Interneuronal Transfer of Human Tau Between Lamprey Central Neurons in situ

Department of Biological Sciences, Center for Cellular Neuroscience and Neurodegeneration Research, University of Massachusetts Lowell, Lowell, MA, USA.
Journal of Alzheimer's disease: JAD (Impact Factor: 4.15). 01/2010; 19(2):647-64. DOI: 10.3233/JAD-2010-1273
Source: PubMed
The mechanisms by which tau-containing lesions are propagated between adjacent and synaptically interconnected parts of the brain are a potentially important but poorly understood component of human tauopathies such as Alzheimer's disease, Pick's disease, and corticobasal degeneration. Since the utility of currently available transgenic models for studying intercellular aspects of tauopathy is limited by their broad patterns of tau expression in the central nervous system, we used an in situ tauopathy model that replicates tau-induced cytodegeneration in identified neurons on a tau-negative background to determine whether tau secretion or interneuronal transfer might play a role in lesion propagation. We found that the N-terminal half of tau is required for tau secretion and is efficiently exported to the extracellular space and adjacent neurons at relatively low levels of overexpression. By contrast, full-length tau is secreted by a separate mechanism that is correlated with phosphorylation of tau at tyrosine 18 and dendritic degeneration, is exacerbated by tauopathy mutations, and blocked by mutations that inhibit tau:tau interactions. Anterograde transneuronal tau movement occurred with the expression of tau containing the P301L tauopathy mutant, but not with wild type tau isoforms. Our results are consistent with recent studies suggesting a role for molecular "templating" in the propagation of neurofibrillary lesions and provide a novel conceptual and experimental basis for studying the mechanisms of interneuronal propagation and toxicity in human neurodegenerative disease.


Available from: Garth F. Hall
Journal of Alzheimer’s Disease 19 (2010) 647–664 647
DOI 10.3233/JAD-2010-1273
IOS Press
Interneuronal Transfer of Human Tau
Between Lamprey Central Neurons
in situ
WonHee Kim
, Sangmook Lee
, Cheolwha Jung
, Ambar Ahmed
, Gloria Lee
and Garth F. Hall
Center for Cellular Neuroscience and Neurodegeneration Research, Department of Biological Sciences,
University of Massachusetts Lowell, Lowell, MA, USA
Department of Internal Medicine, Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa
City, IA, USA
Handling Associate Editor: Thomas Shea
Accepted 19 August 2009
Abstract. The mechanisms by which tau-containing lesions are propagated between adjacent and synaptically interconnected
parts of the brain are a potentially important but poorly understood component of human tauopathies such as Alzheimer’s disease,
Pick’s disease, and corticobasal degeneration. Since the utility of currently available transgenic models for studying intercellular
aspects of tauopathy is limited by their broad patterns of tau expression in the central nervous system, we used an in situ tauopathy
model that replicates tau-induced cytodegeneration in identified neurons on a tau-negative background to determine whether tau
secretion or interneuronal transfer might play a role in lesion propagation. We found that the N-terminal half of tau is required for
tau secretion and is efficiently exported to the extracellular space and adjacent neurons at relatively low levels of overexpression.
By contrast, full-length tau is secreted by a separate mechanism that is correlated with phosphorylation of tau at tyrosine 18
and dendritic degeneration, is exacerbated by tauopathy mutations, and blocked by mutations that inhibit tau:tau interactions.
Anterograde transneuronal tau movement occurred with the expression of tau containing the P301L tauopathy mutant, but not
with wild type tau isoforms. Our results are consistent with recent studies suggesting a role for molecular “templating” in the
propagation of neurofibrillary lesions and provide a novel conceptual and experimental basis for studying the mechanisms of
interneuronal propagation and toxicity in human neurodegenerative disease.
Keywords: Pseudophosphorylation, secretion, tau, tauopathy, transneuronal tau movement, tyrosine phosphorylation
Over the past few years, the mechanisms underlying
tau-mediated toxicity have become increasingly well
understood at the cellular level in experimental models
of human neurodegenerative diseases induced by ab-
normalities in tau biology (tauopathies). It is now clear
that hyperphosphorylation and aggregation of the mi-
crotubule binding region (MTBR) of the tau C-terminal
Corresponding author: Garth F. Hall, Department of Biological
Sciences, University of Massachusetts Lowell, One University Av-
enue, Lowell, MA 01854, USA. Tel.: +1 978 934 2893; Fax: +1 978
934 3044; E-mail: Garth
domain plays an important pathogenic role at the cellu-
lar level [1–3], and it has more recently become appar-
ent that abnormalities involving the N-terminal “pro-
jection” domain are also associated with neuropatholo-
gy in Alzheimer’s disease (AD) patients [4] and medi-
ate at least some neurotoxicity mechanisms in disease
models [5–7]. Unfortunately, our understanding of the
role that these pathological cellular changes actually
play in the development of tauopathies at the organ-
ismal level remains limited by our current ignorance
of whether tau-mediated neurodegenerationpropagates
through the brain via direct interneuronal interactions,
and if so, how this occurs. It seems likely that any
interneuronal toxic interactions mediated by tau will
ISSN 1387-2877/10/$27.50 2010 IOS Press and the authors. All rights reserved
Page 1
648 W. Kim et al. / Secretion and interneuronal movement of tau
prominently involve the tau N-terminal domain, which
mediates the association of tau with the plasma mem-
brane and perimembranous structures [8,9] and is phos-
phorylated by signal transduction kinases [10].
The long term focus on the C-terminal mediat-
ed functions of tau has led to the tacit assumption
that tau is a protein with exclusively cytoskeleton-
associated functions that does not interact with known
mechanisms of protein secretion and consequently that
tau-mediated toxicity mechanisms are essentially cell-
autonomous. However, this assumption has been
called into question by recent studies of other disease-
associated proteins that form the characteristic toxic
aggregates in Parkinson’s disease, AD, and the vari-
ous prion diseases (i.e., α-synuclein, amyloid-β pep-
tide and prion protein, respectively). Each of these pro-
teins plays a central role in diseases in which interneu-
ronal influences appear to be significant, and each has
recently been shown to interact with fyn kinase and be
secreted via unconventional mechanisms from neurons
in model systems [11–14]. Despite some suggestive
recent findings [15,16], it is as yet unclear whether in-
terneuronal tau transfer also plays a role in tau-mediated
neurotoxicity in tauopathies. Moreover, definitively es-
tablishing such a role using tau expression in the central
nervous system (CNS) is needed to replicate features
of human disease in such models [17,18].
In this study, we show that tau is in fact readily and
specifically secreted by plasmid-injected anterior bul-
bar cells (ABCs) to the extracellular space and to adja-
cent and postsynaptic neurons in a non-transgenic low-
er vertebrate model of tauopathy. This model consists
of specific identified neurons (ABCs) in the brain of
the ammocoete sea lamprey (Petromyzon marinus) that
have been induced to express htau via intracellular plas-
mid injection in situ and then degenerate in a stereo-
typed manner that recapitulates the cytopathology seen
in transgenic models and human tauopathies [19–22].
The limitation of htau expression to individual iden-
tified neurons in this model permits us to show un-
ambiguously that tau secretion from ABCs occurs via
mechanisms that are tau specific, mediated by the htau
N-terminal domain and are exacerbated by the presence
of exonic tauopathy mutations. Moreover, we show
that N-terminal htau fragments and full length htau are
transported across the IVth ventricle in lampreys with
ABCs that express htau at low levels, a result that calls
into question the common assumption that the presence
of tau in the cerebrospinal fluid (CSF) of early stage
tauopathy patients is a sign of large-scale antecedent
neuronal death. Our findings thus cast new light on
hitherto little studied mechanisms underlying htau ex-
port to the CSF and intercellular propagation of tauo-
pathic lesions that are likely to be of central importance
to the diagnosis and the eventual treatment of human
Plasmid Constructs
Plasmids used in this study were derived from
pRc/CMVn123c and pRcCMVn1234,as described [19,
20]. Both plasmids lack two N-terminal insert regions
and contain three or four microtubule binding repeat se-
quences respectively. GFP (green fluorescent protein)
is fused to N-terminus of T23 (no N-terminal insert and
three C-terminal repeats) and T24 (no N-terminal insert
and four C-terminal repeats). The exonic tauopathy in-
ducing point mutation P301L was introduced into T24
and the pseudophosphorylating S to D mutations [23]
into T23 constructs using site-directed mutagenesis kit
(Stratagene, LA Jolla, CA). N-terminal and C-terminal
constructs express 1–255 and 211–441 amino acids re-
spectively of T23 [21]. The GFP/T23 bicistronic con-
struct expresses GFP from the SV40 promoter and T23
from the CMV promoter separately [19]. The sequence
of all constructs was verified by direct sequence analy-
sis (see Fig. 1A).
Cell culture and transfection
Mouse neuroblastoma NB2a/d1 cells were provid-
ed by Dr. Thomas Shea (University of Massachusetts,
Lowell). Undifferentiated cells were transfected to
minimize any potential contribution of endogenous tau.
