|Home | About | Journals | Submit | Contact Us | Français|
Obesity is associated with adipose tissue remodeling, characterized by adipocyte hypertrophy and macrophage infiltration. Previously, we have shown that very low density lipoprotein receptor (VLDLR) is virtually absent in preadipocytes but is strongly induced during adipogenesis and actively participates in adipocyte hypertrophy. In this study, we investigated the role of VLDLR in adipose tissue inflammation and adipocyte-macrophage interactions in wild type and VLDLR-deficient mice fed a high fat diet. The results show that VLDLR deficiency reduced high fat diet-induced inflammation and endoplasmic reticulum (ER) stress in adipose tissue in conjunction with reduced macrophage infiltration, especially those expressing pro-inflammatory markers. In adipocyte culture, VLDLR deficiency prevented adipocyte hypertrophy and strongly reduced VLDL-induced ER stress and inflammation. Likewise, cultures of primary peritoneal macrophages show that VLDLR deficiency reduced lipid accumulation and inflammation but did not alter chemotactic response of macrophages to adipocyte signals. Moreover, VLDLR deficiency tempered the synergistic inflammatory interactions between adipocytes and macrophages in a co-culture system. Collectively, these results show that VLDLR contributes to adipose tissue inflammation and mediates VLDL-induced lipid accumulation and induction of inflammation and ER stress in adipocytes and macrophages.
Obesity induces macrophage infiltration into adipose tissue leading to the establishment of chronic inflammation and metabolic dysfunction (1, 2). With obesity, remodeling of adipose tissue creates a new microenvironment where macrophages and adipocytes are in close contact initiating inflammatory signals that exacerbate adipose tissue inflammation (1). In this scheme, saturated fatty acids (FAs)2 released from hypertrophied adipocytes and triglyceride-rich lipoproteins promote the expansion of the pro-inflammatory classically activated macrophages (also known as M1 phenotype) (3–5). Activated M1 macrophages in turn produce pro-inflammatory cytokines and chemokines that exacerbate adipocyte inflammation (6). In this system of interactions, cell membrane receptors in adipocytes and macrophages are crucial because they control fluxes of extracellular mediators into the cell (7, 8) and activate stress signaling pathways, such as the c-Jun N-terminal kinase (JNK) (5, 9) and nuclear factor-κB (NF-κB) (10, 11). Therefore, identification of proteins and cell membrane receptors that are involved in adipocyte-macrophage cross-talk is important to understand the mechanisms underlying adipose tissue inflammation in obesity and may lead to novel therapeutic strategies to prevent or treat obesity-related complications.
Very low density lipoprotein receptor (VLDLR) is a member of the low density lipoprotein receptor (LDLR) family. By opposition to LDLR, VLDLR is highly expressed in adipose tissue, heart, and skeletal muscles but is virtually absent in liver. It binds apolipoprotein E-(apoE) triglyceride-rich lipoproteins, enhances LPL activity (12), and mediates lipid entry into the cell (13, 14). Initially, Frykman et al. (15) reported that VLDLR-null (vldlr−/−) mice are leaner, with normal blood lipids. However, several studies have revealed an increase in plasma TGs after feeding (16) or fasting (17) states. Later studies (12, 18) have established a link between VLDLR expression, lipoprotein lipase activity, and blood triglyceride levels. Recently, we reported that VLDLR expression is virtually undetectable in preadipocytes but increases gradually with differentiation to reach a maximum in mature adipocytes (14, 19). These findings confer to VLDLR a prominent role in lipid uptake and adipocyte maturation (14). Moreover, VLDLR seems to be linked to obesity, because its deficiency protects mice from obesity (17). In agreement with the results in mice, patients with VLDLR mutations have abnormally low body mass index compared with control subjects (20, 21).
In addition to adipocytes, VLDLR is also expressed in immune cells, including macrophages (22, 23), but it is not known whether it regulates adipose tissue remodeling in obesity. In this study, we investigated the role of VLDLR in obesity-related adipose tissue inflammation and adipocyte-macrophage interaction.
Wild type (WT) and VLDLR-null (vldlr−/−) mice were purchased from The Jackson Laboratory. Six-week-old sex-matched mice were fed obesogenic high fat diet (36%, w/w, adjusted calories from fat, Bio-Serv S3282) for 16 weeks. All animal procedures were approved by the Institutional Animal Care and Use Committee, performed at Vanderbilt University and Hackensack University Medical Center, AAALAC-accredited facilities, and the investigation followed the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health.
Nonesterified fatty acids and triacylglycerols were assayed using kits from Wako (Richmond, VA) and Thermo Scientific (Middletown, VA), respectively. Fasting insulin was measured with the ultrasensitive mouse insulin ELISA (R&D Systems, Minneapolis, MN).
The levels of TNF-α and IL-6 in culture medium were tested by ELISA procedures (RayBiotech, Norcross, GA). Macrophage attractant protein-1 (MCP-1) and leptin levels were measured by ELISA kits (R&D systems, Minneapolis, MN). Procedures were performed according to the manufacturer's instructions (24).
For glucose tolerance test (GTT), mice were fasted overnight and injected intraperitoneally with 2 mg of glucose/g of body weight. Tail vein blood glucose was measured at the indicated time using a One-Touch basic glucometer (25). For insulin tolerance test (ITT), mice were fasted for 4 h prior to intraperitoneal injection of insulin (0.75 units/kg). Blood glucose level was tested at the indicated time points.
