A snake venom group IIA PLA2 with immunomodulatory activity induces formation of lipid droplets containing 15-d-PGJ2 in macrophages

Crotoxin B (CB) is a catalytically active group IIA sPLA2 from Crotalus durissus terrificus snake venom. In contrast to most GIIA sPLA2s, CB exhibits anti-inflammatory effects, including the ability to inhibit leukocyte functions. Lipid droplets (LDs) are lipid-rich organelles associated with inflammation and recognized as a site for the synthesis of inflammatory lipid mediators. Here, the ability of CB to induce formation of LDs and the mechanisms involved in this effect were investigated in isolated macrophages. The profile of CB-induced 15-d-PGJ2 (15-Deoxy-Delta-12,14-prostaglandin J2) production and involvement of LDs in 15-d-PGJ2 biosynthesis were also investigated. Stimulation of murine macrophages with CB induced increased number of LDs and release of 15-d-PGJ2. LDs induced by CB were associated to PLIN2 recruitment and expression and required activation of PKC, PI3K, MEK1/2, JNK, iPLA2 and PLD. Both 15-d-PGJ2 and COX-1 were found in CB-induced LDs indicating that LDs contribute to the inhibitory effects of CB by acting as platform for synthesis of 15-d-PGJ2, a pro-resolving lipid mediator. Together, our data indicate that an immunomodulatory GIIA sPLA2 can directly induce LD formation and production of a pro-resolving mediator in an inflammatory cell and afford new insights into the roles of LDs in resolution of inflammatory processes.

Phospholipases A 2 (PLA 2 s) are enzymes that play a key role in a range of cellular processes in physiological and pathological conditions by regulating the release of arachidonic acid (AA), a precursor of distinct classes of lipid mediators such as prostaglandins and leukotrienes, which are key regulators of inflammatory processes 1 . These enzymes have been classified into fifteen groups (I-XV) and subgroups according to nucleotide and amino acid sequence criteria. The groups comprise five distinct types of enzymes: secreted (sPLA 2 ), cytosolic (cPLA 2 ) and calcium independent (iPLA 2 ) PLA 2 s, platelet activating factor (PAF) acetylhydrolases and lysosomal PLA 2 s 2 . Group IIA includes mammalian inflammatory sPLA 2 s and phospolipases present in viperid snake venoms 3 . Crotoxin B (CB), a group IIA sPLA 2 isolated from Crotalus durissus terrificus (Cdt) snake venom, is one of the two subunits that form crotoxin, the major component of Cdt venom. In contrast to most group IIA sPLA 2 s, CB per se displays immunomodulatory effects 4,5 . It is able to inhibit macrophage spreading and phagocytic activities, both of which are associated with increased levels of lipoxin A 4 (LXA 4 ), an important arachidonic acid-derived lipid mediator shown to be active in the resolution phase of inflammation 6 . CB also induces the biosynthesis and release of prostaglandin E 2 (PGE 2 ) and D 2 (PGD 2 ) via activation of the catalytic activity of cyclooxygenase-1 (COX-1) 7 . Although PGE 2 is known to dilate blood vessels, potentiating edema formation, its immunosuppressive activities, including inhibition of the phagocytic functions and microbicidal activity of macrophages, have been well Phospholipase A 2 . CB-sPLA 2 was obtained from crotoxin (CTX) isolated from Cdt snake venom by reversed-phase chromatography 27,28 , using a 201SP54 column (Grace Vydac, CA, USA). Briefly, 1 mg of CTX was loaded onto the column, and fractions of 0.2 mL were eluted with a linear gradient of 0.1% trifluoroacetic acid (TFA) in H 2 O and 0.1% TFA in acetonitrile. CTX was isolated from Cdt venom by conventional gel filtration on Sephadex 75 29 . CB identity was checked by SDS-polyacrylamide gel electrophoresis under non-reducing conditions and by matrix-assisted laser desorption ionization time-of-flight mass spectrometry in an AutoFlex III MALDI-TOF-TOF instrument (Bruker Daltonics, Bremen, Germany). The PLA 2 activity of CB was assayed using the classical egg yolk clearing method 30 . The absence of endotoxin contamination in the CB preparation was demonstrated by the quantitative Limulus Amebocyte Lysate (LAL) test 31 , which revealed undetectable levels of endotoxin (<0.125 EU/mL).
