|Home | About | Journals | Submit | Contact Us | Français|
Prion protein modulates many cellular functions including the secretion of trophic factors by astrocytes. Some of these factors are found in exosomes, which are formed within multivesicular bodies (MVBs) and secreted into the extracellular space to modulate cell-cell communication. The mechanisms underlying exosome biogenesis were not completely deciphered. Here, we demonstrate that primary cultures of astrocytes and fibroblasts from prnp-null mice secreted lower levels of exosomes than wild-type cells. Furthermore, prnp-null astrocytes exhibited reduced MVB formation and increased autophagosome formation. The reconstitution of PRNP expression at the cell membrane restored exosome secretion in PRNP-deficient astrocytes, whereas macroautophagy/autophagy inhibition via BECN1 depletion reestablished exosome release in these cells. Moreover, the PRNP octapeptide repeat domain was necessary to promote exosome secretion and to impair the formation of the CAV1-dependent ATG12–ATG5 cytoplasmic complex that drives autophagosome formation. Accordingly, higher levels of CAV1 were found in lipid raft domains instead of in the cytoplasm in prnp-null cells. Collectively, these findings demonstrate that PRNP supports CAV1-suppressed autophagy to protect MVBs from sequestration into phagophores, thus facilitating exosome secretion.
PRNP (prion protein) is a glycoprotein found attached to the cell membrane through a glycosylphosphatidylinositol anchor.1 Major conformational changes that convert PRNP to its abnormal form (PrPSc) are associated with the generation of prions, the infectious agents that cause transmissible spongiform encephalopathies.2 The mechanisms associated with PrPSc in neurodegeneration underlying transmissible spongiform encephalopathies and the physiological functions of PRNP in the nervous system have been largely explored.3,4 PRNP cellular functions are attributed to its interaction with a set of protein complexes such as LAMA1/laminin-1-GRM/metabotropic glutamate receptors and STIP1 (stress-induced phosphoprotein 1)-CHRNA7/α7 nicotinic acetylcholine receptors.5-7 Some PRNP interactions take place in lipid raft microdomains such as those involving NCAM (neural cell adhesion molecule) and CAV1/caveolin-1, which couple PRNP to the tyrosine kinase FYN (Fyn proto-oncogene).8,9 These complexes modulate neuroprotection against cellular and systemic insults,10-12 participate in cell signaling cascades,13 promote neuritogenesis,14 regulate neuronal plasticity, excitability, memory formation, and consolidation,15 and mediate synaptic vesicle transmission.16
PRNP is abundantly expressed in neuronal cells, but has also been observed in non-neuronal cells.17,18 In astrocytes, PRNP modulates survival and differentiation via the PRKA/protein kinase A and MAPK1/ERK2-MAPK3/ERK1 pathways.19 Some studies have addressed the role of PRNP in neuron-astrocyte communication. For instance, glutamate uptake by astrocytes is dependent on PRNP expression, and may influence neuronal survival.20 In conditioned media from astrocytes, STIP1 is found within MVB-derived extracellular vesicles (EVs) containing classical exosome markers.21 STIP1-containing exosomes are responsible for PRNP-dependent neuronal survival. Accordingly, glioblastomas, which are tumors derived from astrocyte precursors, also secrete STIP1 whose interaction with PRNP at the cell surface modulates the in vitro and in vivo proliferation of these tumor cells.22 Remarkably, conditioned media from prnp-null astrocytes promote a reduced neuronal survival compared to media obtained from wild-type astrocytes.19 Nonetheless, the role of PRNP in the secretion of trophic components, particularly those sorted to exosomes, is still unknown.
EV secretion from cells represents a remarkable system for biomolecule transfer that serves as a general mode of cell-cell communication. EVs are classified into 3 main classes based on their biogenesis: exosomes, microvesicles, and apoptotic bodies.23 In contrast to microvesicles, which are generated by budding from the plasma membrane, exosomes are derived from the endolysosomal pathway.24 The biogenesis of exosomes begins by invagination and budding of the limiting membrane of late endosomes, generating vesicle-laden endosomes, or multivesicular bodies (MVBs).25 The proteins of the ESCRT/Endosomal Sorting Complex Required for Transport are the best-described components contributing to MVB and intralumenal vesicle (ILV) biogenesis. The first complex, ESCRT-0, binds to the cargo on endosomes and with the help of the ESCRTs-I, -II, and –III, promotes cargo accumulation on the endosomal membrane. At the end of sorting, an AAA-type ATPase, VPS4A (vacuolar protein sorting 4 homolog A), disrupts the ESCRT complexes, and the cargo-containing membrane invaginates into the maturing endosome.26,27
Different subpopulations of MVBs coexist simultaneously in cells. MVB-derived ILVs can be secreted as exosomes by MVB fusion to the plasma membrane or can serve as transporters for the delivery of proteins to the lysosome for degradation.24,28 A specific degradation of damaged proteins and organelles occurs in autophagic-lysosomal compartments in a catabolic process called autophagy. This process is controlled by specialized set of gene products that promote the genesis of autophagosomes, a well-characterized structure of the autophagy pathway, substrate recruitment, lysosome–autophagosome fusion, and final degradation of autolysosome contents.29 BECN1/Beclin 1 and ULK1 (unc-51 like kinase 1)-ULK2, are key autophagy-promoting factors that are activated by a complex phosphorylation cascade downstream of AMP-activated protein kinase (AMPK) and MTOR (mechanistic target of rapamycin [serine/threonine kinase]).30 Remarkably, autophagosomes can fuse with endocytic structures such as MVBs.31
prnp-null cells display increased autophagy,32 and several lines of evidence indicate a close relationship between the autophagy pathway and exosome biogenesis and secretion.33,34 Together with the fact that conditioned media from prnp-null astrocytes was only capable of weak neuroprotection,35 these observations guided us to explore the role of PRNP in exosome secretion. Herein, we used primary cultures from astrocytes and fibroblasts as well as plasma from wild-type, prnp-null and PRNP-overexpressing mice to evaluate exosome secretion. PRNP reconstitution in prnp-null cells or PRNP knockdown in wild-type cells was also conducted to address our hypothesis. The analysis of endocytic pathways and specific PRNP partners made it possible to decipher the mechanism associated with the participation of PRNP in the regulation of exosome secretion.
