Loss of TDAG51 Results in Mature-Onset Obesity, Hepatic Steatosis, and Insulin Resistance by Regulating Lipogenesis

Cells and tissues were homogenized in 4× SDS lysis buffer, and protein lysates were subjected to immunoblotting as previously described (23). Antibodies were used to detect the following proteins: TDAG51, SREBP-1 (H160 for detection of precursor p125 and K10 for detection of mature p68), C/EBPα, PPARα (Santa Cruz Biotechnology), Akt, pS473-Akt, Scd-1, fatty acid synthase (FAS), phosphoserine insulin receptor substrate 1 (phosphoserine IRS1), aP2, PPARγ, perilipin, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Cell Signaling), IRS1, phosphotyrosine IRS1 (Invitrogen), and β-actin (Sigma).

Mouse 3T3-L1 cells were cultured, differentiated, and stained with Oil Red O as previously described (23). Pooled protein extracts were obtained from isolated human preadipocytes and their corresponding differentiated adipocytes from four lean female subjects (Zenbio).

Total RNA was isolated using the RNeasy kit (Qiagen) according to the manufacturer’s protocol. Using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems), total RNA was reverse transcribed to obtain cDNA. Quantitative real time-PCR (qRT-PCR) was performed using Fast SYBR Green PCR Master Mix (Applied Biosystems) in the AB7900 HT Fast Real-Time PCR System. For a list of primer sequences see Supplementary Table 1. Data analysis was performed using the ΔΔC(T) method and normalized to 18S, unless otherwise indicated.

Primary preadipocytes were cultured by isolating the stromal-vascular fraction from mouse WAT. In brief, WAT was minced, collagenase (Sigma) digested, and centrifuged to separate the stromal-vascular fraction from the mature adipocyte fraction. The pelleted cells were then plated and induced to differentiate in Dulbecco’s modified Eagle’s medium/F12 containing 2% FBS, 1 μg/mL insulin, 250 nmol/L dexamethasone, 0.5 mmol/L isobutylmethylxanthine, 60 μmol/L indomethacin, and 2.5 μmol/L rosiglitazone. Cells were stained with Oil Red O as previously described (23), or protein lysates were collected for Western blotting. Primary hepatocytes isolated from the liver were plated on collagen-coated six-well plates (BD) and allowed to attach for 2–3 h in Williams’ medium E (Invitrogen), after which medium was replaced with Williams’ medium E containing 10% FBS. Isolated hepatocytes were incubated overnight at 37°C and the experiments were performed the following day.

TDAG51^−/− mice were generated previously (17) and back-crossed onto the C57BL/6 background for at least nine generations and genotyped as described in the Supplementary Data. C57BL/6 or TDAG51^−/− mice were maintained under 12-h light-dark cycles with unlimited access to standard rodent chow and water. At 8 weeks of age, male WT or TDAG51^−/− mice (n = 8) were either continued on the regular chow diet or placed on a high-fat diet (42% of kcal from fat, TD.88137; Harlan Teklad). At 28 weeks of age, mice were fasted for 14 h before they were killed by cervical dislocation under anesthesia. Blood was drawn from the heart, and tissues were immediately collected for histology or biochemical assessment. Only male mice were used in this study. Unless stated otherwise, only chow diet–fed mice were examined. All protocols for animal use and euthanasia were approved by the McMaster University Animal Research Ethics Board.

WT and TDAG51^−/− mice were fasted overnight and given an intravenous injection of 2-(¹⁸F) fluoro-2-deoxyglucose (2-FDG). The mice were kept under heat lamps for 30 min, anesthetized using isoflurane, and scanned on a small animal positron emission tomography (PET) scanner for visualization of glucose uptake. The mice were placed in a single-photon emission computed tomography (SPECT)/CT scanner for determination of body fat content (24).

