Macrophage classical M1 activation via TLR4 triggers a variety of responses to achieve the elimination of foreign pathogens. During this process, there is also an increase in lipid droplets which contain large quantities of triacylglycerol (TAG) and phospholipid (PL). The functional consequences of this increment in lipid mass are poorly understood. Here, we studied the contribution of glycerolipid synthesis to lipid accumulation, focusing specifically on the first and rate-limiting enzyme of the pathway: glycerol-3-phosphate acyltransferase (GPAT). Using bone marrow-derived macrophages (BMDMs) treated with Kdo2-lipid A, we showed that glycerolipid synthesis is induced during macrophage activation. GPAT4 protein level and GPAT3/GPAT4 enzymatic activity increase during this process, and these two isoforms were required for the accumulation of cell TAG and PL. The phagocytic capacity of Gpat3−/− and Gpat4−/− BMDM was impaired. Additionally, inhibiting fatty acid β-oxidation reduced phagocytosis only partially, suggesting that lipid accumulation is not necessary for the energy requirements for phagocytosis. Finally, Gpat4−/− BMDM expressed and released more pro-inflammatory cytokines and chemokines after macrophage activation, suggesting a role for GPAT4 in suppressing inflammatory responses. Together, these results provide evidence that glycerolipid synthesis directed by GPAT4 is important for the attenuation of the inflammatory response in activated macrophages.
Macrophages are versatile cells of the immune system that play indispensable roles in both the innate and adaptive immune responses. They can be activated by exposure to inflammatory stimuli such as lipopolysaccharide (LPS) and thereby acquire biological properties used for microbiocidal activity . LPS, and specifically Kdo2-lipid A (KLA), its active component, is recognized by Toll-like receptor 4 (TLR4) . This TLR activates signaling cascades that induce significant changes in cellular gene expression and lead to phagocytosis activation, increased cytokine secretion, reactive oxygen production, and alterations in cell adhesion [3,4]. Macrophage activation is also accompanied by an increase in the synthesis of structural lipids, including triacylglycerol (TAG) and phospholipid (PL) [5–12]. TAG and cholesterol esters (CEs) accumulate in lipid droplets which are dynamic storage organelles, consisting of a core of neutral lipids surrounded by a monolayer of PL, free cholesterol (Chol), and lipid droplet-associated proteins [13–15]. Little is known about the mechanisms by which TLR activation of macrophages leads to TAG and PL accumulation or whether this accumulation is required for macrophage function. It has been suggested that an increase in TAG synthesis, in addition to the previously reported reduction in TAG lipolysis, may contribute to TAG accumulation in activated macrophages . This process can be fueled by an increase in fatty acid (FA) availability as a consequence of multiple pathways: an increase in FA uptake resulting from the induction of the scavenger receptor CD36, the reduction in FA oxidation due to down-regulation of CPT1a (carnitine palmitoyltransferase 1a), and the increase in glucose uptake due to a higher expression of GLUT1. This last process, together with a decrease in glucose oxidation to CO2, leaves glucose available for de novo FA synthesis [12,16,17]. Additionally, PL accumulation and turnover in activated macrophages is essential for the concomitant phagocytic process and signaling responses [18,19].
The de novo synthesis of glycerolipids (TAG and PL) in mammalian cells begins with the acylation of glycerol-3-phosphate, catalyzed by glycerol-3-phosphate acyltransferase (GPAT, EC 22.214.171.124). This step is believed to be rate-limiting . As occurs in many other lipid metabolic reactions, the activity is catalyzed by several GPAT isoforms. Encoding GPAT isoforms 1–4 are four independent genes, which differ in tissue expression pattern, subcellular localization, fatty acyl-CoA substrate preference, and sensitivity to sulfhydryl reagents such as N-ethylmaleimide (NEM) [21,22]. GPAT1 (the product of the Gpam gene) and GPAT2 are mitochondrial isoforms, whereas GPAT3 (also known as AGPAT9/10) and GPAT4 (also known as AGPAT6) are located on the endoplasmic reticulum . Specific GPAT isoforms are thought to initiate glycerolipid synthesis in different cells; GPAT1 and GPAT4 are predominant in hepatocytes , and GPAT3 is the major isoform in differentiated adipocytes . It is unknown which GPAT isoform accounts for the increase in TAG and PL content in activated macrophages.
