The neutral sphingomyelinase pathway regulates packaging of the prion protein into exosomes

Background: intercellular Results: Chemical and RNAi inhibition of the neutral sphingomyelinase (nSMase) pathway impairs exosome formation and prion packaging. Conclusion: The nSMase pathway regulates exosome formation and packaging of infectious prions. Significance: This reveals a novel role for the nSMase pathway in exosomal prion packaging and identifies a direct pathway which mediates prion transmission. ABSTRACT Prion diseases are a group of transmissible, fatal neurodegenerative disorders associated with the misfolding of the host-encoded prion protein, PrP C , into a disease associated form, PrP Sc . The transmissible prion agent is principally formed of PrP Sc itself and is associated with extracellular vesicles known as exosomes.

of nSMase1 and nSMase2 in mouse neurons using lentivirus-mediated RNAi also decreases exosome release, demonstrating the nSMase pathway regulates the biogenesis and release of exosomes. We also demonstrate that PrP C packaging is dependent on nSMase2, whereas the packaging of disease-associated PrP Sc into exosomes occurs independently of nSMase2. These findings provide further insight into prion transmission and identify a pathway which directly assists exosome-mediated transmission of prions.
Prion diseases, such as scrapie in sheep, bovine spongiform encephalopathy in cattle, and Creutzfeldt-Jakob disease in humans, are a family of invariably fatal neurodegenerative diseases characterised by a progressive loss of neurons and spongiform vacuolation of the brain. According to the protein only hypothesis, the underlying cause of prion disease is misfolding of the normal soluble cellular prion protein, PrP C , into the disease-associated isoform, PrP Sc , which has an increased propensity to aggregate (1). Prion diseases are unique amongst neurodegenerative diseases as they can be acquired through exposure to the infectious agent, giving the disease a transmissible etiology.
The exact mechanism by which prion disease is transmitted intercellularly is yet to be identified, although three mechanisms have been proposed, including: direct cell-cell contact; tunnelling nanotubes, which are fine membrane protrusions connecting the cytoplasm of one cell with a neighbouring cell; and extracellular vesicles such as microvesicles and exosomes (2)(3)(4)(5). Exosomes are small membrane-bound vesicles released by cells, approximately 30-100 nm in diameter, as determined using TEM, and contain proteins, mRNAs and miRNAs (6)(7)(8)(9)(10). Whereas microvesicles are generated via budding at the plasma membrane, exosomes are distinct and are generated within the endosomal system (11). As early endosomes mature into late endosomes, the limiting membrane of late endosomes invaginate to form intraluminal vesicles (ILVs), generating the multivesicular body (MVB) (6). The MVB either fuses with lysosomes, resulting in degradation of the ILVs, a well-known process for down-regulating transmembrane receptors, or it fuses with the plasma membrane, resulting in release of the ILVs as exosomes. Consistent with the endosomal origin, exosomes have a unique protein and lipid composition, and contain proteins such as Tsg101, a component of the endosomal sorting complex required for transport (ESCRT) machinery, and are enriched in raft-associated lipids such as cholesterol, sphingolipids, ceramide and glycerophospholipids (6,7,12).
A link between exosomes and prion disease was first established in rabbit kidney epithelial cells expressing ovine PrP (ROV). The ROV cells release exosomes which contain PrP Sc and are capable of transmitting infection to mice (4). In support of this, exosomes released from a prion-infected mouse neuronal cell line are also capable of transmitting infection to naïve cells and to mice (10,13). Exosomes are also able to transmit infectivity between heterologous cell types, suggesting that exosomes could assist prion transmission from the periphery to the central nervous system (13). In addition to prion disease, exosomes have also been implicated in other neurodegenerative diseases such as Alzheimer's and Parkinson's disease, and Amyotrophic lateral sclerosis, where the disease-associated proteins have also been found in exosomes (14)(15)(16)(17)(18). Furthermore, HIV-1 has also been proposed to hijack the exosomal pathway to produce new viral particles (19).
