Leptin administration restores the altered adipose and hepatic expression of aquaglyceroporins improving the non-alcoholic fatty liver of ob/ob mice

Glycerol is an important metabolite for the control of lipid accumulation in white adipose tissue (WAT) and liver. We aimed to investigate whether exogenous administration of leptin improves features of non-alcoholic fatty liver disease (NAFLD) in leptin-deficient ob/ob mice via the regulation of AQP3 and AQP7 (glycerol channels mediating glycerol efflux in adipocytes) and AQP9 (aquaglyceroporin facilitating glycerol influx in hepatocytes). Twelve-week-old male wild type and ob/ob mice were divided in three groups as follows: control, leptin-treated (1 mg/kg/d) and pair-fed. Leptin deficiency was associated with obesity and NAFLD exhibiting an AQP3 and AQP7 increase in WAT, without changes in hepatic AQP9. Adipose Aqp3 and hepatic Aqp9 transcripts positively correlated with markers of adiposity and hepatic steatosis. Chronic leptin administration (4-weeks) was associated with improved body weight, whole-body adiposity, and hepatosteatosis of ob/ob mice and to a down-regulation of AQP3, AQP7 in WAT and an up-regulation of hepatic AQP9. Acute leptin stimulation in vitro (4-h) induced the mobilization of aquaglyceroporins towards lipid droplets (AQP3) and the plasma membrane (AQP7) in murine adipocytes. Our results show that leptin restores the coordinated regulation of fat-specific AQP7 and liver-specific AQP9, a step which might prevent lipid overaccumulation in WAT and liver in obesity.

FFA and glycerol release from the adipose tissue 24,27 . We 12 and others [28][29][30] have reported that aquaglyceroporins AQP3 and AQP7 facilitate glycerol outflow from adipocytes in response to the lipolysis induced by the β -adrenergic agonist isoproterenol. Thus, in the present study, the direct effect of acute leptin treatment on aquaglyceroporin expression was analyzed by real-time PCR and Western blot in murine differentiated subcutaneous adipocytes. Upon 24-h leptin stimulation, Aqp3 mRNA tended to decrease (P = 0.072) and Aqp7 gene expression was down-regulated (P < 0.05) in murine subcutaneous adipocytes (Fig. 1A,B). Moreover, both AQP3 and AQP7 protein levels were reduced (P < 0.05) after leptin treatment (Fig. 1C,D). To gain more insight into the regulation of aquaglyceroporins by leptin, the subcellular localization of AQP was studied in differentiated 3T3-L1 adipocytes by confocal immunofluorescence microscopy (Fig. 1E). We previously described that after subcellular fractionation of quiescent 3T3-L1 adipocytes, AQP3 was located in the plasma membrane and cytosolic fraction, whereas AQP7 was expressed in the subfractions of lipid droplets and the rest of the cytoplasm 12 . In the present study, we confirmed that, under basal conditions, AQP3 was present mainly in the cell surface, although some punctuate labelling in the cytoplasm could also be observed, while AQP7 resided predominantly in the cytoplasm, surrounding lipid droplets of differentiated 3T3-L1 adipocytes. After 4-h leptin stimulation, AQP3 tended to surround lipid droplets more prominently, whereas AQP7 was translocated to the plasma membrane.
In order to test the functionality of aquaglyceroporins on the lipolytic effect triggered by leptin, murine subcutaneous adipocytes were exposed to leptin 10 nmol/L for 24 h in the presence of HgCl 2 , a nonspecific AQP inhibitor 31 , or to CuSO 4 , a more selective AQP3 inhibitor 32 , prior to determination of glycerol release to the culture media. The inhibition of AQP permeability with 0.3 mmol/L HgCl 2 alone induced a modest decrease in glycerol release in murine subcutaneous adipocytes (control 3.18 ± 0.19 vs. HgCl2 3.06 ± 0.30 mg/dL, P = 0.729). Nonetheless, mercury ions abolished around 50% of the leptin-induced glycerol release in murine subcutaneous adipocytes, while copper ions inhibited approximately 20% of the glycerol release caused by leptin (Fig. 1F). These data suggest that the major glycerol channel in murine adipocytes, AQP7 and, to a lesser extent, AQP3 mediate the glycerol efflux triggered by leptin in fat cells.