Transfection was carried out using Lipofectamine
2000 (Sigma Aldrich) according to the manufacturer’s
protocol. Serum-free medium was replaced with com-
plete medium 24 h after the transfection. Successfully
transfected cells were localized by GFP fluorescence.
Transfection rates under this condition routinely ex-
ceeded 70%, which is comparable to other published
results in this cell line [24]. Culture medium was col-
lected 24 h after medium replacement and cleared by
centrifugation at 10,000 x g at 4
C for 20 min to re-
move cells and cellular debris. To concentrate protein,
Centricon (Millipore, Billerica, MA) was used accord-
ing to the manufacturer’s protocol. NB2a/d1 cells were
lysed in Tris-NaCl (TN) buffer (50 mM Tris-HCl, pH
7.4, 150 mM NaCl, 1% Triton-X, 10% Glycerol, 2 mM
EDTA, protease inhibitor cocktail (Sigma protease in-
hibitors). Cell lysates were clarified by centrifugation
at 10,000 x g for 10 min.
Page 2
W. Kim et al. / Secretion and interneuronal movement of tau 649
Fig. 1. Plasmid microinjection into anterior bulbar cells (ABCs) in situ. A) Human tau constructs used in this study. B) Schematic diagram of
the lamprey brain in dorsal view showing the location and microinjection of anterior bulbar cell (ABC) somata in the hindbrain (inset). C) Dorsal
(top) and transverse (bottom) views of an ABC showing its normal anatomy. Planes of section through (1) the ABC soma and dendritic field and
(2) the axon are shown.
Cell lysates and medium were analyzed by West-
ern blot. After running the gels, samples were trans-
ferred to the polyvinyledene difluoride (PVDF) mem-
brane and incubated with the following primary anti-
bodies: Tau12 (1:10,000), Tau5 (1:10,000), and DM1A
(1:1,000). Horseradish peroxidase conjugated anti-
rabbit (1:10,000) and anti-mouse (1:10,000) antibod-
ies were applied as second antibodies. Western im-
ages were obtained using a chemiluminescence (Pierce,
Rockford, IL).
Surgery and plasmid microinjection into ABCs
Plasmid microinjection was performed as previous-
ly described [20,21]. Briefly, an anesthetized lamprey
hindbrain was exposed and the identified anterior bul-
bar cells were injected with the plasmid with 0.5% fast
green at a final concentration of 1mg/ml (see Fig. 1B).
A total of approximately 1100 lampreys were used for
in situ microinjections, with an expression rate of be-
tween 1–2 anterior bulbar cells per lamprey injected.
A total of 460 tau-expressing cells were identified for
use in this study.
Immunohistochemistry for brightfield microscopy
was done as previously described [20]. Briefly, lam-
prey brains expressing GFP tagged human tau express-
ing cells were identified under the fluorescence and
fixed in FAA (10% formalin, 10% glacial acetic acid,
and 80% ethanol) (see Fig. 1B). Immunohistochem-
istry was performed on 10 µm transverse sections of
paraffin-embedded lamprey heads. Sections contain-
ing somatodendritic regions or axons were stained with
following monoclonal tau antibodies. Tau5 and Tau12
(1:1000; a generous gift from L. Binder, Northwest-
ern University, Chicago, IL) were preliminary used to
identify the tau expressing cells in lamprey brain. For
immunolabeling tyrosine phosphorylation of tau at 18,
monoclonal 9G3 (1:100) was used. Polyclonal anti-
GFP (1:400; Invitrogen, Carlsbad, CA) was used to de-
tect GFP itself or tagged proteins. Appropriate species
specific VECTASTAIN ABC kits (Vector Laborato-
ries, Burlingame, CA) were used and diaminobenzidine
(Sigma Aldrich) was used as a chromagen.
For confocal imaging of multiple immunoprobes,
immunofluorescence was used. Deparaffinized and de-
hydrated sections were placed in pre-heated Tris-EDTA
pH 9.0 buffer (10mM Tris Base, 1mM EDTA Solu-
tion, 0.05% Tween 20, pH 9.0) for unmasking the anti-
gens and epitopes in brain tissue sections. The sec-
tions were then incubated with 0.2% triton in TBS for
20 min for permeabilization. For quenching autoflu-
orescence, 2 mg/ml sodium borohydride in TBS was
applied to the sections for 10 min twice. Each sec-
tion was blocked with 5% goat serum with 0.1% fish
gelatin in TBS for 1 h at room temperature. After
blocking, the sections were incubated overnight at 4
with two or three of the following primary antibod-
ies: Tau5 (1:400), Tau12 (1:400), 9G3 (1:30), and anti-
GFP (1:200). For confocal immunofluorescence imag-
ing, species-specific or mouse IgG subclass specific
secondary antibodies linked to FITC, Rhodamine Red-
X, cy5 (1:200; Jackson ImmunoResearch Laboratories,
Inc., West Grove, PA) were used. Confocal imaging
Page 3
650 W. Kim et al. / Secretion and interneuronal movement of tau
was obtained using a Fluoview 300 (FV300) Confocal
Laser Scanning Microscope (Olympus, Center Valley,
PA). Three dimensional image files were collected, cre-
ated, and analyzed with Image J (v. 1.41x) and Volocity
4.3 (Improvision, Waltham, MA).
Staging and image analysis
The stage of neurodegeneration was scored accord-
ing to previously established criteria based on the pro-
gression of dendritic beading, fragmentation, and loss
in the distal dendritic field. "High stage" or severe-
ly degenerated cells were defined as any stage greater
than 2.5 (most dendrites in pial dendritic field exhibit
beading and fragmentation) [20]. Only focal dendritic
extracellular tau with a diameter greater than 10 µm
and a circular or oval shape was measured to deter-
mine where focal dendritic extracellular tau is frequent-
ly found along degenerating dendrites. The distance
between the point of dendritic degeneration and the cen-
ter of focal dendritic extracellular tau was measured as
follows: The point of dendritic degeneration was des-
ignated as the zero point, positive values were assigned
to points distal to the point of degenerationand negative
values were assigned to points proximal to the point of
degeneration. T24 or P301L tau expressing cells at 10
to 30 days post injection were used for analyzing the
location of focal dendritic extracellular tau.
The secretion of tau protein from both mammalian
neurons in culture and lamprey neurons in situ is
tau-specific and requires the N-terminal projection
We performed experiments in two different model
systems to determine whether human tau can be ex-
ported to the extracellular space, and if so, whether this
occurs via a specific mechanism that could be part of
the normal or pathological biology of tau. First, undif-
ferentiated mouse neuroblastoma NB2a/d1 cells were
transiently transfected with plasmids encoding either
full length (T23 or T24), or a C-terminal half T23 con-
struct (encoding residues 210–441) in which GFP was
fused to the tau N-terminal. A construct encoding the
N-terminal half of T23 (residues 1–255) without the
GFP fusion was also used. GFP and T23 were also
expressed as separate proteins on a bicistronic plasmid
in some experiments (see Methods and Figs 1 and 2).
After 24 h of expression, we collected and immunoblot-
ted cell lysates and centrifugation-concentrated media
samples (Fig. 2A), using mAbs specific for 2 separate
epitopes (i.e., Tau12 (9–18) and Tau5 (210–230)) to
identify full length tau and tau fragments. We found
that NB2a/d1 cells expressing either full length T23 or
the tau N-terminal domain tau secreted Tau12 positive
protein into the culture medium while cultures express-
ing the C-terminal half of tau only did not (Fig. 2A).
This result strongly suggests that the presence of the
N-terminal half of tau (or at least the 9-18 sequence in
the N-terminal domain) is required for tau secretion.
The size range of tau12-positive bands seen suggests
that tau secretion was often accompanied by proteolytic
cleavage. It is unlikely that the presence of the GFP tag
on the C-terminal tau construct had any role in block-
ing tau secretion, since it did not block secretion of
the N-terminal GFP tagged T23. Finally, the complete
absence of immunolabel anti-alpha tubulin (DM1A) in
the culture medium (Fig. 2A), indicates that secretion
is not due to passive spillage of cytosolic proteins from
dead cells into the culture medium, and is consistent
with a specific mechanism for tau secretion in NB2a/d1
In order to determine if tau secretion can also occur in
an in situ tauopathy model, we microinjected plasmids
encoding T23 and T24 with and without GFP fusions
at the N terminus into ABCs, a set of giant, identified
neurons in the lamprey hindbrain (see Fig. 1 and meth-
ods). The lamprey system was chosen for this because
it is entirely negative for tau and MAPs that cross-react
with most tau specific mAbs, permitting us to conclu-
sively determine the origin of any exogenously derived
tau. Plasmids encoding NF-180 (lamprey neurofila-
ment protein) with a GFP fusion and GFP alone were
used to control for tau specificity and overexpression
toxicity. We found that expression of full length tau
isoforms (T23 and T24) resulted in the formation of
Tau12-positive deposits in the extracellular space near
the somata and dendrites of expressing ABCs, while ex-
pression of NF-180 or GFP alone did not (Fig. 2D) [20].