To evaluate if the expression status of PRNP affects exosome secretion, we analyzed conditioned media (CM) samples from primary cultures of astrocytes (Fig. 1A) and fibroblasts (Fig. 1B) collected from wild-type, prnp-null, and PRNP-overexpressing mice. We previously demonstrated that EVs isolated from astrocyte CM typically express exosomal markers and exhibit classical exosome morphology and size as evidenced by electron microscopy and nanoparticle tracking analysis (NTA).21 Additionally, our previous data showed that a dominant-negative variant of VPS4A, an AAA-ATPase essential for MVB biogenesis, reduces the release of EVs into astrocyte CM by 85%, indicating that a large fraction of these EVs are composed of exosomes.21 Our present results pointed out that CM from prnp-null astrocytes derived from ZrchI (prnp0/0) or Npu (prnp−/−) mice contained reduced levels of exosomes compared to the respective controls (Prnp+/+, Prnpwt/wt)(Fig. 1A). Conversely, exosome levels in the CM of astrocytes isolated from mice overexpressing PRNP (TG20) were considerably higher than those of wild-type CM (Fig. 1A).
Exosomes in the CM of mouse embryonic fibroblasts (MEFs) isolated from Prnp+/+, prnp0/0, and TG20 mice were also analyzed. In agreement with the results obtained in astrocytes, prnp0/0 MEFs secreted fewer exosomes, whereas TG20 cells secreted greater levels of exosomes compared to Prnp+/+ MEFs (Fig. 1B). Remarkably, MEFs secreted about 4 times more exosomes than astrocytes.
In order to confirm exosome isolation, we analyzed in our samples the presence of well-established exosome markers. FLOT1 (flotilin 1) and TSG101 (tumor susceptibility gene 101), proteins which are secreted associated with exosomes in several cell types25 were detected by western blot in exosomes derived from astrocytes (Fig. 1C) and fibroblasts (Fig. 1D). However, as expected prnp0/0 astrocytes and fibroblasts showed lower levels of these exosomal markers due to a reduced number of exosomes in the conditioned media from these cells.
Furthermore, we also analyzed the morphology of isolated exosomes from Prnp+/+, prnp0/0 and TG20 astrocytes and MEFs by electron microscopy. The phase-contrast electron micrographs of the exosomes revealed rounded and double-membraned structures with a size of approximately 100–150 nm (Fig. 1E).
Exosomes can also be detected in bodily fluids. Therefore, we measured circulating exosomes in plasma isolated from the mice described above. prnp0/0 animals show a significant decrease in the levels of circulating exosomes, whereas greater levels of exosomes were detected in the plasma of TG20 mice (Fig. 1F). Taken together, our data indicate a correlation between PRNP expression and levels of secreted exosomes.
The NTA software identified and measured particles of the expected exosomal size, averaging 150 nm, and there were no significant differences in exosome size distribution between groups (Fig. S1).
To confirm that exosome secretion is associated with PRNP levels, we knocked down PRNP expression in Prnp+/+ astrocytes using small interfering RNA (siRNA) (Fig. 2A and B) or restored PRNP expression in prnp0/0 astrocytes (Fig. 2D). The levels of exosomes released by PRNP-depleted cells were markedly reduced compared to control cells (Fig. 2C), whereas PRNP re-expression almost completely rescued the secretion of exosomes to Prnp+/+ levels (Fig. 2E). The expression of a PRNP mutant that is correctly inserted into the plasma membrane, but is unable to be endocytosed (*N-PrP-3F4),36 also restored exosome secretion in prnp0/0 astrocytes (Fig. 2E). Conversely, a PRNP mutant lacking the amino-terminal leader peptide resulting in its intracellular retention (EGFP-PRNP[Δ1–22]) failed to restore exosome secretion in prnp0/0 astrocytes (Fig. 2E). Collectively, these results demonstrate that PRNP is required for regulating exosome secretion, and must be localized to the plasma membrane to properly perform this function.
