|
Research Article Hepatic PTEN deficiency improves muscle insulin sensitivity and decreases adiposity in mice
Background & AimsPTEN is a dual lipid/protein phosphatase, downregulated in steatotic livers with obesity or HCV infection. Liver-specific PTEN knockout (LPTEN KO) mice develop steatosis, inflammation/fibrosis and hepatocellular carcinoma with aging, but surprisingly also enhanced glucose tolerance. This study aimed at understanding the mechanisms by which hepatic PTEN deficiency improves glucose tolerance, while promoting fatty liver diseases. MethodsControl and LPTEN KO mice underwent glucose/pyruvate tolerance tests and euglycemic-hyperinsulinemic clamps. Body fat distribution was assessed by EchoMRI, CT-scan and dissection analyses. Primary/cultured hepatocytes and insulin-sensitive tissues were analysed ex vivo. ResultsPTEN deficiency in hepatocytes led to steatosis through increased fatty acid (FA) uptake and de novo lipogenesis. Although LPTEN KO mice exhibited hepatic steatosis, they displayed increased skeletal muscle insulin sensitivity and glucose uptake, as assessed by euglycemic-hyperinsulinemic clamps. Surprisingly, white adipose tissue (WAT) depots were also drastically reduced. Analyses of key enzymes involved in lipid metabolism further indicated that FA synthesis/esterification was decreased in WAT. In addition, Ucp1 expression and multilocular lipid droplet structures were observed in this tissue, indicating the presence of beige adipocytes. Consistent with a liver to muscle/adipocyte crosstalk, the expression of liver-derived circulating factors, known to impact on muscle insulin sensitivity and WAT homeostasis (e.g. FGF21), was modulated in LPTEN KO mice. ConclusionsAlthough steatosis develops in LPTEN KO mice, PTEN deficiency in hepatocytes promotes a crosstalk between liver and muscle, as well as adipose tissue, resulting in enhanced insulin sensitivity, improved glucose tolerance and decreased adiposity. Graphical abstractAbbreviations
Keywords
IntroductionNon-alcoholic fatty liver disease (NAFLD) encompasses a spectrum of liver metabolic disorders, starting with an excessive accumulation of neutral lipids in cytoplasmic droplets of hepatocytes (steatosis), which can then progress towards inflammation, fibrosis and cirrhosis. Obesity and viral infections are common causes of these chronic liver diseases, which are often accompanied by insulin resistance (IR). Indeed, lipotoxicity, resulting from excessive overloading of hepatocytes with lipids, was reported to affect insulin-stimulated signalling pathways that control glucose and lipid metabolism [1]. Hepatic IR is likely to represent a precursor event, leading to systemic and long-standing IR [2]. Uncontrolled hepatic glucose output may indeed induce hyperglycemia and compensatory hyperinsulinemia, favouring IR development in other organs. In turn, insulin-resistant muscle and adipose tissue exacerbate hepatic metabolic disorders, thus nourishing a vicious circle of peripheral IR. Lipotoxicity, inflammation and systemic IR contribute with time to the alteration of pancreatic β-cell function and survival, resulting in their inability to secrete enough insulin to counteract peripheral tissues IR, therefore leading to the development of type 2 diabetes [2] and [3]. In turn, diabetes favours steatosis evolution towards steatohepatitis, fibrosis/cirrhosis and hepatocellular carcinoma, again creating a vicious circle [4]. Insulin signalling is highly regulated at different levels by multiple mechanisms. Among them, the phosphatase and tensin homolog (PTEN) is a dual specificity protein and phosphoinositide phosphatase that dephosphorylates PdtIns(3,4,5)P3, the product of PI3K [5]. By metabolizing PdtIns(3,4,5)P3, PTEN interrupts insulin signalling downstream of PI3K. This PTEN antagonistic effect on PI3K signalling [6] and its nuclear function on chromosomal stability [7] position PTEN as an important tumour suppressor, which is often deleted/mutated or downregulated in human cancers [6]. Alterations of PTEN expression/activity are also expected to deeply affect lipid and glucose