The Schistosoma mansoni lipidome: Leads for immunomodulation

(cid:1) Comprehensive Lipidomics study of Schistosoma mansoni and its life stages. (cid:1) Description of a Schistosoma mansoni lipid database. (cid:1) Identi ﬁ cation of several important immunomodulatory lipids, leads for further research. cercariae, worms and eggs, as well as extracts and excretory/secretory (ES) products of different life cycle stages of S. mansoni . We identi ﬁ ed life cycle stage speci ﬁ c signatures of lipid classes of which cercariae were found to have the most distinct pro ﬁ le. Moreover, we detected several immunolomodulatory oxylipids in the different life cycle stages. Speci ﬁ cally, prostaglandins were found to be most highly enriched in egg preparations, while resolvins were speci ﬁ cally detected in cercariae. Together, the generation of this detailed lipid database of the different life cycle stages of S. mansoni will not only be important for a better understanding of the biology of the parasite itself but also of host-parasite interactions and how that could result in immunomodulation.


Introduction
Schistosoma mansoni is a parasitic trematode causing schistosomiasis in humans and occurs in mostly tropical parts of South America and Africa. According to the WHO more than 61 million people have been treated for this disease in 2014 [1] with estimates of between 20000 and 200000 deaths per year. The life cycle of this parasite is complex and involves both humans as a final host as well as freshwater snails as intermediate host. Infection of humans is initiated by the larval stage of these parasites, termed cercariae, through their ability to penetrate human skin. Upon penetration cercariae lose their tail, allowing their head to develop into schistosomula that migrate through the skin into the circulation. Two to 3 weeks after the initial infection, schistosomula end up in the hepatoportal circulation where they develop into sexually mature adults and pair for sexual reproduction. Subsequently, approximately 6e8 weeks into the infection, females start to produce eggs, of which around 50% break through the wall of the intestine to reach the outside environment. However, a large fraction of the eggs get lodged in the intestinal wall or are instead carried by the blood flow into the liver, where they become trapped in hepatic sinusoids. This accumulation of eggs in tissues is the major cause of pathology in schistosomiasis [2].
The immune responses associated with this infection are equally complex. Infections with cercariae initially trigger an immune reaction characterized by a T helper 1 (Th1) response. However, upon egg production, a strong Th2 response is induced that subsequently orchestrates the development of granulomatous lesions surrounding the eggs. S. mansoni infections are often chronic in nature, which in part is thought to be due to the ability of these parasites to promote regulatory responses that can dampen anti-parasitic type 2 inflammation [3,4]. Importantly, evidence is accumulating that this induction of regulatory immune responses by S. mansoni, may not only impair effector responses directed against the parasite, but may also be beneficial to the host as it can concomitantly lead to bystander suppression of other inflammatory immune responses. For instance, population studies have revealed that these parasitic infections, can reduce various parameters associated with allergic responses [5,6]. These observations have been corroborated in animal models [7,8] and further extended to models of other inflammatory disorders including colitis [9] and type 2 diabetes [10]. This illustrates that different life cycle stages of S. mansoni that infect or develop in the human host, promote distinct immune responses, that are not only important in mediating immunity against the infection and in determining the immunopathological outcome of the disease, but that can also have beneficial effects on development of various unrelated inflammatory disorders. Therefore, it will be important to define the components and antigens from the different life cycle stages that may drive these distinct immune responses in humans. This could be of great value for vaccine development against S. mansoni, but could also contribute to the design of therapeutics to treat inflammatory diseases by exploiting the immunoregulatory potential of some helminth-derived molecules. Thus far, to identify S. mansoniderived molecules with immunomodulatory potential, most studies have focused on proteomics [11,12] or glycomics [13,14]. Much less is known about S. mansoni-derived lipids in this context. Already in the late 1960s studies were undertaken to characterize the major lipid classes present in S. mansoni worms [15]. Subsequent studies revealed that apart from being constituents of membranes, lipid metabolism in S. mansoni also has important roles in development of the different life cycle stages [16] and in evasion of immune responses by adult worms [17] or cercariae [18]. The potential role of lipids interacting with the immune system comes from a study that has directly linked S. mansoni Lyso-PS, to immunomodulation by induction of regulatory T cells [19]. Moreover, there is some evidence to suggest that S. mansoni can produce eicosanoids [20], which are bioactive lipids well recognized for their immunomodulatory capacity. With these studies in mind, it is important to first conduct a comprehensive in-depth characterization of the lipid composition of the different classes of lipids of the distinct life cycle stages of S. mansoni. Studies that have performed lipid profiling of S. mansoni so far, were either based on a limited fingerprinting approach leading to the identification of a very small number of lipids [21e23] or focused specifically on the phospholipid content of only the worm itself [23,24]. Using three complementary highly sensitive MS-based lipidomics platforms, we present a comprehensive lipidomics analysis of S. mansoni cercariae, adult worm and eggs, as well as of typical extracts of cercariae (cercarial antigen, CA), worms (adult worm antigen, AWA) and eggs (soluble egg antigen, SEA) and their excretory and secretory products (ES), each widely used in immunological studies. We identified life cycle stage specific lipid signatures not only for membrane and storage lipids but in particular for bioactive lipid mediators, which provides potentially interesting new leads to study the link between the S. mansoni lipidome and immunomodulation.

