Autotaxin Is Regulated by Glucose and Insulin in Adipocytes

Abstract

Autotaxin (ATX) is an adipokine that generates the bioactive lipid, lysophosphatidic acid. Despite recent studies implicating adipose-derived ATX in metabolic disorders including obesity and insulin resistance, the nutritional and hormonal regulation of ATX in adipocytes remains unclear. The current study examined the regulation of ATX in adipocytes by glucose and insulin and the role of ATX in adipocyte metabolism. Induction of insulin resistance in adipocytes with high glucose and insulin concentrations increased ATX secretion, whereas coincubation with the insulin sensitizer, rosiglitazone, prevented this response. Moreover, glucose independently increased ATX messenger RNA (mRNA), protein, and activity in a time- and concentration-dependent manner. Glucose also acutely upregulated secreted ATX activity in subcutaneous adipose tissue explants. Insulin elicited a biphasic response. Acute insulin stimulation increased ATX activity in a PI3Kinase-dependent and mTORC1-independent manner, whereas chronic insulin stimulation decreased ATX mRNA, protein, and activity. To examine the metabolic role of ATX in 3T3-L1 adipocytes, we incubated cells with the ATX inhibitor, PF-8380, for 24 hours. Whereas ATX inhibition increased the expression of peroxisome proliferator–activated receptor-γ and its downstream targets, insulin signaling and mitochondrial respiration were unaffected. However, ATX inhibition enhanced mitochondrial H₂O₂ production. Taken together, this study suggests that ATX secretion from adipocytes is differentially regulated by glucose and insulin. This study also suggests that inhibition of autocrine/paracrine ATX–lysophosphatidic acid signaling does not influence insulin signaling or mitochondrial respiration, but increases reactive oxygen species production in adipocytes.

Autotaxin (ATX) is a secreted lysophospholipase D that generates lysophosphatidic acid (LPA) by hydrolyzing lysophosphatidylcholine (LPC) (1). Upon binding to G-protein–coupled LPA receptors, six of which have been identified, LPA triggers cellular signaling that influences a variety of biological and pathophysiological processes, including brain development, embryo implantation, blood vessel formation, hair growth, neuropathic pain, fibrosis, cancer, and cardiovascular disease (1, 2). Although LPA can be generated by multiple pathways, the majority of extracellular LPA is produced by ATX (1). ATX is ubiquitously expressed; however, studies using mice with adipocyte-specific ATX deficiency have shown that adipocytes are a major source of circulating ATX (3, 4).

Recently, ATX has also been implicated in disorders of energy metabolism, although its specific role in obesity-related metabolic disease remains, in part, controversial. Prior studies by different groups either suggested that ATX contributes to (4, 5), or protects from (3), diet-induced obesity in mice. Irrespective of the effect of ATX on adiposity, almost all studies to date report that ATX-LPA signaling correlates with and/or contributes to the development of obesity-related insulin resistance and impaired glucose homeostasis in mice and humans. Dusaulcy et al. (3) found that plasma LPA levels and ATX messenger RNA (mRNA) expression in subcutaneous, perigonadal, and perirenal adipose tissue are increased in high-fat diet–fed mice. Obese-diabetic db/db mice, but not streptozotocin-treated (type 1 diabetic) mice, exhibited increased adipose tissue ATX mRNA expression compared with their lean, nondiabetic littermates (6, 7). Moreover, Boucher et al. (7) found that ATX mRNA expression was upregulated in intra-abdominal adipose tissue from obese patients with insulin resistance and impaired glucose tolerance compared with obese patients with normal glucose tolerance. Consistent with these findings, our data show that serum ATX levels correlate with measures of insulin resistance in obese nondiabetic humans (8, 9). Specifically, we demonstrated that serum ATX correlates with measures of obesity, impaired glucose homeostasis, and insulin resistance in older humans (9). Moreover, we identified serum ATX as an independent predictor of glucose infusion rate during a hyperinsulinemic-euglycemic clamp, and homeostatic model assessment of insulin resistance, after controlling for both sex and medication use in this population (9). In an earlier study, we also linked serum ATX to nonalcoholic fatty liver disease (NAFLD), a metabolic complication of excess adiposity (8). We showed that serum ATX is not only increased in severely obese, nondiabetic women with NAFLD, but significantly correlates with measures of adiposity, impaired glucose homeostasis, insulin resistance, and liver damage (8). In addition, log-transformed ATX levels served as an independent predictor of NAFLD in this cohort (8). Taken together, these data strongly suggest that serum ATX is associated with obesity, dysregulated glucose metabolism, and metabolic dysfunction-related comorbidities.

Despite the important role of adipocytes in controlling circulating ATX levels and prior studies implicating ATX in obesity-induced metabolic dysfunction, very little is known about how ATX is regulated in adipocytes, specifically by modulators of energy metabolism. We hypothesized that insulin and glucose would regulate adipocyte ATX expression and secretion and that, conversely, modulation of adipocyte ATX would influence adipocyte insulin signaling and/or mitochondrial energy metabolism. To test this hypothesis, we sought to determine the effects of glucose and insulin on ATX production and secretion in 3T3-L1 adipocytes and adipose tissue explants from mice. Given the apparent role of mitochondrial metabolism in insulin signaling (10), we also sought to determine the effects of ATX inhibition on insulin signaling and mitochondrial metabolism in 3T3-L1 adipocytes. Our data show that secreted ATX activity is acutely upregulated by high glucose and insulin levels in cultured adipocytes. The acute stimulatory effect of insulin on ATX secreted from adipocytes was dependent on PI3Kinase activation, but independent from stimulation of mTORC1. Our data also show that chronically, glucose increases ATX, whereas high insulin levels reduce ATX at the level of mRNA expression, protein expression, and activity in adipocytes. Inhibition of ATX activity did not influence insulin signaling or mitochondrial oxygen consumption but enhanced mitochondrial reactive oxygen species (ROS) production in 3T3-L1 adipocytes.

Taken together, our study contributes to the understanding of ATX regulation by modulators of energy metabolism. This study suggests that glucose and insulin play an important role in regulating ATX secretion from adipocytes, which may underlie changes in circulating ATX levels during feeding/fasting and obesity-related metabolic disease. This study also suggests that autocrine/paracrine ATX-LPA signaling does not directly influence insulin signaling or mitochondrial energy metabolism, but increases ROS production in adipocytes.

Materials and Methods

Chemicals and reagents

Unless otherwise stated, chemicals and reagents were obtained from Sigma. The ATX inhibitor, PF-8380, was prepared in five steps from 2-benzoxazolone and 3-chloropropanoyl chloride as detailed by St-Cœur et al. (11).