homeostasis. Indeed, PTEN heterozygosity and PTEN tissue-specific deletions in muscle or adipose tissue all lead to improved glucose tolerance in healthy or obese/diabetic mice [8], [9] and [10]. However, adding to the complexity of PTEN function, transgenic mice, overexpressing PTEN, display increased energy expenditure and insulin sensitivity [11] and [12]. Regarding the liver, we previously reported that PTEN is downregulated in steatotic livers of obese patients, as well as in rat models of genetic or diet-induced obesity [13]. Likewise, PTEN is downregulated in the liver of patients infected with hepatitis C virus (HCV) [14]. Interestingly, both obesity and HCV infection are associated with the development of steatosis and IR. However, liver-specific PTEN knockout mice (LPTEN KO) exhibit an ambiguous phenotype. Indeed, LPTEN KO mice develop sequentially hepatic steatosis, inflammation/fibrosis and hepatocellular carcinoma with aging, indicating that PTEN plays a crucial role in the development of these pathologies [15] and [16]. Yet, LPTEN KO mice also exhibit an improved glucose tolerance, which is unexpected with NAFLD [15] and [16]. This study aimed at understanding the mechanisms through which liver-specific PTEN deficiency improves glucose tolerance, while promoting NAFLD. Materials and methodsReagents, antibodies, and cell culturesAll reagents, antibodies, commercial kits, cell isolation and cell culture are described in the Supplementary Materials and methods section. AnimalsPtenflox/flox (CTL) and AlbCre-Ptenflox/flox (LPTEN KO) mice generated as previously described [15], were housed at 23 °C; light cycle: 07.00 am–07.00 pm and had free access to water and standard diet. All experiments were conducted in accordance with the Swiss guidelines for animal experimentation and were ethically approved by the Geneva Health head office. 4-month old mice were sacrificed using isoflurane anaesthesia followed by rapid decapitation and blood/tissues were collected and stored at ?80 °C. Metabolic phenotyping, EchoMRI, and CT-scanEnergy expenditure and the respiratory exchange ratio were determined by indirect calorimetry: locomotor activity was recorded by an infrared frame, and food and fluid intake were measured by highly sensitive feeding and drinking sensors. These parameters were measured in mice housed individually in Labmaster metabolic cages (TSE, Bad Homburg, Germany) after 5 days of adaptation prior to recording. Fuel (carbohydrate plus protein vs. fat) oxidation was calculated as described by Bruss et al. [17]. An EchoMRI-700 quantitative nuclear magnetic resonance analyzer (Echo Medical Systems, Houston, TX) was used to measure total fat and lean mass. Distribution, volume and weight of fat depots were analysed by a multidetector CT-scan (Discovery 750 HD, GE Healthcare, Milwaukee, USA) and dissection after sacrifice. For cold exposure, mice were housed in a 6 °C cold room up to 24 h and body temperature was measured at the indicated time points. Glucose, pyruvate tolerance tests (GTT, PTT) and insulin injectionsAfter overnight starvation, mice were administered intraperitoneally with glucose (1.5 g.kg?1) or pyruvate (2 g.kg?1) and glycaemia was measured from tail blood during 2 h. To investigate insulin signalling in organs, mice were injected intraperitoneally with 150 mU/g of insulin (or PBS) 40 min before sacrifice, as previously validated [18]. Euglycemic-hyperinsulinemic clamps4 h fasted mice were anesthetized with intraperitoneal pentobarbital (80 mg.kg?1). As previously described [19], euglycemic-hyperinsulinemic clamps were performed, using insulin infusion at a dose suppressing hepatic glucose production (18 mU.kg?1.min?1), and the glucose infusion rate was measured. At steady state, a bolus of 2-deoxy-D-(1-3H)glucose (30 μCi) was injected to determine the in vivo glucose utilization index of insulin-sensitive tissues. 