Plasma adiponectin (R&D Systems), leptin (Linco), and insulin (Chemicon) concentrations were measured using mouse ELISA kits. Plasma glucose, triglycerides (TGs), cholesterol esters, and free fatty acids (FFAs) were measured using colorimetric assays from Wako Diagnostics. Plasma thiobarbituric acid reactive substances (TBARS) assay was carried out using an OXI-TEK TBARS Kit (Alexis Biochemicals), and plasma alanine aminotransferase (ALT) levels were analyzed using a kit (ThermoDMA).

Plasma cytokines were measured using a bead cytometric array (BD) according to the manufacturer’s protocol.

For glucose tolerance tests (GTTs), mice were fasted overnight and given an intraperitoneal injection of glucose (1.5 mg/g body weight). For insulin tolerance tests (ITTs), mice were fasted for 6 h and intraperitoneally injected with 0.75 units/kg of insulin. Blood was collected from the tail vein, and glucose concentrations were measured using a hand-held glucometer. For in vivo insulin signaling experiments, tissues were collected 15 min after an intraperitoneal injection of insulin (0.75 units/kg) in fasted mice as previously described (25). Metabolic studies using indirect calorimetry were performed using the Columbus Laboratory Animal Monitoring System (CLAMS) as described previously (26). Core body temperature was determined in the CLAMS after subcutaneous implantation, under isoflurane anesthesia, of minimiter probes (Respironics, Inc.) in the dorsal subscapular region.

Interscapular brown adipose tissue (BAT) was excised, cut into 2-mm³ pieces, and immersed in 2% glutaraldehyde in cacodylate buffer. Tissues were processed using standard protocols and embedded in Spurr resin. Ultrathin sections were stained with uranyl acetate and lead citrate and observed under a JEOL 200CX TEMSCAN microscope.

Approximately equivalent weights of frozen liver tissue from mice were homogenized in PBS, and TGs were extracted using the Folch extraction method as previously described (27). Hepatic VLDL-TG secretion was measured as previously described (28). In brief, fasted 8-week-old mice were injected intraperitoneally with 1 g/kg poloxamer-407 dissolved in saline, and blood samples were collected for serum TG measurements at various time points.

For determination of de novo lipogenesis, hepatocytes were incubated with media containing sodium acetate (5 mmol/L) and ³H-acetate (1 µCi/mL; Perkin Elmer) for 1 or 4 h in the presence or absence of 10 nmol/L insulin. Cells were washed twice with PBS, and lipids were extracted by the method of Bligh and Dyer (29). Neutral lipids were quantified as previously described (30). Lipid bands were visualized and scraped into counting vials for measuring radioactivity. Palmitate oxidation was determined as described previously (31). In brief, sodium palmitate (0.5 mmol/L) and ¹⁴C-palmitate (0.5 µCi/mL; Perkin Elmer) were allowed to conjugate to 2% BSA for 2–3 h and then incubated with primary hepatocytes for 4 h, in the presence or absence of 10 nmol/L insulin. Radiolabeled CO₂ was liberated by adding an equal volume of 1 mol/L acetic acid and trapped by the addition of benzethonium hydroxide (1 mol/L) over 90 min. The tube containing trapped ¹⁴CO₂ was placed in a scintillation vial and counted.

Data are expressed as the mean ± SEM unless otherwise indicated. Statistical analyses were performed using unpaired, two-tailed Student t test, and significance was defined as P < 0.05.

TDAG51 protein expression was examined in various tissues from adult C57BL/6 mice. In addition to the liver and lung, which have abundant TDAG51 levels at the mRNA and protein level (15,17), WAT also expressed TDAG51 protein (Fig. 1A). Previous microarray analysis during adipogenesis indicated that TDAG51 mRNA expression is modulated in early adipogenesis (11). Therefore, we carried out qRT-PCR analysis, which confirmed that TDAG51 mRNA was significantly induced in the first 2 h of adipogenesis followed by a sharp decline back to baseline expression (Fig. 1B). We next sought to evaluate the expression profile of TDAG51 at the protein level during adipocyte differentiation. TDAG51 protein expression was induced in the first 4 days poststimulation, followed by a gradual decline after day 4, and diminished expression by day 14 of differentiation, when the vast majority of cells had accumulated large lipid droplets, as shown by Oil Red O staining (Fig. 1C, top). In addition, TDAG51 expression showed an inverse correlation with PPARγ expression during adipogenesis (Fig. 1C).

