Lutzomyia longipalpis Saliva Triggers Lipid Body Formation and Prostaglandin E2 Production in Murine Macrophages

Background Sand fly saliva contains molecules that modify the host's hemostasis and immune responses. Nevertheless, the role played by this saliva in the induction of key elements of inflammatory responses, such as lipid bodies (LB, also known as lipid droplets) and eicosanoids, has been poorly investigated. LBs are cytoplasmic organelles involved in arachidonic acid metabolism that form eicosanoids in response to inflammatory stimuli. In this study, we assessed the role of salivary gland sonicate (SGS) from Lutzomyia (L.) longipalpis, a Leishmania infantum chagasi vector, in the induction of LBs and eicosanoid production by macrophages in vitro and ex vivo. Methodology/Principal Findings Different doses of L. longipalpis SGS were injected into peritoneal cavities of C57BL/6 mice. SGS induced increased macrophage and neutrophil recruitment into the peritoneal cavity at different time points. Sand fly saliva enhanced PGE2 and LTB4 production by harvested peritoneal leukocytes after ex vivo stimulation with a calcium ionophore. At three and six hours post-injection, L. longipalpis SGS induced more intense LB staining in macrophages, but not in neutrophils, compared with mice injected with saline. Moreover, macrophages harvested by peritoneal lavage and stimulated with SGS in vitro presented a dose- and time-dependent increase in LB numbers, which was correlated with increased PGE2 production. Furthermore, COX-2 and PGE-synthase co-localized within the LBs induced by L. longipalpis saliva. PGE2 production by macrophages induced by SGS was abrogated by treatment with NS-398, a COX-2 inhibitor. Strikingly, SGS triggered ERK-1/2 and PKC-α phosphorylation, and blockage of the ERK-1/2 and PKC-α pathways inhibited the SGS effect on PGE2 production by macrophages. Conclusion In sum, our results show that L. longipalpis saliva induces lipid body formation and PGE2 production by macrophages ex vivo and in vitro via the ERK-1/2 and PKC-α signaling pathways. This study provides new insights regarding the pharmacological mechanisms whereby L. longipalpis saliva influences the early steps of the host's inflammatory response.


Introduction
To obtain a blood meal, sand flies locate blood by introducing their mouthparts into the vertebrate host's skin, tearing tissues, lacerating capillaries and creating hemorrhagic pools upon which they feed. During this process, sand flies need to circumvent a number of the host's homeostatic responses, such as activation of blood coagulation cascades, vasoconstriction, platelet aggregation and immune responses [1,2]. In this environment, sand flies evolved an array of potent pharmacologic components with redundant and synergistic activities that subvert the host's physiological responses and favor the blood meal. Intense research using high-throughput analyses has been conducted to identify salivary factors and their biological activities. Lutzomyia (L.) longipalpis, the main vector of visceral leishmaniasis in South America, has been extensively studied. During the inflammatory response, L. longipalpis saliva induces cellular recruitment, modulates both antibody production and the formation of immunocomplexes [3,4], regulates T cell activities and inhibits dendritic cells and macrophages, the latter being preferential host cells for Leishmania [5,6]. There is also evidence that maxadilan, a L. longipalpis salivary protein with vasodilator properties, downregulates LPS-induced TNF-a and NO release through a mechanism dependent on PGE 2 and IL-10 [7]. PGE 2 is an eicosanoid derived from arachidonic acid (AA) metabolism by the enzyme cyclooxygenase (COX). Prostanoids and leukotrienes can be intensely produced by macrophages during inflammatory responses [8], and these mediators are implicated in cellular recruitment and activation. Among the eicosanoids, LTB 4 induces neutrophil recruitment [9], whereas PGE 2 and PGD 2 attract mainly macrophages [10]. Previous studies used different experimental models to show that L. longipalpis saliva induces an influx of neutrophils [11] and macrophages [12], but neither the role of saliva in LTB 4 and PGE 2 release nor the involvement of these mediators in this process has been fully addressed.
Under inflammatory and infectious conditions, prostaglandins and others lipid mediators are mainly produced by cytoplasmic organelles called lipid bodies (LB) [13]. Intense research over the past few years has defined lipid bodies as dynamic cytoplasmic organelles. It has been demonstrated that lipid bodies compartmentalize enzymes involved in the biosynthesis, transport and catabolism of lipids, proteins involved in membrane and vesicular transport and proteins involved in cell signaling and inflammatory mediator production, including eicosanoid-forming enzymes, phospholipases and protein kinases. All of these molecules can be localized into lipid bodies in various cells under a range of activation conditions, suggesting a wide role for lipid bodies in the regulation of cellular lipid metabolism and signaling [13].
Herein, we evaluated the effect of L. longipalpis salivary gland sonicate (SGS) on the induction of LB formation as well as PGE 2 and LTB 4 production in vitro and ex vivo. Moreover, we explored the role of peritoneal macrophages in the production of these lipid mediators in response to L. longipalpis SGS in vitro. Finally, we found that the PGE 2 production induced by L. longipalpis saliva is dependent on intracellular mechanisms involving the phosphorylation of signaling proteins such as PKC-a and ERK-1/2 and subsequent activation of COX-2.

