Loss of macrophage fatty acid oxidation does not potentiate systemic metabolic dysfunction

Abstract

Fatty acid oxidation in macrophages has been suggested to play a causative role in high-fat diet-induced metabolic dysfunction, particularly in the etiology of adipose-driven insulin resistance. To understand the contribution of macrophage fatty acid oxidation directly to metabolic dysfunction in high-fat diet-induced obesity, we generated mice with a myeloid-specific knockout of carnitine palmitoyltransferase II (CPT2 Mϕ-KO), an obligate step in mitochondrial long-chain fatty acid oxidation. While fatty acid oxidation was clearly induced upon IL-4 stimulation, fatty acid oxidation-deficient CPT2 Mϕ-KO bone marrow-derived macrophages displayed canonical markers of M2 polarization following IL-4 stimulation in vitro. In addition, loss of macrophage fatty acid oxidation in vivo did not alter the progression of high-fat diet-induced obesity, inflammation, macrophage polarization, oxidative stress, or glucose intolerance. These data suggest that although IL-4-stimulated alternatively activated macrophages upregulate fatty acid oxidation, fatty acid oxidation is dispensable for macrophage polarization and high-fat diet-induced metabolic dysfunction. Macrophage fatty acid oxidation likely plays a correlative, rather than causative, role in systemic metabolic dysfunction.

infiltration of macrophages into adipose tissue following high-fat diet-induced obesity has been shown to potentiate systemic metabolic dysfunction. Macrophages can contribute to insulin resistance by creating a low-grade chronic inflammatory state induced by secretion of proinflammatory cytokines that inhibit insulin signaling (14, 15, 42). Macrophages can also polarize into distinct phenotypes with characteristic inflammatory states, although these demarcations are less pronounced in human macrophages. In obese and diabetic models, adipose tissue exhibits increased macrophage infiltration, and these cells exhibit a more classically activated (M1) program (28). Macrophage polarization programs are commonly characterized by the distinct metabolic phenotype(s) to which they adapt to carry out their cellular functions. For example, greater anaerobic glucose utilization is a canonical feature of M1-polarized macrophages, while fatty acid β-oxidation has been implicated in the polarization of macrophages to the alternative (M2) phenotype (1). IL-4-stimulated M2 polarization, for instance, increases fatty acid oxidation and the corresponding metabolic gene program (43). Pharmacological inhibitors of fatty acid oxidation, such as the promiscuous epoxide etomoxir, inhibit M2 macrophage polarization (31). One implicated role of fatty acid oxidation in macrophage function is support of the high energy demands of macrophage phagocytosis, since fatty acids released via lipolysis are routed to the mitochondria for fatty acid β-oxidation (2). Others have shown that lysosomal lipolysis is required to facilitate fatty acid oxidation, since inhibition of lipolysis suppressed the M2 program (19). Elevated fatty acid levels in macrophages are also known to play a role in fatty acid-induced inflammation, since it has been shown that fatty acid-induced inflammation can be counteracted by increased triacylglycerol synthesis, reduced lipid transport, or activation of fatty acid oxidation (7, 9, 21).

Mitochondrial long-chain fatty acid β-oxidation requires the transfer of acyl-coenzyme As (acyl-CoAs) from the cytoplasm and into the mitochondrial matrix and occurs via the coordinated action of two enzymes: carnitine palmitoyltransferase (CPT) I and II. CPT1, the rate-limiting enzyme in mitochondrial long-chain fatty acid β-oxidation, generates acylcarnitines that can traverse the mitochondrial membrane. Once inside, CPT2 will generate acyl-CoAs from acylcarnitines to initiate the β-oxidation of long-chain fatty acids to acetyl-CoA. The balance of fatty acid synthesis and fatty acid oxidation is regulated by the rate-determining metabolite in de novo fatty acid synthesis, malonyl-CoA, which can directly inhibit CPT1 activity.

To assess the role of fatty acid β-oxidation in both macrophage polarization and diet-induced insulin resistance and inflammation in obesity, we generated a genetic model with a myeloid-specific deletion of CPT2 (Cpt2), an obligate step in mitochondrial long-chain fatty acid β-oxidation. We found that IL-4-stimulated M2 polarization of bone marrow-derived macrophages (BMDM) did, indeed, increase fatty acid oxidation, in support of this metabolic hallmark of alternatively activated macrophages. However, mitochondrial fatty acid β-oxidation was not required for polarization to the alternative phenotype, as BMDM from mice with a myeloid-specific knockout (KO) of CPT2 (CPT2 Mϕ-KO mice) displayed canonical markers of M2 polarization following IL-4 stimulation. Additionally, we showed that loss of macrophage fatty acid oxidation in vivo does not alter the progression of high-fat diet-induced inflammation, oxidative stress, or glucose intolerance. These data suggest that although IL-4-stimulated alternatively activated macrophages upregulate fatty acid oxidation, fatty acid oxidation is dispensable for macrophage polarization and high-fat diet-induced metabolic dysfunction. We conclude that macrophage fatty acid oxidation likely plays a correlative, rather than causative, role in systemic metabolic dysfunction.

Experimental Procedures

Animals and diets.

