| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
MRC Centre for Neurodegeneration Research, Institute of Psychiatry, King’s College London, UK
1Correspondence: MRC Centre for Neurodegeneration Research, Department of Neuroscience (Box 037), Institute of Psychiatry, King’s College London, De Crespigny Park, Denmark Hill, London SE5 8AF, UK. E-mail: ritchie.williamson{at}iop.kcl.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: Alzheimer’s disease • ADDLs • protein aggregation • tau
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
. ADDLs have been reported to cause neuronal and synaptic dysfunction and to specifically target synapses in both neuronal culture and AD brain (4)
and to alter synaptic integrity (5
, 6)
. More specifically, soluble Aβ correlates better than plaque load with cognitive impairment (7)
and synaptic loss (8)
.
Recent reports have emphasized the importance of lipid rafts/detergent resistant membranes (DRMs) in the biogenesis and accumulation of amyloid protein (9
10
11)
. Increasing evidence now shows that Aβ can associate with lipid rafts and components of DRMs isolated from human and rodent brain as well as from cultured cells. Aβ was found to be tightly associated with GM1 ganglioside (GM1), and it was originally postulated that this may act as a seed for its accumulation (12)
. Subsequently, Aβ and phosphorylated tau were shown to accumulate in the DRMs from AD brain and mouse models of AD over expressing mutant APP (13
, 14)
, implicating lipid rafts in the process of Aβ aggregation.
Studies into the neurotoxic mechanisms induced by Aβ highlighted the potential importance of the nonreceptor tyrosine kinase fyn, with the finding that hippocampal slice cultures from fyn–/– mice are resistant to the toxic properties of ADDLs (15)
and that Aβ induces activation of fyn in neuronal cultures (16)
. The critical role of fyn extends to in vivo mouse models of Aβ-induced synaptotoxicity since synapse loss and cognitive decline in animals over expressing human amyloid precursor protein (hAPP mice) was reduced in animals lacking fyn (hAPP/fyn–/–) (17)
. Together these results suggest that fyn has an important role in amyloid-induced neurotoxicity.
Since a proportion of cellular fyn is located in DRMs, we undertook to investigate changes in DRM proteins in neurons exposed to ADDLs. Here, we describe a novel action of Aβ on neuronal plasma membranes in which Aβ binds to the neuronal membrane, redistributes over time and colocalizes with markers of lipid rafts; induces their aggregation; and accumulates in isolated DRMs. We also show here that ADDLs cause excess fyn to be recruited to DRMs and conversely that fyn expression is required for Aβ toxicity irrespective of whether fibrillar Aβ or ADDLs are employed. Furthermore, the ADDL-induced recruitment of fyn to DRMs is accompanied by the further recruitment of tau to DRMs. Attachment and accumulation of Aβ on the neuronal membrane was in itself not sufficient to cause cell death but required the fyn-dependent redistribution of Aβ and associated alterations in DRM protein composition.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The following primary antibodies were used against antigens: fyn, flotillin-1 (BD Transduction Laboratories, Lexington, KY, USA); Aβ species (6E10; Signet Laboratories, Dedham, MA, USA); phosphotyrosine (4G10; Upstate, Lake Placid, NY, USA); tau (Dako Diagnostics, Ely, UK); and actin (Abcam, Cambridge, UK).
Aβ preparations and treatments
Aβ1–42 peptide (California Peptide Research, Napa, CA, USA) was used to prepare ADDLs and fibrillar Aβ, following the method as described by Klein et al. (18)
, except that phenol red free Neurobasal medium (Life Technologies, Paisley, UK) was used in place of phenol red free F12 medium.
Mice
Fyn–/– mouse strain B6;129S7-Fyntm1Sor and the approximate control wild-type strain B6129SF2/J were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). Western blot analysis of brain from the fyn–/– and wild-type mice confirmed the absence of fyn expression but normal expression of the related protein tyrosine kinase, src, in the knock-out animals (Fig. 1
A, insert).
|
Primary cortical and hippocampal neuronal cultures
Neuronal cultures were prepared from day 18 rat embryos and from day 16 mouse embryos as described previously (16)
. Primary cortical cells were plated onto poly-L-lysine (10 µg/ml) coated 10-cm dishes at 9 x 106 cells/dish. Primary hippocampal cells were plated onto poly-L-lysine coated glass coverslips in 6-well tissue culture plates at 1.2 x 105 cells/well or 3 x 106 cells/dish were plated onto 10 cm dishes. Cultures were maintained in Neurobasal medium (without phenol red) containing B27 supplement, 2 mM glutamine, and 20 µg/ml gentamicin solution. Primary cortical cells were cultured for 7 days, and primary hippocampal cells were cultured for 14–21 days before being used for the treatments described.
