Lipid overload in obesity and type 2 diabetes is associated with adipocyte dysfunction, inflammation, macrophage infiltration, and decreased fatty acid oxidation (FAO). Here, we report that the expression of carnitine palmitoyltransferase 1A (CPT1A), the rate-limiting enzyme in mitochondrial FAO, is higher in human adipose tissue macrophages than in adipocytes and that it is differentially expressed in visceral vs. subcutaneous adipose tissue in both an obese and a type 2 diabetes cohort. These observations led us to further investigate the potential role of CPT1A in adipocytes and macrophages. We expressed CPT1AM, a permanently active mutant form of CPT1A, in 3T3-L1 CARΔ1 adipocytes and RAW 264.7 macrophages through adenoviral infection. Enhanced FAO in palmitate-incubated adipocytes and macrophages reduced triglyceride content and inflammation, improved insulin sensitivity in adipocytes, and reduced endoplasmic reticulum stress and ROS damage in macrophages. We conclude that increasing FAO in adipocytes and macrophages improves palmitate-induced derangements. This indicates that enhancing FAO in metabolically relevant cells such as adipocytes and macrophages may be a promising strategy for the treatment of chronic inflammatory pathologies such as obesity and type 2 diabetes.
- type 2 diabetes
- fatty acid oxidation
obesity has reached epidemic proportions worldwide, leading to severe associated pathologies such as insulin resistance, type 2 diabetes (T2D), cardiovascular disease, Alzheimer's disease, hypertension, hypercholesterolemia, hypertriglyceridemia, nonalcoholic fatty liver disease, arthritis, asthma, and certain forms of cancer (12).
Over the last two decades, adipose tissue has gained crucial importance in the mechanisms involved in obesity-related disorders. The energy-storing white adipose tissue (WAT) is well vascularized and contains adipocytes, connective tissue, and numerous immune cells such as macrophages, T and B cells, mast cells, and neutrophils that infiltrate and increase their presence during obesity (22). Macrophages were the first immune cells reported to participate in obesity-induced insulin resistance (56). This highlights their pathological role in adipose tissue in addition to their traditional involvement in tissue repair and in response to dead and dying adipocytes (5, 14). Fat is an active endocrine tissue that secretes hormones such as leptin, adiponectin, or resistin and inflammatory cytokines such as TNF-α, IL-6, IL-1β, etc. in response to several stimuli. It is therefore a complex organ controlling energy expenditure, appetite, insulin sensitivity, endocrine and reproductive functions, inflammation, and immunity (53).
The pathophysiology of obesity-induced insulin resistance has been attributed to ectopic fat deposition (39), increased inflammation, endoplasmic reticulum (ER) stress (16, 42), adipose tissue hypoxia (15) and mitochondrial dysfunction (32), and impaired adipocyte expansion and angiogenesis (50, 51, 54). In obesity, fatty acids (FA) together with other stimuli such as ceramide, various PKC isoforms, proinflammatory cytokines and reactive oxygen species (ROS), and ER stresses activate JNK, NF-κB, RAGE, and TLR pathways both in adipocytes and macrophages triggering inflammation and insulin resistance (43).
Strenuous efforts are being made by the research community to elucidate the mechanisms involved in the pathophysiology of obesity-related disorders. However, an alternative strategy could be to act upstream by preventing the accumulation of lipids and the progression of obesity. In addition to reducing caloric intake, a potential effective approach to combating obesity would be to increase energy expenditure in key metabolic organs such as adipose tissue. Obese individuals and those with T2D are known to have lower fatty acid oxidation (FAO) rates and lower electron transport chain activity in muscle (17, 19, 37), together with higher glycolytic capacities and enhanced cellular FA uptake compared with nonobese and nondiabetic individuals (44). Thus, any strategy able to eliminate the excess of lipids found in obesity could be beneficial for health. Lipid levels can be reduced by inhibiting synthesis and transport or by increasing oxidation; here, we focus on the latter.
Malonyl-CoA, derived from glucose metabolism and the first intermediate in lipogenesis, regulates FAO by inhibiting carnitine palmitoyltransferase 1 (CPT1). This makes CPT1 the rate-limiting step in mitochondrial FA β-oxidation. Thus, in high-energy conditions, malonyl-CoA inhibits oxidation, diverting FAs' fate into TG accumulation. There are three CPT1 isoforms, with different tissue expressions: CPT1A (liver, kidney, intestine, pancreas, ovary, and mouse and human WAT), CPT1B (brown adipose tissue, skeletal muscle, heart, and rat and human WAT), and CPT1C (brain and testis) (2, 36). The fact that CPT1 controls FAO makes it a very attractive target to reduce lipid levels and fight against obesity and T2D. Despite their excess fat, obese individuals have reduced visceral WAT CPT1 mRNA and protein levels (20). This prompted our group and others to overexpress CPT1 in liver (26, 29, 45), muscle (3, 33, 40), and white adipocytes (9), which led to a decrease in triglyceride (TG) content and an improvement in insulin sensitivity.
