Fluorescent Labeling of Helminth Extracellular Vesicles Using an In Vivo Whole Organism Approach

In the last two decades, extracellular vesicles (EVs) from the three domains of life, Archaea, Bacteria and Eukaryotes, have gained increasing scientific attention. As such, the role of EVs in host-pathogen communication and immune modulation are being intensely investigated. Pivotal to EV research is the determination of how and where EVs are taken up by recipient cells and organs in vivo, which requires suitable tracking strategies including labelling. Labelling of EVs is often performed post-isolation which increases risks of non-specific labelling and the introduction of labelling artefacts. Here we exploited the inability of helminths to de novo synthesise fatty acids to enable labelling of EVs by whole organism uptake of fluorescent lipid analogues and the subsequent incorporation in EVs. We showed uptake of 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (DOPE-Rho) in Anisakis spp. and Trichuris suis larvae. EVs isolated from the supernatant of Anisakis spp. labelled with DOPE-Rho were characterised to assess the effects of labelling on size, structure and fluorescence of EVs. Fluorescent EVs were successfully taken up by the human macrophage cell line THP-1. This study, therefore, presents a novel staining method that can be utilized by the EV field in parasitology and potentially across multiple species.


Introduction
Extracellular vesicles, small membranous vesicles that contain a cargo of bioactive molecules are released from organisms spanning all three domains of life [1,2], including parasitic helminths and their hosts [3,4]. The discovery and prediction of miRNAs targeting host genes in EVs released by helminths [5,6] suggest that helminth-derived EVs contribute to host-parasite interactions and may modulate host immune responses, presenting potential translational applications for helminth EVs [3,4,[7][8][9]. Central to the study of the biological function of EVs is the accurate determination of cellular uptake or biodistribution. For this, the labelling techniques of EVs are required but this is not facile and post-isolation labelling techniques can introduce artefacts, through the co-isolation of micelles [10,11] or modulation of the EV surface [12] thereby obscuring accurate assessment of the mechanisms, rate of EV uptake in host cells and biodistribution in vivo.
Expression of known EV proteins tagged with fluorescent proteins, such as green fluorescent protein (GFP), provides a specific method of labelling EVs [13]. However, such techniques introducie

Larval Uptake of Fluorescent Lipid Analogues in Vitro
Hatched T. suis (L1) were divided into three groups, first group designated live uptake, second group designated passive uptake and last group a negative control.
The group designated passive uptake were placed on dry ice for 15 min to euthanize the larvae. 4 µM 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-lissamine rhodamine B sulfonyl (DOPE-Rho) (Avanti Polar Lipids, Alabaster, AL, USA) in DMEM were added to the groups except for the negative control to which was just added DMEM. They were then incubated for 2 h. Anisakis spp. (L3), were incubated with 0, 1, 4 or 8 µM DOPE-Rho for 5 min or 16 h. Larvae were harvested at the indicated time points, washed three times in PBS prior to fixation in 10 % formalin for T. suis and 4% paraformaldehyde (PFA) for Anisakis spp. T. suis larvae were washed 3 times after fixation and counter-stained with Hoechst-33342 nuclear stain. Larvae were imaged using a Leica DM 2000 LED fluorescent microscope (Leica microsystems, Copenhagen, DK) and images were processed in ImageJ 1.52a (NIH).

Anisakis spp. In Vivo EV Labelling
Anisakis spp. L3 larvae were maintained in PBS with Anti/Anti throughout culture at 37 • C and 5% CO 2 . Three Anisakis spp. larvae per well were incubated with 0, 1, 4 or 8 µM DOPE-Rho in PBS with Anti/Anti for 16 h. The larvae were then washed with PBS to remove excess dye, given new PBS with Anti/Anti and incubated for 48 h and the supernatant was harvested as conditioned PBS. Dye controls were made in parallel without larvae in the wells and dye solution was directly subjected to EV isolation.

Anisakis spp. Extracellular Vesicle Isolation
EVs from conditioned PBS were isolated by sequential differential centrifugation: 300× g for 10 min, 2000× g for 15 min, 10,000× g for 30 min followed by 110,000× g (40.800 RPM TI-50 rotor) (Beckman Coulter, Brea, CA, USA) for 90 min to isolate EVs. EVs were washed once by re-suspending the pellet in PBS and centrifuged once more at 110,000× g (40.800 RPM TI-50 rotor) for 90 min. Washed EVs were resuspended in PBS for subsequent characterization and uptake studies. Dye controls were subjugated to the same differential centrifugation. The total protein concentration of intact EVs was assessed using BCA assay as per the manufacturer's instructions (Thermo Fischer, Waltham, MA, USA). Fluorescence of isolated EVs and dye controls in PBS were assessed using a DS-11 spectrophotometer and fluorometer (DeNovix, Wilmington, DE, USA) with excitation at 525 nm and emission 565-615 nm.