DRM isolation
All procedures were performed on ice, and all buffers were kept at 0–4°C for the duration of the procedure. For whole cell lipid raft isolation 3 x 106 (hippocampal) or 9 x 106 (cortical) cells were lysed directly in 1% (w/v) CHAPSO in MBS buffer (25 mM MES, 150 mM NaCl, pH 6.5) containing 10 mM MgCl2, 10 mM NaF, 2 mM Na3VO4, 1 mM EGTA, 5 mM DTT, and 0.2 mM PMSF. The lysate was homogenized by Dounce homogenization (18 strokes) and incubated on ice for 30 min. The homogenate (1 ml) was then mixed with 1 ml of 90% (w/v) sucrose in MBS buffer and placed in a 12 ml ultracentrifuge tube. A discontinuous 5–35–45% sucrose gradient was formed by layering 6 ml of 35% (w/v) sucrose in MBS solution on top of the 2 ml homogenate, followed by 4 ml 5% (w/v) sucrose in MBS solution. The sample was then centrifuged at 39,000 rpm for 18 h at 4°C in a Beckman SW41 rotor (Beckman Instruments, Fullerton, CA, USA). A light scattering band at the 5–35% interface was identified that was enriched in flotillin, indicating the presence of DRMs. Fractions (12x1 ml) were collected from the top of each gradient, and DRMs in fractions 4–5 were concentrated by diluting in 10 ml MBS buffer containing 10 mM NaF and 2 mM Na3VO4 and by centrifugating at 39,000 rpm for 1 h in a Beckman SW41 rotor. The DRM-containing pellet was solubilized in 100 µl 20 mM Tris/8 M urea pH 7.4 containing 10 mM NaF, 2 mM Na3VO4, 5 mM DTT, and 0.2 mM PMSF. For neuronal plasma membrane lipid raft isolation, 6 x 106 hippocampal cells were lysed directly in HBL buffer (10 mM Tris, 10 mM NaCl, pH 7.4) containing 3 mM MgCl2, 10 mM NaF, 2 mM Na3VO4, and 1 mM EGTA. The lysate was incubated on ice for 10 min and homogenized by Dounce homogenization (40 strokes). The homogenate was then centrifuged at 375 g (av) at 4°C for 5 min. The pellet containing nuclear material was discarded, and the supernatant was centrifuged at 100,000 g (av) at 4°C for 30 min. The resulting pellet was washed in HBL buffer and centrifuged 100,000 g (av) at 4°C for 30 min 2 more times. The final membrane-containing pellet was then resuspended in 1% (w/v) CHAPSO in MBS buffer, and lipid rafts were isolated as described above.
Immunofluorescence microscopy
Hippocampal neurons grown on coverslips were fixed in 4% (w/v) paraformaldehyde (PFA) in phosphate-buffered saline (PBS) for 10 min at ambient temperature and blocked in 5% (v/v) fetal calf serum/0.2% (w/v) Tween 20 in Tris-buffered saline (TBS) for 30 min. Primary antibody was diluted in blocking solution and incubated for 1 h. Primary antibodies were detected using Alexa Fluor® conjugated goat anti-mouse or goat anti-rabbit antibodies (Invitrogen, Paisley, UK). For staining of surface GM1, nonpermeablized paraformaldehyde-fixed cells were incubated with CTXβ conjugated to FITC (0.1 µg/ml for 30 min). Images were captured using a Zeiss LSM 510 META confocal microscope (Carl Zeiss, Oberkochen, Germany). Images of amyloid puncta were quantified using LSM Image Examiner (Zeiss).
Toxicity assays
Lactate dehydrogenase release into the supernatant was assayed as described previously (16)
. MTS reduction was assayed using the CellTiter 96TM AQueous One Solution Cell Proliferation Assay (Promega UK Ltd., Southampton, UK) according to the manufacturer’s instructions, absorbance was read at 490 nm using a Perkin Elmer 1420 multichannel counter (Perkin Elmer, Beaconsfield, UK).
