Biochem. J. (2011) 435, 723–732 (Printed in Great Britain) doi:10.1042/BJ20101680 723

Carnitine palmitoyltransferase 1A prevents fatty acid-induced adipocytedysfunction through suppression of c-Jun N-terminal kinaseXuefei GAO*1, Kuai LI*†1, Xiaoyan HUI‡, Xiangping KONG§, Gary SWEENEY‖, Yu WANG‡, Aimin XU‡, Maikun TENG†,Pentao LIU¶ and Donghai WU*2

*Key laboratory of Regenerative Biology, Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou, China, †School of Life Sciences, University ofScience and Technology of China, Hefei, China, ‡Department of Medicine, Pharmacology and Pharmacy, University of Hong Kong, Hong Kong, China, §458th Hospital, Guangzhou,China, ‖Department of Biology, York University, Toronto, Canada, and ¶Wellcome Trust Sanger Institute, Hinxton, Cambridge, U.K.

The adipocyte is the principal cell type for fat storage. CPT1(carnitine palmitoyltransferase-1) is the rate-limiting enzyme forfatty acid β-oxidation, but the physiological role of CPT1 inadipocytes remains unclear. In the present study, we focused on thespecific role of CPT1A in the normal functioning of adipocytes.Three 3T3-L1 adipocyte cell lines stably expressing hCPT1A(human CPT1A) cDNA, mouse CPT1A shRNA (short-hairpinRNA) or GFP (green fluorescent protein) were generated andthe biological functions of these cell lines were characterized.Alteration in CPT1 activity, either by ectopic overexpressionor pharmacological inhibition using etomoxir, did not affectadipocyte differentiation. However, overexpression of hCPT1Asignificantly reduced the content of intracellular NEFAs (non-esterified fatty acids) compared with the control cells whenadipocytes were challenged with fatty acids. The changeswere accompanied by an increase in fatty acid uptake and adecrease in fatty acid release. Interestingly, CPT1A protectedagainst fatty acid-induced insulin resistance and expression of

pro-inflammatory adipokines such as TNF-α (tumour necrosisfactor-α) and IL-6 (interleukin-6) in adipocytes. Further studiesdemonstrated that JNK (c-Jun N terminal kinase) activity wassubstantially suppressed upon CPT1A overexpression, whereasknockdown or pharmacological inhibition of CPT1 caused asignificant enhancement of JNK activity. The specific inhibitorof JNK SP600125 largely abolished the changes caused bythe shRNA- and etomoxir-mediated decrease in CPT1 activity.Moreover, C2C12 myocytes co-cultured with adipocytes pre-treated with fatty acids displayed altered insulin sensitivity. Takentogether, our findings have identified a favourable role for CPT1Ain adipocytes to attenuate fatty acid-evoked insulin resistance andinflammation via suppression of JNK.

Key words: adipocyte, carnitine palmitoyltransferase 1A(CPT1A), c-Jun N-terminal kinase (JNK), fatty acid, insulinresistance, pro-inflammatory adipokine.

INTRODUCTION

Adipocytes, traditionally thought to be an inert cell type for lipidstorage, are now recognized as an essential player orchestratingthe overall energy metabolism of the body [1]. Insulinsensitivity, adipokine production and inflammatory status inadipocytes are central to global energy homoeostasis. Adipocytedysfunction underlies the initiation and development of a panelof metabolic disorders, including insulin resistance, hypertension,atherosclerosis and stroke. It is now evident that pro-inflammatoryadipokines secreted from adipocytes systemically control glucoseand lipid metabolism in the body. In obese subjects, the expansionof adipose tissue leads to the secretion of excess amounts offatty acids and pro-inflammatory adipokines, such as TNF-α(tumour necrosis factor-α), A-FABP (adipocyte fatty acid-bindingprotein), MCP-1 (monocyte chemoattractant protein-1), PAI-1(plasminogen-activator inhibitor-1), lipocalin-2 and RBP4(retinol-binding protein 4), to name a few, but also causes adecrease in the production of adiponectin, an insulin-sensitizingadipokine with anti-diabetic, anti-inflammatory, and cardiopro-tective and vasculoprotective properties [2]. Additionally, insulinresistance in adipocytes manifests as a blunted GLUT4 (glucose

transporter 4) translocation in response to insulin, which isfollowed by the whole-body glucose intolerance [3].

How abnormal adipokine production is evoked in the context ofobesity and subsequently leads to insulin insensitivity still needsto be fully elucidated. Obesity begins with the overexpansionof adipose tissues both in cell number (hyperplasia) and size(hypertrophy), which is essentially a result of an imbalancebetween lipid storage and lipid utilization. Although it isgenerally believed that enlarged adipose tissue is predisposedto adipocyte dysfunction, the exact role of fatty acid oxidation inadipocyte functioning remains unclear.

CPT1 (carnitine palmitoyltransferase 1) is the key regulatoryenzyme in fatty acid oxidation. It is anchored in the outermembrane of mitochondria and catalyses the formation of long-chain acyl-carnitine, which is enabled to traverse the innermitochondrial membrane and thus committed to β-oxidation inthe mitochondria. There are three genes that code for CPT1:CPT1A, which is the most abundant form in liver, CPT1B, whichis the major form in muscle, and CPT1C, which is mainly presentin the brain. Deletion of CPT1A or CPT1B in mouse is lethal inthe early development [4,5], therefore the exact role of CPT1Aor CPT1B in energy homoeostasis remains unresolved. However,

Abbreviations used: A-FABP, adipocyte fatty acid-binding protein; C/EBPα, CCAAT/enhancer-binding protein α; CPT1, carnitine palmitoyltransferase 1;DMEM, Dulbecco’s modified Eagle’s medium; FBS, fetal bovine serum; GFP, green fluorescent protein; hCPT1A, human CPT1A; hCPT1A-adipocyte,adipocyte overexpressing hCPT1A; IL-6, interleukin 6; IRS1, insulin receptor substrate 1; JNK, c-Jun N-terminal kinase; KRH, Krebs–Ringer–Hepes; LCA-CoA, long-chain acyl-CoA; NEFA, non-esterified fatty acid; NF-κB, nuclear factor κB; PB, PiggyBac; PPARγ, peroxisome-proliferator-activated receptor γ;shRNA, short-hairpin RNA; shRNA-adipocyte, adipocyte with CPT1A knocked down by shRNA; TNF-α, tumour necrosis factor-α; WT, wild-type.