In the endocytic pathway, as early endosomes transition into late endosomes, exosomes form by inward budding of the late endosome/MVB.37,38 Therefore, we examined if PRNP levels affect the distribution of different endocytic compartments. We assessed the expression of RAB5, an early endosomal marker, RAB7, a late endosomal marker and RAB11A, a recycling endosome marker in astrocytes isolated from Prnp+/+ and prnp0/0 mice by immunofluorescence. RAB5 staining in early endosomes appeared punctate and diffuse throughout the cytoplasm (Fig. S2A), and RAB7 staining in late endosomes was more concentrated around the nucleus (Fig. S2B), whereas RAB11A appeared in a punctate pattern in the cell periphery (Fig. S2C). There was no alteration in the distribution of these endosomal compartments between Prnp+/+ and prnp0/0 astrocytes.
We further analyzed the distribution of the late endosomal phospholipid lysobisphosphatidic acid (LBPA), which is incorporated into the ILVs of MVBs.39 Confocal imaging revealed that Prnp+/+ astrocytes exhibited LBPA staining in a punctate pattern, especially in perinuclear areas, suggesting that LBPA is abundant in ILVs and therefore labels mature MVBs in these cells. Conversely, poor staining was observed in prnp0/0 cells (Fig. 3A). Transmission electron microscopy confirmed the presence of multivesicular structures with the characteristic MVB morphology in Prnp+/+ cells (Fig. 3D). However, MVBs were not found in prnp0/0 cells.
The fusion of the MVB with the plasma membrane allows the release of exosomes into the extracellular space, but alternatively, MVBs may fuse with the lysosome to facilitate degradation of their content.40 Staining of Prnp+/+ cells with antibodies against LAMP1 (lysosomal-associated membrane protein 1), a lysosomal/late endosome marker and LAMP2, a specific lysosomal marker, showed that lysosomes were widely dispersed throughout the cell. Conversely, lysosomes from prnp0/0 cells were larger and aggregated in the perinuclear region. Additionally, there was a large increase in the amount of LAMP1 and LAMP2 in prnp0/0 cells (Fig. 3B and C). These results suggest that PRNP decreases the late endosomal trafficking of ILV-containing MVBs toward lysosomes. It has been demonstrated that under conditions of serum deprivation, autophagy is activated in prnp0/0 hippocampal neurons compared to Prnp+/+ neurons.41 Furthermore, lysosome activity during the course of autophagy requires the fusion between autophagosomes and lysosomes.42 Transmission electron microscopy confirmed that a large number of autophagic structures (phagophores, autophagosomes, and late-autophagic compartments) were present in prnp0/0 astrocytes compared to Prnp+/+ cells (Fig. 3E).
To further explore the role of PRNP in autophagy and its correlation with exosome secretion, we monitored autophagosome biogenesis under different cellular conditions in cells expressing different levels of PRNP and compared it with the levels of exosome secretion. Autophagosome biogenesis was evaluated by fluorescence microscopy (Fig. 4A and B) of MAP1LC3B (microtubule-associated protein 1 light chain 3), a protein involved in phagophore formation.43 During autophagy, the cytosolic form of MAP1LC3B (MAP1LC3B-I) is conjugated to phosphatidylethanolamine to generate MAP1LC3B-II, which is then recruited to phagophore membranes.44 The expression construct encoding EGFP-MAP1LC3B was transfected into cells to visualize MAP1LC3B without any additional labeling. The number of EGFP-MAP1LC3B-II puncta was quantified in Prnp+/+, prnp0/0 and TG20 astrocytes maintained in 10% fetal bovine serum (FBS), under serum-free conditions (starved), or after treatment with 100 nM rapamycin. Cellular extracts from cells treated with the same conditions were used to evaluate MAP1LC3-II expression by western blot (Fig. 4C and D).
In normal culture conditions (10% FBS), prnp0/0 astrocytes showed higher levels of EGFP-MAP1LC3B puncta and MAP1LC3-II levels, whereas lower levels of these markers were observed in TG20 cells when compared to wild-type astrocytes (Fig. 4A and D). Under serum-starved and rapamycin-treated conditions, Prnp+/+ astrocytes exhibited a large increase in EGFP-MAP1LC3B puncta (Fig. 4A and B) as well as higher expression of MAP1LC3-II (Fig. 4C and D), which is indicative of autophagy induction. The number of EGFP- MAP1LC3B puncta (Fig. 4A and B) and MAP1LC3-II levels (Fig. 4C and D) did not change in prnp0/0 or TG20 cells after starvation and rapamycin treatments. In prnp0/0 cells autophagy levels were already high even in unstimulated conditions, suggesting that these cells already achieved a plateau in autophagy induction. Conversely, TG20 cells were completely resistant to stress-induced autophagy. These results highlight the fundamental role of PRNP in the induction of autophagy.