Macrophage harvesting. Elicited macrophages were harvested four days after i.p. injection of 1 mL of 3% Brewer thioglycolate medium. The animals were killed under CO 2 atmosphere, and the cells were harvested by washing the peritoneal cavities with 3 mL of PBS, pH 7.2, containing 10 IU/mL heparin. Aliquots of the washes were used for total cell counts in a Neubauer chamber after dilution (1:20, v/v) in Turk's solution (0.2% crystal violet dissolved in 30% acetic acid). Differential cell counts were performed on smears stained with Hema-3 and examined under light microscopy. More than 95% of the cell population consisted of macrophages according to conventional morphological criteria. The remaining peritoneal wash was centrifuged at 500 g for 6 min (4 °C), and the cell pellets were used for subsequent studies after suitable dilutions.
Cytotoxicity assay. Cytotoxicity of CB and all the pharmacological agents used toward thioglycolate-elicited macrophages was evaluated using a tetrazolium-based assay (MTT). In brief, 2 × 10 5 macrophages/well in RPMI-1640 medium supplemented with 40 µg/mL gentamicin sulfate and 2 mM L-glutamine were plated in 96-well plates and incubated with 100 μL of selected concentrations of CB (0.1-0.8 μM) diluted in medium or with the same volume of medium alone (control) for 1, 3, 6, 12 or 24 h at 37 °C in a humidified atmosphere of 5% CO 2 . MTT (5 mg/mL) was dissolved in PBS and filtered to sterilize it and remove a small amount of insoluble residue present in some batches. Stock MTT solution (10% in culture medium) was added to all wells in each assay, and the plates were incubated at 37 °C for 3 h. Next, a volume of 100 μL of DMSO was added to the wells and mixed thoroughly at room temperature for 30 min. Absorbance at 540 nm was then recorded in a microtiter plate reader. Results were expressed as percentage of viable cells, and the control cells (incubated with medium alone) were considered 100% viable.
Lipid droplet staining and quantification. Analysis of LD numbers was performed in osmium-stained cells. In brief, macrophages (2 × 10 5 cells) adhered to glass coverslips were fixed in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer, pH 7.2, for 15 min and stained with OsO 4 (osmium tetroxide). The coverslips were then rinsed in 0.1 M phosphate buffer, stained in 1% OsO 4 (30 min), rinsed in deionized H 2 O, immersed in 1.0% thiocarbohydrazide (5 min), rinsed again in 0.1 M phosphate buffer, re-stained with 1% OsO 4 (3 min), rinsed with H 2 O and then dried and mounted. The morphology of the fixed cells was examined, and round osmiophilic structures were identified as LDs, which were then counted under phase-contrast microscopy using a 100x objective lens in 50 consecutively scanned macrophages in each coverslip. For assays with fluorescent-labeled LDs, macrophages (2 × 10 5 cells) adhered to glass coverslips were incubated with Nile Red staining solution freshly prepared in 0.1 M phosphate buffer (10 μg/mL) for 20 min at room temperature and washed with phosphate buffer. After several washes the coverslips were mounted with Fluoromount-G and examined under a confocal microscope (Zeiss LSM 510 Meta confocal microscope). Immunocytochemistry