Mice
Inbred male C57BL/6 mice, age 6-8 weeks, were obtained from the animal facility of Centro de Pesquisas Gonçalo Moniz, Fundação Oswaldo Cruz (CPqGM-FIOCRUZ, Bahia, Brazil). All experimental procedures were approved and conducted according to the Animal Care and Using Committee of the FIOCRUZ.

Sand flies and preparation of salivary glands
Adult Lutzomyia longipalpis captured in Cavunge (Bahia, Brazil) were reared at the Laboratório de Imunoparasitologia/CPqGM/ FIOCRUZ (Bahia, Brazil) as described previously [3]. Salivary glands were dissected from 5-to 7-day-old L. longipalpis females under a Stemi 2000 Carl Zeiss stereoscopic microscope (Göttingen, Germany) and stored in groups of ten pairs in 10 mL of endotoxin-free PBS at 270uC. Immediately before use, the glands were sonicated with a Branson Sonifier 450 (Danbury, CT) and centrifuged at 10,0006 g for four minutes. The supernatant from salivary gland sonicate (SGS) was used for experiments. The level of LPS contamination of L. longipalpis SGS preparations was determined using a commercially available LAL Chromogenic Kit (Lonza Bioscience, Walkersville, MD); negligible levels of endotoxin were found in the salivary gland supernatant (0.1 gg/mL). We measured 0.7 micrograms of protein in an amount equivalent to 0.5 pair of salivary glands and used SGS dilutions (2.0-0.2 pairs) in our experiments [14].

Leukocyte recruitment to the peritoneal cavity
To assess the leukocyte recruitment induced by L. longipalpis SGS, we used the well-established peritoneal model of inflammation because the peritoneal cavity is a self-contained and delineated compartment and thus provides a large number of post-stimulus leukocytes. As previously established in the air pouch murine model [12] and peritoneal cavity (unpublished data), a 0.5pair dose of SGS was used for the leukocyte recruitment assay. C57BL/6 mice were inoculated i.p. with 0.1 mL of L. longipalpis SGS (0.5 pair/cavity), endotoxin-free saline (negative control) or 0.1 mL of LPS (20 mg/mL, positive control). At 1, 3 and 6 h poststimulus, leukocytes inside the peritoneal cavity were harvested by

Author Summary
After the injection of saliva into the host's skin by sand flies, a transient erythematous reaction is observed, which is related to an influx of inflammatory cells and the release of various molecules that actively facilitate the blood meal. It is important to understand the specific mechanisms by which sand fly saliva manipulates the host's inflammatory responses. Herein, we report that saliva from Lutzomyia (L.) longipalpis, a widespread Leishmania vector, induces early production of eicosanoids. Intense formation of intracellular organelles called lipid bodies (LBs) was noted within those cells that migrated to the site of saliva injection. In vitro and ex vivo, sand fly saliva was able to induce LB formation and PGE 2 release by macrophages. Interestingly, PGE 2 production induced by L. longipalpis saliva was dependent on intracellular mechanisms involving phosphorylation of signaling proteins such as PKC-a and ERK-1/2 and subsequent activation of cyclooxygenase-2. Thus, this study provides new insights into the pharmacological properties of sand fly saliva and opens new opportunities for intervening with the induction of the host's inflammatory pathways by L. longipalpis bites.
Sand Fly SGS Triggers Eicosanoid Production www.plosntds.org injection and recovery of 10 mL of endotoxin-free saline. Total counts were performed on a Neubauer hemocytometer after staining with Turk's solution. Differential cell counts (200 cells total) were carried out microscopically on cytospin preparations stained with Diff-Quick.