To generate a myeloid lineage-specific loss-of-function of Cpt2, we bred Cpt2^lox/lox mice to lysozyme2-Cre transgenic mice (3, 31). Mice were housed in ventilated racks with a 14:10-h light-dark cycle and fed a standard chow diet (Etruded Flobal Rodent Diet, Harlan Laboratories). For the diet study, mice were fed a 10% low-fat diet (D12450J, Research Diets) or a 60% high-fat diet (D12492, Research Diets) from 6 to 18 wk of age (12 wk on the diet). At week 10 of the diet, mice were subjected to a glucose tolerance test (GTT): glucose (0.75 g/kg ip) was injected, and tail blood glucose was measured at 0, 15, 30, 60, and 120 min. At week 11 of the diet, an insulin tolerance test (ITT) was performed: insulin (0.6 U/kg ip) was injected, and tail blood glucose was measured at 0, 15, 30, 60, and 90 min (Nova Max Plus). Body weight was measured on a weekly basis. Body fat and lean mass of 18-wk-old mice were measured via magnetic resonance imaging analysis (Minispec MQ10). Liver, spleen, and gonadal white adipose tissue (gWAT) were collected from all mice and frozen in liquid nitrogen. Serum was also collected from all mice, and free glycerol and triacylglycerol (Sigma), β-hydroxybutyrate (StanBio), total cholesterol (Wako), and nonesterified fatty acid (Wako) were measured colorimetrically. All procedures were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the approval of the Johns Hopkins Medical School Animal Care and Use Committee.

Cell culture.

BMDM were obtained from bone marrow isolated from femurs and tibias of control or CPT2 Mϕ-KO mice and cultured on petri dishes in RPMI medium supplemented with 20% fetal bovine serum, 10 ng/ml recombinant macrophage colony-stimulating factor (eBioscience), 1% penicillin-streptomycin, and 2 mM l-glutamine. Differentiation of macrophages occurred over 7 days, and cells were plated on day 7 or 8 for use in downstream applications. For classical or alternative macrophage activation, BMDM were stimulated with murine IFNγ (100 U/ml; PeproTech) + LPS (5 ng/ml; Sigma) or murine IL-4 (10 ng/ml) for 24 h, respectively. Transfection of 2.5 μg/ml polyinosinic:polycytidylic acid [poly(I:C)] was complexed at a ratio of 2:1 Lipofectamine 2000 to nucleic acid for 2 h. Free fatty acid (FFA, 100 μM) was conjugated to BSA at a 2:1 molar ratio (oleate-to-palmitate) for 2 h.

Metabolic flux assays.

For [1-¹⁴C]oleate oxidation, 5 × 10⁵ control or CPT2 Mϕ-KO BMDM were placed in T-25 flasks. After 24 h of IL-4 pretreatment, BMDM were supplemented with 0.1 mCi of [1-¹⁴C]oleate, 100 mM l-carnitine hydrochloride (Sigma), and 0.2% BSA. For [U-¹⁴C]glucose oxidation, 1 × 10⁶ control or CPT2 Mϕ-KO BMDM were placed in T-25 flasks. After 24 h of pretreatment with IL-4, cells were supplemented with 0.1 μCi of [U-¹⁴C]glucose, 1.25 mM glucose, 2 mM l-glutamine, and 0.25 mM sodium pyruvate in glucose-free DMEM. The flasks were sealed with a rubber stopper containing a hanging well filled with filter paper and incubated at 37°C for 4 h. CO₂ was trapped by addition of 150 μl of 1 M NaOH to the filter paper in the center of the well and 200 μl of 1 M perchloric acid to the reaction mixture. The samples were then incubated at 55°C for 1 h, the filter paper was placed in scintillation fluid, and counts were normalized to total protein.

To determine acetate incorporation into total lipid, BMDM were plated in 24-well dishes. After 24 h of pretreatment with IL-4, BMDM were labeled with 0.3 μCi/ml [³H]acetate in RPMI medium for 4 h. Total lipids were extracted with 2:1 chloroform-methanol via the Folch method, and radioactivity was counted by liquid scintillation. Counts were normalized to total protein.

To measure glucose uptake, BMDM were plated in 24-well dishes. After 24 h of pretreatment with IL-4, BMDM were incubated for 2 h in Krebs-Ringer-HEPES (KRH) buffer (20 mM HEPES, 136 mM NaCl, 4.7 mM KCl, 1.25 mM MgSO₄, 1.25 mM CaCl₂, and 0.1% BSA) and then labeled with 0.5 μCi of [1,2-³H]2-deoxyglucose, along with a final concentration of 6.5 mM 2-deoxyglucose, for 10 min. After incubation, cells were washed four times with KRH buffer and lysed in 0.3 ml of 1× PBS + 1% Triton X-100. Total lysate was added to the scintillation fluid, and counts were normalized to total protein.

Western blotting.

Control or CPT2 Mϕ-KO BMDM were collected after IL-4 stimulation and homogenized with RIPA buffer (50 mM Tris·HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, and 0.25% deoxycholate) with protease inhibitor cocktail (Roche), and the insoluble debris was pelleted at 13,000 g for 15 min at 4°C. The protein concentrations of lysates were determined by BCA assay (Thermo Scientific), and 30 μg of lysate were separated by Tris-glycine SDS-PAGE (10% polyacrylamide). Proteins were transferred to nitrocellulose membranes (Protran BA 83, Whatman), blocked in 3% BSA-TBST (Tris-buffered saline with Tween 20), and incubated with primary antibodies overnight. The blots were probed with the following antibodies: Cpt2 (catalog no. PA5-12217, Thermo Scientific), Hadha (catalog no. GTX101177, GeneTex), Acot1 (catalog no. ab133948, Abcam), Pgc1a (catalog no. ab54481, Abcam), Acot2 (catalog no. SAB2100030, Sigma), Acsl1 (catalog no. 4047, Cell Signaling Technology), Acot7 [affinity-purified antibody (5)], Aco2 (catalog no. 6922, Cell Signaling Technology), Mcad (catalog no. GTX32421, GeneTex), Atp5a, Mtco1, and Uqcrc2 (MitoProfile total OXPHOS; catalog no. ab110413, Abcam), Atp5a (catalog no. ab14748, Abcam), and Hsc70 (catalog no. sc-7298, Santa Cruz Biotechnology). Cy3-conjugated anti-mouse (Invitrogen) or HRP-conjugated anti-mouse (GE Healthcare) or anti-rabbit (GE Healthcare) secondary antibodies were used appropriately.