Remarkably, autophagy levels (Fig. 4B) in cells under normal growth or under stress conditions were always inversely correlated with exosome secretion (Fig. 4E), confirming the existence of a PRNP-dependent autophagy-exosome axis. These results indicate that PRNP levels are inversely correlated with autophagosome formation and autophagy.
To confirm these observations, we depleted BECN1 in prnp0/0 cells to decrease the rate of autophagy. BECN1 participates in the regulation of autophagy and is required with its partner PIK3C3/VPS34 for the initiation of autophagosome formation.45 We confirmed the knockdown of BECN1 in prnp0/0 astrocytes transfected with 3 different Becn1-specific siRNAs (Fig. 5A and B). In parallel, we assessed the expression of MAP1LC3B-II, which correlates with the number of autophagosomes, in BECN1-depleted cells by western blot. MAP1LC3-II levels were decreased in BECN1-depleted astrocytes compared to nontransfected or control siRNA-transfected prnp0/0 cells (Fig. 5A and C), confirming that reduced BECN1 expression impairs autophagy. Remarkably, BECN1-depleted cells exhibited increased levels of exosomes secreted into the CM compared to control prnp-null cells (Fig. 5D). Taken together, these results indicate that PRNP is a negative modulator of autophagy, and that autophagy induction in prnp-null astrocytes disrupts exosome secretion.
Recently, it was reported that the octapeptide repeat domain of PRNP may play a pivotal role in controlling autophagy in prnp-null neuronal cells.41 To test if this region is also important for exosome secretion in astrocytes, prnp0/0 cells were transfected with constructs expressing GFP fused to full-length PRNP (PRNP-GFP), truncated PRNP mutants lacking the octapeptide repeat domain (GFP-PRNP[Δ51–90] and GFP-PRNP[Δ32–135]), and a truncated PRNP mutant lacking the hydrophobic domain (GFP-PRNP[Δ105–128]). The expression of each GFP-PRNP construct was confirmed by western blot (Fig. 6A). Cells expressing full-length PRNP or PRNP(Δ105–128) showed a reduction in autophagy as measured by MAP1LC3A/B-II levels (Fig. 6A and B) and by the formation of the ATG12 (autophagy-related 12)–ATG5 complex (Fig. 6A and C), which is involved in expansion of the phagophore, and is negatively regulated by CAV1.46 Remarkably, restoration of full-length PRNP or PRNP lacking the hydrophobic domain rescued the ability of prnp0/0 cells to secrete exosomes (Fig. 6D). Conversely, autophagy inhibition was not observed in cells expressing PRNP(Δ51–90) or PRNP(Δ32–135) (Fig. 6A, B and C), and these mutants failed to rescue exosome secretion (Fig. 6D). These results suggest that the octapeptide repeat domain of PRNP is important for attenuating autophagic flux and facilitating exosome secretion.
The PRNP octapeptide repeat domain is critical for some important biological functions, such as Cu2+ binding47 and PRNP internalization.48 It has been reported that the interaction between PRNP and CAV1 at the plasma membrane occurs through this region.9,49 Interestingly, CAV1 impairs the interaction between ATG5 and ATG12 in the cytoplasm, thereby suppressing autophagy.46 We thus investigated if CAV1 is affected by PRNP expression.
CAV1 protein levels were similar in astrocytes derived from Prnp+/+ or prnp0/0 mice (Fig. 7A). Lipid rafts are often enriched in CAV1, which is thought to play roles in cholesterol movement and the scaffolding of signaling molecules.50 Despite being an integral membrane protein a fraction of CAV1 is present in cells as a soluble cytoplasmic protein or associated with other organelles.51 Thus, depending on the cell location, CAV1 has distinct functions. Therefore, we characterized the cellular distribution of CAV1 by isolating lipid rafts via discontinuous sucrose density gradient ultracentrifugation. In Prnp+/+ cells, CAV1 was located in lipid raft domains (fractions 4–5) identified by the presence of FLOT1. CAV1 was also detected in high-density nonlipid raft fractions (fractions 7–12) corresponding to the cytoplasm, suggesting that CAV1 is shuttled back and forth between the cell surface and cytoplasm (Fig. 7B). Conversely, CAV1 was enriched in lipid raft domains (fractions 3–5), but was absent in high-density nonlipid raft fractions in prnp0/0 cells (Fig. 7B). Quantification of CAV1 levels in lipid raft domains revealed significantly higher levels in prnp0/0 cells than in Prnp+/+ cells (Fig. 7B). This result suggests that in the absence of PRNP, CAV1 may be arrested in lipid rafts of the plasma membrane instead of internalized. This would directly affect the inhibitory role of cytoplasmic CAV1 on ATG12–ATG5 engagement and autophagy stimulation, as observed in Fig. 6.