Given the proposed role of exosomes in facilitating transmission of disease-related proteins involved in neurodegeneration, elucidating pathways which regulate exosome protein packaging could be valuable for understanding disease progression and pathogenesis (20). The most common mechanism of protein targeting into ILVs is mono-ubiquitination and recruitment of the ESCRT machinery (21,22). However, ubiquitin-independent mechanisms have also been described, including lipid-raft-association, higher-ordered oligomerisation, association with tetraspanin-enriched domains, and the generation of ceramide via the neutral sphingomyelinase (nSMase) pathway (23)(24)(25)(26). These alternative pathways are dependent on a reduction in the diffusion rate of cargo proteins, either through association with other proteins or with microdomains within the limiting membrane of endosomes, leading to retention of the proteins and subsequent incorporation into budding ILVs. The nSMase pathway was first demonstrated to assist ILV formation and packaging of the proteolipid protein in mouse oligodendroglial cells, and have since then being found to also regulate packaging of miRNA into ILVs (27). Furthermore, nSMase-dependent exosome biogenesis has also been linked to Alzheimer's disease as the exosomes were proposed to promote clearance of Aβ by microglia (28).
Four mammalian nSMase isoforms have been identified: nSMase1, nSMase2, nSMase3 and the most recently identified mitochondrial-associated nSMase (MA-nSMase) (29)(30)(31)(32). Both nSMase1 and nSMase2 have been reported to have nSMase activity in vitro, although only nSMase2 has nSMase activity in vivo. With exosomes implicated as vehicles for intercellular transmission of prions, we aimed to investigate a role for the nSMase pathway in exosome biogenesis and packaging of PrP into exosomes by inhibiting the nSMase pathway using a chemical inhibitor, GW4869, and using lentivirus-mediated RNAi to silence nSMase1 and nSMase2.

Maintenance and infection of cultured cells lines with
prions, and the cell blot assay -GT1-7 cells were cultured in Opti-MEM (Life Technologies, Melbourne, VIC, Australia), and infected with a mouse-adapted strain of human prions (M1000), as previously described (10). At 6 passages post infection, stable infection of cells was assessed using a cell blot assay. Cells were seeded onto sterile plastic cover slips (Thermanox; Nunc-Thermo, Scoresby, VIC, Australia) in 6-well plates (Nunc-Thermo) with 250,000 cells/well and grown to confluence before assay was carried out as previously described (10). Generation of lentiviral supernatants and stable transduction of GT1-7 cells -A panel of 3 (A-C) and 4 (A-D) commercial lentiviruses for murine nSMase1 and nSMase2, respectively, GIPZ Non-silencing lentiviral control and TRC lentiviral pLKO.1 empty vector control were purchased as glycerol stocks from Open Biosystems (Millennium Science, Surrey Hills, VIC, Australia). The construct ID and mature antisense sequences are: nSMase1A (V2LMM_27471; TAAGCTCTTAAGCTCTGTC); nSMase1B (V2LMM_29867; TCAGACTGTAGACATGCTG); nSMase1C (V2LMM_38704; ACAAAGTAGCCATGAGGGC); nSMase2A (TRCN0000099415; AAATCGTGACTGAAAGAATGG); nSMase2B (TRCN0000099416; ATACAGCAGCTTGCTGTTAGC); nSMase2C (TRCN0000099418; ATGTAGATCTTGATCTGAGGC); nSMase2D (TRCN0000099419; ATAGAGACCGTTTGTGTCCAG). Plasmid DNA was isolated using the PureYield™ Plasmid Midiprep System (Promega, Fitchburg, WI, USA) according to manufacturer's instructions. The 2 nd generation lentiviral packaging plasmids, psPAX2 (Addgene plasmid 12260; Addgene, Cambridge, MA, USA) and pMD2.G (Addgene plasmid 12259; Addgene) were used to generate lentiviral particles in HEK293FT cells. 2.5 µg pMD2.G, 6.5 µg psPAX2 and 2 µg lentiviral construct were combined and used to transfect a 75 cm 2 flask (Nunc-Thermo) of HEK293FT cells using Lipofectamine 2000 (Life Technologies), according to manufacturer's instructions. 