For CPT2 protein, mitochondria from BMDM were isolated as follows. Control and CPT2 Mϕ-KO BMDM were plated, each on a 10-mm² tissue culture plate. On the following day, medium was removed, and cells were washed with 10 ml of 1× PBS. On removal of 1× PBS, 5 ml of ice-cold IB_c buffer (10 mM Tris-MOPS, pH 7.4, 1 mM EGTA-Tris, pH 7.4, and 0.2 M sucrose) were added to each plate, and cells were dislodged from the plate by scraping. Cell suspensions were transferred to a 7-ml glass homogenizer (pestle A) that had been chilled on ice and homogenized with 45 strokes total. The homogenate was transferred to a centrifuge tube and spun at 600 g for 10 min at 4°C. After centrifugation, the supernatant was collected and transferred to a new tube, and the remaining pellet was resuspended in 5 ml of fresh ice-cold IB_c buffer and then mixed on a vortex for 30 s on the high setting. The resuspended pellet was returned to the tube containing the supernatant from the first spin to yield a total volume of 10 ml, and the contents of the tube were centrifuged at 600 g for 10 min at 4°C. The supernatant was collected and transferred into a new tube and spun at 7,000 g for 10 min at 4°C, while the pellet containing unbroken cells and nuclei was discarded. The supernatant was discarded, and the pellet was washed with 200 μl of ice-cold IB_c buffer and then transferred to a microcentrifuge tube and spun at 7,000 g for 10 min at 4°C. The supernatant was discarded, and the pellet was resuspended in 200 μl of ice-cold IB_c buffer and spun at 10,000 g for 10 min at 4°C. The supernatant was discarded, and the pellet was resuspended in 60 μl of RIPA buffer with protease inhibitor cocktail, and the insoluble debris was pelleted at 13,000 g for 15 min at 4°C. The protein concentrations of lysates were determined by BCA assay (Thermo Scientific).

Steady-state metabolite analysis with ¹H-NMR spectroscopy.

Culture medium was removed, and 1 × 10⁶ cells were quenched and scraped in 1 ml of ice-cold methanol, transferred to a 2-ml Eppendorf tube, snap-frozen in liquid nitrogen for 10 min, thawed on ice for 10 min, and vortexed at 12,000 g for 10 min at 4°C. Supernatant was transferred to a 2-ml Eppendorf tube, and the pellet was stored at −80°C for protein quantification. Freeze-thaw and vortex cycles were repeated twice: first with 500 μl of methanol and then with 400 μl of 80:20 methanol-water. Supernatants from each step were collected in the same tube and dried in a Speed-Vac overnight. Dried samples were reconstituted with 20 mM phosphate buffer containing 0.1 mM trimethylsilylpropanoic acid (TMSP) as an internal reference and 0.1 mM NaN₃, and pH was adjusted to 7.4 ± 0.1.

¹H spectra of the cell extracts were recorded on a NMR spectrometer (Avance III 500-MHz, Bruker Instrument, Mannheim, Germany) operating at 499.9 MHz and equipped with a room temperature quadruple nuclei probe. Typical ¹H spectra were acquired using presaturation solvent suppression pulse sequence (noesyprld). Acquisition parameters were set as follows: spectral width of 8012.820 with 64,000 data points, 512 scans, with a relaxation delay of 7 s for a total collection time of 1.14 h. Samples were automatically tuned and matched and were shimmed to the TMSP signal. Spectra were exported into Bruker format and processed with Chenomx NMR Suite 8.2 Professional (Chenomx, Edmonton, AB, Canada). The TMSP signal (0.0 ppm) was used as a reference peak, spectra were manually phase-corrected, and spline function was applied for the baseline correction. Metabolites were profiled and quantified using a built-in Chenomx 500-MHz library. Data were exported to an Excel sheet for further data analysis.

Flow cytometry.

Adipose tissue was prepared for flow cytometry as previously described (30). Antibodies used for flow cytometric analysis are as follows: CD301 PerCP (Biolegend), CD206 FITC (Biolegend), CD11b AF700 (BD Biosciences), Siglec-F PE-CF594 (BD Biosciences), CD45 BV510 (BD Biosciences), F4/80 BV786 (BD Biosciences), Ly6G BV421 (BD Biosciences), MHC II BV605 (BD Biosciences), TNFα FITC (Biolegend), and IFNγ PE (BD Biosciences).

Analysis of gene expression.

Total RNA was isolated using TRIzol reagent followed by the RNeasy Mini Kit (Qiagen). RNA was reverse-transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosciences). The cDNA was diluted to 2 ng/ml and amplified by specific primers in a 20-μl reaction using SsoAdvanced SYBR Green Supermix (Bio-Rad). Gene expression was analyzed in a real-time PCR detection system (CFX Connect, Bio-Rad). For each gene, mRNA expression was calculated as 2^ΔCT relative to ribosomal protein L22, Gapdh, or cyclosporine A expression.