The aim of the present study was to determine the contribution of glycerolipid synthesis to the increased lipid content in activated macrophages, to study the role of the GPAT isoforms in this processes, and to elucidate the relevance of the ER isoforms to macrophage function, including phagocytic capacity and the ability to produce cytokines and chemokines.
Materials and methods
All chemicals were purchased from Sigma unless otherwise indicated.
Animal care and bone marrow-derived macrophage isolation and culture
Animal protocols were approved by the University of North Carolina at Chapel Hill Institutional Animal Care and Use Committee. Mice were housed in a pathogen-free barrier facility on a 12-h light/dark cycle with free access to water and food. Gpat4−/− mice [25–27] were backcrossed at least eight times onto a C57BL6/J background. Gpat3−/− mice in a C57BL6/J background were provided by Pfizer . To generate bone marrow-derived macrophages (BMDMs), mice were killed with 250 mg/kg Avertin, and bone marrow was collected from the femur and tibia using 2% FBS–PBS. Bone marrow cells were cultured in 10-cm non-surface-treated dishes for 6 days in 4.5 g/l glucose RPMI medium with 100 units/ml penicillin and 100 µg/ml streptomycin, 2 mM glutamine, 1 mM sodium pyruvate, and 20 ng/ml of M-CSF (macrophage colony-stimulating factor; Biolegend). Fresh media were added after 3 days. On day 6, BMDMs were counted, seeded in culture plates, and allowed to attach overnight. Cells were either untreated (naïve unstimulated) or activated following LIPID MAPS protocols (http://www.lipidmaps.org/protocols/PP0000001000.pdf)  with 100 ng/ml KLA (Avanti Polar Lipids) in RPMI supplemented with 10% heat-inactivated FBS during the indicated times. We used FBS in the media because lipopolysaccharide-binding protein, present in the serum, is required to mediate KLA and TLR4 interaction and achieve classical macrophage activation [28–30].
Quantitative real-time PCR
Total RNA was isolated from cells using TRIZOL (Life Technologies) following the manufacturer's instructions, and 1 µg of RNA was used for cDNA synthesis employing the High-Capacity Reverse Transcription Kit (Applied Biosystems). A 1/10–1/7 cDNA dilution was used for the QT-PCR with IQ Sybr Green Super Mix (Bio-Rad). A complete list of primers used are listed in Supplementary Table S1. The thermal profile was 50°C for 10 min, 95°C for 5 min, followed by 40 cycles of 95°C for 30 s, 59°C for 1 min, and 72°C for 30 s, on a Stratagene Mx3000P apparatus. RNA expression of the gene of interest was quantified in triplicate using the ΔCt method and normalized to β-actin housekeeping gene using Qbase software.
Samples of 40–60 μg of total membrane protein fractions were separated by SDS–PAGE (12% gel), transferred to a PVDF membrane (Bio-Rad Laboratories), and probed with a 1 : 2000 dilution of anti-GPAT3 antibody (Sigma HPA029414), anti-GPAT4 antibody (Origine TA309568), or with 1 : 5000 of the anti-β-actin antibody (Sigma A2228) as the loading control overnight. The specificity of GPAT3 and GPAT4 antibodies was validated using total membrane fractions from livers of Gpat3 and Gpat4-null mice (Supplementary Figure S1). Membranes were then washed extensively and probed during 2 h with a 1 : 5000 dilution of horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG antibody (Thermo-Pierce). For chemiluminescent detection, the membranes were incubated with the Super Signal detection kit (Thermo-Pierce).
Lipid droplet staining
Cells were grown on glass coverslips, rinsed with phosphate-buffered saline (PBS), fixed with 2% formaldehyde for 1 h at RT, and then incubated with 3 mg/ml of Oil Red O solution for 1 h at RT. Coverslips were then rinsed once with 60% isopropanol and three times with 0.25 M sucrose before mounting with 1 : 1 glycerol : H2O. Images were taken with a 100× oil immersion lens. Total neutral lipid content was quantified extracting Oil Red O stain with isopropanol for 30 min and the absorbance at 490 nm was measured.
After cell treatments, lipids were extracted , dried under N2 gas, resuspended in t-butanol : methanol : Triton X-100 (3 : 1 : 1, v/v), and analyzed colorimetrically using the Serum Triglyceride Determination Kit (Sigma) following the manufacturer's instructions. Experimental readings were normalized to protein content.