Animals

C57BL6J mice were procured from The Jackson Laboratory. Mice were housed on a 12-hour light:12-hour dark cycle with ad libitum access to chow diet (LD5001 from Laboratory diet with 13.5 kcal% from fat) or high-fat–high-sucrose (HFHS) diet (12451 from Research Diets with 45 kcal% from fat and 17 kcal% from sucrose) and water. Nine- to 10-week-old male mice were randomly assigned to chow or HFHS cohorts and fed for 16 weeks. Mice were euthanized either in the fed state (1-hour food withdrawal) or following a 16-hour fast. Blood was spun at 2000 × g for 15 minutes at 4°C to collect serum, which was frozen and stored at –80°C until further use. All protocols involving mice were approved by the Dalhousie University Institutional Animal Care and Use Committee.

Cell culture

As previously described, 3T3-L1 cells (ATCC) were grown and differentiated to mature adipocytes, with minor modifications (12). Briefly, 3 × 10⁵ 3T3-L1 cells were seeded in 35-mm dishes and maintained in Dulbecco’s modified Eagle medium (DMEM) containing high glucose (25 mM) concentrations (DMEM-HG; Hyclone Laboratories) supplemented with 10% fetal bovine serum (FBS; Seradigm). Two days postconfluence (day 0), cells were differentiated in DMEM-HG containing 10% FBS, 10 µg/mL insulin from bovine pancreas, 0.4 µg/mL dexamethasone and 0.5 mM 3-isobutyl-1-methylxanthine. After 2 days (day 2), the media was changed to DMEM-HG supplemented with 10% FBS and 10 µg/mL insulin. At day 4, the media was changed to DMEM-HG containing 10% FBS and 0.5 µg/mL insulin. After day 6, cells were maintained in DMEM-HG containing 10% FBS. Experiments were performed with adipocytes at days 8 to 9. Treatment of adipocytes with different glucose and/or insulin concentrations was performed either for 2 to 6 hours (acute exposure) or up to 30 hours (chronic exposure).

Insulin resistance was induced by 24-hour exposure to high glucose and insulin, as previously described (13). Briefly, adipocytes were washed once in phosphate-buffered saline (PBS) and incubated in 1 mL of DMEM-1X media (Thermo Fisher Scientific) supplemented with 4.5 g/L glucose (25.0 mM, Amresco), 0.5% [weight-to-volume ratio (w/v)] fatty acid–free (FAF) bovine serum albumin (BSA), 110 mg/mL sodium pyruvate (Sigma) and 100 nM insulin for 24 hours. Insulin-sensitive (IS) 3T3-L1 adipocyte controls were cultured in DMEM-1X supplemented with 1.1 g/L glucose (6.1 mM), 0.5% (w/v) FAF-BSA and 110 mg/mL sodium pyruvate. Where indicated, 1 µM rosiglitazone was added to the media of insulin-resistant (IR) or IS adipocytes and cells were incubated for 24 hours. Following incubation, media aliquots were collected and stored at –80°C until further analysis. Adipocytes were washed once in PBS and acutely stimulated with 20 nM insulin in 1 mL DMEM-1X + 1.1 g/L glucose for 15 minutes. Cells were washed and scraped in ice-cold PBS. Cells were subsequently pelleted through centrifugation at 10,000 × g for 10 min at 4°C, flash frozen in liquid nitrogen and stored at –80°C until further use. Cell pellets were lysed in 100 µL of lysis buffer (20 mM Tris-HCl pH 7.5, 5 mM EDTA, 10 mM Na₄P₂O₇, 100 mM NaF, 1% NP-40) containing 2 mM sodium orthovanadate, 2 mM protease inhibitor cocktail (P8340, Sigma), and 100 µg/mL phosphatase inhibitor cocktail (524628, Calbiochem). Protein concentrations in cell lysates were quantified colorimetrically using a bicinchoninic acid protein assay kit (Thermo Scientific) and BSA as standard.

To examine how glucose and insulin regulate ATX secretion acutely (6 hours) and chronically (30 hours), 3T3-L1 adipocytes were washed once in PBS and preincubated for 30 minutes with 1 mL of DMEM-1X + 4.5 g/L glucose + 0.5% (w/v) FAF-BSA and 110 mg/mL sodium pyruvate with the following chemical inhibitors, as indicated: 5 mg/mL actinomycin D, 1 µM wortmannin, 100 nM rapamycin (Alfa Aesar), a combination of 5 µg/mL brefeldin A and 5 µM monensin (Enzo), or 10 µg/mL cycloheximide (Biovision). Following preincubation, 100 nM insulin or PBS was added to adipocytes (t = 0). For acute stimulations, media aliquots of 100 µl were collected and stored at 2, 4, and 6 hours following incubation with PBS or insulin. Due to the short half-lives of wortmannin, brefeldin A/monensin and cycloheximide, an equal volume of DMEM-1X supplemented with 4.5 g/L glucose and 0.5% (w/v) FAF-BSA plus indicated inhibitors was added to cells after an aliquot of media was removed. For chronic stimulation, inhibitors were added every 6 hours. ATX activity in the media was adjusted for dilution factor. Adipocytes were collected and homogenized as described previously. All cell culture data reflect at least three independent experiments.

Subcutaneous adipose tissue explants

Adipose tissue explants were prepared as previously described (14). Briefly, subcutaneous adipose tissue (SCAT) from female C57Bl6 mice was surgically removed, washed once in PBS and incubated in prewarmed (37°C) DMEM-1X for 30 minutes. SCAT explants were cut into pieces (∼20 mg) using scissors and incubated in 100 μL DMEM-1X containing 2% FAF-BSA, 110 mg/mL sodium pyruvate and indicated glucose concentrations for 8 hours. Thereafter, media aliquots were collected and stored at –80°C until further analysis. SCAT explants were homogenized in ice-cold lysis buffer, incubated for 30 minutes on ice and spun at 14,000 × g for 30 minutes at 3°C. The resulting supernatants were spun again at 14,000 × g for 30 minutes at 3°C. Lysate protein concentrations were quantified using a bicinchoninic acid protein assay kit and BSA as standard.