Here, we show that CPT1A expression was higher in human adipose tissue macrophages than in mature adipocytes and that it was differentially expressed in visceral vs. subcutaneous adipose tissue. To further investigate the role of CPT1A in both adipocytes and macrophages, we used a permanently active mutant form of CPT1A, CPT1AM, which is insensitive to its inhibitor malonyl-CoA (27), to achieve continuous oxidation of lipids. When cells were incubated with palmitate to mimic obesity, CPT1AM restored most of the palmitate-induced imbalances. An increase in FAO in adipocytes and macrophages reduced TG content and inflammatory levels, improved insulin sensitivity in adipocytes, and reduced ER stress and ROS damage in macrophages.
MATERIALS AND METHODS
Human cohorts: selection of patients.
Adipose tissue was selected from an adipose tissue biobank collection of the University Hospital Joan XXII (Tarragona, Spain). All subjects were of Caucasian origin and reported that their body weight had been stable for at least 3 mo before the study. They had no systemic disease other than obesity or T2D, and all had been free of any infections in the previous month before the study. Liver and renal diseases were specifically excluded by biochemical work-up. Appropriate Institutional Review Board approval and adequate biobank informed consent were obtained from all participants. Biobanking samples included plasma and total and fractionated adipose tissue from subcutaneous and visceral origin. All patients had fasted overnight before collection of blood and adipose tissue samples. Visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) samples were obtained during surgical procedures that included laparoscopic surgery for hiatus hernia repair or cholecystectomy. Samples were selected according stratification by age, sex, and BMI and grouped into two cohorts:
Subjects were classified by BMI according to the World Health Organization criteria (WHO, 2000). The study included 19 lean, 28 overweight, and 15 obese nondiabetic subjects matched for age and sex (Table 1).
Patients were classified as having T2D according to American Diabetes Association criteria (1997). Variability in metabolic control was assessed by stable glycated hemoglobin A1c (HbA1c) values during the previous 6 mo. These criteria having been gathered, there were 11 T2D subjects. As a control group, we selected 36 subjects without diabetes from the obesity cohort, matched for age, BMI, and sex (Table 2). No patient was being treated with thiazolidinedione.
Height was measured to the nearest 0.5 cm and body weight to the nearest 0.1 kg. BMI was calculated as weight (kilograms) divided by height (meters) squared. Waist circumference was measured midway between the lowest rib margin and the iliac crest.
Collection and processing of human samples.
Samples from VAT (omental) and SAT (anterior abdominal wall) from the same individual were obtained during abdominal elective surgical procedures (cholecystectomy or surgery for abdominal hernia). All patients had fasted overnight at least 12 h before the surgical procedure. Blood samples were collected before the surgical procedure from the antecubital vein, 20 ml of blood with EDTA (1 mg/ml), and 10 ml of blood in silicone tubes. Collected blood (15 ml) was used for the separation of plasma. Plasma samples were stored at −80°C until analytic measurements were performed; 5 ml of blood with EDTA was used for the determination of HbA1c. Adipose tissue samples were collected, washed in PBS, immediately frozen in liquid N2 and stored at −80°C.
Adipose tissue fractionation.
Adipose tissue biopsies were processed immediately. The adipose tissue was finely diced into small pieces (10–30 mg), washed in PBS, and incubated in Medium 199 (Life Technologies) supplemented with 4% BSA plus 2 mg/ml collagenase type I (Sigma) for 1 h in a shaking water bath at 37°C. After digestion, mature adipocytes (ADI) were separated from tissue matrix by filtration through a 200-μm mesh fabric (Spectrum Laboratories). The filtrated solution was centrifuged for 5 min at 1,500 g. The mature adipocytes were removed from the top layer and the stromal vascular fraction (SVF) cells remained in the pellet. Cells were washed four times in PBS and processed for RNA and protein extraction.
Glucose, cholesterol, and TG plasma levels were determined in an autoanalyzer (Hitachi 737, Boehringer Mannheim) using the standard enzyme methods. High-density lipoprotein (HDL) cholesterol was quantified after precipitation with polyethylene glycol at room temperature (PEG-6000). Plasma insulin was determined by radioimmunoassay (Coat-A-Count insulin; Diagnostic Products). Nonesterified free fatty acid (NEFA) serum levels were determined in an autoanalyzer (Advia 1200, Siemens) using an enzymatic method developed by Wako Chemicals. Plasma glycerol levels were analyzed by using a free glycerol determination kit, a quantitative enzymatic determination assay (Sigma-Aldrich). Intra- and interassay CVs were less than 6% and less than 9.1%, respectively. The degree of insulin resistance was determined by homeostasis model assessment (HOMA), as [glucose (mmol/l) × insulin (mIU/l)]/22.5. (24).
Five-micrometer sections of formalin-fixed paraffin-embedded adipose tissue were deparaffinized and rehydrated prior to antigen unmasking by boiling in 1 mM EDTA, pH 8. Sections were blocked in normal serum and incubated overnight with rabbit anti-CPT1A (Sigma-Aldrich) at 1:50 dilution. Secondary antibody staining was performed using a VECTASTAIN ABC kit (Vector Laboratories) and detected with diaminobenzidine (Vector Laboratories). Sections were counterstained with hematoxylin prior to dehydration and coverslip placement and examined under a Nikon Eclipse 90i microscope. As a negative control, the procedure was performed in the absence of primary antibody.