To confirm impaired CAV1 internalization in prnp0/0 cells, we examined the uptake of cholera enterotoxin B subunit (CTxB), which uses caveolae membrane invaginations to become internalized.50 Labeling of cells with Alexa Fluor 488-conjugated CTxB revealed it to be bound to the plasma membrane and internalized as punctate endosomal structures in Prnp+/+ cells (Fig. 7C). However, CTxB remained closer to the cell surface in prnp0/0 cells (Fig 7C). Quantification of the fluorescent signal confirmed that prnp0/0 cells internalized less CTxB than Prnp+/+ cells (Fig. 7D). Taken together, these results suggest that PRNP regulates the internalization and localization of CAV1 in lipid raft domains, which ultimately affects autophagy activation.
In this study, we demonstrated that PRNP is a positive regulator of exosome secretion. PRNP appears to control the distribution of CAV1 between lipid raft domains on the cell membrane and the cytoplasm where CAV1 can function to impair the ATG12–ATG5 complex and thus to inhibit autophagy progression. Under these conditions, MVBs form and exosomes are secreted. In the absence of PRNP at the membrane, CAV1 internalization is inhibited, ATG12–ATG5 complexes are formed, and autophagy is stimulated. Autophagosomes fuse to MVBs and both are delivered to lysosomes, thus sequestering the cellular source of exosomes and abolishing their secretion into the extracellular milieu (Fig. 8).
We investigated the underlying mechanisms by which PRNP regulates exosome secretion. We first found that this process requires the presence of PRNP at the plasma membrane, as a PRNP mutant lacking the membrane localization signal was incapable of restoring exosome secretion in prnp-null cells. Conversely, the internalization of PRNP into intracellular compartments was unnecessary for this function, thus indicating that exosome secretion regulation primarily occurs at the cell surface. In fact, the role of membrane proteins in exosome secretion has been recently explored. The transmembrane heparansulphate proteoglycans of the syndecan family are connected by the cytoplasmic soluble protein SDCBP/syntenin (syndecan binding protein) to PDCD6IP/ALIX (programmed cell death 6 interacting protein), an auxiliary component of the ESCRT machinery that supports endosomal membrane budding.52
Our initial observations suggested that in the absence of PRNP, the endocytic compartments of astrocytes were irregularly distributed, as seen by the reduction in ILVs and MVBs, the enhancement of lysosomal morphology, and the presence of autophagosomes. These data comply with the findings of other groups demonstrating that autophagy is upregulated in PRNP-deficient cells.53,54 Moreover, our findings are consistent with a connection between the endocytic pathway, autophagy, and exosome biogenesis and secretion.33,34 It is well known that autophagosome maturation requires ESCRT function, and that autophagosomes are able to fuse with endocytic vesicles (as MVBs) or lysosomes, which contain the hydrolytic enzymes required to degrade autophagosomal content.31,55-57 Additionally, diverse conditions that stimulate autophagy, such as serum starvation, rapamycin treatment, or MAP1LC3-II overexpression, inhibit exosome release.33 Our experiments demonstrating that exosome secretion in prnp-null cells was rescued by either PRNP re-expression or autophagy inhibition via BECN1 knockdown highlight, for the first time, that PRNP regulates exosome secretion by modulating autophagy induction.
Autophagy can be regulated by many different mechanisms,58 but the identification of the link between autophagy induction and PRNP is essential for understanding how PRNP facilitates exosome secretion. Our group proposed that PRNP acts to organize signaling molecules into lipid rafts due to its ability to interact with and modulate the activity of several membrane proteins.59 CAV1 is one well-described PRNP ligand that interacts with the octapeptide repeat domain to activate signal transduction events downstream of the kinases FYN and LYN (Lyn proto-oncogene Src family tyrosine kinase).49,60 Therefore, our findings that CAV1 is localized primarily within lipid rafts instead of the cytoplasm in prnp-null cells are in accordance with these previous studies. Remarkably, CAV1 has been described as an important negative regulator of autophagosome formation,61-62 given its ability to impair ATG12–ATG5 engagement.46 The introduction of full-length PRNP, but not the deletion mutants lacking the octapeptide repeat domain, in prnp-null astrocytes prevented cytoplasmic CAV1 function. Consequently, ATG12–ATG5 complex formation was impaired leading to diminished autophagy and increased exosome release. Finally, it is important to note that PRNP is internalized by caveolae.63 However, our results suggest that PRNP internalization is not necessary for controlling exosome secretion.
Primary astrocytes actively participate in neural development and communicate extensively with neurons.64 Exosomes secreted by glial cells not only contribute to the development of the central nervous system (CNS), but also promote the repair of neuronal injuries and regeneration of peripheral nerves.65,66 Glial cells may release SYN1 (synapsin I) into exosomes in response to stressful conditions, such as the oxidative stressor ischemia, to modulate neuronal outgrowth and neuron-glia interactions.67 In the peripheral nervous system, Schwann cells release exosomes that are specifically internalized by axons, resulting in increased axonal regeneration after sciatic nerve injury.68 Additionally, exosomes derived from multipotent mesenchymal stromal cells (MSCs) improve the functional recovery of neurons after stroke in a rat model.69
Notably, our present findings demonstrate that astrocytes and fibroblasts derived from mice overexpressing PRNP exhibited higher levels of secreted exosomes, indicating a general role for PRNP in regulating exosome secretion independent of cell type. Furthermore, plasma from these mice contained altered levels of exosomes, implying that exosome secretion in other cell types is also affected by PRNP expression.