ATX activity assay

ATX activity in serum or conditioned media was determined using FS-3 (Echelon), a fluorogenic LPC analog (15). 10 µl of serum or media were incubated with 10 µL of 100 µM FS-3 in 80 µL of freshly prepared assay buffer (50 mM Tris, 140 mM NaCl, 5 mM KCl, 1 mM CaC`l₂, 1 mM MgCl₂, pH 8.0). Samples were incubated at 37°C for 2 hours, during which fluorescent measurements were taken every 5 minutes. ATX activity was quantitated by measuring the rate of linear increase in fluorescence at 528 nm with excitation at 485 nm and was expressed as relative fluorescence units/(minutes × milligrams of cellular protein).

Immunoblotting analysis

The 3T3-L1 cell lysates were subjected to SDS-PAGE and proteins were transferred onto a nitrocellulose membrane. Proteins were visualized using a reversible protein stain (Memcode, Pierce, Thermo Fisher Scientific) and membranes were incubated with the following primary antibodies: anti-ATX (Cayman Chemicals, 10005375), anti-adiponectin (Novus, NBP2-22450), anti-pAkt Ser⁴⁷³ (Cell Signaling, 9271), anti-Akt (Millipore, 05-591), anti-pP70S6K Thr³⁸⁹ (Cell Signaling, 9234), anti-p70S6K (Cell Signaling, 2708), anti-peroxisome proliferator–activated receptor-γ (PPAR-γ; Cell Signaling, 2435), anti-CD36 (gift from Dr. Ong), and anti-glucose transporter 4 (Glut4; Millipore, 07-140). Immunoblots were developed using the Western Lightning Plus-ECL enhanced chemiluminescence substrate (Perkin Elmer). Densitometric analysis was performed using Image Lab software (Bio-Rad).

RNA extraction, reverse transcription, and gene expression analysis

RNA was isolated from 3T3-L1 cells following indicated treatments using RIBOZOL (Amresco) and chloroform following the manufacturer’s directions. The RNA was resuspended in 50 µL nuclease-free water (Ambion). The quality and quantity of RNA was assessed using a QIAxcel Advanced System (Qiagen) and QIAxcel RNA QC Kit v2.0 (Qiagen) according to the manufacturer’s instructions. Complementary DNA (cDNA) was synthesized using qScript cDNA supermix (Quanta Biosciences) from 500 ng of RNA. Quantitative polymerase chain reactions (PCRs) were carried out in 96-well plates on a ViiA7 real-time PCR machine (Thermo Fisher Scientific) and contained 2 μL of cDNA template, 5 µL of SYBR green Low ROX PCR supermix (Thermo Fisher Scientific), 0.25 µM for each forward and reverse primer, and nuclease-free water in a total of 10 µL per reaction. Primer sequences used to amplify ATX were ATX-F 5′-CTTGTGAAACGTTACGCTA-3′ and ATX-R 5′-CTTCATTATCTGATCGGTGT-3′. In addition, primers used for the reference genes, ribosomal protein L27 and 41 were Rpl27F 5′-ACGGTGGAGCCTTATGTGAC-3′, Rpl27R 5′-TCCGTCAGAGGGACTGTCTT-3′, Rpl41F 5′-GCCATGAGAGCGAAGTGG-3′, and Rpl41R 5′-CTCCTGCAGGCGTCGTAG-3′. ATX mRNA levels were determined using Biogazelle qbase+ software and presented as fold change.

Mitochondrial analysis

Respiratory oxygen flow in permeabilized 3T3-L1 adipocytes was measured in high resolution using the Oxygraph-2k (OROBOROS Instruments). Samples (400,000 cells/mL) were assessed in 2 mL of 37°C, air O₂-saturated respiration assay buffer [buffer Z; recipe described previously (16)], and reoxygenated as necessary throughout the protocol. Instrumental background O₂ consumption was corrected using equations determined under the same parameters used for experimental data collection. Tests were conducted prior to experimental data collection, in which it was determined that 2 µg/mL digitonin was optimal for permeabilizing the cells in this study. The respirometric protocol involved the sequential addition of substrates and inhibitors and/or titration of substrates, as follows: 5 mM ADP, 4 mM malate, 2 µg/mL digitonin, 5 mM lactate, 2.5 mM NAD⁺, 5 mM NAD⁺, 10 mM lactate, 20 mM lactate, 5 mM pyruvate, 0.5 µM carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP; an uncoupling protonophore), 1.0 µM FCCP, 1.5 µM FCCP, 2.0 µM FCCP, 5 mM glutamate, 10 mM glutamate, 10 mM succinate, 25 µM palmitoyl-carnitine, 0.5 µM rotenone, 2.5 µM antimycin A, 500 µM N,N,N',N'-tetramethyl-p-phenylenediamine dihydrochloride + 2 mM ascorbate, 200 mM sodium azide.

H₂O₂ production was monitored in permeabilized adipocytes simultaneously with respirometry using an amperometric add-on module to the OROBOROS Oxygraph-2k. Prior to the first substrate addition of the respirometric protocol, 10 µM Amplex Red Ultra (Life Technologies) and 1 U/mL horseradish peroxidase were added to the oxygraph chambers containing the cells. H₂O₂ calibration experiments were performed under the same parameters used for experimental data collection. The relationship between H₂O₂ and fluorescence intensity was linear to at least 0.6 µM.

Statistical analysis

Results are expressed as mean ± standard error of the mean. Comparisons between two groups were performed using an unpaired, two-tailed Student’s t test. Comparisons between multiple groups were performed using a paired or unpaired one- or two-way analysis of variance followed by a Tukey or Bonferroni post hoc test, as appropriate. All statistical analysis was performed using Prism (GraphPad Software). P values of less than 0.05 were considered statistically significant.

Results

ATX is regulated by acute and chronic nutritional stimuli in vivo

It remains unclear how ATX is regulated by short-term (i.e., feeding/fasting) and long-term changes in nutritional status (i.e., obesity). Prior studies examining the regulation of ATX in murine adipose tissue during obesity provided conflicting results (3, 4, 7). To test whether serum ATX is regulated by acute nutritional changes in vivo, we isolated serum from fed and 16-hour–fasted C57Bl6 mice and measured ATX activity [Fig. 1(a)]. ATX activity was markedly decreased in serum from fasted mice compared with fed mice, suggesting acute nutritional regulation of ATX in vivo. To examine whether this effect persisted even after the induction of obesity, we subjected C57Bl6 mice to 16 weeks of HFHS or chow diet feeding. HFHS-fed mice displayed a 44% increase in body weight compared with chow-fed mice [Fig. 1(b)]. We have previously shown that this feeding regimen not only leads to obesity, but impaired glucose homeostasis and cardiac dysfunction (17). Serum ATX activity was upregulated in obese HFHS-fed mice compared with chow-fed mice [Fig. 1(c)]. Fasting led to a decrease in serum ATX activity in both chow-fed and HFHS-fed mice [Fig. 1(c)]. Because it has recently been suggested that lipids contained in serum may interfere with the fluorogenic ATX activity assay using FS-3 (18), we also confirmed these nutritional changes in ATX activity by quantifying choline released from LPC (data not shown). Taken together, these data suggest that acute fasting decreases serum ATX activity, whereas the induction of obesity using a diet rich in fat and sucrose increases serum ATX in mice.