In order to determine whether protein export to the ex-
tracellular space is in fact a specific property of tau
or is the consequence of tau-induced neurotoxicity, we
expressed T23 and GFP separately in the same lamprey
neurons using a bicistronic plasmid. Adjacent sections
immunolabeled with Tau12 or anti-GFP showed that
untagged tau was exported to the extracellular space
in a manner similar to that seen with monocistronic
expression, while GFP was not (Fig. 2C–D). Specific
secretion of tau12-immunopositive protein could occur
Page 4
W. Kim et al. / Secretion and interneuronal movement of tau 651
Fig. 2. Secretion of human tau to the extracellular space is tau specific and can occur in the absence of tau-mediated toxicity. A) Undifferentiated
mouse neuroblastoma NB2a/d1 cells were transfected with plasmids encoding either N-terminus, GFP fusion C-terminus, GFP+T23 (bicistronic),
GFP fusion T23, or GFP alone. Cell lysates and media concentrates were immunoblotted with mAbs Tau12 (to identify N-terminal and full length
tau) and with Tau5 (for C-terminal tau). Much of the tau secreted from cells expressing full length T23 (both bicistronic and GFP fused) consists
of what appears to be a common cleavage fragment (arrow). The bottom blot shows clearly that no α-tubulin (mAb DM1A) has been released
into the medium, suggesting that passive release of cytoskeletal proteins from degenerating or dead cells did not occur. B) ABCs expressing
the N-terminal half (1–255) of T23 tau show dendritic gross morphology (left, center) and somatodendritic α-tubulin content (right). Secreted
tau diffuses to a broad portion of the CNS axis in the hindbrain and is efficiently taken up by adjacent neuronal somata (left, s, center, double
asterisk) even at relatively low levels of tau in the expressing cell (left, asterisk). Note also that the dendrites of this cell (left, arrows) contain
relatively little tau and appear healthy and unbeaded. This is similarly true at higher levels of tau expression (center image) where sufficient tau is
expressed to accumulate in distal dendrites. Even in this case, the microtubule network in the soma (right, asterisk) and dendrites (right, arrows)
appear to be intact relative to the tubulin immunolabel present in adjacent somata (right, double asterisk). C) A total of 150 lamprey neurons
expressing NF-180, tau, and GFP were immunostained, densitometrically analyzed (Scion Image) and quantitatively assessed on a per cell basis
(chi square test) to demonstrate that tau secretion from ABCs is specific to tau. Expression of NF-180 (lamprey neurofilament) or GFP alone in
ABCs fail to cause extracellular immunolabeling for these proteins, even when GFP is expressed in the same cell with tau as a separate protein
from a bicistronic plasmid. These results were highly significant (p < 0.01, chi square test, asterisks). D) Representative sections from the
analysis shown in C. ABCs expressing full-length tau (center left and rightmost images) show tau that has been secreted to the extracellular space
as either a diffuse deposit in the vicinity of the soma (asterisks) and/or focal peridendritic deposits (arrows), while NF-180 expressing ABCs does
not exhibit extracellular immunolabel (left). The center panels of D show adjacent sections immunolabeled with Tau12 and anti-GFP through the
soma and dendrites of an ABC expressing GFP and T23 separately from a single plasmid. Note that T23, but not GFP, is selectively exported to
the extracellular space. Scale Bars: B 100 µm, D 50 µm.
at relatively low levels of tau overexpression and in the
absence of overt toxicity to the expressing cell, even
at higher levels of expression (Fig. 2B). These results
from ABCs are thus consistent with the results from
NB2a/d1 cells and confirmed that tau secretion is both
specific to tau and that it can occur under physiologi-
cally relevant conditions.
Diffuse and focal extracellular tau deposits were
distinguished from one another by their appearance
and site of secretion from tau-expressing ABCs
We defined two types of extracellular tau deposits
generated by tau expression in ABCs that appeared to
be produced by distinct protein secretion mechanisms
based 1) on their distribution and appearance in sec-
tions and 2) on the way that tau deletions and modi-
fications affected their formation. We defined extra-
cellular tau deposits that were 1) homogenous and 2)
distributed in a perisomatic pattern with respect to the
secreting ABC as “diffuse”. Tau immunolabel in dif-
fuse deposits typically became less concentrated with
increasing distance from the soma of the expressing
ABC (Fig. 3). They were occurred with both low and
high expression levels and in association with ABCs
in any stage of degeneration. In contrast, tau deposits
that were localized to specific sites near the dendrites
or axons of tau-expressing ABCs were defined as “fo-
Page 5
652 W. Kim et al. / Secretion and interneuronal movement of tau
Fig. 3. ABCs expressing full length WT and mutant htau isoforms secrete tau in two distinct patterns. All sections shown were immunolabeled
with Tau12. A) By 10 days post plasmid injection, “diffuse” extracellular tau deposits appear as gradients of immunolabel (arrows, bottom
image) extending away from the somata of ABCs (asterisks) expressing T24 (top) and P301L (bottom). Diffuse secreted tau readily crosses
into the somata of adjacent and even contralateral non-expressing neurons (caret). Note that while the top image shows only diffuse label, the
bottom image in A also shows peridendritic foci (f) that clearly originate from sites in the dendrites (bottom, dotted arrows). as well which are
quite distinct from the perisomatic label (d). B) Diffuse and focal extracellular tau deposits at 20 days post plasmid injection closely resemble
those seen at 10 days. As in A, diffuse deposits can be seen in a perisomatic gradient (top, arrows). Focal deposits are localized to discrete sites
within the dendrites that form the foci of generation for extracellular tau (top, carets). Note that these foci tend to occur near the point of onset
of dendritic degeneration (bottom, dotted line). C) Top: High magnification view of a peridendritic focal deposit (asterisk). Note the granular
punctate label especially near the center of the deposit. Bottom: Punctate label can often be seen very near to beaded dendrites (arrows) distal to
the point of degeneration onset, and sometimes also from dendritic branch points (caret). Note that most of this label is located near the beads
rather than the interbead regions. Scale bars: A (top) 100 µm, A (bottom), B 50 µm, C (top) 20 µm and C (bottom) 10 µm.
cal” (Fig. 3). Unlike diffuse deposits, focal deposits
were clearly associated in some way with the extent of
tau induced degeneration in the ABCs that gave rise to
them [3,22].
Secretion of tau to diffuse and focal deposits occurs by
distinct mechanisms differentiated by the presence of a
functional C-terminal MTBR domain
In order to identify specific functional elements in
tau responsible for producing the distinctive appear-
ance and distribution of diffuse and focal tau deposits,
we compared the effects of expressing the N (1–255)
and C (211–441) terminal halves of T23 in ABCs at
10 days post injection on the pattern of secreted tau.
We found that ABCs expressing the N-terminal tau
construct secreted extracellular tau profusely, but in an
exclusively diffuse pattern, whereas tau secretion was
entirely absent with the expression of the C-terminal
construct (Fig. 4). By 10 days post plasmid injec-
tion, a significantly higher proportion of N-terminus-
expressing ABCs exhibited extracellular tau immuno-
label than was true of ABCs expressing full length con-
structs (Fig. 4), and that this secretion was already iden-
tifiable at 5 days post injection, even though htau was
expressed at a low level at that time. Unlike full length
T23, expression of the tau N terminal domain was nev-
er seen to induce obvious cytotoxicity (Fig. 2B). These
results are similar to the results we obtained in NB2a/d1
cells and together they clearly indicate that tau secretion
is not an overexpression artifact or a consequence of
tau-induced neurotoxicity (Figs 2A–B and 4A). They
also suggested that a tau-specific mechanism for the
efficient secretion of tau resides in the N-terminal half
of tau, while the C-terminal domain may in fact inhibit
secretion. By contrast, the C-terminal half of T23 ap-
pears to be incapable of supporting secretion by itself,
but its removal prevents events necessary for focal se-
cretion in particular. Next, we examined whether the
ability of tau to bind to either microtubules or to other
tau molecules via its C-terminal MTBR domain plays
a role in either the inhibition of secretion overall or the
Page 6
W. Kim et al. / Secretion and interneuronal movement of tau 653
Fig. 4. Focal but not diffuse tau secretion requires the presence of a functional C-terminal domain. A) ABCs microinjected with plasmid
encoding mutant tau (P301L), pseudophosphorylated tau (S262D/S356D), N-terminus, and C-terminus were immunolabeled with Tau12 or Tau5.
While S262D/S356D tau and N-terminal tau (at 10 days and 5 days post injection respectively) show only diffuse tau (asterisks) that has moved
into adjacent neuronal somata (carets), P301L mutant tau-expressing cells exhibit focal extracellular tau deposits (arrows). Expression of the
C-terminus construct (residues 211–441) does not result in either focal or diffuse secretion. B) A total of 374 ABCs expressing tau constructs
were examined and scored on a per cell basis for the presence of diffuse or focal tau deposits at 10 days post injection. Removal of the C-terminal
half of tau increases diffuse secretion and completely abolishes focal secretion (double asterisks p < 0.05, chi square test), while removal of
the N-terminal half of tau abolishes all tau secretion (p < 0.01). More than 40 percent of cells expressing full-length tau (T23 and T24) and
tauopathy mutant tau produce both focal and diffuse extracellular tau deposits. Pseudophosphorylated tau (S262D/S356D) significantly reduces
focal extracellular deposits (single asterisk, p < 0.05). Scale Bar: 50 µm.
specific prevention of focal deposit development. To do
this, we expressed T23 that encodes a tau pseudophos-
phorylated [23] at residues 262 and 356 within the MT-
BR (S262D/S356D), and was thus incapable of bind-
ing to either microtubules or other tau molecules [25].