24 h after transfection, the media is replaced with fresh DMEM (Life Technologies) supplemented with 10% (v/v) heat inactivated FCS, 1 mM sodium pyruvate (Life Technologies), 1x MEM Non-essential Amino Acids, 100 units of penicillin and 100 µg/ml streptomycin as Penicillin-Streptomycin 100x and 1x GlutaMAX. Lentiviral supernatants were harvested at 48 h posttransfection and centrifuged for 5 min, 1,400 g, then aliquoted and stored at -80 °C. GT1-7 cells were seeded into 6-well plates (Nunc-Thermo) at 400,000 cells/well and allowed to recover for 24 h before being overlaid with 1 ml of lentiviral supernatant diluted in 1 ml complete Opti-MEM culture media. 24 h after transduction, the lentiviral supernatant is replaced with fresh complete culture media and 0.2 µg/ml puromycin was introduced at 48 h post-transduction. The degree of KD achieved was assessed at 3 passages posttransduction. RNA isolation and quantitative Real Time RT-PCR -Total RNA were extracted from cells using the miRNeasy Mini Kit (Qiagen, Hilden, Germany) according to manufacturer's protocol. 2 µg RNA was converted to cDNA using High Capacity cDNA Kit (Life Technologies) according to manufacturer's instructions. Real Time RT-PCR samples were prepared using TaqMan® Gene Expression Master Mix (Life Technologies) and murine specific TaqMan® Gene Expression Assays (Life Technologies) for Prnp (Mm00448389_m1), nSMase1 (Mm01188195_g1) and nSMase2 (Mm00491359_m1), with endogenous controls for murine Tbp (Mm00446971_m1) and Hprt1 (Mm00446968_m1). Samples were prepared on the QIAgility liquidhandling robot (QIAGEN) and run on a StepOnePlus qRT-PCR instrument (Life Technologies) using manufacturer's recommended cycling conditions. Data was analysed by the comparative ΔΔCt method as previously described (34). Exosome isolation -For exosome collection, 175 cm 2 flasks (Nunc-Thermo) were seeded with 7 x 10 6 cells and allowed to recover for 24 h. For stably transduced cells, the culture media is replaced with exosome-free media (complete media containing exosome-depleted FCS) and cultured for 2 days, without puromycin. For GW4869 treatment, the culture media is replaced with exosome-free media containing 4 µM GW4869 and cultured for 2 days. Exosomes were isolated from the culture supernatant using differential ultracentrifugation, as previously described (10). Exosome pellets were resuspended using 70 µl filtered PBS. Transmission Electron Microscopy -10 µl aliquots of exosomes were fixed with 10 µl 2% (v/v) EM grade glutaraldehyde and incubated for 30 min, room temperature. A 6 µl aliquot of fixed exosome solution was then absorbed onto glow-discharged 300 mesh carbon coated formvar-copper TEM grids (ProSciTech, SA, Australia) and analysed using TEM, as previously described (10). Acetylcholinesterase (AChE) assay -AChE activity assay was carried out as previously described (35)(36)(37). Briefly, freshly collected exosomes in 70 µl PBS were diluted 1:10 with PBS, and 50 µl of the diluted samples were aliquoted into a clear 96-well colorimetric plate in triplicate. The reaction was initiated by adding 50 µl AChE assay solution (1.25 mM acetylthiocholine, 0.1 mM 5,5-dithio-bis(2nitrobenzoic acid), PBS pH7.4). The plate was wrapped in foil and incubated for 10 min, 37 °C. Absorbance at 450 nm was measured every 30 s for 10 min, room temperature, using a Varioskan® multimode plate reader. The rate of change (per min) was determined over the 10 min using the SkanIt Software 2.4.3 RE (Thermo Scientific, Australia), adjusted to be rate of change per ml, and then scaled by 1000 for data handling. qNano -Scanning ion occlusion sensing analysis using the qNano instrument (Izon, Christchurch, New Zealand) was carried out as previously described, with minor alterations (38). Cell culture supernatant was spiked with 40x qNano supplement (3.2 M KCl, 100 mM EDTA pH 8.0, 0.1% Tween20) to reach a final concentration of 1x. The supernatant was cleared using a 0.22 µm filter (Merck Millipore, Darmstadt, Germany). The samples were vortexed for 30 sec immediately prior to measurement. Protease digestion, SDS-PAGE, and immunoblotting -Cells were lysed on ice for 20 min using ice-cold lysis buffer (150 mM NaCl, 50 mM Tris pH 7.4, 1% (v/v) Triton X-100, 1% (w/v) sodium deoxycholate containing cOmplete protease inhibitor cocktail (Roche, Castle Hill, NSW, Australia)). Exosomes were lysed on ice for 20 min using 30 µl lysis buffer. Total protein concentrations were determined using the bicinchoninic acid (BCA) protein assay (Pierce, Thermo Scientific), according to manufacturer's protocol. 