Nanoparticle Tracking Analysis
Extracellular vesicle hydrodynamic radius and concentration were assessed using an NS300 (Malvern Pananalytical, Malvern, UK). Isolated EVs or dye only controls were diluted in filtered (0.2 µm) PBS and analysed immediately hereafter. The following conditions were maintained for all analyses of EVs: particles per frame of 20-100, camera level-15, detection threshold-5 and syringe pump speed of 10 µL/s. Three one-minute videos were captured per sample prior to analysis using Nanosight NTA 3.4.003 software. Hydrodynamic size is given as the mode of 3 measurements in nm (±SEM) and concentration as the mean of 3 measurements in particles/mL (±SEM).

Cryo-Transmission Electron Microscopy
Samples were defrosted and subjugated to short centrifugation to remove aggregates introduced by the freeze-thaw cycle with a Qualitron DW-41. Samples were prepared in an automated plunge freezer system (Leica Microsystems, Wetzlar, Germany). A 4 µL drop of the sample was dispersed on a glow discharged lacey formvar carbon-coated copper grids (Ted Pella Inc., Redding, CA, USA), blotted with a filter paper and then plunged into liquid ethane (approximately −183 • C). The vitrified specimens were thereafter stored in liquid nitrogen (−196 • C) prior to imaging. A Fischione Model 2550 cryo transfer tomography holder was used to transfer the specimen into the electron microscope, JEM 2200FS (JEOL, Tokyo, Japan), equipped with an in-column energy filter (Omega filter), which allows zero-loss imaging. The acceleration voltage was 200 kV and energy-filtered images were digitally recorded with a TVIPS F416 camera using SerialEM under low dose conditions with a 25 eV slit in place. Images were processed in ImageJ 1.52a (NIH).

Proteomic Analysis of Extracellular Vesicles
50 µL (>10 µg total protein) from each sample was lysed in 0.1% ProteaseMax in 0.1 M TEAB; sonicated for 2 min in water sonicator and denaturized for 10 min at 60 • C and stored at −80 • C until further processed. Protein concentration was estimated by protein absorbance at 280 nm on a Nanodrop 1000 UV-vis spectrophotometer (Thermo Scientific, Waltham, MA, USA) using the extinction coefficient of bovine serum albumin as a reference (Thermo Scientific Pierce BSA Protein; 2 mg/mL). Samples were reduced by incubating with 10 mM tris (2-carboxyethyl) phosphine (Thermo Scientific, Waltham, MA, USA) and 50 mM chloroacetamide (Sigma Aldrich, St. Louis, MO, USA) final concentration at 37 • C for 30 min. The samples were subsequently digested overnight with 1µg of sequencing grade modified trypsin in 0.1 M TEAB (Promega, Madison, WI, USA). Samples were acidified with 0.1% trifluoroacetic acid and reduced by vacuum centrifugation. The reduced samples were dissolved in 30 µL 2% acetonitrile; 0.1% formic acid; 0.1% trifluoroacetic acid and sonicated for 5 min in a water bath. An aliquot corresponding to 200 ng of tryptic peptides were used for analysis in triplicate (n = 3) for quantitative analysis. The samples were separated on a Dionex RSLC UPLC system (Thermo Scientific, Waltham, MA, USA) with uPAC 50 cm analytical column with precolumn (Pharmafluidics, Ghent, Belgium). The samples were loaded at 5 min at 10 uL per min and the mobile phase was ramped over 30 min at a constant flowrate of 700 nL/min from 98% solvent A (0.1% formic acid) and 2% solvent B (0.1% formic acid in acetonitrile) to 45% solvent B in 40 min. Eluted peptides were directly introduced to the coupled ThermoSci QE HF-X mass spectrometer (Thermo scientific, Bremen, Germany) by a picotip emitter for electrospray ionization (New objective, Woburn, MA, USA). The mass spectrometer was operated in positive mode using a data-dependent acquisition method with the following settings: mass range m/z 375-1200; MS1-scan resolution 120,000; MS2-scan resolution 30,000; isolation window m/z 1.6 and NCE 28. Peptide hits were searched against Anisakis simplex UniProt protein entries (UP000036680; 20,879 entries; released 05/2019) using standard settings in Maxquant v.1.6.12.0 [28]. Proteins of interest were analysed using Blast2Go [29] for BLASTP (NCBI).