Over the past few years, exosomes have emerged as important mediators of intercellular communication that facilitate the transmission of physiological signals between cells to regulate a diverse range of biological processes. Exosomes are also prominent mediators of neurodegenerative diseases. These vesicles can carry aggregated proteins or prions to recipient neurons to induce toxic effects. PrPSc is actively released into the extracellular environment associated with exosomes, and promotes the distribution of prions throughout the organism to cause prion diseases.70 Additionally, autophagy can play a protective role in the clearance of pathological PrPSc accumulated within neurons. Remarkably, defective autophagy may contribute to the formation of spongiform changes in prion disease.32 Interestingly, TG20 mice succumb to terminal disease more rapidly than wild-type mice,71 which is consistent with the lower autophagy activity observed in cells from these animals. Moreover, decreased autophagy in TG20 cells may contribute to the regulation of infection and spreading of PrPSc due to their improved secretion of exosomes.
In conclusion, our findings show for the first time that PRNP modulates the endocytic pathways that control exosome secretion. Given the pivotal role of exosomes in physiological and pathological conditions, targeting the mechanisms by which exosomes are secreted offers important therapeutic possibilities. Furthermore, the unpredicted function of PRNP in facilitating exosome secretion in addition to its convenient localization at the cell membrane highlight its potential promise as an amenable candidate for the development of therapeutic strategies seeking to target exosome delivery.
A polyclonal antibody against recombinant mouse PRNP was produced in prnp-null (prnp0/0) mice72 and was used at a dilution of 1:500. Primary antibodies against anti-LBPA (6C4, 1:250; Echelon Biosciences Inc., Z-SLBPA), anti-TUB1A1 (1:10,000; Sigma-Aldrich,T-9026), anti-RAB5 (1:100; Abcam, ab18211), RAB7 (1:100; Abcam, ab50533), Anti-RAB11A antibody (1:100; Abcam, ab65200) anti-LAMP1 (1:250; Abcam, ab25245), anti- LC3A/B (D3U4C) (1:1,000; Cell Signaling Technology, 12741), anti-BECN1 (1:1,000; Cell Signaling Technology, 3738), anti-ATG12 (D88H11, 1:1,000; Cell Signaling Technology, 4180), anti-LAMP2 (ABL-93, 1:50; Abcam, ab25339), anti-TSG101 (4A10, 1:500, Abcam, ab83) and anti-FLOT1 (1:1000; Merck-Millipore, Ab9292) were used in western blot and immunofluorescence experiments. Secondary antibodies included: goat anti-mouse Alexa Fluor 488 (1:1000; ThermoFisher Scientific, A-11001), goat anti-rabbit Alexa Fluor 488 (1:1000; ThermoFisher Scientific, A-11034) and goat anti-rat Alexa Fluor 488 (1:1000; ThermoFisher Scientific, A-11006). Expression vectors encoding GFP-PRNP, GFP-PRNP(Δ1–22), GFP-PRNP(Δ32–135), GFP-PRNP(Δ51–90), and GFP-PRNP(Δ105–128) were constructed and expressed as previously described.73,74 Plasmids for PRNP-3F4 and *N-PRNP-3F36 were kindly provided by Dr. R. Morris (King's College, London, England).
Two independent lines of prnp-null mice were used for astrocyte isolation, designated in this study as prnp0/0 (ZrchI)75 and prnp−/− (Npu).76 ZrchI prnp0/0 mice were provided by Dr. C. Weissmann (The Scripps Research Institute, Jupiter, FL, USA). Control mice (Prnp+/+) were generated by crossing F1 siblings from matings between 129/Sv and C57BL/6J mice. Npu prnp−/− mice were provided by Drs. B. Chesebro and R. Race (Rocky Mountain Laboratories, National Institute of Allergy and Infectious Diseases, Hamilton, MT, USA), and are descendants from the Npu line backcrossed to C57BL/10 mice over at least 20 generations. Prnpwt/wt mice were used as the respective control. PRNP-overexpressing mice (TG20) originally described77 were provided by Dr. C. Weissmann (The Scripps Research Institute, Jupiter, FL, USA). The Principles of Laboratory Animal Care (National Institutes of Health publication number 85–23, 1996) was strictly followed in all experiments, and all protocols were approved by the A.C. Camargo Cancer Center Animal Care and Use Committee (project number 058/13).