1 These authors contributed equally to this work.2 To whom correspondence should be addressed (email wu_donghai@gibh.ac.cn).

c© The Authors Journal compilation c© 2011 Biochemical Society

724 X. Gao and others

the essential role of CPT1A has been inferred by several studies.In patients with unusual spontaneous mutations in the CPT1Agene, hepatic deficiency of CPT1A leads to recurrent episodes ofhypoketotic hypoglycaemia, hepatomegaly, seizures and coma.In the pancreas, inhibition of fatty acid oxidation by a CPT1-specific inhibitor, etomoxir [6], or by non-metabolizable fattyacid analogues [7] changes GSIS (glucose-stimulated insulinsecretion). Additionally, blockade of CPT1 activity by its selectiveinhibitors protected hearts from fatty acid-induced ischaemicinjury [8–10]. Alteration of hypothalamic CPT1 activity bymolecular and pharmacological approaches significantly affectsendogenous glucose production [11,12]. In addition, studies byGao et al. [13] and Wolfgang et al. [14] have shown that mice withCpt1c deleted are more sensitive to high-fat diet-induced insulinresistance.

CPT1 is expressed in adipocytes [15–17], yet little is knownabout the functional role of CPT1 in this cell type, and whetherCPT1 serves as an active regulator for adipocyte function is yet tobe elucidated. Thus it is relevant and important to examine the roleof CPT1 in adipocytes, where fat storage and accumulation arebelieved to be the main fate for fatty acids. In the present study, weexamined the alterations in the normal functioning of adipocytesafter using strategies to provide gain or loss of CPT1 function.In addition, the underlying mechanisms for the observed effectswere also investigated. Our findings demonstrate that CPT1 is akey regulator in adipocyte inflammation and insulin signalling.

EXPERIMENTAL

Cell culture and treatment

Mouse 3T3-L1 pre-adipocytes were cultured and differentiatedinto mature adipocytes using a standard protocol. Briefly, at2 days post-confluence (day 0), cells were cultured in DMEM(Dulbecco’s modified Eagle’s medium) supplemented with10% FBS (fetal bovine serum) (Gibco), 0.5 mM 1-methyl-3-isobutylxanthine (Sigma), 1 μM dexamethasone (Sigma) and10 μg/l insulin (Sigma) for 48 h. For days 2–4, the medium wassupplemented only with 10 μg/ml insulin. The cells were thenswitched to DMEM containing 10% FBS until fully differentiatedat day 10. For fatty acid treatment, equal molar quantities ofpalmitate and oleate (Sigma) were dissolved in 95 % ethanol at60 ◦C and then mixed with pre-warmed fatty acid-free BSA (3:1molar ratio) to yield a fatty acid stock concentration of 5 mM.Fatty acid was added to the medium and incubated with cells for6 h. Etomoxir (Sigma) and SP600125 (Merck) were dissolved inDMSO and added to the culture medium as indicated.

Generation of stable cell lines for hCPT1A (human CPT1A)overexpression and endogenous CPT1A knockdown

cDNA encoding hCPT1A and GFP (green fluorescent protein)were cloned into a PB (PiggyBac) expression vector andnamed PB-CAG-hCPT1A-IRES-puro and PB-CAG-GFP-IRES-puro respectively.

For RNAi (RNA interference) experiments, annealedoligonucleotides were cloned into the BamHI/XhoI sites of thepRNAT-U6.2/Lenti vector (GenScript) and were designed totarget the coding sequence of mouse CPT1A from nucleotides290 to 318. The sequence of the oligonucleotides was: 5′-GGAT-CCCGACTCACGATGTTCTTCGTCTGGCTTGACATTGATA-TCCGTGTCAAGCCAGACGAAGAACATCGTGAGTTTTTT-TCCAACTCGAG-3′. Thereafter the 700-bp PCR fragmentcontaining the U6.2 promoter and CPT1A shRNA (short-hairpinRNA) construct were produced and subcloned into PB-CAG-

GFP-IRES-puro vector before the CAGG promoter and behind3′-LTR (long terminal repeat) to form PB-shRNA-CAG-GFP-IRES-puro.

To obtain stable cells, 3×106 3T3-L1 pre-adipocytes wereharvested and electroporated using an Amaxa Protocol T-030 with 2 μg of plasmid (PB-CAG-hCPT1A-IRES-puro, PB-shRNA-CAG-GFP-IRES-puro and PB-CAG-GFP-IRES-puro)mixed with 1 μg of PB-transposase vector plasmid respectively.After 24 h of culture, the selection medium containing 14 μg/mlpuromycin (Merck) was added for 2 weeks with the mediumexchanged every 2–3 days. Isolated cell clones were picked andexpanded for use in subsequent experiments.

Real-time PCR analysis

Total RNA was isolated from cells using the TRIzol® reagent(Invitrogen) and reverse-transcribed using the Superscript IIIreverse transcriptase kit (Invitrogen). For real-time PCR analysis,cDNA samples were used in a quantitative PCR in the presenceof fluorescent dye SYBR®Green (Bio-Rad). The following PCRconditions were applied: 5 min at 95 ◦C, and 40 rounds of 10 s at95 ◦C, 20 s at 60 ◦C and 1 s at 70 ◦C each. After each elongationstep, the reaction was quantified in a reading step and theproduct quality was tested by melting curve analysis. Relativeabundance of mRNA was calculated after normalization to 18SRNA. The sequences for the primers used in the present studyare shown in Supplementary Table S1 (at http://www.BiochemJ.org/bj/435/bj4350723add.htm).

Analysis of cellular glucose and fatty acid uptake

Following treatment with 0.5 mM fatty acid (palmitate/oleatemixture, two of the most abundant nutritional fatty acids) for6 h, adipocytes were glucose-starved for 30 min in KRH (Krebs–Ringer–Hepes) buffer [120 mM NaCl, 4.7 mM KCl, 2.2 mMCaCl2, 10 mM Hepes, 1.2 mM KH2PO4, 1.2 mM MgSO4 and0.1% BSA (pH 7.5)]. After incubation for 15 min withoutor with insulin (100 nM), the tracer [1,2-3H]deoxy-D-glucose(0.93 GBq/mmol; Amersham Pharmacia) was then added for15 min and glucose uptake was assayed in triplicate for eachcondition. Fatty acid uptake assays were initiated by incubatingthe cells for 20 min in KRH buffer containing 5.4 mM glucoseand [1–14C]oleic acid (2.04 GBq/mmol; Amersham Pharmacia)bound to fatty acid-free BSA. The cells were washed andcellular incorporated 3H and 14C radioactivity was determinedby liquid-scintillation counting. The abundance of radioactivitywas normalized to protein content.