Extracted lipids were separated by thin layer chromatography (TLC) on silica gel G150 20 × 20 cm 250 µm plates in heptane : isopropyl ether : acetic acid (60 : 40 : 4, v/v/v) next to a standard curve. The plates were exposed for 5 s to 5% sulfuric acid in PBS and then heated at 100°C for 20 min to visualize the lipids. The intensity of the spots was quantified and compared with the standard curve using Image J software. Experimental readings were normalized to protein content.
GPAT activity assays
GPAT activity was assayed in total membrane fractions (30–50 µg of protein) . Assays contained 0.8 mM glycerol-3-phosphate–3 µCi [3H] glycerol-3-phosphate , 75 mM Tris–HCl (pH 7.4), 4 mM MgCl2, 2 mg/ml BSA, 8 mM NaF, 1 mM dithiothreitol, and 82.5 µM of palmitoyl-CoA. The assay was performed at 23°C for 10 min in the presence or absence of 1 mM NEM to inhibit the microsomal isoforms. Radioactivity in chloroform-soluble reaction products was quantified by liquid scintillation counting after washing with 1% perchloric acid. Microsomal GPAT activity was estimated by subtracting the NEM-resistant activity (GPAT1) from the total. All assays measured initial rates. [3H]Glycerol-3-phosphate was synthesized enzymatically .
Cell radiolabeling and lipid analysis
Cells were seeded in 12-well plates and attached overnight, then incubated for 24 h with 100 ng/ml KLA in medium supplemented with 10% FBS. During the last 3 h of treatment, trace amounts of [14C]acetate (1 µCi/well, 2.17 mM) or [14C]oleate (0.5 µCi/well, 100 µM) (PerkinElmer) were added in a final volume of 0.5 ml of routine medium supplemented with 0.25% BSA and 1 mM carnitine when [14C]oleate incorporation was assayed. After treatment, the medium was removed, and cells were washed once with 1% BSA in ice-cold PBS before lipids were extracted . Lipid extracts and pure standards were separated by TLC on silica-gel G150 20 × 20 cm 250 µm plates in a two-phase system: chloroform : methanol : ammonium hydroxide (65 : 25 : 4, v/v) run to 50% of the plate and then dried and followed by heptane : isopropyl ether : acetic acid (60 : 40 : 4, v/v) run to the top of the plate. Radiolabeled lipids were quantified with a Bioscan AR-2000 imaging scanner. Lipid standards were visualized by iodine staining. Experimental readings were normalized to protein content.
Cells were plated on 96-well plates, grown overnight, and incubated for 8 h with 100 ng/ml KLA in medium supplemented with 10% FBS. This time point was chosen based on GPAT expression and activity maximum peaks. After 4 h incubation with KLA, the cells were starved with low glucose media (DMEM or RPMI) supplemented with 10% FBS in the presence of KLA and in the presence or absence of etomoxir (100 µM) for 4 h. During the last 2 h of incubation, a phagocytosis assay was performed by adding the pHrodo® Red E. coli BioParticles® Conjugate for Phagocytosis (Molecular Probes, Life Technologies) to the cells. Fluorescence was measured (560/585) (Ex/Em) and the phagocytic capacity was calculated following the manufacturer's instructions.
Cytokine and chemokine expression and release measurement
Cells were treated for 8 h with KLA in medium supplemented with 10% FBS. This time point was chosen based in GPAT4 protein, expression and activity maximum peak. Cells were collected, RNA extracted, and retro-transcribed, and cytokine expression was analyzed using QT-PCR. Media were also collected and centrifuged to eliminate cell debris. Cytokines and chemokines were assayed using LUMINEX MAGPIX technology and MILLIPLEX magnetic bead-based multi-analyte panels from EMD Millipore Company following manufacturer's instructions.
Macrophage activation stimulates lipid droplet formation and increases TAG and PL content
To study glycerolipid dynamics after macrophage activation, we treated mouse BMDMs with KLA for different time periods. Activation was confirmed measuring Tnfa expression [35,36], which increased 30-fold (Figure 1A). Compared with the non-activated macrophages, lipid droplet area increased 2-fold in activated cells (Figure 1B,C). TAG content increased 1.5-fold after 24 h KLA treatment (Figure 1D). Similarly, PL content also increased ∼50% after KLA treatment (Figure 1E). Similar results were obtained with the murine RAW 264.7 cell line (Supplementary Figure S2). These results show that KLA activates macrophages and that the lipid droplet area correlates with the increase in cell content of TAG and PL.
