Fetal bovine serum (FBS), Dulbecco’s modified Eagle’s medium (DMEM), 1% (v/v) streptomycin/penicillin and TRIzol reagent were from Gibco Invitrogen, (Grand Island, NY, USA); the Superscript III First-strand Synthesis System was from Promega (Madison, WI, USA). Rapamycin (RAP) was purchased from Merck Bioscience (Darmstadt, Germany). PA (P0500), bovine serum albumin (BSA) (albumin, endotoxin), chloroquine (CQ), 4-phenyl butyrate (4-PBA), 3-methyladenine (3-MA), SP600125 and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were from Sigma-Aldrich (St. Louis, MO, USA). The PageRuler™ Prestained Protein Ladder was from Thermo Scientific (Kalamazoo, MI, USA). Complete protease inhibitor mixture and immunoblot polyvinylidene difluoride (PVDF membranes were from Roche Diagnostics (Barcelona, Spain). RIPA lysis buffer and the BCA protein assay kit were from Beijing ComWin Biotech Co., Ltd. (Beijing, China). Western Chemiluminescent HRP substrate was purchased from Millipore (Billerica, MA, USA). Rabbit anti-microtubule-associated protein 1 light chain 3 (LC3) 1:1,000 (no. 2775), rabbit anti-activating transcription factor 4 (ATF4) 1:1,000 (no. 11815), rabbit anti-C/EBP homologous protein (CHOP) 1:1,000 (no. 5554), rabbit anti-p-eukaryotic translation initiation factor 2α (eIF2α) 1:1,000 (no. 3398), rabbit anti-p-c-Jun N-terminal kinase (JNK) and total JNK 1:1,000 (no. 9912) were from Cell Signaling Technology (Beverly, MA, USA). Goat polyclonal anti-BiP 1:200 (sc-1050) and goat polyclonal anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) 1:200 (sc-48166) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA).

The 3T3-L1 cells were obtained from the American Type Culture Collection (ATCC; Rockville, MD, USA). The cells were seeded and fed every 2 days in DMEM containing 25 mM glucose supplemented with 50 U/ml penicillin, 50 μg/ml streptomycin, 100 mM MEM sodium pyruvate and 10% FBS. The cells were grown under 5% CO₂ at 37°C. At confluence, differentiation was induced by the addition of medium containing 0.5 mM isobutylmethylxanthine, 1 μM dexamethasone (Sigma-Aldrich) and 10 μg/ml insulin. After 48 h, this mixture was replaced with fresh medium, and this was changed every 2 days. On days 6–8 after the induction of adipocyte differentiation, the cells were used for the experiments. PA/BSA conjugates were prepared as previously described (14). PA was dissolved at 70°C in 0.1 M NaOH to obtain a 100 mM stock solution. A 5% (w/v) solution of free fatty acid (FFA)-free BSA was prepared in double distilled water. Subsequently, a 5 or 10 mM PA/BSA mixture was prepared by suitable combination of the 2 above-mentioned solutions. Finally, the mixture was further diluted in serum-free medium to obtain the required final concentrations of 0.5 mM PA/0.5% BSA or 1.0 mM PA/0.5% BSA.

After the indicated treatments, mature 3T3-L1 adipocytes were fixed in phosphate buffer (pH 7.4) containing 2.5% glutaraldehyde and 2% paraformaldehyde at room temperature for 60 min. The cells were post-fixed in 1% OsO₄ at room temperature for 60 min, dehydrated through graded ethanol solutions, and embedded in Quetol 812 (Nisshin EM Co., Tokyo, Japan). Areas containing cells were block-mounted and cut into 70 nm sections that were stained with uranyl acetate (saturated aqueous solution) and lead citrate and examined under a transmission electron microscope (H-7100; Hitachi, Ibaraki, Japan).

The cells were grown on round glass coverslips in 35 mm cell culture dishes. Following a 20-min fixation with pre-chilled methanol, the coverslips were washed with phosphate-buffered saline (PBS) and permeabilized with 0.2% Triton X-100-PBS for 30 min. The coverlips were then incubated with rabbit polycolonal LC3 primary antibodies (1:200) at 37°C for 90 min in the dark, followed by 3 washes (10 min each) in PBS. Subsequently, the coverlips were incubated with goat-anti rabbit IgG-FITC secondary antibodies (1:1,000) at 37°C for 1 h in dark and washed 3 times (10 min each time) in PBS. Digital images were obtained at an original magnification of x20 using a Nikon C1 confocal microscope, with NIS-Element software (Nikon, Melville, NY, USA).