Samples of epididymal white adipose tissue were fixed by immersion in 4% paraformaldehyde in 0.1 m phosphate buffer. Five-micrometer-thick serial sections were obtained from paraffin-embedded tissues. To assess adipose tissue morphology, sections were stained by hematoxylin and eosin (H&E) as described elsewhere (26). Fat cell size estimation was based on ×100 magnification of histological sections images (Zeiss digital camera, New York). Images were converted into a binary format with ImageJ (version 1.47, National Institutes of Health) and compared with the original images to ensure an accurate conversion. A total of 500 cells was measured in five different fields from four mice. The data were first averaged per section and then per animal. The distribution of adipocyte size was determined by relative frequencies of adipocytes having a size within a regular interval. Other sections were used for immunostaining of adipose tissue macrophages (ATMs) using specific marker F4/80 according to the procedure described by Ueda et al. (26). Briefly, sections were incubated overnight (4 °C) with anti-mouse F4/80 primary antibody (1:100; Serotec) followed by the application of biotinylated HRP-conjugated rat anti-goat IgG secondary antibody (Vector Laboratories, Burlingame, CA). Primary antibody was omitted from negative controls. Positive controls were also used to test the antibody specificity. Histochemical reactions were performed using the EnVision Doublestain System (Dako Cytomation) and counterstained with hematoxylin. Sections were examined with light microscopy, and digital images were used to evaluate macrophage contents (8). The number of crown-like structures (CLS) was determined in five different ×100 magnified fields from four mice per group. Data were expressed as number of crown-like structures per low power field (CLS/LPF). For immunofluorescence staining of macrophages, sections were incubated with rabbit polyclonal anti-CD11c (BD Biosciences) followed by the application of appropriate secondary antibody. Stained sections were examined under a fluorescence microscope (Olympus BX51, Goettingen, Germany).
All procedures for cultures of adipose tissue explants were carried out using a sterile technique as described earlier (27). Epididymal fat from WT and vldlr−/− mice fed a high fat diet was minced and washed with HEPES/salts buffer. Then, ~500-mg samples were placed in 2 ml of serum-free M199 medium with 1% BSA under 95:5 O2/CO2 with daily medium replenishment (27). Medium was collected from explant cultures to test chemokine secretion using ELISA kits and free fatty acid release using an enzymatic kit (25).
Isolation of adipocytes and adipose tissue stromal vascular fraction (SVF) was performed according to our previous procedure (14, 19). Briefly, freshly collected epididymal fat pads were excised and minced in PBS with calcium chloride and 1% BSA. Tissue suspensions were rinsed with PBS to remove erythrocytes and free cells. Then, tissue pieces were digested in Krebs buffer containing 1 mg/ml type I collagenase (Worthington) and 1% fatty acid-free BSA (Sigma) for 20 min at 37 °C with shaking. The suspension was filtered through a sterile 100-μm nylon mesh and centrifuged at 500 × g for 5 min. After centrifugation, buoyant adipocytes were isolated, washed twice with Krebs-Ringer HEPES (KRH) buffer, and collected as the adipocyte fraction. The remaining pellet representing the SVF was washed three times with Krebs buffer and further incubated with red blood cell lysis buffer to remove red blood cells. This fraction was used to test gene expression by qPCR and macrophage-specific marker expression with flow cytometry.
After washing, cells of the SVF were resuspended in 300 μl of staining buffer (PBS containing 2% FBS) containing FcBlock (BD Biosciences) and stained with conjugated antibodies, all on ice for 30 min and in the dark (9). Antibodies used for staining surface antigens were CD11c-EP, F4/80-FITC (both from E-Bioscience, San Diego), and CD206-APC (BioLegend, San Diego). For each sample, triplicates of cell preparations were double-stained with CD11c-EP and F4/80-FITC or with CD206-APC and F4/80-FITC. The samples were then processed with an LSRII flow cytometer (BD Biosciences), and data were analyzed using Cytopaint Classic software (Tree Star, Inc.). FSc-A/FSc-H gates were used to identify single cells. Fluorescence intensity of samples was normalized using isotype control antibodies, and cells were incubated in the absence of conjugated antibodies. Live cells were identified with propidium iodide staining (Invitrogen).
Primary adipocytes were prepared from adipose tissue digest as described above. After washing, adipocyte fractions were examined by microscopy to ensure proper isolation and cell count numbers before experiments. Glucose uptake was measured using tritiated 2-deoxyglucose as described previously (28). Briefly, freshly isolated adipocytes were incubated in Krebs-Ringer phosphate (KRP) buffer (136 mmol/liter NaCl, 4.5 mmol/liter KCl, 1.25 mmol/liter CaCl2, 1.25 mmol/liter MgCl2, 0.6 mmol/liter Na2HPO4, 0.4 mmol/liter NaH2PO4, 10 mmol/liter HEPES, and 0.1% BSA) at 37 °C for 20 min with or without insulin (100 μm). Then 10 μl of KRP containing 0.25 mmol/liter unlabeled 2-deoxyglucose and 0.1 μCi of tritiated 2-deoxyglucose (PerkinElmer Life Sciences) was added to each preparation, and incubation was continued for another 10 min. Cells were then washed with ice-cold PBS and lysed in 0.1 m NaOH. Radioactivity was counted in scintillation counter (PerkinElmer Life Sciences) and corrected using unlabeled cell blanks.
Peritoneal macrophages (PM) were harvested from high fat diet-fed WT and vldlr−/− mice by lavaging the peritoneal cavity with PBS (29). Macrophages were then plated on 12-well plates (1.5 × 106 cells per well) in RPMI 1640 medium supplemented with 10% FBS, 50 IU/ml penicillin, 50 μg/ml streptomycin, and 2 mm l-glutamine. Four hours after plating, cells were washed three times with PBS to remove the nonadherent cells. Adherent cells were then incubated for an additional 48 h at 37 °C before treatment.
Migration of PM was measured in a modified Boyden chamber migration assay using Transwell inserts with an 8 μm porous membrane (Corning) as described by Patsouris et al. (30). To compare the chemotactic effects of adipose tissues of WT and vldlr−/− mice, conditional medium was prepared from adipose tissue explants cultured for 24 h in M199 medium as described above (27). Cells were loaded into the migration chamber and conditional medium from WT (CMWT) or vldlr−/− (CMvldrko) adipose tissues were loaded in the lower chamber. After allowing cell migration for 16 h, cells were removed from the upper side of the membranes, and nuclei of migratory cells on the lower side of the membrane were stained with 4′,6-diamidino-2-phenylindole (DAPI). The nuclei were visualized by fluorescence microscopy, and the average number of migratory cells was determined from averaging four fields (30).
Buoyant VLDL particles (d <1.006 g/ml) were isolated form pooled plasma of WT mice fed high fat diet by ultracentrifugation (31). VLDL preparation was extensively dialyzed and diluted in PBS and then protein and triglyceride concentrations were measured (31). Fatty acid composition of VLDL lipids was determined by gas liquid chromatography (GLC) according to the one-step transesterification procedure as described elsewhere (31). VLDL preparations were used within a period of 24 h.