Lipid droplet formation, TAG, and PL content in KLA-stimulated macrophages.
Glycerolipid synthesis increases during macrophage activation
The higher glycerolipid content observed after KLA activation could be the consequence of increased lipid uptake, a decrease in lipid catabolism, and/or the triggering of pathways of de novo synthesis. To determine the contribution of glycerolipid biosynthesis, we analyzed the incorporation of [14C]acetate and [14C]oleate into total lipids. Incorporation of [14C]acetate indicates the use of de novo synthetized FA, whereas [14C]oleate shows the incorporation of exogenous pre-formed FA. We specifically chose oleic acid as a substrate because, unlike saturated  and polyunsaturated fatty acids , it does not activate macrophages. The incorporation of both substrates into total lipids increased ∼35–45% after 24 h KLA treatment (Figure 2). When analyzing the distribution of radioactivity in different lipids, PC and TAG showed the highest increases in incorporation (∼2-fold) of both substrates after activation (Figure 2). [14C]Acetate incorporation into free fatty acids (FFAs) increased, but no increase was observed in the cholesterol or CE fraction. Similar results were observed in RAW 264.7 cells (Supplementary Figure S3). Our results indicate that glycerolipid synthesis is induced during KLA activation and that FAs derived from both de novo synthesis and exogenous sources are used. Thus, the increase in TAG and PL content observed in KLA-activated macrophages (Figure 1 and Supplementary Figure S2) can be attributed, at least partially, to de novo fatty acid and glycerolipid synthesis.
[14C]acetate and [14C]oleate incorporation into lipids after KLA activation.
GPAT3/GPAT4 activity increases during macrophage activation
To determine if GPAT is contributing to the increase in glycerolipid synthesis during macrophage activation, we analyzed the expression (protein and mRNA) of the four GPAT isoforms (1–4) and the GPAT enzymatic activity after different time periods of KLA treatments. GPAT1 expression was low and did not change upon activation, and GPAT2 expression was not detectable, even at the mRNA level (data not shown). However, the two microsomal proteins GPAT3 and GPAT4 were highly expressed in BMDM and the expression of GPAT4 increased >2-fold after 8 h KLA treatment (Figure 3A). Consistent with the higher GPAT4 content, total and NEM-sensitive GPAT activity (i.e. GPAT3 plus GPAT4) was ∼4-fold higher, whereas NEM-resistant GPAT activity corresponding to GPAT1 did not change (Figure 3B). Gpat3 and Gpat4 mRNA expression decreased during activation (Figure 3C), suggesting that GPAT3 and/or GPAT4 may have been activated post-transcriptionally or post-translationally, as previously reported . These results suggest that GPAT4 may be the GPAT isoform that mainly contributes to the initiation of glycerolipid synthesis during macrophage activation, although GPAT3 may have a secondary role.
Effect of macrophage activation on GPAT3 and GPAT4 expression and activity.
GPAT3 and GPAT4 contribute to the increase in lipid droplet, TAG, and PL after KLA activation
To study the contributions of GPAT3 and GPAT4 to the cellular content of lipid droplets, after macrophage activation, we used BMDM from Gpat3−/− and Gpat4−/− mice, as well as the Wt BMDM as controls. The absence of GPAT3 and GPAT4 mRNA and protein levels was monitored by qPCR and Western blot (Supplementary Figure S1). Interestingly, the increase in lipid droplet area observed in control cells after activation (Figure 1B,C) was abrogated in Gpat3−/− and Gpat4−/− macrophages (Figure 4A,B). TAG and PL content increased ∼2-fold in activated wt BMDM, but did not change in Gpat3−/− and Gpat4−/− cells after activation (Figure 4C,D). Similar results were observed in LD, TAG, and PL content in activated RAW cells in which Gpat3 was ∼70% silenced using a specific shRNA (shGpat3 RAW cells) compared with the SCR (Scramble) controls in which we used an shRNA that does not target any known sequence (Supplementary Figures S4 and S5). It is important to mention that no increase in GPAT3 or GPAT4 expression was observed when the other isoform was absent or silenced (Supplementary Figure S6). To evaluate if other enzymes involved in the glycerolipid synthesis pathway are compensating the lack of Gpat3 and Gpat4, we quantified Dgat1 (diacylglycerol O-acyltransferase 1) and Lpin1 mRNA under basal conditions and after 8 h KLA activation. Under basal conditions, only Dgat1 expression was higher in Gpat3-null BMDM, whereas Lpin1 did not change. Neither gene product increased after KLA activation (Supplementary Figure S7). These results demonstrate that not only GPAT4, but also GPAT3, are required for the increase in lipid droplets, TAG, and PL after macrophage activation.