Lipid body staining and quantification
Cells harvested by peritoneal lavage 1, 3, 6 or 24 h after i.p. injection of 0.1 mL of L. longipalpis SGS (0.5 pair/cavity), endotoxin-free saline or LPS (20 mg/mL) were centrifuged at 4006 g and the lipid bodies within the leukocytes were stained with BODIPY 493/503 (5 ug/mL) according to Plotkowisk et al. [15]. Samples were analyzed using a FACSort flow cytometer from Becton Dickinson Immunocytometry Systems (San Jose, CA) and by fluorescence microscopy.
Macrophages adhered to coverslips within 24-well plates were fixed with 3.7% formaldehyde and stained with osmium tetroxide as described previously [16]. The morphology of the fixed cells was observed, and lipid bodies were counted by light microscopy with a 100x objective lens in 50 consecutively scanned macrophages.

Resident peritoneal macrophage harvesting and treatments
For in vitro assays, macrophages were obtained by peritoneal lavage with cold RPMI 1640. Then, cells were centrifuged at 4006 g for 10 minutes. Macrophages (3610 5 /well) were cultured in 1 mL of RPMI 1640 medium supplemented with 1% Nutridoma-SP, 2 mM L-glutamine, 100 U/mL penicillin and 100 mg/mL streptomycin in 24-well plates for 24 hours. Next, the macrophages were stimulated with different doses of L. longipalpis SGS (0.2, 0.5, 1.0, 1.5, 2.0 pairs/well). In some experiments, LPS (500 ng/well) was used as a positive control. One, 6, 24, 48 and 72 hours after stimuli, supernatants were collected and cells were fixed with 3.7% formaldehyde. For inhibitory assays, macrophages were pretreated for one hour with 1 mM NS-398, a COX-2 inhibitor; 20 gM BIS, a PKC inhibitor; or 50 mM PD98059, an ERK-1/2 inhibitor. Then, the cells were stimulated with SGS (1.5 pairs/well) or medium containing vehicle (DMSO) for 24 hours, and the supernatants were collected for eicosanoid measurement. Cell viability as assessed by trypan blue exclusion was always greater than 95% after the end of treatment.
Immunofluorescence for COX-2 and PGE-synthase Resident peritoneal macrophages were cultured on coverslips in the presence of L. longipalpis SGS (1.5 pair/well) as described above. After 24 h, the cells were washed twice with 500 ml of HBSS 2/2 and immediately fixed with 500 mL of water-soluble EDAC (1% in HBSS 2/2 ), used to cross-link eicosanoid carboxyl groups to amines in adjacent proteins. After 15 min of incubation at room temperature (RT) with EDAC to promote both cell fixation and permeabilization, macrophages were then washed with HBSS 2/2 and incubated with 1 mM BODIPY 493/503 for 30 min. Then, the cover slips were washed with HBSS 2/2 and incubated with mouse anti-COX-2 (1:150) or anti-PGE-synthase (1:150) for 1 h at RT. MOPC 21 (IgG1) was used as a control. After further washes, cells were incubated with biotinylated goat anti-rabbit IgG secondary Ab, washed twice and incubated with avidin conjugated with PE for 30 min. The cover slips were then washed three times and mounted in Vectashield medium containing DAPI (Vector Laboratories, Burlingame, CA). The samples were observed by fluorescence microscopy and images were acquired using the software Image-Pro Plus (Media Cybernetics, Silver Spring, MD).