To test if VLDLR deficiency modulates the effects of VLDL on macrophages, we challenged peritoneal macrophages collected from WT and vldlr−/− mice with VLDL. Accordingly, macrophages seeded in 12-well plates were treated with 0, 50, 100, or 200 μg/ml VLDL based on triglyceride concentration. Cells were then incubated for a period of 24 h after which cells were used to examine lipid contents. Cells were fixed with 4% formalin and then stained for 30 min with LipidTOX (Invitrogen) for neutral lipids followed by DAPI for nucleus staining. Cells were then examined with fluorescence microscope (Olympus BX51) equipped with Olympus C-3030 digital Olympus camera. Additional incubations were used to collect the cells in lysis buffer for triglyceride quantification with enzymatic kits (Wako) or frozen for Western blotting and qPCR analyses.
Culture and differentiation of preadipocytes were performed as described in our previous procedure (14). Briefly, preadipocytes collected in adipose SVF were plated in DMEM/F12 (1:1) medium supplemented with 10% FBS, 33 μm biotin, 17 μm pantothenic acid, 50 IU/ml penicillin, 50 μg/ml streptomycin and 2 mm l-glutamine. After confluence was reached, preadipocytes were differentiated for 4 days in differentiation medium containing 0.02 μm insulin, 25 nm dexamethasone, 0.5 mm 3-isobutyl-1-methylxanthine (IBMX), 10 μg/ml transferrin, and 0.2 nm thyroid hormone T3 (Sigma). Mature adipocytes were identified with the presence of lipid droplets and Oil Red O staining.
Because VLDL particles are the natural ligands of VLDLR (13), we questioned whether VLDLR mediates the effects of VLDL on lipid accumulation and inflammation in adipocytes. Preadipocytes isolated from fat pads of WT and vldlr−/− mice were cultured and differentiated as described above. At day 12 of differentiation, mature adipocytes were cultured with VLDL at the indicated doses for an additional 48 h. The cells were then used to examine triglyceride accumulation in adipocytes by enzymatic kit reaction and Oil Red O staining (19). Additional preparations were also used to collect cells in lysis buffer for Western blotting and qPCR analyses.
Double labeling of VLDL was performed with a two-stage procedure according to our previous procedure (14). First, [3H]TG VLDL labeling was performed endogenously in WT mice with injection of 100 μCi of [3H]palmitate (PerkinElmer Life Sciences) into the tail vein (14). Blood was collected 60 min later, and serum samples were subjected to ultracentrifugation to obtain [3H]TG-labeled VLDL at density d <1.016 g/liter. To examine the efficiency of the labeling, VLDL lipids were extracted and separated by thin layer chromatography (tlc) (32), and radioactivity associated with lipid classes was counted in scintillation counter (14). The bulk of 3H counts (about 89%) was associated with TG fraction. In a second stage, [3H]TG-labeled VLDL were iodinated in vitro with 125I using the iodine monochloride method of McFarlane (33) as reported previously by us (14, 31). Double-labeled 3H,125I-labeled VLDL particles were collected after passage through size exclusion Sephadex G-25 PD-10 column, and extensive dialysis and filtration through a 0.45-μm filter. More than 98% of the radioactivity was precipitated by trichloroacetic acid, and less than 4% of the radioactivity was extracted by diethyl ether. 125I associated with VLDL was counted in an LKB gamma counter. To compare the uptake of VLDL with BSA-bound FA, [3H]palmitate was complexed in vitro to BSA at a molar ratio 5:1 as described earlier (27, 28). WT and vldlr−/− adipocytes cultured in 12-well plates were differentiated to reach maturity. At day 12 of differentiation, cells were cultured in serum-free medium with 1% BSA. Then 100 μl of medium containing either [3H]palmitate complexed to albumin (1 μCi/ml) or 3H,125I-labeled VLDL ( 0.5 μCi of 3H and 1.5 × 106 cpm 125I) were added separately. Control cells were cultured in an equal volume of serum-free medium with 1% BSA. Cultures were continued for an additional 3 h at 37 °C. Then cells were extensively washed with ice-cold PBS containing 1% albumin and collected in lysis buffer. Aliquots were counted for 125I radioactivity in a gamma counter. Additional aliquots were used for lipid extraction and 3H counting in a scintillation counter. Separation of major lipid classes was also performed with tlc on silica gel plates and using a solvent mixture containing petroleum ether, diethyl ether, and glacial acetic acid (70:30:1, v/v/v) (14, 32). This procedure gave clear separation between major lipid fractions (32). Identified lipid spots were then scraped and counted in scintillation liquid. Protein contents were measured (14) and used to calculate rates of uptake. Preliminary experiments indicated that uptake and incorporation of VLDL-derived labels in newly formed lipids in adipocytes was better detected after 3 h of incubation. We also examined the uptake of labeled VLDL and BSA-bound palmitate over a short time of incubation (60 s) using the procedure described above with the exception that total 125I and 3H in adipocytes was measured without tlc separation. These preliminary results showed that recovery of VLDL-derived labels in both WT and vldlr−/− adipocytes were very low and did not yield meaningful information. We also questioned the possibility of differential label recycling between cells according to the genotypes. Preliminary experiments designed to examine rates of FA oxidation in adipocyte cultures showed comparable results between WT and vldlr−/− cells.
To examine whether VLDLR deficiency affects VLDL-induced ER stress, WT and vldlr−/−adipocytes differentiated in vitro were cultured with VLDL (300 μg/ml), thapsigargin (10 μg/ml), or both. Control cells were treated with an equal volume of vehicle. Then cells were collected for Western blotting and qPCR analysis to examine ER stress markers.
Co-cultures were performed in indirect co-culture Transwell system according to the procedure of Suganami et al. (3). Briefly, preadipocytes were seeded in 12-well plates and differentiated for 12 days as described above. The PM (1 × 105 cells/well) were plated and grown in the top inserts. Co-cultures were treated with 300 μg/ml VLDL for an additional 24 h after which co-culture medium and adipocytes in the bottom chambers were collected. Medium was used to measure total production of IL-6, TNF-α, and MCP-1 by ELISA, and adipocyte lysate was used to examine protein levels by Western blotting and gene expression by qPCR.