Animal and parasite materials
The full life cycle of the Puerto Rican strain of S. mansoni used in this study was maintained in Biomphalaria glabrata as snail intermediate host and male Golden hamsters (RjHan: AURA, Javier labs) as mammalian definitive host. For infections, male hamsters 9.5 weeks old and weighing ±100 g were exposed to an aqueous suspension containing 1200 cercariae of S. mansoni freshly shed by B. glabrata infected hamsters were maintained in a controlled environment at room temperature (RT) with normal night and daylight cycle. All hamster experiments were performed in accordance with local government regulations, and the EU Directive 2010/63EU and Recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes and approved by the CCD. Preparation of parasite samples for lipidomics analysis is described below. For each sample three biological replicates were generated to account for possible parasite batch variations.

Cercariae
S. mansoni-infected B. glabrata snails kept in trays were transferred to a cup containing 30 mL of mineral water, and exposed to light at 28 C for 2 h to trigger cercariae shedding. Cercariae were sedimented by cooling on ice, and stored in water at À80 C until use. Alternatively, as described below, live cercariae were cultured to produce cercarial ES products.

Eggs
Eggs were obtained as described before (Dalton et al. 1997), with some modification. Briefly, the livers of 10 hamsters that were infected with S. mansoni 6 weeks before, were homogenized in 300 mL of wash buffer (1.7% NaCl PBS, to prevent hatching of eggs). The suspension was poured into a 1-L screw cap bottle containing 800 mL wash buffer, collagenase B (400 mg), streptomycin and penicillin (750 mL 200.000 U/ml, each) and gently stirred at 37 C overnight. Then the suspension was sequentially sieved through 355 and 200 mm sieves, rinsed with wash buffer, and transferred to 50 mL tubes. After centrifugation for 5 min at 400 g, RT the pellets were collected, resuspended in wash buffer, transferred to new 50 mL tubes followed by centrifugation as above. These steps were repeated until the supernatant became clear, with intermediate application to an 80 mm sieve. The isolated eggs were resuspended in wash buffer containing EGTA and EDTA to inactivate collagenase, washed three times with this buffer at 60 g for 3 min, RT. The eggs were collected and counted prior to a final wash with wash buffer at 1000 rpm, for 1 min, RT. Aliquots of approximately 50.000 eggs were stored at À80 C in PBS until use. In addition, live eggs were cultured to generate egg ES, as described below.