Chronic leptin administration in vivo reduces adiposity in parallel to a decrease in aquaglyceroporins AQP3 and AQP7 in adipose tissue. Leptin is an adipokine that reduces food intake and increases energy expenditure to maintain energy balance 33 . As expected, leptin-deficient ob/ob mice exhibited severe obesity and hyperphagia (Table 1). Chronic leptin treatment corrected the obese phenotype of ob/ob mice, as evidenced by the lower body weight as well as epididymal, subcutaneous and perirenal fat mass via the reduction of food intake and the increase in rectal temperature. In the present study, chronic leptin administration was associated with a decrease in circulating FFA and glycerol, pointing to a lower lipolytic rate in leptin-treated animals. (E) Immunocytochemical detection of the AQP3 and AQP7 proteins in differentiated murine 3T3-L1 adipocytes (day 10) under basal conditions (upper panels) and after the stimulation for 4 h with leptin (10 nmol/L) (lower panels). Images were taken from the basal planes of the cells. Representative images of at least three separate experiments are shown. (F) Glycerol release from murine subcutaneous adipocytes under basal conditions (control) and after leptin (10 nmol/L)-induced stimulation without or with preincubation with HgCl 2 (0.3 mmol/L), a nonspecific AQP inhibitor, or with CuSO 4 (0.1 mmol/L), a selective AQP3 inhibitor. Differences between groups were analyzed by Student's t test or one-way ANOVA followed by Tukey's test. *P < 0.05 vs. control unstimulated cells; † †P < 0.01 vs. adipocytes stimulated with leptin. To analyze the potential involvement of aquaglyceroporins in the changes observed on adiposity after chronic exogenous leptin administration (4 weeks), we first assessed the gene and protein expression of AQP3 and AQP7 in subcutaneous WAT of the experimental groups by real-time PCR, Western blot and immunohistochemistry (Fig. 2). As illustrated in Fig. 2A,B, the tissue distribution of AQP3 and AQP7 showed a predominant immunostaining in the stromovascular fraction and lower expression in mature adipocytes, as previously reported by our group and others 12,34 . In the multiple lineal regression analysis, AQP3 and AQP7 protein levels in subcutaneous WAT contributed independently to 51.0% (P < 0.05) and 51.2% (P < 0.05) to the circulating glycerol concentrations after controlling for body weight, suggesting an important role of these aquaglyceroporins in glycerol efflux from adipose tissue.
Leptin deficiency was associated with higher mRNA and protein levels of AQP3 and AQP7 in subcutaneous WAT (Fig. 2C-F). In line with these results, Aqp3 and Aqp7 mRNA levels were positively associated with markers of adiposity [body weight (r = 0.33, P = 0.025 and r = 0.44, P = 0.001) or subcutaneous WAT/body weight (r = 0.33, P = 0.025 and r = 0.53, P = 0.001)] and hepatosteatosis [liver/ body weight (r = 0.36, P = 0.013 and r = 0.35, P = 0.010 and intrahepatic TG (r = 0.40, P = 0.006 and r = 0.53, P < 0.001)]. No differences in the transcript levels of Aqp3 and Aqp7 were detected after leptin administration, but a tendency towards a down-regulation of both glycerol channels was observed in leptin-treated ob/ob mice. Nonetheless, at the protein level, both leptin administration and caloric restriction reduced (P < 0.05) AQP3 and AQP7 in subcutaneous WAT of wild type and ob/ob mice.
Exogenous leptin replacement reduces the hepatic steatosis of ob/ob mice and upregulates AQP9 expression in the liver. Leptin-deficient ob/ob mice showed an increased (P < 0.0001) liver weight that was significantly reduced (P < 0.0001) by either caloric restriction or leptin replacement (Fig. 3A). Histological sections of leptin-deficient ob/ob mice were characterized by the presence of severe macrovesicular steatosis, but not advanced inflammation/fibrosis, that was completely reverted after leptin administration for 28 days (Fig. 3E). The analysis of intrahepatic triacylglycerol content revealed elevated TG levels (P < 0.001) in the liver of ob/ob mice that was prevented by leptin treatment (P < 0.05), but not by caloric restriction (Fig. 3B).
