-3-phosphate acyltransferase; SREBP, sterol regulatory element binding transcription factor. *P < 0.05, **P < 0.005 compared with overnight-unfed control ASO-treated mice fed HFD.

β-Catenin ASO mice have decreased liver DAGs and enhanced hepatic insulin signaling

DAG can activate novel PKC isoforms (θ, in muscle; ε, in liver) and impair insulin signaling (28). Therefore, we considered if the improvements in insulin action and decrease in DAG content with β-catenin ASO could be attributed to alterations in PKC activation and insulin signaling in vivo. The decrease in hepatic DAG content (Fig. 4C, D) was associated with a significant decrease in PKCε activation (29) (Fig. 5B), reflected by a 20% decrease in the membrane:cytosol ratio. Akt activation was assessed by comparing the degree of Akt phosphorylation in response to insulin in both control ASO- and β-catenin ASO-treated animals. There was 64% increase in insulin-stimulated pAKT:AKT ratio in β-catenin ASO-treated mice compared with controls, consistent with improved hepatic insulin signaling (Fig. 5A).

Figure 5.

Hepatic insulin signaling is augmented with ASO-mediated knockdown of β-catenin expression. A) Immunoblots of liver tissue from mice treated with control or β-catenin ASO for 4 wk in the basal or clamped state showing increased ratio of phosphorylated AKT2 to total AKT with insulin stimulation in the β-catenin knockdowns. B) Immunoblots of liver tissue from mice treated with control or β-catenin ASO for 4–5 wk showing decreased membrane to cytosol PKCε ratio in the β-catenin knockdowns (n = 4 per group). P value calculated by 2-way ANOVA. *P < 0.05 vs. control ASO; ***P < 0.001 vs. control ASO.

DISCUSSION

β-Catenin, a final downstream mediator of canonical Wnt signaling, regulates cell differentiation and growth partly through its regulation of lipid metabolism (30). Based on the involvement of Wnt in hepatic and systemic metabolism, we hypothesized that β-catenin may regulate glucose and lipid metabolism in diet-induced insulin resistance. Here we report a novel role of β-catenin as a regulator for hepatic lipid metabolism and propose a mechanism by which knockdown of β-catenin may prevent lipid-induced insulin resistance.

Recent studies examined the role of the Wnt pathway in regulating hepatic glucose metabolism in several different models. Liu et al. used mice with floxed β-catenin alleles that were fed a HFD for 10 wk and received adenoviral Cre-recombinase to selectively decrease β-catenin expression only in the liver (14). The phenotype was then assessed 1 wk after the injection. This acute liver-specific deletion of β-catenin resulted in lower fasting plasma glucose concentrations and improved glucose tolerance despite increased hepatic steatosis on a HFD. These changes were attributed to reduction in nuclear FoxO1 and decreased expression of hepatic gluconeogenic enzymes as well as improvements in insulin signaling following loss of β-catenin (14). Interestingly, these changes occurred despite increased accumulation of liver TG, although other lipid intermediates were not assessed. Another study (31) used liver-specific β-catenin knockout and transgenic mice to determine the effect of manipulating liver metabolic zonation in the overfed state. The knockout mice had lower hepatic TG and improved hepatic insulin sensitivity on HFD, similar to our results. However, the knockout mice also had decreased weight gain on HFD compared with the transgenic and wild-type mice, perhaps due to decreased intestinal fat absorption, which could have contributed to the overall phenotype. Furthermore, although the knockout mice exhibited improved hepatic steatosis, there was evidence of increased apoptosis. The complete lack of any hepatic β-catenin protein in this model may have possibly contributed to increased liver susceptibility to injury (15, 32).