Oxidative damage.

Thiobarbituric acid-reactive substances (TBARS) were determined in serum, gWAT, and liver from high-fat diet-fed CPT2 Mϕ-KO and control male littermate mice. The gWAT and liver were prepared by homogenization of 25 mg of tissue in 1× RIPA buffer with protease inhibitors (Complete Mini, Roche). Homogenates were centrifuged at 1,600 g for 10 min, and 25 μl of the supernatant were used to determine oxidative damage via a TBARS assay kit (Cayman Chemical).

Statistical analyses.

Pair-wise comparisons were calculated using a two-tailed Student’s t-test. Multiple comparisons were calculated using ordinary two-way ANOVA. Two-way ANOVA with repeated measures was utilized for GTT and ITT data.

RESULTS

Generation of mice with a macrophage-specific loss of mitochondrial fatty acid β-oxidation.

Previously, we generated a mouse model with a conditional loss-of-function allele for Cpt2, an obligate step in mitochondrial long-chain fatty acid β-oxidation (23–25). To produce mice with a Cpt2 loss-of-function in macrophages, we bred Cpt2^lox/lox mice to myeloid-specific lysozyme2-Cre transgenic mice (3, 31). The resulting myeloid-specific Cpt2 KO (CPT2 Mϕ-KO) mice showed decreased mRNA and protein levels of Cpt2 in BMDM (Fig. 1, A and B). Despite reduced Cpt2 levels, we did not observe changes in expression of various mitochondrial proteins in control or IL-4-stimulated CPT2 Mϕ-KO BMDM (Fig. 1C). Consistent with the notion of increased fatty acid oxidation as a metabolic hallmark of alternatively activated macrophages, we observed a ~2.5-fold increase in oxidation of [1-¹⁴C]oleic acid in control BMDM in the presence of IL-4 stimulation. As expected, CPT2 Mϕ-KO BMDM showed a suppressed ability to oxidize oleic acid in the presence and absence of IL-4 stimulation (Fig. 1D). In addition, CPT2 Mϕ-KO BMDM oxidized [U-¹⁴C]glucose to the same extent as control BMDM in the absence of IL-4 stimulation. IL-4-stimulated CPT2 Mϕ-KO BMDM showed slightly elevated, but not statistically significant, levels of glucose oxidation (Fig. 1E). On measurement of [³H]acetate incorporation into the total lipid fraction, we observed no alternations in de novo fatty acid synthesis in the presence and absence of IL-4 (Fig. 1F). To determine changes in glucose uptake between control and CPT2 Mϕ-KO BMDM, we used [1,2-³H]2-deoxyglucose. CPT2 Mϕ-KO BMDM showed no alternations in glucose uptake in IL-4-stimulated and unstimulated conditions (Fig. 1G). In summary, these data demonstrate generation of mice with a defect in macrophage mitochondrial fatty acid oxidation and show that loss of fatty acid oxidation does not significantly alter glucose uptake and oxidation or de novo fatty acid synthesis in IL-4-stimulated CPT2 Mϕ-KO BMDM.

Fig. 1.

Generation of mice with a macrophage-specific loss of mitochondrial fatty acid β-oxidation. A: mRNA for carnitine palmitoyltransferase (CPT) II (Cpt2) in bone marrow-derived macrophages (BMDM) from control [wild-type (WT)] mice and BMDM from mice with a specific knockout of Cpt2 [CPT2 Mϕ-KO (KO), n = 3]. B: Western blot for Cpt2 in control and CPT2 Mϕ-KO BMDM. C: Western blot of mitochondrial proteins in control and CPT2 Mϕ-KO BMDM in the presence and absence of IL-4 stimulation (n = 2). D: oxidation of [1-¹⁴C]oleic acid to ¹⁴CO₂ in control and CPT2 Mϕ-KO BMDM (n = 6). E: oxidation of [U-¹⁴C]glucose to ¹⁴CO₂ in control and CPT2 Mϕ-KO BMDM (n = 6). F: incorporation of [³H]acetate into lipid in control and CPT2 Mϕ-KO BMDM (n = 6). G: [1,2-³H]2-deoxyglucose uptake in control and CPT2 Mϕ-KO BMDM (n = 6). H: steady-state metabolite concentrations in IL-4- or LPS-IFNγ-stimulated control and CPT2 Mϕ-KO BMDM (n = 5–6). Values are means ± SE. *P < 0.05, ***P < 0.001 for genotype comparison. #P < 0.05, ###P < 0.001 for treatment comparison.

To gain more insight into the role of fatty acid oxidation in stimulus-induced macrophage metabolism, we turned to ¹H-NMR-based metabolomics. We assessed the steady-state concentrations of metabolites in an unbiased fashion in control and CPT2 Mϕ-KO BMDM in the naïve state or treated with IL-4 or LPS-IFNγ. Consistent with a switch to anaerobic glucose metabolism, which is characteristic of LPS-IFNγ-induced M1 macrophages, we observed a significant increase in steady-state lactate concentrations in control and CPT2 Mϕ-KO BMDM treated with LPS-IFNγ, but not in IL-4-treated BMDM (Fig. 1H). However, we did not observe genotypic differences within any treatment, suggesting that the loss of macrophage fatty acid oxidation does not significantly alter the steady-state concentrations of metabolites in central carbon metabolism. These data show that macrophages can greatly alter their metabolic program, depending on external stimuli; however, a defect in mitochondrial fatty acid oxidation does not significantly alter these basic metabolic phenotypes in cytokine-stimulated BMDM.