We next analyzed the expression of AQP9, the primary route for glycerol uptake in murine hepatocytes, by real-time PCR, Western blot and immunohistochemistry. As previously described by our group 22 , two immunoreactive bands of 30-32 kDa, corresponding to the core and N-glycosylated forms of AQP9 protein, respectively, were observed in the immunoblots (Fig. 3D). Leptin deficiency was associated with similar expression of AQP9 mRNA and whole (glycosylated and non-glycosylated) AQP9 protein signal than that observed in wild type mice, with leptin administration and caloric restriction increasing (P < 0.05) AQP9 gene and protein expression (Fig. 3C,D). Aqp9 gene expression was positively associated with markers of adiposity [body weight (r = 0.60, P < 0.001) or subcutaneous WAT/ body weight (r = 0.44, P = 0.002)] and hepatic steatosis [liver/body weight (r = 0.69, P < 0.001) and intrahepatic TG (r = 0.28, P < 0.05)]. Liver sections showed a strong immunoreactivity for AQP9 after leptin infusion, which was mainly localized in the plasma membrane of hepatocytes around the central veins (Fig. 3E). Positive association of PPARγ with changes observed in the expression of aquaglyceroporins in adipose tissue and liver after leptin replacement. Peroxisome proliferator-activated receptor γ (PPARγ ) represents a well-known lipogenic factor and, importantly, putative peroxisome proliferator response elements (PPRE) are present in the promoters of Aqp3 and Aqp7 genes 35,36 . In line with the observed excess adiposity and hepatic steatosis, leptin-deficient mice exhibited higher Pparg mRNA levels in the adipose tissue and liver that were reduced by leptin replacement and, to a lesser extent, by caloric restriction (Fig. 4A,B). As expected, gene expression levels of Pparg in subcutaneous WAT and liver were positively associated with markers of obesity [body weight (r = 0.43, P < 0.001 and . Moreover, a strong positive association was found between Pparg transcript levels and Aqp7 mRNA in the adipose tissue as well as with Aqp9 mRNA in the liver (Fig. 4C,D). Pparg mRNA was also correlated with Aqp3 gene expression in subcutaneous WAT but to a lower extent (r = 0.43, P < 0.001).
To gain further insight into the plausible association of PPARγ with these glycerol channels after leptin treatment, we examined the effect of leptin stimulation on basal and PPARγ agonist rosiglitazone-induced expression of aquaglyceroporins in murine subcutaneous differentiated adipocytes and AML12 hepatocytes. As expected, rosiglitazone stimulation for 24 h upregulated 1.4-and 2.0-fold the transcription of Pparg gene in murine subcutaneous adipocytes and AML12 hepatocytes, respectively, although no  hepatocytes. The gene expression in vehicle-treated wild type mice or unstimulated cells was assumed to be 1. Differences between groups were analyzed by two-way ANOVA or one-way ANOVA followed by Tukey's post-hoc test, where appropriate. *P < 0.05; **P < 0.01 vs. control unstimulated cells.
statistical differences between groups were found in fat cells (P = 0.269) (Fig. 4E,F). Moreover, the treatment with this TZD also increased the transcription of Aqp7 in subcutaneous fat cells and of Aqp9 in AML12 hepatocytes (Fig. 4G,H). The co-incubation with leptin tended to reduce both basal and TZD-induced mRNA expression of Pparg and Aqp7 genes in subcutaneous adipocytes, although changes fell out of statistical significance (P = 0.083 and P = 0.125, respectively). A similar trend was observed for the effect of leptin on basal and rosiglitazone-induced expression of Aqp3 gene in subcutaneous adipocytes (control 1.0 ± 0.4 A.U.; leptin 0.4 ± 0.1 A.U.; TZD 4.3 ± 1.6 A.U.; TZD + leptin 2.7 ± 0.8 A.U.; P = 0.003). However, leptin co-treatment induced a slight down-regulation of Pparg transcript levels (P = 0.292), while increasing (P < 0.05) the transcription of Aqp9 in AML12 hepatic cells.