In support of the findings in 3T3-L1 cells, an inverse correlation between TDAG51 and PPARγ expression and/or adipogenesis was also observed in mouse embryonic fibroblasts induced to differentiate into adipocytes (Fig. 1D). We next examined protein extracts from pooled human preadipocytes and their corresponding differentiated mature adipocytes, which also showed disappearance of TDAG51 in mature adipocytes (Fig. 1E). These findings indicate that TDAG51 is downregulated in differentiated lipid-laden adipocytes, suggesting that TDAG51 may play an important role in adipogenesis.

To reconcile the temporal difference in the expression of TDAG51 mRNA and protein during adipogenesis, 3T3-L1 cells were treated with cycloheximide to block protein translation. As shown in Fig. 1F, relative to other proteins that are crucial for adipogenesis, such as PPARγ and C/EBPβ, TDAG51 protein is considerably more stable.

In order to examine the role of TDAG51 in adipogenesis, primary stromal-vascular cells (SVCs) were isolated from WT and TDAG51^−/− WAT and induced to differentiate. TDAG51^−/− SVCs exhibited earlier and/or greater expression of adipogenic markers such as PPARγ2, aP2, FAS, perilipin, and C/EBPα on day 2 (Fig. 2A). By day 5, most of these adipogenic markers reached similar levels in WT and TDAG51^−/− SVCs. However, despite the plateau in expression of these adipogenic markers, TDAG51^−/− SVCs exhibited increased lipid accumulation, as determined by Oil Red O staining (Fig. 2B), and enhanced adiponectin secretion, a marker of mature adipocytes (Fig. 2C).

To complement these findings, both silencing/short interfering (si)RNA and short hairpin (sh)RNA targeting mouse TDAG51 were used to knockdown TDAG51 expression in 3T3-L1 cells. Similar to TDAG51^−/− SVCs, siTDAG51-transfected cells had significantly greater lipid accumulation compared with siControl-transfected cells (Supplementary Fig. 1A and B). An shTDAG51 construct was used, leading to a stable 30% knockdown in TDAG51 mRNA expression (Supplementary Fig. 1C). Even this modest decrease in mRNA expression led to a striking decrease in TDAG51 protein (Supplementary Fig. 1D) and resulted in significantly greater adiponectin secretion from shTDAG51 cells on day 3 (Supplementary Fig. 1E). Expression of C/EBPδ and C/EBPβ mRNA was higher within the first 8 h of differentiation but dropped off more quickly, and similarly mRNA expression of preadipocyte factor-1 (Pref-1) declined more quickly in shTDAG51 cells on day 1 as compared with shControl cells (Supplementary Fig. 1F). Analysis of late markers of adipogenesis, such as expression of PPARγ, aP2, and C/EBPα mRNA, indicated that they were expressed earlier (day 2) and to a greater extent in shTDAG51 cells (Supplementary Fig. 1G).

