Impaired Adipose Tissue Expansion Caused by LXR Activation is Associated with Insulin Resistance in HFD Mice

Liver X receptors (LXR) are deemed as potential drug targets for atherosclerosis, whereas a role in adipose tissue expansion and its relation to insulin sensitivity remains unclear. To assess the metabolic effects of LXR activation, C57BL/6 mice on a high-fat diet (HFD) were treated with the dual LXRα/β agonist T0901317 (30 mg/kg per day) for 3 weeks. Differentiated 3T3-L1 was used for analysing the effect of T0901317 on glucose uptake.T0901317 reduced fat mass, accompanied by a massive fatty liver and lower adipokine levels in circulation of HFD mice. Increased adipocyte apoptosis and macrophage infiltration were found in epididymal fat of T0901317-treated HFD mice. In addition, T0901317 treatment promoted basal lipolysis, but blunted the anti-lipolytic action of insulin. Furthermore, LXR activation antagonized PPARγ target genes in epididymal fat and PPARγ-PPRE binding activity in 3T3-L1 adipocytes. Although the glucose tolerance was comparable to that in vehicle-treated HFD mice, the insulin tolerance was significantly decreased in T0901317-treated HFD mice, indicating decreased insulin sensitivity by T0901317 administration, and which was further supported by impaired insulin signalling found in epididymal fat and decreased insulin-induced glucose uptake in 3T3-L1 by T0901317administration. These findings reveal that LXR activation impairs adipose expansion which contributes to decreased insulin sensitivity.


Introduction
Adipose tissue is a key regulator of energy balance, playing an active role in lipid storage and buffering, and synthesizing and secreting a wide range of fatty acids and adipokines into the circulation that influence systemic metabolism [1]. There is strong evidence showing that dysfunction of adipose tissue plays a critical role in the development of insulin resistance and diabetes mellitus [2][3]. A classical paradigm is that the more adipose tissue, the higher the prevalence of metabolic diseases, and it is this relationship that has interested researchers [4]. However, recent studies have suggested that the factor linking obesity and insulin resistance may not be the absolute amount of fat accumulated but the mismatch between energy surplus and storage capacity [5].
Allowing adipose tissue to store more lipids may prevent secondary metabolic complications caused by lipids being deposited in non-adipose organs. Thus, the ability of adipose tissue to expand and match the storage needs of energy surplus may be a key determinant in protection against the metabolic syndrome associated with obesity [6].
Adipose tissue mass is determined by processes governing adipocyte size and number [7]. The size of adipocytes increases because of increased storage of triglycerides from dietary sources or endogenous lipogenic pathways, whereas adipocyte number increases as a result of increased proliferation and differentiation [8].
Decreases in adipose tissue mass may involve the loss of lipids through lipolysis and the loss of mature fat cells through apoptosis [9]. Peroxisome proliferator-activated receptor (PPARγ) is a critical regulator of adipose tissue mass. The activation of PPARγ leads to adipocyte differentiation and fatty acid storage [10][11]. Moreover, the expansion of adipose tissue associated with obesity may be based on a hyperplasic response of the adipose tissue regulated by PPARγ rather than just on hypertrophy of the mature adipocytes, thus resulting in adipose tissue with smaller but more numerous adipocytes. These smaller adipocytes retain insulin sensitivity with the secretion of insulin-sensitizing adipokines. This occurs in a mouse model that is heterozygous for PPARγ, which shows improved insulin sensitivity and protection from lipotoxicity despite increased fat mass [12][13].
Liver X receptors (LXR), including two isoforms of α and β, are members of the nuclear receptor family.
Most studies of LXR have been performed in non-adipose tissue, showing a positive role of LXR in central metabolic pathway regulation including cholesterol, glucose and lipid homeostasis; however, results in recent years suggest that LXR may have important modulatory roles in adipose tissue [14][15]. However, studies have reported inconsistent or even contradictory results with regard to the role of LXR in regulating adipogenesis, lipogenesis and insulin sensitivity [16][17]. Recently, Beaven et al [18] have shown that loss of LXR impairs hepatic lipogenesis, accompanied by a reciprocal increase in adipose lipid storage, by promoting adipose PPARγ pathway activity, indicating possible cross-talk between LXR and PPARγ in adipose tissue. Cross-talk between LXR and PPARγ in adipose tissue is also supported by our previous study that showed an antagonizing effect of LXR on PPARγ in the regulation of adiponectin expression [19].
Activation of PPARγ in adipose tissue by thiazolidinedione ligands promotes adipose lipid storage and secondarily increases insulin sensitivity in liver and muscle [20]. However, the role of LXR in the alteration of adipose expansion and its relation to insulin resistance has not been studied. Furthermore, whether the possible interference of the PPARγ pathway in adipose tissue by LXR activation results in impairment of adipose expansion and thus impairs insulin sensitivity needs to be clarified in order to comprehensively understand the net outcome of the advantageous and disadvantageous effects of LXR activation.