Discussion
Adipocyte lipolysis is the process that controls the breakdown of TG into glycerol and FFA, which are released into the circulation and used as energy substrates in metabolic organs 7,37 . AQP3 and AQP7 facilitate glycerol outflow from adipocytes in response to β -adrenergic receptor-stimulated lipolysis via its translocation from the cytosolic fraction (AQP3) or lipid droplets (AQP7) to the plasma membrane 12,28,29 . Basal lipolytic activity of adipocytes is conditioned not only by catecholamines, but also by other factors, such as atrial natriuretic peptides, insulin, leptin, adenosine, tumor necrosis factor α (TNF-α ) or neuropeptide Y, among others 7 . The adipokine leptin exerts an autocrine/paracrine lipolytic effect on murine adipocytes 27 . In this sense, acute leptin treatment (1 h) reportedly increases basal lipolysis of wild type and ob/ob mice 27 . Here, we found that acute leptin treatment (4 h) stimulated AQP3 translocation from the plasma membrane to lipid droplets, a step that might reflect the glycerol efflux from lipid droplets after lipolytic response in differentiated subcutaneous murine adipocytes. Upon leptin stimulation, AQP7 was translocated from lipid droplets to the plasma membrane, and this finding suggests that this glycerol channel constitutes the main gateway for glycerol secretion to the bloodstream. Thus, we speculate that acute leptin treatment induces the translocation of AQP3 and AQP7 to lipid droplets and the plasma membrane, respectively, to facilitate glycerol mobilization after lipolysis. Nonetheless, the existence of further operative glycerol channels in subcutaneous adipocytes cannot be discarded. Obesity is associated with increased lipolysis due to higher lipolytic activity of β 3 -adrenergic receptors and reduced anti-lipolytic action of insulin, leading to elevated circulating concentrations of FFA and glycerol 38,39 . In the present study, we found that leptin-deficient obese ob/ob mice showed increased circulating glycerol together with higher subcutaneous fat expression of AQP3 and AQP7. Both chronic leptin treatment and caloric restriction significantly decreased circulating glycerol and AQP3 and AQP7 proteins in subcutaneous adipose tissue in ob/ob mice. The adipose tissue is composed not only by adipocytes, but also by SVFCs (i.e., macrophages, T lymphocytes, endothelial cells, fibroblasts, vascular smooth muscle cells or mesenchymal stem cells). Because SVFCs might contribute to the reduction of aquaglyceroporins in adipose tissue, we also studied the direct effect of leptin treatment on differentiated murine subcutaneous adipocytes. In line with the results obtained with the whole adipose tissue, 24-h leptin treatment decreased the gene and protein expression of AQP3 and AQP7 of differentiated murine subcutaneous adipocytes. In this regard, in a previous study, we found that in vitro 24-h leptin treatment downregulated AQP7 protein expression in differentiated human adipocytes via the PI3K/Akt/mTOR signalling pathway 12 . Taken together, both in vivo chronic leptin administration and caloric restriction limit glycerol release from adipocytes through the down-regulation of AQP3 and AQP7, suggesting a negative feedback regulation in lipolytic states to maintain intracellular glycerol and, therefore, to avoid the depletion of fat stores (Fig. 5).