Cell viability was examined for each of the experimental conditions indicated above (explants and primary cell cultures) using MTT (Cayman Chemical, Ann Arbor, MI) and lactate dehydrogenase (LDH) (Biovision, Milpitas, CA) assays. LDH is an intracellular enzyme but could be liberated in the medium when cells are injured. Therefore, activity of LDH in the medium could be used as an index of cell injury and death. Cell and explant cultures designed for cell viability assays were treated exactly as described for each experiment. Following each treatment, cultures were incubated with DMEM containing FBS for an additional time (4 and 24 h for MTT and LDH assay, respectively). MTT assay was performed according to the manufacturer's instructions. For LDH activity assay, medium was harvested after centrifugation at 1200 rpm for 5 min, and aliquots were used for colorimetric reaction and reading of absorbance at 440 nm.
These procedures were performed exactly as described in our previous procedure (14). Briefly, equal amounts of proteins were subjected to SDS-PAGE and electrotransfer to Immobilon-P membranes (Millipore, Billerica, MA). Membranes were incubated with the following primary antibodies: phospho-SAPK/JNK (Thr183/Tyr185), phospho-c-Jun (Ser63), total c-Jun, total SAPK/JNK, phospho-p38 MAPK (Thr180/Tyr182), total p38 MAPK, phospho-Akt (Ser473), Akt, p-PERK (Thr980), phospho-eIF2α (Ser51), total eIF2α (Cell Signaling Technology), PERK, CHOP (Santa Cruz Biotechnology), and phospho-IRS-1 (Tyr608) (Millipore). The membranes were incubated with the appropriate second antibody followed by detection with chemiluminescent substrate products (Pierce) (24). Band intensity for each protein was analyzed by densitometry (ImageJ version 1.37), and corrections were made using β-actin intensity reading. For total IRS-1 and phosphorylated-IRS-1, immunoprecipitation was performed prior to immunoblotting. Briefly, tissue lysates were incubated with anti-IRS-1 antibody at 4 °C for 1.5 h and then rotated with protein G-agarose (Invitrogen) overnight at 4 °C. The protein G-agarose beads were washed three times with cold lysis buffer. Protein G-agarose beads were then suspended in SDS sample buffer and incubated at 100 °C for 5 min prior to electrophoresis and immunoblotting.
Total RNA was extracted using a total RNA fatty and fibrous tissue pack (Bio-Rad) according to the manufacturer's protocol. Complementary DNA was synthesized from 1 μg of total RNA with iScript reverse transcriptase (Bio-Rad). qPCR was performed using SYBR Green Supermix with iTaqDNA polymerase on the IQ5 thermocycler (Bio-Rad), as described earlier (24) Oligonucleotides (Table 1) were designed, optimized, and efficiently tested before using. For each primer set, the annealing temperature was optimized, and the solitary product formed was confirmed through the melt curve analysis in each PCR run. The PCR program included 36 cycles of heat denaturing at 95 °C for 30 s and annealing at 52–62 °C for 30 s, followed by a melt curve cycle. qPCR data were obtained as CT values, where CT was defined as the threshold cycle of PCR where products amplify exponentially. As an internal control, β-actin expression was measured in parallel. The difference in the CT values (ΔCT) was derived from the specific gene tested and CT of the control gene (β-actin) according to the following equation: 2[CT actin − CT target gene] (24). Final results are presented as relative expression (ΔCT).
Statistical significance was determined by two-tailed Student's t test. The significance of the difference in mean values between more than two groups was evaluated by one-way analysis of variance, followed by Student's t test. All significant differences (p < 0.05) are given in the figures and/or figure legends. Statistical analyses were carried out using Graph Pad Prism 4 (GraphPad Software). Values are expressed as mean ± S.E.
For both genotypes, body and organ weights and blood parameters were consistently higher in male than in female mice (Table 2), and there was no sex-specific impact of VLDLR deficiency on these parameters. The starting body weights (age 6 weeks) of WT and vldlr−/− mice were not significantly different. However, high fat diet feeding resulted in significantly lower body weight gain in female and male vldlr−/− mice than their counterpart WT mice. Consistent with reduced body weight, gonadal, peri-renal, and inguinal fat weights were lower in female (50–57%) and male (40–50%) vldlr−/− mice (Table 2). Absolute liver weights of WT and vldlr−/− mice were comparable, but the ratio liver weight-to-body weight was significantly (p < 0.01) higher in vldlr−/− mice (0.062 ± 0.01) than WT mice (0.045 ± 0.002). Autopsy examination also revealed a clear reduction of fat adjoining muscles of the hindlimb of vldlr−/− mice. In addition, measurements of triglyceride content in muscle lipid extracts showed significant (p < 0.01) reduction of triglyceride content in gastrocnemius muscle of vldlr−/− mice (22.5 ± 1.8 μg/mg protein) compared with that of WT mice (66.8 ± 7.4 μg/mg protein). These results suggest that the magnitude of difference in body weight between vldlr−/− and WT mice was an indication of reduced whole body adiposity, including fat deposition in adipose tissue and muscles. In addition, vldlr−/− mice exhibit a higher level of plasma triglycerides but normal levels of free fatty acids and glucose (Table 2). They also had lower fasting blood insulin and leptin as well as a greater blood glucose clearance (GTT) in response to a bolus of glucose injection (Fig. 1A). Increased whole body insulin sensitivity was also confirmed by ITT showing stronger reduction of blood glucose in response to insulin injection (Fig. 1B). In view of the results of whole body glucose clearance (GTT and ITT), we tested insulin sensitivity in adipose tissue. Insulin-stimulated glucose uptake was about 3-fold higher in adipocytes of vldlr−/− mice compared with WT mice (Fig. 1C) consistent with increased insulin-stimulated phosphorylation of phosphatidylinositol 3-kinase/Akt (Akt) and insulin receptor substrate 1 (IRS1) (Fig. 1, D and E). To determine whether the difference of fat mass between WT and vldlr−/− mice was due to changes of adipocyte cell size (hypertrophy), cell number (hyperplasia), or a combination of both, sections of epididymal adipose tissues were examined (Fig. 1, F–I). The size distribution of adipocytes shows markedly larger adipocytes of WT mice (maximum diameter at 140–160 μm) compared with vldlr−/− adipocytes (maximum diameter at 80–100 μm). However, the number of adipocytes/mm2 was about 30% higher in vldlr−/− than WT mice.