Macrophage mitochondrial fatty acid β-oxidation is not required for alternative activation in vitro.

Previously, we showed that etomoxir, a commonly used pharmacological inhibitor of fatty acid oxidation, is capable of blocking the M2 polarization of control and CPT2 Mϕ-KO BMDM (31). To further understand the role of macrophage fatty acid β-oxidation in M2 and M1 macrophage polarization, we determined the expression of canonical macrophage polarization genes after 24 h of stimulation with IL-4 and LPS-IFNγ in control and CPT2 Mϕ-KO BMDM. IL-4-treated CPT2 Mϕ-KO BMDM displayed no significant changes in the M2 polarization markers Mgl2, Retnla, and Il-10, while Arg1 expression levels were slightly increased by ~1.5-fold. In addition, there were no increases in mRNA levels of the M1 markers Tnf-a, Nos2, Mcp1, and Il-6 in IL-4 alternatively activated CPT2 Mϕ-KO BMDM, suggesting that loss of macrophage fatty acid oxidation does not drive them to exhibit a more classically (M1) activated program. LPS-IFNγ-treated CPT2 Mϕ-KO BMDM showed slight (<1.5-fold) increases in expression of the M1 polarization markers Tnf-a and Nos2 but no significant changes in Mcp1 or Il-6 levels (Fig. 2A). We further assessed expression levels of various genes involved in the fatty acid synthesis and oxidation pathways in the presence or absence of IL-4 or LPS-IFNγ, as we previously observed significant misregulation of these genes in other models of Cpt2 deficiency (23–25). No genotypic differences within treatments were observed in the fatty acid synthesis genes Acc1, Fasn, and Scd1 or in the transcriptional regulators Srebf1, Pparα, and Pgc1α (Fig. 2B). Among fatty acid oxidation genes, both Cpt1a and Hadha were upregulated by <1.5-fold in IL-4-stimulated CPT2 Mϕ-KO BMDM (Fig. 2C). Finally, we assessed the gene expression of various oxidative stress markers and found that only Cox2, which is known to produce inflammatory mediators, was increased nearly twofold in LPS-IFNγ-stimulated CPT2 Mϕ-KO BMDM (Fig. 2D). Together these data suggest that LPS-IFNγ-stimulated CPT2 Mϕ-KO BMDM exhibit a small potentiation of the inflammatory response in vitro, but such a response is not observed in IL-4 alternatively activated CPT2 Mϕ-KO BMDM. Additionally, these data show that IL-4-stimulated M2 polarization can occur in CPT2 Mϕ-KO BMDM, despite a significant suppression of their ability to undergo mitochondrial fatty acid oxidation.

Fig. 2.

Macrophage mitochondrial fatty acid β-oxidation is not required for alternative activation in vitro. A–D: quantitative RT-PCR analysis of M1 and M2 macrophage markers, fatty acid synthesis and transcriptional regulator genes, fatty acid oxidation genes, and oxidative stress genes in control and CPT2 Mϕ-KO BMDM in the presence and absence of IL-4 or LPS-IFNγ. Values are means ± SE (n = 6). *P < 0.05, **P < 0.01, ***P < 0.001.

Loss of macrophage fatty acid oxidation does not affect body composition.

It has previously been shown that macrophage-specific deletions of metabolic regulators and inflammatory mediators can result in physiological alterations in body composition (20, 33, 34). Changes in body composition were assessed in CPT2 Mϕ-KO male mice fed a low- or high-fat diet for 12 wk. Body mass, fat mass, and lean mass of CPT2 Mϕ-KO mice were comparable to littermate controls fed either diet (Fig. 3, A and B). To assess changes in adiposity, we examined the wet weight of inguinal and gonadal white adipose depots and, in addition, assessed changes in liver wet weight, since the liver contains its own subset of resident macrophages, i.e., Kupffer cells. No changes in white adipose depots or liver wet weights were observed in CPT2 Mϕ-KO mice fed either diet (Fig. 3C). Consistent with this observation, histological examination of gWAT and the liver showed no apparent genotype-specific alterations. The degree of crowning in gWAT was similar in CPT2 Mϕ-KO and high-fat diet-fed control mice, and the amount of lipid droplets in the liver in CPT2 Mϕ-KO mice was comparable to that in high-fat diet-fed control mice (Fig. 3, E and F). Lastly, we analyzed serum metabolite levels in high-fat diet-fed control and CPT2 Mϕ-KO mice. Results indicated no alteration in the levels of β-hydroxybutyrate, triacylglycerides, nonesterified FFA, cholesterol, or glycerol in the serum of CPT2 Mϕ-KO mice (Fig. 3D). In summary, loss of macrophage fatty acid oxidation does not potentiate diet-dependent changes in body composition, adiposity, or systemic metabolic dysfunction.

Fig. 3.