Phosphorylation of these sites has also been shown to
be necessary for axonogenesis [26], suggesting that this
manipulation would preserve the centrifugal tau trans-
port of tau and thus be a fairly specific test of the role
played by MTBR-mediated tau interactions with tau
and possibly tau:heparan sulfate proteoglycans [25] in
tau secretion. We found that the S262D/S356D mu-
tations caused a marginal increase in the incidence of
diffuse deposits while abolishing focal deposits entire-
ly (Fig. 4), effects that were very similar to those of
removing the entire C-terminal domain. This suggests
that the C-terminal domain is not needed for the pro-
duction of diffuse deposits, and may in fact serve to
hinder their formation, while focal deposit formation
requires tau to have a functional MTBR.
Diffuse and focal secretion are respectively associated
with transneuronal tau movement into adjacent and
postsynaptic neurons
In addition to finding two functionally distinct modes
of tau secretion from ABCs, we distinguished two types
of secretion-associated movement of tau into other neu-
rons (i.e., transneuronal movement) in this study. Tau
that was secreted in the diffuse pattern was often tak-
en up by neurons with somata adjacent to expressing
ABCs and tau that was transported centrifugally along
axons (but not dendrites) was sometimes transferred to
neurons that were postsynaptic to tau-expressing ABCs
in what appeared to be a transsynaptic pattern associat-
ed (but not always coincident) with periaxonal focal tau
secretion (Fig. 5C). While transneuronal movement in-
to adjacent neurons occurred relatively frequently when
tau was secreted in the diffuse pattern, anterograde
transneuronal movement was uncommon, and was on-
ly observed in ABCs expressing tauopathy mutant tau
“Diffuse” tau movement into adjacent neurons ap-
peared to be a byproduct of the diffusion of tau label
down a concentration gradient where neuronal somata
happened to be present. It was seen with all of the con-
structs whose expression resulted in diffuse tau secre-
tion, and appeared to be favored by any circumstances
that favored diffuse extracellular tau formation, such as
the inactivation (i.e., by the S262D/S356D mutation)
or absence of the tau C-terminal domain (as with N-
terminal tau). Diffuse tau deposits were never associ-
ated with signs of toxicity in adjacent neurons. In most
cases, diffusely secreted tau moved into and across the
cytoplasm of nearby neurons, glia, and even the IVth
ventricle without interruption of the gradient (Fig. 5B).
Tau was not significantly concentrated in the cytoplasm
of neurons and glia above the levels present in the ad-
jacent extracellular space, and was usually excluded
Page 7
654 W. Kim et al. / Secretion and interneuronal movement of tau
Fig. 5. Transneuronal tau movement to adjacent and postsynaptic neurons are both favored by the P301L tauopathy mutation but appear to occur
via distinct mechanisms. A) ABCs expressing T24 (left column) and P301L (right column) at 10 days post plasmid injection (top photos) produce
gradients of diffuse extracellular tau, which can cross the midline (dotted line). Transventricular movement of diffuse tau was restricted to P301L
tau at 20 days or more ppi (bottom left). B) A total of 310 microinjected lamprey neurons expressing T24 or P301L tau post injection were scored
for the presence of diffuse Tau12 label contralateral to expressing cells (i.e., transventricular tau) and/or focal peridendritic or periaxonal deposits.
At 20 days and later ppi, a significantly higher percentage of P301L expressing ABCs showed more transventricular and periaxonal extracellular
tau than at 10 days (asterisks, p < 0.05 chi square test), while T24 expressing ABCs show little transventricular and periaxonal extracellular
tau. C) Left: Axons of most tau-expressing ABCs showed normal cross sectional morphology Axons (a) of most when immunolabeled with
Tau12, or tau5. The section shown was immunolabeled with Tau12. Center left: Periaxonal tau surrounding an axonal profile (asterisk) in the
caudal hindbrain is distributed radially around the axon (arrows) and is excluded from a nearby axonal profile (a) and dendrites of a nearby
neuron (caret). Center Right and Right: Brightfield/DAB image (center right) and confocal 3D rendering of a fluorescent image (right) of two
examples of anterograde transneuronal movement from axons (marked by asterisks) of ABCs expressing P301L tauopathy mutant T24 into what
appear to be postsynaptic neurons (arrows) with relatively little periaxonal extracellular label, suggesting a transsynaptic mechanism. The pattern
of transcellular movement appears distinct from that seen with the movement of perisomatic diffusely secreted tau. Note that in the brightfield
image, tau appears in perinuclear vesicular bodies in what appear to be 2 different neurons, suggesting an endocytotic transfer mechanism. The
cleft-like space between the ABC axon and the adjacent cell in the confocal reconstruction (caret) is consistent with this possibility, and appears
to be the typical “en passant” synaptic junction made by these neurons in the caudal hindbrain [6]. n = nucleus. Scale bars: 20 µm (A); 100 µm
only from the nuclei of neurons and glia (Figs 4A and
5B). Our observations suggest overall that diffuse ex-
tracellular tau moves freely but passively across plas-
ma membranes, possibly with the aid of an unknown
tau-specific facilitated transport mechanism.
By contrast, “anterograde” transneuronal movement
of tau into neurons that are postsynaptic to ABCs [26]
was only seen in a minority of cases (8 of 48) in which
focally secreted periaxonal deposits were identified,
and then only when the P301L tauopathy mutant was
present (Fig. 5A). As was the case with tau movement
into adjacent cells, anterograde tau movement was well
correlated with the incidence of periaxonal tau secre-
tion, which was also heavily favored by the presence
of the P301L mutation (Fig. 5B). However, unlike tau
movement into adjacent neurons, anterogradely trans-
ferred tau moved vectorially from the tau expressing
axon to postsynaptic neurons, often with very little
concomitant periaxonal secretion (Fig. 5C). In addi-
tion, anterogradely transferred tau was often not even-
Page 8
W. Kim et al. / Secretion and interneuronal movement of tau 655
Fig. 6. Extracellular tau deposits are favored by the presence of the P301L tauopathy mutation and are spatiotemporally correlated with dendritic
degeneration. A) The incidence of extracellular tau was highly correlated with high stages (stage 2.5+) of degeneration in cell expressing both
T24 and P301L. B) ABCs expressing wild type and P301L mutant T24 were scored for the number and of focal peridendritic deposits present and
their radial location relative to the most proximally located beaded dendrite at 10 to 30 days post injection. The presence of diffuse tau was scored
as well on a per cell basis. We found that the number of ABCs with focal (but not diffuse) tau deposits was increased by the P301L mutation tau at
10 days post injection (p < 0.05, chi square test, single asterisk). By 20 days post injection and later, the incidence of both focal and total (focal
+ diffuse) deposits was significantly increased in P301L–expressing ABCs (p < 0.001, chi square test, double asterisks, p < 0.05, chi square
test, single asterisk above bar respectively). Overall, the P301L mutation selectively increased the incidence of focal relative to diffuse deposits
over time. C) and D) Locations of focal extracellular tau deposits were measured relative to the radial distance from the point of degeneration
(dotted line) along a proximal-distal axis, with the distance in microns proximal to the point of degeneration being negative () and distance
distal positive (+). The distribution of focal deposits seen in T24 and P301L expressing cells from the point of degeneration was quantified to
investigate the relationship between the focal extracellular tau and dendritic neurodegeneration. More than 50% of focal peridendritic deposits
were centered within 15 µm of the point of degeneration (POD). A total of 310 ABCs expressing T24 or P301L mutant tau were examined and
analyzed for both sets of experiments shown. Scale bar: 50 µm.
ly distributed in the postsynaptic neuron, localizing to
what appeared to be perinuclear vesicular organelles
(Fig. 5C), suggesting that the mechanisms responsible
for the transfer of diffuse tau into adjacent neurons and
anterograde tau movement are distinct from one anoth-
er. This difference is further emphasized by our com-
plete failure to identify cases of retrograde tau transfer
from the dendrites of tau-expressing ABCs into neurons
presynaptic to ABCs (not shown).