To determine whether protection against high fat diet-induced obesity in vldlr−/− mice paralleled a resistance to inflammation, we evaluated macrophage content and inflammation markers in adipose tissue. In agreement with previous findings (2, 9, 34), staining of adipose tissue section of WT mice with F4/80 showed a high proportion of macrophage accumulation and crown-like structures (17 ± 3 CLS/LPF) (Fig. 2, A and C). Immunostaining of adipose tissue of vldlr−/− mice revealed a substantial reduction of F4/80 stained crown-like macrophage structures (5 ± 2 CLS/LPF) and also the presence of stained macrophages scattered in between adipocytes (Fig. 2, B and D). These results are consistent with the measurement of adipokine production results (Fig. 2E) showing markedly lower secretion of TNFα (−67%), IL-6 (−61%), MCP-1 (−58%), and leptin (−48%) in the culture of adipose tissue explants of vldlr−/− mice compared with those of WT mice. The release of free fatty acids (FFA) was also strongly reduced (−60%) in adipose tissue of vldlr−/− mice. The expression of inflammation markers was also tested in adipocyte fraction and SVF separately. The results presented in Fig. 2F show that expression of Il-6, Tnfα, and Ccl2 (MCP-1), and leptin were about 2–3-fold lower in vldlr−/− adipocytes than WT mice. Similarly, mRNA abundance of Il6, Tnfα, and Ccl2 (MCP-1) was markedly reduced in SVF of vldlr−/− mice (Fig. 2G). These results show that VLDLR deficiency reduced inflammation in both adipocytes and SVF cells.
Macrophage phenotype is a determinant factor in adipose tissue inflammation (9, 34). Therefore, we examined the presence of the pro-inflammatory (M1) and anti-inflammatory (M2) macrophages in adipose tissue of WT and vldlr−/− mice. Histological analysis of ATMs was conducted by immunostaining of the M1 macrophage marker CD11c (Fig. 3, A and B). As expected, adipose tissue of high fat diet-fed WT mice exhibited the presence of a large number of CD11c-positive ATMs, most of which were organized as clusters around adipocytes (Fig. 3A). In adipose tissue of vldlr−/− mice, the number of CD11c-positive ATMs was evidently reduced and mostly scattered without specific localization (Fig. 3B). To quantify the proportion of the M1 and M2 ATMs, SVF isolated from epididymal adipose tissue of WT and vldlr−/− mice were double-stained with F4/80 and CD11c (M1) or F4/80 and CD206 (M2) antibodies and analyzed by flow cytometry. Equal numbers of ATMs of WT and vldlr−/− samples were enumerated, and the proportion of M1 and M2 was calculated (Fig. 3, C–F). From total macrophages (ATMs), the proportion of F4/80 and CD11c double-positive (M1) macrophages was predominant in WT mice (68 ± 9%) (Fig. 3C, top right quadrant), and the proportion of M2 represented a minor fraction (4 ± 1%) (Fig. 3E, top right quadrant). The profile of ATMs in vldlr−/− mice was shifted toward an anti-inflammatory phenotype with a reduced number of F4/80 and CD11c double-labeled ATMs (17 ± 7%) (Fig. 3D, top right quadrant) and increased numbers of F4/80 and CD206 double-positive macrophages (57 ± 8%) (Fig. 3F, top right quadrant). These results are consistent with reduced expression of pro-inflammatory markers TNF-α and IL-6 in the SVF-containing macrophages of vldlr−/− mice shown in Fig. 2G. The expression of additional markers of the M1 and M2 phenotypes was also tested (Fig. 3, G and H). VLDLR deficiency induced significant reduction of the expression of M1 phenotype markers interleukin 1β (Il-1β), inducible nitric-oxide synthase (iNos or Nos2), and Cd11c genes (Fig. 3G) while increasing the expression of the M2 phenotype markers interleukin 10 (Il-10), Cd206 (also called mannose receptor C type 1), and arginase-1 (Arg1) (Fig. 3H). These results show that VLDLR deficiency shifted the ratio of M1/M2 ATMs toward a less inflammatory profile consistent with the results of chemokine production (Fig. 2E) suggesting that reduction of the M1 phenotype among the ATMs of vldlr−/− mice contributed to lower secretion of pro-inflammatory adipokines.
Because high fat diet did not increase macrophage content in adipose tissue of vldlr−/− mice, we questioned if VLDLR expression regulates macrophage chemotaxis. Using the same experimental procedure, we tested the effects of adipose tissue-derived factors on the chemotactic response of WT and vldlr−/− macrophages. Accordingly, peritoneal macrophages were isolated from WT and vldlr−/− mice, and their chemotactic response was tested using conditional medium prepared from cultures of adipose tissues of WT and vldlr−/− mice. As shown in Fig. 4, conditional medium from WT adipose tissue culture (CMWT) produced the highest migration of macrophages of WT and vldlr−/− mice. By contrast, chemotaxis induced by conditional medium of adipose tissue of vldlr−/− mice (CMvldlrko) was markedly lower than that induced by CMWT. These results indicate that factors produced by adipose tissue of WT mice, and not that of vldlr−/− mice, induced stronger chemotaxis response of macrophages independent of vldlr expression in macrophages.