Western blotting analysis
Macrophages were treated or not with SGS (1.0 pair/well) for 40 min. Next, the cells were washed once with phosphate-buffered saline, homogenized in lysis buffer containing phosphatase inhibitors (10 mM TRIS-HCl, pH 8.0, 150 mM NaCl, 0.5% v/ v Nonindet-P40, 10% v/v glycerol, 1 mM DTT, 0.1 mM EDTA, 1 mM sodium orthovanadate, 25 mM NaF and 1 mM PMSF) and a protease inhibitor cocktail (Roche, Indianapolis, IN). Protein concentrations were determined using the method of Lowry et al. [17] with BSA as the standard. Total proteins (20 mg) were then separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described previously [18] and transferred onto nitrocellulose membranes. The membranes were blocked in Tris-buffered saline (TBS) supplemented with 0.1% Tween 20 (TT) plus 5% BSA for 1 h before incubation overnight in the primary rabbit anti-mouse PKC-a and anti-ERK-1/2 (1:1,000) antibodies. After removal of the primary antibody and washing five times in TT, the membranes were incubated in the secondary antibody conjugated to peroxidase (1:10,000) for 1 h. Washed blots were then incubated with an ECL chemiluminescence kit (Amersham, UK). The membranes were discharged and immunoblotted again using primary rabbit anti-mouse phosphorylated-PKC-a and ERK-1/2 (1:1,000) antibodies according to the manufacturer's instructions (Amersham, UK).
Quantification of the level of proteins in the western blotting membranes was determined by densitometry. Briefly, bands were scanned and processed using Adobe Photoshop 5.0 software (Adobe Systems Inc.), and arbitrary values for protein density were estimated. Ratios between phosphorylated and unphosphorylated proteins were obtained to calculate the difference between groups.

PGE 2 and LTB 4 measurement
C57BL/6 mice were inoculated i.p. with 0.1 mL of L. longipalpis SGS (0.5 pair/cavity), endotoxin-free saline or 0.1 mL of LPS (500 gg/mL). At 1, 3 and 6 h post-stimulus, leukocytes were harvested by peritoneal washing with HBSS 2/2 and 1610 6 cells/mL were resuspended in HBSS +/+ and stimulated with A23187 (0.5 mM) for 15 min [16]. The reactions were stopped on ice, and the samples were centrifuged at 5006 g for 10 min at 4uC. Supernatants from leukocytes re-stimulated ex vivo or those of in vitro assays were collected for measurement of PGE 2 and LTB 4 by enzyme-linked immunoassay (EIA) according to the manufacturer's instructions (Cayman Chemical, Ann Arbor, MI).

Statistical analysis
The in vivo assays were performed using at least five mice per group. Each experiment was repeated at least three times. Data are reported as the mean and standard error of representative experiments and were analyzed using GraphPad Prism 5.0 software. Disparities in leukocyte recruitment, lipid bodies and lipid mediator quantification were explored using Student's t test. Means from different groups from the in vitro assays were compared by ANOVA followed by Bonferroni's test or a posttest for linear trends. Differences were considered statistically significant when p#0.05.

Lipid bodies and eicosanoids in leukocytes recruited by L. longipalpis SGS
To measure the leukocyte recruitment induced by SGS, we injected 100 mL of saline or SGS (0.5 pair/cavity), and 1, 3 and 6 hours after injection, we enumerated total leukocytes recruited to the peritoneal cavity. Most of the cells recruited were mononuclear cells and neutrophils (Figure 1). In this context, SGS induced mononuclear cell recruitment for 3 hours (Figure 1  A and B) and neutrophil recruitment for over 6 hours ( Figure 1A-C) of stimulation when compared with the saline group. Other cell populations (eosinophils and mast cells) were not altered after SGS stimulation, and there was no variation in these numbers over time (Figure 1). The peritoneal cell population in unstimulated animals (time zero) was composed of mononuclear cells (2.985610 4 60.027) and negligible amounts of neutrophils (0.018610 4 60.027). At this time, macrophages are the major cells within    Sand Fly SGS Triggers Eicosanoid Production www.plosntds.org the mononuclear population in the peritoneal cavity besides lymphocytes, which represent ,10% of mononuclear cells (data not shown). As shown in Figure 2, SGS administration led to enhanced PGE 2 ( Figure 2A) and LTB 4 ( Figure 2B) release within those cells recruited to the peritoneal cavity.
Because LBs are sites of eicosanoid production [19], we evaluated LB formation in leukocytes recruited to the peritoneal cavity by FACs using the neutral lipid probe BODIPY 493/503. The kinetics of LB formation was evaluated at 1, 3, 6 and 24 hours after SGS stimulation by measuring mean fluorescence intensity (MFI). SGS increased MFI in mononuclear but not in polymorphonuclear cells after 3 and 6 hours, (Figure 3A and B) compared with the saline group. Histograms ( Figure 3C and D) and fluorescence microscopic images (Figures 3E and F) at the 3-hour time point confirmed these effects of SGS on macrophages.