LXR activation reduces fat mass accompanied by hepatomegaly and hypertriglyceridemia in mice fed on HFD
To determine the consequence of LXR activation in obesity, C57/BL6 mice were fed a high-fat diet for 12 weeks to induce obesity and subsequently treated for 3 weeks with dual LXRα/β agonist T0901317. Fig. 1a shows growth curves of C57/BL6 mice fed a high-fat diet or ND over a 15 week period. Mice on a high-fat diet had higher body weight than those on a ND from 8 weeks (week 8: ND=26.1±0.7 g; HFD=28.5±0.6 g; HFD+T0901317=28.7±0.8 g, P<0.05), whereas the body weight of HFD mice in the presence or absence of 3 weeks of T0901317 (30 mg/kg per day) treatment was comparable (week 15: HFD=31.4±0.7 g; HFD+T0901317=31±0.8 g, P>0.05, Fig. 1a). Consistent with the body weight, the food intake between them was also similar (Fig. 1b). Unexpectedly, T0901317 administration significantly decreased the fat mass of the epididymal (Fig. 1c), peri-renal and inguinal tissue, whereas liver weight in T0901317-treated HFD mice was doubled than that of the liver in HFD mice with pale appearance (Fig. 1d and Fig. 1e). In the histological analysis, the adipocyte size was significantly reduced in the T0901317-treated HFD mice ( Fig. 1f and Fig. 1g).
In consistent to the enlarged liver, histological staining with H&E ( Fig. 1h) and red oil (Fig. 1i) confirmed substantial lipid accumulations in the livers of T0901317-treated HFD mice. The serum TG level was significantly higher (Fig. 1j); whereas adipokines of adiponectin and leptin secretion into the circulation were significantly lower in T0901317-treated HFD mice than those in HFD mice ( Fig. 1k and Fig. 1l).

LXR activation promotes adipocyte apoptosis and inflammation in epididymal fat
In order to clarify whether the morphological changes in adipose tissue of T0901317-treated mice were associated with decreased cell numbers, apoptotic cells in the adipose tissue of epididymal fat were assessed.
There was no difference in apoptotic nuclei numbers in epididymal fat between mice on ND and HFD with vehicle treatment; however, increased apoptosis was indicated by a 3.5-fold increase in TUNEL immunoreactivity in nuclei in epididymal fat of T0901317-treated HFD mice as compared with that in vehicle HFD mice ( Fig. 2a and Fig. 2b). Consistently, protein levels of cleaved caspase 9, caspase 3, and PARP, typical molecules of the apoptotic pathway, were all increased by T0901317 treatment ( Fig. 2c and Fig. 2d). As seen in However, there was no difference in total collagen deposition or expressions (except MMP-2) between HFD mice with or without treatment T0901317 ( Fig. 2g and Fig. 2h).