analysis. Detection of PLIN2 in CB-stimulated macrophages was performed by PLIN2 immunostaining. Cells were fixed in 2% PFA, permeabilized with 0.2% Triton-X 100 in 0.1 M phosphate buffer and blocked with 0.5% normal donkey serum in 0.1 M phosphate buffer for 90 min. After PBS washes, macrophages were incubated for 1 h with guinea pig polyclonal anti-mouse PLIN2 (1:2000) diluted in 0.1 M phosphate buffer with 0.2% Triton-X 100. After three washes with PBS (10 min each), the preparations were incubated with secondary FITC-conjugated donkey anti-guinea pig antibody (1:500) in the dark for 1 h. After the washes, the slides were mounted with Fluoromount-G and examined under a confocal laser scanning microscope (Zeiss LSM 510 Meta). For analysis of COX-1and 15-d-PGJ 2 immunostaining, the cells were fixed and permeabilized in 1% EDAC 45  Western blotting. For PLIN2 detection, whole cells extracts obtained by lysing the pellets with 100 μL of sample buffer (0.5 M Tris-HCl, pH 6.8, 20% SDS, 1% glycerol, 1 M β-mercaptoethanol, 0.1% bromophenol blue) and boiling them for 10 min were used. Samples were resolved by SDS polyacrylamide gel electrophoresis (SDS-PAGE) on a 10% bis-acrylamide gel overlaid with a 5% stacking gel. Proteins were then transferred Scientific RepoRts | 7: 4098 | DOI:10.1038/s41598-017-04498-8 to a nitrocellulose membrane (GE Healthcare, Buckinghamshire, UK) using a Mini Trans-Blot ® (Bio-Rad Laboratories, Richmond, CA, USA). The membranes were blocked for 1 h with 5% nonfat dry milk in TTBS (20 mM Tris, 100 mM NaCl and 0.5% Tween 20) and incubated with a primary antibody against PLIN2 (Abcam) overnight at 4 °C or with β-actin (Sigma) for 1 h at room temperature. They were then washed and incubated with the appropriate horseradish peroxidase conjugated anti-rabbit IgG secondary antibody (1:1000 dilution, 1 h, room temperature). To analyze phosphorylated forms of PI3K, MEK1/2, ERK1/2 and JNK, the cells were lysed using RIPA ® (Sigma Aldrich) with a protease inhibitor cocktail and phosphatase inhibitors (sodium fluorate 10 mM and sodium orthovanadate 1 mM) and then dissolved (1:5 v/v) in the sample buffer and boiled for 10 min. After the proteins had been separated by SDS-PAGE (12%) and transferred to a nitrocellulose membrane, as described above, the phosphorylated and total non-phosphorylated proteins were detected using polyclonal antibodies to PI3K, MEK1/2, ERK1/2 and JNK (1:1000) and total kinases (1:650) overnight at 4 °C, followed by incubation with horseradish peroxidase-conjugated anti-rabbit IgG (1:1000, 1 h, room temperature). Detection was by the enhanced chemiluminescence (ECL) method according to the manufacturer's instructions (GE Healthcare, Buckinghamshire, UK). Band densities were quantified with a GS 800 Densitometer (Bio-Rad Laboratories, Richmond, CA) using Molecular Analyst ® image analysis software (Bio-Rad Laboratories, Richmond, CA). Statistical Analysis. Data are expressed as the mean ± standard error of mean (SEM) of at least three independent experiments. Multiple comparisons among groups were performed with one-way ANOVA followed by Tukey's test. Probabilities of less than 5% (p < 0.05) were considered statistically significant.