Western blotting
Protein concentrations were quantified by the method of Bradford (19)
. Equal amounts of total protein were resolved by SDS-PAGE using 10% (w/v) polyacrylamide. Separation of Aβ species was performed using precast 16% Tris-glycine gels (Invitrogen). Proteins were transferred to nitrocellulose (Schleicher & Scheull, Dassel, Germany) and immunodetection was performed using Alexa Fluor conjugated goat anti-mouse or IRDyeTM800 conjugated affinity-purified anti-Rabbit IgG (Rockland Immunochemicals Inc, Gilbertsville, PA, USA) in conjunction with an Odyssey Infrared Imaging System (Li-Cor Biosciences, Lincoln, NE, USA). Image analysis and quantification measurements were performed using the Odyssey Infrared Imaging System application software (Li-Cor Biosciences).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) as there was no significant reduction in cell viability in neurons from fyn–/– mice. In addition, the resistance to Aβ toxicity was not restricted to ADDL treatment since exposure of fyn–/– mouse neurons to fibrillar Aβ25–35 or fibrillar Aβ1–42 did not reduce cell viability up to 48 h treatment while exposure of wild-type mouse neurons resulted in a marked reduction in cell viability (Supplemental Fig. 1B). Again ADDLs proved to be the more potent neurotoxic form of Aβ.
As fyn is a known lipid raft protein we first looked for evidence of an association between exogenously applied Aβ with markers of lipid rafts. Hippocampal neurons from fyn–/– mice were exposed to ADDLs for 24 h, and the cells were stained with antibody 6E10 for Aβ (red) and cholera toxin β-subunit (CTXβ) conjugate to FITC (green), which binds GM1, a marker of lipid rafts/DRMs (Fig. 1B
, left panel). As expected, Aβ-positive puncta appeared to accumulate on or near neuronal processes and there appeared to be a uniformity of both size and distribution of puncta over the culture surface with a paucity of colocalization (yellow). However, identical treatment of hippocampal neurons from wild-type mice resulted in a distinctly different pattern of 6E10 labeling (Fig. 1B
, right panel). An obvious reduction was noted in the number of 6E10 labeled puncta, which varied considerably in size and included many large 6E10 positive puncta (arrows), suggestive of assemblies of smaller puncta. Also evident was an increase in the colocalization of Aβ with GM1 (arrowheads). These results suggested that Aβ interaction with the neuronal membrane was not merely a stochastic accumulation of amyloid on the cell surface but that a fyn-dependent cell membrane event gave rise to these larger assemblies. Furthermore, accumulation of Aβ on the cell surface was not sufficient to cause cell death in the absence of fyn.
Aβ colocalizes with neuronal lipid raft constituents in a time-dependent manner
To examine the underlying mechanisms more closely, and to confirm that this effect was not specific to the background strain of mice used but conserved in different rodent models, primary rat brain hippocampal neurons were treated with ADDLs for 1 min and 1 h and the cells were stained with 6E10 and CTXβ (Fig. 2
). Aβ was not uniformly distributed over the neuronal processes; only a minority were stained intensely. Furthermore, Aβ labeling was most intense on neuronal processes proximal to the cell body while distal processes were less immunoreactive. As is apparent in Fig. 2
, minimal coincident staining (yellow) with both markers was found after 1 min ADDL exposure. However, after 1 h ADDL exposure a striking increase in the amount of Aβ colocalized with GM1 (Fig. 2
, arrowheads) was found, demonstrating that exogenously applied Aβ interacts with the neuronal cell surface at sites containing lipid raft components. Furthermore, the increased colocalization of Aβ with GM1 over time suggests lipid raft components act as sites of Aβ accumulation.
|
Redistribution of Aβ on neuronal processes
Owing to the time-dependent increase in colocalization of Aβ and GM1, we investigated the dynamics of Aβ binding to neurons. Hippocampal neurons were exposed to ADDLs, fixed at various times postexposure and stained with 6E10 and CTXβ (Fig. 3
A). After 1 min exposure, Aβ was detected on neuronal processes in the form of puncta and again was not uniformly distributed on the cell surface. 6E10 labeled small puncta with a mean ± SE of 49 ± 5 puncta/100 µm2 neurite. After 1 h ADDL exposure a significant reduction in the number of 6E10 labeled puncta/100 µm2 neurite to 31 ± 7 puncta was found; however, the remaining puncta had increased in size. The number of labeled puncta continued to decrease progressively over the course of 24 h, with a concomitant increase in puncta size (Fig. 3A, B
). The increase in size of puncta was also observed with CTXβ labeling (not shown), demonstrating coalescence of lipid rafts in response to Aβ.
|
We established that the increase in puncta size was likely due to a cell-driven process and not dependent on a continuous accumulation of Aβ sequestered from the media. Thus, hippocampal neurons were exposed to ADDLs for 10 min, after which the medium was removed and replaced with conditioned medium lacking added Aβ. The cells were fixed 12 h later and stained for Aβ and GM1 ganglioside as above. A significant decrease in the number of puncta was found, to 24 ± 4 puncta/100 µm2 neurite when compared to the 1-min time point (P<0.05) (Fig. 3B
). This density is similar to that observed on neurons continuously exposed to Aβ for 6 h, implying that extracellular Aβ is not the principal driving force but that the redistribution of puncta is most likely through cell-driven processes initiated by application of exogenous Aβ.