Liver steatosis is a multi-factorial disease where abnormal TG accumulation in the hepatocytes can be triggered by metabolic, toxic, pharmacological or viral insults across a genetic predisposition 1,2 . Glycerol-3-phosphate constitutes a key metabolite for de novo synthesis of TG and derives from glycolysis, glyceroneogenesis as well as recycling of glycerol by GK 40,41 . AQP9 represents the main facilitative pathway for glycerol uptake as a substrate for gluconeogenesis and lipogenesis in hepatocytes [15][16][17] . Interestingly, a decrease in hepatic AQP9 and glycerol permeability has been observed in murine and human NAFLD, suggesting a defensive mechanism to prevent further development of hyperglycemia and hepatosteatosis 19,22,23 . Moreover, a dysregulation of AQP9 has been observed in several hepatic inflammatory derangements, such as extrahepatic cholestasis, alcoholic steatohepatitis and NASH 23,[42][43][44] . However, little is known about the regulation of AQP9 in the context of NAFLD/NASH. In the present study, AQP9 was mainly localized in the sinusoidal domain of the plasma membrane of hepatocytes, which is in agreement with previous results 45,46 including ours 12,23 . Leptin-deficient mice, a murine model of NAFLD, displayed macrovesicular steatosis without changes in hepatic AQP9 mRNA and protein. In a previous study, a lower expression of AQP9 was found in liver samples of ob/ob mice 22 . In this regard, AQP9 expression in the liver is influenced by the degree of hepatic steatosis and inflammation 23 that might change the expression of this aquaglyceroporin during the ongoing NAFLD in adult ob/ob mice. Short-term leptin administration has been reported to exert profound effects on hepatic lipid metabolism of ob/ob mice by reducing de novo lipogenesis via repressing acetyl-CoA carboxylase (ACC), fatty acid synthase (FAS) or stearoyl-coenzyme A desaturase 1 (SCD1) expression, and through the activation of β -oxidation by increasing the transcript levels of acetyl-coenzyme A acetyl-transferase 1 (ACAT1) or carnitine palmitoyl transferase 1 (CPT1) 47 . We herein show that chronic leptin administration completely rescues the hepatosteatosis of ob/ob mice as evidenced by the normalization of intrahepatocellular Scientific RepoRts | 5:12067 | DOi: 10.1038/srep12067 hepatocytes and liver morphology. Moreover, a valuable result of this work regards the up-regulation of AQP9 after chronic leptin treatment in wild type and ob/ob mice. Taken together, similar or lower levels of AQP9 associated to leptin deficiency appear to reflect a defensive cell reaction of the steatotic hepatocyte. Interestingly, chronic leptin administration not only rescues the fatty liver, but also increases AQP9 in order to facilitate glycerol import into hepatocytes for maintaining the glycemia as well as an appropriate lipid metabolism.
The adipose tissue and liver from leptin-deficient ob/ob mice showed an induction of PPARγ , which is a critical transcription factor for the development of obesity and hepatic steatosis as previously reported by other authors [48][49][50] . In a previous study of our group 51 , we found that the downregulation of PPARγ in adipose tissue and liver of diet-induced obese rats after bariatric surgery was strongly associated with a reduction in the transcription of aquaglyceroporins in these tissues. Nonetheless, the molecular mechanisms underlying this association were unclear. In the present study, we found that chronic leptin administration significantly decreased Pparg transcript levels in parallel with the improvement of adiposity and fatty liver. Interestingly, the promoters of Aqp3 and Aqp7 genes present putative PPRE with the expression of these aquaglyceroporins being up-regulated by PPARγ agonists 35,36 . In line with this observation, Pparg transcript levels were positively correlated with Aqp3 and Aqp7 in adipose tissue, but also with Aqp9 in the liver. Moreover, leptin co-treatment tended to reduce the transcription of PPARγ and AQP7 induced by rosiglitazone stimulation, a well-known PPARγ -selective agonist, in murine subcutaneous adipocytes. Our results are in agreement with other reports showing that pioglitazone and rosiglitazone administration to rodents increase the expression of AQP7 in adipose tissue 35,52 . However, leptin increased both basal and rosiglitazone-induced transcription of Aqp9 in AML12 hepatocytes, despite inducing a slight reduction Pparg mRNA levels in these hepatic cells. Thus, the mild action of leptin on rosiglitazone-induced up-regulation of aquaglyceroporins in adipocytes and hepatocytes suggests that other upstream molecules in addition to PPARγ might be involved in the regulatory effect of this adipokine.