CB induces LDs formation in macrophages. To investigate whether CB would induce formation of
LDs in isolated macrophages, these cells were incubated for 1 h with CB at non cytotoxic concentrations ranging from 0.1 to 0.8 µM. Lack of cytotoxicity with these concentrations of CB was previously determined (1-24 h time intervals) by MTT assay. As shown in Fig. 1A, incubation of macrophages for 1 h with CB at concentrations ranging from 0.2 to 0.8 µM, but not 0.1 µM, induced a significant increase in the number of LDs compared with control cells incubated with culture medium alone. Maximal LD numbers were observed with a concentration of 0.8 µM. This CB-induced effect was not observed in Ca 2+ -free, EGTA-containing medium, in a medium in which Ca 2+ had been replaced by Sr 2+ or when CB was inactivated by p-bromophenacyl bromide (p-BPB) (Fig. 1B), indicating that the catalytic activity of this sPLA 2 is essential for induction of LD formation under the present experimental conditions. To determine the time-course of CB-induced LD formation, a submaximal concentration of this sPLA 2 was used (0.4 µM), and the number of LDs after 1-24 h of incubation was determined. As shown in Fig. 1, CB caused a significant increase in the number of LDs after 1-24 h incubation compared with control cells. The greatest number of LDs was detected after 12 h of incubation (Fig. 1C). As illustrated in Fig. 1D, control macrophages stained with OsO 4 showed very few inclusions in the cytoplasm. In contrast, the cytoplasm in CB-stimulated macrophages was packed with the osmiophilic organelles, which can be seen as dark punctate structures in Fig. 1E, F and G. Taken together, these findings indicate that CB is able to induce LD biogenesis in cultured macrophages in a rapid-onset effect that is dependent on the incubation time and catalytic activity of this enzyme.

CB induces protein expression of PLIN2 and its recruitment to LDs in macrophages. PLIN2
is a fatty acid-binding protein that plays an important role in LD assembly and the development of foamy macrophages 18 . In an attempt to better understand the stimulatory effect of CB that leads to LD formation in macrophages, we investigated whether CB induces expression of PLIN2. While this revealed increased levels of PLIN2 expression in cell extracts stimulated with CB for 1, 3 and 12 h, but not for 6 and 24 h, extracts from unstimulated cells (controls) failed to show significant levels of PLIN2 for any of the incubation periods tested ( Fig. 2A, B and Supplementary Fig. S1). These findings demonstrate that CB is able to up-regulate PLIN2 expression at translational levels and suggest a mechanism for the increase in LB formation detected in CB-stimulated macrophages. Considering that the increase of PLIN2 protein expression occurred since the first hour of stimulation by CB it is possible that the amount of PLIN2 needed to supply the increased assembly of LDs induced by CB was enough at 6 h of stimulation leading to the transient lack of increase of PLIN2 protein expression seen at 6 h of stimulation by this sPLA 2 . At 24 h, however, the absence of PLIN2 protein expression was accompanied by a significant decay of LD numbers, suggesting that CB-induced effect falls into decline. To further identify the mechanisms involved in CB-induced LD formation, the ability of CB to recruit PLIN2 into cytosol LDs was investigated. As illustrated in Fig. 2C, macrophages stimulated with CB (0.4 µM) for 3 h exhibited strong fluorescent staining (green) for PLIN2 with a punctate cytoplasmic pattern that was absent in the unstimulated control cells. Fluorescent Nile Red-labeled neutral lipid inclusions (LDs) overlapping with stained cytoplasmic PLIN2 were also visualized 3 h after CB-induced stimulation indicating that PLIN2 co-localizes to LDs. As expected, no significant staining was detected in control macrophages. These data demonstrate the ability of CB to recruit PLIN2 from cell membranes to form new LDs in macrophages.
Distinct signaling pathways are involved in CB-induced LD formation. Regulation of LD formation induced by diverse inflammatory agents involves stimulus-specific signaling pathways 46 . Participation of protein kinases in the signaling triggered by sPLA 2 s has been previously reported 47,48 . Therefore, the critical signaling proteins involved in CB-induced LD biogenesis were investigated using pharmacological approaches. The effects of selected pharmacological treatments were evaluated 3 h after incubation of macrophages with CB.