To confirm that the reduced number and growth in size of puncta was an intrinsic property of neurons and was not due to aggregation of Aβ in tissue culture media, coverslips without neurons were incubated in culture medium in the presence of Aβ at 37°C for 24 h. Coverslips were then fixed and labeled with 6E10. A diffuse layer of 6E10 immunoreactivity and no puncta were observed. Further, ultracentrifugation of the medium also failed to sediment significant Aβ (not shown). Thus the observed puncta are not simply large aggregates of Aβ that form in tissue culture medium, which with time adhere to the cell surface, neither are they the result of a continuous recruitment of Aβ from the media. Rather they must have been produced by a cell-surface redistribution and concentration on the ADDL bound initially.
To confirm biochemically the association of Aβ with cell surface lipid rafts, primary rat hippocampal cultures were exposed to ADDLs for various times and DRMs isolated and separated by 16% Tris-glycine SDS-PAGE. Figure 3C
, lane 8 illustrates a typical profile of Aβ species present in the ADDL preparation with the major species being monomeric Aβ accompanied by significant amounts of trimers and tetramers and lower amounts of dimers. There was only a trace of larger aggregates, demonstrating that the material was a typical ADDL preparation. Hippocampal neurons incubated with this preparation for short times (1–10 min) yielded principally monomeric Aβ in the DRM fraction (Fig. 3C
), but after 1 h exposure there was a significant increase in dimeric Aβ species and at longer times both trimeric and tetrameric Aβ species were present (Fig. 3C
). Quantitation of the Western blots revealed a 4-fold increase in Aβ content of DRMs between 1 min and 1 h after exposure to ADDLs. The amount of Aβ detected within the isolated DRMS continued to increase from 1 h (283±19.8) to 12 h (646±150) post-exposure to ADDLs (Fig. 3D
). The 4-fold increase in Aβ content within DRMs after 1 h exposure paralleled the increased colocalization of accumulated Aβ and GM1 (Fig. 2)
. These observations imply that underlying fyn-dependent compositional and structural changes in DRMs in response to ADDLs are likely, since these phenomena are absent in fyn–/– neurons.
ADDL treatment of neurons induces recruitment of fyn and tau into DRMs and alterations in cytoskeletal composition
Hippocampal neurons were treated with ADDLs, and DRMs were analyzed for the presence of fyn (Fig. 4
A). A rapid but transient increase was found in the amount of fyn detected in DRMs on ADDL exposure for 2 min that was partially reduced after 10 min exposure. Mirroring the movement of fyn into the DRM, tau content of DRMs increased in response to ADDLs (Fig. 4A
). Quantitation of Western blots revealed a transient 1.5-fold (± 0.2) increase in the amount of tau present in DRMs after 1 min of treatment with ADDLs. The increase in tau peaked at 2 min (2.3±0.2) and was reduced to near basal by 10 min exposure to ADDL (Fig. 4B
). The parallel movement of fyn and tau into and out of the DRMs suggests a possible interaction between tau and fyn. To further test a putative association between fyn and tau levels in DRMs, we treated cells with 2-bromopalmitate, an inhibitor of protein palmitoylation, which is required for fyn’s localization in DRMs. This treatment resulted in a progressive reduction in both fyn and tau in DRMs, demonstrating that the association of tau with DRMs may be directly regulated by fyn (Supplemental Fig. 2).