L. longipalpis SGS induces macrophage PGE 2 production via the COX-2 enzyme
Prostaglandins are produced by cyclooxygenases, which occur in constitutive (COX-1) and inducible (COX-2) forms [20]. We investigated the expression and subcellular localization of COX-2 within SGS-stimulated macrophages. Immunofluorescence microscopy revealed the presence of COX-2 ( Figure 5A-C) and PGE-synthase ( Figure 5D-F) within LBs in macrophages stimulated with SGS.
Next, we measured PGE 2 and LTB 4 production in the supernatant of macrophage cultures. SGS induced PGE 2 production starting at 1.0 pair/well ( Figure 6A), whereas LTB 4 was not detectable under any conditions (data not shown). As expected, PGE 2 production by macrophages stimulated with SGS was reduced to basal levels when the cells were pre-incubated with NS-398, a COX-2 inhibitor ( Figure 6B). Thus, the PGE 2 production in peritoneal macrophages induced by SGS occurs in newly formed lipid bodies and is dependent on COX-2.

SGS induces PGE 2 production via PKC-a and ERK-1/2
Multiple pathways are involved in the signaling for PGE 2 production [13]. Recently, ERK and PKC-a were shown to be involved in COX-2 activity [21]. We observed that SGS activated both ERK (Figure 7A and C) and PKC-a phosphorylation ( Figure 7B and D), but it did not alter the levels of the unphosphorylated proteins. To investigate whether these kinases are involved in the induction of PGE 2 production by SGS, we pretreated macrophages with bisindolylmaleimide I (BIS I) and PD98059, PKC-a and ERK-1/2 inhibitors, respectively ( Figure 8A-B). Inhibition of both enzymes completely abrogated PGE 2 production induced by SGS ( Figure 8A-B). In sum, these results suggest that PKC-a and ERK-1/2 are involved in the PGE 2 production induced by SGS.