Adult worms
Mixed male and female adult worms were obtained from hamsters sacrificed 6 weeks after infection with S. mansoni, through perfusion of the hepatic portal system and mesenteric veins. Collected live worms were washed with perfusion buffer (Dulbecco's buffered saline and sodium citrate) to remove blood and debris. Worms were then washed with PBS and stored at À80 C until use, or cultured to collect worm ES as described in following section. Frozen cercariae or eggs were thawed and manually homogenized in a sterilized glass homogenizer on ice. Homogenates were subsequently subjected to sonification (six rounds of 30 s, with 20 s intervals) (Branson sonifier) on ice. The extracts were kept overnight at À80 C. Following thawing, extracts were centrifuged twice cat 13000 rpm, 4 C for 25 min. The supernatants were collected and filter-sterilized. Protein concentration was determined using BCA (Pierce BCA Protein assay kit, Thermo Scientific) according to the manufacturer's protocol. The protein concentrations of CA and SEA were adjusted to 80 mg/mL and samples were stored at À80 C until use.

Adult worm antigen (AWA).
Worms were lyophilized, suspended in cold PBS and manually homogenized in a sterilized glass homogenizer. The process was continued as for CA and SEA preparation.

Egg ES (EES).
Approximately 300.000 freshly isolated eggs were resuspended in 25 mL 20% percoll and centrifuged at 500 g for 6 min. The eggs were centrifuged for 4 min, at 1100 rpm, RT for four times: two times with 2 mL PBS (1.7% NaCl); one time with RPMI (Gibco) supplemented with penicillin and streptomycin; and one time with up to 40 mL of medium (RPMI supplemented with fungizone, penicillin and streptomycindreferred as E-medium). The eggs were resuspended in 1 mL of E-medium in 24 well plates and incubated for 48 h at 37 C 5% CO 2 . About 800 mL supernatant was collected (EES).

Worm ES (WES).
Live worms were gently washed twice with S-medium (M199 medium (Gibco) supplemented with ABAM (Sigma) and HEPESdreferred as S-medium). In total, approximately 200 mixed male and female worms were resuspended in 20 mL of S-medium and placed in a 75 cm 3 flask. The flask was incubated at 37 C 5% CO 2 for 48 h and the supernatant (WES) was collected.
All ES products were kept at À80 C until further use. Prior to lipidomics analysis, ES preparation concentrations were adjusted to 80 mg/mL of protein. For protein concentration determination by BCA, the medium used for culturing the cercariae, eggs and worms was taken as background.

Sample preparation
Three different types of samples were investigated in this study: 1) whole parasite life cycle stages (cercariae, worms and eggs); 2) water soluble extracts including: CA, AWA, SEA; and 3) ES products (CES, EES and WES). Whole parasite life cycle stages were extracted using 2-propanol after homogenization in a bullet blender for further analysis using mass-spectrometry. In the case of extracts or ES products, 20 mL of sample was directly used for the below described procedures. Briefly: to the parasite materials sample in 150 mL phosphate buffered saline was added 5 stainless steel beads and the sample homogenized for 5 min in a bullet blender. Subsequently 300 mL 2-propanol was added, the sample was vigorously shaken and centrifuged at 16.100 g for 5 min. The supernatant was transferred to a glass vial. The residual sample was re-extracted using 300 mL of 2-propanol and the organic extracts combined.
The combined organic extracts were dried under a gentle stream of nitrogen and dissolved in 200 mL 2-propanol, stored at À80 C until analysis.

GC-MS analysis for total fatty acid analysis
GC-MS analysis was carried out as described elsewhere [26,27], with some modifications. Briefly, to 20 mL sample in a glass vial was added 10 mL of 10 M NaOH and 250 mL acetone. The vial was flushed with nitrogen, tightly closed and incubated at 60 C for 1 h. After cooling to room temperature 10 mL of gcIS solution, followed by the addition of 100 mL of a 172 mM solution of PFBBr in acetone. The samples were subsequently incubated for 30 min at 60 C, 250 mL of water and 500 mL of n-hexane was added and the hexane layer transferred to an autosampler vial. GC-ECNI-MS analysis was carried out on a Bruker scion TQ (Bruker, Bremen, Germany) using methane (99.9995%) as CI gas, equipped with an Agilent VF-5MS (30 m Â 0.25 mm Â 0.25 mm) column. The temperature program was as follows: 1 min 50 C, linear increase at 40 C/min to 60 C, held 3 min at 60 C, linear increase at 25 C/min to 237 C, linear increase at 3 C/min to 250 C, linear increase at 25 C/min to 315 C held for 1.55 min. The transfer line and ionization source temperature were 280 C. The pressure of the chemical ionization gas was set at 15 psi. The carrier gas was helium (99.9990%) at a flow rate of 1.2 mL/min.