Effect of GPAT3 and GPAT4 knockout on lipid droplet, TAG, and PL accumulation after KLA treatment.
GPAT3 and GPAT4 are required for enhanced glycerolipid synthesis in activated macrophages
To determine the contribution of GPAT3 and GPAT4 to total and NEM-sensitive GPAT activity, specific activity was measured in Gpat3−/− and Gpat4−/− BMDM. Results obtained under basal conditions demonstrated that GPAT4 is the major contributor to NEM-sensitive GPAT activity and the responsible for the increment in GPAT activity upon KLA activation (Figure 5 and Supplementary Figure S8A). NEM-resistant GPAT activity did not change in the different genotypes (Supplementary Figure S8A). Similar results were found in shGpat3 RAW cells, where NEM-sensitive GPAT activity was 20% lower than SCR controls after 8 h KLA treatment and NEM-resistant GPAT activity did not change (Supplementary Figure S8B).
Effect of GPAT3 and GPAT4 knockout on NEM-sensitive GPAT activity after KLA treatment.
Because both glycerolipid content and NEM-sensitive GPAT activity were lower in activated macrophages with absence of GPAT3 or GPAT4 expression, we measured the ongoing rate of glycerolipid synthesis after incubating Gpat3−/− and Gpat4−/− macrophages with [14C]acetate or [14C]oleate. Gpat3−/− and Gpat4−/− activated BMDM showed ∼42% and ∼30% less radiolabeled acetate incorporation into total lipids, respectively (Figure 6A). In both cases, [14C]acetate incorporation into the PC and TAG fractions was reduced. The uptake and incorporation of [14C]oleate into total lipids was also lower in activated Gpat3−/−and Gpat4−/− cells, showing the higher changes in the PC and TAG fractions (Figure 6B). Similar results were observed in shGpat3 RAW cells [14C]acetate and [14C]oleate incorporation (Supplementary Figure S9).
Effect of GPAT3 and GPAT4 knockout in the de novo glycerolipid synthesis after KLA treatment.
These results show that deficiency of either GPAT3 or GPAT4 in macrophages results in decreased GPAT activity and reduced glycerolipid synthesis during activation.
GPAT3 or GPAT4 deficiency affects phagocytosis during macrophage activation
Macrophage phagocytosis is an essential biological process in host defense and requires large amounts of energy and the activation of specific signals. To determine whether newly synthesized glycerolipids function as an energy source for this process or have an alternative role, such as providing FA for the production of signaling molecules, we analyzed the effects of β-oxidation inhibition on phagocytosis. To achieve this, we measured cell uptake of Escherichia coli conjugated with a pH-sensitive dye in the presence or absence of etomoxir (CPT1a inhibitor). KLA treatment caused 2.85-fold increase in the phagocytic capacity of wt BMDM (Figure 7). When β-oxidation was inhibited 85% by etomoxir treatment during macrophage activation (Supplementary Figure S10), the phagocytic capacity was reduced by 15% in BMDM (Figure 7). These results show that FA oxidation contributes only minimally to the energy supply that feeds phagocytosis after macrophage activation. We also studied the effect of the absence of GPAT3 or GPAT4 on phagocytosis. Gpat3−/− and Gpat4−/− macrophages had reduced phagocytic capacity after activation with KLA alone or KLA in the presence of etomoxir, suggesting that these two GPAT isoforms are needed to synthesize the glycerolipid pool required for phagocytosis after macrophage activation. Consistent results were observed in shGpat3 RAW cells (Supplementary Figure S11).
Effect of the lack of GPAT3 or GPAT4 in phagocytosis capacity after macrophage activation.
