RETRACTED ARTICLE: The Adipocyte Na/K-ATPase Oxidant Amplification Loop is the Central Regulator of Western Diet-Induced Obesity and Associated Comorbidities

Obesity has become a worldwide epidemic. We have previously reported that systemic administration of pNaKtide which targets the Na/K-ATPase oxidant amplification loop (NKAL) was able to decrease systemic oxidative stress and adiposity in mice fed a high fat and fructose supplemented western diet (WD). As adipocytes are believed to play a central role in the development of obesity and its related comorbidities, we examined whether lentiviral-mediated adipocyte-specific expression of NaKtide, a peptide derived from the N domain of the alpha1 Na/K-ATPase subunit, could ameliorate the effects of the WD. C57BL6 mice were fed a WD, which activated Na/K-ATPase signaling in the adipocytes and induced an obese phenotype and caused an increase in plasma levels of leptin, IL-6 and TNFα. WD also decreased locomotor activity, expression of the D2 receptor and tyrosine hydroxylase in brain tissue, while markers of neurodegeneration and neuronal apoptosis were increased following the WD. Selective adipocyte expression of NaKtide in these mice fed a WD attenuated all of these changes including the brain biochemical alterations and behavioral adaptations. These data suggest that adipocyte derived cytokines play an essential role in the development of obesity induced by a WD and that targeting the adipocyte NKAL loop may serve as an effective therapeutic strategy.

Obesity is a major international health challenge as it is often associated with other comorbidities including (but not limited to) metabolic syndrome, cardiovascular diseases, chronic kidney disease, non-alcoholic steatohepatitis (NASH), and neurodegenerative disorders [1][2][3][4] . It is estimated that by the year 2030, 38% of the world's adult population will be overweight and 20% will be obese 5 . Although traditionally the adipocyte had been thought to play a passive role in the process of obesity, recent studies have established adipocyte malfunction as playing a causative role in the pathogenesis of obesity, associated systemic oxidant stress and the comorbidities of obesity including insulin resistance, accelerated cardiovascular disease and NASH [2][3][4] . Evidence suggests that deranged mitochondrial function and chronic inflammation in the adipocytes further contribute towards the pathogenesis of obesity 6,7 .
Our group has demonstrated that the Na/K-ATPase functions as a scaffolding protein, affecting a signal cascade that also serves to amplify oxidants in a feed forward manner [8][9][10][11][12][13][14][15][16][17][18] . Activated Na/K-ATPase signaling has been demonstrated in adipose, hepatic, renal, and cardiac tissues, thus implying a role for Na/K-ATPase signaling in various disease models [2][3][4]19 . Specifically, we have demonstrated that reactive oxygen species (ROS) can both activate this Na/K-ATPase signal cascade as well as generate additional ROS through downstream consequences opeN There are amendments to this paper R E T R A C T E D A R T I C L E www.nature.com/scientificreports www.nature.com/scientificreports/ of this pathway 20,21 . NaKtide, a specific peptide antagonist of Src kinase, was derived from the α1 subunit of the NaK-ATPase, consisting of Ser 415 to Gln 434. NaKtide or a cell permeant derivative, pNaKtide which was created by merging NaKtide with a 13 amino acid TAT leading sequence, prevents the activation of Src which is normally regulated by the α1 subunit of the Na/K-ATPase 19,22 . Since we have demonstrated that systemic administration of pNaKtide was able to reduce oxidative stress, obesity and atherosclerosis induced by a western diet composed of high fat and supplemental fructose 2,3,19 , we were interested in using this molecular strategy to examine the specific tissues involved in the pathogenesis of obesity. Lentiviral vectors have evolved over the last decade as a promising approach to target a gene of interest using a promoter for a specific cell type 23 . Utilization of the lentiviral vector has been safe and highly effective in experimental models [23][24][25][26][27] . We therefore chose this strategy to test the hypothesis that adipocyte oxidant stress caused by activation of adipocyte Na/K-ATPase signaling might play a central role in the development and maintenance of obesity and its associated comorbidities.