Extracellular Vesicle Uptake in Human Macrophage-Like THP-1 Cells
THP-1 cells (ATCC, Manassas, VA, USA) were maintained in RPMI-1640 (Biowest, Nuaillé, France) with 10% FBS (Sigma Aldrich, St. Louis, MO, USA) and 0.05 mM β-mercaptoethanol (Sigma Aldrich, St. Louis, MO, USA) at 37 • C and 5% CO 2 . Cells were activated with 100 ng/mL phorbol 12-myristate-13-acetate (PMA) (Sigma Aldrich, St. Louis, MO, USA) prior to seeding of 200,000 cells per well in chamber slides for 24 h. Cells were rested in complete media (without PMA) for a further 24 h prior to uptake studies. DOPE-Rho labelled EVs were added at low and high dose, 2 µg and 8 µg total EV protein per well and cells incubated overnight in RPMI with 5% EV-depleted FBS (Thermo Scientific, Waltham, MA, USA). Unlabeled EVs from Anisakis spp. were added at a concentration of 12 µg per well as a negative control. Cells were then washed once in PBS, fixed with 10% formalin prior to washing, counterstaining using Hoechst-33342 nuclear stain and mounting in fluoroshield gold (Thermo Scientific, Waltham, MA, USA). Uptake was assessed using a Leica DM 2000 LED fluorescent microscope (Leica microsystems, Copenhagen, DK). Images were processed in ImageJ 1.52a (NIH).

Fluorescent Lipid Analogue Uptake by T. suis L1 Larvae
Hatched T. suis L1 larvae were incubated with 4 µM DOPE-Rho in culture media for 2 h prior to washing, fixation, nuclear counterstaining and analysis by fluorescent microscopy. Fluorescent lipid was taken up by live larvae although stain intensity varied amongst larvae, all showed strong fluorescence in the mouth region ( Figure 1A,B). No unspecific fluorescence was observed in non-labelled control larvae ( Figure 1C,D). To determine if uptake was a passive or active process, larvae were killed by freeze-thaw and incubated for 2 h with 4 µM DOPE-Rho and analysed. The fluorescent signal was detected in the mouth region of dead larvae, albeit at a much lower intensity, but in contrast to live larvae fluorescence was not distributed elsewhere ( Figure 1E,F). labelled control larvae ( Figure 1C,D). To determine if uptake was a passive or active process, larvae were killed by freeze-thaw and incubated for 2 h with 4 µM DOPE-Rho and analysed. The fluorescent signal was detected in the mouth region of dead larvae, albeit at a much lower intensity, but in contrast to live larvae fluorescence was not distributed elsewhere ( Figure 1E,F).

Cryo-TEM of Anisakis spp.
EVs were heterogeneous in size, and most had a corona of surface molecules (Figure 4). The samples were quite dilute with the highest EV number and absence of aggregates, in the labelled samples derived from larvae incubated with 1 µM DOPE-Rho ( Figure 4). The EVs derived from larvae incubated with 4 and 8 µM DOPE-Rho contained a high degree of aggregates and few EVs ( Figure  S2).

EVs from Anisakis spp. Assessed by Quantitative Proteomics to Identify EV Candidate EV Markers
Stringent filtering by inclusion criteria requiring at least two unique peptides + razor in triplicates of both day 3 and day 5 harvested EVs. In the absence of prior knowledge of Anisakis spp.

Cryo-TEM of Anisakis spp.
EVs were heterogeneous in size, and most had a corona of surface molecules (Figure 4). The samples were quite dilute with the highest EV number and absence of aggregates, in the labelled samples derived from larvae incubated with 1 µM DOPE-Rho ( Figure 4). The EVs derived from larvae incubated with 4 and 8 µM DOPE-Rho contained a high degree of aggregates and few EVs ( Figure S2).

Cryo-TEM of Anisakis spp.
EVs were heterogeneous in size, and most had a corona of surface molecules ( Figure 4). The samples were quite dilute with the highest EV number and absence of aggregates, in the labelled samples derived from larvae incubated with 1 µM DOPE-Rho ( Figure 4). The EVs derived from larvae incubated with 4 and 8 µM DOPE-Rho contained a high degree of aggregates and few EVs ( Figure  S2).