Figure 1.

ATX is regulated by acute and chronic nutritional stimuli in mice. (a) Serum ATX activity in 8-week-old chow-fed male C57Bl6 mice following a 1-hour food withdrawal (fed) or 16-hour fasting (n = 4). (b) Body weight and (c) serum ATX activity in 25- to 26-week-old male C57Bl6 mice fed a chow or HFHS diet for 16 weeks (n = 10 to 15). Statistical analysis was performed using an unpaired two-tailed Student’s t test (a, b) or a two-way analysis of variance followed by a Tukey’s multiple comparison test (c); ***P < 0.001; ****P < 0.0001; ^####P < 0.0001 vs chow.

ATX is increased in insulin-resistant 3T3-L1 adipocytes exposed to high glucose and high insulin concentrations

Adipose tissue is a major source of circulating ATX, accounting for ∼40% of serum ATX and LPA in mice (3, 4). We used 3T3-L1 adipocytes, which display a marked increase in secreted ATX activity during differentiation [Fig. 2(a)], to examine the regulation of ATX in vitro. To determine whether the upregulation of ATX activity in serum from obese mice [Fig. 1(c)] can be mimicked by exposing adipocytes to an obese-diabetic milieu, we incubated differentiated 3T3-L1 adipocytes in the presence of high levels of glucose (25 mM) and insulin (100 nM) for 24 hours [Fig. 2(b–e)]. Subsequent insulin signaling analysis confirmed that adipocytes exposed to high glucose and insulin concentrations were IR because acute insulin stimulation failed to increase Akt phosphorylation at Ser⁴⁷³ [Fig. 2(b) and 2(c)], consistent with a previous study (13). Protein expression of ATX in the media and cell lysates was markedly increased in IR adipocytes compared with the IS control cells [Fig. 2(b) and 2(d)]. Upregulation of ATX protein expression was paralleled by a corresponding increase in ATX activity in the media from IR compared with IS adipocytes [Fig. 2(e)]. As expected, adiponectin levels in the media were reduced in IR adipocytes compared with control cells [Fig. 2(b) and 2(d)]. Consistent with a prior study demonstrating that the insulin sensitizer and PPAR-γ agonist, rosiglitazone, diminishes ATX mRNA levels in 3T3F442A adipocytes (7), our data show that incubation with rosiglitazone decreases ATX mRNA and secreted ATX protein levels and activity in both IS and IR 3T3-L1 adipocytes [Fig. 2(f–h)]. Taken together, these data suggest that chronic exposure of 3T3-L1 adipocytes to conditions that mimic an obese-diabetic environment (i.e., high glucose and insulin concentrations) lead to the upregulation of ATX expression and corresponding increase in secreted ATX activity, whereas insulin sensitization via PPAR-γ activation prevented these effects.

Figure 2.

ATX is upregulated in IR 3T3-L1 adipocytes. (a) Secreted ATX activity during the course of 3T3-L1 adipocyte differentiation from preadipocytes (day 0) to mature, differentiated adipocytes (day 8). (b, c) Differentiated 3T3-L1 adipocytes were incubated in presence of 25 mM glucose and 100 nM insulin for 24 hours (IR) and induction of insulin resistance was determined through immunoblotting analysis of Akt phosphorylation at Ser⁴⁷³ following stimulation with 20 nM insulin for 10 minutes. IS 3T3-L1 adipocytes, incubated with 6 mM glucose and in the absence of insulin, were used as controls. (b, d) Protein levels of secreted and cellular ATX (ATX-S, ATX-C) and adiponectin (Adn-S, Adn-C). (e) Secreted ATX activity in the media of IS and IR 3T3-L1 adipocytes. (f) ATX/Enpp2 mRNA, (g) secreted protein, and (h) activity following incubation of IS and IR adipocytes with 1 µM rosiglitazone (Rosi) or DMSO [control (Ctrl)]. Statistical analysis was performed using one-way (a) or two-way analysis of variance followed by a Tukey’s multiple comparison test (c, d, f–h), or an unpaired two-tailed Student’s t test (e). **P < 0.01; ***P < 0.001; ****P < 0.0001; ^###P < 0.001; ^####P < 0.0001 vs IS Ctrl; n = 9 from at least three independent experiments. PS, protein stain.

Glucose and insulin differentially regulate ATX in adipocytes

To differentiate the effect of glucose and insulin, per se, on the regulation of ATX expression and activity in adipocytes, we incubated 3T3-L1 adipocytes with either no glucose, low (6 mM) glucose or high (25 mM) glucose concentrations in the presence or absence of 100 nM insulin for 2 hours, 6 hours, and 30 hours and determined ATX mRNA levels, cellular ATX protein levels, and secreted ATX activity [Fig. 3(a–i)]. At 2 hours, ATX mRNA, protein, and activity were similar across all groups [Fig. 3(a–c)]. At 6 hours, high (25 mM) glucose but not low (6 mM) glucose significantly upregulated ATX mRNA, which was paralleled by a mild increase in ATX protein and activity [Fig. 3(d–f)]. Insulin increased ATX activity independent of glucose at 6 hours [Fig. 3(f)]. Interestingly, insulin-stimulated upregulation of secreted ATX was paralleled by a decrease in ATX mRNA levels in the high glucose group [Fig. 3(d)], whereas cytosolic ATX protein levels were not influenced by insulin at 6 hours [Fig. 3(e)]. At 30 hours, both low (6 mM) and high (25 mM) glucose concentrations increased ATX mRNA, protein, and activity [Fig. 3(g–i)]. The glucose-induced upregulation of ATX mRNA, protein, and activity was inhibited by the mRNA synthesis inhibitor actinomycin D [Fig. 4(a–d)], suggesting that stimulation of ATX mRNA synthesis is necessary for glucose-induced upregulation of ATX activity. Moreover, incubation of cells with cycloheximide, an inhibitor of protein synthesis, or a combination of brefeldin A and monensin, inhibitors of protein secretion through the classical secretory pathway, diminished ATX activity in the high glucose group comparable to control [Fig. 4(d)], suggesting that the upregulation of ATX by glucose also involves the classical secretory pathway and synthesis of new ATX protein. As expected, cycloheximide decreased whereas brefeldin A/monensin increased cytosolic ATX protein levels [Fig. 4(b) and 4(c)]. Interestingly, a 30-hour incubation with 100 nM insulin significantly reduced ATX mRNA levels, cytosolic ATX protein levels and ATX activity independent of glucose [Fig. 3(g–i)]. To examine whether the effect of insulin on ATX is concentration dependent, we incubated adipocytes with either 0, 0.01, 0.1, 1, 10, or 100 nM insulin for up to 30 hours and determined secreted ATX activity. Insulin concentrations ranging from 0.1 to 100 nM gradually increased ATX activity for up to 16 or 24 hours compared with controls incubated in the absence of insulin [Fig. 4(e) and 4(f)]. At 30 hours, 100 nM insulin reduced ATX activity although lower insulin concentrations had no effect on ATX activity [Fig. 4(e) and 4(f)].