Five-micrometer sections of formalin-fixed paraffin-embedded adipose tissue were blocked in normal serum and incubated overnight with rabbit anti-CPT1A antibody (Sigma-Aldrich) at 1:50 dilution and with mouse anti-CD68 (Santa Cruz Biotechnology) at 1:50 dilution, washed, and visualized using Alexa fluor 546 goat anti-rabbit, and Alexa fluor 488 goat anti-mouse antibodies, respectively (1:500, Molecular Probes). The slides were counterstained with DAPI (4,6-diamidino-2-phenylindole) to reveal nuclei and were examined under a Nikon Eclipse 90i fluorescent microscope. As a negative control, the assay was performed in the absence of primary antibody.
Sodium palmitate, sodium oleate, BSA, and l-carnitine hydrochloride were purchased from Sigma Aldrich. DMEM, FBS, and penicillin-streptomycin mixture were purchased from Life Technologies.
Murine 3T3-L1 CARΔ1 preadipocytes, kindly given by Dr. Orlicky (Department of Pathology, UCHSC at Fitzsimons, Aurora, CO), were cultured and differentiated into mature adipocytes following the published protocol (31). Mature adipocytes were used for experiments at day 8 postdifferentiation. Murine RAW 264.7 macrophages were obtained from ATCC and were maintained in DMEM supplemented with 10% heat-inactivated FBS and 1% penicillin-streptomycin mixture. Simpson-Golabi-Behmel Syndrome (SGBS) human cells were cultured and differentiated to adipocytes as previously described (55).
At day 8 of differentiation, 3T3-L1 CARΔ1 cells were infected with adenoviruses AdGFP [100 moi (multiplicity of infection)] and AdCPT1AM (13) (100 moi) for 24 h in serum-free DMEM, and then the medium was replaced with complete DMEM for additional 24 h. RAW 264.7 macrophages were infected with AdGFP (100 moi) and AdCPT1AM (100 moi) for 2 h in serum-free DMEM and then replaced with complete medium for an additional 72 h. Adenovirus infection efficiency was assessed in AdGFP-infected cells (see Fig. 3, A and B). The same batch of adenoviruses stored in 50-μl aliquots was used throughout the experiments.
Sodium palmitate was conjugated with FA-free BSA in a 5:1 ratio to yield a stock solution of 2.5 mM (40). Cells were incubated with 0.3 or 1 mM of this solution for 24 h (3T3-L1 CARΔ1 adipocytes) or 0.3, 0.5, or 0.75 mM for 24, 18, or 8 h (RAW 264.7 macrophages), respectively.
Adipocyte and macrophage viability.
3T3-L1 CARΔ1 adipocytes and RAW 264.7 macrophages were infected as previously described and incubated for 24 h with 1 or 0.3 mM palmitate, respectively. Cells were washed twice with PBS and lifted from the surface with trypsin followed by 2 min of incubation at 37°C. Trypsinization was stopped with 10% FBS-containing medium, and equal volumes of cell suspension were mixed with 0.4% Trypan blue staining. Trypan blue-positive and -negative cells were counted using a Neubauer chamber for adipocytes and a Countess Automated Cell Counter (Invitrogen) for macrophages. Percentage of viability was determined normalizing viable cells of each group to viable cells of BSA GFP group. Statistical significance was assessed using two-way ANOVA of three individual experiments (*P < 0.05).
Mitochondria-enriched fractions were obtained from cell culture grown in 10-cm2 dishes, and CPT1 activity was measured by a radiometric method as described (13).
Total oleate oxidation was measured in 3T3-L1 CARΔ1 adipocytes and RAW 264.7 macrophages grown in 25-ml flasks, differentiated, and infected as described above. The day of the assay, cells were washed in KRBH-0.1% BSA, preincubated at 37°C for 30 min in KRBH-1% BSA, and washed again in KRBH-0.1% BSA. Cells were then incubated for 3 h (3T3-L1 CARΔ1 adipocytes) or 2 h (RAW 264.7 macrophages) at 37°C with fresh KRBH containing 11 mM glucose, 0.8 mM carnitine, and 0.2 mM [1-14C]oleate (PerkinElmer). Oxidation was measured as described (29). The scintillation values were normalized to the protein content of each flask.
Cells were grown in 12-well plates, differentiated, and infected as described above. After 24 h (3T3-L1 CARΔ1 adipocytes) or 18 h (RAW 264.7 macrophages) of FA treatment, cells were collected for lipid extraction following a protocol of Gesta et al. (10) with minor modifications: after cell lysis with 0.1% SDS, 1/2/0.12 (vol/vol/vol) methanol-chloroform-0.5 M KCl solution was added, the two phases were separated by centrifugation, and the upper phase was dried with N2. Finally, lipids were resuspended in 100% isopropanol, and TGs were quantified using a TG Ttriglyceride kit (Sigma) according to the manufacturer's instructions. Protein concentrations were used to normalize sample values.
Oil red O staining.
RAW 264.7 macrophages grown on coverslips were infected as described above and incubated with 0.75 mM palmitate for 18 h. After this time, cells were rinsed twice with PBS, fixed in 10% paraformaldehyde for 30 min at room temperature, and washed again with PBS. Then, cells were rinsed with 60% isopropanol for 5 min to facilitate the staining of neutral lipids and were stained with filtered Oil red O working solution (0.3% Oil red O in isopropanol) for 15 min. After several washes with distilled water, the coverslips were removed and mounted on a drop of mount medium. The intracellular lipid vesicles stained with Oil red O were identified by their bright red color under the microscope.