The extent of diffuse deposits and the persistence of
focal deposits over time were both increased by the
presence of the P301L tauopathy mutation
As a first step toward clarifying the role played by
the P301L tauopathy mutation in both tau secretion and
Page 9
656 W. Kim et al. / Secretion and interneuronal movement of tau
tau-mediated toxicity, we asked whether 1) the extent
of diffuse deposits across the midline and/or 2) the in-
cidence and distribution of focal extracellular tau de-
posits are correlated with the presence of the P301L
tauopathy mutation, which has been shown to favor tau
aggregation [27] and accelerates tau-induced neurode-
generation in lamprey neurons [20]. ABCs expressing
T24 and P301L mutant tau at 10, 20, and 30 days post
injection were analyzed to see whether the number of
cells in which diffuse tau deposits cross the midline
(i.e., transventricular tau, Fig. 5B), and the incidence
of periaxonal extracellular tau deposits was correlated
with the presence of the P301L tauopathy mutation. At
10 days post plasmid injection, the number of cells pro-
ducing transventricular and periaxonal extracellular tau
was not significantly different between T24 and P301L
mutant tau-expressing cells. However, the incidence
of transventricular and periaxonal tau deposits was sig-
nificantly increased over time by the presence of the
P301L mutation (p < 0.001, chi square test, Fig. 5C).
Similarly, focal peridendritic extracellular deposits in
P301L-expressing ABCs were significantly increased
over the number seen in T24-expressing ABCs at 20
and 30 days post injection, but not at 10 days (Fig. 6).
We also found that the incidence of focal peridendrit-
ic deposits tau was correlated with high stages (2.5 or
greater) of neurodegeneration (p < 0.001, chi square
test, Fig. 6A). We quantified the distribution of focal
deposits (10 µm across or larger) in T24 and P301L-
expressing ABCs with respect to the point of onset of
dendritic degeneration (Fig. 6D) and found that more
than 50 percent of focal deposits were localized to with-
in 15 µm of the point of degeneration, irrespective of
the actual stage of degeneration (Fig. 6C-D).
In an effort to determine whether the greater inci-
dence of focal deposits at late timepoints seen with
P301L tau expression was correlated with progressive
changes in their appearance and immunostaining pat-
terns, we used the fact that the degeneration induced by
tau expression in ABCs is progresses stereotypically
from distal to proximal dendrites [20], permitting us
to assume that more distally located focal extracellu-
lar tau deposits (inset 2 in Fig. 7B) are secreted earli-
er than more proximal focal extracellular tau deposits
(inset 1 in Fig. 7B). When dendritic sections through
ABCs expressing P301L mutant tau were double im-
munostained with Tau5 and anti-GFP, we found that
more distally located focal extracellular tau deposits
were more granular in appearance, more Tau5 positive
and much less anti-GFP positive than were more prox-
imal ones in the same section. These changes are con-
sistent with a progressive degradation of focal extra-
cellular deposits over time and suggest that the greater
resistance to proteolysis of aggregated P301L tau [1,2]
might be responsible for its increased persistence over
time in focal deposits.
Both dendritic degeneration and the site of focal tau
secretion are spatiotemporally correlated with the
distribution of 9G3-positive tau, while diffuse
secretion is not
Since the presence of the tau N terminus is required
for both diffuse and focal secretion, we investigated
the possibility that these modes of tau secretion might
be distinguished from one another by differential phos-
phorylation at tyrosines within the N-terminal known
to be phosphorylated by fyn kinase [8]. We used a mAb
specific for the fyn phosphorylation site at tyr18 of tau
(9G3) and examined the pattern of 9G3 immunolabel
produced by expression of tau that 1) could only pro-
duce diffuse tau secretion and could not mediate tau
transport (i.e., the N-terminal (1–255) tau domain), and
2) could only produce diffuse tau secretion but which
can undergo dendritic transport (S262D/S356D), and
full length wild type or P301L tau, which can be se-
creted in focal as well as diffuse patterns and which
can be transported. We found that secreted N-terminal
tau was completely 9G3 negative, while dendritically
transported wild type or P301L tau was preferential-
ly transported to the dendrites, where it accumulated
near the point of degeneration onset (Fig. 7, panels A
and B, respectively). S262D/S356D tau was efficient-
ly and selectively labeled with 9G3 only when it was
transported distally (Fig. 7A). In all cases, secreted tau
was 9G3 negative, but tau located at the immediate site
of origin for focal deposits (which interestingly was
usually at or near the point of degeneration onset) was
strongly 9G3 positive, suggesting that the secretion of
focal tau deposits might be mechanistically linked with
both the phosphorylation of tau at tyr18 and with events
immediately associated with dendritic degeneration.
Immunolabel for the GFP tag is selectively absent
from diffuse but not focal tau deposits secreted by
ABCs expressing full length tau
Finally, we examined the possibility that diffuse-
ly secreted tau might be generated from the cleavage
of N-terminal fragments from full length tau species,
since the characteristic functional differences between
diffuse and focal deposits described above are consis-
Page 10
W. Kim et al. / Secretion and interneuronal movement of tau 657
Fig. 7. Phosphorylation at tyrosine18 is associated with transport of tau along ABC dendrites, dendritic degeneration and tau secretion to
peridendritic focal, but not diffuse extracellular deposits. A) Adjacent sections through ABCs expressing N-terminal fragment (1–255, left) and
pseudophosphorylated (S262D/S356D, right). While both S262D/S356D and N-terminal tau contain the tyr18 site and ABCs expressing these
proteins produce only diffuse tau deposits, only S262D/S356D tau is immunopositive for 9G3 (phosphorylated tyr18 - rightmost panel in each
pair). S262D/S356D tau is also much more efficiently transported to distal dendrites than is N-terminal htau (left panels, compare single to
double asterisks). B) Left: Volocity image of a section through a P301L mutant tau-expressing cells immunolabeled with 9G3 (green) and a
polyclonal antiserum against the GFP tag (red). 9G3 immunoreactivity in the dendrites is maximal at the point of onset of dendritic degeneration
(shown by carets). Right: Image of a triple labeled (T12-green, 9G3-blue, and GFP-red) dendrite from a P301L mutant tau-expressing cell at
the point of degeneration was viewed at high opacity so as to reveal only tau near the plasma membrane of the degenerating dendrite. Note that
focal extracellular tau is generated primarily from the first “bead” (asterisk) after the point of degeneration (arrow), not from more distal beads
(double asterisk). The tau phosphorylated at tyr18 (9G3) is localized to the surface of the dendrite near where focal htau is being secreted. C)
shows a dendrite from a P301L-expressing ABC with distally (D inset 2) and proximally (P inset 1). The distal (and presumably older)
deposit shown consists mostly of Tau5-positive aggregates (carets) and appears more aggregated than the more proximal deposit. There is also a
selective loss of the diffuse, GFP-positive immunolabel seen in more proximally located deposits that suggests ongoing conformational changes
or truncation of focally secreted tau with time. D) Focal tau deposits differ significantly from diffuse deposits in their immunolabel patterns.
Diffuse extracellular tau deposits are negative for anti-GFP, while focal peridendritic deposits (f) are strongly positive for anti-GFP, suggesting
that they differ in proteolytic cleavage or folding patterns. Scale Bars: 50 µm (A and B left); 20 µm (B right, C left, D); 5 µm (C right).
Page 11
658 W. Kim et al. / Secretion and interneuronal movement of tau
tent with the possibility that the diffuse secretion pat-
tern results from the partial or complete removal of
the C terminal domain. We compared the overall im-
munolabel intensity of diffusely and focally secreted
tau directly using multi-labeling with Tau12, an mAb
specific for the extreme N-terminal (residues 9–18),
and Tau5, which is specific for the central region of
the tau molecule (residues 215–235) and the GFP tag
polyclonal via indirect immunofluorescence. We found
that diffuse extracellular tau deposits were always im-
munolabeled with Tau12, and usually not with anti-
GFP (Fig. 7D), whereas focal deposits labeled strongly
with both Tau12 and the GFP polyclonal. Tau5 label
was typically much weaker or absent than Tau12 on
diffuse deposits (not shown) and was present on more
condensed focal deposits (described above). The ab-
sence of GFP and in many cases Tau5 from diffuse
deposits secreted by ABCs is quite consistent with the
characteristics of secreted tau produced by full length
tau expression in NB2a/d1 cells (Fig. 2A), and both
are consistent with (but do not demonstrate) proteolytic
cleavage prior to secretion.
In this study, we identified and characterized two
distinct mechanisms of tau secretion to the extracel-
lular space and to adjacent and postsynaptic neurons
in the lamprey ABC tauopathy model. Both secretion
mechanisms are tau-specific and require the presence
of the tau N-terminal domain, but only focal secretion
is clearly correlated with the phosphorylation of tau at
tyr18. We also found that the secretion of diffuse htau
from ABCs is enhanced (and focal deposits abolished)
when the MTBR region in the C-terminal domain of tau
is either removed or inactivated by pseudophosphoryla-
tion at residues 262 and 356. Diffuse tau secretion can
occur in the absence of high expression levels and/or
the presence of degenerative changes and can occur
from cultured mammalian (NB2a/d1) cells as well as
from ABCs. By contrast, the accumulation of focal tau
deposits near ABC dendrites and axons is highly cor-
related with high tau expression, high 9G3 immunola-
bel and dendritic degeneration near the site of secretion
and is enhanced by the presence of the P301L tauopa-
thy mutation and abolished by a mutation that blocks
the formation of tau:tau aggregates [28]. Thus: a) the
prior phosphorylation of tau at tyr18; b) the presence
and functionality of the MTBR domain; and c) the ten-
dency of tau species to form tau:tau aggregates appear
to be key factors determining the type of extracellu-
lar deposits produced and the patterns of transneuronal
movement seen with htau overexpression. These data
are summarized in Table 1.