Results
Effect of lenti-adiponectin-GFP-NaKtide on adiposity, metabolic balance, and locomotion in C57BL6 mice fed a WD. The effectiveness and specificity of the NaKtide lentiviral-construct were evaluated by immunofluorescence studies in C57BL6 mice fed a WD. Since adiponectin is expressed specifically in adipocytes, the lentiviral construct with NaKtide driven by an adiponectin promoter was used to achieve NaKtide expression specifically in adipocytes (Fig. S1). Results showed that NaKtide expression was exclusively present in the adipose tissue and not visibly expressed in liver and brain tissues (Fig. S2). Further, our results showed that mice fed a WD exhibited a significant increase in body weight over a period of 12 weeks compared to the mice on normal chow (Fig. 1A). The increase in the body weight was significantly attenuated by the transduction of lenti-adiponectin-GFP-NaKtide (Fig. 1A). Our results also demonstrated a significant increase in the visceral and subcutaneous fat, liver weight, and heart weight in the mice fed a WD, which was markedly reduced with lenti-adiponectin-GFP-NaKtide treatment (Table 1). Food intake and energy intake did not differ amongst the different experimental groups (Fig. 1B,C). Energy expenditure was determined as heat production rate in units of kcal/kg/day 3 . The results showed significant reduction in the energy expenditure of mice fed a WD compared to control group; this energy expenditure was significantly increased with lenti-adiponectin-GFP-NaKtide treatment (Fig. 1D). Further, our results showed that oxygen consumption in mice fed a WD alone (2491 ± 149 mL/ kg/hr) was significantly (p < 0.01) decreased as compared to control (3104 ± 88 mL/kg/hr). Treatment with lenti-adiponectin-GFP-NaKtide (3146 ± 232 mL/kg/hr) significantly (p < 0.01) improved the oxygen consumption as compared to WD alone. Since obesity has been associated with decreased locomotion 3,28 , we further looked at the effect of lenti-adiponectin-GFP-NaKtide on movement. Locomotor activity was measured as previously described 3 . Mice fed a WD showed decreased locomotion, and this was normalized by treatment with the lenti-adiponectin-GFP-NaKtide (Fig. 1E).

R E T R
www.nature.com/scientificreports www.nature.com/scientificreports/ Effect of lenti-adiponectin-GFP-NaKtide on adipocyte phenotype and systemic inflammatory profile in C57BL6 mice fed a WD. The WD altered the morphological phenotype of visceral adipocytes with dramatic increases in the size and fat content of these cells. Mice treated with the lenti-adiponectin-GFP-NaKtide showed an improved adipocyte phenotype as evidenced by significantly increased numbers of adipocyte, but significant decreases in adipocyte cell area, when compared to mice fed a WD alone ( Fig. 2A-C). Mice injected with lenti-adiponectin-GFP-NaKtide and fed the normal chow diet showed no significant differences in any of the aforementioned measurements from the control mice. There was also no significant difference noted between mice injected with the empty vector control, the lenti-adiponectin-GFP lacking the NaKtide sequence that were fed a WD when compared to mice fed a WD alone. We next evaluated the markers directly tied to the altered adipocyte phenotype that plays a causal role in the aggravation of systemic oxidative stress. Our Western blot analysis showed the upregulated expression of lipogenic marker, fatty acid synthase (FAS) and adipogenic markers, peroxisome proliferator-activated receptor gamma (PPARγ) and mesoderm specific transcript gene (MEST) in WD fed mice. This upregulation was significantly attenuated in lenti-adiponectin-GFP-NaKtide transduced mice (Fig. 2D-F). Since adipose mitochondria participate in energy expenditure, we measured genes of mitochondrial biogenesis in adipose tissue. Our results showed a significant down regulated expression of peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), a major regulator of mitochondrial biogenesis, in mice fed a WD, as compared to control. This down regulation was prevented by lenti-adiponectin-GFP-NaKtide treatment (Fig. 2G). Similarly, the expression of genes related to mitochondrial biogenesis and the "browning" phenomenon of adipose tissue, including expression of mitofusin (MFN) 1 and 2, and sirtuin 1 (Sirt1) were significantly down regulated in mice fed a WD. Importantly, the expression of the aforementioned genes were all significantly increased upon treatment with lenti-adiponectin-GFP-NaKtide (Table 1). Further, our results showed that the expression level of the inflammatory cytokine, tumor necrosis factor alpha (TNFα), was significantly attenuated by treatment with lenti-adiponectin-GFP-NaKtide (Table 1). Leptin, a hormone secreted by adipocytes, is a potent inducer of ROS generation by promoting inflammation and oxidative stress 29,30 . Mice fed a WD exhibited upregulated levels of leptin, as compared to control. This was significantly attenuated by lenti-adiponectin-GFP-NaKtide treatment ( Table 1).