To gain further information about the role of vldlr expression in inflammation, we analyzed the levels of proinflammatory kinases JNK and p38 MAPK in freshly isolated adipocytes. In adipocytes of vldlr−/− mice, the ratio of phosphorylated to total JNK and p38 MAPK was about 45% lower than WT mice (Fig. 5, A and B). We also tested the activity of JNK through the measurement of the JNK reaction product c-Jun phosphorylation, which showed a strong reduction in adipocytes of vldlr−/− mice (Fig. 5, A and B). Because inflammation is linked to endoplasmic reticulum stress (6), we questioned if VLDLR expression also affected ER stress. The levels of phosphorylation of p-PERK and eukaryotic initiation factor 2α (p-eIF2α) as well as the level of CHOP were about 2-fold lower in adipocytes of vldlr−/− mice compared with those of WT mice (Fig. 5, C and D). Moreover, mRNA abundance of genes linked to ER stress, including stress sensors glucose-regulated protein 78 (Grp78), activating transcription factors 4 (Atf4) and 6 (Atf6), ER-resident chaperones protein-disulfide isomerase (Pdi), and endoplasmic reticulum DnaJ homolog 4 (Erdj4), was markedly reduced in vldlr−/− adipocytes (Fig. 5E). Abundance of mRNA of spliced X-box-binding protein 1, a downstream target of ATF6, was also strongly reduced (−76%) (Fig. 5E). These findings indicate that VLDLR deficiency suppressed the induction of inflammation and ER stress by high fat diet feeding.
To examine if VLDLR involvement in adipocyte inflammation is dependent on exogenous lipid, we challenged adipocytes with increasing amounts of VLDL and tested lipid content and inflammation markers. VLDL overload in culture medium strongly increased triglyceride content in WT adipocytes but was less effective in vldlr−/− adipocytes (Fig. 6, A and B). In fact, TG content of WT adipocytes increased in a dose-dependent manner to reach a maximum with the dose of 300 and 400 μg/ml VLDL, which is about 3-fold higher in WT adipocytes than vldlr−/− adipocytes (Fig. 6A). These results are consistent with Oil Red O staining showing lower lipid staining in vldlr−/− adipocytes (Fig. 6B). Interestingly, high doses of VLDL (300 and 400 μg/ml) markedly increased the phosphorylation of JNK and p38 MAPK in WT adipocytes compared with vldlr−/− adipocytes (Fig. 6, C and D). These data are consistent with high expression of IL-6 (Fig. 6E) and MCP-1 (Fig. 6F) in WT adipocytes. These results indicate that VLDL load increased WT adipocyte hypertrophy and simultaneously induced inflammation markers. By contrast, the response of vldlr−/− adipocytes to VLDL load was blunted with lower lipid accumulation and reduced inflammation markers. In light of these results, we questioned if VLDLR mediates VLDL-induced ER stress. Thus, we tested the levels of ER stress markers in adipocytes challenged with VLDL and thapsigargin separately or in combination. Thapsigargin is a pharmacological agent known to rapidly elicit ER stress by inhibiting the ER Ca2+ pump (35). Adipocytes of WT and vldlr−/− responded in a similar fashion to the treatment with thapsigargin showing strong increase of ER stress markers CHOP and phosphorylated eIF-2α (Fig. 7, A and B). These markers were also increased in WT adipocytes challenged with VLDL but were markedly blunted in vldlr−/− adipocytes. The combination of thapsigargin and VLDL induced CHOP and eIF-2a phosphorylation in adipocytes independent of VLDLR expression suggesting that VLDLR is not required for pharmacological induction of ER stress but could mediate VLDL-induced ER stress. It is noteworthy to mention that GLC analysis of VLDL fatty acids showed that the proportion of saturated fatty acids was the highest (59%), although the proportions of monounsaturated and polyunsaturated fatty acids were ~31 and 10%, respectively.
Delivery of VLDL-derived lipids to adipocytes may occur either as FFA liberated locally after the action of LPL or as whole remnant particles. In addition to VLDL-derived lipids, albumin-bound FAs that circulate in the bloodstream represent another source of lipids readily taken by adipocytes without LPL action. To evaluate the role of VLDLR in these pathways, we examine the uptake of VLDL and albumin-bound FAs. Adipocytes were incubated with 3H,125I-double-labeled VLDL to examine the uptake of VLDL-derived lipids as well as remnant particles. The counts of 125I represent the uptake of whole remnant particles, whereas that of 3H corresponds to total uptake, including whole particle and FFA liberated after LPL action. Irrespective of the genotype, the amount of 3H radioactivity was significantly higher than 125I (Fig. 8A). In addition, the amounts of 3H and 125I recovered in vldlr−/− adipocytes were about 3-fold lower than WT adipocytes (Fig. 8A). However, the relative distribution (%) of VLDL-derived 3H into the cellular lipids was similar in WT and vldlr−/− adipocytes (Fig. 8B). These data suggest that VLDLR deficiency reduced the uptake of both VLDL-derived lipids and remnant particles but did not alter incorporation of VLDL-derived FAs into cellular lipids. The role of VLDLR in albumin-bound FA uptake was also examined using [3H]palmitate complexed to albumin. The amount of [3H]palmitate recovered in WT and vldlr−/− adipocytes after 2 h (Fig. 8C) or after 60 s of incubation (results not shown) was comparable. Moreover, incorporation of [3H]palmitate into cellular lipids of WT and vldlr−/− adipocytes was similar (results not shown).
Palmitate is known to induce a pro-inflammatory response in adipocytes (10, 11). We therefore examine the effects of palmitate treatment on WT and vldlr−/− adipocytes. Cells were challenged with various doses of palmitate, and inflammation markers were tested (Fig. 8, D–G). Palmitate loads increased phosphorylation of JNK (Fig. 8, D and E) and expression of IL-6 (Fig. 8F) and MCP-1 (Fig. 8G) in WT and vldlr−/− adipocytes in a similar fashion. These results show that VLDLR deficiency was not a significant factor in palmitate-induced inflammation in adipocytes.
It has been reported that VLDL particles induce lipotoxicity and inflammation in macrophages (29). Thus, we questioned whether VLDLR mediates VLDL-induced inflammation in macrophages. Accordingly, we challenged peritoneal macrophages isolated from WT and vldlr−/− mice with VLDL and examined lipid accumulation and inflammation indicators (Fig. 9). VLDL treatments of WT and vldlr−/− macrophages increased triglyceride content in a dose-dependent manner (Fig. 9A). For each dose of VLDL, the increments of TG contents were consistently higher in WT macrophages than in vldlr−/− macrophages. These results are also confirmed by specific staining of intracellular neutral lipids showing a stronger staining in WT macrophages (Fig. 9B). In addition, expression of Tnfα and Il6 (Fig. 9C) and the levels of phosphorylation of JNK and p38 MAPK (Fig. 9, D and E) were markedly reduced in macrophages of vldlr−/− mice compared with those of WT mice. These results indicate that VLDLR deficiency reduced lipid accumulation and inflammation in macrophages.