Analysis of intracellular protein oxidation.
RAW 264.7 macrophages were cultured in 12-well plates and infected as described before. After FA treatment, cell extracts were prepared and analyzed for protein oxidative modifications (i.e., carbonyl group content) with a OxyBlot Protein Oxidation Detection kit (Millipore), following the manufacturer's instructions.
Western blot analysis.
3T3-L1 CARΔ1 adipocytes and RAW 264.7 macrophages were cultured in 12-well plates, differentiated, and infected as described above. Cells were collected in lysis buffer (RIPA), and protein concentration was determined using a BCA protein assay kit (Thermoscientific). An equal amount of protein from whole cell lysates was resolved by 8% SDS-PAGE and transferred to PVDF membranes (Millipore). Signal detection was carried out with the ECL immunoblotting detection system (GE Healthcare), and the results were quantitatively analyzed using Image Quant LAS4000 Mini (GE Healthcare). The following antibodies were used: CPT1A [1/6,000 (13)], β-actin (I-19; 1/4,000, Santa Cruz), Akt and pAkt (Ser473; 1/1,000, Cell Signaling), C/EBP homologous protein (CHOP; GADD 153, 1/200; Santa Cruz) and insulin receptor-β (IRβ; 1/1,000; Santa Cruz). Human fat tissue was homogenized in RIPA buffer as previously described (34). Protein extracts (10–20 μg) were loaded, resolved on 10% SDS-PAGE, and transferred to Hybond ECL nitrocellulose membranes. Membranes were stained with 0.15% Ponceau red (Sigma-Aldrich) to ensure equal loading after transfer and then blocked with 5% (wt/vol) BSA in TBS buffer with 0.1% Tween 20. Immunoblotting was performed with 1:2,000 goat anti-human CPT1A (Abcam). Blots were incubated with the appropriate IgG-HRP-conjugated secondary antibody. Immunoreactive bands were visualized with an ECL-plus reagent kit (GE Healthcare). Optical densities of the immunoreactive bands were measured using Image J analysis software.
Analysis of mRNA expression by quantitative real-time PCR.
Total RNA was extracted from cultured cells grown in 12-well plates using Illustra Mini RNA Spin kit (GE Healthcare), and cDNA was obtained using a Transcriptor First Strand cDNA Synthesis kit (Roche). Quantitative real-time PCR was performed using a SYBR Green PCR Master Mix Reagent Kit (Life Technologies). Levels of mRNA were normalized to those of β-actin and expressed as fold change. Forward/reverse primers for several genes were used (Table 3; other sequences are available upon request):
Frozen human adipose tissue (400–500 mg) was homogenized with an Ultra-Turrax 8 (Ika). Total RNA from adipose biopsies, SVF, and isolated adipocytes were extracted by using an RNeasy Lipid Tissue Midi Kit (QIAGEN) following the manufacturer's instructions, and total RNA was treated with 55 U of RNase-free DNase (QIAGEN) to avoid contamination with genomic DNA. Between 0.2 and 1 μg of total RNA was reverse-transcribed to cDNA using TaqMan reverse transcription reagents (Applied Biosystems) and subsequently diluted with nuclease-free water (Sigma) to 20 ng/μl cDNA. For adipose tissue gene expression analysis, real-time quantitative PCR was performed, with duplicates, on a 7900HT Fast Real-Time PCR System using commercial Taqman Assays (Applied Biosystems). SDS software 2.3 and RQ Manager 1.2 (Applied Biosystems) were used to analyze the results with the comparative threshold cycle (CT) method (2ΔΔCT). CT values for each sample were normalized with an optimal reference gene (cyclophilin) after testing of three additional housekeeping genes: β-actin and RNA 18S. A panel of genes involved in the adipocyte differentiation and metabolism was selected in the study of CPT1A gene expression (Table 4).
Cytokine measurement in culture media.
Cytokine protein levels in culture medium of 3T3-L1 CARΔ1 adipocytes and RAW 264.7 macrophages were measured by Luminex technology with a MILLIPLEX Analyzer Luminex 200x Ponenet System (MCYTOMAG-70K-08 Mouse Cytokine MAGNETIC Kit; Merck Millipore).
Analysis of cellular redox status.
To detect ROS (superoxide) formation, MitoSOX Red (M36008, Life Techonologies) fluorescence was measured by flow cytometry. RAW 264 cells were infected with 100 moi AdCPT1AM (or AdGFP as control) for 48 h; then, 16 h prior to ROS measurement, macrophages were treated with 0.75 mM palmitate BSA-conjugated (or with BSA as control). Medium was removed, and cells were incubated for 30 min with PBS containing 5 μM MitoSOX Red. The labeled macrophages were washed three times with HBSS-Ca-Mg, pelleted, resuspended in 300 μl of HBSS-Ca-Mg, and fixed by adding 1.2 ml of absolute ethanol and keeping them at −20°C for 5 min. Cells were pelleted again and resuspended in HBSS-Ca-Mg containing 3 μM DAPI to mark their nuclei. Then macrophages were analyzed by flow cytometry (Gallios Cytometer, Beckman-Coulter). The fluorescence intensity of MitoSOX Red was measured using excitation at 510 nm and emission at 580 nm.