|
The alteration in tau and fyn levels in DRMs in response to Aβ led us to search for changes in the amounts of other known DRM-associated proteins. The membrane protein flotillin-1, a well-characterized protein marker for DRMs, is involved in signal transduction, vesicle trafficking, and axonal regeneration (20)
. ADDL treatment of hippocampal neuronal cultures caused a decrease in the amount of flotillin-1 that partitioned into DRMs (Fig. 4A
). This reduction was detected as early as 1 min after ADDL exposure and continued to decrease up to 10 min ADDL treatment. To determine if the reduction in flotillin-1 within DRMs was due to lateral translocation to non-DRM regions of the plasma membrane, DRMs were prepared from hippocampal plasma membrane preparations derived from neurons treated for 5 min with ADDLs or from untreated control cultures. As expected, flotillin-1 in the plasma membrane DRMs (fractions 4 and 5) was reduced in the ADDL-treated neurons (Fig. 4C
) compared to control DRMs. A corresponding increase in flotillin-1 in non-DRM fractions (fractions 11 and 12) demonstrated that flotillin-1 migrates from the lipid raft region into nonraft regions in ADDL-treated neurons. As flotillin-1 has been reported to be directly involved in the regulation of the membrane-associated actin cytoskeleton, we examined if the amount of actin was altered in DRMs in response to ADDLs. We found reduced amounts of actin in DRMs from hippocampal neurons treated with ADDLs, similar to the trend observed with flotillin-1 (Fig. 4A
), suggesting an immediate impact on local cytoskeletal dynamics in response to Aβ.
ADDLs induce increased tyrosine phosphorylation of DRM-associated proteins in rat brain hippocampal neurons
Since fyn is a tyrosine kinase and actin reorganization has been reported on increased tyrosine phosphorylation, we probed Western blots of DRMs isolated from primary rat brain hippocampal neurons after exposure to ADDLs with the pan-phosphotyrosine monoclonal antibody, 4G10 (Fig. 5
). As expected, DRMs (fractions 4 and 5) from control neurons were rich in phosphotyrosine proteins, consistent with them having a role in signal transduction. Within 1 min of ADDL treatment, increases in 4G10 labeling of several protein bands were apparent (Fig. 5A
, bottom panel). In addition, distinct phosphotyrosine immunoreactive bands appeared (Fig. 5A
, bottom panel, asterisk) and disappeared (Fig. 5A
, top panel, asterisk) in ADDL-treated DRMs. To examine the time course of changes in tyrosine phosphorylation in hippocampal DRM proteins in response to ADDLs, neurons were treated and their DRMs isolated and concentrated, and equal amounts of their proteins were separated by SDS-PAGE. Blots were then probed for phosphotyrosine. Figure 5B
shows a rapid but transient increase in total phosphotyrosine content of DRMs. Total phosphotyrosine labeling was maximal after 1 to 2 min exposure, decreasing to near basal (untreated control) levels after 10 min ADDL exposure.
|
We then examined the basal level of tyrosine phosphorylation of DRM proteins isolated from fyn–/– cortical neurons and found that it was considerably lower than those isolated from wild-type cortical neurons (Fig. 5C
, left panel). These results suggest that fyn is a major tyrosine kinase in neuronal DRMs and that absence of fyn reduces the tyrosine phosphorylation state of multiple DRM proteins. Furthermore, exposure of fyn–/– cortical neurons to ADDLs did not result in an increase in the levels of phosphotyrosine of DRM proteins (Fig. 5C
, right panel) similar to that observed in primary rat hippocampal neurons (Fig. 5B
). In contrast, wild-type mouse cortical neurons exposed to ADDLs exhibited a similar pattern of increased tyrosine phosphorylation of membrane domain proteins with concomitant increases in tau and fyn content similar to that displayed by rat hippocampal neurons (not shown), thus further demonstrating that fyn directly recruits tau to DRMs in response to ADDL exposure.
Secondary changes in DRMs in response to ADDLs
As the above actions of ADDLs on DRMs were rapid and transient, and membrane accumulation of Aβ increased significantly after 1 h, we examined the effect of longer-term exposure of ADDLs on hippocampal neurons up to the time when cell death was first detectable. Long-term exposure of hippocampal neurons to ADDLs gave an overall increase in tyrosine phosphorylation of proteins within DRMs (Fig. 6
A). Again, an immediate increase in phosphotyrosine levels was found after 1 min exposure to ADDLs, which peaked at 6 h and decreased to below initial levels by 48 h, suggesting a biphasic response to Aβ exposure. DRM tau levels similarly increased in a biphasic manner following an initial transient rise peaking at 5 min and a secondary elevation after 6 h. In addition, of the two major tau-positive bands routinely detected in DRMs from hippocampal cultures, there was a specific increase in intensity of the slower migrating species, suggestive of increased tau phosphorylation. The very marked elevation of DRM-associated tau at 6 h treatment was still present after 24 h exposure but returned to basal levels after 48 h. In contrast, after an initial reduction at 10 min (Fig. 4)
, flotillin levels remained constant up to 48 h exposure (not shown). All of the changes in DRM composition in response to ADDLs occurred before cell death commenced (Fig. 6B
).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and actin filament remodeling has previously been reported in response to Aβ (22
, 23)
. Besides the requirement of fyn for Aβ neurotoxicity, tau is also necessary in mediating Aβ-toxicity (24
, 25)
. Tau can also induce alterations in the actin cytoskeleton, resulting in subsequent neurodegeneration; this is potentiated by Aβ (26)
. Furthermore, tau has been reported to enhance the kinase activity of fyn in tau-mediated actin rearrangement (27)
. Taken together, fyn and tau may act in concert to link external Aβ accumulation to cytoskeletal dysregulation and our observations reported here are consistent with this occurring in lipid rafts. Our results would thus position fyn upstream of tau in the neuronal response to ADDLs, and we propose that the observed accumulation of fyn and tau in lipid rafts may facilitate local actin cytoskeleton rearrangement that may underlie the observed redistribution of Aβ on the cell surface and ultimately lead to cell death.
Our findings have a number of major implications. First, we have identified a mechanism whereby Aβ accumulation in lipid raft domains is distinct from Aβ generation in lipid rafts. Second, a clear interplay exists between nonraft domains and raft domains on the plasma membrane in the cells response to Aβ. Cholesterol-lowering strategies aimed at reducing Aβ generation may ultimately bias Aβ-membrane interactions toward nonraft membrane domains with possibly detrimental effects. Cholesterol has been shown to be an important modulator of Aβ binding and neurotoxicity, having both a protective effect on the toxic actions of Aβ and reducing Aβ-membrane interactions (28
, 29)
. This protection by cholesterol was also shown with an unrelated peptide aggregate in a panel of different cell lines (30)
. We hypothesize that trafficking of Aβ on the cell surface may be a physiological response to compartmentalize Aβ into cholesterol-rich areas where its toxic effect may be muted. However, increased accumulation of Aβ in lipid rafts over time, driven by increased levels of Aβ, may have a detrimental effect on the other cellular functions of lipid rafts such as the maintenance of synapses (31)
. Last, accumulation of Aβ on the cell surface occurred in the presence or absence of fyn. However, the absence of fyn prevented redistribution of Aβ on the neuronal surface and conferred immunity from toxicity. This identifies the mechanism of redistribution as being integral to the toxic effects of Aβ, and such a mechanism may have parallels in other protein misfolding neurodegenerative diseases.
The inherent toxicity of a variety of protein aggregates is suggestive of a common mechanism underlying protein misfolding diseases (32)
and is supportive of an emerging concept in the study of protein misfolding disease, that many neurodegenerative disorders share a common mechanistic process (33)
. Nucleation-dependent polymerization on plasma membranes has been reported for a number of amyloidogenic proteins, and it has been suggested that it is the process of polymerization of Aβ, rather than the oligomeric form itself, that generates neurotoxicity (34)
. Considerable evidence now indicates that GM1 ganglioside, a prominent DRM component, can mediate such polymerization and toxicity (35
36
37)
. Given the numerous reports that GM1 and isolated brain membranes can mediate polymerization (38
, 39)
, it is tempting to speculate that this initial neurotoxic event occurs on membranes, which would be consistent with our observation. The finding that components of lipid rafts could be coisolated with amyloid fibrils extracted from tissue in a variety of human conditions (40)
hints at a common cellular origin of these deposits.
We have identified lipid rafts as providing both a cell surface platform for the accumulation of Aβ and an intracellular microenvironment affording the interaction of normally spatially separate proteins. While accumulation of Aβ on the cell surface is associated with Aβ neurotoxicity, it is not sufficient to cause neuronal death. Aβ-induced alterations in DRMs facilitated by fyn underlie both the subsequent redistribution of Aβ and cell death, and hence may be key events in the early pathogenesis of AD.
|
ACKNOWLEDGMENTS |
|---|
Received for publication October 23, 2007. Accepted for publication November 20, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
9 from at least 2 separate experiments. *P < 0.001; Mann-Whitney U test. B) Immunofluorescent labeling of hippocampal cultures exposed to 10 µM ADDLs for 24 h and labeled with antibodies to Aβ (red) and GM1 ganglioside (green); yellow represents colocalization (arrowheads). Scale bars = 20 µm.
) and ADDL-treated (10 µM; 5 min) (
) neurons, expressed as the percentage of total membrane flotillin. All blots are representative of at least three independent experiments using 3 different preparations of ADDLs.