Astrocyte primary cultures were prepared from the cerebral hemispheres of embryonic day (E) 17 wild-type, prnp-null, and TG20 mice. Briefly, single-cell suspensions were obtained by dissociating cerebral hemispheres in Dulbecco's modified Eagle's medium (DMEM-low glucose) (ThermoFisher Scientific, 31600034) supplemented with 40 mg garamycin (Nova Farma, 42859) and 3 mM sodium bicarbonate (Sigma Aldrich, S5761). Cells were seeded on pre-coated poly-L-lysine plates and grown in DMEM containing 10% FBS (ThermoFisher Scientific, 12657). The medium was changed every 2 d. For transfection experiments, confluent astrocytes were incubated with 10 µg of Lipofectamine 2000 (ThermoFisher Scientific, 11668019) and 5 µg of plasmid DNA in OPTI-MEM (ThermoFisher Scientific, 51985034) for 4 h, and afterwards, the medium was changed to DMEM containing 10% FBS.
MEFs were prepared from E14 embryos. The head and organs were dissected, and fetal tissue samples were rinsed in phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4, 1.47 mM KH2PO4, pH 7.4), minced, trypsinized for 10 min at 37°C, and subsequently dissociated in DMEM. Cells were seeded on pre-coated poly-L-lysine plates and grown in DMEM containing 10% FBS. The medium was changed every 2 d.
CM was obtained as previously described.35 Confluent astrocytes or fibroblasts grown in 100-mm culture dishes were washed 3 times with PBS and incubated with serum-free medium for 48 h. CM was collected on ice and pre-cleared from cell debris by sequential centrifugation (1,500 xg for 10 min, 4,500 xg for 10 min, and 10,000 xg for 30 min). Exosomes were obtained by ultracentrifugation at 100,000 xg for 16 h using a SW 40 Ti rotor (Beckman-Coulter, 331302). The pelleted fraction was resuspended in PBS and centrifuged again at 100,000 xg for 16 h.
Mice were bled by cardiac puncture, and blood samples were centrifuged twice at 1,500 xg for 15 min at room temperature to obtain platelet-free plasma. The plasma was further centrifuged at 10,000 x g for 30 min at 4°C to remove cell debris, followed by ultracentrifugation at 100,000 xg for 2 h. The pellet was resuspended in PBS and ultracentrifuged again at 100,000 xg for 2 h.
Pellets of exosomes isolated from CM or plasma were resuspended in 1 ml of PBS. The number of particles and particle size were measured using a nanoparticle tracking analysis device NanoSight LM10 coupled to a CCD camera and a laser emitting a 60-mW beam at 405 nm (Malvern, United Kingdom). Video acquisitions were performed in 5 records of 60 s using the following parameters: shutter = 604, gain = 100, and threshold = 10. At least 1,000 particles were tracked in each sample.
To silence PRNP, wild-type astrocytes were transfected with target-specific siRNAs and a scrambled siRNA as a control. The target sequence for mouse Prnp was: 5′GGUUUUUGGUUUGCUGGGCTT3′. The scrambled control siRNA sequence was: 5′UGUUGUUGGCGUUUGUGGCTT3′. To silence BECN1 3 different siRNAs targeting mouse Becn1 mRNA and nontarget siRNA were used (Life Technologies, 4390771). Astrocytes were seeded on 100-mm plates and were 70–90% confluent at the time of transfection. 200 pmol of RNA duplexes were transfected using Lipofectamine 2000 reagent according to the manufacturer's instructions. The efficacy of siRNA-mediated knockdown was confirmed by isolating total cell lysates from siRNA-transfected or nontransfected cells 24 h after transfection and performing SDS-PAGE and western blot analysis with the specific antibodies listed under Reagents.
Astrocytes in culture were fixed in 4% paraformaldehyde (PFA; Polysciences Inc., 18814):2% glutaraldehyde (Sigma-Aldrich, 340855) in 0.1 M phosphate buffer, pH 7.4. After fixation, the samples were rinsed several times with PBS followed by post-fixation in 1% osmium tetroxide (EMS, 19150) in phosphate buffer for 1 h. The samples were rinsed with PBS for 15 min and dehydrated through a series of graded ethanol washes ranging from 70% to 100%. The samples were then immersed in an ethanol:epon mixture (1:1) and polymerized in pure epon (Polysciences, Inc. 02334-500) at 60°C for 48 h. Ultrathin sections were stained with uranyl acetate and lead citrate and imaged on a JEOL 1200 EX II transmission electron microscope (JEOL, USA).
Isolated exosomes were deposited onto formvar-carbon-coated electron microscopy grids (EMS, FCF200H-Cu), fixed with a mixture of 2% paraformaldehyde and 0.125% glutaraldehyde for 20 min. A 100-µL drop of PBS was placed on a sheet of parafilm and grids were transferred with the sample membrane side facing down using clean forceps for 2 min. The grids were kept wet on the side of the membrane during all steps, but dry on the opposite side. The grids were transferred to a 50-µL drop of 1% glutaraldehyde for 5 min before transferring to a 100-µL drop of distilled water for 2 min. This was repeated 7 times for a total of 8 water washes. To contrast the samples, grids were transferred to a 50-µL drop of uranyl-oxalate solution (4% uranyl acetate [EMS, 22400-4]; 0.15 M oxalic acid [Sigma-Aldrich, 75688] pH 7), for 5 min before transferring to a 50-µL drop of methyl-cellulose-UA, (a mixture of 4% uranyl acetate and 2% methyl cellulose [Sigma-Aldrich, M6385] in a ratio of 100 µL:900 µL), for 10 min, placing the grids on a glass dish covered with parafilm on ice. The grids were removed with stainless steel loops and excess fluid blotted gently on Whatman no.1 filter paper. Grids were left to dry and stored in appropriate grid storage boxes, then observed under a JEOL 1200 EX II transmission electron microscope.