Lack of GPAT4 increases pro-inflammatory cytokine and chemokine release during macrophage activation
Activation of macrophages results in the induction and release of cytokines and chemokines. It has been proposed that cytokines are stored in lipid droplets and that some phospholipids like PC might be essential for cytokine release [40,41,42]. Thus, we studied the effect of absent GPAT3 or GPAT4 on cytokine and chemokine expression and release. The lack of GPAT3 did not produce a consistent effect on cytokine and chemokine expression and release (Supplementary Figure S12). In contrast, when Gpat4−/− BMDMs were activated, the synthesis and release of pro-inflammatory signal molecules increased significantly (Figure 8). These results suggest that GPAT4 is required to regulate the production of pro-inflammatory signal molecules.
Effect of the lack of GPAT4 in pro-inflammatory cytokine and chemokine expression and release during macrophage activation.
Conclusions and discussion
Previous studies have shown that activation of TLRs causes TAG and PL to accumulate in macrophages [10–12,43,44]. The purpose of the present study was to elucidate the contribution of glycerolipid synthesis to lipid accumulation, to determine which of the four GPAT isoforms is activated when macrophages are treated with KLA, and to learn whether TAG and PL accumulation is required for cytokine synthesis and release, and for phagocytosis. We used BMDM from Gpat3 and Gpat4-null mice as well as the stable macrophage cell line, RAW 264.7. As previously reported, our data show that glycerolipid accumulates during macrophage activation and that macrophages can incorporate both exogenous and de novo synthetized FAs into TAG and PL [12,45]. More precursors were incorporated into PL than TAG, consistent with the need for PL to surround the droplets that have increased their total area, and more importantly, to contribute to the increase in cell membrane surface during the enlargement of activated macrophages [18,46,47]. The importance of PC synthesis for the inflammatory response was observed in myeloid cells null for CCT-alpha, which catalyzes the rate-limiting step in PC synthesis . Recently, it was reported that the expression of the choline transporter CTL1 was up-regulated in BMDM after activation. Inhibiting choline uptake increased the secretion of pro-inflammatory cytokines in response to LPS, indicating a functional role linking PL metabolism and inflammation .
The induction of de novo FA synthesis during macrophage activation has been previously reported [49,50]; based on our results, we postulate that GPAT4 is probably responsible for the majority of the increase in GPAT NEM-sensitive activity in activated macrophages, as the absence of GPAT4 causes the GPAT NEM-sensitive activity to fall to control levels while the effect of GPAT3 deficiency on this enzymatic activity is unremarkable (Figure 5). Nevertheless, a clear and similar effect on lipid synthesis and accumulation is observed in both Gpat3- and Gpat4-deficient macrophages. This is not the first observation that minimal variations in GPAT activity in GPAT-deficient cells lead to striking phenotypic differences. For example, in Gpat4-null mice, changes in GPAT activity in brown adipose were not significant, even though the mice showed altered growth, heat wasting, and brown adipose FA oxidation . On the other hand, the GPATs may be able to acylate other substrates in addition to glycerol-3 phosphate. Both GPAT3 and GPAT4 were initially thought to be acylglycerol phosphate acyltransferases (AGPATs), and GPAT3 has AGPAT activity . Thus, the phenotypic effects of GPAT3-deficient macrophages might be due to an alternate enzymatic activity, such as AGPAT. The fact that only the two endoplasmic reticulum isoforms (GPAT3 and GPAT4), and not the mitochondrial isoforms (GPAT1 and GPAT2), appear to be responsible for glycerolipid synthesis induction in activated macrophages may explain data that show an increase in some classes of phosphatidic acid (the product of GPAT). These phosphatidic acid species (32 : 1, 34 : 1, 36 : 0, 40 : 3, 40 : 4 and 40 : 6 plus PC 38:6) were increased specifically in the endoplasmic reticulum and not in mitochondria after macrophage activation .
The induction of GPAT4 protein and activity was not the consequence of a transcriptional up-regulation. In a previous report, it was shown that oleate loading causes GPAT4 to re-localize from the endoplasmic reticulum to the lipid droplets in Drosophila cells. This relocalization was proposed to play a role in LD growth because it was only detected in a subpopulation of active and growing lipid droplets . A similar regulatory mechanism involving translocation between organelles was observed for rat liver GPAT1; the most active pool is located within the outer mitochondrial membrane, whereas the proteins localized in mitochondria-associated vesicles are less active . Therefore, the induction of ER-GPAT activity could be the consequence of protein translocation.