Infiltration of macrophages into adipose tissue promotes interactions between macrophages and adipocytes, and both types of cells act in a synergistic manner to enhance adipose tissue inflammation (3, 36). Investigations using in vitro co-culture systems have demonstrated reciprocal impacts of adipocytes and macrophages and also highlighted the contributions of these cells in inflammation (3, 36). To determine the impact of VLDLR expression on the interactions between cells, adipocytes and macrophages were cultured in Transwell indirect co-culture systems with VLDL treatments. Preadipocytes isolated from WT (ADWT) and VLDLR-deficient (ADvldlrko) mice were differentiated in vitro to reach maturity, a stage at which they were co-cultured with peritoneal macrophages collected from WT (PMWT) and VLDLR-deficient (PMvldlrko) mice. Adipocytes and macrophages cultured separately (single cultures) were used as controls. Production of IL-6 and MCP-1 was tested in the culture medium, and expression of inflammatory markers was measured in adipocytes grown in the underlying wells. Production of IL-6 and MCP-1 was strongly reduced in single cultures of adipocytes (AD) and macrophages (PM) of vldlr−/− mice (Fig. 10, A and B, open bars). For each adipokine, the sum of production by adipocytes and macrophages in single cultures was calculated and presented as AD + PM to compare with co-cultures as indicated below. The results generated from separate cultures of adipocytes and peritoneal macrophages showed that VLDLR deficiency regulates inflammation in both types of cells. So the next question was to examine whether VLDLR expression also modulates cross-talk between adipocytes and macrophages. To this end, we performed co-cultures of peritoneal macrophages and adipocytes with mixed genotypes, and we measured production and expression of inflammatory factors. As shown in Fig. 10, C and D, secretion rates of IL-6 and MCP-1 were the highest in co-cultures containing WT adipocytes and macrophages (ADWT − PMWT, black bars) and the lowest in co-cultures containing vldlr−/− adipocytes and macrophages (ADvldlrko − PMvldlrko, open bars). In the co-cultures containing a mix of cell phenotypes (PMvldlrko − ADwt and PMwt − ADvldlrko, gray bars), IL-6 and MCP-1 secretion rates were intermediate between the lowest and highest values. In addition to the results reported here, we also tested the production of TNF-α in the same culture medium. The results for TNF-α were similar to those reported for IL-6 and MCP-1; TNFα production was the highest in co-cultures of WT cells (ADWT − PMWT) and the lowest in co-cultures of vldlr−/− cells (ADvldlrko − PMvldlrko). Comparisons between the results of singles cultures and co-cultures are possible because experimental conditions of both systems were comparable, including an equal number of cells and similar conditions of culture. For comparative purposes, the data generated from single and co-culture systems were plotted with the same scale. This presentation allows straightforward comparison between the sum of production (AD + PM) of single cultures (Fig. 10, A and B) and the results of co-cultures (Fig. 10, C and D). Interestingly, the levels of production of IL-6 and MCP-1 were strikingly higher (+71 and +104%, respectively) in ADWT − PMWT co-cultures (black bars on right side of Fig. 10, C and D) than the sum of production (AD + PM) of WT cells in single cultures. By contrast, the difference between co-cultures of ADvldlrko − PMvldlrko and the sum of production by vldlr−/− cells in single cultures (AD + PM) was noticeably smaller (+44 and +35% for IL-6 and MCP-1, respectively (open bars on right side of Fig. 10, C and D). These results suggest that VLDLR expression promoted synergistic inflammation response in adipocyte-macrophage co-cultures. We also questioned whether the cell genotype-related difference in chemokine production was an indication of intracellular inflammation. Accordingly, we tested the activity of pro-inflammatory kinase JNK in adipocytes (Fig. 10E). Western blot analysis revealed a high level of JNK phosphorylation in ADWT co-cultured with PMWT macrophages compared with those cultured with PMvldlrko and vldlr−/− adipocytes cultured with PMWT or PMvldlrko. Moreover, analysis of IL-6 expression with qPCR shows stronger mRNA abundance in ADWT co-cultured with PMWT compared with ADWT co-cultured with PMvldlrko (Fig. 10F). Il6 expression in ADvldlrko co-cultured with PMWT or PMvldlrko was clearly lower than that in ADWT. These results indicate that VLDLR deficiency strongly reduced inflammation in adipocytes in response to macrophage signals.
In these studies, we asked the question whether VLDLR plays a role in high fat diet-induced inflammation in adipose tissue. The results revealed that vldlr deficiency reduced inflammation and ER stress in adipose tissue, concomitantly with reduced adiposity and insulin resistance. Moreover, adipocytes lacking vldlr are less prone to VLDL-induced inflammation and had a lower ability to induce in vitro macrophage chemotaxis. Finally, VLDLR deficiency reduced the synergistic inflammatory interactions between macrophages and adipocytes.