The insulin resistance in mice fed a WD was reversed by treatment with lenti-adiponectin-GFP-NaKtide (Fig. 3A). Inflammatory cytokines are both indicative of high levels of oxidative stress as well as contribute to the production of oxidants 2 . Mice fed a WD showed significantly upregulated plasma levels of the inflammatory cytokines TNFα, monocyte chemoattractant protein-1 (MCP1), and interleukin-6 (IL-6) ( Fig. 3B-D). Treatment with the lenti-adiponectin-GFP-NaKtide ameliorated the increases in the plasma concentrations of these inflammatory cytokines. Our results showed increases in plasma leptin concentrations in mice fed a WD. This was attenuated by treatment with lenti-adiponectin-GFP-NaKtide (Fig. 3E).

Effect of lenti-adiponectin-GFP-NaKtide on Na/K-ATPase/Src signaling cascade in C57BL6
mice fed a WD. Protein carbonylation is an established method for assessing oxidative stress 31,32 . Our results showed a significant increase in protein carbonylation, measured by 2,4-dinitrophenylhydrazine (DNP), in mice fed a WD which was attenuated by treatment with the lenti-adiponectin-GFP-NaKtide (Fig. 2H). Expression of the α1 subunit of the Na/K-ATPase was significantly down regulated in visceral fat of mice fed a WD. This was prevented by lenti-adiponectin-GFP-NaKtide treatment (Fig. 2I). Conversely, the up-regulated expression of the α2 subunit in mice fed a WD was significantly attenuated by lenti-adiponectin-GFP-NaKtide (Fig. 2J). Treatment with lenti-adiponectin-GFP-NaKtide also prevented Src activation in mice fed a WD (Fig. 2K).

Effect of lenti-adiponectin-GFP-NaKtide on neurodegeneration in C57BL6 mice fed a WD.
Obesity and systemic oxidative stress have been implicated with neurodegenerative disorders, so we next examined the effect of lenti-adiponectin-GFP-NaKtide on these parameters. Tyrosine hydroxylase (TH) is the rate-limiting enzyme in the production of dopamine; the expression of both TH and the type 2 dopamine receptor (D2) correlate with locomotor activity 33 . TH staining showed decreased expression of TH in the prefrontal cortex of the brain in mice fed a WD; western blot analysis of TH showed significant decrease in brains of mice fed WD (Fig. 4A,B).