LC-MS/MS analysis of eicosanoids and docosanoids
Eicosanoid and docosanoid analysis was carried out as described elsewhere [28]. For liquid samples 20 mL sample was mixed with 60 mL MeOH and 2 mL oxIS solution, the sample was centrifuged for 10 min at 16.100 g and 4 C, 50 mL of the supernatant was dried under a gentle stream of nitrogen and re-dissolved in 50 mL 40% MeOH for injection. Targeted lipidomics analysis was carried out as follows: analysis was achieved using a QTrap 6500 mass spectrometer in negative ESI mode (ESI-) (Sciex, Nieuwerkerk aan den Ijssel, The Netherlands), coupled to a LC system employing two LC-30AD pumps, a SIL-30AC autosampler, and a CTO-20AC column oven (Shimadzu, 's-Hertogenbosch, The Netherlands). The employed column was a Kinetex C18 50 Â 2.1 mm, 1.7 mm, protected with a C8 precolumn (Phenomenex, Utrecht, The Netherlands), kept at 50 C. The following binary gradient of water (A) and MeOH (B) with 0.01% acetic acid was used: 0 min 30% B, held for 1 min, then ramped to 45% at 1.1 min, to 53.5% at 2 min, to 55.5% at 4 min, to 90% at 7 min, and to 100% B at 7.1 min, held for 1.9 min. The injection volume was 40 mL and the flow rate 400 mL/min. The MS was operated under conditions as described before [29]. For quantification calibration lines, constructed with standard material were used and only peaks with a signal to noise (S/N) > 10 were quantified. For analytes where no calibration line was used, area ratios were used and a S/N > 3 was used as a detection limit. A list of all monitored oxylipids, corresponding lipidmaps IDs and calibration ranges can be found in supplementary (Table S1).
The injection volume was 20 mL. Confirmatory analysis on a Triple TOF 6600 using a representative cercaria sample was carried out in ESIþ and ESI-as described in supplementary table S3.  Table 1.

Lipid identification
Within Data Analysis homemade Visual Basic scripts were used to filter all collected MS/MS spectra according to each lipid class separately. The results were exported to mgf format, resulting in a single mgf file per sample per lipid class. Each mgf file was used to search the LipidBlast databases [32]. LipidBlast contains several different libraries, some contain multiple lipid classes others are specific for one lipid class. The top results of all database searches were stored and further processed in R (CRAN R, version 3.3.2). For each sample type three biological sample replicates were measured and only lipids detected in all three samples were deposited in our S. mansoni lipid database (supplementary Table S2). In the case of uncommon lipids such as plasmenyl/plasmanyl species and very long chain LPC (VLC LPC) lipids, the following steps were undertaken in order to ensure unambiguous identification. For VLC LPC, we manually inspected (tandem) MS spectra in the ESI þ mode, ensuring matching molecular masses with the FiehnO lipid database in MS-DIAL (v2.74), next we ensured the presence of the LPC specific ions m/z 104 and 184 (PC species) and finally we inspected the tandem MS spectra obtained from the analysis of a representative cercaria sample in the ESI-mode for the presence of the corresponding fatty acid fragment (please see supplementary material S1 and S3 for several examples). The obtained data were then matched to the other biological samples analyzed in the ESI þ mode, taking high resolution mass and retention time into account. In case of plasmenyl/plasmanyl PC and LPC species we manually inspected all tandem MS spectra in the ESIþ and ESImode (supplementary table S3) and compared measured MS/MS spectra in the negative mode with predicted spectra from Lipidmaps. In addition, we carried out hydrolysis experiments according to [33]. For calculating % lipid composition of all samples we used the MS signals as follows: from the MS/MS data m/z and retention time for each lipid were extracted. Subsequently a homemade Visual Basic script was used to automate the following steps within DataAnalysis: create extracted ion chromatogram (EIC) from the m/z with a narrow window of 5 mDa, set retention time window around the retention time and finally carry out peak detection and integration. Results of each sample were exported to a comma separated (csv) file. All result files were collected and further processed in R.