Numerous studies (1, 2, 34) have shown that obesity induces local inflammation in adipose tissue and that cells of the innate immune system, particularly macrophages, are crucially involved in adipose tissue inflammation and metabolic dysfunction. In this study, we provide new findings showing that VLDLR deficiency reduced inflammation and ER stress in adipose tissue and simultaneously decreased adipose tissue macrophages, especially those expressing markers of the M1 phenotype (Fig. 3). In agreement with previous reports (14, 15, 17), our in vivo data also showed that mice lacking vldlr are protected from high fat diet-induced obesity. These findings open the question whether reduced inflammation in adipose tissue of vldlr−/− mice was secondary to the protection from obesity or a reflection of VLDLR interaction with inflammatory pathways. Accordingly, we examined the inflammatory response of adipocytes challenged with VLDL overload. The results show that VLDLR deficiency strongly decreased VLDL-induced lipid accumulation and simultaneously reduced inflammation and ER stress (Figs. 6 and and7).7). VLDLR deficiency did not, however, mitigate pharmacological induction of ER stress as seen in adipocytes treated with thapsigargin (Fig. 7). These results suggest that induction of inflammation is linked with excess lipid accumulation and adipocyte hypertrophy but not as a separate action of VLDLR on the ER stress pathways. Moreover, comparative results of adipocytes cultured with labeled VLDL particles or albumin-bound palmitate show that VLDLR deficiency reduced the uptake of both VLDL-derived fatty acids and remnant particles (Fig. 8A) but did not alter the uptake of albumin-bound palmitate (Fig. 8C). Previously, we have shown that adipocyte uptake of oleic acid and expression of CD36 and fatty acid transport protein, proteins known to be involved in FA uptake, was not affected by VLDLR deficiency (14). Moreover, VLDLR deficiency is associated with reduced LPL activity and lipolysis of TG-rich lipoproteins (12, 16). Taken together, these results suggest that FA uptake per se was not compromised by VLDLR deficiency. Reduction of TG lipolysis and remnant particle uptake was rather the primary cause of lower recovery of VLDL-derived lipids in adipocytes. These results are also in agreement with previous work showing that in vivo clearance of blood TG-rich lipoproteins was delayed in vldlr−/− mice (14, 13, 16) and that recovery of chylomicrons and VLDL-derived lipids in adipose tissue was markedly reduced (14). Altogether, these findings show that VLDLR impacts TG-rich lipoprotein catabolism at different levels. First, VLDLR is required for optimal functioning of LPL either by increasing LPL and TG-rich lipoprotein interactions at the capillary surface (12) or by serving as a helper for the transcytosis of LPL (18). Second, VLDLR binds and internalizes TG-rich lipoprotein particles through specific binding of apolipoprotein E, as has been shown in isolated cells (13, 14). Through these actions, VLDLR plays a prominent role in lipid uptake and adipocyte hypertrophy. Ultimately, adipocytes loaded with lipids exceeding its ability of storage or oxidation display signs of stress and inflammation (10, 11, 37). Of note, silencing apoE, which is the natural ligand of VLDLR, reduces diet-induced adipose tissue expansion and inflammation (38, 39). Moreover, apoE in VLDL increases VLDL uptake (38) and adipocyte hypertrophy (40). Similarly, suppression of apoE expression in adipocyte is associated with a reduction of adipocyte size and decline of VLDLR expression, suggesting a link between reduced VLDL uptake and VLDLR expression (41, 42). The expressions of apoE as well as that of LDLR-related protein and LDLR, receptors that could bind VLDL, were not altered by VLDLR deficiency (results not shown). Therefore, the changes of lipid uptake and inflammation in adipocytes reported in this study are linked to vldlr expression and not to these proteins.
With the onset of obesity, expansion of adipose tissue is accompanied by the accumulation of macrophages (2, 43), and some reports have proposed a primary role for adipocyte-derived mediators in the initiation of macrophage migration (9, 44). In these studies, we found that vldlr deficiency greatly reduced adipocyte size and simultaneously reduced ATM content. This is also consistent with the low production of chemokines in adipose tissue of vldlr−/− mice (Fig. 2). In addition, the in vitro data presented in Fig. 4 show that chemotactic responses of WT and vldlr−/− macrophages were similar and that conditional medium from WT adipose tissue, and not that from vldlr−/− adipose tissue, markedly increased chemotaxis of macrophages. These results corroborate the hypothesis that signals from adipose tissue play a major role in obesity-induced macrophage accumulation. In addition to adipose tissue-derived signals, circulating factors in the blood have been suspected to initiate macrophage activation and migration (30, 45, 46). In this regard, the question of whether plasma triglycerides could affect monocyte/macrophage activation and promote their migration has been the subject of previous investigations (44, 47, 48), and there is some evidence in the literature suggesting that blood lipids could be linked to ATM accumulation. In fact, plasma triglyceride levels have been shown to positively correlate with ATM markers in humans (49). Contrary to these reports, hypertriglyceridemia in the vldlr−/− mouse was not associated with increased ATM content. However, our results are in agreement with a previous study in LDLR-null mice showing that obesity and not hyperlipidemia correlates with adipose tissue macrophage content (50). Although VLDLR deficiency did not influence the chemotactic response of peritoneal macrophages to adipose tissue signals, it did reduce the inflammatory response of peritoneal macrophages to VLDL load (Fig. 9). Therefore, we cannot exclude the possibility that VLDLR expression could impact macrophage polarization at later stages after migration into adipose tissue.
Our studies also provide new evidence that VLDLR modulates the expression of inflammatory markers in macrophages and adipocytes in single and co-culture systems. In fact, the expressions of chemokines Il6 and Ccl2 (MCP-1) were markedly higher (+71 and 104%) in the co-culture system of adipocytes and macrophages of WT mice (Fig. 10, C and D) than WT cells cultured separately (Fig. 10, A and B). In co-cultures of adipocytes and macrophages of vldlr−/− mice, the rate of chemokine secretion was only 35–44% above the sum of vldlr−/− adipocytes and macrophages single cultures. These results indicate that VLDLR deficiency not only reduced inflammation in adipocytes and macrophages but also markedly diminished the synergistic interactions between adipocytes and macrophages. At present, it is not clear how VLDLR expression regulates the interactions between these cells. One possibility could be linked to the simultaneous effect of VLDLR in both adipocytes and macrophages. Previous studies (3), using similar co-culture models, have demonstrated that a paracrine loop involving pro-inflammatory mediators derived from adipocytes and macrophages act on both types of cells and amplify the inflammatory changes. In this study, excess chemokines produced by WT cells in co-cultures most likely act on both types of cells through a paracrine loop leading to the activation of multiple signaling pathways that exacerbate inflammation. By contrast, VLDLR deficiency reduces inflammation in both adipocytes and macrophages and consequently prevented additive effects of inflammatory mediators.
In conclusion, our findings show that VLDLR expression plays a prominent role in high fat diet-induced adipose tissue inflammation. Moreover, VLDLR mediates VLDL-induced lipid accumulation and induction of inflammation in adipocytes and macrophages.
*This work was supported by development grants from Hackensack University Medical Center and American Heart Association Award AHA0730356N.
2The abbreviations used are: