However, in some circumstances MVBs, instead of fusing with the lysosome, fuse with the plasma membrane and release their inner vesicles, or exosomes. In this paper, Mikael Simons...  Read more

However, in some circumstances MVBs, instead of fusing with the lysosome, fuse with the plasma membrane and release their inner vesicles, or exosomes. In this paper, Mikael Simons and colleagues propose for the first time a mechanism for the inward budding of MVBs’ limiting membrane that precedes exosome formation. The authors convincingly demonstrate that, at least in an oligodendrocyte cell line, ceramide is responsible for the inward budding and sorting of an exosome’s cargo into MVBs’ inner vesicles.

The authors designed an interesting and unconventional assay to study MVBs’ inner vesicle formation by fluorescence microscopy and overexpression of constitutively active Rab5. This assay allowed the visualization and quantification of inner vesicles in the lumen of enlarged endosomes, which enabled the identification of this new mechanism. Unfortunately, the authors did not show electron microscopy data confirming the role of ceramide in the formation of MVBs’ inner vesicles in non-overexpressing cells. Further investigation is necessary before one can generalize this new mechanism to all cell types, since oligodendrocytes are very specialized and their trafficking may well be optimized to produce myelin. Indeed, the literature is becoming richer with examples of cell types that adapt and create specialized compartments that bypass the more general endocytic pathway to perform their specialized tasks, as for example in melanosome biogenesis in melanocytes (Raposo and Marks, 2007).

Nevertheless, this paper brings a novel insight into the important role that lipid metabolism plays in protein trafficking. Moreover, the main players of this paper—MVBs, exosomes, and ceramide—have all been implicated in Alzheimer disease pathogenesis. Gunnar Gouras and others have demonstrated that MVBs are the main site of β amyloid accumulation in neurons (Gouras et al., 2005). More recently, in the Gouras lab, we showed that the sorting of membrane receptors for degradation at MVBs is altered by β amyloid (Almeida et al., 2006). Kai Simons’s group has reported that exosome secretion contributes to the extracellular accumulation of β amyloid (Rajendran et al., 2006), and Mark Mattson’s group observed that ceramide levels are increased in Alzheimer disease (Cutler et al., 2004). Therefore, I think that it will be very interesting to investigate if this mechanism of ceramide-dependent exosome formation exists in neurons. If it exists, does it mean that the increased levels of ceramide in Alzheimer disease are involved in upregulating the secretion of β amyloid via exosome release? If so, how does that contribute to the progression of the disease?

References:
Hanson PI, Roth R, Lin Y, Heuser JE. Plasma membrane deformation by circular arrays of ESCRT-III protein filaments. J Cell Biol. 2008 Jan 28;180(2):389-402. Abstract

Raposo G, Marks MS. Melanosomes--dark organelles enlighten endosomal membrane transport. Nat Rev Mol Cell Biol. 2007 Oct;8(10):786-97. Abstract

Gouras GK, Almeida CG, Takahashi RH. Intraneuronal Abeta accumulation and origin of plaques in Alzheimer's disease. Neurobiol Aging. 2005 Oct;26(9):1235-44. Abstract

Almeida CG, Takahashi RH, Gouras GK. Beta-amyloid accumulation impairs multivesicular body sorting by inhibiting the ubiquitin-proteasome system. J Neurosci. 2006 Apr 19;26(16):4277-88. Abstract

Rajendran L, Honsho M, Zahn TR, Keller P, Geiger KD, Verkade P, Simons K. Alzheimer's disease beta-amyloid peptides are released in association with exosomes. Proc Natl Acad Sci U S A. 2006 Jul 25;103(30):11172-7. Abstract

Cutler RG, Kelly J, Storie K, Pedersen WA, Tammara A, Hatanpaa K, Troncoso JC, Mattson MP. Involvement of oxidative stress-induced abnormalities in ceramide and cholesterol metabolism in brain aging and Alzheimer's disease. Proc Natl Acad Sci U S A. 2004 Feb 17;101(7):2070-5. Abstract

View all comments by Claudia Almeida

Author Q&A with Mikael Simons and Lawrence Rajendran. Questions by Gabrielle Strobel.