LXR activation increases lipolysis and inhibits PPARγ mediated transcriptional activity
To better understand the mechanism by which T0901317 treatment decreases fat mass and causes hyperlipidemia, we next studied the basal lipolysis, as well as insulin induced anti-lipolytic action. In addition to HFD, fasting FFA levels in circulation was further increased by T0901317 administration in HFD-mice (Fig. 3a).
Perilipin1 (PLIN1), the most abundant lipid droplet-coating protein blocking basal lipolysis in adipocytes, was consistently decreased in epididymal fat of HFD mice, with the lowest level in T0901317-treated HFD mice ( Fig.   3b and Fig. 3c). Furthermore, mice on ND exhibited a nearly 35% drop in fasting FFA, reflecting insulin's ability to inhibit lipolysis. However, the FFA suppression to exogenous insulin was blunted in vehicle-treated HFD mice (P=0.076), which was further impaired by T0901317 administration (Fig. 3a). Insulin is known to suppress lipolysis by inactivating HSL; we thus investigated the phosphorylated forms of HSL levels in epididymal fat in HFD mice administrated with insulin. Indeed, HSL phosphorylation (s660) levels (but not s563 and s565) and its upstream protein of ATGL were significantly higher in epididymal fat of T0901317-treated HFD mice ( Fig. 3d and Fig. 3e).
We next measured the mRNA levels of several genes involved in lipogenesis in epididymal fat. As expected, HFD increased expression of the genes SREBP-1c and Fasn, instead of PPARγ, ACC and PCG-1α in epididymal fat. Surprisingly, T0901317 treatment further induced the expression of SREBP-1c, Fasn and PPARγ without affecting ACC and PCG-1α despite the fat mass reduction in HFD mice (Fig. 3f).
PPARγ is known to be a powerful promoter of lipogenesis and adipogenesis in adipose tissue, and it plays a critical role in adipose expansion. Based on the contrary pattern of PPARγ and fat mass reduction, we next investigated the effect of T0901317 administration on PPARγ target genes. As shown in Fig PPARγ-mediated transcriptional activity, and this finding was supported in vitro study. Fig.3h showed that the binding of PPARγ to PPRE was increased with PPARγ agonst of pioglitazone treatment, whereas T0901317 treatment decreased the activity in presence or absence of pioglitazone in 3T3-L1 adipocytes analyzed by electrophoretic mobility shift assays (EMSA).

LXR activation induces lipid accumulation in the liver
Generally, increased lipolysis and FFA secretion from adipose tissue leads to ectopic fat accumulation, and which was indeed found in T0901317 treated HFD mice showing enlarged liver and lipid accumulation (Fig. 1d,   1h and 1i). Consistent with lipid deposition, liver lipogenic genes of SREBP-1c and Fasn were significantly increased. In contrast to epididymal fat, the LPL gene, responsible for hydrolysis of lipoprotein-bound triglycerides supplying fatty acids for liver, was significantly induced by T0901317 administration. By contrast, CD36, the gene for fatty acid transport, was not changed by T0901317 treatment (Fig. 4a), suggesting that a decreased fatty acid influx in adipose tissue might facilitate more fatty acid flow to the liver resulting in triglyceride synthesis. Furthermore, genes responsible for fatty acid oxidation including CPT1α and PPARα were decreased, supporting lipid deposition in the liver (Fig. 4b).