To study the effect of loss of TDAG51 on weight regulation in vivo, 8-week-old chow diet–fed male TDAG51^−/− mice were monitored over 20 weeks and compared with age- and sex-matched heterozygous TDAG51^+/− and WT mice. Body weight was not significantly different between genotypes at 8 weeks of age (Fig. 3A and Supplementary Fig. 2). However, over time, in comparison with WT mice, TDAG51^−/− mice exhibited a significant increase in body weight (Fig. 3A), which was not observed for TDAG51^+/− mice (Supplementary Fig. 2). There was no significant difference in tissue weights (Supplementary Fig. 3A and B) except for epididymal WAT (eWAT), where a significant increase was observed in TDAG51^−/− mice (Fig. 3B). CT scans in 28-week-old mice showed that TDAG51^−/− mice have a fivefold increase in total body WAT (Fig. 3C), with fat deposition in both the visceral and subcutaneous depots (Fig. 3C, top). A significant increase in adiposity was also observed in the bone marrow of TDAG51^−/− mice (Supplementary Fig. 4A and B). Histological examination of the eWAT revealed that TDAG51^−/− adipocytes were significantly enlarged compared with WT controls (Fig. 3D). In support of enhanced adipogenic and lipogenic potential observed in cultured preadipocytes (Fig. 2A–C), mRNA expression of several adipogenic genes (aP2, PPARγ, and C/EBPβ) was significantly increased in WAT from TDAG51^−/− mice (Fig. 3E). These findings demonstrate that absence of TDAG51 enhances adipocyte lipid accumulation, leading to WAT expansion and increased body mass in TDAG51^−/− mice.

Consistent with increased adiposity, TDAG51^−/− mice had increased levels of systemic oxidative stress (TBARS) (Table 1), reaching levels similar to age-matched diet-induced obese mice (Supplementary Table 2), as well as increased plasma leptin and reduced adiponectin concentrations, compared with WT mice (Table 1). Plasma inflammatory markers (tumor necrosis factor-α [TNF-α], monocyte chemoattractant protein-1 [MCP-1], and IL12p70) tended to be higher in TDAG51^−/− mice, but did not reach statistical significance (Supplementary Fig. 5A). Furthermore, staining of eWAT sections to detect macrophages revealed an increase in the number of crown-like structures in comparison with WT mice (Supplementary Fig. 5B). These findings are consistent with the notion that increased adiposity leads to adipose macrophage infiltration.

We next examined metabolic parameters using indirect calorimetry. Food consumption in WT and TDAG51^−/− mice was not significantly different (Fig. 4A), and although there was a trend toward lower locomotor activity in TDAG51^−/− mice, the difference did not reach statistical significance (Fig. 4B,). However, TDAG51^−/− mice exhibited a decrease in metabolic rate as indicated by a significant reduction in oxygen consumption (Vo₂) during both day and night hours (Fig. 4C). Core body temperature was also lower in TDAG51^−/− mice during the night, which corresponded with their lower Vo₂ (Fig. 4D).

We explored whether changes in BAT, an important regulator of thermogenesis in mice, could explain the difference in energy expenditure. Interscapular BAT from WT and TDAG51^−/− mice was examined to assess cell morphology, and we found an overall increase in lipid droplet size in BAT from TDAG51^−/− mice (Fig. 4E). We next examined expression of thermogenic and mitochondrial genes in BAT from WT and TDAG51^−/− mice. No significant differences were found in the expression of UCP1, CIDEA, PGC1α, CPT1b, or PPARγ, and we did not detect any differences in the expression of components of the electron transport chain (Supplementary Fig. 6A). Furthermore, no overall difference in UCP-1 immunostaining of BAT paraffin sections was observed between genotypes (Supplementary Fig. 6B). To assess mitochondrial morphology and size, BAT was further examined using transmission electron microscopy; however, no striking differences in mitochondria shape, size, or cristae were observed between genotypes (Supplementary Fig. 6C). Given that TDAG51 expression is ∼90% lower in BAT compared with WAT of WT mice (Supplementary Fig. 6D), it is possible that BAT function in TDAG51^−/− mice would not be directly affected by the absence of TDAG51. These findings suggest an increase in ectopic lipid deposition in BAT; however, BAT function does not appear to be significantly affected.