Analysis and identification of lipid species
Using three lipidomics platforms, which were: QToF based LC-MS/MS for major lipid classes, GC-MS for total fatty acid (FA) analysis and QTrap based LC-MS/MS for eicosanoid and docosanoid analysis, we investigated occurrence and quantity of several lipid classes in different life-cycle stages of S. mansoni that could be of relevance to interaction with the human host. The investigated lipid classes, the employed platform and the number of detected species are given in Table 2.

Global lipid class composition
It was our aim to investigate the lipid content of the different S. mansoni life cycle stages, not restricting our investigation to phospholipids, as done previously [23,24], but also to include neutral lipids such as sterols, sterol esters and TG. Typical lipidomic profiles (ESI þ mode) obtained for cercariae, worms and eggs are shown in Fig. 1. As can be appreciated from Fig. 1, PC and TG lipids are the predominantly detected lipid classes. Specifically, we found PC(34:1), PC(36:1), and PC(36:2) to be the major phospholipid species present in all life cycle stages. The fractional composition shown in Fig. 2 indicates that particularly worms and eggs present similar lipid compositions while cercariae have a somewhat different lipid composition. A more detailed depiction comparing the overlapping individual lipid species within different lipid classes between the distinct life-cycle stages is shown in Fig. 3. Interestingly, we could only detect DG lipids in cercariae, but not in the eggs or worms. The same holds true for the species of lysophospholipids (LPC and LPE), which were mainly seen in cercariae and eggs but hardly in worms. On the other hand, there were lower amounts of neutral lipids of the CE and TG type in cercariae. Besides several very common lipids containing oleic or palmitic acid we could also detect some uncommon lipids, such as very long chain LPCs or PCs (i.e. PC(50:6) or LPC(26:2)) ( Figure S1). Further validation of the identity of these uncommon lipid species using the FiehnO lipid database in MS-DIAL (v2.74) confirmed the identity of LPC(O-16:0) and LPC(P-18:0) (see supplementary Table S3). However, in case of the identified plasmenyl PC, hydrolysation experiments did not result in changes in the observed signals, leading us to question the initial identification of these species by Lipidblast and to remove these species from the list of identified lipids. A list of all identified lipids as well as the relative abundance of each lipid species of the total lipid pool from each life cycle stage can be found  in supplementary Table S2.

Total fatty acid composition
Next we investigated the total FA composition of the different S. mansoni life cycle stages. As can be seen from Table 3, particularly worms contained significant amounts of FA. As expected from the generally low solubility of FA in aqueous media, much lower amounts of FA per mg protein were obtained from AWA and WES.