Tau secretion from ABCs in situ and from NB2a/d1
cells in culture is physiological and likely to be
relevant to normal tau biology and tau pathobiology
in humans
Before considering possible specific roles that se-
creted tau might play in the pathogenesis of human
tauopathies, it is important to establish clearly that the
findings of this study are relevant to human tau biol-
ogy in general, since tau is known as a microtubule-
associated protein, has as yet no known function as a
secreted protein, and since tau secretion per se has not
been shown to occur in neurodegenerative tauopathies.
Moreover, much of the secretion data that we obtained
(at least in ABCs) was from tau-overexpressing cells
that exhibited clear signs of degeneration (i.e., progres-
sive dendritic beading and fragmentation), which raises
the theoretical possibility that some or all of the “se-
cretion” observed may have been either artifactual or
the consequence of a non-specific release of tau due
to toxicity caused by the overexpression of exogenous
protein [29]. Our contention that tau secretion from
ABCs and NB2a/d1 cells reflects a hitherto undescribed
aspect of tau biology that is relevant to the neuropatho-
genesis of human disease is primarily based on the ob-
servation that secretion of the tau N-terminal domain
occurs at relatively low levels of intracellular tau accu-
mulation and in the absence of toxic cellular changes
(Fig. 2A–B). Secretion of the tau N-terminal domain
is both specific and efficient enough to prevent its den-
dritic accumulation in all but the highest expressing
ABCs. Since dendritic microtubules of ABCs are in-
tact and their morphology normal and unbeaded, even
in relatively high-expressing cells, and since tubulin is
retained within NB2a/d1 cells expressing and secreting
the tau N-terminal domain (Fig. 2A–B), this clearly in-
dicates that tau secretion is independent of tau toxicity.
Toxicity mediated by either the tau N-terminal domain
or by full length isoforms has been shown to result in
dendritic microtubule loss and degeneration in a variety
of tauopathy models [7,16] and such changes are corre-
lated with tau lesions in human disease [30]. Moreover,
since both diffuse and focal tau secretion are abolished
by the removal of the tau N terminus (Fig. 2), it seems
likely that even focal secretion at the point of degen-
eration in the dendrites is not merely a passive conse-
Page 12
W. Kim et al. / Secretion and interneuronal movement of tau 659
Table 1
Summary of the characteristics that distinguish extracellular and transneuronal tau species in the lamprey tauopathy model
at 262 and 356
at tyrosine 18
(9G3+ tau)
Other Immunolabel pattern
(1–255) (211–441) Tau12 Tau5 Anti-GFP
Required Inhibited No clear
Increased Not correlated Occurs at
low tau
+ + + + +
Required Required Greatly
Inhibited Correlated Reduced
by C-
++ ++
Required Inhibited Increased Increased Not correlated ++ + +
Not tested Not tested Required Not observed Not tested ++ ++ ++
Characteristic differences of diffuse and focal extracellular deposits and the two types of transneuronal movement associated with them (adjacent
and anterograde, respectively) are summarized. The effects of the presence of the N- and C-terminal domains of tau and the P301L tauopathy
mutation. Since anterograde transneuronal movement was observed in lamprey neurons expressing P301L mutant tau, but not in other full
length tau isoforms, it is shown as requiring P301L; truncated and pseudophosphorylated tau species were not examined for this study. The
effect of pseudophosphorylating residues 262 and 356 in the tau MTBR (S262D/S356D), which abolishes tau:tau, tau:tubulin, and tau:polyanion
interactions via the C-terminal MTBR domain required for tau aggregation, microtubule binding, and tau binding to heparan sulfate proteoglycans
is also shown. The degree of spatiotemporal correlation of each secretion pattern with tyrosine phosphorylation of tau at residue 18 (i.e.,
immunolabeling with the 9G3 mAb) is also shown. The characteristic immunolabel pattern of secreted and transneuronally transported tau with
the Tau12 and Tau 5 mAbs and the anti-GFP polyclonal is shown on a semiquantitative and comparative basis (i.e., . +, ++, + + +).
quence of dendritic integrity loss, but requires specific
events mediated by the tau N-terminal domain in order
to occur, although the nature of these events is not yet
Both diffuse perisomatic and focal peridendritic or
periaxonal tau secretion from overexpressing ABCs
can be readily distinguished from the non-specific re-
lease of misprocessed exogenous protein due to over-
expression which can occur via the “unfolded protein
response” or UPR. This mechanism is potentially rele-
vant to tau secretion since the UPR can result in the se-
cretion of exogenous proteins such as GFP [29]. How-
ever, we have already shown that the accumulation of
“misprocessed” proteins due to overexpression in some
high expressing ABCs is itself not associated with neu-
rodegeneration and is readily identified as such by its
stereotypic localization to the dorsal surface of the so-
ma and its lack of specificity to tau [20]. While it
is unclear whether such aggregates are ever secreted
from ABCs, all localized tau accumulations from the
region of the dorsal soma were excluded from consid-
eration in this study. Furthermore, since the dorsal so-
ma is anatomically entirely separate from the extracel-
lular and transneuronal deposits that were observed, it
is hard to see how artifactual secretion from this area
due to overexpression could account in any way for our
Based on the above considerations, we conclude that
tau secretion probably plays some role in natural tau
biology under at least some circumstances and that it
is worth asking whether it is relevant to the pathogene-
sis of human diseases involving tau, especially in light
of the transneuronal links between sequentially affect-
ed neurons in such diseases. While tau secretion as
such has not yet been shown to play a role in human
tauopathies, it is likely that this is at least partly due
to the inherent difficulty of demonstrating secretion
as such in human disease and in transgenic tauopathy
models due to the presence of background tau levels
throughout the CNS. There is also a natural tendency
to assume (in the absence of direct evidence) that 1)
tau has no functions other than its known cytoskeleton-
related ones, and 2) therefore that any tau observed in
the extracellular space or CSF humans suffering from
AD or other tauopathies must be due to passive tau
release from dying neurons. The results of this and
other recent studies suggesting a role for interneuronal
tau transfer in tauopathy pathogenesis that were con-
ducted in cell culture [16] and murine transgenic [15]
model systems suggest that these assumptions should
be re-examined.
Diffuse and focal tau secretion appear to occur via
separate but related mechanisms While there have been
prior reports of both focal [21] and diffuse [31] tau
Page 13
660 W. Kim et al. / Secretion and interneuronal movement of tau
secretion in the lamprey tauopathy model, the cur-
rent study is the first characterization of tau secretion
that is sufficiently detailed to permit the considera-
tion of specific cellular mechanisms by which such
secretion might occur. Since tau is a cytoplasmic
protein and thus lacks classic “leader” and lipidation
consensus sequences that would target it to the clas-
sical brefeldinA/tunicamycin-sensitive secretion path-
way, its secretion presumably occurs via one of sever-
al alternative secretion routes that have recently been
shown to be responsible for the export of a wide variety
of proteins under diverse circumstances [32]. Our re-
sults show that both diffuse and focal secretion mech-
anisms require the presence of an unknown feature of
the tau N-terminal “projection” domain and that the ex-
tracellular accumulation of peridendritic focal deposits
requires in addition that the C-terminal domain (which
contains the microtubule binding region of tau) be 1)
present and 2) capable of participating in one of several
possible interactions with targets in addition to micro-
tubules, including other tau molecules [33] or any of
a variety of biological polyanions that include heparan
sulfate proteoglycans [34,35] that are blocked by phos-
phorylation or pseudophosphorylation of residues 262
and 356 [25,28, 36]. We also showed that the phos-
phorylation of the tau N terminus at tyr18 is blocked
by removal of the C-terminal domain, but not by pseu-
dophosphorylation at residues 262 and 356, suggesting
that phosphorylation of this site might be necessary for
the centrifugal transport of tau into the dendrites, but
is not itself sufficient to cause the formation of focal
Diffuse secretion
We can thus clearly begin to differentiate two gener-
al mechanisms of tau secretion, each of which appears
to be associated with a related mechanism of interneu-
ronal tau transfer. One of these (diffuse secretion from
the soma, resulting in tau transfer to adjacent neurons)
appears to be a relatively simple “default” mechanism
that depends on the presence of an as yet unknown fea-
ture of the tau N-terminal domain (or at least the region
of htau near the Tau12 epitope at 9–18). Our current
results tell us only that this mechanism does not require
either phosphorylation of tyr18, microtubule-mediated
somatofugal transport or the presence or functionality
of the C-terminal domain of tau, and is not obviously
correlated with tau-mediated toxicity. The diffuse se-
cretion mechanism must therefore be entirely unrelated
to tau oligomer or aggregate formation and its attendant
cytotoxicity [2,3], which is consistent with our prior
finding in the lamprey model that the inactivation of
tau:tau aggregation in vivo with a low molecular weight
agent is neuroprotective and greatly enhances the ex-
tent of both diffuse secretion and the uptake of tau by
adjacent neurons [31]. The fact that tau N-terminal
cleavage fragments are themselves frequently toxic [5–
7] also suggests that the apparent lack of toxicity of the
tau N-terminal domain to expressing ABCs might be
due to the efficient removal of the secreted tau via dif-
fusion. We also found that the diffuse secretion mecha-
nism is sufficiently selective and efficient to export tau
from ABCs that are only expressing moderate levels of
tau and to generate a gradient-like pattern of extracel-
lular distribution around the somata of tau expressing
ABCs. This last suggests that diffusely secreted tau
moves away from the soma by a passive mechanism
(presumably diffusion), and does not bind effectively
to and become immobilized by extracellular molecules
(such as heparan sulfate proteoglycans) known to bind
to the tau MTBR [27,47]. Since expression of full
length as well as with N-terminal domain tau constructs
in ABCs also results in some diffuse secretion, it is
possible that the C-terminal domain is inactivated or
proteolytically cleaved prior to secretion (as suggested
by the tau secretion pattern from NB2a/d1 cells; see
the Western blot in Fig. 2). While the characteristic
differences in immunolabel between diffuse and focal-
ly secreted tau (Fig. 7, Table 1) and the enhancement
of diffuse pattern secretion by C-terminal domain in-
activation or removal (Fig. 2A) are consistent with this
possibility, it is unclear whether the MTBR removal or
the inactivation of MTBR-mediated binding is required
before full length tau can be secreted via the diffuse
Focal secretion
While the cellular mechanisms responsible for gen-
erating focal deposits around the dendrites and axons of
tau-expressing ABCs appear considerably more com-
plex than those involved in diffuse secretion, the results
of this study allow us to characterize focal secretion in
a number of important ways with respect to its clear
differences from diffuse tau secretion. The secretion of
tau to focal deposits, like diffuse tau secretion, requires
an unknown element in the N-terminal domain to be
present in order for secretion to occur at all. In addition,
focal deposit formation requires 1) the transport of tau
from the soma to the site of secretion from dendrites or
axons and 2) the immobilization of tau in the extracel-
Page 14
W. Kim et al. / Secretion and interneuronal movement of tau 661
lular space near the site of secretion to form the nidus
of the deposit. Our results show that the focal secretion
pathway is favored by the presence of tauopathy mu-
tations and that the site of maximal focal tau secretion
from the dendrites is spatiotemporally correlated with
the presence of phosphorylation at tyr18 (9G3 label),
the onset of dendritic degeneration and circumstances
(tauopathy mutation, the presence of the MTBR) that
permit and/or favor htau aggregation. This suggests
plausible mechanisms by which both tau transport and
the immobilization of secreted tau might occur, and al-
so why focal deposit formation is so tightly linked to
degeneration, especially in the dendrites.