Western blot and ELISA analyses

Isolated cell lysates were resolved by SDS/PAGE, electroblottedon to PVDF membranes (Millipore) and immunoblottedwith antibodies, including rabbit anti-Myc, rabbit anti-Akt,rabbit anti-(phospho-Akt), rabbit anti-JNK (c-Jun N-terminalkinase) and mouse anti-(phospho-JNK) (all from Cell SignalingTechnology). Bound antibodies were detected using ECL(enhanced chemiluminescence) reagents (GE Healthcare). TNF-α levels in the medium were determined using mouse-specificELISA kits (Pierce Endogen), according to the manufacturer’sprotocols.

Mitochondrial CPT1 activity assay

Mitochondria of 3T3-L1 cells were prepared using the proceduredescribed previously [18]. Mitochondrial protein concentrations

c© The Authors Journal compilation c© 2011 Biochemical Society

Carnitine palmitoyltransferase 1A in adipocytes 725

Figure 1 Generation of stable 3T3-L1 adipocytes

The 3T3-L1 pre-adipocytes were transfected with PB vectors and selected for stable cell lines as described in the Experimental section. The cell lines were then differentiated and analysed. (A) Celllysates from WT and Myc-tagged hCPT1A-adipoctes were subjected to Western blot analysis with an anti-Myc antibody. (B) Mitochondria were fractionated from WT adipocytes, hCPT1A-adipocytesand shRNA-adipocytes and the CPT1 activity was determined. (C) The lipid accumulation in indicated naıve adipocytes was determined quantitatively by Oil Red O staining. (D) Relative mRNAabundance of PPARγ , C/EBPα and A-FABP/aP2 were determined by real-time PCR. Results are representative of three independent experiments and are means +− S.D., n = 4. *P < 0.05 comparedwith WT adipocytes.

were determined using Bradford assay and CPT1 activity wasassayed as described in [13].

Analysis of lipolysis and cellular fatty acid content

After treatment with 0.5 mM fatty acid for 6 h, adipocytes werestarved in low-serum medium (0.1%) overnight and incubatedwith DMEM containing 450 μmol/l BSA and 25 mM Hepes.After incubation for 3 h, the culture medium was completelyremoved for NEFA (non-esterified fatty acid) and glycerol assays.For the cellular NEFA content assay, cellular lipids were extractedwith chloroform/methanol (2:1, v/v). The organic phase was driedand re-solubilized in a minimal volume of chloroform. NEFA andglycerol concentrations were determined using colorimetric kits(NEFA and Glycerol kits; Kexin). Protein concentrations wereused to normalize the sample values. For Oil Red O staining,differentiated adipocytes in 24-well plates were fixed in freshformalin solution and stained with Oil Red O dye (Amresco). Thelipid was quantified by extracting the dye with propan-2-ol anddetermining the absorbance at 520 nm.

Co-culture of differentiated 3T3-L1 adipocytes with C2C12myocytes

Mouse C2C12 cells were maintained at a subconfluent density inDMEM supplemented with 10% FBS. To induce differentiation,confluent cells grown on a transwell plate were shifted toa medium containing DMEM supplemented with 2% horseserum for 6 days. After in vitro differentiation of C2C12, themyocytes were washed twice with PBS and then incubated inDMEM containing 10% FBS. Differentiated adipocytes grownon individual membrane inserts were pre-treated with fatty acidor BSA for 6 h and subsequently transferred to the cultureplates containing the differentiated myocytes. This resulted inan assembly of the two cell types sharing the culture medium,but separated by the membrane of the insert. The distance fromthe bottom of the culture dish to the membrane was 0.9 mm.

Co-culture was conducted for 18 h and the glucose uptake assayand Western blot analysis of C2C12 myocytes were performed asdescribed above.

Statistical analysis

Comparisons were analysed by unpaired two-tailed Student’st test. All calculations were performed using SPSS version 13.0for Windows. A P value of < 0.05 was considered significant.

RESULTS

CPT1A expression does not influence differentiation of 3T3-L1adipocytes

In order to investigate the biological function of CPT1Ain adipocytes, we adopted a PB delivery system [19] todeliver hCPT1A cDNA (hCPT1A-adipocytes) and shRNAagainst endogenous mouse CPT1A (shRNA-adipocytes) intoproliferating 3T3-L1 pre-adipocytes and selected for stablytransduced cell lines with puromycin. Control cells were 3T3-L1 pre-adipocytes transfected with the PB vector inserted withGFP [WT (wild-type)], ruling out the possibility of the PBinsertion and/or antibiotic selection-mediated toxicity as thesource of the observed phenotypic variation. The results showedthat transfection of CPT1A and mouse shRNA effectively resultedin the overexpression or knocking-down of CPT1 in 3T3-L1pre-adipocytes respectively compared with control (Figure 1Aand Supplementary Figure S1A at http://www.BiochemJ.org/bj/435/bj4350723add.htm), as determined by both Westernblotting and quantitative real-time PCR. Consistently, CPT1activities mirrored the changes in CPT1A expression (Figure 1B).

Next, we investigated whether CPT1A regulated adipogenesisin 3T3-L1 cells. hCPT1A, shRNA and WT 3T3-L1 pre-adipocyteswere subjected to adipogenic induction medium and theiradipogenic differentiation was monitored. Neither overexpressionnor knockdown of CPT1A led to any obvious changes in

c© The Authors Journal compilation c© 2011 Biochemical Society

726 X. Gao and others

Figure 2 Etomoxir does not affect differentiation of 3T3-L1 pre-adipocytes

Oil Red O staining and real-time PCR were used to determined the lipid content (A) and relative mRNA abundance of adipogenic gene expression (B) in the indicated adipocytes. Results arerepresentative of three independent experiments and are means +− S.D. Eto, etomoxir.

lipid droplet accumulation (Figure 1C). Consistently, quantitativereal-time PCR analysis revealed that expression of some keyadipogenic transcription factors, including PPARγ (peroxisome-proliferator-activated receptor γ ), C/EBPα (CCAAT/enhancer-binding protein α) and their target gene A-FABP/aP2, were notinfluenced (Figure 1D).