Q: What is the gist of your new study?
A: Exosomes are small vesicles with a diameter of approximately 50-100 nm that are secreted by a number of different cells. Exosomes are thought to derive from intraluminal vesicles of multivesicular endosomes that are able to fuse with the plasma membrane, leading to the release of these exosomes into the extracellular milieu, where they may function in a multitude of intercellular signaling processes. In this study, we analyzed the molecular mechanisms of exosome release. As a model system, we used an oligodendroglial cell line to analyze how these intraluminal vesicles are generated and how they segregate from the intraluminal vesicles destined for cargo degradation in lysosomes. We found that exosomal cargos are segregated into distinct subdomains on the endosomal membrane and that the transfer of exosome-associated domains into the lumen of the endosome required the sphingolipid ceramide. Purified exosomes were...  Read more

Author Q&A with Mikael Simons and Lawrence Rajendran. Questions by Gabrielle Strobel.

Q: What is the gist of your new study?
A: Exosomes are small vesicles with a diameter of approximately 50-100 nm that are secreted by a number of different cells. Exosomes are thought to derive from intraluminal vesicles of multivesicular endosomes that are able to fuse with the plasma membrane, leading to the release of these exosomes into the extracellular milieu, where they may function in a multitude of intercellular signaling processes. In this study, we analyzed the molecular mechanisms of exosome release. As a model system, we used an oligodendroglial cell line to analyze how these intraluminal vesicles are generated and how they segregate from the intraluminal vesicles destined for cargo degradation in lysosomes. We found that exosomal cargos are segregated into distinct subdomains on the endosomal membrane and that the transfer of exosome-associated domains into the lumen of the endosome required the sphingolipid ceramide. Purified exosomes were enriched in ceramide, and the release of exosomes was reduced after the inhibition of neutral sphingomyelinases. To directly explore the role of ceramide in the budding of a liquid-ordered lipid phase, we prepared giant unilamellar vesicles that form two different lipid phases. After adding sphingomyelinase, small vesicles start to bud, only from the liquid-ordered phase, and to accumulate into the lumen of the large lipid vesicles. We conclude that ceramide, possibly due to its cone-shaped structure, induces budding of membrane to produce intraluminal vesicles in endosomes that are subsequently secreted as exosomes.

Q: This is beautiful basic cell biology research. Could it be relevant to AD? Why?
A: In this work we show that ceramide is responsible for the generation of one class of exosome. Previous work from us has shown that Aβ that is formed in endosomes is then transported to MVBs. The fusion of multivesicular bodies with the plasma membrane releases the intraluminal vesicles of MVBs as exosomes. In the absence of extracellular plaques, Aβ is sequestered in MVBs in mice that overexpress Arc/Swe APP (Rajendran et al., 2007). The fact that these mice exhibit behavioral and cognitive deficits similar to that of other AD mice suggests that Aβ in the MVBs is the one to be blamed. Work from the labs of Gunnar Gouras and Frank LaFerla has stressed the importance of intracellular amyloid for a long time. So, in the light of our new findings that ceramide is responsible for formation of exosomes, and also taking into account that ceramide is elevated in AD, one may speculate that ceramide increases Aβ release by exosomes or Aβ aggregation in MVBs. The reader should be aware that we have made this observation only in oligodendrocytes, and further work in hippocampal neurons and other cells will show if exosomal MVBs are generated via a ceramide-mediated mechanism.

Q: Some data exist on ceramide elevation in AD brain. What implications do they have in light of your current finding?
A: Mark Mattson’s lab showed that AD brains have elevated levels of ceramide. Work from Dora Kovacs’s lab shows that ceramide levels stabilize BACE and are necessary for Aβ generation. One should also note that oxidative stress activates sphingomyelinases and increases ceramide levels. Hence, one could assume that ceramide, for example, produced during oxidative stress, accelerates Aβ production in multivesicular bodies and Aβ-release via the exosomal pathway. There is also work indicating that the Aβ peptides could directly modulate lipid metabolism. Work from Tobias Hartmann’s group shows that Aβ42 activates neutral sphingomyelinase to produce ceramide. In this paper the authors also show that inhibition of neutral sphingomyelinase reduces Aβ levels (Grimm et al., 2005). It seems that ceramide regulates Aβ production, and Aβ itself is part of a feedback limiting the production of Aβ. The whole scenario is still very speculative, and more work is required in this direction. It also is not clear whether ceramide has any role in the regulation of Aβ release via exosomes. And we do not know to what extent Aβ is secreted in association with exosomes in vivo, and whether this fraction is relevant to the aggregation process.