100 µg protein were treated with proteinase K (PK; 25 µg/ml), as previously described (10). 40 µg protein were incubated on ice for non-PK-treated samples. Ice-cold methanol (4x volume of lysate) was added to the samples, vortexed briefly, and then stored overnight at -20 °C. Protein precipitates were pelleted by centrifugation for 15 min, 16,100 g, room temperature, and the pellets were air-dried for 30 min. Pellets were resuspended with 20 µl 1x SDS-PAGE sample buffer, vortexed, heated for 8 min, 100 °C, then centrifuged for 2 min, 10,000 g, prior to SDS-PAGE and immunoblotting analysis, as previously described (10). Toxicity assay -GT1-7 cells were seeded into µClear black 96-well plates with clear bottom (Greiner Bio-One, Frickenhausen, Germany) at 12,000 cells/well and allowed to attach overnight. Cells were treated with DMSO-only, 0.5 µM, 1 µM, 2µM and 4 µM GW4869 for 48 h, and toxicity was determined using the LIVE/DEAD® Viability/Cytotoxicity Kit (Life Technologies), according to manufacturer's instructions. 15 min prior to the assay, 1% (v/v) Triton X-100 (Ajax Finechem; Thermo Scientific), diluted in complete Opti-MEM, was added to positive control cells to induce cell death. Data was analysed according to manufacturer's protocol and represented as % live cells compared to DMSO-only control. Lipidomic analysis -GT1-7 cells, were seeded with 7 x 10 6 cells into 175 cm 2 flasks (Nunc-Thermo) and allowed to recover for 24 h. Cells were treated with DMSO-only or 4 µM GW4869 and cultured for 2 days in exosome-free media. Cells were washed warm media and PBS. Cells were then detached in PBS, pelleted by centrifuagtion, snap frozen in liquid nitrogen and stored at -80 °C before lipidomic extractions. Lipids were extracted using monophasic extraction with chloroform:methanol:water (1:2:0.6,vol/vol/vol). After removal of insoluble material by centrifugation (15,000g, 15 min), extracts were dried under nitrogen and dried cellular extracts were resuspended in 100 uL of butanol/methanol (1:1, vol/vol) containing 5 μM ammonium formate. Cellular lipids were separated using an Agilent LC 1200 and lipids were analysed by electrospray ionisation-mass spectrometry (ESI-MS) using an Agilent Triple Quad 6460 (Mulgrave, Australia), as previously described (39). The molecular species of the lipid classes were identified using precursor ion scanning from 100 to 1000 m/z in positive ion mode, ceramide (m/z 264.6). The peak area of each lipid species was normalised to abundance per cell. Image processing and statistical analysis -Densitometric analysis was carried out using ImageJ (40). Statistical analyses were performed using GraphPad Prism 6 (GraphPad, San Diego, CA, USA). Data is analysed as follows: qRT-PCR and toxicity assay data are shown as mean of triplicate and quadruplicate experiments, respectively, ± SEM, and statistical significance tests were performed with two sample t tests; AChE, qNano, and densitometric analyses data are shown as mean change in comparison to control of triplicate experiments ± SEM, and statistical significance tests were performed with one sample t tests to compare each value against a standardised control of 100%. Advice was sought from The University of Melbourne's Statistical Consulting Centre, and it was advised that where there is a difference observed for samples, but the p value is not < 0.05, it is still acceptable to note that there is a change. Statistical significance is defined as * p < 0.05, ** p < 0.01 and *** p < 0.001.