Tau destined for focal secretion appears to be
centrifugally transported in vesicles by a MT-based
mechanism in ABCs
The mechanism by which tau is transported to its
point of secretion in the axon and dendrites in ABCs
must account for the tight correlations observed be-
tween focal secretion and the point of onset of dendritic
degeneration and the accumulation of 9G3-positive tau
at the apparent point of secretion. Our observations
suggest that this transport is active, since tau accumu-
lates in and tau secretion occurs preferentially from the
distalmost intact parts of dendrites (i.e., the dendritic
tips in intact dendrites, and the point of degeneration
onset in degenerating dendrites). Moreover, the trans-
ported tau exhibits a punctate distribution and colocal-
izes with fluorescent dextran in tau-expressing ABCs,
suggesting that it is carried in vesicles which accumu-
late in the “beads” of degenerating dendrites that stain
poorly for tubulin [20]. The correlation between lo-
cal failure of MT-based vesicular transport and accu-
mulation of tau-containing vesicles in beads with tau
secretion is consistent with our observations that focal
secretion is more associated with beads than from in-
terbead sections of dendrite (Figs 3C, 7C) and that the
actual ability to bind to microtubules is not required for
transport, since 262/356 pseudophosphorylated tau is
efficiently transported into distal dendrites (Fig. 7A–B).
While tau is not generally known to be associated
with cellular trafficking vesicles, it does localize to
vesicular compartments in neurons such as the Golgi,
lysosomes, and autophagosomes in neurodegenerative
tauopathies such as AD [37,38]. The phosphorylation
of tau at tyr18 by fyn or related tyrosine kinases pre-
sumably occurs within or in association with lipid-raft
domains of the plasma membrane in human brain [39]
an event which appears to be exacerbated in the patho-
genesis of AD [10]. Fyn mediated signal transduction
typically leads to oligomerization and endocytosis of
downstream elements [40–42], suggesting that tau that
has been phosphorylated (and possibly oligomerized)
by fyn might plausibly be endocytosed via a similar
mechanism, especially if synergistic interactions lead-
ing to oligomerization occur between tau and the prion
protein or α-synuclein, two other raft-associated pro-
teins that are phosphorylated by fyn and synergistically
oligomerize with tau [43–45]. The correlation that we
observed between the presence of the P301L mutation,
which favors tau oligomerization, and the incidence of
both focal secretion and interneuronal tau transfer is
consistent with this.
Tau is immobilized in focal deposits via the activity of
its C-terminal MTBR
The accumulation of tau in focal perineuritic de-
posits ostensibly requires that tau, normally a small sol-
uble protein, be prevented from diffusing away from
the site of secretion, as appears to happen with diffuse
tau secretion. We observed that more distally locat-
ed (and hence presumably older) “focal” tau deposits
were more granular in appearance and less immunopos-
itive for the GFP tag than were more proximal deposits
(Figs 6, 7C), suggesting that focal tau deposits are in-
deed relatively immobile once secreted and are prote-
olyzed in place over time. The observation that focal
tau deposits increase in prevalence over time (Fig. 6),
are favored by the P301L tauopathy mutation, and are
prevented by pseudophosphorylation at residues 262
and 356 (Table 1) all indicate that the functionality of
the C-terminal MTBR of tau is essential for tau immo-
bilization and suggest that this might dependent on ei-
ther tau:tau interactions leading to the formation of in-
soluble aggregates, and/or interactions with extracellu-
lar polyanionic molecules such as heparan sulfate pro-
teoglycans [35,46], leading to the binding of tau to the
extracellular matrix.
The immobilization of tau in perineuritic deposits
may also contribute to another characteristic feature
of tau in focal deposits, i.e., its tight association with
local dendritic degeneration. While it seems plausi-
ble that ongoing degeneration could favor local tau se-
cretion if it is accompanied by localized high [Ca
several studies over the past few years have suggested
that extracellular tau deposits are themselves toxic, and
that this toxicity is exacerbated by the P301L tauopa-
thy mutation [47,48]. If so, this could fuel a toxicity
feedback loop involving tau secretion and toxicity that
Page 15
662 W. Kim et al. / Secretion and interneuronal movement of tau
in turn could drive progressive dendritic degeneration.
Such a mechanism would account for the exacerbation
of tau toxicity by the P301L mutation of extracellu-
larly applied tau in cultured mammalian neurons and
its increased aggregation state and longevity in focal
peridendritic extracellular deposits (Figs 5, 6D).
Tau movement into postsynaptic neurons
Just as the movement of tau from expressing ABCs
into adjacent neurons appear to be either a consequence
or a special case of diffuse secretion, the secretion of
tau into periaxonal deposits (Fig. 4B) and the specific
anterograde movement of tau from expressing ABCs
into what appear to be postsynaptic neurons in the cau-
dal hindbrain (Fig. 5C) are both dependent on the same
major factor (i.e., the presence of a functional MTBR)
and are both highly dependent on the presence of the
P301L tauopathy mutation. Since the P301L muta-
tion increases the tendency of tau to oligomerize but
inhibits tau:microtubule binding, it seems likely that
both tau secretion and its uptake into postsynaptic cells
might involve tau oligomerization, possibly in the con-
text of local synapse associated [Ca++] fluxes [16,27].
The recent demonstration by Frost and co-workers [16]
that both oligomerization and the P301L mutation in-
crease tau uptake into neuronal cells in culture is con-
sistent with this possibility, which also suggests that
the P301L-induced transsynaptic movement of toxicity
recently observed in an in situ mouse model [15] might
be due to actual transneuronal tau protein transfer.