We next analyzed the functional role of the glycerolipids that had accumulated after macrophage activation. We demonstrated that FAs stored in LD are not preferentially used for energy production via β-oxidation because inhibiting oxidation with etomoxir blocked phagocytosis only minimally. This observation showing that fatty acid β-oxidation is not the primary source of energy for phagocytosis agrees with previous studies that postulate that glucose rather than lipids is the main substrate used to support phagocytosis energetically . However, glycerolipid accumulation is necessary for successful phagocytosis since activated Gpat3−/− and Gpat4−/− BMDMs have their phagocytic capacity reduced by 20–30%.
Our results show that the lack of GPAT4 increases the release of several pro-inflammatory cytokines and chemokines after macrophage activation. Thus, GPAT4 could act by suppressing inflammatory responses typical of M1 macrophage activation. Precisely, how GPAT4 affects M1 macrophage activation is unclear, as is its potential effect on chronic infection or basal immune function. M1 activation is influenced by the levels of specific intracellular lipids (FFAs, acyl-CoAs, and glycerolipids), which regulate a wide variety of signaling pathways [56–59]. The capacity of macrophages to incorporate FAs into TAG via GPAT4 might reduce the pool of a specific FA and diminish its signaling role, thereby modulating its potential for M1 activation. Increased esterification of intracellular FAs into the TAG pool may act as a ‘sink’ to lower their flux into pathways that affect the levels of signaling metabolites linked to inflammation (ceramides, eicosanoids, diacylglycerol, and lysophosphatidic acid). This mechanism might trap specific FAs in TAG, thereby reducing inflammation; such an effect has been previously proposed for DGAT1, the enzyme that catalyzes the final step in TAG synthesis . However, in Gpat4-null BMDM, the expression of Dgat1 was not up-regulated. The possible role of TAG in macrophage function emerged from experiments performed in ATGL-deficient mice. Atgl−/− BMDMs store more TAG and also present an M-2 anti-inflammatory-like phenotype , further supporting that the increase in LD TAG synthesis (via Gpat4 or Dgat1) or the decrease in TAG hydrolysis (via Atgl) protects from the inflammatory response induced by FA. The fact that both Gpat4−/− and Atgl−/− show impaired phagocytosis  and exacerbated pro-inflammatory cytokine release suggests that the balance between TAG synthesis and degradation is essential for macrophage function. On the other hand, we cannot exclude the possibility that GPAT4 could affect a pool of signaling lipids such as LPA and PA. This occurs in Lpin1-deficient macrophages, in which decreased production of DAG and decreased MAPK activation alter cytokine expression and release .
The mechanisms by which TLR stimulation interacts with the lipid metabolism in macrophages are poorly understood. An important link is that macrophages treated with LPS increase the expression and activation of the sterol regulatory element-binding protein-1a (SREBP1a), which up-regulates the transcription of enzymes required for de novo FA and TAG synthesis . Thus, the induction of enzymes that capture these newly synthetized FAs and direct them into pathways that synthesize glycerolipids might be a compensatory strategy to avoid the increase in potentially toxic cellular FAs .
In summary, the present study demonstrates that glycerolipid synthesis is induced during macrophage activation and that GPAT3 and GPAT4 are required for the synthesis of lipid droplets, TAG, and PL that accumulate during this process. In addition, our results show that the down-regulation of GPAT3 and GPAT4 impairs phagocytosis and the production of signaling molecules, providing evidence that glycerolipid synthesis is critical for macrophage function.
acylglycerol phosphate acyltransferases
bone marrow-derived macrophages
carnitine palmitoyltransferase 1
diacylglycerol O-acyltransferase 1
free fatty acids
thin layer chromatography
I.Y.Q. and A.L.S. performed research. I.Y.Q., M.P.-M., R.A.C., and M.R.G.-B. designed research. I.Y.Q. and M.R.G.-B. prepared illustrations. I.Y.Q., M.P.-M., R.A.C. and M.R.G.-B. wrote the manuscript.
This work was supported by Agencia Nacional de Promoción Científica y Tecnológica, Argentina PICT 3214 and Consejo Nacional de Investigaciones Científicas y Tecnológicas, Argentina PIP0310 (M.R.G.-B.) and a grant from the National Institutes of Health, USA DK56598 (R.A.C.).
We thank Mario Raul Ramos for the illustrations, Marianela Santana and Gillermina Mangione for their technical assistance and Dr Pamela Young for her technical advice and insightful discussions. We also thank Pfizer for providing the Gpat3−/− mice.
The Authors declare that there are no competing interests associated with the manuscript.