Isolating exosomes from prion-infected GT1-7 cells -
The GT1-7 mouse hypothalamic neuronal cell line was used as a model of prion disease as it expresses endogenous levels of PrP C and is one of few cell lines capable of sustaining persistent prion infection (41,42). Non-and M1000 prion-infected GT1-7 cells were generated and the infection was confirmed through detection of PrP Sc by the cell blot assay (Fig. 1A). Exosomes isolated from non-and prion-infected GT1-7 cells were characterised by probing for the presence of exosomal protein markers, Tsg101 and flotillin-1, and the absence of cytosolic contaminants markers such as GM130 (Golgi), Bcl-2 (mitochondria) and nucleoporin (nucleus) (Fig. 1B). In addition, both PrP C and PrP Sc are detected within exosomes (Fig. 1C). Morphological analysis using TEM reveals a homogenous population of round membrane-bound vesicles, approximately 85 nm in diameter (Fig. 1D). However, reports using techniques such as cryo-EM and nanoparticle tracking analysis to visualise the hydrated structure of exosomes indicate that exosomes are slightly larger, with an approximate diameter of 80-160 nm (10,43). In agreement with this, qNano analysis of the conditioned media from non-and prioninfected cells reveal that the exosomes are approximately 116 nm in diameter (Fig 1E).
Inhibition of the nSMase pathway with GW4869 impedes exosome biogenesis and PrP packaging -GW4869, a nSMase inhibitor, was used to investigate the role of the nSMase pathway in exosome biogenesis and PrP packaging (26). GW4869 was tested for toxicity to GT1-7 cells and analysis of the percentage live cells relative to vehicle-only control reveal no toxic effects for all concentrations tested ( Fig.  2A). GW4869 was thus used at a concentration of 4 µM, which is comparable to the 5 µM used in the study which reported a link between the nSMase pathway and packaging of the proteolipid protein into exosomes (26). GC-MS analysis of ceramide levels in control and GW4869-treated cells reveal a reduction for several ceramide species, confirming the inhibitory activity of the compound on the nSMase pathway (Fig.  2B).
The effect of GW4869 on exosome biogenesis was assessed initially using the qNano instrument to quantitate exosomes in culture supernatant. qNano uses scanning ion occlusion sensing technology to determine particle size and concentration (44). While the size distribution of non-and prion-infected exosomes do not change with treatment (Fig. 2C, D), the amount of exosomes present is decreased upon GW4869 treatment (Fig. 2E). To further validate this finding, exosomes isolated from culture supernatant were quantitated using the acetyl cholinesterase (AChE) activity assay, which provides a quantitative measure for exosomes via detection of acetyl choline present on the surface of exosomes (36,37). In agreement, exosome release is also decreased in both non-and prion-infected cells upon GW4869 treatment (Fig. 2F). This indicates that the nSMase pathway also regulates exosome biogenesis in GT1-7 cells, similar to oligodendroglial cells (26).
The effect of GW4869 on the levels of cellular and exosomal PrP C and PrP Sc were also examined. Analyses of vehicle-and GW4869-treated cells show that treatment with the GW4869 decreases both cellular PrP Tot (total PrP prior to PK treatment) and PrP Sc levels in prion-infected cells (Fig. 3A and 3B). In exosomes, a significant decrease is observed for PrP C from non-infected cells, while PrP Tot and PrP Sc is decreased in exosomes from infected cells (Fig. 3A  and 3B). Treatment of both non-and prion-infected cells with GW4869 also reduced the levels of exosome markers Flotillin-1 and Tsg101 (Fig 3C and 3D), in agreement with a reduction in exosome release. Together, these findings indicate that packaging of PrP into exosomes is regulated by the nSMase pathway.
Depletion of nSMase1 and nSMase2 impairs exosome biogenesis -To further investigate the nSMase pathway, two isoforms, nSMase1 and nSMase2, were targeted for RNAi. As GT1-7 are neuronal cells, lentivirus-mediated RNAi was used to generate stably knocked down cells in order to maximise RNAi efficiency. A panel of 3 commercial lentiviruses were used for nSMase1 (nSMase1A-C) and 4 were used for nSMase2 (nSMase2A-D). RNAi was quantitated using qRT-PCR for nSMase1 (Fig. 4A) and nSMase2 (Fig.  4B). The levels of nSMase1 KD were also quantitated via immunoblotting, the same could not to be carried out for nSMase2 due to the lack of a suitable antibody. (Fig. 4C). Lentiviruses producing successful KD were used for ensuing analyses.