Cells were cultured on glass coverslips, fixed in 4% PFA for 20min at room temperature, washed in PBS, blocked in a 5% bovine serum albumin (Sigma-Aldrich, A4503) solution in PBS, permeabilized in 0.5% Triton X-100 (Sigma-Aldrich, T8787), incubated with the primary antibodies listed under Reagents (diluted in PBS containing 0.5% bovine serum albumin), and then stained with the appropriate secondary antibodies, DRAQ5 (1:500; ThermoFisher Scientific, 62252), and rhodamine-phalloidin (1:50; ThermoFisher Scientific, R415). Coverslips were mounted onto glass slides using FluorSave reagent (Calbiochem, 345789). All fluorescent images were acquired on a Leica TCS SP5 confocal microscope (Leica Microsystems, Wetzlar, Germany). To quantify fluorescence levels, a single in-focus plane was acquired. An outline was drawn around each cell and circularity, area, mean fluorescence, and several adjacent background readings were measured using ImageJ (v1.48, NIH). The total corrected cellular fluorescence (TCCF) was calculated as the integrated density (area of selected cell × mean fluorescence of background readings).78
Cells were washed 3 times with pre-warmed PBS and then incubated in serum-free DMEM at 37°C for 24 h in the presence or absence of 100 nM rapamycin (Sigma-Aldrich, R0395). The medium was collected and exosomes were isolated as described above. Cell extracts were used for western blot analyses.
Cells were cultured on glass coverslips and tranfected with EGFP-LC3 plasmid (AddGene, 11546, deposited by Karla Kirkegaard). Transfected cells were fixed in 4% PFA for 20 min at room temperature and then rinsed with PBS. The nuclei were stained with DRAQ5. Slides were mounted with FluorSave mounting medium and examined by fluorescence microscopy. To quantify autophagy activation, at least 150 EGFP-LC3-expressing cells were analyzed and the number of puncta per cell, based on EGFP expression, was determined by using the ‘analyze particles' function in ImageJ.
Primary astrocytes were lysed for 10 min in cold TNE buffer (10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA) containing 0.5% Triton X-100, and Protease Inhibitor Cocktail (Roche, 11836153001). Extracts were centrifuged at 1,300 xg for 5 min. One ml of the cleared supernatant fraction was mixed with 85% sucrose (Merck Millipore, 107651) in TNE buffer and layered on the bottom of a Polyallomer 12-ml centrifuge tube (Beckman Coulter, 331372). The lysate was overlaid with 4 ml of 35% sucrose in TNE buffer, and finally with 4 ml of 5% sucrose in TNE buffer. The samples were centrifuged in a SW41 Ti rotor at 260,110 xg for 18 h at 4°C. At the end of the run, 1-ml fractions were collected from the top of the gradient and proteins were precipitated with 15% trichloroacetic acid (Sigma-Aldrich, T6399). The samples were then subjected to 10% SDS PAGE and western blot analysis.
Lipid raft labeling was performed using the Vybrant Alexa Fluor 488 Lipid Raft labeling kit according to the manufacturer's protocol (Thermo Fisher Scientific, V34403). Briefly, live cells were washed with ice cold PBS, labeled with Alexa Fluor 488-conjugated CTx-B, and crosslinked with anti-CTx-B antibody in serum-free media for 15 min at 4°C. Cells were shifted to 37°C for 10, 30, 45, and 60 min, washed with PBS twice, fixed in 4% PFA, stained with DRAQ5 to mark nuclei, and mounted onto glass slides before imaging on the Leica SP5 confocal system. For quantitative analysis, the fluorescence of internalized CTx-B was measured using ImageJ software (n = 100 cells).
All experiments were repeated 3 to 5 times. Statistical analyses were performed using GraphPad Prism 5 (La Jolla, CA). Experimental groups were compared using one-way or 2-way ANOVA followed by Tukey's post hoc test or the Student t test.
No potential conflicts of interest were disclosed.
We are very grateful to Dr. Luis Lamberti da Silva for reagents and scientific discussion, Dr. Renato Arruda Mortara for LAMP2 antibody and Dr. Richard Morris for PRNP plasmids.
This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP 09/14027-2) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq 467566/2014-3). Fellowships from FAPESP to MS (2010/19200-1) B.R.R (2012/19019-0), are gratefully acknowledged.