EVs from Anisakis spp. Assessed by Quantitative Proteomics to Identify EV Candidate EV Markers
Stringent filtering by inclusion criteria requiring at least two unique peptides + razor in triplicates of both day 3 and day 5 harvested EVs. In the absence of prior knowledge of Anisakis spp.

EVs from Anisakis spp. Assessed by Quantitative Proteomics to Identify EV Candidate EV Markers
Stringent filtering by inclusion criteria requiring at least two unique peptides + razor in triplicates of both day 3 and day 5 harvested EVs. In the absence of prior knowledge of Anisakis spp. EV markers, identified proteins, excluding proteases, were compared to selected helminth EV proteomes in the literature (Table 1). Proteins 14-3-3 and HSP70 that are commonly observed in the proteome of mammalian EVs and also the studies referenced in Table 1 were detected in the EV enriched fraction of Anisakis spp. GTP binding proteins such as ras-like-protein 3 and ras-like GTP-binding protein RhoA that are implicated in vesicle biogenesis and transport were also detected in Anisakis spp. EV enriched samples. Antioxidant protein superoxide dismutase was also detected in EVs from Anisakis spp (Table 1). Pepsin inhibitor [32] 1 References for homologous proteins identified in EV proteomes of H. polygyrus [5], F. hepatica [30,31], T. muris [32], A. suum [6], S. mansoni [33] and T. circumcincta [34]. 2 Chloride channel exc identified in [31], C-exl 1 identified in [6]. Alternative chloride channels identified e.g., CLCN7 [29,30]. 3 Identified in references as lysosomal acid phosphatase or acid phosphatase.

Uptake of DOPE-Rho Labelled EVs in THP-1 Cells
To further evaluate this EV labelling method, uptake of labelled Anisakis spp. EVs was assessed in PMA activated THP-1 cells. As Cryo-EM determined that the EVs from Anisakis spp. labelled with 1 µM DOPE-Rho were the purest, this was used for uptake study. Fluorescent EVs were added at 2 µg or 8 µg (total protein of non-lysed EVs) per well and uptake was assessed using fluorescence microscopy after overnight incubation (

Discussion
The addition of fluorescent lipid analogue, DOPE-Rho, to culture media was actively taken up in vivo and subsequently incorporated by larvae of nematodes from two differing classes of nematodes. Passive uptake was observed in dead T. suis larvae albeit reduced when compared to live larvae. Passive uptake is in line with Furlong et al. (1992) who saw rapid uptake through the outer membrane of S. mansoni and accumulation in the oesophageal gland and gut with headgroup labelled phosphoethanolamine [22]. Although the outer surface of nematodes (cuticle) differs from trematodes (tegument) previous studies determined that uptake was mediated via regions of the surface membrane, where lipids can diffuse easily and may explain uptake in dead larvae [35]. Helminths produce and secrete several lipid-binding proteins (LBP) [17,19] that are implicated in the appropriation of host fatty acids and lipids, which could enable lipid-binding to dead larvae. As per the observations of Furlong et al. (1995) in S. mansoni cercariae and schistosomula, we also observed accumulation of fluorescent phospholipid at the surface, in the gut [23], and evidence for ingestion of the lipids with staining likely corresponding to the mouth and excretory pore of Anisakis spp. While EV subpopulations, such as exosomes and microvesicles, are intensively researched from mammalian sources [1], little is known about the biogenesis and subpopulations of EV derived from parasites. Nevertheless, promiscuous labelling as we have performed would be presumed to label every EV subtype secreted from the parasite.