Using AChE assay, exosomes from nSMase1 and nSMase2 KD cells were quantified following isolation from the culture supernatant. In agreement with GW4869 treatment, exosome release from nonand prion-infected nSMase1B and nSMase1C KD cells are decreased, indicating that nSMase1 regulates exosome biogenesis (Fig. 5A). Similarly, exosome release is also decreased from non-and prion-infected nSMase2A and nSMase2B KD cells, albeit the decrease is less in the latter (Fig. 5B). Together, these results confirm a role for the nSMase pathway in modulating exosome biogenesis in GT1-7 cells.
The nSMase pathway regulates PrP packaging into exosomes -The effect of nSMase1 and nSMase2 KD on the levels of cellular and exosomal PrP were also analysed. For both nSMase1B and nSMase1C KD, no significant changes are detected for the levels of cellular PrP C , PrP Tot and PrP Sc (Fig 6A and 6B). Similarly, no significant changes are detected for exosomal PrP C , PrP Tot and PrP Sc . However, an apparent decrease is observed for PrP Tot and PrP Sc in prioninfected nSMase1C exosomes, suggesting that the KD slightly hinders packaging of PrP into exosomes. In agreement, the levels of exosome markers Flotillin-1 and Tsg101 also show an apparent decrease in prioninfected nSMase1C exosomes ( Fig. 6C and 6D). Together, these results indicate that nSMase1 could play a minor role in exosome protein packaging in prion-infected cells, which correlates with the reported activity of nSMase1 only under stressed cellular conditions (45,46).
Analyses of nSMase2A and nSMase2B KD also indicated no changes in cellular levels of PrP C , PrP Tot and PrP Sc (Fig. 7A and 7B). However, a decrease is observed for PrP C in nSMase2B exosomes. While a slight decrease in PrP C is also observed for nSMase2A exosomes, the decrease did not achieve statistical significance. Most significantly, no changes are observed for PrP Tot or PrP Sc in prion-infected exosomes, suggesting that the nSMase pathway is altered in prion-infected cells such that nSMase2 no longer mediates exosomal packaging of PrP Sc . Analysis of exosome markers Flotillin-1 and Tsg101 levels in non-and prion-infected exosomes also supported nSMase2B KD reducing exosome biogenisis in non-infected cells, with no changes observed for nSMase2B KD in prion-infected cells (Fig. 7C and  7D). While the effect of nSMase2 KD on PrP levels in non-infected cells and exosomes is in agreement with GW4869 treatment, the effects of nSMase2 KD on PrP levels in prion-infected cells differs from that of GW4869 treatment. This discrepancy suggests that there could be compensation between the nSMase isoforms. Together, these findings suggest that the nSMase pathway regulates exosome protein packaging in non-infected GT1-7 cells, with the process predominantly mediated by the nSMase2 isoform. However, upon prion infection, the pathway is altered such that nSMase2 no longer governs sorting of the infectious prion protein PrP Sc .

DISCUSSION
Multiple studies have now shown that the prion protein is secreted from cells in exosomes, highlighting a novel role for these extracellular vesicles in intercellular dissemination of prions (4,10,13,47). Despite the growing number of reports establishing an association of PrP with exosomes, the cellular pathways which regulate packaging of PrP has eluded identification. We aimed to address this by investigating the potential involvement of the nSMase pathway in exosome biogenesis and protein packaging. Using chemical inhibition and RNAi, we have shown that the nSMase pathway regulates exosome biogenesis in mouse neuronal cells. In addition, the nSMase pathway also modulates PrP packaging into exosomes. While nSMase2 appears to be the predominant isoform responsible for PrP C packaging, an alternative isoform could be responsible for packaging PrP Sc into exosomes.