The differences between these 2 prior studies and the present study are likely due to the specific models. We used a specific ASO to decrease β-catenin expression over a 4-wk high-fat feeding period [as opposed to the end of a 10-wk HFD, as done by Liu et al. (14)]. And, although we did see a reduction in phosphoenolpyruvate carboxykinase expression (but not G6Pase expression), there was no change in fasting plasma glucose concentrations or in basal rates of endogenous glucose production. Instead, the primary effect of β-catenin ASO was a reduction in specific liver lipid metabolites with improved hepatic insulin sensitivity and insulin-stimulated whole body glucose metabolism. Although there were no detectable increases in insulin-stimulated skeletal muscle glucose uptake, there were significant increases in WAT and cardiac glucose uptake. The underlying mechanism for this improvement in cardiac glucose uptake is not clear. β-Catenin ASO did not decrease target expression in cardiac tissue, suggesting that this is an indirect effect of β-catenin ASO at other sites. And, in contrast to the study performed by Behari et al., β-catenin ASO did not alter body weight or composition (32). Metabolic cage studies revealed that these mice had similar energy expenditure and food intake arguing against any alterations in energy balance or fat absorption. Finally, one additional difference is the selective decrease of β-catenin protein expression in the cytoplasmic pool without significant changes in the membrane bound β-catenin. Although ASOs decrease mRNA expression relatively quickly, the difference in β-catenin protein expression in these 2 pools suggest that either protein turnover is very different between these 2 pools, possibly due to increased stability of the membrane pool, or there is preferential binding by E-cadherin of any β-catenin protein that is synthesized from a reduced pool of mRNA (33). Although further studies are required to determine why the membrane-bound pool of β-catenin is protected from the effects of ASO, it is clear that the cytosolic pool is decreased following β-catenin ASO treatment.

The main effect of β-catenin ASO was the alteration in hepatic lipid metabolism. β-Catenin ASO decreased hepatic TG content in HFF mice and improved hepatic insulin sensitivity. Intracellular TGs are unlikely to be direct mediators of insulin resistance. Other bioactive lipid metabolites, such as DAGs (34) and ceramides (27), have been implicated in impairing insulin signaling and causing insulin resistance. DAGs specifically can activate PKCε in the liver, which impairs insulin receptor kinase activity (35). This mechanism has also been translated to humans where hepatic DAGs are more closely correlated with hepatic insulin resistance than any other hepatic lipid species (36, 37). Intracellular DAGs are present in membrane, cytosolic, and ER/lipid droplet pools, but only specific pools may be important in the pathogenesis of insulin resistance. This is demonstrated in mice treated with an ASO against comparative gene identification-58 (CGI-58, also called αβ hydrolase domain containing-5), an activator of adipose TG lipase. CGI-58 knockdown worsened hepatic steatosis but did not affect insulin action. This apparent disconnect between hepatic steatosis and hepatic insulin action was attributed to a redistribution of intracellular DAG, with a marked accumulation of DAG in the lipid droplet, but a reduction in membrane DAG (38). With β-catenin ASO, although there was no change in the cytosolic DAG content, there was a reduction in membrane DAG content. The reduction in membrane DAG prevented activation of PKCε and this consequently preserved insulin signaling, as evidenced by the enhanced insulin-stimulated phosphorylation of Akt2.

We propose that β-catenin ASO decreases cellular DAG and TG content primarily because of reduced synthesis of these species. There was decreased expression of mtGPAT and DGAT2, 2 key enzymes involved in the esterification of fatty acids onto a glycerol backbone to create TG. MtGPAT is hypothesized to be the rate-controlling enzyme in the glycerol-3-phosphate pathway, catalyzing the esterification of LCCoA with glycerol 3-phosphate to produce 1-acyl-glycerol 3-phosphate (39), whereas DGAT2 regulates the terminal esterification of LCCoA with DAG to produce TGs, the final committed step in TG synthesis in the liver (40). Genetic deletion of mtGPAT results in partitioning of fatty acids away from DAG and TG synthesis and toward oxidation or phospholipid synthesis (41). This abrogates ectopic buildup of DAGs and TGs in HFF rodents, although overexpression of mtGPAT leads to hepatic steatosis (39). Suppression of DGAT2 with ASO has been shown to prevent accumulation of DAGs and TGs by decreasing lipogenesis and protecting rodents from fat-induced insulin resistance (42). Both DGAT (43) and GPAT (44) appear to be regulated by β-catenin, possibly due to decreased activation of key transcription factors (e.g., peroxisome proliferator activated receptor α or γ) (43).