Pharmacological inhibition of endogenous CPT1 activity does notinfluence differentiation of 3T3-L1 pre-adipocytes

To exclude the possibility that lipid accumulation in 3T3-L1 cellswas altered by the cloning procedures rather than the stablyexpressing constructs, we tested the effects of an irreversibleCPT1-specific inhibitor etomoxir [8] on cell differentiation andlipid accumulation in naıve 3T3-L1 pre-adipocytes. Similar toprevious reports [19a], etomoxir significantly inhibited CPT1activity in a dose-dependent manner, as determined by theamount of CO2 released from adipocytes (Supplementary FigureS1B). However, results for lipid accumulation (Figure 2A) andadipogenic marker gene expression (Figure 2B) indicated thatthe differentiation of 3T3-L1 pre-adipocyte was not affectedby etomoxir treatment. These results demonstrated that CPT1expression is not critically involved in adipocyte differentiationin 3T3-L1 adipocytes.

CPT1A alters lipid metabolism in adipocytes

Considering CPT1A is the key enzyme in fatty acid catabolism,we sought to investigate whether CPT1A interferes with lipidstorage and metabolism in adipocytes. To address this question,the release of glycerol and NEFAs, which are two major productsof lipolysis, was determined under both normal and fatty acid-stimulated conditions. There were no significant differences inglycerol release among WT adipocytes, hCPT1A-adipocytes andshRNA-adipocytes (Figure 3A), suggesting that lipid hydrolysisis not significantly affected by CPT1 activity under both basaland fatty acid-treatment conditions. NEFA release was increasedin WT adipocytes pre-treated with fatty acid. In contrast, therelease of NEFAs was largely unaltered in 3T3-L1 adipocytesoverexpressing hCPT1A, whereas knockdown of CPT1A causedan even higher level of NEFA release into cell culture medium(Figures 3A and 3B) in response to pre-treatment with fattyacid. Fatty acid uptake was also examined and we found thathCPT1A-adipocytes exhibited a much higher level of fattyacid uptake than did control adipocytes, whereas in shRNA-

adipocytes this was reduced significantly compared with theWT adipocytes (Figure 3C). In agreement with this observation,triacylglycerol content was enhanced in hCPT1A-adipocytes andattenuated in shRNA-adipocytes respectively, as compared withWT adipocytes. This difference was not observed under BSAtreatment conditions (results not shown). Taken together, theuncoupling between glycerol and NEFA release and varied fattyacid uptake in tight association with CPT1A expression levels inadipocytes suggested that CPT1A interferes with the intracellularlipid availability. To test this hypothesis, the intracellular lipidswere extracted and the amount of NEFAs was measured. Wefound that treatment of fatty acid increased the intracellular levelof NEFAs (Figure 3D). However, this effect was completelyabolished in hCPT1A-adipocytes. Conversely, knockdown ofCPT1A caused an even higher level of NEFAs inside the cell(Figure 3D), which is similarly observed in adipocytes treatedwith the CPT1-specific inhibitor (Supplementary Figures S1C–S1F), clearly indicating the relevance of CPT1 in the regulationof lipid flux in adipocytes.

CPT1A rescued fatty acid-induced insulin resistance in adipocytes

Fatty acids have a well established role as a principal mediator ofadipocyte dysfunction, including insulin resistance and activationof inflammatory responses. In line with our finding that CPT1Apromoted the depletion of intracellular fatty acid, it is highlypossible that overexpression of CPT1A in adipocytes mayrectify fatty acid-induced adipocyte dysfunction. To this end,we investigated the effect of altered CPT1 activity on adipocyteinsulin sensitivity. We evaluated basal and insulin-stimulatedglucose uptake without (BSA control) or with fatty acid treatment.

Under normal conditions, there were no significant differencesin both basal and insulin-stimulated glucose uptake and activationof Akt signalling among the three cell types used in the presentstudy (results not shown). Incubation with fatty acid attenuatedinsulin-induced glucose uptake in WT mature 3T3-L1 adipocytes(Figure 4A). Activation of Akt was also significantly decreasedby ∼30% (Figure 4B). However, overexpression of hCPT1Arendered cells resistant to the impairment of glucose uptakeand Akt phosphorylation by fatty acids (Figures 4A and 4B).In contrast, knockdown of endogenous CPT1A resulted in aneven more blunted response to insulin compared with WT cells.These results demonstrated that maintaining adequate CPT1Aactivity in adipocytes can ameliorate fatty acid-induced insulinresistance.

c© The Authors Journal compilation c© 2011 Biochemical Society

Carnitine palmitoyltransferase 1A in adipocytes 727

Figure 3 CPT1A affects lipid storage and transport in adipocytes

Stable adipocyte cell lines were pre-incubated with BSA or 0.5 mM fatty acid (FA) for 6 h and lipid metabolism was examined. Cellular efflux of glycerol (A) and NEFA (B), and fatty acid uptake (C)in these stable adipocytes were assayed. (D) Intracellular lipids were extracted and NEFA content was measured. Results are representative of three separate experiments and are means +− S.D., n =4. *P < 0.05 compared with WT adipocytes treated with fatty acid; †P < 0.05 compared with counterpart adipocytes treated with BSA.

Figure 4 hCPT1A-overexpressing adipocytes are more resistant to fatty acid-induced insulin resistance

(A) Serum- and glucose-starved 3T3-L1 adipocytes were treated with 0.5 mM fatty acid (FA) for 6 h prior to insulin stimulation. Basal and insulin-stimulated 2-deoxyglucose uptake was assayed asdescribed in the Experimental section. (B) Under insulin-stimulated conditions, phospho-Akt (p-AKT) and total Akt (t-AKT) in the WT adipocytes, hCPT1A-adipocytes and shRNA-adipocytes wereevaluated by Western blot analysis and quantitative analysis of the phospho-Akt/total Akt ratio was performed (lower panel). Results are representative of three independent experiments and aremeans +− S.D., n = 4–5. *P < 0.05 compared with WT adipocytes treated with fatty acid; †P < 0.05 compared with counterpart adipocytes under BSA treatment.