In agreement with Trajkovic et al., treatment with GW4869 impairs exosome biogenesis and protein packaging, and this is further supported by the findings obtained using RNAi (26). Silencing of nSMase1 and nSMase2 decreases exosome release, indicating that exosome biogenesis in GT1-7 cells is regulated by the nSMase pathway. In addition, it also reveals that nSMase1 and nSMase2 share redundancy with respect to regulating exosome biogenesis. Comparison of the effect of nSMase1 and nSMase2 KD on exosome release from non-and prion-infected cells reveal an intriguing relationship. nSMase2 KD causes a larger decrease of exosome release from non-infected cells than nSMase1 KD. Conversely, nSMase1 KD causes a larger decrease of exosome release from infected cells than nSMase2 KD. This suggests that while the two isoforms share functional redundancy in terms of exosome formation, the relative contribution of the two isoforms towards this process varies depending on prion infection.
A similar scenario is observed with packaging of proteins into exosomes. While exosomal protein packaging in non-infected cells is not dependent on nSMase1, in infected cells, loss of nSMase1 decreases the amount of PrP Tot and PrP Sc in exosomes, albeit the decrease does not achieve statistical significance. Nonetheless, this suggests that prion infection alters the nSMase pathway such that nSMase1 plays a small role in protein packaging in prion-infected cells. The change with nSMase1 suggests that this isoform displays more nSMase activity in prion-infected cells than in non-infected cells. This is in accordance with the reported nSMase activity for nSMase1 under conditions of stress (45,46). Likewise, a change in nSMase2-regulated packaging of proteins is also observed upon infection, lending support to the proposal that the nSMase pathway is altered in prioninfected cells. Both nSMase1 and nSMase2 therefore mediate exosome biogenesis and the relative contribution of these isoforms differ depending on the prion-infected state of the cell.
While neither nSMase1 nor nSMase2 KD produces appreciable decreases of PrP Tot or PrP Sc in exosomes, treatment with GW4869 significantly impairs sorting of these proteins into exosomes in infected cells. This difference also suggests that there could be compensation occurring between nSMase2 and an alternative nSMase isoform, one which is also susceptible to inhibition by GW4869. GW4869 is proposed to inhibit the nSMase pathway by targeting nSMase2 and inhibiting activation of the isoform by anionic phospholipids (48). Co-incidentally, MA-nSMase has also been reported to be susceptible to activation by anionic phospholipids (32). Thus, GW4869 could also negatively regulate MA-nSMase activity, although how MA-nSMase could regulate exosome formation and protein packaging, given that it is located in the mitochondria, is enigmatic. Perhaps, upon prion infection, nSMase2 and MA-nSMase are redistributed subcellularly, resulting in a switch of exosome biogenesis regulation (Fig. 8). Indeed, mitochondrial structural abnormalities have been observed in prion-infected hamsters and mice (49)(50)(51)(52). Components of the mitochondria were detected in lysosomes, and as the endosomal system also sorts proteins destined for lysosomal degradation, it is possible that MA-nSMase could traverse the endosomal system in prion-infected cells en route to the lysosome. In doing so, it could also regulate exosome formation and protein packaging, and thus compensate for disruptions in nSMase2 activity. As tools for studying MA-nSMase become available, the localisation of this isoform in prion-infected cells and contribution towards exosome biogenesis and protein packaging will require investigation.
The roles of nSMase1 and nSMase2 in exosomal protein packaging is different and is also likely to be affected by subcellular localisation. Whereas nSMase1 is located in the ER and the nuclear matrix, nSMase2 is reported to be located in the Golgi, but can also be shuttled between the Golgi and plasma membrane via anterograde and retrograde transport (53)(54)(55). Once exported to the plasma membrane, nSMase2 activity is enhanced by anionic phospholipids in the membrane (56). Internalisation and retrograde transport of nSMase2, however, negatively regulates the activity of the enzyme (54). As nSMase2 transits the endosomal compartments, the isoform is able to hydrolyse sphingomyelin, leading to production of ceramide and initiation of exosome formation and protein packaging. Since the activity of nSMase2 is regulated by intracellular trafficking, it can be speculated that prion infection can abolish the role of nSMase2 in exosome protein packaging by modifying nSMase2 trafficking, diverting it away from the plasma membrane and endosomal compartments. Similarly, nSMase1 could also contribute towards PrP Sc packaging into exosomes as a result of altered subcellular trafficking due to cellular stress (Fig. 8). ER stress has previously been reported in prioninfected cells, and this could give rise to relocation of nSMase1 from the ER to the lysosome via the endosomal system (45,46). In addition, nSMase1 is also reported to display sphingomyelinase activity in vivo under stressed conditions, consistent with the observed role of nSMase1 in regulating PrP packaging into exosomes in prion-infected cells.