In contrast to the reduction in DAGs and TGs, hepatic LCCoA content was not significantly altered by the β-catenin ASO treatment. This was somewhat surprising as β-catenin ASO decreased mtGPAT expression and mice with a genetic deletion of mtGPAT accumulate hepatic LCCoA when fed a HFD (39). We did not observe any changes in the expression of key proteins that regulate lipid oxidation or hepatic fat transport. Instead, we hypothesize that there could be redirection of fatty acids into a different lipid synthesis pathway, such as ceramide synthesis. We found an increase in liver ceramide content with β-catenin knockdown and increase in the expression of SPT, the rate-controlling enzyme in ceramide synthesis (45). SPT activity is mainly regulated by influx of saturated fat (46), suggesting redistribution of fatty acids in the β-catenin ASO group toward ceramide synthesis. Ceramides have also been implicated as mediating insulin resistance (47). However, in our study, hepatic insulin sensitivity was improved despite the accumulation of ceramides, thus dissociating ceramide buildup from insulin resistance in this particular experiment. This disassociation between hepatic ceramide content and insulin sensitivity has been observed in human studies as well (36, 37). Taken together, these data may suggest that β-catenin ASO leads to a rerouting of intermediate lipid metabolites from DAG and TG synthesis to ceramide synthesis.

β-Catenin ASO also targets adipose tissue, as reflected by the decrease in adipose tissue β-catenin mRNA. Although we were unable to detect differences in WAT β-catenin protein in whole cell lysates, there were some changes in adipocyte function with β-catenin ASO treatment. Notably, there was a modest 30% increase in adiponectin. Adiponectin has been suggested to improve insulin sensitivity (48, 49), possibly by decreasing ceramide synthesis (50). However, as described above, although there were improvements in hepatic insulin action, hepatic ceramide content was actually increased. Changes in adipose lipolysis may alter hepatic glucose and lipid metabolism. Lipolysis could drive substrate-mediated esterification of lipids into DAG and TG (51). In addition, hepatic β oxidation of adipose-derived fatty acids into acetyl-CoA could allosterically activate pyruvate carboxylase and regulate gluconeogenesis (52, 53). However, the decrease in serum NEFA was relatively small in comparison to the changes in liver DAG and TG and was not associated with decreases in LCCoA or ceramide. In addition, the slight reduction in fasting NEFA concentration was not associated with reduction in basal glucose concentration or the basal rate of glucose production. Finally, although there were significant improvements in insulin-stimulated WAT glucose uptake, this is only a minor contribution to the improvements in whole body glucose uptake. Taken together, the changes in adipose function in this model likely contribute very little to the overall metabolic changes following β-catenin ASO treatment.

Mutations in the Wnt signaling pathway have been associated with metabolic disease and agents that modify these pathways have been proposed as novel treatments for metabolic disorders (54). The studies presented here examine the role of β-catenin, a key target of the Wnt signaling pathways. Knockdown with ASO protected HFF mice from hepatic steatosis and hepatic insulin resistance, revealing a potential novel role for β-catenin as a regulator of lipid metabolism, specifically regulating pathways of fatty acid esterification in the liver. Furthermore, these studies have provided insights in the role of specific lipid metabolites in the pathogenesis of insulin resistance, supporting the role of DAG as a mediator for insulin resistance and provide evidence that ceramide accumulation is not sufficient to cause insulin resistance.

Glossary

ASO

antisense oligonucleotide

DAG

diacylglycerol

DGAT

acyl-CoA:diacylglycerol acyltransferase

HFD

high-fat diet

HFF

high-fat-fed

IPGTT

intraperitoneal glucose tolerance test

LCCoA

long-chain fatty acyl-CoA

mtGPAT

mitochondrial acyl-CoA:glycerol-sn-3-phosphate acyltransferase

NAFLD

nonalcoholic fatty liver disease

NEFA

nonesterified fatty acids

SPT

serine palmitoyl transferase

TG

triglyceride

WAT

white adipose tissue