Loss of macrophage fatty acid oxidation does not affect body composition. A: body weight of control and CPT2 Mϕ-KO male mice fed a low-fat diet (LFD) or a high-fat diet (HFD, n = 10–14). B: EchoMRI measurement of body composition of control and CPT2 Mϕ-KO mice fed a low- or high-fat diet (n = 10–14). C: inguinal white adipose tissue (iWAT), gonadal white adipose tissue (gWAT), and liver wet weight of control and CPT2 Mϕ-KO male mice fed a low- or high-fat diet (n = 10–14). D: serum metabolites [β-hydroxybutyrate (β-HB), triacylglycerol (TAG), nonesterified fatty acid (NEFA), cholesterol, and glycerol] in control and CPT2 Mϕ-KO male mice fed a high-fat diet (n = 7). E: hematoxylin-eosin-stained sections of gWAT from control and CPT2 Mϕ-KO male mice fed a low- or high-fat diet. Scale bar = 200 μm. F: hematoxylin-eosin-stained sections of gWAT from liver of control and CPT2 Mϕ-KO male mice fed a low- or high-fat diet. Scale bar = 200 μm. Values are means ± SE.

Macrophage fatty acid β-oxidation is not required for polarization in vivo.

Metabolic dysfunction in adipose tissue macrophages has been shown to drive macrophage inflammation and alter polarization (22, 28). To determine whether macrophage fatty acid oxidation is a metabolic regulator of alternative activation, gWAT, liver, and spleen depots of low- or high-fat diet-fed control and CPT2 Mϕ-KO mice were analyzed for mRNA expression of several pan-macrophage markers, genes preferentially expressed in classically activated macrophages, and genes preferentially expressed in alternatively activated macrophages. In gWAT, we observed significant increases in mRNA abundance of the pan-macrophage markers F4/80, Cd68, and Cd11b in high-fat diet-fed mice compared with low-fat diet-fed mice, in support of increased macrophage infiltration in obese and diabetic models. We did not, however, observe genotypic differences in the expression of F4/80, Cd68, and Cd11b in gWAT within diets between control and CPT2 Mϕ-KO mice. Furthermore, we did not observe changes in mRNA expression of Tnf-a, Nos2, Mcp1, or Il-6 between genotypes fed a low- or a high-fat diet, suggesting that CPT2 Mϕ-KO mice do not exhibit more inflammation than the control littermates. Additionally, the abundance of mRNA of genes preferentially expressed in alternatively activated macrophages, such as Arg1, Mgl2, Retnla, and Il-10, did not change in CPT2 Mϕ-KO mice compared with control mice (Fig. 4A). These same markers were assessed in the liver and spleen of low- and high-fat diet-fed mice, since both tissues contain their own subset of resident macrophages. The mRNA abundance of all macrophage markers assessed in the liver and spleen remained constant between genotypes fed a low- or a high-fat diet, suggesting that CPT2 Mϕ-KO macrophages resident in either tissue do not exhibit changes in inflammation or macrophage polarization (Fig. 4, B and C). To further evaluate macrophage polarization in adipose tissue of high-fat diet-fed control and CPT2 Mϕ-KO mice, we assessed expression of the alternative activation markers CD206 and CD301 in macrophages isolated from gWAT by flow cytometry. In agreement with the mRNA data, intracellular cytokine staining revealed no changes in expression of CD206, CD301, or the inflammatory markers TNFα and IFNγ between genotypes (Fig. 5). Together these data demonstrate that impaired macrophage fatty acid oxidation does not appear to be a driver of alternative activation, nor does it potentiate inflammation in gWAT of high-fat diet-fed mice.

Fig. 4.

Macrophage fatty acid β-oxidation is dispensable for polarization in vivo. A–C: quantitative RT-PCR analysis of macrophage markers in gWAT(n = 6), liver (n = 6), and spleen (n = 6) of control and CPT2 Mϕ-KO male mice fed a low- or high-fat diet. Values are means ± SE. #P < 0.05; ##P < 0.01; ###P < 0.001 for treatment comparison.

Fig. 5.

Loss of macrophage fatty acid β-oxidation does not affect polarization of adipose tissue macrophages from high-fat diet-fed mice. A and B: total adipose cell number and percentage of adipose tissue macrophages in gWAT of control and CPT2 Mϕ-KO male mice fed a high-fat diet. C–F: percentage of CD206⁺, CD301⁺, TNFα⁺, and IFNγ⁺ adipose tissue macrophages in control and CPT2 Mϕ-KO male mice fed a high-fat diet. Values are means ± SE (n = 5).

Loss of macrophage fatty acid oxidation does not potentiate oxidative stress in vivo.

Obesity is associated with an increase in reactive oxygen species and oxidative stress (8, 16, 27). Previously, we showed that adipocyte fatty acid oxidation was required to potentiate high-fat diet-induced oxidative stress in adipose tissue (25). To determine whether macrophage fatty acid oxidation could also potentiate oxidative stress, we assessed mRNA expression levels of several oxidative stress markers in gWAT. Expression levels of Sod1, Sod2, Cox2, Nqo1, Nox4, and Duox1 in CPT2 Mϕ-KO mice did not differ from those in control littermates fed a low- or high-fat diet (Fig. 6A). In agreement with the mRNA data, there were no changes in lipid peroxidation in gWAT, liver, or serum of CPT2 Mϕ-KO mice fed a high-fat diet (Fig. 6B). Therefore, the loss of macrophage fatty acid oxidation does not potentiate oxidative stress under low- or high-fat diet conditions.

Fig. 6.