Q: Where does the retromer intersect with the endosome/exosome/MVB membranes you analyzed? (For a review out this week, see Small, 2008).
A: Retromers play a crucial role in retrieving the players of APP processing, such as APP/BACE, from the early endosomes and sort it to the trans-Golgi network TGN. Any defect in this sorting (via the retromer interaction with SorLA) accumulates APP/BACE in early endosomes, and this would lead to an increase in Aβ production. Whether there is a direct link between the retromer and transport to MVBs that secrete their content as exosomes is not known. One could only assume that in early endosomal membranes, there must be a way to regulate the sorting of different cargos either to the TGN or to MVBs that end up in lysosomes, and to such vesicles that eventually secrete their content as exosomes. The aspect of sorting several cargoes at the early endosomal level is simply fascinating. Clearly, more work is needed on this front.

Q: Do changes in ceramide levels affect Aβ generation, and if so, where?
A: Yes, there are data showing that ceramide increases Aβ generation. This seems to occur by stabilizing BACE by increasing the half-life of BACE (Puglielli et al., 2003). Consistent with this result, another paper reports that inhibition of neutral sphingomyelinase reduces Aβ levels (Grimm et al., 2005). Ceramide is known to induce the formation of larger membrane domains, named ceramide-enriched membrane domains or platforms (for review, see Grassme et al., 2007). These domains may facilitate Aβ generation by BACE. It is possible that these domains are formed within the endosomal membrane where BACE cleavage is known to occur. However, one should not forget that ceramide is an important signaling molecule, and many of its effects could also be explained by triggering various signal transduction pathways.

Q: How do your data inform the debate about extracellular vs. intraneuronal Aβ accumulation early on in AD?
A: A recent PNAS paper (Khandogin and Brooks, 2007) showed that oligomerization of Aβ could already happen at the place where it is produced, i.e., in early endosomes. Oligomerization of the natural amyloid protein Pmel17 occurs in melanosomes, an MVB-related organelle. In a process with potential similarities to APP, Pmel17 undergoes proteolytic processing in endosomes, and the fibril formation of the cleaved product also occurs in multivesicular bodies (Raposo and Marks, 2007). Recent work from our group on Aβ in MVBs that we did in collaboration with Roger Nitsch’s group also points to the same direction (Rajendran et al., 2007). Moreover, the gradual accumulation of the oligomers in the extracellular medium and the release of exosomal vesicles via the fusion of MVBs could drastically reduce the threshold for fibril formation (Yuyama et al., 2007).

Q: You worked in an oligodendrocyte line. What about neurons?
A: In this study, we used an oligodendroglial cell line because one of Mika Simons’s main interests is to understand the mechanisms of myelin biogenesis. Myelin has a special lipid composition rich in galactosylceramide and cholesterol. Because they produce myelin, oligodendrocytes synthesize large amounts of these lipids. We found that the formation of exosomes depends on these lipids, and it is therefore not clear whether neurons form similar types of exosomes. In our paper, we did not make any connection to Alzheimer disease, but it is clear that one needs to look at neurons to study this question. We have not studied exosome release in primary neurons, but this is something we plan to do.

Q: What is the next question to pursue in exploring the relevance of this work to age-related neurodegeneration?
A: There is increasing evidence that intracellular traffic could contribute to neurodegeneration, and trafficking deficits are observed in these diseases. Mutations in CHMP2B, a gene responsible for encoding a component of the MVB-associated ESCRTIII complex, were causally associated to frontotemporal dementia (van der Zee et al, 2008). Mark Mattson’s group showed that ceramide levels are elevated in ALS brain samples. It is also interesting to note that Cu/Sn superoxide dismutase is secreted via exosomes (Gomes et al., 2007). In addition, cells release prions in association with exosomes (Fevrier et al., 2004). So there is increasing evidence that intra-endosomal trafficking and exosomes may have a role in neurodegeneration, but a conclusive statement can only be made in some years to come.

ARF: We thank you for this interview.