Data are expressed as means ± SE and analyzed statistically using Student's t-test (column analysis) or two-way ANOVA (grouped analysis). All figures and statistical analyses were generated using GraphPad Prism 6. P < 0.05 was considered statistically significant. For human data, statistical analyses were performed with SPSS 12.0. Results are expressed as means ± SD. The nonnormally distributed variables are represented as the median (interquartile range). Categorical variables are reported by number (percentages). Student's t-test analysis was used to compare the mean value of normally distributed continuous variables. Variables with a non-Gaussian distribution were analyzed using a nonparametric test (Kruskal-Wallis or Mann-Whitney test for independent samples or Wilcoxon test for related samples when necessary). Associations between continuous variables were sought by correlation analyses. Finally a stepwise multiple linear regression analysis was performed to determine independent variables associated with CPT1A gene expression levels in SAT and VAT depots. Results are expressed as unstandardized coefficient (B), and 95% confidence interval for B [95% CI(B)]. Differences are considered significant if a computed two-tailed probability value (P) is < 0.05.
CPT1A expression pattern in human adipose tissue from obese and diabetic patients. Visceral and subcutaneous adipose tissue (VAT and SAT, respectively) were analyzed from both an obesity cohort (lean, overweight and obese patients) and a T2D cohort (control and T2D patients). Tables 1 and 2 show the phenotypic and metabolic characteristics and CPT1A expression levels of the subjects. No differences in CPT1A gene expression levels either in SAT or in VAT depots were observed when comparing with the nonobese or the nondiabetic counterparts (Fig. 1, A and B; Tables 1 and 2). However, in the obesity cohort, CPT1A mRNA expression was significantly higher in lean and overweight VAT than in SAT (Fig. 1A); this difference was lost in the obese patients. These results were corroborated by Western blot with human adipose tissue of several lean and obese individuals (Fig. 1, C and D, P = 0.015). Similar results were obtained in the T2D cohort, where control subjects showed significantly higher CPT1A mRNA levels in VAT vs. SAT (Fig. 1B); however, this difference disappeared in T2D patients. Despite T2D patients showing a trend to express higher CPT1A levels in SAT and VAT compared with controls (the opposite of that in the obese subjects), this difference was nonsignificant. Since the CPT1B isoform is also expressed in human adipose tissue, we analyzed CPT1B mRNA (Fig. 1, E and F) and protein (data not shown) levels in human VAT and SAT of the obesity and the T2D cohort. No differences were seen among the groups.
To establish the main relationship between CPT1A gene expression and key adipocyte genes involved in differentiation and metabolic pathways, we explored a panel of genes (see materials and methods) both in SAT and in VAT depots of the obesity cohort. Results are shown from those genes that changed the most (up or down; Tables 5 and 6). Simple association analysis showed an inverse correlation between CPT1A and peroxisome proliferator-activated receptor-γ (PPARγ) in SAT (r = −0.38, P = 0.002; Table 5). Positive CPT1A correlation in both VAT and SAT was found with 1-acylglycerol-3-phosphate O-acyltransferase 5 (AGPAT5; phospholipid biosynthesis), sterol regulatory element binding transcription factor 1 (SREBF1; glucose and lipid metabolism), B cell CLL/lymphoma 2 (BCL2; antiapoptosis), and CD163 (macrophage marker) (Table 5).
To study the main determinants of CPT1A gene expression levels, a stepwise multiple regression analysis was performed, including the aforementioned bivariate associations and confounding factors such as BMI, age and sex. This model showed that SAT CPT1A was positively associated with AGPAT5, SREBF1, and CD163 and that VAT CPT1A was positively correlated with SREBF1 and CD163 and negatively with age and PPARγ (Table 6). The inverse relationship between CPT1A and PPARγ was corroborated with the human adipocyte cell line SGBS. CPT1A mRNA expression dropped to a new steady state in adipocytes that was 11% of its expression in fibroblasts (data not shown).
CPT1A is highly expressed in human adipose tissue macrophages.
To determine the cellular distribution of CPT1A gene and protein in human adipose tissue biopsies, we performed qRT-PCR and immunostaining analysis on both adipose and SVF. CPT1A mRNA levels were 42.6- and 43.4-fold increased in the SVF compared with mature adipocytes in both human SAT (P < 0.05) and VAT (P < 0.05), respectively (Fig. 2A). Immunohistological examination of SAT from obese subjects revealed CPT1A+ cells mostly in the SVF (Fig. 2B). Immunofluorescence detection showed a bright staining pattern in cells resembling adipose tissue macrophages. Costaining analysis using CPT1A and CD68 (a macrophage marker) antibodies confirmed the expression of CPT1A in macrophages (Fig. 2C). Macrophages seemed to localize forming the so-called “crown-like structures” surrounding the adipocytes.
CPT1AM-expressing adipocytes show enhanced FA oxidation and reduced TG content.
To further study the role of CPT1A in adipocytes and macrophages, we decided to continue with in vitro studies. Since 3T3-L1 adipocytes are inefficiently infected with adenovirus, we decided to use the high-infection efficiency white adipocyte cell culture line, 3T3-L1 CARΔ1 adipocytes (31) (Fig. 3A). Cells were transduced for the first time with adenoviruses carrying the CPT1AM gene or GFP as a control. Interestingly, CPT1AM-expressing adipocytes were partially protected from palmitate-induced cell death (Fig. 3C).