To investigate the role of PKC and PI3K in CB-induced LD formation, macrophages were treated with selective PKC and PI3K inhibitors (H7, GF109203X and LY294002, CAL101, respectively) or their vehicles (controls) prior to incubation with CB, followed by quantification of osmium-stained LDs. As seen in Fig. 3A and B, both PKC inhibitors (H7 and GF109203X) and PI3K inhibitors (LY294002 and CAL101) abolished the increased LD formation in CB-stimulated macrophages compared with vehicle-treated CB-stimulated macrophages. Next, we investigated whether the effect of CB on LD formation is related to MAPKs p38 MAPK , ERK1/2 or JNK. As demonstrated in Fig. 3C and D, neither ERK1/2 inhibitors (PD98059 and ERK inhibitor), nor p38 MAPK inhibitors (SB202190 and SB203580) changed the number of LDs in CB-stimulated macrophages compared with vehicle-treated macrophages stimulated with CB. Furthermore, Fig. 3E shows that pharmacological inhibition of cells with JNK inhibitors (JNK inhibitor II and SP600125) abrogated LD formation in CB-stimulated macrophages compared with untreated control cells. The next protein evaluated was MEK1/2, an important upstream protein of the MAPK signaling pathway. U0126, a MEK1/2 inhibitor, reduced the increase in LD formation in CB-stimulated macrophages compared with vehicle-treated control cells (Fig. 3E), reinforcing the role of the MAPK family in CB-induced LD formation in macrophages. Activation of the proteins PKC, PI3K, MEK1/2 and JNK was confirmed by evaluating the extent of phosphorylation (Fig. 4A, B, C and D). CB induced a significant increase in phosphorylation of these proteins only at the 5-minute time point.

PLD and iPLA 2 are involved in CB-induced LD formation in macrophages.
It has been reported that phospholipase D (PLD) and phospholipase C (PLC) play important roles in the signal transduction pathways associated with LD formation 49 . As seen in Fig. 5A, pre-treatment of macrophages with U73122, an inhibitor of PLC, did not have an effect on CB-induced LD formation in comparison with control cells. However, when the cells were incubated with compound FIPI, an inhibitor of PLD (Fig. 5B), LD formation induced by CB was abrogated compared with controls. These results indicate that PLD is critical for the stimulatory effect of CB on LDs formation in macrophages, whereas PLC does not contribute to this CB-induced effect. Additionally, it is known that the intracellular PLA 2 s, such as calcium independent PLA 2 (iPLA 2 ) and cytosolic PLA 2 (cPLA 2 ) have a role in LDs formation 47,50 . Moreover, a cross talk between sPLA 2 s and intracellular PLA 2 s for production of prostaglandins and LDs has been described 7,47 . Therefore, to verify whether iPLA 2 and/or cPLA 2 would be committed to CB-induced LD formation, macrophages were treated with low micromolar concentrations of Pyr-2, a specific inhibitor of cPLA 2 α, or with either BEL or FKGK11, both inhibitors of iPLA 2 . As shown in Fig. 5C, treatment of macrophages with BEL or FKGK11, but not Pyr-2, caused a significant reduction in the number of LDs in CB-stimulated macrophages compared with CB-stimulated vehicle-treated cells. These data indicate that iPLA 2 , but not cPLA 2 plays a role in LD formation induced by CB in macrophages. 15-d-PGJ 2 , which is a PGD 2 -derived prostanoid 51 , has been described as a pro-resolving lipid mediator, which is involved in the switch from the acute into the regenerative phases of inflammation 52 . We have previously reported that CB elicits production of high levels of PGD 2 by isolated macrophages 7 . On this basis, we evaluated the ability of CB to induce the release of 15-d-PGJ 2 from macrophages in culture. 15-d-PGJ 2 concentration was measured in supernatants from macrophages incubated with either a sub-cytotoxic concentration of CB (0.4 μM) or RPMI alone (control). As shown in Fig. 6A, incubation of macrophages with CB induced a significant increase in 15-d-PGJ 2 concentration in the supernatants from 1 up to 24 h, with levels ranging from 1545 ± 149.5 to 2283 ± 340.5 pg/mL (p < 0.05). The corresponding control values ranged from 110.1 ± 11.5 to 1333 ± 179.3 pg/mL. These results are evidence of the ability of CB to induce generation of 15-d-PGJ 2 . Considering that (i) cyclooxygenases are key enzymes in the prostaglandin biosynthetic pathway and (ii) previous data from our group showing that COX-1,   bands (E, F, G and H). The densities (in arbitrary units) were normalized with those of β-actin. Results are expressed as mean ± SEM for three experiments. *p < 0.05 compared with RPMI-stimulated cells.