Loss of macrophage fatty acid oxidation does not potentiate oxidative stress. A: quantitative RT-PCR analysis of oxidative stress genes in gWAT of control and CPT2 Mϕ-KO male mice fed a low- or high-fat diet (n = 6). B: thiobarbituric acid-reactive substances assay of gWAT, liver, and serum of control and CPT2 Mϕ-KO male mice fed a high-fat diet (n = 5). MDA, malondialdehyde. Values are means ± SE. #P < 0.05, ###P < 0.001 for treatment comparison.

Loss of macrophage fatty acid oxidation does not alter the progression of high-fat diet-induced glucose intolerance.

Adipose tissue resident macrophages, a major source of inflammatory mediators, have been suggested to play a role in regulating insulin sensitivity (44, 46). To determine whether macrophage fatty acid oxidation plays a role in promoting insulin resistance and glucose intolerance, we assessed glucose and insulin tolerance in low- and high-fat diet-fed male mice. Low- and high-fat diet-fed CPT2 Mϕ-KO mice did not exhibit changes in glucose tolerance during the GTT, nor did they display impaired glucose clearance in response to insulin administration in the ITT, compared with littermate controls (Fig. 7). GTT and ITT data are consistent with the in vivo gene expression profile in adipose tissue of CPT2 Mϕ-KO mice, which showed that neither inflammatory nor oxidative stress markers were upregulated. Therefore, the data show that loss of macrophage fatty acid oxidation does not alter high-fat diet-induced systemic insulin resistance or glucose intolerance.

Fig. 7.

Loss of macrophage fatty acid oxidation does not alter progression of high-fat diet-induced insulin resistance. A and B: intraperitoneal glucose tolerance test (ipGTT) and intraperitoneal insulin tolerance test (ipITT), including area under the curve and area above the curve, respectively, for control and CPT2 Mϕ-KO male mice fed low-fat (n = 10–13) and high-fat (n = 8–10) diets. Values are means ± SE.

Loss of macrophage fatty acid oxidation does not induce a mitochondrial DNA stress response.

It has been shown that loss of mitochondrial DNA (mtDNA) can initiate a mtDNA stress response, characterized by induction of IFN-stimulated genes (ISGs) and other antiviral signaling elements (45). Previously, we showed that loss of fatty acid oxidation in adipocytes potentiates the loss of mtDNA, as these mice exhibited increased expression of several ISGs in brown adipose tissue (23). To assess whether macrophages with a loss in fatty acid oxidation exhibited a similar response, we analyzed expression levels of ISGs in control and CPT2 Mϕ-KO BMDM in the presence and absence of poly(I:C). Interestingly, the subset of ISGs, which included Usp18, Irf7, Ifi44, Ifit1, Ifit3, Isg15, and Viperin, remained unchanged in CPT2 Mϕ-KO BMDM compared with control BMDM in the presence and absence of poly(I:C) (Fig. 8). Other inflammatory genes, such as Tnf-a and Il-6, also remained unchanged in CPT2 Mϕ-KO BMDM compared with control BMDM. Only Il-1b, a mediator of inflammatory responses, was upregulated nearly twofold in poly(I:C)-treated, but not untreated, CPT2 Mϕ-KO BMDM. These data suggest that the loss of mitochondrial fatty acid oxidation in BMDM does not potentiate mtDNA stress.

Fig. 8.

Loss of macrophage fatty acid oxidation does not induce a mitochondrial DNA stress response. Quantitative RT-PCR analysis of interferon-stimulated genes in control and CPT2 Mϕ-KO BMDM in the presence and absence of polyinosinic:polycytidylic acid [poly(I:C)]. Values are means ± SE (n = 6). ***P < 0.001.

Loss of macrophage fatty acid β-oxidation does not result in FFA-induced inflammation.

The expression of several inflammatory markers, such as Cox2 and Il-6, has been shown to be regulated by FFA in a Toll-like receptor 4-dependent manner in macrophages (26, 39). To examine whether loss of macrophage fatty acid oxidation produces an inflammatory response after exposure to FFA, we treated control and CPT2 Mϕ-KO BMDM with 100 μM FFA. Although FFA treatment promoted the gene expression of several inflammatory markers, including Cox2 and Il-6, we did not observe genotype-specific changes between wild-type and KO BMDM treated with FFA (Fig. 9). These findings support previous reports showing FFA-induced inflammatory signaling in macrophages and show that loss of macrophage fatty acid β-oxidation does not play a significant role in potentiating an inflammatory response.

Fig. 9.

Loss of macrophage fatty acid β-oxidation does not result in free fatty acid-induced inflammation. Quantitative RT-PCR analysis of inflammatory response genes in control and CPT2 Mϕ-KO BMDM in the presence and absence of 100 μM FFA. Values are means ± SE (n = 6).

DISCUSSION

Traditionally, macrophages have been known to express varying functional programs in response to a wide range of microenvironmental signals, such as products of activated T helper 1- or 2-type lymphocytes and natural killer cells (11). Although the distinction between different macrophage functional programs is dependent on tissue of residence and tissue state in vivo, numerous differentially expressed markers have been identified in order to distinguish and classify macrophage polarization phenotypes in vitro and in vivo (41). Macrophages are capable of adopting distinct metabolic programs; thus, macrophage phenotypes can also be distinguished via shifts in metabolism. However, whether these metabolic changes contribute to macrophage polarization and macrophage function remains unclear. For example, induction of specific M2 genes can occur via IL-4-stimulated activation of the Akt-mTORC1 pathway, which regulates a subset of M2 genes via a signaling pathway that leads to histone modification of these genes (4). Of particular interest is whether macrophage fatty acid oxidation has a causal or a correlative role in macrophage polarization and function. Here we show that macrophage fatty acid oxidation is not required to promote alternative activation, nor is it necessary to potentiate high-fat diet-induced inflammation, oxidative stress, or insulin resistance. These data suggest that while fatty acid oxidation is tightly correlated with these phenotypes, macrophage fatty acid oxidation does not play a causative role.