LXR activation induces insulin resistance in epididymal fat and in the whole body
The change of morphology and metabolism of adipose tissue is usually associated with impaired glucose utilisation in adipose tissue, leading to impaired glucose homeostasis in the whole body; therefore, we studied whether LXR affected the glucose response and insulin sensitivity in T090131-treated HFD mice.
Fifteen weeks of HFD successfully induced diabetes and insulin resistance in C57/BL6 mice; these mice showed a significantly higher glucose and insulin response in IPGTT than mice on ND (Fig. 5a-d). Although the glucose response was comparable in IPGTT between T0901317-treated and vehicle-treated HFD mice ( Fig. 5a and Fig. 5b), the insulin response of T0901317-treated mice was significantly higher than that of vehicle-treated mice, implicating a decrease in insulin sensitivity with T0901317 treatment (Fig. 5c and Fig. 5d). The decreased 9 insulin function was further confirmed by ITT. As shown in Fig. 5e and Fig. 5f, the area under curve of glucose was significantly higher in T0901317-treated HFD mice after insulin overloading, indicating decreased whole-body insulin action by LXR activation.
To further clarify the effect of LXR on glucose uptake in adipose tissue, we studied the glucose utilization in vitro. In differentiated 3T3-L1 adipocytes, both 1 and 10 μM T0901317 treatments decreased basal glucose uptake, as well as diminished the glucose uptake increase stimulated by 100 nM insulin (Fig. 5g).
In accord with these results, the typical markers of insulin signalling activity, Akt phosphorylation and Glut-4 protein translocation to the membrane after insulin loading, were markedly decreased in EP fat of T0901317-treated mice ( Fig. 5h and Fig. 5i). Consistent with increased insulin levels, histological analysis revealed enlarged and an increased number of pancreatic islets in T0901317-treated mice ( Fig. 5j and Fig. 5k).

Discussion
The present study demonstrates that LXR activation causes a reduction in adipose tissue mass but results in massive fatty livers, accompanied by increased adipocyte apoptosis and lipolysis, as well as decreased PPARγ mediated transcriptional activity in adipose tissue. Moreover, impaired adipose expansion by LXR activation is associated with decreased insulin signalling in adipose tissue and decreased insulin sensitivity of the whole body in HFD-fed mice. [18,21] who described that global LXR deletion in the setting of obesity shifts the programme of de novo lipogenesis from the liver to adipose tissue, we found an opposite phenotype, showing that LXR activation caused reductions in fat pads but liver enlargement and fatty steatosis. Loss of adipocytes through apoptosis by LXR activation might be an important process leading to fat mass reduction in the present study. Apoptosis of adipose tissue is relatively poorly studied compared to that in other tissues; yet increased adipocyte apoptosis has been recently proposed to contribute to obesity, and to differences in regional fat distribution or expansion and insulin resistance, in both obese animals and humans [9,[22][23]. As shown in the present study and others [9,[22][23], increased adipocyte apoptosis usually results in macrophage infiltration and inflammation, which are associated with insulin resistance. Furthermore, insulin resistance induced by inflammatory factors is reversed by interference with apoptosis initiation via CASP3/7 inhibition. This finding is supported by Archer et al [26] who showed increased SREBP-1c expression in visceral fat but lower visceral fat mass. These results suggest that at least in rodent adipose tissue, LXR and SREBP-1c are not the primary regulator of lipogenesis; this conclusion is supported by data from SREBP-1c -/mice and mature adipocytes isolated from LXRα/β -/mice, both of which showed a similar level of lipogenesis to that in their wild-type littermates [27].