CB induces production of 15-d-PGJ 2 via COX-1 in macrophages.
possible role of CB-induced LDs as intracellular sites for the synthesis of lipid mediators, specially 15-d-PGJ 2 , macrophages stimulated or not by CB were incubated with EDAC to immobilize the newly synthesized eicosanoids 45 and immunostained with antibodies against 15-d-PGJ 2 and against COX-1, and LDs were stained with Nile Red. As illustrated in Fig. 7A, macrophages stimulated with CB (0.4 µM) for 3 h exhibited a (green) cytoplasmic staining pattern for 15-d-PGJ 2 . This pattern was not observed in the unstimulated control cells, which exhibited diffuse staining. Fluorescent Nile Red-labeled LDs were also visualized 3 h after stimulation by CB and were virtually absent in unstimulated control macrophages. Overlapping images show that stained cytoplasmic 15-d-PGJ 2 matched perfectly with neutral lipid inclusions in CB-stimulated macrophages indicating that 15-d-PGJ 2 co-localizes to LDs. In light of our present results showing that formation of LDs induced by CB was inhibited by PKC and JNK, the role of LDs in the synthesis of 15-d-PGJ 2 induced by CB was further investigated pretreating macrophages with either PKC inhibitors (H7 and GF109203X) or JNK inhibitors (JNK inhibitor II and SP600125), which abolished CB-induced LD formation, or their vehicles (controls) prior to incubation with CB for 3 h, followed by quantification of 15-d-PGJ 2 levels in cell supernatants. As seen in Fig. 7B, treatment of cells with PKC inhibitors abolished the increased 15-d-PGJ 2 production in CB-stimulated macrophages compared with vehicle-treated CB-stimulated macrophages. Moreover, Fig. 7C demonstrates that pharmacological inhibition of cells with JNK inhibitors abrogated 15-d-PGJ 2 production in CB-stimulated macrophages compared with untreated control cells. In addition, as illustrated in Fig. 8A, macrophages stimulated with CB (0.4 µM) for 3 h exhibited a punctate cytoplasmic fluorescence (green) for COX-1. This pattern could not be observed in the unstimulated control cells, in which diffuse staining was observed. Fluorescent Nile Red-labeled LDs were also visualized 3 h after stimulation by CB and were virtually absent in unstimulated control macrophages. Overlapping images show that stained cytoplasmic 15-d-PGJ 2 and COX-1 matched perfectly with neutral lipid inclusions in CB-stimulated macrophages. Considering these results we investigated whether COX-1 contributes to formation of LDs induced by CB. As demonstrated in Fig. 8B, 1 h pretreatment of macrophages cultures with valeryl salicylate (450 μM) or SC-560 (0.5 μM), both specific COX-1 inhibitors, did not affect LDs formation induced by CB after 3 h of stimulation in comparison with control cells, pretreated with vehicle and stimulated with CB, indicating that COX-1 isoform is not involved in formation of LDs induced by CB. In the present experimental condition, this isoform is committed to 15-d-PGJ 2 synthesis. Taken together these findings indicate that CB has the capability to induce the synthesis of a pro-resolving lipid mediator (15-d-PGJ 2 ) and its compartimentalization, in association with the prostaglandin-forming enzyme COX-1, in LDs. These data indicate the role of LDs as a relevant site for the synthesis and accumulation of 15-d-PGJ 2 in macrophages stimulated by CB.