2-deoxy-D-(1-3H)glucose-6-phosphate in peripheral tissues was measured using a liquid scintillation analyzer (Tri-Carb 2900TR, Perkinelmer, MA, USA). Histological analysesTissues were fixed in 4% paraformaldehyde and 6 μm thin sections were stained with haematoxylin/eosin for morphological investigations. Quantifications were performed using the Metamorph software. Plasma and tissue analysesPlasma triglycerides (TGs) were determined by an automated Abott Architect analyzer (Abott Architect, Paris, France). Plasma glucose, insulin, non-esterified fatty acids (NEFA), lactate, ketone bodies and FGF21 levels, as well as liver content of TGs, glycogen and ketone bodies were measured with commercial kits. Real-time PCRRNA was extracted using Trizol according to the manufacturer’s instructions. 1 ug of RNA was reverse transcribed using a VILO kit. Quantitative RT-PCRs were performed using a SYBR green detector on a StepOne PCR system (Life Technologies, Carlsbad, USA). Primer sequences are listed in Supplementary Table 1. Western blot analysesHomogenized cells/tissues were lysed in ice-cold RIPA buffer. Proteins were resolved by 5–20% gradient SDS-PAGE and blotted onto nitrocellulose membranes. Proteins were detected with specific primary antibodies and HRP-conjugated secondary antibodies using chemoluminescence. Quantifications were performed using the ChemiDoc? XRS from Biorad (Cressier, Switzerland) and the Quantity One? Software. Statistical analysisResults are expressed as means ± SEM of at least 3 independent experiments or at least 4 different animals per group. Results were analysed by Student’s t test or two-way ANOVA followed by a Sidak’s multiple comparisons test when more than 2 groups or multiple time points were analysed. Values were considered significant when ?p <0.05, ??p <0.01 or ???p <0.001. ResultsHepatic steatosis in LPTEN KO mice is associated with increased glycolysis but decreased gluconeogenesis and glucose outputAs previously reported, LPTEN KO mice have an increased liver weight related to triglyceride (TG) accumulation in hepatocytes ( Supplementary Fig. 1 and [16]). We found two mechanisms contributing to excessive TG accumulation in the liver of LPTEN KO mice. First, mRNA expression levels of FA transporters, in particular Cd36, Fatp3, and Fabp1, were significantly upregulated, suggesting increased FA uptake from the bloodstream by PTEN-deficient hepatocytes ( Supplementary Fig. 2). Secondly, critical effectors promoting de novo lipogenesis were strongly overexpressed in the liver of LPTEN KO mice. In particular, the mRNA expression of Fas, Acc1, Acc2, Scd1, Pparγ, and Srebp1c was upregulated. In addition, protein expression of key enzymes involved in FA biosynthesis, i.e. FA synthase (FAS) and acetyl-CoA carboxylase (ACC), was also strongly increased in PTEN-depleted hepatic tissue. On the contrary, the general expression pattern of rate-limiting key enzymes controlling hepatic FA oxidation, lipolysis and export, as well as cholesterol metabolism, mainly remained unchanged with the exception of a few enzymes weakly up- or downregulated ( Supplementary Fig. 2). Hepatic steatosis is usually tightly associated with IR [20]. However, glucose tolerance tests (GTT) indicated that LPTEN KO mice paradoxically exhibited an improved glucose tolerance associated with hypoinsulinemia ( Supplementary Fig. 1C and D). The liver appeared to contribute in two ways to this improved glucose tolerance of LPTEN KO mice. Consistent with a boosted de novo lipogenesis ( Supplementary Fig. 2), glucose utilization was strongly promoted, as indicated by an increased mRNA expression of enzymes regulating glycolysis ( Fig. 1A). Secondly, the mRNA expression of key factors promoting gluconeogenesis, i.e. Pepck and Pgc1-α, was downregulated in LPTEN KO mice, suggesting an impairment of de novo glucose synthesis ( Fig. 1B). This was further confirmed by pyruvate tolerance tests (PTTs), showing that the pyruvate-dependent hepatic glucose output was strongly abrogated in LPTEN KO mice ( Fig. 1C). Although de novo glucose production was inhibited in LPTEN KO mice, we did not observe any changes, neither in the liver glycogen content, nor in ketone bodies, which can arise from increased pyruvate production through the glycolytic oxidative pathway. Plasma lactate levels remained unchanged as well ( Fig. 1D and E).