Pro-inflammatory cytokine expression and secretion aresuppressed upon CPT1A overexpression

We next examined whether manipulation of CPT1A activity inadipocytes led to alterations in inflammation in adipocytes. Toaddress this question, we quantified the expression levels of two

pro-inflammatory adipokines, TNF-α and IL-6, by quantitativereal-time PCR. As shown in Figure 5(A), fatty acid treatmentcaused a significant elevation in mRNA levels of both TNF-α andIL-6. The changes were largely abolished in hCPT1A-adipocytes.In contrast, shRNA-adipocytes were more sensitive to fatty acid-evoked up-regulation of these two pro-inflammatory adipokines.

c© The Authors Journal compilation c© 2011 Biochemical Society

728 X. Gao and others

Figure 5 CPT1A expression alters pro-inflammatory adipokine production

WT adipocytes, hCPT1A-adipocytes and shRNA-adipocytes were incubated with BSA or fatty acid (FA) for 6 h and harvested for analysis 12 h later. (A) The relative mRNA abundance of TNF-α andIL-6 were measured by real-time PCR. (B) TNF-α protein was measured in conditioned medium by ELISA. Results are representative of three separate experiments and are means +− S.D., n = 3–5.*P < 0.05 compared with WT adipocytes treated with fatty acid; †P < 0.05 compared with counterpart adipocytes treated with BSA.

Figure 6 CPT1A exerts anti-inflammatory and insulin-sensitizing effects through suppression of JNK

(A) Western blot analysis of phospho-JNK (p-JNK) and total JNK (t-JNK) in the three 3T3-L1 adipocyte cell lines (upper panel) and quantitative analysis of the phospho-JNK/total JNK ratio wasalso performed (lower panel). (B, C) WT and shRNA-adipocytes were incubated with or without SP600125 and insulin-stimulated glucose uptake (B) and the phospho-Akt/total Akt ratio (C) wereanalysed. (D) The mRNA level of TNF-α and IL-6 in the indicated adipocytes pre-treated with fatty acid (FA) and 5 μM SP600125 were measured by real-time PCR. Results are representative of threeindependent experiments and are means +− S.D., n = 4–5. *P < 0.05 compared with WT adipocytes treated with fatty acid; †P < 0.05 compared with counterpart adipocytes treated with BSA.

c© The Authors Journal compilation c© 2011 Biochemical Society

Carnitine palmitoyltransferase 1A in adipocytes 729

Figure 7 Pharmacological inhibition of CPT1 activity deteriorates fatty acid-induced insulin insensitivity and inflammation through activation of JNK

(A) Western blot analysis of phospho-JNK (p-JNK) and total JNK (t-JNK) in naıve adipocytes treated with DMSO and etomoxir (upper panel) in the absence or presence of fatty acid (FA) andquantitative analysis of the phospho-JNK/total JNK ratio were performed (lower panel). Under insulin stimulation, naıve adipocytes were treated with DMSO, etomoxir (Eto; 40 μM) or etomoxir(40 μM) plus SP600125 (5 μM) and glucose uptake was determined (B), and phosphorylation of Akt (C) was analysed by Western blotting (upper panel), with the phospho-Akt/total Akt ratiodetermined quantitatively (lower panel). (D) The mRNA levels of TNF-α and IL-6 in the indicated naıve adipocytes with different treatments were measured by real-time PCR. Results are representativeof three separate experiments and are means +− S.D., n = 3–4. *P < 0.05 compared with DMSO control adipocytes treated with fatty acid; †P < 0.05 compared with counterpart adipocytes treatedwith BSA.

The expression of TNF-α was also confirmed by ELISA(Figure 5B).

CPT1A exert its beneficial effects in adipocytes via suppressionof JNK

It is well established that fatty acids play a negative role ineliciting pro-inflammatory responses primarily via the activationof JNK and NF-κB (nuclear factor κB), which in turn induce the

expression of various pro-inflammatory cytokines. In addition,activation of JNK has been shown to block insulin signallingvia the phosphorylation of IRS1 (insulin receptor substrate 1)and then Akt [20]. We therefore investigated whether thesemechanisms participated in CPT1A-related insulin resistance andinflammation in adipocytes. Consistent with previous reports [21],we found that incubation with fatty acids led to a significantactivation of JNK, as assessed by the ratio of phosphorylationof JNK1/2 to total JNK1/2 (Figure 6A). More interestingly,this effect was attenuated considerably in hCPT1A-adipocytes

c© The Authors Journal compilation c© 2011 Biochemical Society

730 X. Gao and others

Figure 8 Up-regulation of CPT1A in adipocytes protects co-cultured C2C12 myocytes from fatty acid-induced insulin resistance

C2C12 myocytes were co-cultured with adipocytes (pre-treated with fatty acid or BSA for 6 h) for 24 h as indicated in the Experimental section before the cells were harvested for analysis. (A) Thebasal and insulin-induced glucose uptake in C2C12 myocytes was measured. (B) Western blot analysis of phospho-Akt (p-AKT) and total Akt (t-AKT) in co-cultured C2C12 myocytes (upper panel),and the results were quantified (lower panel). Results are representative of three separate experiments and are means +− S.D., n = 4. *P < 0.05 compared with myocytes co-cultured with WTadipocytes pre-treated with fatty acid; †P < 0.05 compared with counterpart myocytes co-cultured with adipocytes pre-treated with BSA.

and enhanced in shRNA-adipocytes respectively (Figure 6A).Furthermore, inhibition of JNK by its selective inhibitorSP600125 potently protected against CPT1A-deficiency-relatedinsulin resistance in shRNA-adipocytes. To be more specific,the impairment of insulin-induced glucose uptake (Figure 6B)and Akt signalling (Figure 6C) by CPT1A knockdown wascompletely reversed upon pre-incubation with 5 μM SP600125.In addition, inhibition of JNK activity in shRNA-adipocytesabolished the fatty acid-induced increase in TNF-α and IL-6gene expression (Figure 6D). On the other hand, NF-κB wasalso activated following fatty acid treatment, but there was nosignificant difference among the three adipocyte cell lines used(results not shown). Hence these results suggested that JNK playsa central role that exacerbated adipocyte dysfunction in responseto decreased CPT1A expression.