Oxylipid composition
Finally we analyzed the composition of oxylipids of the different life cycle stages and parasite extracts. Table 4 shows that large amounts of polyunsaturated fatty acids (PUFAs) as well as downstream Cyclo-oxygenase (COX) and Lipoxygenase (LOX) derivatives could be detected in several parasite-derived materials. Particularly, high levels of PUFA precursors such as arachidonic acid (AA), linoleic acids (LA) and docosahexaenoic acid (DHA) were found in all preparations. Furthermore, the overall composition of COX and LOX derivatives of cercariae and eggs was largely similar but distinct to that of worms. However, an interesting difference between cercariae and eggs was the fact that we could detect several proresolving lipids, including Protectin D1 (PD1) and Resolvin E2 (RvE2), as LOX derivatives specifically in cercariae, but not in eggs (Table S2). Fig. 4 illustrates the identification of these substances in the investigated samples and demonstrates that our structural assignments were based on several layers of analytical characteristics. Firstly, we monitored characteristic tandem mass spectrometric transitions (MRM), secondly we compared RRT with synthetic standard material and lastly we investigated the observed tandem mass spectra for the presence of several substance specific fragment ions.
Conversely, we detected relatively high concentrations of COX products, including prostaglandin D2 (PGD 2 ) and E2 (PGE 2 ), in eggs and egg-derived antigen preparations (Table 4). Moreover, apart from PGE 2 itself, several isomers were detected in high amounts in all egg-derived materials. The presence of geometric isomers can hamper substance identification, which is illustrated in Fig. 5. When monitoring the PGE 2 characteristic MRM transition m/z 351->271 several closely eluting peaks were obtained. Due to the fact that geometric isomers such as the D-and E-series prostaglandins listed in Fig. 5 give rise to almost identical tandem mass spectra we had to rely on RRTs for substance identification. We compared the RRTs with the ones obtained for synthetic standard material as listed in Fig. 5, showing the presence of several isomeric prostaglandins particularly present in eggs and egg-derived materials (SEA and EES).
Finally, worms and AWA were found to have a particular enrichment for leukotriene B4 (LTB 4 ). Although this was also detected in eggs, the signal found in SEA and cercariae could be attributed to isomeric 5S,12S-diHETE, which was revealed by small RRT differences between the two components, as described by us previously [30,34]. Concentrations and area ratios of all detected oxylipids can be found in Table 4 and supplementary data (Table S2).