Fasting plasma glucose and insulin levels were significantly elevated in TDAG51^−/− mice at 28 weeks of age (Fig. 5A and B). In young 8-week-old WT and TDAG51^−/− mice, an age where whole-body, liver, and WAT weights were indistinguishable (Fig. 3A and Supplementary Fig. 7A), GTTs and ITTs revealed no differences between genotypes (Supplementary Fig. 7B and C). However, at 28 weeks of age, when TDAG51^−/− mice exhibited a significant increase in WAT and body weight, TDAG51^−/− mice were also significantly more glucose intolerant (Fig. 5C) and insulin resistant (Fig. 5D) compared with WT mice. The efficiency of glucose uptake in various tissues was examined using PET scanning, which revealed a significant decrease in glucose uptake in skeletal muscles and WAT of TDAG51^−/− mice (Fig. 5E). WAT from TDAG51^−/− mice also showed decreased insulin receptor signaling, as indicated by reduced tyrosine (Y608) phosphorylation of IRS-1 and increased phosphorylation on serine (S612), which inhibits insulin receptor signaling (Fig. 5F). However, at least within this time interval, insulin-stimulated Akt phosphorylation and SREBP-1 activation were similar in WAT from WT and TDAG51^−/− mice. Consistent with the presence of increased adiposity, decreased glucose disposal, and insulin resistance (IR), plasma FFAs were also significantly elevated in TDAG51^−/− mice (Table 1).

To investigate the cause of hyperinsulinemia and IR in TDAG51^−/− mice, pancreatic islets were histologically assessed using insulin and glucagon staining. No discernible differences in islet size or shape were observed between genotypes (Fig. 5G). Overall, these findings demonstrate that TDAG51^−/− mice display an age-related impairment in insulin sensitivity that is likely related to increased adiposity in these mice.

Given the high levels of TDAG51 expression in the liver, we sought to determine if TDAG51 expression was altered in fatty livers by examining three mouse models of nonalcoholic fatty liver disease. Dietary, genetic, and chemically induced steatosis were all highly correlated with reduced TDAG51 protein expression (Fig. 6A). Interestingly, absence of TDAG51 increased hepatic lipid accumulation and steatosis in 28-week-old mice on chow diet, as determined by histological staining and biochemical measurement of liver TG (Fig. 6B and C), a phenotype that was exacerbated in TDAG51^−/− mice fed a high-fat diet for 20 weeks (Fig. 6B). Given that significant metabolic differences existed between genotypes on a standard chow diet, only chow diet–fed mice were further examined. To rule out the possibility that loss of TDAG51 resulted in hepatic TG accumulation due to impaired hepatic VLDL-TG secretion, serum TG concentrations were measured over the course of 4 h after injection of poloxamer-407, an inhibitor of lipoprotein lipase. There were no significant differences in the rate of hepatic TG secretion (Fig. 6D).

TDAG51^−/− livers showed increased staining for proteins covalently adducted with 4-hydroxynonenal, a marker of lipid peroxidation (Fig. 6E), and an increased number of granulomas indicative of inflammation (Fig. 6F), both of which are consistent with greater steatosis. However, there were no signs of hepatocellular injury as measured by plasma ALT (Table 1), and hepatic cholesterol content was not affected (data not shown). In agreement with increased hepatic TG in TDAG51^−/− livers, there was a marked increase in insulin-stimulated SREBP-1 activation, a master regulator of hepatic lipogenesis (32–34), as indicated by reduced expression of the precursor form of SREBP-1 (p125) and enhanced expression of the mature form (p68) (Fig. 6G), which was not affected under basal fasted conditions (Supplementary Fig. 8A). Insulin-injected TDAG51^−/− livers also had reduced serine 473 phosphorylation of Akt, suggestive of hepatic IR (Fig. 6G). Expression of lipogenic enzymes and SREBP-1 target genes, FAS and stearoyl-CoA desaturase-1 (Scd-1), were induced in livers from TDAG51^−/− mice under basal fasting conditions (Fig. 6H). Furthermore, preliminary microarray gene expression analysis of liver from chow diet–fed WT and TDAG51^−/− mice (K.N.M., R.C.A., unpublished data) revealed an increase in >10 SREBP target genes in TDAG51^−/− livers (Supplementary Table 3).