The coordinated regulation of adipose and hepatic aquaglyceroporins is extremely relevant to maintain the control of fat accumulation and glycemia (Fig. 5) 12,18 . We herein report, for the first time, that chronic leptin administration regulates the altered expression of the adipose glycerol channels AQP3 and AQP7 and the liver-specific AQP9 in leptin-deficient obese ob/ob mice. Since glycerol is a key metabolite for lipid accumulation in fat depots and liver, the improvement of glycerol availability might be involved in the beneficial effects of leptin on obesity and NAFLD. Nonetheless, future in vivo studies are needed to fully demonstrate the requirement of AQP proteins for the improvement of these pathologies. Moreover, the time functional link between the regulation of AQP and leptin-dependent changes in lipid flux at the clinical level require the exact characterization of NAFLD and more advanced liver damage stages in patients with respect to weight changes and diet.

Methods
Animals. Ten-week-old male wild type (C57BL/6J) (n = 30) and genetically obese ob/ob mice (C57BL/6J) (n = 30) (Harlan Laboratories Inc., Barcelona, Spain) were housed in a room with controlled temperature (22 ± 2 °C), and a 12:12 light-dark cycle (lights on at 08:00 am). Wild type and ob/ob mice were divided in control, leptin-treated (1 mg/kg/d) and pair-fed groups (n = 10 per group), as previously described 26 . The control and pair-fed groups received vehicle (PBS), while leptin-treated groups were intraperitoneally administered with leptin (Bachem, Bubendorf, Switzerland) twice a day at 08:00 and 20:00 for 28 days. Control and leptin-treated groups were provided with water and food ad libitum with a rodent maintenance diet (12.1 kJ: 4% fat, 82% carbohydrate and 14% protein, Diet 2014S, Teklad Global Diets, Harlan, Barcelona, Spain), while the daily food intake of the pair-fed groups was matched to the amount eaten by the leptin-treated groups the day before to discriminate the inhibitory effect of leptin on appetite. All experimental groups had an isoproteic intake consuming similar amounts of sodium and phytates 53 . Body weight and food intake were daily registered and rectal temperature was measured using a thermoprobe (YSI 4600 Series Precision Thermometers, YSI Temperature, Dayton, OH, USA) at the end of the experiment. Animals were sacrificed on the 28 th day of treatment by CO 2 inhalation. Epididymal, subcutaneous and perirenal white adipose tissue (WAT) as well as the liver were rapidly dissected out, weighed, frozen in liquid nitrogen, and stored at − 80 °C until mRNA and protein extraction. A piece of the tissues was fixed in 4% formaldehyde for immunohistochemical analyses. All experimental procedures conformed to the European Guidelines for the Care and Use of Laboratory Animals (directive 2010/63/EU) and were approved by the Ethical Committee for Animal Experimentation of the University of Navarra (041/08).
Blood and tissue assays. Blood assays were determined as previously described 26  RNA extraction and real-time PCR. RNA isolation and purification was performed as described earlier 19 . Transcript levels for Aqp3 (NM_016689.  Table S1) were designed using the software Primer Express 2.0 (Applied Biosystems) and purchased from Genosys (Sigma). Primers or TaqMan ® probes encompassing fragments of the areas from the extremes of two exons were designed to ensure the detection of the corresponding transcript avoiding genomic DNA amplification. The cDNA was amplified at the following conditions: 95 °C for 10 min, followed by 45 cycles of 15 s at 95 °C and 1 min at 59 °C, using the TaqMan ® Universal PCR Master Mix (Applied Biosystems). The primer and probe concentrations were 300 and 200 nmol/L, respectively. All results were normalized for the expression of 18 S rRNA (Applied Biosystems), and relative quantification was calculated using the Δ Δ Ct formula 19 . Relative mRNA expression was expressed as fold expression over the calibrator sample. All samples were run in triplicate and the average values were calculated.
Immunohistochemistry. The immunodetection of AQP3, AQP7 and AQP9 in histological sections of subcutaneous adipose tissue and liver was performed by the indirect immunoperoxidase method 12 .