Discussion
In the current study we performed an in-depth lipidomic analysis of the different life cycle stages of S. mansoni using three complementary MS-based platforms. In addition, we defined the lipid profile of aqueous extracts from different life stages of S. mansoni, which are CA, AWA and SEA, as well as their excretory and secretory products (ES), which are all commonly used antigen preparations in immunoparasitological studies. In total we identified more than 350 lipid species including several bioactive molecules, which we organized in a S. mansoni lipid database (supplementary Table S2). For FA, eicosanoids and docosanoids we have generated quantitative data, while a compositional analysis with relative abundances was carried out for higher order lipids. Our data reveal that there are life-stage specific lipid signatures not only for structural lipids but also for bioactive lipid mediators that may be important in host-pathogen interactions.
Global lipid class composition analysis revealed that PC(34:1), PC(36:1), and PC(36:2) are the predominant phospholipid species present in all life cycle stages. In line with this work, Schariter et al. have described the presence of PC in cercariae [35]. Moreover, our work is also consistent with earlier work on adult worms by Retra et al. [23] which also covered phosphoinositol (PI) and phosphatidylserine (PS) lipids and, as such, our work can be seen as complementary to theirs. Earlier work has found that a S. mansonispecific PS fraction conditioned DCs to induce both Th2 and IL-10producing T cells via TLR2 activation [19]. However, we were unable to detect significant amounts of PS in any of our preparations, this is likely due to the fact that our analytical approach focused on the ESIþ mode to detect PS while Retra et al. use the ESI-mode in their study. Furthermore, we observed that the global composition of the major lipid classes present in worms and eggs were similar, while cercariae were found to have a distinct profile, with less neutral lipids such as CE and TG. This specific difference is likely to be a direct result of the fact that because S. mansoni is unable to synthesize fatty acids and sterols de novo [36], it relies on scavenging of lipid precursors from its host to generate complex lipids such as phospholipids and TG [37], a process that free swimming cercariae cannot resort to. In addition our observation that cercariae contain much lower amounts of a large range of free FA we quantified, would support the idea that cercariae, compared to worms and eggs that reside in the host, have limited lipid accessibility.
Oxylipids such as docosanoids and eicosanoids have pleiotropic functions that include shaping the function of immune cells [38]. Hence, mapping and analysis of these known immunomodulatory lipid classes including their precursors was an important part of the present study. We found that cercariae as well as eggs and the antigen preparation derived from these life cycle stages, but less so adult worms, contain significant amounts of PUFAs, such as n3-PUFA (DHA, eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA)) and n6-PUFA (AA), as well as many of their bioactive derivatives that are known to be generated through COX and LOX activity. While cercariae have been shown before to produce various eicosanoids [39], this has not been described for the egg stage. An interesting observation is that despite the absence of COX enzymes in the S. mansoni genome [20,40] we detected various prostaglandins in both cercariae and eggs. This apparent discrepancy has been reported by others as well [41], and it has been postulated that this could be explained by an alternative mechanism for the generation of prostaglandin-like compounds, involving the non-enzymatic generation of molecules called isoprostanes, by auto-oxidation of AA by free-radicals. Indeed, we found evidence for production of isoprostanes in these parasites, which would allow for prostaglandin synthesis in the absence of COX activity. Of the prostaglandins we found PGD 2 and PGE 2 to be the most highly abundant species present in cercariae and eggs. Although PGD 2 may exert pro-inflammatory or anti-inflammatory effects in different biologic systems, in the context of S. mansoni it has primarily been linked to limiting the ability of the host to mount a protective immune response against cercariae [42], through its ability to suppress migration of Langerhans cells in the skin [18]. Whether PGD 2 produced by eggs has similar immunomodulatory effects remains to be determined. PGE 2, like PGD 2 , can have proinflammatory or anti-inflammatory effects depending on the context [43]. One particularly interesting property of PGE 2 is that is has been shown to condition DCs to prime Th2 responses [44]. This property, together with our observation that specifically eggs and SEA, which are well-known for their Th2 polarizing potential, contain high levels of PGE 2 , would make it conceivable that PGE 2 contributes to Th2 induction by S. mansoni eggs. Further studies would be needed to test this hypothesis.
In addition, we found various potentially immunomodulatory LOX derived products to be present in the different life cycle stages. For instance, in line with other studies [20], the AA-derived LOX product 15-HETE was detected. 15-HETE is known to be a ligand for peroxisome proliferator-activated receptors [45], which when activated in immune cells generally lead to suppression of proinflammatory effector responses by these cells. As such, one could speculate that release of 15-HETE by the parasite may locally contribute to suppression of immune responses. Finally, our analyses for the first time reveal that cercariae contain several n3-PUFA Table 3 A heat-map of total fatty acids composition in ng per mg of protein.
Each column is coloured separately. CA¼ Cercarial Antigen; CES¼ Cercarial ES; AWA ¼ Adult Worms Antigen; WES¼ Worms ES; SEA¼ Soluble Eggs Antigen; EES ¼ Egg ES. derived LOX products including RvE2, PDX and PD1. These docosanoids are specialized pro-resolving mediators that can suppress inflammatory immune responses and are important for promoting resolution of inflammation [46]. Once cercariae have penetrated the skin, they stay in the skin for 3e4 days before they enter the circulation. We hypothesize that production of these pro-resolving mediators during this period increases their chances of survival as these factors can suppress recruitment of immune cells and anticercarial immune responses. More in depth studies are warranted to further explore the role of resolvin production by S. mansoni life cycle stages in host-pathogen interactions and subversion of host immunity.
In conclusion, we have generated a comprehensive dataset defining the lipidome of different life cycle stages of S. mansoni in unprecedented detail. Our analysis has revealed several important differences between the lipid composition of the different life cycle stages of S. mansoni, that provide new insights into the biology of the parasite itself as well as into how the various life cycle stages interact with and may modulate the host immune system. Moreover, an important observation was that lipid composition of parasite life cycle stages and the corresponding extracts were highly similar, which validates the use of those widely used parasite preparations as models to study host-parasite interactions in immunological studies. Altogether, we believe this work will be a highly useful resource to many researchers in the field of parasitology and immunoparasitology and may serve as a starting point for the identification of immunomodulatory lipids that could be used or targeted, to shape anti-parasite immune responses with regards to vaccination or to develop lipid-based therapeutics to treat inflammatory disorders.

Competing financial interests
The authors declare not to have any competing financial Table 4 A heat-map of quantified oxylipids in S. mansoni preparations in ng per sample.