R E T R
www.nature.com/scientificreports www.nature.com/scientificreports/ Treatment with lenti-adiponectin-GFP-NaKtide negated these changes. Furthermore, expression of D2 receptor in the brain was significantly down regulated in mice fed a WD. This was also improved by the treatment with the lenti-adiponectin-GFP-NaKtide (Fig. 4B). Further, our results demonstrated an increase in protein carbonylation in the brain tissues of mice fed a WD; this was attenuated by the lenti-adiponectin-GFP-NaKtide (Fig. 4C). Our results also showed decreased expression of postsynaptic density protein 95 (PSD95), a marker of synaptic plasticity 34 in the brain tissues of WD mice. Treatment with lenti-adiponectin-GFP-NaKtide improved this PSD95 expression (Fig. 4D). Tau, a marker of tangled neuron tracks and a hallmark of neurodegenerative disease 34 , was up regulated in the brains of mice fed WD and markedly reduced by the lenti-adiponectin-GFP-NaKtide (Fig. 4E). The TUNEL assay demonstrated that apoptosis was increased in the brain tissue of mice fed WD. Apoptosis was also attenuated by the lenti-adiponectin-GFP-NaKtide treatment (4F, G).
Effect of lenti-adiponectin-GFP-NaKtide on hepatic histology, inflammation, and fibrosis in C57BL6 mice fed a WD. Next, we aimed to determine whether NaKtide targeted specifically to adipocytes was able to attenuate the development of nonalcoholic steatohepatitis (NASH). H&E staining of liver sections from C57BL6 mice fed a WD showed inflammation and increased lipid accumulation in the liver as compared to the control group (Fig. S3A). Treatment with lenti-adiponectin-GFP-NaKtide exhibited decreased lipid and inflammatory cell infiltration. We note that the viral transduction did not lead to demonstrable hepatic NaKtide expression (Fig. S1) implying an indirect effect. Mice fed a WD had significantly increased lipid accumulation in the liver compared to mice fed a normal chow diet as demonstrated by Oil Red O staining. Administration of lenti-adiponectin-GFP-NaKtide decreased lipid accumulation in mice fed a WD (Fig. S3B). CD36, a fatty acid transport protein, contributes to the progression of NASH 3 . Our results showed that CD36 mRNA expression was decreased by lenti-adiponectin-GFP-NaKtide treatment as compared to the mice fed a WD (Fig. S3C). Similarly, TNFα and F4/80 mRNA expression, markers of inflammation and macrophage/kupffer cells infiltration, were also increased in mice fed a WD as compared to control mice (Fig. S3D,E). These changes were attenuated by the lenti-adiponectin-GFP-NaKtide treatment. Furthermore, mRNA expression of hepatic matrix metalloproteinases (MMP) 2 and 9, genes related to fibrogenesis, were also elevated in mice fed a WD, and these increases were also attenuated with lenti-adiponectin-GFP-NaKtide treatment (Fig. S2F,G). These findings indicate that the lenti-adiponectin-GFP-NaKtide not only improves the metabolic profile in adipocytes, but also has profound effects on the liver through indirect mechanisms.
Correlational analysis of markers associated with reprogramming of adipocyte phenotype, NASH and neurodegeneration in C57BL6 mice fed a WD. First, we examined the degree of correlation between the various measurements performed in our study. These data are summarized in Fig. 5A. Our results showed a number of strong correlations. In particular, it was clear that the plasma levels of inflammatory cytokines correlated with locomotor activity (Fig. 5A). When we performed multiple linear regression analysis, it was clear that a model consisting of plasma TNFα and IL-6 predicted the locomotor activity with excellent accuracy (R 2 = 0.84) (Fig. 5B). We have also included a webpage illustrating heat map and corresponding box plates as Fig. S4.  Effect of lenti-adiponectin-GFP-NaKtide on neurodegeneration in brain tissue in C57BL6 mice fed a WD. Representative and quantitative analysis based on tyrosine hydroxylase staining (A), images taken with 20X objective lens; scale represents 100 µm. Immunoblot analysis for tyrosine hydroxylase (TH) (B). Protein carbonylation levels with Coomassie staining as a loading control (C). Immunoblot analysis of D2 receptor with data shown as mean band density normalized to GAPDH (D). Immunoblot analysis for marker of synaptic plasticity, PSD95 (E), and marker of tangles in neuron tracks, Tau (F). Representative images and quantification of the TUNEL assay in brain tissue (G-H). All gels have been cropped above and below the band, and the full blots have been included in Supplementary Fig. S6

Discussion