To determine whether these results were secondary to the insulin-resistant phenotype or intrinsic to the loss of TDAG51, primary hepatocytes from 6-week-old WT and TDAG51^−/− mice, which displayed no differences in body weight or metabolic phenotype, were examined. TDAG51^−/− hepatocytes responded normally to insulin (Supplementary Fig. 8B) but had enhanced basal and insulin-stimulated lipogenesis (rate of newly synthesized fatty acid incorporation into TG) (Fig. 7A). Palmitate oxidation, on the other hand, was not different between genotypes when examined basally or upon insulin stimulation in primary hepatocytes (Fig. 7B). Furthermore, levels of PPARα, a transcription factor involved in fatty acid catabolism, was not affected by the absence of TDAG51 in liver tissue (Supplementary Fig. 8A). Consistent with enhanced de novo lipogenesis, primary hepatocytes from TDAG51^−/− mice expressed higher mRNA levels of the SREBP-1c target genes glucokinase (GK) and acetyl CoA carboxylase 1 (ACC1) (Fig. 7C). These results in cultured hepatocytes are in full agreement with our findings in livers from TDAG51^−/− mice, suggesting that deletion of TDAG51 promotes lipogenesis at least in part through increased SREBP-1 activation and its downstream target gene expression.

Dysregulation of the pathways involved in adipogenesis, lipid metabolism, and energy homeostasis can lead to the development of obesity, hepatic steatosis, diabetes, and dyslipidemia. Here, we present novel evidence that TDAG51 plays a significant role in the regulation of energy and lipid metabolism. First, the findings of this study show for the first time that TDAG51 is expressed in WAT and that its protein expression is modulated during adipogenesis. Given that TDAG51 protein expression is minimal or absent in mature mouse and human adipocytes, TDAG51 may be a useful preadipocyte marker. Second, deletion of TDAG51 enhances lipid accumulation during adipogenesis and accelerates fatty acid/TG synthesis in primary hepatocytes. Third, the absence of TDAG51 leads to an age-associated increase in whole-body adiposity and hepatic TG accumulation, and decreases metabolic rate and insulin sensitivity in standard chow diet–fed WT mice. Notably, the phenotype observed in TDAG51^−/− mice resembles several aspects of high-fat diet–fed WT mice. Finally, TDAG51 protein expression is inversely correlated with hepatic steatosis.

The age-dependent increase in adiposity, hepatic TG accumulation, and systemic IR in TDAG51^−/− mice are suggestive of a chronic and cumulative effect, likely attributable to a decrease in metabolic rate and increased lipogenesis. Excessive accumulation of TG can culminate in the expansion of WAT, ectopic lipid accumulation, and eventually IR in combination with chronic inflammation and oxidative stress (35,36), which is consistent with the phenotype observed in older TDAG51^−/− mice. Moreover, when cells become insulin resistant, not all pathways are blunted as lipogenic pathways remain active, a concept known as selective IR (37). The reduced Akt phosphorylation in combination with increased SREBP-1 activation seen in livers from insulin-injected TDAG51^−/− mice is consistent with selective IR. The absence of TDAG51 in mice culminated in increased TG deposition in the liver. Numerous studies show a strong correlation between fatty liver and whole-body IR (38–42), suggesting that increased hepatic steatosis and selective IR in the liver of TDAG51^−/− mice may be contributing to the systemic IR phenotype in these mice. Given that SREBP-1 is a master regulator of hepatic de novo lipogenesis (32–34) and an increase in SREBP-1 target gene expression was observed in TDAG51^−/− livers, the absence of TDAG51 likely leads to hepatic steatosis through increased SREBP-1 activity. Consistent with this result, primary TDAG51^−/− hepatocytes exhibited a significant enhancement in the rate of lipogenesis, which correlates with the strong association between TDAG51 reduction and fatty livers. Taken together, these data suggest a role for TDAG51 in the negative regulation of SREBP-mediated lipogenesis and prevention of fatty liver.