CPT1A mRNA, protein, and activity levels were increased in CPT1AM-expressing adipocytes compared with GFP control cells (Fig. 4, A–C). CPT1AM-expressing adipocytes retained most of the CPT1 activity after incubation with the CPT1A inhibitor malonyl-CoA (Fig. 4C). The FA oxidation (FAO) rate was concordantly enhanced (1.37-fold increase, P < 0.05) in CPT1AM-expressing adipocytes (Fig. 4D). FA undergoing β-oxidation yield acetyl-CoA moieties that have two main possible fates: 1) complete oxidation to CO2 and ATP production or 2) conversion to ketone bodies (mainly in the liver). Here, the total FAO rate was calculated as the sum of acid-soluble products plus CO2 oxidation. CPT1AM expression blocked the palmitate-induced increase in TG content (Fig. 4E).
Enhanced adipocyte FAO improves insulin sensitivity and reduces inflammation.
We examined the effect of increased FAO on insulin sensitivity and inflammatory responses in 3T3-L1 CARΔ1 adipocytes infected with AdCPT1AM. Palmitate-induced decrease in insulin-stimulated Akt phosphorylation and insulin receptor-β (IRβ) protein levels was partially restored in CPT1AM-expressing adipocytes (Fig. 4, F–H). Palmitate-induced increase of proinflammatory markers [IL-1β, monocyte chemoattractant protein-1 (MCP-1), and IL-1α] mRNA and protein levels was blunted in CPT1AM-expressing adipocytes (Fig. 4, I–K). Several palmitate concentrations and times of incubation were used to better fit the different dose and time responses of the cytokines and parameters measured. Consistent with previous studies (9, 11), palmitate incubation raised cytokines expression two- to threefold.
Increased FAO in CPT1AM-expressing macrophages protects from palmitate-induced TG accumulation.
Since CPT1A was highly expressed in the SVF, and particularly in macrophages, of human adipose tissue, we decided to further analyze the effect of increased FAO on cultured macrophages. RAW 264.7 macrophages were efficiently infected with AdCPT1AM (Fig. 3B). CPT1AM-expressing macrophages were protected from palmitate induced cell death (Fig. 3D). CPT1AM-expressing macrophages showed a 2.4-fold (P < 0.01) increase in CPT1A mRNA levels, a 6.6-fold (P < 0.01) increase in protein levels, and a 2.2-fold (P < 0.05) increase in activity levels (Fig. 5, A–C). In addition, we showed that malonyl-CoA did not inhibit CPT1 activity in CPT1AM-expressing macrophages (Fig. 5C). CPT1AM-expressing macrophages showed a 1.5-fold increase in FAO rate compared with GFP control cells (Fig. 5D, P < 0.05) and a total restoration in palmitate-induced enhancement of TG content (Fig. 5, E and F).
Enhanced macrophage FAO reduced inflammation, ER stress, and ROS damage.
Palmitate-induced increase in proinflammatory cytokines [TNF-α, MCP-1, IL-1β, Toll-like receptor 4 (TLR-4), and IL-12p40], and ER stress markers (CHOP, GRP78, protein disulfide isomerase (PDI), and ER degradation enhancing α-mannosidase-like protein (EDEM)] mRNA and protein levels were blunted in CPT1AM-expressing macrophages (Fig. 6, A, B, D, and E). Consistent with previous studies (18, 47, 48), palmitate incubation raised cytokines expression two- to threefold. No differences were seen in anti-inflammatory markers such as IL-10, Mgl-1, and IL-4 in CPT1AM-expressing cells incubated with or without palmitate (Fig. 6C). Incubation with etomoxir, a permanent inhibitor of CPT1A, counteracted the reduction of MCP-1 expression seen in CPT1AM-expressing cells incubated with palmitate (data not shown). We also studied the effect of enhanced FAO in RAW 264.7 macrophages on palmitate-induced ROS damage by protein carbonyl content analysis. Palmitate-induced ROS damage was reduced in CPT1AM-expressing macrophages (Fig. 6F). This reduction was not detected when ROS (superoxide) was directly measured using the MitoSOX Red probe (Fig. 6G).
The obesity epidemic has put a spotlight on adipose tissue as a key player in obesity-induced insulin resistance (38). Obese individuals and those with T2D have lower FAO rates (17, 19, 37). Although these data were reported in skeletal muscle, we expected to see reduced CPT1A expression levels in the adipose tissue of both obese and T2D patients. However, no differences were seen in CPT1A mRNA expression between the obese or T2D patients and their respective controls either in VAT or in SAT. Other authors have reported a decrease in VAT CPT1 mRNA and protein levels in obese individuals (20). However, those authors did not specify which of the CPT1 isoforms was measured in VAT, CPT1A or CPT1B. We showed that CPT1A expression is higher in adipose tissue macrophages than in mature adipocytes. Since the obese adipose tissue has higher infiltration of immune cells such as macrophages, we postulate that the putative decrement of CPT1A expression in obese individuals could be compensated for by increased expression from the infiltrated macrophages and thus that no differences are seen between the groups. The CPT1B isoform is also expressed in human adipose tissue, and it has been shown to raise FAO in metabolic tissues such as skeletal muscle (3). Thus, we measured mRNA and protein levels in the obese and T2D cohorts. However, no differences were seen among the groups, indicating that CPT1B expression is not changed by obesity and T2D.