Consistent with the findings of Beaven et al and Korach-André and co-workers
However, increased basal lipolysis supported by increased fasting NEFA levels in the circulation by T0901317 treatment might contribute to reduced adipocyte size and fat mass by LXR activation. Similar treatment with GW3965 resulted in smaller fat cells [26], indicative of increased triglyceride utilisation. An effect of LXR on basal lipolysis is also supported by findings in human adipocytes and adipose tissue LXRα knockout (ATaKO) mice [28][29]. Down-regulation of lipid droplet-coating proteins of PLIN1, as found in our present study might be the molecular mechanism, for low levels or absence of PLIN1 has been implicated in enhanced spontaneous lipolysis both in mice and humans [30][31][32]. In addition, we found that the insulin induced anti-lipolytic action was significantly blunted by LXR activation. Insulin is known to suppress lipolysis by inactivating HSL [33]. In T0901317-treated HFD mice, we did find HSL phosphorylation and its upstream molecular of ATGL were higher than those in vehicle-treated HFD mice with same dose of insulin administration, supporting the reduction of fat mass was associated with lipolysis. Furthermore, expression levels of CD36 and LPL (proteins involved in lipid clearance from the circulation and deposition in adipose tissue as triglycerides) were decreased, which might contribute to increased TG and NEFA levels in the circulation and decreased fat mass in T0901317-treated HFD mice. PPARγ is critically required for adipose tissue expansion by increasing lipogenesis and adipocyte proliferation. Indeed, consistent with reductions in fat mass and increases in adipocyte apoptosis, we found that PPARγ mediated transcriptional acitivity was inhibited by LXR, which was supported by the reduced PPARγ-PPRE binding activity in T0901317-treated 3T3-L1 cell, as well as the inhibited PPARγ target genes expression in epididymal fat in T0901317-treated HFD mice. Although the mechanism of LXR interfering with the PPARγ pathway was not further studied, the fact that LXR shares its heterodimerising partner (that is RXR) with PPARs and sometimes competes for the same DNA response elements suggests that LXR could affect PPAR signalling [34][35].
Impaired adipose expansion was generally associated with ectopic lipid accumulation and insulin resistance.
Indeed in the present study, moderate hepatomegaly and steatosis, as well as higher insulin levels, decreased insulin sensitivity and enlarged pancreatic islets were consistent with other animal models of impaired adipose expansion and human lipoatrophy or lipodystrophy [36][37]. The role of LXR activation in fatty acid synthesis in the liver is well established [38][39]. Indeed, our study also found that typical genes of lipogenesis, including SREBP-1c and FAS, were significantly induced by T0901317 administration. Importantly, we further found increased LPL and decreased CTP1α and PPARα in the liver, suggesting an increased capacity of fatty acid intake and decreased fatty acid oxidation in the liver, which might also be another mechanism for hepatic steatosis by LXR activation. It is worth pointing out that pancreatic islets in T0901317-treated mice generated double the insulin levels, which at least suggests insulin resistance instead of a direct stimulating effect of LXR activation, for blood glucose responses in T0901317-treated mice were comparable instead of lower than those in HFD mice, and this finding was supported by ITT results which showed that T0901317-treated mice had a blunted glucose response. Decreased insulin efficiency was most likely associated with increased NEFA levels and decreased levels of the insulin-sensitising adipokines adiponectin and leptin.
Glucose uptake in adipose tissue impacts on whole-body glucose homeostasis. The effect of T0901317 on glucose utilisation in adipocyte is not consistent [40][41][42]. In our study, we found that T0901317 decreased both basal and insulin-stimulated glucose uptake in 3T3-L1 cells. Decreased glucose uptake in adipose tissue was supported by insulin-induced Akt activity and Glut-4 membrane translocation in T0901317-treated mice. This result is consistent with the in vitro findings of other studies either using the same cell model or in primary human adipocytes [40,42]. While other studies have shown unchanged or increased basal glucose uptake, the discrepancies between these studies could result from differences in agonists, in vitro cell models and treatment conditions. Despite decreased glucose uptake in adipocytes, whole-body glucose homeostasis was not changed in T0901317-treated mice. This finding probably resulted from the compensatory increase in insulin secretion and pancreatic islet enlargement. The relatively stable glucose metabolism might also be due to decreased gluconeogenesis by LXR activation [41,[43][44]. A few studies addressing the role of LXR in carbohydrate metabolism have shown improved glucose tolerance by LXR activation in vivo; however, no study has investigated the fat mass change. No change of systemic glucose homeostasis was shown in ob/ob female mice with T0901317 treatment in the study by Archer et al, in which they also found decreased visceral fat mass with GW3965 treatment [26]. Thus, the adipose tissue loss in these studies might be the reason for the discrepancies. the liver and aberrant PPARγ signalling in adipose tissue, leading to unexpected metabolic consequences when administering LXR activators in vivo.

Reagents
The dual LXRα/β agonist T0901317 was purchased from Cayman Chemical Company ( After 12 weeks, the HFD group was further randomised into two groups of 12 mice. HFD mice were treated for 3 weeks with 30 mg/kg T0901317 per day (n = 12) or the vehicle (3% DMSO in PBS, n = 12) by i.p injection.
The ND control group also received the same vehicle treatment by i.p injection. Body weight was recorded once a week and food intake was monitored every day.