TDAG51 has several important domains, including a PH-like domain, which appears to be a likely candidate for functional significance in the context of cell signaling and metabolism. For instance, a recent study identified a novel PH domain–containing protein involved in the regulation of adipocyte insulin signaling (43). Indeed, TDAG51 was identified as a specific IGF-1 target gene in NIH-3T3 fibroblasts and was shown to mediate the effects of IGF-1 on cell survival (44). Furthermore, previous studies suggest that extracellular signal–related kinase (ERK) and p38 mitogen-activated protein kinase (p38 MAPK) signaling can modulate TDAG51 expression (44,45). However, the exact cellular mechanisms by which IGF-1, ERK, or p38 MAPK induces TDAG51 expression are unknown.

PH domains play an important role in recruiting target proteins to cell membranes through their binding interactions with phospholipids such as phosphatidylinositol (3-5)-triphosphate (PIP3) and phosphatidylinositol (4,5)-bisphosphate (PIP2) (46,47). Interestingly, TDAG51 has a high stringency PIP3-binding site in its PH-like domain, as predicted by ScanSite (48). Furthermore, the PH-like domain of TDAG51 exhibits a high degree of homology to that of phosphatidylinositol 3-kinase (PI3K) enhancer (PIKE), as predicted by the Protein Homology/Analog Y Recognition Engine (Phyre2) (49). Interestingly, PIKE has recently been implicated in playing a role in adipogenesis, WAT regulation, and insulin sensitivity, and in contrast to TDAG51^−/− mice, PIKE^−/− mice are protected from diet-induced weight gain and IR (50). Thus, we are currently exploring whether the PH-like domain of TDAG51 is indeed mediating its effects on the regulation of lipogenesis and whole-body energy metabolism and whether the PI3K/Akt pathway may be involved.

Collectively, the results of our study illustrate that TDAG51 plays an important role as a regulator of carbohydrate and lipid metabolism, at least in part through regulating SREBP activity, which may contribute to the phenotypes observed with the metabolic syndrome. Understanding this newly identified cellular factor in lipid metabolism and adipogenesis as well as its precise regulation under normal and pathophysiological conditions will be crucial for the development of effective treatments against diabetes and obesity.

R.C.A. and G.R.S. are supported by grants from the Heart and Stroke Foundation of Ontario (NA-6024) and the Canadian Institutes of Health Research (CIHR) (MOP-111239 and MOP-67116). Financial support from St. Joseph’s Healthcare Hamilton is also acknowledged. K.N.M. is supported by National Institute of Diabetes and Digestive and Kidney Diseases grant R01-DK-074487-01. S.B. is supported by a Heart and Stroke Foundation of Canada Doctoral Award. M.D.F. is a CIHR Banting Postdoctoral Fellow. G.R.S. is a Canada Research Chair in Metabolism and Obesity. R.C.A. is a Career Investigator of the Heart and Stroke Foundation of Ontario (CI-7408) and holds the Amgen Canada Research Chair in Nephrology.

No potential conflicts of interest relevant to this article were reported.

S.B. performed experiments, collected and interpreted data, and wrote the manuscript. Š.L., M.D.F., R.P., and E.G.L. performed experiments, contributed to data interpretation, and edited the manuscript. H.J. and R.J.F. assisted with experiments. K.N.M., G.R.S., and R.C.A. contributed to data interpretation and discussion and reviewed the manuscript. R.C.A. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors thank Dr. Y. Choi (University of Pennsylvania, Philadelphia, PA) for kindly providing the TDAG51^+/− mice. The authors acknowledge the help of K. Karimi (McMaster University) with the BD cytometric bead array and are grateful to C. Saab and R. Rhem (McMaster University) for their technical assistance with the animal imaging.