We found that, in insulin-sensitive individuals (control and overweight patients from the obese cohort and control patients from the T2D cohort), CPT1A mRNA expression was higher in VAT than in SAT. However, no differences between VAT and SAT were seen in the more insulin-resistant individuals with a more proinflammatory environment: obese and T2D patients. A similar phenomenon was described for T regulatory cells, described to have anti-inflammatory properties and to improve obesity-induced insulin resistance (7). Those authors reported that the VAT and SAT of healthy individuals had similar low numbers of T regulatory cells at birth, with a progressive accumulation over time in the VAT, though not in the SAT. Our results suggest a CPT1A expression balance between SAT and VAT depots that may be disturbed in obese and T2D patients. The difference in CPT1A expression between these two fat depots is potentially crucial, given the association of VAT, but not SAT, with insulin resistance (1, 52). It might indicate, in healthy individuals, a potential protective role of CPT1A in the more insulin-resistant associated VAT.
Gene expression analysis revealed a negative association between CPT1A and the adipocyte marker of differentiation PPARγ. This is consistent with the fact that while white adipocytes mature they shift their lipid preferences to storage rather than oxidation. Aging was associated with reduced CPT1A expression in VAT. This might reflect a potential protective role of CPT1A expression in VAT that is lost with age. Considering that VAT accretion is a hallmark of aging and especially that it is a stronger risk factor for comorbidities and mortality (23), we speculate a favorable role of enhanced CPT1A expression in age metabolic decline and related pathological conditions. Positive correlation in both VAT and SAT CPT1A was found with AGPAT5, SREBF1, Bcl2, and CD163. These results may indicate a potential role of CPT1A in lipid biosynthesis processes (AGPAT5), glucose and lipid metabolism (SREBF1), and protecting adipose tissue from apoptosis (Bcl2). The positive association between CPT1A and CD163 (macrophage marker) was not surprising given the higher CPT1A expression in macrophages than in adipocytes (Fig. 2).
We are aware that many of the aforementioned associations may be secondary to obesity or T2D and that no causal relationship may be inferred with this study design. To prove the causality of some of these observations, we performed in vitro studies directly targeting adipocytes and macrophages to burn off the excess lipids through an increase in FAO. We used the high-infection efficiency adipocyte cell line 3T3-L1 CARΔ1 (31) to express for the first time CPT1AM through adenoviral infection. Noteworthy, white adipocytes are designed to store lipids rather than to oxidize them. Thus, CPT1 activity in WAT is lower than in other tissues (6). However, CPT1AM-expressing adipocytes showed a 4.3-fold increase in CPT1 activity that was not inhibited despite incubation with high concentrations of malonyl-CoA. Since increased lipid accumulation, inflammation, ER stress, and ROS-induced protein damage trigger metabolic diseases, we decided to measure TG content, inflammation, ER stress, and ROS damage as important mechanisms that could explain the potential protective effect of CPT1AM expression. Enhanced FAO led to complete restoration of TG content, improved insulin signaling (measured as pAkt), increased IRβ expression and cell viability, and reduced inflammation in palmitate-incubated CPT1AM-expressing adipocytes. CPT1AM-expressing adipocytes showed a general improvement in lipid-induced derangements as a consequence of increased FA flux through mitochondria. However, enhanced FA flux in the absence of a concomitant dissipation of FAO metabolites has been associated with increased ROS damage (35) and inflammation (8, 21, 43). Interestingly, although no differences were seen in ER or oxidative stress (data not shown), CPT1AM-expressing adipocytes showed a significant decrease in proinflammatory mediators such as IL-1β and MCP-1. The favorable role of CPT1A in adipocytes to attenuate FA-evoked insulin resistance and inflammation has been also described to act via suppression of JNK (9). These results suggest that factors other than a FAO increase per se are responsible for ROS production and inflammation. Accumulation of toxic substances (diacylglycerol or ceramides) (49), hypoxia (15), as well as cytokines (42) might participate in the induction of ROS damage and the inflammatory state. Several researchers have demonstrated that enhanced FAO through CPT1A or CPT1AM expression results in a decrease in relevant lipid mediators involved in inflammation and insulin resistance such as diacylglycerol, intracellular NEFAs, free FA, ceramides, and TG (3, 9, 13, 26, 29, 40, 45). Although some authors (3) did not see changes in skeletal muscle acylcarnitines' profile, our group has shown an increase in several acylcarnitines in CPT1AM-expressing neurons (25).
FA undergoing β-oxidation yield acetyl-CoA moieties that have two main possible fates: 1) entry to the TCA cycle for complete oxidation and ATP production or 2) conversion to ketone bodies (mainly in the liver). We observed increased FAO to CO2 and acid-soluble products in CPT1AM-expressing adipocytes and macrophages. CPT1AM expression in liver has been shown to enhance ATP and ketone body production with no changes in glucose oxidation (13, 29). All together, this indicates a metabolic rate switch toward FA.