Discussion
The addition of fluorescent lipid analogue, DOPE-Rho, to culture media was actively taken up in vivo and subsequently incorporated by larvae of nematodes from two differing classes of nematodes. Passive uptake was observed in dead T. suis larvae albeit reduced when compared to live larvae. Passive uptake is in line with Furlong et al. (1992) who saw rapid uptake through the outer membrane of S. mansoni and accumulation in the oesophageal gland and gut with headgroup labelled phosphoethanolamine [22]. Although the outer surface of nematodes (cuticle) differs from trematodes (tegument) previous studies determined that uptake was mediated via regions of the surface membrane, where lipids can diffuse easily and may explain uptake in dead larvae [35]. Helminths produce and secrete several lipid-binding proteins (LBP) [17,19] that are implicated in the appropriation of host fatty acids and lipids, which could enable lipid-binding to dead larvae. As per the observations of Furlong et al. (1995) in S. mansoni cercariae and schistosomula, we also observed accumulation of fluorescent phospholipid at the surface, in the gut [23], and evidence for ingestion of the lipids with staining likely corresponding to the mouth and excretory pore of Anisakis spp. While EV subpopulations, such as exosomes and microvesicles, are intensively researched from mammalian sources [1], little is known about the biogenesis and subpopulations of EV derived from parasites. Nevertheless, promiscuous labelling as we have performed would be presumed to label every EV subtype secreted from the parasite.
The modal size of vesicles, 140.5 nm, harvested from Anisakis spp. conditioned media is consistent with the size profile of mammalian small EVs [36] and that previously reported for EVs from A. suum, T. suis, Fasciola hepatica, Brugia malayi and Heligmosomoides polygyrus [6,[37][38][39]. Of note, the incubation of Anisakis spp. with the fluorescent lipid analogue, DOPE-Rho, did not alter size distribution of EVs from Anisakis spp. at 1-4 µM DOPE-Rho concentration. However, the size distribution of vesicles harvested from 8 µM incubated Anisakis spp. indicated that the sample was not as pure and suggested altered release or toxicity at this concentration. The concentration of vesicles isolated from non-labelled and 4 µM DPPE-Rho labelled were unchanged, however, a small decrease in vesicles was observed in 1 µM DPPE-Rho EVs. Whilst equal numbers of worms were included per group variations in EV concentration resulting from differences in larvae biomass, and therefore EV production cannot be discounted. However, the greatly reduced number of vesicles in media of 8 µM DOPE-Rho incubated Anisakis spp. combined with the reduced lipid uptake in these larvae suggests potential toxicity at this highest concentration. Viability of larvae were only assessed visually, with no reductions in movement observed, but effects on metabolism at this concentration are unknown. Furthermore, the larvae were only kept alive until the harvest of the EVs was performed, prolonging incubation might have revealed further toxicity of the 4 and 8 µMDOPE-Rho concentration.
Using Cryo-TEM we visualized EVs from Anisakis spp. with high resolution in their native state with labelling at 1 µM showing the highest purity and EV concentration. Whether the impurities at higher concentrations DOPE-Rho are due to lipid mediating cytotoxicity, and thereby release of dead cells and cuticle, is unknown. The combination of lipid uptake, NTA and cryo-TEM was essential in determining the optimal concentration of DOPE-Rho used in uptake studies and highlights the risks of relying upon single characterisation techniques in EV biology.
To our knowledge, this study presents the first proteomic study of EVs from Anisakis spp. and indeed proteins heavily represented in helminth EV proteomes were identified in Anisakis spp. EV enriched samples, including 14-3-3 and HSP70 that are almost ubiquitous in proteomic studies of both mammalian [40,41] and helminth EVs [9,42]. Proteases were highly prominent in this analysis, an observation that is consistent with that previously observed for helminth derived EVs [5,6,[30][31][32]43]. Given the absence of a priori knowledge of Anisakis spp. EV markers and a lack of commercial antibodies targeting previously proposed EV markers of closely related helminths, we propose that the proteomics approach employed and identification of common helminth EV proteins is sufficient to confirm the presence of intact EVs in isolated fractions, when combined with cryo-electron microscopy.
This proof-of-concept study has shown that labelled fluorescent lipid analogues are internalised by nematodes from two classes. Furthermore, using Anisakis spp. as a model system, this study has demonstrated that the subsequently released EVs contain fluorescent lipids and their uptake in cells can be assessed by fluorescence microscopy. With the increasing attention on the importance of EVs in host-parasite interaction, this method could not only be used to label all EV subtypes released by nematodes, but may also be applied over a wide range of helminth species. This method could likewise be adapted for the labelling of protozoan EVs, although they both scavenge and synthesize fatty acids de novo, fluorescent fatty acids have been shown to be taken up [44][45][46] and have been employed in the tracking of protozoans intracellularly [46] and trematodes in vivo [21]. The choice of fluorophore-conjugated fatty acid or lipid, uptake duration and concentration should be optimised for each helminth species studied but we propose this may provide a unique method for labelling of EVs in situ for subsequent harvest and use in functional assays that are highly suited to helminth biology.