Pharmacological inhibition of endogenous CPT1 activityexacerbates fatty acid-induced insulin insensitivity andinflammation

To rule out the possibility that the cellular cloning procedureis responsible for the observed phenotypes in our clonedadipocyte lines, we treated the naıve 3T3-L1 adipocytes withthe CPT1-specific inhibitor etomoxir. As shown in Figure 7,etomoxir significantly enhanced the phosphorylation of JNK1/2(Figure 7A) and exacerbated the negative effects of fatty acidson insulin sensitivity (Figures 7B and 7C) and expression of pro-inflammatory cytokines (Figure 7D). Similarly, the JNK inhibitorSP600125 completely reversed the impairment in insulin-inducedglucose uptake (Figure 7B) and Akt signalling (Figure 7C). Again,the fatty acid-induced increase in TNF-α and IL-6 gene expressionin etomoxir-treated adipocytes was abolished by inhibiting JNKactivity with SP600125 (Figure 7D).

Adipocyte CPT1A modulates insulin sensitivity in co-culturedC2C12 myocytes

Adipose tissue has been well recognized as one of the largestendocrine organs in the body. We next investigated whetheralteration in CPT1A expression within adipocytes systemicallyaffected other tissues or cell types. Differentiated C2C12myocytes were co-cultured with the three adipocyte cell linesand the insulin signalling in the cultured myocytes was examined.Consistent with a previous report [22], myocytes co-cultured withadipocytes showed a decreased insulin sensitivity (results notshown). There was no significant difference in basal and insulin-stimulated glucose uptake (Figure 8A) and phosphorylationstatus of Akt (results not shown) among C2C12 myocytes co-cultured with the three adipocyte cell lines. However, whenthese adipocytes were pre-treated with fatty acid, the C2C12myocytes co-cultured with WT adipocytes displayed a modestreduction in insulin-stimulated glucose uptake (Figure 8A),whereas the effect was significantly alleviated in myocytes co-cultured with hCPT1A-adipocytes and aggravated in myocytesco-cultured with shRNA-adipocytes (Figure 8A). The insulin-stimulated phosphorylation of Akt in differentiated myocytesco-cultured with fatty acid-pre-treated adipocytes mirrored thechanges in glucose uptake (Figure 8B).

DISCUSSION

Adipocytes are no longer regarded as an inert site for fatstorage and we designed the present study on the plausiblehypothesis that fatty acid metabolism in adipocytes intricatelyregulates the function of the cell. Fatty acids, which are elevatedin obese individuals, have been firmly established as one ofthe major factors triggering insulin resistance in peripheralmetabolic tissues. Although the detailed mechanisms are still

c© The Authors Journal compilation c© 2011 Biochemical Society

Carnitine palmitoyltransferase 1A in adipocytes 731

Figure 9 Proposed model for the protective role of CPT1A in fatty acid-induced insulin resistance in adipocytes

TG, triacylglycerol.

unclear, mounting evidence suggests that fatty acid overloadcauses intracellular accumulation of fatty acid-derived metabolicproducts [23,24]. These derivatives activate the key moleculesin pro-inflammatory pathways, such as JNK, which blunt insulinsignalling via direct phosphorylation on the threonine residue inIRS1 and then Akt [20,25]. In the present study, we show that thekey enzyme in fatty acid oxidation, CPT1A, protected adipocytesfrom fatty acid-induced insulin resistance and inflammation viasuppression of JNK activity.

The accumulation of fatty acid derivatives, such as in the acyl-CoA form, are capable of activating JNK, and this mechanismis of great significance in the pathogenesis of fatty acid-induced insulin resistance and inflammation in adipocytes. Thisis supported by our present findings that the differences amongthree cell lines with various degrees of CPT1A activity weremore evident when the cells were challenged with fatty acid. Wehypothesize (Figure 9) that, under conditions of lipid overload,CPT1 catalyses the transesterification reaction between LCA-CoA (long-chain acyl-CoA) and acyl-carnitine esters, and thuschanges the availability of LCA-CoA inside the cell. In addition,it is also key substrate for the synthesis of several complex lipidderivatives which act as key signalling molecules in the activationof inflammation and insulin resistance. Notably, it has beenshown previously that knockdown of LCA-CoA synthase, a keyenzyme for LCA-CoA synthesis, remarkably attenuated insulinsignalling in 3T3-L1 adipocytes [26]. Furthermore, knockdownof LCA-CoA synthase in adipocytes leads to reduced fatty aciduptake, which is largely reminiscent of our present findings thatCPT1 overexpression caused an increase in fatty acid uptake.There are also reports showing that mitochondrial oxidationdysfunction in adipocytes has a close correlation with thestates of insulin resistance and was considered as a target fortreatment of diabetes [27,28]. CPT1 is a mitochondrial oxidativeenzyme, thus lowered CPT1 expression might contribute tothe pathogenesis of obesity-related inflammation and insulinresistance.

Our present co-culture study demonstrated that elevatedexpression of CPT1A in adipocytes also increased insulinsensitivity in the co-cultured muscle cells, as compared withmyocytes co-cultured with WT adipocytes. This paracrineeffect was possibly mediated by the pro-inflammatory cytokinesand fatty acids secreted from adipocytes. Indeed, the masterrole of adipocytes in systemic glucose and lipid metabolismhas been implicated by several adipocyte-specific knockoutmice models [29,30]. Adipocyte-specific knockout or transgenicanimal models for CPT1 would be especially informative toelucidate its biological actions in vivo.

In summary, our present study has uncovered a pro-tective role for CPT1 in fatty acid-evoked adipocytedysfunction. Pharmacological activation of CPT1 might representa promising strategy for the prevention and treatment of obesity-related metabolic diseases.

AUTHOR CONTRIBUTION

Xuefei Gao, Kuai Li and Xiaoyan Hui designed and performed the experiments. GarySweeney, Yu Wang and Aimin Xu contributed to writing the manuscript. XiangpingKong, Maikun Teng and Pentao Liu contributed to discussion about the manuscript.Donghai Wu designed the experiments, and wrote and edited the manuscript prior tosubmission.