These results indicate that FA synthesis is fostered by an increase in glucose utilization, whereas inhibition of gluconeogenesis and glucose output contributes to the improved glucose tolerance in LPTEN KO mice. LPTEN KO mice display enhanced systemic insulin sensitivity and insulin-stimulated glucose uptake in skeletal muscleAlthough impaired hepatic gluconeogenesis and glucose output likely contribute to the improved glucose tolerance of LPTEN KO mice, insulin sensitivity and glucose uptake by peripheral organs, i.e. skeletal muscle and adipose tissues, are also important potential mechanisms to be considered. To address this issue, we first examined the phosphorylation/activation of AKT, a major insulin signalling effector, in peripheral organs of LPTEN KO vs. CTL mice injected with insulin. As shown in Fig. 2A, LPTEN KO mice, stimulated with insulin, displayed lower activation of the insulin receptor (INSR) in the liver, but higher basal and insulin-stimulated AKT phosphorylation, due to a lack of the PTEN antagonistic effect on PI3K signalling. Surprisingly, although PTEN expression was not altered in non-hepatic and metabolically active tissues of LPTEN KO mice ( Supplementary Fig. 1E and F) and despite the presence of higher TG levels in muscle ( Supplementary Fig. 5), phosphorylation of the INSR and its downstream effector, AKT, was significantly increased in muscles of LPTEN KO mice, indicating muscle insulin hypersensitivity ( Fig. 2B). Contrasting with skeletal muscle, AKT phosphorylation in white adipose tissue (WAT) of LPTEN KO mice was reduced, although INSR phosphorylation was unaffected ( Fig. 2C). To further evaluate the influence of skeletal muscle on the glucose tolerance of LPTEN KO mice, we performed euglycemic-hyperinsulinemic clamps under conditions of complete suppression of hepatic glucose production. We observed that the glucose infusion rate (GIR) measured at the end of the clamps was highly increased in LPTEN KO mice, confirming enhanced peripheral insulin sensitivity ( Fig. 2D). Consistent with this observation, insulin-induced glucose uptake was increased in almost all skeletal muscle types examined, while it remained unaffected in WAT ( Fig. 2E). These results indicate that hepatic PTEN-deficiency induces muscle insulin hypersensitivity, which importantly contributes to the improved glucose tolerance observed in LPTEN KO mice. Hepatic PTEN deletion decreases lipid storage in fat depotsAnalysis of the overall phenotypic characteristics of LPTEN KO mice showed that their body weight and food intake were unaltered compared to CTL mice ( Fig. 3A and B). The same was also the case for locomotor activity, energy expenditure and the respiratory exchange ratio (RER) measured by indirect calorimetry, as well as thermal regulation upon cold exposure ( Supplementary Fig. 3). Furthermore, when body composition was assessed by EchoMRI analysis, the total lean and fat mass remained similar in both groups ( Fig. 3C). However, LPTEN KO mice exhibited marked hepatic steatosis, as previously reported ( Supplementary Fig. 1 and [16]). In view of these results, the normal overall fat content of LPTEN KO mice suggested a decrease in adipose tissue depots. We therefore used quantitative CT-scan imaging to measure fat depot volumes. Data obtained indicated that the volumes of interscapular, subcutaneous and intraperitoneal WAT depots were drastically reduced in LPTEN KO mice ( Fig. 3D). This was further confirmed by accurate dissection and weighing of all visible fat depots in CTL and LPTEN KO mice ( Fig. 3E). On the contrary, the volume and weight of major brown adipose tissue depots (e.g. interscapular) were unchanged in LPTEN KO as compared to CTL mice.
We then analysed the mRNA expression of major effectors involved in lipid metabolism of mesenteric WAT (a representative WAT in the context of metabolic diseases) and did not detect any important and significant change in the expression of key enzymes regulating FA uptake or lipolysis (Supplementary Fig. 4A and C), consistent with the absence of significant changes in plasma TG and non-esterified fatty acid (NEFA) levels (Supplementary Fig. 5A). Although the RER (ratio of VCO2/VO2) tended to decrease and fat oxidation to increase (Supplementary Fig. 3C and D) during the diurnal period, the expression of critical rate-limiting enzymes, controlling fatty acid β-oxidation, was not significantly altered (Supplementary Fig. 4C). However, Western blot analyses of FAS and ACC, two major enzymes required for FA biosynthesis, revealed a decrease in the expression of these proteins in WAT of LPTEN KO mice with no change in their respective mRNAs ( Supplementary Fig. 4B). Reduction in FA esterification was also suggested by a decreased mRNA expression of Gpat1 ( Supplementary Fig. 4A). Together, these data indicate that PTEN-deficiency in the liver decreases adiposity through crosstalk mechanisms between the liver and WAT, preventing FA synthesis and esterification in WAT. Browning of mesenteric WAT occurs in LPTEN KO mice and correlates with increased FGF21 production and release by the liverRecent evidence indicates that browning of WAT (appearance of brown-like adipocytes called “beige cells” in WAT) may also importantly contribute to a decreased adiposity and improved metabolic status in mice [21]. Beige cells with high FA oxidation capacity in WAT are differentiated from classical white adipocytes by mainly two specific characteristics: (i) the expression of the uncoupling protein Ucp1 and (ii) multilocular lipid droplet structures, instead of a single large lipid droplet [21]. As shown in Fig. 4A, Ucp1 expression was increased in specific WAT depots of LPTEN KO mice, such as the mesenteric depot, indicating the presence of beige cells. Further histological analyses of this fat depot confirmed the presence of beige adipocytes foci, characterized by the presence of multilocular lipid droplets in LPTEN KO ( Fig. 4B). |
|
|