Monocytes were the first immune cells reported to infiltrate obese adipose tissue, differentiate to macrophages, produce inflammatory cytokines, and trigger insulin resistance (56, 57). Thus, we examined whether CPT1AM expression could play a protective role in obesity-induced macrophage derangements. We found that, in human WAT, CPT1A is highly expressed in SVF compared with adipocytes. This happened in both human VAT and SAT. A closer histological and immunofluorescence examination showed that macrophages present in the adipose tissue expressed CPT1A. This does not rule out CPT1A expression in other immune cells also present in the adipose tissue such as T and B cells, T regulatory cells, and mast cells.
Given the high CPT1A expression in human adipose tissue macrophages, we decided to study the effect of CPT1AM in RAW 264.7 macrophages. A permanently enhanced FAO rate in CPT1AM-expressing macrophages led to a complete restoration of palmitate-induced increase in TG content and a decrease in inflammation and ER and oxidative stress without affecting cell viability. Recent data show that FAO is capable of regulating the degree of acyl chain saturation in ER phospholipids (28). Since increasing the degree of saturation in ER phospholipids has been described to directly activate ER stress and inflammation (28), this might provide a mechanistic link to how FAO alleviates ER stress under palmitate loading. Thus, enhancing CPT1A expression in macrophages may be a potential approach to fight against obesity-induced disorders.
In conclusion, we have shown that CPT1A expression was higher in human adipose tissue macrophages than in mature adipocytes and that it was differentially expressed in VAT vs. SAT. Further in vitro studies demonstrated that an increase in FAO in lipid-treated adipocytes and macrophages reduced TG content and inflammatory levels, improved insulin sensitivity in adipocytes, and reduced ER stress and ROS damage in macrophages. Adipocyte-specific knockout or transgenic animal models for CPT1A would be especially relevant to elucidate its potential protection against obesity-induced insulin resistance in vivo. Our data support the hypothesis that pharmacological or genetic strategies to enhance FAO may be beneficial for the treatment of chronic inflammatory pathologies such as obesity and T2D.
This study was supported by the Spanish Ministry of Science and Innovation (Grants SAF2010-20039 and SAF2013-45887-R to L. Herrero, SAF2011-30520-C02-01 to D. Serra, PI11/00085 to J. J. Vendrell, SAF2012-33014 to B. Peral, SAF2012-36186 to S. Fernández-Veledo, SAF2012-30708 to M. Vázquez-Carrera, SAF2011-23626 to F. Villarroya, and doctoral fellowships to M. I. Malandrino and J. F. Mir), by the CIBER Fisiopatología de la Obesidad y la Nutrición (Grant CB06/03/0001 to D. Serra), and CIBER Diabetes y Enfermedades Metabólicas Asociadas (Grant CB07/08/0003 to M. Vázquez-Carrera), Instituto de Salud Carlos III, by the European Union (BetaBat project FP7-277713 to F. Villarroya), by the European Foundation for the Study of Diabetes (EFSD)/Lilly and EFSD/Janssen-Rising Star research fellowships to L. Herrero, and by a L'Oréal-UNESCO “For Women in Science” research fellowship to L. Herrero. S. Fernández-Veledo acknowledges support from the “Miguel Servet” tenure track program (CP10/00438) from the Fondo de Investigación Sanitaria and cofinanced by the European Regional Development Fund. M. Weber is a recipient of the Ciência sem Fronteiras-CNPq fellowship (237976/2012-9).
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: M.I.M., R.F., M.W., M.C.-D., J.F.M., L.V., X.E., M.G.-S., B.P., L.S., S.F.-V., N.C., M.V.-C., F.V., J.J.V., D.S., and L.H. conception and design of research; M.I.M., R.F., M.W., M.C.-D., J.F.M., L.V., X.E., M.G.-S., B.P., and L.S. performed experiments; M.I.M., R.F., M.W., M.C.-D., J.F.M., L.V., X.E., M.G.-S., B.P., L.S., M.V.-C., F.V., J.J.V., D.S., and L.H. analyzed data; M.I.M., R.F., M.W., M.C.-D., J.F.M., L.V., X.E., M.G.-S., B.P., L.S., S.F.-V., N.C., M.V.-C., F.V., J.J.V., D.S., and L.H. interpreted results of experiments; M.I.M., R.F., M.W., M.C.-D., L.V., X.E., M.G.-S., B.P., L.S., J.J.V., and L.H. prepared figures; M.I.M., R.F., M.C.-D., J.F.M., X.E., B.P., J.J.V., and L.H. drafted manuscript; M.I.M., R.F., M.C.-D., J.F.M., X.E., B.P., N.C., J.J.V., D.S., and L.H. edited and revised manuscript; M.I.M., R.F., M.W., M.C.-D., J.F.M., L.V., X.E., M.G.-S., B.P., L.S., S.F.-V., N.C., M.V.-C., F.V., J.J.V., D.S., and L.H. approved final version of manuscript.
We thank Prof. F. G. Hegardt and Dr. G. Asins for helpful comments and suggestions, A. Orozco for technical assistance, and R. Rycroft from the Language Service of the University of Barcelona for valuable assistance in the preparation of the English manuscript. We also thank D. Orlicky for kindly providing 3T3-L1 CARΔ1 adipocytes.
- Copyright © 2015 the American Physiological Society