Additional support for the proposed role of the nSMase pathway in PrP packaging into exosomes can be taken from Naslavsky et al., in which decreasing sphingomyelin levels were linked to increasing PrP Sc levels (57). Based on the results presented here, decreasing sphingomyelin levels will impair the nSMase pathway and thus reduce exosome biogenesis and packaging of PrP Sc . Consequently, PrP Sc will accumulate inside the cell and stimulate further prion conversion, thereby increasing PrP Sc levels. Packaging and release of PrP Sc via exosomes could therefore be considered as beneficial for the cell. In fact, neuronderived exosomes have been reported to be internalised and degraded by microglia, which suggests that exosomes could represent an efficient mechanism for clearance of PrP Sc (58). Furthermore, exosomes could also facilitate clearance of other disease-associated proteins, such as Aβ, as exosomes generated in an nSMase2-dependent manner has been found to promote clearance of Aβ by microglia (28). Since PrP C has been proposed to function as a receptor for Aβ, it can be speculated that exosome-mediated clearance of extracellular Aβ could be facilitated by nSMase2-regulated incorporation of PrP C into exosomes.
Together, these data have provided further support for a role for the nSMase pathway in exosome biogenesis and also demonstrated a novel role for this pathway in regulating PrP C and PrP Sc packaging into exosomes. We also show that while nSMase2 is the predominant isoform mediating this process, other nSMase isoforms could also contribute, depending on the prion-infected status of the cell. Comprehensive analyses of all the nSMase isoforms in non-and prioninfected cells is thus required to fully characterise the roles of each isoform in exosome biogenesis and protein packaging. Understanding how the nSMase pathway modulates this process not only has important implications for understanding the unique transmissible nature of prions, but will also have important implications for other diseases such as Alzheimer's disease. The decrease in non-infected exosomes is close to, but did not achieve statistical significance (p < 0.061) (F) AChE quantitation of isolated exosomes, confirming a decrease in exosome release upon GW4869 treatment. Data presented as mean ± SEM, n = 3, * p < 0.05, ** p < 0.01, *** p < 0.001.   The decrease for noninfected nSMase1B exosomes is close to, but did not achieve statistical significance (p < 0.094) (B) AChE quantification of exosomes from nSMase2A and nSMase2B, also revealing a decrease in exosome release upon nSMase2 KD, although the decrease is slightly less from prion-infected than non-infected cells. The decrease for non-infected nSMase2B and prion-infected nSMase2A exosomes is close to, but did not achieve statistical significance (p < 0.078 and p < 0.061, respectively). Data presented as mean ± SEM, n = 3, * p < 0.05. FIGURE 6. Depletion of nSMase1 does not significantly impair exosomal PrP packaging. (A) Immunoblot analyses of PrP in nSMase1B and nSMase1C cells and exosomes. (B) Densitometric analyses show that cellular PrP C , PrP Tot and PrP Sc levels remain unchanged in nSMase1B and nSMase1C. While exosomal PrP C levels remain unchanged, exosomal PrP Tot and PrP Sc levels decrease slightly in prion-infected nSMase1C exosomes, although this is not statistically significant (p < 0.142 and p < 0.214, respectively). (C) Immunoblot analyses of Flotillin-1 and Tsg101 in nSMase1B and nSMase1C exosomes. (D) Densitometric analyses show that while exosomal Flotillin-1 levels remain unchanged in non-infected nSMase1B and nSMase1C exosomes, a small decrease is observed in prion-infected nSMase1C exosomes, although this is not statistically significant (p < 0.115). A similar decrease is also observed for Tsg101 in prion-infected nSMase1C exosomes (p < 0.140). Data presented as mean ± SEM, n =3.  Treatment of infected cells with GW4869 impedes sorting of PrP Sc into exosomes, indicating that compensatory mechanisms exist between the nSMase isoforms. Possible mechanisms include the cellular redistribution of MA-nSMase or nSMase1 from the mitochondria or endoplasmic reticulum, respectively, to the endosomal system in prioninfected cells (dotted line).