Culture and differentiation of 3T3-L1 cells
3T3-L1 cells were purchased from the American Type Culture Collection (ATCC, Rockefeller, MD,USA) and were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS (Bio-rad, Hercules, CA, USA) and 100 IU/ml penicillin/streptomycin at 37°C in an atmosphere of 5% CO2 and 95% humidity. Two days post-confluent cells (designated as Day 0) were induced to differentiate into adipocytes by the addition of differentiation mixture with DMEM containing 10% FBS, 10 μg/mL insulin, 0.5 mM 3-isobutyl-1-methylxanthine (IBMX) and 1 μM dexamethasone. Two days later, culture medium was changed to DMEM supplemented with 10% FBS and 10 μg/mL insulin for 2 days. The medium was then replaced every other day with DMEM containing 10% FBS for different periods until day 10.

In vivo glucose homeostasis assays
After 3 weeks of treatment, intraperitoneal glucose tolerance tests (IPGTT) and intraperitoneal insulin tolerance tests (ITT) were carried out. Half of the mice (n = 5-6 per group) were injected i.p. with glucose (1.5 g/kg body weight) following an overnight fast and blood glucose levels were tested from tail blood using One Francisco, CA, USA). All kits were used according to the manufacturer's protocols.

Histological analysis of adipose tissue and liver
Adipose, liver and pancreatic tissues fixed in 4% formaldehyde were then embedded in OCT compound and cut into sections (thickness, 4 μm) according to a standard protocol. The sections were stained with hematoxylin and eosin (H&E) and examined under a light microscope. Size of adipocytes was analysed using Image Pro-plus

Quantitative real-time RT-PCR
Total RNA was isolated using TRIzol (Invitrogen, Grand Island, NY, USA) and reverse-transcribed with random hexamers by using TaqMan reverse-transcription reagents kit (Applied Biosystems Inc., San Francisco, CA, USA) according to the manufacturer's protocol. Real-time PCR was performed using the 7500 real-time PCR system (Applied Biosystems) and SYBR Green qPCR Kit (TaKaRa, Dalian, China). Relative expression was normalised to that of GAPDH as an internal control for quantification of individual mRNA species and calculated using the formula 2(−ΔΔCt). Primer sets were listed in Table S1.

Electrophoretic mobility shift assay
Differentiated 3T3-L1 cells were treated with DMSO, 10 μM T0901317 and 3 μM pioglitazone or co-treated with 10 μM T0901317 and 3 μM pioglitazone for 24 hr, and nuclear protein was extracted and quantified. EMSA/Gel-Shift binding buffer and nuclease-free water for 20 min at room temperature in a total volume of 10 µl. Unlabelled oligo probe was used for the cold probe competitive reaction, and unlabelled mutant olig probe was used as the mutation probe for the cold competitive reaction. Anti-rabbit PPARγ antibody (Cell Signaling Technology, Denver, MA, USA) was added for super-shift reaction. The reaction mixture was subjected to electrophoresis (100 V in 0.5 × Tris-buffered EDTA solution at room temperature) using 8% nondenaturing polyacrylamide gels, then transferred to a positively charged nylon membrane (Beyotime Biotechnology, Haimen, China) , UV cross-linked and blocked. After incubation with 5 μl streptavidin-HRP conjugate, immunoreactive proteins were detected using a chemiluminescent ECL assay kit (Millipore).

Statistical analysis
Data are expressed as means ± SEM (in vivo studies) or means ± SD (in vitro studies). Differences between the means of individual groups were analysed with independent t-tests or one-way ANOVA and LSD multiple range tests. Two-way repeated measures were used for comparisons between glucose and insulin levels of IPGTT or ITT using the statistical software package SPSS 16.0. A significant difference was defined as P < 0.05.
Each in vitro experiment was conducted in triplicate.