Figure 3.

Glucose and insulin differentially regulate ATX in adipocytes. 3T3-L1 adipocytes were exposed to 0, 6 or 25 mM glucose in the presence or absence of 100 nM insulin for 2 hours (a–c), 6 hours (d–f), or 30 hours (g–i), and (a, d, g) ATX/Enpp2 mRNA, (b, e, h) cellular ATX (ATX-C) protein levels, and (c, f, i) secreted ATX (ATX-S) activity were determined. Statistical analysis was performed using a two-way analysis of variance followed by a Tukey’s multiple comparison test. *P < 0.05; **P < 0.01; ****P < 0.0001; ^#P < 0.05; ^###P < 0.001; ^####P < 0.0001 vs untreated controls; n = 9 from at least three independent experiments. PS, protein stain.

Figure 4.

Glucose increases ATX/Enpp2 mRNA synthesis and insulin regulates ATX activity in a time- and concentration-dependent manner. (a–d) 3T3-L1 adipocytes were incubated with media containing 0 or 25 mM glucose for 30 hours. Where indicated, adipocytes were additionally incubated with either 5 µg/mL actinomycin D (ACT), 10 µg/mL cycloheximide (CHX), a combination of 5 µg/mL brefeldin A and 5 µM monensin (B/M) or DMSO/methanol [control (Ctrl)]. (a) ATX/Enpp2 mRNA, (b, c) cytosolic ATX (ATX-C) protein expression, (d) secreted ATX (ATX-S) activity. (e, f) 3T3-L1 adipocytes were incubated in media containing 25 mM glucose and either 0, 0.01, 0.1, 1, 10, or 100 nM insulin for 0, 1, 4, 16, 24, or 30 hours and secreted ATX activity was determined. (g) SCAT explants from 25- to 40-week-old chow-fed female C57Bl6 mice were incubated in the presence of 0, 6, or 25 mM glucose for 8 hours and secreted ATX activity was determined. Statistical analysis was performed using (a, c–f) two-way or (g) one-way analysis of variance followed by a Tukey’s multiple comparison test. (a, c, d, g) *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ^#P < 0.05; ^####P < 0.0001 vs no-glucose controls. (e) ⁺⁺⁺⁺P < 0.0001 for 0.01 nM vs 0 nM insulin; ^####P < 0.0001 for 0.1 nM vs 0 nM insulin. (f) ^##P < 0.01 and ^####P < 0.0001 for 1 nM vs 0 nM insulin; ⁺⁺⁺⁺P < 0.0001 for 10 nM vs 0 nM insulin; ***P < 0.001 for 100 nM vs 0 nM insulin. (a, c–f) n = 9 from at least three independent experiments. (g) n = 5 to 10 from five to 10 mice. PS, protein stain.

To determine whether glucose modulates ATX also in whole adipose tissue, we incubated SCAT explants from C57Bl6 mice with either low or high glucose (6 or 25 mM) for 8 hours and determined secreted ATX activity compared with no glucose [Fig. 4(g)]. Similar to the regulation of ATX by glucose in 3T3-L1 adipocytes, ATX activity was upregulated by glucose in a concentration-dependent manner in SCAT explants, with an approximate threefold increase observed in the high-glucose group [Fig. 4(g)]. Taken together, these data suggest that glucose increases ATX mRNA expression, protein expression, and activity in a time- and concentration-dependent manner in 3T3-L1 adipocytes. These data also suggest that exposure to high insulin concentrations elicits a biphasic response in 3T3-L1 adipocytes—insulin initially increases secreted ATX activity; however, long-term incubation with insulin decreases ATX activity due to downregulation of ATX mRNA and protein expression.

Acute stimulation of ATX secretion by insulin is dependent on PI3Kinase but not mTOR activation

To examine potential mechanisms that underlie the transient stimulatory effect of insulin on ATX secretion, 3T3-L1 adipocytes were incubated with wortmannin, a PI3Kinase inhibitor, or rapamycin, an inhibitor of mTORC1, in the presence or absence of 100 nM insulin for 6 hours. As expected, wortmannin diminished insulin-stimulated phosphorylation of Akt at Ser⁴⁷³ and p70S6K at Thr³⁸⁹ [Fig. 5(a)]. Wortmannin also blunted the increase in ATX activity induced by insulin, suggesting that PI3Kinase activation was essential for insulin stimulation of ATX secretion [Fig. 5(e)]. This was mirrored by decreased secreted and cellular ATX protein levels in the media of wortmannin-treated cells [Fig. 5(a–d)]. Incubation with rapamycin led to diminished phosphorylation of p70S6K at Thr³⁸⁹, a known mTOR target site, whereas Akt phosphorylation at Ser⁴⁷³ was unchanged [Fig. 5(a)]. In contrast to PI3Kinase inhibition, inhibition of mTOR signaling using rapamycin had no substantial effect on ATX activity or protein at baseline or in the presence of insulin [Fig. 5(a–e)]. Incubation of adipocytes with brefeldin A/monensin blunted secreted ATX activity and protein levels in the absence and presence of insulin [Fig. 5(c–d)], which resulted in an increased accumulation of cellular ATX protein levels [Fig. 5(a) and 5(b)]. Incubation of adipocytes with cycloheximide reduced cellular ATX protein levels, which was accompanied by a drastic reduction in secreted ATX activity and protein in presence and absence of insulin [Fig. 5(a–e)]. Taken together, these data suggest that acute insulin-stimulated ATX secretion is PI3Kinase-dependent, but does not require mTOR activation. These data also suggest that ATX secretion at baseline and following acute insulin stimulation requires the classical secretory pathway and synthesis of new ATX protein.