Discussion
Group IIA secreted phospholipases A 2 are enzymes whose proinflammatory properties have been well established 53,54 . However, CB, a group IIA sPLA 2 isolated from Cdt snake venom, interferes with the development of the inflammatory response by blocking important stages in this process, such as leukocyte recruitment and macrophage activation 4,5 . In the present study, the ability of this GIIA sPLA 2 to directly stimulate formation of LDs in macrophages and the mechanisms involved in this phenomenon were demonstrated. This effect was time-dependent, had a very fast onset with a maximum at 12 h, and persisted up to the 24-hour time-point. These data agree with previous reports from our group showing that two proinflammatory secreted PLA 2 s from Bothrops snake venom are able to induce LD formation in isolated macrophages 47,48 . These results are interesting as formation of this organelle in leukocytes is generally associated with a variety of phlogistic stimuli 13,55,56 , including stimuli by proinflammatory group IIA sPLA 2 s 47,48 . In addition, to our knowledge, this is the first study showing that LD formation can be induced by group IIA sPLA 2 that does not induce an inflammatory response.
PLIN2 is a LD marker protein and plays an important role in LD assembly and foam cell formation 18-20 by its properties to bind lipids with high affinity and increase the uptake of long chain fatty acids 57 . Several studies have shown that PLIN2 expression has been directly related to enhanced neutral lipid storage and biogenesis of LDs in a variety of cell types 13,21,46,47 . Our results show that CB induces PLIN2 recruitment from its membrane pools, suggesting a role for this protein as a nucleation site for formation of new LDs under the stimulus by this sPLA 2 . Moreover, up-regulation of PLIN2 expression observed within 12 h of stimulation by CB suggests a mechanism for LD formation in CB-stimulated macrophages. These results differ from those obtained with MT-III, a pro-inflammatory PLA 2 isolated from Bothrops snake venom, which induced an increase in PLIN2 protein expression in macrophages only 6 h after they were stimulated 47 .
Next, we investigated the mechanisms by which CB stimulates LD formation in macrophages by focusing on major downstream signaling protein kinases that have previously been shown to participate in LD biogenesis induced by inflammatory and infectious stimuli. Our results revealed rapid activation of the signaling proteins PKC, PI3K, MEK1/2 and JNK, but not p38 MAPK or ERK1/2, in macrophages activated by CB. While inhibition of PKC, PI3K, MEK1/2 and JNK abolished the CB-induced increase in LD formation, indicating that these signaling proteins contribute to this increase, p38 MAPK and ERK1/2 do not play a role in CB-induced LD formation. In light of these results, it is reasonable to suggest that PKC and PI3K are crucial in LD formation induced by venom sPLA 2 s since both proteins are involved in LD formation induced by both Crotalus (CB) and Bothrops sPLA 2 s (MT-III and MT-II) 47,48 . Our finding that CB-induced LD formation requires activation of the PI3K pathway is in accordance with data in the literature showing that this signaling protein plays a key role in pathways that up-regulate proteins associated with lipid accumulation such as peroxisome proliferator-activated receptor-gamma (PPAR-γ) and PLIN2 58,59 . In this context, our finding that CB-induced LD formation in macrophages is largely dependent on both the MEK1/2 and JNK pathways agrees with the literature, which indicates that MEK1/2 and JNK are important signaling proteins implicated in foam-cell formation 60,61 .