View all comments by Gabrielle Strobel

Schneider et al. have added a new protagonist in this interplay among APP, the secretases, and cholesterol. In this paper, they indeed describe a potentially important link between AD and flotillin. Flotillin-1 and -2 are proteins anchored at the cell membrane. They are associated with lipid rafts, the 50 to 100 nm large microdomains, enriched in cholesterol, which seem to "float" over the membrane glycerophospholipids. In lipid rafts, the diffusion coefficients are smaller than in non-raft domains: molecules are less...  Read more

Schneider et al. have added a new protagonist in this interplay among APP, the secretases, and cholesterol. In this paper, they indeed describe a potentially important link between AD and flotillin. Flotillin-1 and -2 are proteins anchored at the cell membrane. They are associated with lipid rafts, the 50 to 100 nm large microdomains, enriched in cholesterol, which seem to "float" over the membrane glycerophospholipids. In lipid rafts, the diffusion coefficients are smaller than in non-raft domains: molecules are less mobile and can probably more easily interact. This has suggested a new way of regulating enzymatic reactions: β-cleavage of APP, for instance, could be regulated by the movement, and collision of rafts containing either APP or BACE, and by their encounter (Ehehalt et al., 2003).

Schneider et al. show that APP endocytosis is reduced when production of flotillin-2 is inhibited by siRNA. Using the modern tools of fluorescent microscopy, they conclude that flotillin-2 is associated with an increased clustering of APP at the cell surface, which can be mimicked by antibodies: a primary antibody that labels APP promotes clustering of APP when a bivalent secondary antibody is applied, an effect that is not obtained if the secondary antibody is monovalent. The antibody mediated clustering also promotes APP endocytosis. The decrease in APP endocytosis induced by depletion of the cholesterol content of the membrane can be overcome by antibody mediated clustering—an observation that suggests that cholesterol also plays a role in APP clustering. APP endocytosis requires clathrin, as Schneider et al. demonstrate with the use of dominant positive mutants involving proteins of the clathrin pathway, a pathway that could appear more complex than presently thought: transferrin endocytosis, which is also clathrin dependent, is indeed affected by an epsin 1 mutant while APP endocytosis is not.

These results have now to be transferred to human pathology: was the antibody used to detect the flotillin accumulation in tangle-bearing neurons (Girardot et al., 2003) really specific for flotillin-1, or did it also recognize flotillin-2 that, at least in the cell systems used by Schneider et al., seems to be the active molecule? How and why could flotillin accumulate in tangle-bearing neurons? Could this accumulation interfere with Aβ production? Much remains to be done, on the other hand, concerning the clathrin pathway and AD lesions. Literature on these topics is rather old and should be reanalyzed in view of these new results.

The data of Schneider et al. show that the understanding of Alzheimer pathogenesis needs more than enzymatic activities in a test tube: the dynamic topography of the protagonists is crucial, as is the biophysics of cell membrane that allow or prevent the gathering of the protagonists.

We are personally involved in research that aims to understand the relationship between membrane cholesterol and the clathrin mediated pathway. Our current results in HEK293 cell cultures also show that increasing plasma membrane cholesterol promotes APP endocytosis and the production of Aβ through a clathrin dependent pathway. Since enlargement of Rab 5-positive endosomes are the earliest appearing Alzheimer disease’s specific cellular pathology that is also modulated by APP gene dosage (Nixon et al., 2004), is Alzheimer disease a disease of the endocytosis machinery?

References:
Ehehalt R, Keller P, Haass C, Thiele C, Simons K. Amyloidogenic processing of the Alzheimer beta-amyloid precursor protein depends on lipid rafts. J Cell Biol. 2003 Jan 6;160(1):113-23. Abstract

Girardot N, Allinquant B, Langui D, Laquerrière A, Dubois B, Hauw JJ, Duyckaerts C. Accumulation of flotillin-1 in tangle-bearing neurones of Alzheimer's disease. Neuropathol Appl Neurobiol. 2003 Oct;29(5):451-61. Abstract

Nixon RA. Niemann-Pick Type C disease and Alzheimer's disease: the APP-endosome connection fattens up. Am J Pathol. 2004 Mar;164(3):757-61. Abstract

View all comments by Charles Duyckaerts
View all comments by Marie-Claude POTIER