Figure 5.

Acute insulin stimulation of ATX secretion is PI3Kinase-dependent and mTOR-independent. 3T3-L1 adipocytes were incubated with media containing 25 mM glucose in the presence or absence of 0 or 100 nM insulin for 6 hours. Where indicated, adipocytes were additionally treated with either 1 µM wortmannin (W), 100 nM rapamycin (R), 10 µg/mL cycloheximide (CHX), or a combination of 5 µg/mL brefeldin A and 5 µM monensin (B/M). (a, b) Cellular ATX (ATX-C) protein expression, (c, d) secreted ATX (ATX-S) protein expression, and (e) activity. Statistical analysis was performed using one-way analysis of variance followed by a Tukey’s multiple comparison test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 vs untreated controls; ^##P < 0.01; ^###P < 0.001; ^####P < 0.0001 for insulin plus chemical inhibitors vs insulin minus chemical inhibitors; n = 9 from at least three independent experiments. PS, protein stain.

Inhibition of ATX activity does not influence insulin signaling or mitochondrial oxygen consumption but enhances mitochondrial reactive oxygen species production in 3T3-L1 adipocytes

To examine the role of ATX in adipocyte metabolism, we determined the expression of proteins involved in fat and glucose metabolism and insulin signaling in IS and IR 3T3-L1 adipocytes incubated in the presence or absence of the ATX inhibitor, PF-8380 (11). Secreted ATX activity was blunted in both IS and IR adipocytes following a 24-hour incubation with PF-8380 [Fig. 6(a)], whereas cytosolic and secreted ATX protein levels remained unchanged [Fig. 6(b–d)]. Consistent with previous studies suggesting that the ATX-LPA signaling pathway inhibits PPAR-γ (3, 19), inhibition of ATX activity increased protein levels of PPAR-γ and the PPAR-γ targets, adiponectin, CD36, and Glut4 [Fig. 6(b, f–j)] in IS cells. However, the expression of these proteins was not affected by ATX inhibition in IR cells [Fig. 6(b, f–j)]. ATX inhibition was also unable to rescue the decrease in PPAR-γ, adiponectin, CD36, and Glut4 protein levels in IR compared with IS cells [Fig. 6(b, f–j)]. Despite increased PPAR-γ and PPAR-γ target protein expression following ATX inhibition, insulin-stimulated phosphorylation of Akt at Ser⁴⁷³ was similar between IS 3T3-L1 adipocytes incubated with and without ATX inhibitor [Fig. 6(b) and 6(e)]. Moreover, 24-hour ATX inhibition did not alter impaired Akt phosphorylation in IR cells [Fig. 6(b) and 6(e)].

Figure 6.

Inhibition of ATX activity upregulates PPAR-γ but does not influence insulin signaling in 3T3-L1 adipocytes. Differentiated 3T3-L1 adipocytes were incubated with 6 mM glucose (IS) or 25 mM glucose and 100 nM insulin for 24 hours (IR) in the presence of 5 µM PF-8380 or DMSO [control (Ctrl)]. (a) ATX activity and (b, c) secreted (ATX-S) and (b, d) cellular ATX (ATX-C) protein levels, (b, e) Akt phosphorylation at Ser⁴⁷³, and protein expression of (b, f) PPAR-γ, (b, g) secreted adiponectin (Adn-S), (b, h) cellular adiponectin (Adn-C), (b, i) CD36, and (b, j) Glut4 were determined. Statistical analysis was performed using two-way analysis of variance followed by a Tukey’s multiple comparison test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ^#P < 0.05; ^##P < 0.01; ^###P < 0.001; ^####P < 0.0001 vs IS; n = 12 from at least three independent experiments.

To examine whether ATX influences adipocyte metabolism at the level of mitochondria, we incubated IR 3T3-L1 adipocytes with PF-8380 for 24 hours and determined respiratory oxygen flow in permeabilized cells [Fig. 7(a)]. The multisubstrate protocol revealed no substantial differences in respiratory oxygen consumption among treatments in the adipocytes [Fig. 7(a)]. However, ATX inhibition led to significantly greater rates of H₂O₂ emission in IR cells [Fig. 7(b)]. When expressed as a percentage of total oxygen consumed, H₂O₂ production was significantly greater in the PF8380-treated cells following addition of the mitochondrial complex III inhibitor, antimycin A [Fig. 7(c)]. Interestingly, lactate oxidation elicited rates of H₂O₂ production that were decreased upon addition of pyruvate [Fig. 7(b)]. Taken together, these data suggest that ATX inhibition for 24 hours stimulates PPAR-γ in IS but not IR 3T3-L1 adipocytes and does not affect insulin-stimulated Akt phosphorylation. Our data also suggest that mitochondrial oxygen consumption remains unchanged following ATX inhibition, however, ROS production is enhanced.

Figure 7.

Inhibition of ATX activity does not alter mitochondrial respiration but increases ROS production. Differentiated 3T3-L1 adipocytes were incubated with 6 mM glucose (IS) or 25 mM glucose and 100 nM insulin for 24 hours (IR) in the presence of 5 µM PF-8380 or DMSO [control (Ctrl)]. (a) Mitochondrial O₂ flow, (b) H₂O₂ production, and (c) H₂O₂ production expressed as a percentage of O₂ consumed in permeabilized adipocytes as determined by high-resolution respirometry and fluorometry. Statistical analysis was performed using two-way analysis of variance followed by a Bonferroni multiple comparison test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; n = 4 to 6 from four independent experiments.