The cells were washed with PBS and lysed in lysis buffer (0.5% Triton X-100, 10 mM HEPES pH 7.9, 50 mM NaCl, 100 mM EDTA, 0.5 M sucrose) containing 0.1% protease inhibitor cocktail (Roche Diagnostics). The lysates were then incubated on ice for 30 min and centrifuged at 8,000 × g for 10 min. Equal amounts of protein were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; 10–15%), transferred onto PVDF membranes. Molecular weights were estimated by comparison with a pre-stained protein ladder. Non-specific binding was blocked using 5% skim milk. The membranes were then incubated with specific primary antibodies overnight at 4°C. The membranes were washed with PBS-Tween-20 and incubated with peroxidase-conjugated secondary antibodies [anti-rabbit IgG, HRP-linked sntibody 1:5,000 (no. 7074); Cell Signaling Technology; and mouse anti-goat IgG, HRP-linked antibody 1:5,000 (no. sc-2354); Santa Cruz Biotechnology, Inc.] Protein bands were detected using the Western Chemiluminescent HRP substrate (Millipore). Immunoblots were quantified by densitometric analysis using ImageTool 3.0 software. The quantification of protein phosphorylation was normalized to the corresponding total protein expression, and the relative expression level of a certain protein was normalized to GAPDH.

Total RNA was extracted using TRIzol reagent (Invitrogen); samples of 1 μg total RNA were reverse transcribed into cDNA using a cDNA synthesis kit (Promega). Quantitative PCR (qPCR) was performed using an ABI 7900 sequencer (Life Technologies, Carlsbad, CA, USA) with the SYBR-Green PCR Master Mix (Promega) and d(N)₆ random hexamer with primers obtained from Invitrogen. The thermocycling parameters were as follows: 95°C for 10 min, 40 cycles of 95°C for 15 sec, 60°C for 1 min. Each sample was run in triplicate and normalized against 36B4 RNA. The fold changes were determined using the ΔΔCt method. The primers (mouse) used were as follows: IL-6 forward, 5′-CTGGGAAA TCGTGGAAATG-3′ and reverse, 5′-CCAGAGGAAATTTT CAATAGGC-3′; MCP-1 forward, 5′-AGCCAACTCTCACTG AAGCCA-3′ and reverse, 5′-AGTAGCAGCAGGTGAGT GGG-3′; and 36B4 forward, 5′-CGACCTGGAAGTCCAA CTAC-3′ and reverse, 5′-ATCTGCTGCATCTGCTTG-3′.

Cell viability was assessed by MTT assay. Briefly, mature adipocytes were seeded in a 96-well plate and incubated under different conditions. The culture medium was then removed, and 200 μl/well were diluted in MTT solution (0.5 mg/ml) in PBS for 1 h at 37°C. During the incubation time, MTT was converted to an insoluble formazan. At the end of the incubation time, the medium containing MTT was aspirated, and each well was gently washed with PBS. Dimethyl sulfoxide (DMSO) was then used to solubilize the precipitated formazan, and the concentration was determined using a microplate reader at OD 570 nm.

The results are expressed as the means ± SEM. Comparisons of a single variable in >2 groups were analyzed by one-way ANOVA followed by Tukey’s multiple comparison tests (GraphPad Prism). Statistical analysis was performed using the paired and unpaired t-test between 2 groups, using SPSS 12.0 software (SPSS, Inc., Chicago, IL, USA). Values of P<0.05 were considered to indicate statistically significant differences.

As the lipidation of LC3 and its association with autophagsome membranes has been established as a useful marker of autophagy (15), we detected LC3 expression by western blot analysis and fluorescence microscopy. A reliable marker of autophagy is the conversion of the ATG protein, LC3, from a soluble form (LC3-I) to a lipidized form (LC3-II), which stably associates with the membranes of autophagosomes (16). This conversion can be detected by measuring the accumulation of the LC3-II form. Our results demonstrated that treatment with PA increased LC3-II formation in response to treatment with PA (0.5 or 1.0 mM) for 12 h (Fig. 1A). RAP, a known mammalian target of rapamycin (mTOR) pathway inhibitor, is known to induce autophagy (17). RAP was used as a positive control. We observed significant LC3-II accumulation following treatment with RAP (100 nM) for 12 h (Fig. 1A). The number of LC3-II puncta was increased and the LC3-II puncta were clearly visible in the PA (0.5 mM)-treated group compared with the control group (cells treated with 0.5% BSA), shown as green fluorescent granules in the cytoplasm, mainly around the nucleus. The number of LC3-II puncta was also increased in the cells treated with RAP (Fig. 1B).