Discussion
Sand fly saliva triggers an inflammatory response characterized by cellular influx followed by hemostatic and immune mechanism suppression. Nevertheless, the role of sand fly saliva in eicosanoid production during the early steps of the innate immune response is poorly understood. In inflammatory conditions, eicosanoids are mostly produced in cytoplasmic organelles called lipid bodies (LBs), which are formed in leukocytes and other cells involved in the inflammatory and infectious responses to several stimuli [13]. Herein, we showed that L. longipalpis saliva induces lipid body formation and PGE 2 production in peritoneal macrophages ex vivo and in vitro via kinase phosphorylation and COX-2 activation.
Previous investigations have demonstrated that sand fly saliva plays an important role in cellular recruitment in multiple experimental models [3,9,11,12], including in vivo sand fly bites [22]. Herein, we confirmed previous reports that L. longipalpis SGS induces an inflammatory infiltration composed mainly of macrophages and neutrophils. Moreover, we showed that the cellular recruitment induced by L. longipalpis saliva is concomitant with PGE 2 and LTB 4 production. In this scenario, lipid mediators Sand Fly SGS Triggers Eicosanoid Production www.plosntds.org could be triggering cellular recruitment. Secretion of LTB 4 by resident macrophages plays an important role in neutrophil migration [23]. In addition, lipopolysaccharides induce macrophage migration via prostaglandin D 2 and prostaglandin E 2 [10].
Prostaglandin E 2 is an abundant eicosanoid produced by inflammatory cells, and it is known to exert anti-inflammatory and vasodilator effects. PGE 2 is found in Ixodes scapularis saliva and is also implicated in the immunomodulatory activity of tick saliva on dendritic cell and macrophage activation [24]. Furthermore, previous studies using saliva from several Phlebotomus species have suggested that the anti-inflammatory properties of sand fly saliva could be attributed to PGE 2 and IL-10 released by dendritic cells [9,25]. In these studies, the cellular recruitment induced by OVA stimulation was abrogated by saliva from various sand fly species [9,25], which was associated with an anti-inflammatory profile dependent on the production of IL-10, IL-4 [25] and PGE 2 [9]. Intriguingly, maxadilan, a vasodilator peptide with immunomodulatory activities present in L. longipalpis saliva, is able to induce LPS-activated macrophages to release PGE 2 via COX-1, an enzyme that is constitutively active [7]. In the present study, we showed that L. longipalpis SGS triggers PGE 2 production in resident macrophages by an inducible pathway, since this effect was completely abrogated when the cells were incubated in the presence of NS-398, a COX-2 inhibitor. Nevertheless, whether sand fly saliva contains other molecules involved in PGE 2 production or pharmacological amounts of this mediator similarly to tick saliva remains unknown.
Our study is the first to establish a direct link between L. longipalpis saliva, eicosanoid production and lipid body formation. Under inflammatory and infectious conditions, lipid mediators are mainly produced within LBs, which compartmentalize both the substrate and the enzymatic machinery required for eicosanoid production [13]. In this regard, the enzymes COX and 5-LO have been localized to lipid bodies in various inflammatory cells by the use of multiple techniques including fluorescence microscopy [13]. Previous studies have shown that various inflammatory and infectious stimuli are able to trigger LB formation in macrophages [13,19]. Our findings demonstrate that SGS induces LB formation in macrophages in vivo and in vitro, suggesting that L. longipalpis saliva acts directly on these cells, but not on neutrophils. Indeed, L. longipalpis SGS triggered LB formation in macrophages committed to PGE 2 production via COX-2 and PGE-synthase.
Data regarding the direct effects of sand fly salivary compounds on host signaling pathways cells are scarce. The extracellular signal-regulated kinases (ERKs) and protein kinase C (PKC) are among the key enzymes implicated in signaling pathways of diverse cellular responses, including eicosanoid production. The MAP kinases ERK1 and ERK2 induce activation of cPLA2, an enzyme that hydrolyzes arachidonic acid, which is metabolized to  prostaglandin H2 by COX [13]. Previous studies have demonstrated the compartmentalization of MAP kinases and cPLA2 at arachidonate-enriched lipid bodies [26,27], as well as COX-2 and PGE-synthase [16,28,29]. Herein, it is shown for the first time that L. longipalpis SGS triggers ERK-1/2 and PKC-a phosphorylation in macrophages. Other studies have shown that COX-2 activation and PGE 2 production in LPS stimulated-macrophages is dependent on the phosphorylation of protein kinases such as PKC-a [21] and ERK-1/2 [30]. We showed that the PGE 2 production induced by SGS is dependent on both ERK-1/2 and PKC. This association between the activation of kinases and the metabolism of eicosanoids within lipid bodies may serve to enhance rapid eicosanoid production in response to extracellular stimuli such as sand fly saliva. Of note, in addition to their role in regulating the host response to infection by modulating inflammatory mediator production, lipid bodies may also serve as rich sources of nutrients for intracellular pathogens, thus favoring intracellular pathogen replication [31,32].
In brief, the present work provides new insights into the mechanisms involved in macrophage responses to L. longipalpis saliva, including LB formation and the signaling pathways that trigger PGE 2 release. Although the roles of the newly formed LBs and PGE 2 induced by sand fly saliva in the pathogenesis of leishmaniasis have not yet been addressed, several studies have shown that PGE 2 is essential to the infection of macrophages [33,34] and parasite dissemination after infection [35]. The induction of PGE 2 production by sand fly saliva demonstrated herein can influence the initial steps of host infection by favoring less intense macrophage activation. Our group and others have been providing strong evidence that saliva components are immunogenic and have potential as markers of exposure to sand fly vectors [36][37][38][39]. Further studies are required to determinate if the immunization based on components of vector saliva interferes in eicosanoid production with consequences for the host's immune response and the transmissibility of the parasite.