Discussion

Adipose-derived ATX has been implicated in metabolic disorders including obesity and insulin resistance (3, 4, 6–9). However, the regulation of ATX in adipocytes remains incompletely understood. Specifically, it is unclear whether ATX secretion from adipocytes is influenced by modulators of energy metabolism. In this study we show that serum ATX activity is not only increased by chronic overfeeding in mice but responds acutely to nutritional stimuli as was evidenced by a downregulation of ATX activity upon fasting. In addition, ATX secretion from cultured adipocytes and explanted adipose tissue was acutely (6 hours) stimulated by high levels of glucose and insulin in a synergistic manner. Moreover, we demonstrate that acute insulin stimulation of ATX secretion from adipocytes is mediated by PI3Kinase but not mTOR signaling. Our data also suggest that upregulation of ATX secretion by glucose and insulin requires the classical secretory pathway and synthesis of new ATX protein. In addition, glucose also increased ATX mRNA synthesis. Interestingly, although the stimulatory effect of glucose on ATX was also observed following prolonged incubation (30 hours), chronic stimulation with high levels of insulin decreased ATX mRNA, protein, and activity, suggesting that insulin at concentrations mimicking an obese-IR milieu has a biphasic effect on ATX secretion in adipocytes. Our data also show that pharmacological inhibition of ATX for 24 hours leads to the upregulation of PPAR-γ in IS, but not IR adipocytes, and does not influence insulin-stimulated Akt phosphorylation. Furthermore, ATX inhibition did not alter mitochondrial O₂ consumption; however, it led to enhanced H₂O₂ production. Taken together, this study suggests that ATX secretion from adipose tissue is influenced acutely and chronically by changes in glucose homeostasis, which may underlie its regulation during feeding/fasting and obesity/insulin resistance. This study also suggests that ATX inhibition for 24 hours does not influence insulin signaling or mitochondrial respiration, but increases ROS production. The latter may suggest that autocrine or paracrine ATX-LPA signaling does not play a role in modulating insulin sensitivity and energy metabolism in adipocytes. However, it is possible that ATX-LPA signaling-mediated inhibition of mitochondrial ROS emission influences adipocyte differentiation (4, 19, 20) and adipocyte oxidative stress during obesity (21).

Previous studies examining the regulation of ATX in adipocytes have shown that ATX is markedly upregulated during adipocyte differentiation (4, 6). Ferry et al. (6) observed that ATX/Enpp2 mRNA expression increases during differentiation in cultured murine 3T3F442A adipocytes, peaking at 10 days following differentiation start. The increase in ATX mRNA expression coincided with elevated LPA levels in the media, consistent with an upregulation of ATX secretion (6). Differentiation-dependent upregulation of ATX mRNA was also observed in primary mouse preadipocytes with a peak at days 6 to 12, whereas it returned to lower levels as adipocytes matured and became hypertrophic (4, 6). Consistent with these studies, we demonstrated a marked upregulation of secreted ATX activity during 3T3-L1 adipocyte differentiation starting from day 2 of differentiation.

The regulation of ATX during diet-induced obesity remains controversial as prior studies suggested that ATX is either upregulated (3), downregulated (4), or unchanged (7) in obese mice fed a high-fat diet. This discrepancy was attributed to mouse strain differences, starting age of mice for high fat feeding, and possible environmental factors (4). It is plausible that diet composition and feeding duration also contributed to these differences. We showed that serum ATX activity is upregulated in obese mice with impaired glucose tolerance fed HFHS diet in both fed and 16-hour–fasted states, suggesting that ATX secretion from adipocytes is increased in our obesity model. Upregulation of ATX during diet-induced obesity in mice corresponds with prior studies from our group showing a positive correlation between serum ATX levels and measures of adiposity (body mass index, waist circumference), as well as impaired glucose homeostasis/insulin resistance (fasting glucose, fasting insulin, glucose levels 2 hours following an oral glucose tolerance test, glucose infusion rate during a hyperinsulinemic euglycemic clamp, homeostatic model assessment of insulin resistance, and quantitative insulin sensitivity check index) in relatively large human cohorts of 60 to 101 nondiabetic individuals (8, 9).

A study by Boucher et al. (7) suggested that two modulators of insulin sensitivity, tumor necrosis factor-α (TNF-α) and rosiglitazone, influence ATX expression in adipocytes. Specifically, although incubation of 3T3F442A adipocytes with the proinflammatory cytokine TNF-α, which promotes insulin resistance, led to an increase in ATX mRNA expression, the insulin-sensitizing drug rosiglitazone decreased ATX mRNA levels (7). We demonstrated similar results in 3T3-L1 adipocytes where incubation with rosiglitazone decreased ATX mRNA, protein, and activity in both IS and IR cells. These findings suggest that ATX expression is upregulated during insulin resistance-promoting conditions in adipocytes and can be decreased using insulin-sensitizing drugs. Consistent with this concept, ATX secretion is markedly increased following a 24-hour incubation of 3T3-L1 adipocytes with high concentrations of insulin and glucose, which coincided with the development of insulin resistance in these cells. Interestingly, both insulin and glucose upregulated ATX secretion in adipocytes acutely, and fasting decreased serum ATX activity in vivo. These data suggest that ATX secretion is modulated by short-term changes in energy metabolism. PI3Kinase activation appears to be critical for acute insulin-stimulated increases in ATX secretion, which is dependent, in part, on new protein synthesis and the classical secretory pathway. The latter observation is consistent with a previous study showing that secretion of human ATX expressed in HEK293T cells is blocked by the Golgi-disturbing agents, brefeldin A and monensin (22). Interestingly, prolonged stimulation with high levels of insulin led to the downregulation of ATX in adipocytes, which appeared to be driven by a reduction in ATX mRNA expression. Although the mechanisms for the insulin-mediated reduction in ATX mRNA expression have not been examined in this study, it is conceivable that ATX mRNA synthesis is under the control of transcription factors such as Forkhead box–containing protein O subfamily (FoxO) proteins, which are inhibited by insulin receptor-Akt signaling (23).

In contrast to insulin, high levels of glucose increased ATX mRNA synthesis in adipocytes both acutely and chronically, which coincided with increased ATX protein expression and secreted ATX activity involving new protein synthesis and the classical secretory pathway, respectively. The mechanisms underlying the stimulatory effect of glucose on ATX secretion remain unclear, but may involve the induction of an inflammatory response and/or the generation of ROS, processes that are triggered by exposure of adipocytes to hyperglycemic conditions (24). It has previously been shown that high (25 mM) glucose concentrations in the media activate the proinflammatory transcription factor nuclear factor-κB (NF-κB) in vascular smooth muscle cells, which was paralleled by increased superoxide production. Hyperglycemia-induced upregulation of NF-κB appeared to be dependent on protein kinase C activation, a mechanism that has been suggested to contribute to diabetes-related vascular smooth muscle cell injury (24). Han et al. (25) also reproduced these findings in 3T3-L1 adipocytes, where exposure to high glucose led to NF-κB transactivation while transactivation of PPAR-γ, an anti-inflammatory transcription factor, was reduced. Interestingly, NF-κB was required for TNF-α–induced upregulation of ATX expression in hepatoma cells (26). A potent role of proinflammatory cytokines in the regulation of ATX has also been demonstrated by Benesch et al. (18) as both TNF-α and interleukin-β−1 were able to negate the negative feedback inhibition of ATX expression by LPA and sphingosine 1-phosphate in thyroid cancer cells. Likewise, proinflammatory toll-like receptor 4/interferon signaling increased ATX expression via the JAK-STAT and PI3Kinase/Akt pathway in monocytic THP-1 cells (27). A role for the PI3Kinase/Akt signaling pathway in NF-κB activation via I-κB kinases and FOXO1 homeostatic model assessment of insulin resistance, and quantitative insulin sensitivity check index has also been established in nonadipocyte cell types (28, 29). Therefore, it is tempting to speculate that the regulation of ATX by glucose and insulin converges at the level of NF-κB activation, which should be examined in future studies. In addition to the induction of adipocyte inflammation and ROS production, glucose could also regulate ATX secretion via modulation of protein glycosylation. It has previously been suggested that N-glycosylation is required for ATX secretion and activity in 3T3F442A adipocytes (30). Specifically, blockade of N-glycosylation using tunicamycin or deletion of glycosylation sites reduced ATX secretion (30).