In this study, we targeted the Na/K-ATPase signaling antagonist NaKtide specifically to adipocytes in mice fed a WD using a lentivirus employing adiponectin to drive its expression, thus expressing the NaKtide only in adipocytes. In our previous studies, systemic pNaKtide treatment attenuated obesity, as well as aging, NASH, and atherosclerosis [2][3][4] . In this study, our results showed that targeting of NaKtide specifically to adipocytes was still able to reverse WD-induced changes in metabolic profile, oxidative stress, and inflammation. This highlights the importance of the Na/K-ATPase oxidant amplification loop within the adipocyte as well as the importance of adipocyte biology itself. There is accumulating evidence suggesting that adipocyte mitochondria might play an important role in the development of dysfunctional adipocytes 35 . Our results showed that PGC1α levels, the master regulator of mitochondrial biogenesis and adaptive thermogenesis 36 , increased upon adipocyte-targeted NaKtide expression. Furthermore, the expression of genes associated with mitochondrial biogenesis and "browning" of fat, including MFN1, MFN2, and SIRT1, were increased by NaKtide treatment. Adipocytes also showed increased protein carbonylation in mice fed a WD. This was decreased by NaKtide treatment, indicating a reduction in oxidative stress within the adipose tissue.
Our results also showed that expression of the α1 subunit of the Na/K-ATPase was significantly downregulated in visceral fat by the WD. We have shown previously that oxidative stress can induce endocytosis of α1; we therefore propose that WD increases oxidative stress, increasing α1 endocytosis [37][38][39] . Some of the endocytosed α1 is recycled to the surface, while some is destroyed through endosomal and proteolytic pathways. Previous evidence has shown that ROS can accelerate the degradation of the Na/K-ATPase [40][41][42][43][44] . Hence, we propose that increased oxidative stress induced by WD leads to a decrease in total adipocyte α1 expression in mice fed a WD. Interestingly, expression of α2 subunit of Na/K-ATPase was noted to move in the opposite direction as α1; this may be related to a compensatory mechanism that has yet to be elucidated. Treatment with pNaKtide also blocked Na/K-ATPase-regulated Src activation in mice fed a WD.
Studies in mouse models of obesity have shown that increased systemic oxidative stress is due to an altered adipocyte phenotype 45,46 . Comorbidities such as insulin resistance and dyslipidemia also contribute to systemic oxidative stress 47 . Attenuation of the NKAL in the adipocytes during exposure to the WD ameliorated oxidant stress within these cells as well as the WD-induced hypertension, dyslipidemia, and insulin resistance. In particular, the increases in plasma levels of leptin and inflammatory cytokines (i.e., IL-6, TNFα, and MCP1) in mice fed a WD were dramatically attenuated by NaKtide expression in adipocytes. This suggests a central role of the Na/K-ATPase signaling loop within adipocytes in the development of obesity and systemic oxidative stress in response to a WD.
Perhaps of greater interest, we observed profound changes in the locomotor activity of animals with the switch to a WD, which were dramatically attenuated by adipocyte NaKtide treatment. Specifically, animals given the WD demonstrated marked reductions in activity level and calculated energy expenditure. Such changes were also observed in previous reports, even in a mouse strain which did not gain excessive amounts of weight 3 . Previously, we found that systemic pNaKtide prevented these decreases in activity, and we found that NaKtide expression limited to the adipocytes accomplished the same effect. We found that these behavioral changes were strongly associated with changes in brain D2 receptor expression and tyrosine hydroxylase expression, confirmed through both immunofluorescence and western blot. All of these changes appeared to be linked to adipocyte cytokine production. In fact, mouse activity could be predicted with a linear combination of plasma TNFα and IL6 levels

R E T R
www.nature.com/scientificreports www.nature.com/scientificreports/ with an R2 of 84%. In addition to these functional effects, marked changes in brain biochemistry and morphology were also noted with the western diet, and again, adipocyte NaKtide treatment ameliorated these, which included protein carbonylation, expression of PSD95 and pTau as well as neuronal apoptosis. The implications of these observations to cognitive and emotional disturbances associated with obesity and other neurodegenerative disorders including aging related cognitive impairment have yet to be explored. We also note that virtually all consequences of obesity such as dyslipidemia, insulin resistance and NASH were also attenuated by adipocyte NaKtide expression.