FUNDING

This work was supported, in part, by the National Basic Research Program of China(973 Program) [grant numbers 2011CB504004, 2010CB945500], the National NaturalScience Foundation of China [grant numbers 31000353, 30970637], and the KnowledgeInnovation Program of the Chinese Academy of Sciences and Guangzhou Administrationof Science and Technology [grant number 2007Z2-E4021].

c© The Authors Journal compilation c© 2011 Biochemical Society

732 X. Gao and others

REFERENCES

1 Hajer, G. R., van Haeften, T. W. and Visseren, F. L. (2008) Adipose tissue dysfunction inobesity, diabetes, and vascular diseases. Eur. Heart J. 29, 2959–2971

2 Hotamisligil, G. S. (2006) Inflammation and metabolic disorders. Nature 444, 860–8673 Abel, E. D., Peroni, O., Kim, J. K., Kim, Y. B., Boss, O., Hadro, E., Minnemann, T.,

Shulman, G. I. and Kahn, B. B. (2001) Adipose-selective targeting of the GLUT4 geneimpairs insulin action in muscle and liver. Nature 409, 729–733

4 Ji, S., You, Y., Kerner, J., Hoppel, C. L., Schoeb, T. R., Chick, W. S., Hamm, D. A.,Sharer, J. D. and Wood, P. A. (2008) Homozygous carnitine palmitoyltransferase 1b(muscle isoform) deficiency is lethal in the mouse. Mol. Genet. Metab. 93, 314–322

5 Nyman, L. R., Cox, K. B., Hoppel, C. L., Kerner, J., Barnoski, B. L., Hamm, D. A., Tian, L.,Schoeb, T. R. and Wood, P. A. (2005) Homozygous carnitine palmitoyltransferase 1a (liverisoform) deficiency is lethal in the mouse. Mol. Genet. Metab. 86, 179–187

6 Chen, S., Ogawa, A., Ohneda, M., Unger, R. H., Foster, D. W. and McGarry, J. D. (1994)More direct evidence for a malonyl-CoA-carnitine palmitoyltransferase I interaction as akey event in pancreatic β-cell signaling. Diabetes 43, 878–883

7 Bliss, C. R. and Sharp, G. W. (1992) Glucose-induced insulin release in islets of youngrats: time-dependent potentiation and effects of 2-bromostearate. Am. J. Physiol. 263,E890–E896

8 Lopaschuk, G. D., Wall, S. R., Olley, P. M. and Davies, N. J. (1988) Etomoxir, acarnitine palmitoyltransferase I inhibitor, protects hearts from fatty acid-inducedischemic injury independent of changes in long chain acylcarnitine. Circ. Res. 63,1036–1043

9 Molaparast-Saless, F., Liedtke, A. J. and Nellis, S. H. (1987) Effects of the fatty acidblocking agents, oxfenicine and 4-bromocrotonic acid, on performance in aerobic andischemic myocardium. J. Mol. Cell. Cardiol. 19, 509–520

10 Paulson, D. J., Noonan, J. J., Ward, K. M., Stanley, H., Sherratt, A. and Shug, A. L. (1986)Effects of POCA on metabolism and function in the ischemic rat heart. Basic Res. Cardiol.81, 180–187

11 He, W., Lam, T. K., Obici, S. and Rossetti, L. (2006) Molecular disruption of hypothalamicnutrient sensing induces obesity. Nat. Neurosci. 9, 227–233

12 Obici, S., Feng, Z., Arduini, A., Conti, R. and Rossetti, L. (2003) Inhibition ofhypothalamic carnitine palmitoyltransferase-1 decreases food intake and glucoseproduction. Nat. Med. 9, 756–761

13 Gao, X. F., Chen, W., Kong, X. P., Xu, A. M., Wang, Z. G., Sweeney, G. and Wu, D. (2009)Enhanced susceptibility of Cpt1c knockout mice to glucose intolerance induced by ahigh-fat diet involves elevated hepatic gluconeogenesis and decreased skeletal muscleglucose uptake. Diabetologia 52, 912–920

14 Wolfgang, M. J., Kurama, T., Dai, Y., Suwa, A., Asaumi, M., Matsumoto, S., Cha, S. H.,Shimokawa, T. and Lane, M. D. (2006) The brain-specific carnitinepalmitoyltransferase-1c regulates energy homeostasis. Proc. Natl. Acad. Sci. U.S.A. 103,7282–7287

15 Brown, N. F., Hill, J. K., Esser, V., Kirkland, J. L., Corkey, B. E., Foster, D. W. and McGarry,J. D. (1997) Mouse white adipocytes and 3T3-L1 cells display an anomalous pattern ofcarnitine palmitoyltransferase (CPT) I isoform expression during differentiation:inter-tissue and inter-species expression of CPT I and CPT II enzymes. Biochem. J. 327,225–231

16 Doh, K. O., Kim, Y. W., Park, S. Y., Lee, S. K., Park, J. S. and Kim, J. Y. (2005) Interrelationbetween long-chain fatty acid oxidation rate and carnitine palmitoyltransferase 1 activitywith different isoforms in rat tissues. Life Sci. 77, 435–443

17 Esser, V., Brown, N. F., Cowan, A. T., Foster, D. W. and McGarry, J. D. (1996) Expressionof a cDNA isolated from rat brown adipose tissue and heart identifies the product as themuscle isoform of carnitine palmitoyltransferase I (M-CPT I): M-CPT I is the predominantCPT I isoform expressed in both white (epididymal) and brown adipocytes. J. Biol. Chem.271, 6972–6977

18 McGarry, J. D., Mills, S. E., Long, C. S. and Foster, D. W. (1983) Observations on theaffinity for carnitine, and malonyl-CoA sensitivity, of carnitine palmitoyltransferase I inanimal and human tissues: demonstration of the presence of malonyl-CoA in non-hepatictissues of the rat. Biochem. J. 214, 21–28

19 Wang, W., Lin, C., Lu, D., Ning, Z., Cox, T., Melvin, D., Wang, X., Bradley, A. and Liu, P.(2008) Chromosomal transposition of PiggyBac in mouse embryonic stem cells. Proc.Natl. Acad. Sci. U.S.A. 105, 9290–9295

19a Wensaas, A. J., Rustan, A. C., Lovstedt, K., Kull, B., Wikstrom, S., Drevon, C. A. andHallen, S. (2007) Cell-based multiwell assays for the detection of substrate accumulationand oxidation. J. Lipid Res. 48, 961–967

20 Solinas, G., Naugler, W., Galimi, F., Lee, M. S. and Karin, M. (2006) Saturated fatty acidsinhibit induction of insulin gene transcription by JNK-mediated phosphorylation ofinsulin-receptor substrates. Proc. Natl. Acad. Sci. U.S.A. 103, 16454–16459