Previously, we showed that use of the pharmacological fatty acid oxidation inhibitor etomoxir effectively blocks M2 polarization of wild-type and CPT2 KO macrophages, suggesting that macrophage fatty acid oxidation is dispensable for M2 polarization (31). Even though etomoxir and other small-molecule inhibitors of fatty acid oxidation are effective, they have not been demonstrated to be specific. Consistent with these observations is our current finding that M2 gene expression remains unaltered in IL-4-stimulated CPT2 Mϕ-KO BMDM, despite decreased levels of fatty acid oxidation. Furthermore, we have found no evidence to support a role for macrophage fatty acid oxidation in M2 macrophage polarization in a diet-dependent manner in adipose tissue of CPT2 Mϕ-KO mice in vivo. This observation is potentially important, since it is well known that a high-fat diet can promote a phenotypic switch in adipose tissue macrophages to an M1 state (44).

One surprising finding of this study was the absence of a mtDNA stress response in CPT2 Mϕ-KO BMDM. Loss of adipocyte fatty acid oxidation has been shown to generate a mtDNA stress response in brown adipose tissue (23) that is characterized by the induction of ISGs, which are known to exert a wide range of antiviral effector functions (38). We hypothesized that perhaps Cpt2-deficient macrophages would exhibit a similar response, since macrophages play a prominent role in the innate immune response. Interestingly, BMDM from CPT2 Mϕ-KO mice did not exhibit changes in ISG expression, even after viral simulation with poly(I:C). Treatment with poly(I:C), which has been shown to activate the NLRP3 inflammasome (6, 36), did result in a slight induction of Il-1b in CPT2 Mϕ-KO BMDM. In addition, LPS-IFNγ-primed CPT2 Mϕ-KO BMDM showed induction of both Il-1b and Cox2. It is intriguing that one report showed the ability of Cox2 to regulate NLRP3 inflammasome-derived IL-1β production in macrophages (17). Given these observations, it is tempting to speculate whether macrophage fatty acid oxidation has some unknown function in regulating the NLRP3 inflammasome, either directly, via Cox2, or indirectly. One published report showed that fatty acid oxidation can promote NLRP3 inflammasome activation; however, this was shown to occur via Nox4-mediated Cpt1a induction (29). Thus, further investigation may be warranted, and the genetic model presented here may be useful in determining whether stimulation of LPS-primed CPT2 KO BMDM with NLRP3 inflammasome activators, such as, nigericin, ATP, and silica, results in measurable changes in NLRP3 inflammasome activation markers (12, 29).

Several metabolic regulators known to drive alternative macrophage polarization, including Stat6, AMPK, Pparγ, Pparδ, and Pgc1β, have been identified (18, 20, 33–35, 37, 43). Therefore, macrophage metabolism likely still plays a role in polarization. Like the fatty acid oxidation program, all these metabolic regulators of M2 polarization are induced by IL-4 stimulation (11, 18, 20, 33, 34, 43). However, unlike mice lacking macrophage Pparγ or Pparδ, which were shown to be insulin-resistant (13, 20, 33), high-fat diet-fed mice lacking macrophage Cpt2 did not exhibit changes in glucose tolerance or insulin sensitivity. The lack of a potentiated glucose intolerance in CPT2 Mϕ-KO mice is supported by the absence of inflammation and oxidative stress in these mice, both of which can potentiate insulin resistance in high-fat diet-fed obese models (8, 10, 40). Therefore, our studies suggest that macrophage fatty acid oxidation does not appear to have a function in potentiating the progression of glucose intolerance or insulin resistance in type 2 diabetes.

In conclusion, our findings indicate that macrophage long-chain fatty acid β-oxidation is not essential for potentiating metabolic dysfunction characteristic of high-fat diet-fed, obese animals. Interestingly, there are data indicating that macrophage fatty acid oxidation is essential to energetically support M2 function in defense against parasites (19, 32). A genetic model such as the one employed here may be useful and, in fact, necessary to define the requirements of metabolic pathways to macrophage or other immune cell function.

GRANTS

This work was supported in part by National Institute of Neurological Disorders and Stroke Grant R01 NS-072241 and American Diabetes Association Grant 1-16-IBS-313 to M. J. Wolfgang. E. Gonzalez-Hurtado was supported by Johns Hopkins Postbaccalaureate Research Education Program Grant R25 GM-109441. M. R. Horton was supported in part by National Heart, Lung, and Blood Institute Grants PO1 HL-010342 and R21 HL-111783.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.G.-H., J.L., J.C., E.S.S.A., and S.L.C. performed the experiments; E.G.-H., J.L., J.C., E.S.S.A., S.L.C., M.R.H., and M.J.W. analyzed the data; E.G.-H., E.S.S.A., S.L.C., M.R.H., and M.J.W. interpreted the results of the experiments; E.G.-H. prepared the figures; E.G.-H. drafted the manuscript; E.G.-H. and M.J.W. edited and revised the manuscript; E.G.-H., J.L., J.C., E.S.S.A., S.L.C., M.R.H., and M.J.W. approved the final version of the manuscript; M.R.H. and M.J.W. conceived and designed the research.