Together, our results suggest a central role for the Na/K-ATPase oxidant amplification loop within adipocytes in the pathogenesis of obesity induced by a WD as well as the commonly associated comorbidities. Somewhat surprisingly, behavioral changes could be attributed to this deranged adipocyte metabolism which, in turn, were very strongly associated with the levels of cytokines known to be produced by adipocytes. As we review these data, it appears not only that targeting the Na/K-ATPase oxidant amplification loop may be promising in terms of efficacy as well as limiting off-target effects. If these findings can be confirmed in humans, a number of therapeutic strategies might be envisioned ranging from surgical control of fat mass, molecular biological strategies described in this report to simple dietary manipulations.

Materials and Methods experimental design. All animal studies were approved by the Marshall University Animal Care and Use
Committee in accordance with the National Institutes of Health (NIH) Guide for Care and Use of Laboratory Animals. C57BL6 mice (Male, 8 weeks) were purchased from Hilltop Lab Animals. After arriving to the Robert C. Byrd Biotechnology Science Center Animal Resource Facility (ARF) mice were placed in cages, and fed normal chow diet with ad libitum access to water. Animals were randomly divided into five groups (12-18/group) as follows: 1. Control (normal chow) 2. NaKtide (normal chow + lenti-adipo-GFP-NaKtide) 3. Western Diet (WD) 4. WD + GFP (WD + lenti-adiponectin-GFP) 5. WD + NaKtide (WD + lenti-adiponectin-GFP-NaKtide). Western diet (WD) containing fructose is a well-known inducer of diet-induced obesity 2,48,49 . WD was purchased commercially from Envigo (Indianapolis, IN), and contained 42% fat, 42.7% carbohydrate, and 15.2% protein, yielding 4.5 KJ/g specific energy. Fructose was purchased commercially from Alfa Aesar (Ward Hill, MA). Fructose water was made at a concentration of 42 g/L, yielding 0.168 KJ/mL of specific energy. Mice in the WD group were given WD and ad libitum access to fructose water.
Lentiviral construct with NaKtide driven by an adiponectin promoter were constructed by VectorBuilder Inc. and were used in mice to achieve NaKtide expression specifically in adipose tissue (Fig. S2). Lentiviral construct featured the adiponectin promoter driving expression of the NaKtide cassette linked by means of a 2 A peptide to eGFP for bicistronic expression. The NaKtide sequence used for this vector was ATGAGCGCCACCTGGCTGGCCCTGAGCAGGATCGCCGGTCTTTGCAACAGGGCCGTGTTCCAG. Lentivirus (100 uL, 2 × 10 9 TU/mL in saline) with NaKtide and its counterpart Lenti-GFP (driven by adiponectin promoter) were injected into mice intraperitoneally (IP) at week 0. Groups 2 and 5 received an injection of lenti-adiponectin-GFP-NaKtide, and group 4 was given an injection of lenti-adiponectin-GFP. Body weight was measured weekly for 12 weeks, as well as food and water intake. At the time of sacrifice, the body weight, visceral and subcutaneous fat weight, liver weight, heart weight, kidney weights, and brain weights were measured. Blood samples were collected for determination of leptin and inflammatory cytokine levels. Tissues were flash frozen in liquid nitrogen and maintained at −80 °C, preserved in OCT for sectioning, or placed in paraformaldehyde for paraffin embedding.