21 Nguyen, M. T., Satoh, H., Favelyukis, S., Babendure, J. L., Imamura, T., Sbodio, J. I.,Zalevsky, J., Dahiyat, B. I., Chi, N. W. and Olefsky, J. M. (2005) JNK and tumor necrosisfactor-α mediate free fatty acid-induced insulin resistance in 3T3-L1 adipocytes. J. Biol.Chem. 280, 35361–35371

22 Dietze, D., Koenen, M., Rohrig, K., Horikoshi, H., Hauner, H. and Eckel, J. (2002)Impairment of insulin signaling in human skeletal muscle cells by co-culture with humanadipocytes. Diabetes 51, 2369–2376

23 Petersen, K. F., Dufour, S., Befroy, D., Garcia, R. and Shulman, G. I. (2004) Impairedmitochondrial activity in the insulin-resistant offspring of patients with type 2 diabetes. N.Engl. J. Med. 350, 664–671

24 Chavez, J. A. and Summers, S. A. (2003) Characterizing the effects of saturated fatty acidson insulin signaling and ceramide and diacylglycerol accumulation in 3T3-L1 adipocytesand C2C12 myotubes. Arch. Biochem. Biophys. 419, 101–109

25 Hirosumi, J., Tuncman, G., Chang, L., Gorgun, C. Z., Uysal, K. T., Maeda, K., Karin, M.and Hotamisligil, G. S. (2002) A central role for JNK in obesity and insulin resistance.Nature 420, 333–336

26 Lobo, S., Wiczer, B. M. and Bernlohr, D. A. (2009) Functional analysis of long-chainacyl-CoA synthetase 1 in 3T3-L1 adipocytes. J. Biol. Chem. 284,18347–18356

27 Choo, H. J., Kim, J. H., Kwon, O. B., Lee, C. S., Mun, J. Y., Han, S. S., Yoon, Y. S.,Yoon, G., Choi, K. M. and Ko, Y. G. (2006) Mitochondria are impaired in the adipocytes oftype 2 diabetic mice. Diabetologia 49, 784–791

28 Maassen, J. A., Romijn, J. A. and Heine, R. J. (2007) Fatty acid-induced mitochondrialuncoupling in adipocytes as a key protective factor against insulin resistance and β celldysfunction: a new concept in the pathogenesis of obesity-associated type 2 diabetesmellitus. Diabetologia 50, 2036–2041

29 He, W., Barak, Y., Hevener, A., Olson, P., Liao, D., Le, J., Nelson, M., Ong, E., Olefsky,J. M. and Evans, R. M. (2003) Adipose-specific peroxisome proliferator-activatedreceptor γ knockout causes insulin resistance in fat and liver but not in muscle. Proc.Natl. Acad. Sci. U.S.A. 100, 15712–15717

30 Polak, P., Cybulski, N., Feige, J. N., Auwerx, J., Ruegg, M. A. and Hall, M. N. (2008)Adipose-specific knockout of raptor results in lean mice with enhanced mitochondrialrespiration. Cell Metab. 8, 399–410

Received 8 November 2010/11 February 2011; accepted 24 February 2011Published as BJ Immediate Publication 24 February 2011, doi:10.1042/BJ20101680

c© The Authors Journal compilation c© 2011 Biochemical Society

Biochem. J. (2011) 435, 723–732 (Printed in Great Britain) doi:10.1042/BJ20101680

SUPPLEMENTARY ONLINE DATACarnitine palmitoyltransferase 1A prevents fatty acid-induced adipocytedysfunction through suppression of c-Jun N-terminal kinaseXuefei GAO*1, Kuai LI*†1, Xiaoyan HUI‡, Xiangping KONG§, Gary SWEENEY‖, Yu WANG‡, Aimin XU‡, Maikun TENG†,Pentao LIU¶ and Donghai WU*2

*Key laboratory of Regenerative Biology, Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou, China, †School of Life Sciences, University ofScience and Technology of China, Hefei, China, ‡Department of Medicine, Pharmacology and Pharmacy, University of Hong Kong, Hong Kong, China, §458th Hospital, Guangzhou,China, ‖Department of Biology, York University, Toronto, Canada, and ¶Wellcome Trust Sanger Institute, Hinxton, Cambridge, U.K.

Figure S1 Pharmacological inhibition of CPT1 affects lipid storage and transport in adipocytes

(A) The relative level of endogenous CPT1A mRNA was assessed in shRNA-adipocytes and WT adipocytes by real-time PCR. (B) The irreversible CPT1-specific inhibitor etomoxir inhibited theoxidation of [14C]oleic acid in 3T3-L1 adipocytes. (C–E) Cellular efflux of NEFAs (C) and glycerol (D), and fatty acid uptake (E) in the indicated adipocytes were assayed. (F) Intracellular lipids wereextracted and NEFA contents were measured. Results are representative of three separate experiments and are means +− S.D. (n = 3–4). *P < 0.05 compared with WT adipocytes or naıve adipocytestreated with DMSO and fatty acid; †P < 0.05 compared with counterpart adipocytes treated with BSA.

1 These authors contributed equally to this work.2 To whom correspondence should be addressed (email wu_donghai@gibh.ac.cn).

c© The Authors Journal compilation c© 2011 Biochemical Society

X. Gao and others

Table S1 Primer pairs used in RT–PCR

Name GenBank® accession number Forward primer (5′→3′) Reverse primer (5′→3′)

CPT1A NM_013495 TTGGGCCGGTTGCTGAT GTCTCAGGGCTAGAGAACTTGGAA18S RNA XR_034450 GTAACCCGTTGAACCCCATT CCATCCAATCGGTAGTAGCGTTGPPARγ NM_001127330 CGCTGATGCACTGCCTATGA TGCGAGTGGTCTTCCATCACFABP4 NM_024406.2 AAGGTGAAGAGCATCATAACCCT TCACGCCTTTCATAACACATTCCCEBPα NM_007678.3 GTGTGCACGTCTATGCTAAACCA GCCGTTAGTGAAGAGTCTCAGTTTGTNF-α NM_013693.2 AGGGATGAGAAGTTCCCAAATG CCTCCACTTGGTGGTTTGCTAIL-6 NM_031168.1 ACCACGGCCTTCCCTACTTC CAGAATTGCCATTGCACAACTC

Received 8 November 2010/11 February 2011; accepted 24 February 2011Published as BJ Immediate Publication 24 February 2011, doi:10.1042/BJ20101680

c© The Authors Journal compilation c© 2011 Biochemical Society