To gain insight into the morphological changes induced by PA, electron microscopy was performed using the mature 3T3-L1 adipocytes. Treatment with PA (0.5 mM) for 12 h induced the formation of autophagosomes, which were recognized at the ultrastructural level as double-membrane vacuolar structures containing visible cytoplasmic contents. Autolysosomes were recognized as single-membrane vacuolar structures containing high-density materials and some contained multivesicular body-like vesicles (Fig. 1C). These characteristics were also observed in the cells treated with RAP (100 nM) for 12 h. Abundant lipid droplets, but rare evidence of autophagosomes were observed in the control cells (Fig. 1C).

ER stress is a potential molecular mechanism of lipotoxicity, and we thus examined whether PA induces ER stress in mature adipocytes. We observed an increase in ER stress protein markers, evidenced by an increase in the expression of BiP, CHOP and eIF2α phosphoralytion, as well as in JNK phosphoralytion following treatment with PA (0.5 or 1.0 mM) for 12 h (Fig. 2). These results suggest the activation of the PKR-like endoplasmic reticulum kinase (PERK)-associated unfolded protein response (UPR) pathways, as well as inositol-requiring enzyme 1 (IRE1) pathways. Increased CHOP expression occurs downstream of the main pathways activated following ER stress, namely PERK, ATF6 and IRE1 (18). In addition, electron microscopy was performed on the PA-treated adipocytes (for 12 h) and revealed a dilated ER (Fig. 1C, red arrow).

We further addressed the cause-effect relationship among ER stress and autophagy. The phosphorylation of eIF2α and the ATF4 protein expression level were increased following treatment with PA (0.5 and 1.0 mM) for 6 h (Fig. 3). No obvious accumulation of LC3-II was observed following treatment with PA for 6 h, but only at 12 h of treatment. Based on this observation, we then investigated the association between ER stress and autophagy by treating the adipocytes with 4-PBA, an ER stress inhibitor. ER stress was ameliorated by treatment with 4-PBA, as evidenced by a decrease in the expression level of the ER stress marker, BiP (Fig. 4). Western blot analysis revealed that LC3-II accumulation was decreased in the presence of 4-PBA compared to treatment with PA alone (Fig. 4). These results indicate a potential causal involvement of ER stress in the induction of autophagy induced by PA.

To address the role of PA-induced autophagy in mature adipocytes, the cells were co-treated with CQ, a lysosome inhibitor, to block the autophagic flux, as previously described (19). The pharmacological inhibition of autophagy by CQ significantly increased the expression levels of the PA-induced ER stress markers, BiP and CHOP, indicating the extent of cellular stress when PA-induced autophagy was inhibited (Fig. 5A). Conversely, treatment with RAP to induce autophagy reversed these effects in response to PA (Fig. 5A). In addition, the expression levels of the ER stress markers were also increased by treatment with CQ without PA treatment, indicating that basal autophagy plays a role against cellular stress (Fig. 5A). The detrimental effects of the inhibition of PA-induced autophagy were confirmed using a cell viability assay. Cell viability significantly decreased by co-treatment with PA and CQ (Fig. 5B). Treatment with CQ alone also decreased cell viability, indicating that basal autophagy plays a role against cell death. These results demonstrate that the activation of autophagy is an adaptive response to PA and functions as a mechanism for cellular self-protection against ER stress and cell death.

Pre-loading of the cells with 0.5 mM/l of the saturated fatty acid, PA, for 12 h resulted in an increase in the mRNA expression levels of the inflammatory cytokines, IL-6 and MCP-1, in the mature adipocytes (Fig. 6A and C). In order to determine the role of autophagy in PA-induced inflammatory cytokine expression, we pre-treated the adipocytes with the autophagy inhibitor, 3-MA, for 1 h followed by treatment with PA for 12 h. 3-MA effectively inhibited the induction of autophagy by PA, as evidenced by the decrease in LC3-II accumulation (Fig. 6E). RT-qPCR revealed that treatment with 3-MA further increased the mRNA expression levels of MCP-1 and IL-6 compared to treatment with PA alone (Fig. 6A and C). Since the induction of autophagy by ER stress is in part mediated by JNK activation (20), we then examined the role of JNK in PA-induced autophagy in mature adipocytes using the specific JNK inhibitor, SP600125. As anticipated, the enhanced accumulation of LC3-II was decreased by pre-treatment with SP600125 compared to treatment with PA alone (Fig. 6F). Importantly, the expression of IL-6 and MCP-1 was further increased by SP600125 in the PA-treated adipocytes (Fig. 6B and D). These results demonstrate that PA-induced autophagy is partially JNK-dependent and that the activation of JNK-dependent autophagy plays a role in limiting inflammatory cytokine expression.