Indirect calorimetry and locomotor activity. At the end of the 12-week period, energy expenditure, oxygen consumption, and locomotor activity were measured as described previously 3 . All mice were acclimatized to the cages for 24 hours prior to an additional 48 hours of recordings under the regular 12-hour light/dark cycle.
Glucose tolerance test. At the end of the 12-week period, mice were fasted for 8 hours. After the fasting period, a 10% glucose solution (2 g/kg body weight) was injected intraperitoneally. Samples were taken from the tail vein at 0, 30, 60, 90, and 120 minutes after glucose injection. Blood glucose was measured using the Accutrend Sensor glucometer.
Cytokine and leptin measurements. Plasma TNFα, IL-6, MCP-1, leptin were measured using an ELISA assay kit according to manufacturer instructions (Abcam Cambridge, MA) as reported previously 3 .
RNA extraction and real time pCR. Total RNA was extracted from adipose and hepatic tissue using RNeasy Protect Mini Kit (QIAGEN, Maryland) as described previously 4 . Total RNA was analyzed by a quantitative real time polymerase chain reaction. Real time PCR was done using SYBR Green PCR Master Mix on a 7500 HT Fast Real-Time PCR System (Applied Biosystems, US). Each reaction was done in triplicate. All experimental samples were normalized using GAPDH. Specific primers were used for MFN 1 and 2, Sirt3, TNFα, Leptin, CD36, MMP2 and 9 and F4/80.

Western blot analysis.
Visceral adipose tissue and brain tissue were pulverized with liquid nitrogen and placed in RIPA homogenization buffer. Homogenates were centrifuged, the supernatant was isolated, and immunoblotting performed. Adipose tissue was used for determination of α1/2 subunits of the Na/K-ATPase, FAS, PPARγ, MEST, and PGC1α as previously reported 20,50 . Brain tissue was used for expression of tyrosine hydroxylase (TH), dopamine 2 (D2) receptors, PSD95, p-Tau, and Tau expression. D2 receptor antibody and tyrosine hydroxylase was from Abcam (Cambridge, MA) and PSD95, p-Tau, and Tau from Cell Signaling (Danvers, MA).

R E T R
www.nature.com/scientificreports www.nature.com/scientificreports/ Measurement of c-src phosphorylation. Whole cell lysates from visceral adipose tissue were prepared with RIPA buffer. Activation of c-Src was determined as described previously 20 . Polyclonal anti-Src [pY418] phospho-specific antibody was from Invitrogen (Camarillo, CA). Monoclonal antibody against total c-Src was from Santa Cruz (Santa Cruz, CA). After immunoblotting for phospho-Src, the same membrane was stripped and blotted for total c-Src. Activation of c-Src was expressed as the ratio of phospho-Src/c-Src with measurements normalized to 1 for control samples.
Assessment of protein carbonylation. Whole cell lysates from visceral adipose tissue and brain tissue were prepared with RIPA buffer and western blotting for protein carbonylation assay was done as previously described 20 . 2,4-dinitrophenylhydrazine (DNPH) and antibody against 2,4-dinitrophenyl (DNP) hydrazone derivatives were purchased from Sigma-Aldrich. The signal density values of control samples were normalized to 1 with Coomassie blue staining.
Immunofluorescence studies in adipose, hepatic, and brain tissues. Adipose tissue, hepatic tissue, and brain tissue were frozen in OCT compound, cut into 6 um sections and mounted on slides. The sections were fixed with 4% PFA for 15 minutes, washed once with PBS, probed with 1:100 primary antibody dilution and 1:1000 secondary antibody (Alexa 455 Red), and then mounted with DAPI solution and coverslips. Expression of lenti-adiponectin-EGFP was determined using a Nikon Eclipse 80i microscope equipped with a Nikon camera head DS-Fi1 (Nikon, Japan). Expression of lenti-adiponectin-EGFP-NaKtide was determined using an RFP filter on Nikon Eclipse 80i microscope equipped with a Nikon camera head DS-Fi1 (Nikon, Japan). Protocol was adapted from Alexa-fluor-555 Cell Analysis from Thermo Fisher, and IHC protocols (IHC WORLD Life Science Products).
Hematoxylin and eosin staining. Visceral adipose tissue and hepatic tissue was cut into 6 µm sections and stained with hematoxylin and eosin for histological analysis as previously described 3,4 . Adipocyte number and area was determined using ImageJ software (NIH) 51 . In brief, H&E stained adipose tissue images (20X magnification) were opened in ImageJ. Images were changed to 8 bit, enhanced through Threshold, Dilate, and Watershed functions, converted to mask using the binary function, and enhanced. Black and white images were compared to the original image to ensure accurate representation of adipocytes. The total number and size of adipocytes was calculated using the "analyze particles" command. To calculate the number of adipocytes, the size of the sample was normalized based on the number of pixels present in the image and the magnification. This was consistent and used for all subsequent images. The number of total adipocytes in the distribution is calculated from this for scale.
oil Red o staining. Liver tissues, frozen in OCT compounds, were cut into 6 µm sections and Oil Red O staining was done as described previously 3 . Images were taken on a Nikon Eclipse 80i microscope equipped with a Nikon camera head DS-Fi1 (Nikon, Japan). For quantitative analysis, total area of the red pixels on stained tissue was measured by Image J software (NIH). The data was expressed as percentage of the Oil Red O stained area with suspect to total area. tyrosine hydroxylase staining. Brain tissue preserved in paraffin was cut into 6 µm sections and mounted on slides. Deparaffinization, rehydration, and antigen retrieval were performed on the slides. After drying, slides were stained with anti-Tyrosine Hydroxylase (Abcam, US) overnight, following modified manufacturer protocol 52 . Imaging was done with a Nikon camera head DS-Fi1 (Nikon, Japan). For quantitative analysis, total area of the green pixels on stained tissue was measured by ImageJ software (NIH). The data was expressed as percentage of the green stained area with respect to total area. TUNEL assay for DNA damage detection. DNA double-strand breaks were detected in frozen mouse brain tissue using the Click-iT ® Plus TUNEL Assay (Thermo Fisher Scientific Inc., US) according to the manufacturer's protocol. Specimens were mounted using VECTASHIELD ® mounting medium and counterstained with 50 mg/ml DAPI. Images were collected using a digital inverted microscope that spanned the entire area of the tissues on the cover slip. The percentage of TUNEL-positive cells was calculated based on the number of positively stained cells divided by the total number of cells multiplied by 100 as reported previously 4 . statistical analyses. The plots used were combination of scatter and boxplots. The boxplot displays the distribution of the data. It visualizes five summary statistics (the median, two hinges and two whiskers), and all "outlying" points individually 53 . The boxplot was defined by: Notch lower = median − 1.58 * interquartile range (IQR)/sqrt (n) 54 Middle = median Notch upper = median + 1.58 * IQR/sqrt (n); Lower hinge is the 25 th percentile Upper hinge is the 75 th percentile Data were analyzed and presented with the statistical program R using the packages ggplot2 and doBy. Multiple groups of parametric data (passed Shapiro-Wilk normality test) were compared with one way analysis of variance (ANOVA) and individual group means compared with an unpaired t-test employing the Holm correction for multiple comparisons. Nonparametric data were analyzed with the Wilcoxon rank sum test, also employing the Holm correction for multiple comparisons. All data comparisons are presented at the NS, p < 0.05 and p < 0.01 levels. Correlation analysis was performed using the R package Psycho. The fit of a plane based on TNFα and IL6 to locomotion was performed using linear regression and displayed with the R package plot3D.