To examine the autocrine/paracrine role of ATX in adipocyte insulin signaling and mitochondrial metabolism, we incubated 3T3-L1 adipocytes with PF-8380, a potent ATX inhibitor [in vitro half-maximal inhibitory concentration for murine ATX is 1.16 nM (31)], which diminished secreted ATX activity without affecting protein content. A 24-hour incubation with PF-8380 led to the upregulation of PPAR-γ and PPAR-γ targets in IS cells, consistent with prior studies suggesting that the ATX-LPA signaling pathway inhibits PPAR-γ in preadipocytes in vitro (19) and in adipose tissue in vivo (3), although Nishimura et al. (4) reported decreased PPAR-γ expression in preadipocytes from heterozygous ATX knockout mice at baseline and following diet-induced obesity. Because ATX protein levels did not change in response to ATX inhibition, our data suggest that ATX-mediated inhibition of PPAR-γ in adipocytes is at least in part due to LPA production and signaling. However, it remains to be determined whether activity-independent roles of ATX (32, 33) could also contribute to this effect. Further studies employing siRNA-mediated ATX knockdown in differentiated adipocytes would allow us to distinguish between LPA-dependent and LPA-independent roles of ATX in adipocytes. In contrast to the PF-8380-induced PPAR-γ stimulation in IS adipocytes, ATX inhibition was not able to rescue the downregulation of PPAR-γ following induction of insulin resistance. It is possible, however, that more prolonged ATX inhibition is necessary to influence PPAR-γ in IR cells, which should be tested in future studies. Despite the upregulation of PPAR-γ, at least in IS cells, inhibition of ATX activity did not influence insulin-stimulated Akt phosphorylation in IS cells or ameliorate impaired Akt phosphorylation in IR cells. These data suggest that autocrine/paracrine ATX-LPA signaling does not directly alter adipocyte insulin signal transduction. Given the apparent role of mitochondrial metabolism in insulin signaling (10)—particularly mitochondrial H₂O₂ emission (34, 35), we sought to address the potential effect of ATX inhibition on mitochondrial respiration and H₂O₂ emission in adipocytes. Examination of mitochondrial metabolism in IR cells incubated with ATX inhibitor for 24 hours suggested that ATX does not directly impact mitochondrial oxygen flow in the presence of glucose and fatty acid metabolites. However, ATX inhibition increased mitochondrial ROS production in response to all substrates examined. Since prior studies in nonadipocyte cells have shown that the ATX-LPA signaling pathway increases the expression of antioxidant genes and protects from oxidative stress (36, 37), it is possible that a reduction of antioxidant proteins in response to inhibition of ATX activity increases ROS in adipocytes.

Taken together, this study provides further insight into the regulation of ATX. Our data confirm that ATX is modulated by short-term and long-term changes in nutritional status. This study also shows that cellular signaling triggered by glucose and insulin plays an important role in regulating ATX secretion from adipocytes, which may underlie changes in circulating ATX levels during feeding/fasting and obesity-related metabolic disease. Moreover, our data suggest that autocrine/paracrine ATX-LPA signaling does not directly influence insulin signaling and mitochondrial respiration in adipocytes, but may influence ROS production.

Abbreviations:

 
• ATX

  autotaxin

 
• BSA

  bovine serum albumin

 
• cDNA

  complementary DNA

 
• DMEM

  Dulbecco’s modified Eagle medium

 
• DMEM-HG

  DMEM containing high glucose concentrations

 
• FAF

  fatty acid–free

 
• FBS

  fetal bovine serum

 
• FCCP

  carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone

 
• Glut4

  glucose transporter 4

 
• HFHS

  high-fat–high-sucrose

 
• IR

  insulin resistant

 
• IS

  insulin sensitive

 
• LPA

  lysophosphatidic acid

 
• LPC

  lysophosphatidylcholine

 
• mRNA

  messenger RNA

 
• NAFLD

  nonalcoholic fatty liver disease

 
• NF-κB

  nuclear factor-κB

 
• PBS

  phosphate-buffered saline

 
• PCR

  polymerase chain reaction

 
• PPAR-γ

  peroxisome proliferator–activated receptor-γ

 
• ROS

  reactive oxygen species

 
• SCAT

  subcutaneous adipose tissue

 
• TNF-α

  tumor necrosis factor-α.

Acknowledgments

Thanks to Mary Weir and Dr. Lester Perez for technical assistance.

This work was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (RGPIN-2014-04454) and grants from the Banting Research Foundation, the New Brunswick Health Research Foundation (NBHRF), and the New Brunswick Innovation Foundation (NBIF) to P.C.K. D.A.K. would like to acknowledge NSERC, the Nova Scotia Health Research Foundation (NSHRF), and Canada Foundation for Innovation (CFI). M.T. would like to acknowledge the contribution of NBIF, CFI, and Université de Moncton. T.P. would like to acknowledge NSERC, the Canadian Diabetes Association (CDA), NBHRF, NBIF, and CFI. K.D. is a Killam Scholar and NSERC Vanier Scholar.

Author contributions: K.D. and P.C.K. designed research; K.D. performed experiments; K.D. and P.C.K. analyzed and interpreted the data and wrote the manuscript; D.A.K. contributed to mitochondrial experiments and data analysis; M.T. synthesized ATX inhibitor; E.E.K. provided intellectual inputs; T.P. provided intellectual inputs and technical assistance. All authors edited the manuscript.

Disclosure Summary: The authors have nothing to disclose.

Author notes

Address all correspondence and requests for reprints to: Petra C. Kienesberger, PhD, Dalhousie Medicine New Brunswick, Department of Biochemistry and Molecular Biology, Dalhousie University, 100 Tucker Park Road, Saint John, New Brunswick E2L 4L5, Canada. E-mail: [email protected].