The literature shows that group IIA sPLA 2 s interact with intracellular PLA 2 s in the biosynthesis of lipid mediators from fatty acids 6 and also that LD formation involves the activation of intracellular PLA 2 s, such as iPLA 2 and cPLA 2 , as well as other classes of intracellular phospholipases, including PLC and PLD [47][48][49][50] . Our finding that inhibition of iPLA 2 by BEL and FKGK11 reduced CB-induced LD biogenesis indicates that this intracellular PLA 2 contributes to the effect of CB. However, although cPLA 2 has been reported to be relevant to LDs formation induced by diverse PAMPS stimuli and a snake venom inflammatory sPLA 2 25, 47 this intracellular PLA 2 does not participate in the genesis of LDs induced by CB in macrophages. A recent study showed the important role of iPLA 2 in fatty acid metabolism and triacylglycerol formation 62 . Since CB induces the release of high levels of arachidonic acid in macrophages 7 , our data suggest that upon stimulation by CB this fatty acid contributes to the synthesis of triacylglycerol, which in turn is an important component of LDs. However, the mechanism by which CB activates iPLA 2 is unknown and will be investigated in future studies by our group. Additionally, our data show that PLD, but not PLC, is an important step in the signaling pathways involved in the CB-induced effect. It has been shown that PLD acts on the membrane lipid phosphatidylcholine to produce the signal molecule phosphatidic acid (PA), which is a negatively charged phospholipid whose small headgroup promotes membrane curvature, allowing lipid accumulation at this local and facilitating the budding of LDs 63 . In addition, PA can be converted to bioactive lipids, such as diacylglycerol and lysophosphatidic acid 64 . Therefore, any or all of these mechanisms may act in conjunction to mediate LD formation. This is the first demonstration that a group IIA sPLA 2 crosstalks with PLD, activating LD formation. This finding agrees with an earlier study which showed that activation of PLD 2 was induced by a murine group IIA sPLA 2 , indicating that there is crosstalk between PLA 2 s and PLD 65 .
A previous study by our group showed that CB induces production of high levels of PGD 2 by macrophages 7 . It is known that PGD 2 can be chemically modified by non-enzymatic dehydration reactions, leading to production of the cyclopentenone prostaglandin 15-deoxy-Δ 12, 14 PGJ 2 (15-d-PGJ 2 ), an important lipid mediator involved in the resolution phase of inflammation 11,12 . In line with our previous findings, the results of the present study show that CB-stimulated macrophages release significant amounts of 15-d-PGJ 2 at all the time points, in a process dependent on COX-1 pathway and the terminal PGDS. Furthermore, our data show that (i) CB-induced LDs compartmentalizes both COX-1 and 15-d-PGJ 2 and (ii) CB-induced 15-d-PGJ 2 is abrogated when LD formation Values represent means ± SEM for three to five animals. *p < 0.05 compared is inhibited by pretreatment of macrophages with inhibitors of PKC and JNK, upstream and downstream signaling proteins, respectively, that were shown to inhibit LD formation induced by CB. In light of these data, and considering that COX-1, but not COX-2, is activated by CB in macrophages 7 , and as LDs are important organelles for eicosanoid synthesis 66,67 , it is reasonable to suggest that CB-induced LDs are sites for 15-d-PGJ 2 synthesis via the COX-1 enzymatic system. To our knowledge, this is the first demonstration that LDs induced by a group IIA sPLA 2 are equipped with a COX-1 enzymatic system and are involved in synthesis of a lipid mediator implicated in inflammatory resolution, such as 15-d-PGJ 2 . In addition, it should be noted that release of the anti-inflammatory mediator 15-d-PGJ 2 is in line with the immunomodulatory effects of CB and sheds new light on the mechanisms by which whole Crotalus snake venom and CB, its major sPLA 2 , can reduce or abolish inflammatory events 5,68 .
In conclusion, our data show that group IIA sPLA 2 CB directly activates peritoneal macrophages to form LDs by a mechanism that is dependent on increased expression and distribution of PLIN2 and requires activation of PKC, PI3K, MEK1/2 and JNK. We also showed that this venom sPLA 2 crosstalks with iPLA 2 and PLD to induce LDs formation. Our results also indicate for the first time that LDs can constitute platforms implicated in the synthesis of anti-inflammatory lipid mediators such as 15-d-PGJ 2 in macrophages stimulated by a sPLA 2 . Therefore, the LDs, classicaly recognized as relevant organelles of the acute phase of inflammation, can be involved in both development and resolution of the inflammatory process. Finally, our results reinforce previous findings that phospholipases A 2 from snake venom can induce LDs formation in immunocompetent cells and afford new insights into the roles of LDs in the inflammatory process.