Implications of tau secretion and transneuronal
movement for the pathogenesis of human tauopathies
Until very recently, little attention has been given to
the possibility that interneuronal transfer of tau protein
between live CNS neurons is an important aspect of
tauopathy pathogenesis. This is probably due both to
the longstanding assumption that tau is an exclusive-
ly cytoskeletal protein and to the difficulty of clear-
ly demonstrating in human patients and murine trans-
genic models that extracellular or CSF tau is due to tau
secretion from live neurons, rather than just a passive
consequence of neuronal death and lysis. Moreover,
tau lacks the usual features typically associated with
secreted proteins, such as an N terminal hydrophobic
“leader” sequence and lipidation sites, and has hereto-
fore never been observed to be secreted from healthy
cells. However, the stereotypic progression of neurofib-
rillary lesions within the brains of patients suffering
from AD and other tauopathies over time [47–51] are
clearly consistent with the interneuronal propagation of
tau pathology between adjacent and synaptically inter-
connected neurons and suggest that tau secretion lead-
ing to interneuronal transfer might become sufficiently
prominent under disease conditions to play a significant
role in tauopathy pathogenesis. The secretion and in-
terneuronal transfer of tau from expressing neurons by
multiple specific pathways described here in the lam-
prey tauopathy model constitutes the first clear demon-
stration that tau secretion can occur via a tau-specific
mechanism from live neurons that overexpress human
tau at relatively low levels and that the secreted tau can
be taken up into adjacent and postsynaptic neurons in
an in situ tauopathy model. This finding is consistent
with and complementary to recent studies in cell culture
and transgenic mouse models, which have respective-
ly demonstrated that tau uptake [16] and transsynaptic
transfer of tau-mediated toxicity (but not necessarily of
tau protein itself [15]) is possible and that it is either
permitted or enhanced by the presence of the P301L
tauopathy mutation. Our results show conclusively that
the actual transfer of tau protein between neurons in the
CNS in situ can occur and thus that “templating” inter-
actions between normal and misfolded proteins might
mediate lesion propagation in tauopathy pathogenesis
in a manner similar to what is thought to occur in prion
It should be noted that since diffuse and focal secre-
tion mechanisms appear to be favored by (respectively)
either tau cleavage or aggregation of tau, it is possible
that neither one normally operates at an appreciable
rate in humans, and may be imperceptible in the neu-
ropathology of tauopathy patients due to rapid clear-
ance to the CSF of secreted tau that is not already high-
ly aggregated (such as the tau in “ghost” tangles). The
early appearance of tau species in the CSF in AD and
other tauopathies [52] suggests that tau secretion might
play an active role in the pathogenesis of these diseases,
instead of being just a marker of antecedent neuronal
death. In particular, the generation of transventricular
deposits from non-degenerating or mildly degenerated
ABCs (as shown in Fig. 5) calls into question the com-
mon assumption that the presence of tau or tau frag-
ments in the CSF is always indicative of extensive prior
neuronal loss in humans. This finding in the lamprey
model suggests a plausible reason why increased CSF
tau levels might occur prior to the onset of archetypical
AD dementia without requiring the invocation of mas-
sive (and seemingly improbable) neuronal death and
lysis very early in the disease course [52]. The finding
Page 16
W. Kim et al. / Secretion and interneuronal movement of tau 663
that the export of N-terminal fragments (which may
correspond to diffuse extracellular tau secretion) and
full length htau can occur by distinct pathways in an in
situ model may also be relevant to observations that the
ratio of N-terminal fragments to full length htau in the
CSF varies characteristically between different human
tauopathies [53,54] and thus may provide insights into
mechanistic distinctions between the pathogeneses of
these disorders in the future.
This work was supported by NIH grants R01
AG13919 to GFH and R44 AG018661 to GFH in col-
laboration with Neuronautics Inc., Chicago IL. We
would like to acknowledge the excellent technical con-
tributions of Dr. Jun Yao to some of the experiments
described in this paper, and the generous contributions
of Tau1,Tau12 and Tau5 from Dr. Lester Binder and
PHF1 from Dr. Peter Davies.
Authors’ disclosures available online (http://www.j-
[1] Lee VM, Goedert M, Trojanowski JQ (2001) Neurodegenera-
tive tauopathies. Annu Rev Neurosci 24, 1121-1159.
[2] Avila J (2006) Tau phosphorylation and aggregation in
Alzheimer’s disease pathology. FEBS Lett 580, 2922-2927.
[3] Mocanu MM, Nissen A, Eckermann K, Khlistunova I, Biernat
J, Drexler D, Petrova O, Schonig K, Bujard H, Mandelkow
E, Zhou L, Rune G, Mandelkow EM (2008) The potential for
beta-structure in the repeat domain of tau protein determines
aggregation, synaptic decay, neuronal loss, and coassembly
with endogenous Tau in inducible mouse models of tauopathy.
J Neurosci 28, 737-748.
[4] Bhaskar K, Yen SH, Lee G (2005) Disease-related modifica-
tions in tau affect the interaction between Fyn and Tau. J Biol
Chem 280, 35119-35125.
[5] Park SY, Ferreira A (2005) The generation of a 17 kDa neuro-
toxic fragment: an alternative mechanism by which tau medi-
ates beta-amyloid-induced neurodegeneration. J Neurosci 25,
[6] Corsetti V, Amadoro G, Gentile A, Capsoni S, Ciotti MT, Cen-
cioni MT, Atlante A, Canu N, Rohn TT, Cattaneo A, Calissano
P (2008) Identification of a caspase-derived N-terminal tau
fragment in cellular and animal Alzheimer’s disease models.
Mol Cell Neurosci 38, 381-392.
[7] King ME, Kan HM, Baas PW, Erisir A, Glabe CG, Bloom
GS (2006) Tau-dependent microtubule disassembly initiated
by prefibrillar beta-amyloid. J Cell Biol 175, 541-546.
[8] Lee G, Newman ST, Gard DL, Band H, Panchamoorthy G
(1998) Tau interacts with src-family non-receptor tyrosine ki-
nases. J Cell Sci 111 (Pt 21), 3167-3177.
[9] Brandt R, Leger J, Lee G (1995) Interaction of tau with the
neural plasma membrane mediated by tau’s amino-terminal
projection domain. J Cell Biol 131, 1327-1340.
[10] Lee G, Thangavel R, Sharma VM, Litersky JM, Bhaskar K,
Fang SM, Do LH, Andreadis A, Van Hoesen G, Ksiezak-
Reding H (2004) Phosphorylation of tau by fyn: implications
for Alzheimer’s disease. J Neurosci 24, 2304-2312.
[11] Fevrier B, Vilette D, Archer F, Loew D, Faigle W, Vidal M,
Laude H, Raposo G (2004) Cells release prions in association
with exosomes. Proc Natl Acad Sci U S A 101, 9683-9688.
[12] Caughey B, Baron GS, Chesebro B, Jeffrey M (2009) Get-
ting a grip on prions: oligomers, amyloids, and pathological
membrane interactions. Annu Rev Biochem 78, 177-204.
[13] Lee HJ, Patel S, Lee SJ (2005) Intravesicular localization and
exocytosis of alpha-synuclein and its aggregates. J Neurosci
25, 6016-6024.
[14] Rajendran L, Honsho M, Zahn TR, Keller P, Geiger KD,
Verkade P, Simons K (2006) Alzheimer’s disease beta-amyloid
peptides are released in association with exosomes. Proc Natl
Acad Sci U S A 103, 11172-11177.
[15] Clavaguera F, Bolmont T, Crowther RA, Abramowski D,
Frank S, Probst A, Fraser G, Stalder AK, Beibel M, Staufen-
biel M, Jucker M, Goedert M, Tolnay M (2009) Transmission
and spreading of tauopathy in transgenic mouse brain. Nat
Cell Biol 11, 909-913.
[16] Frost B, Jacks RL, Diamond MI (2009) Propagation of tau
misfolding from the outside to the inside of a cell. J Biol Chem
284, 12845-12852.
[17] Gotz J, Deters N, Doldissen A, Bokhari L, Ke Y, Wiesner A,
Schonrock N, Ittner LM (2007) A decade of tau transgenic
animal models and beyond. Brain Pathol 17, 91-103.
[18] Hall GF, Yao J (2005) Modeling tauopathy: a range of com-
plementary approaches. Biochim Biophys Acta 1739, 224-239.
[19] Hall GF, Chu B, Lee G, Yao J (2000) Human tau filaments
induce microtubule and synapse loss in an in vivo model of
neurofibrillary degenerative disease. J Cell Sci 113 (Pt 8),
[20] Lee S, Jung C, Lee G, Hall GF (2009) Exonic point muta-
tions of human tau enhance its toxicity and cause characteris-
tic changes in neuronal morphology, tau distribution and tau
phosphorylation in the lamprey cellular model of tauopathy. J
Alzheimers Dis 16, 99-111.
[21] Hall GF, Yao J, Lee G (1997) Human tau becomes phospho-
rylated and forms filamentous deposits when overexpressed in
lamprey central neurons in situ. Proc Natl Acad Sci U S A 94,
[22] Hall GF, Lee VM, Lee G, Yao J (2001) Staging of neurofib-
rillary degeneration caused by human tau overexpression in a
unique cellular model of human tauopathy. Am J Pathol 158,
[23] Leger J, Kempf M, Lee G, Brandt R (1997) Conversion of
serine to aspartate imitates phosphorylation-induced changes
in the structure and function of microtubule-associated protein
tau. J Biol Chem 272, 8441-8446.
[24] Chan WK, Dickerson A, Ortiz D, Pimenta AF, Moran CM,
Motil J, Snyder SJ, Malik K, Pant HC, Shea TB (2004)
Mitogen-activated protein kinase regulates neurofilament ax-
onal transport. J Cell Sci 117, 4629-4642.
[25] Mukrasch MD, Biernat J, von Bergen M, Griesinger C, Man-
delkow E, Zweckstetter M (2005) Sites of tau important for ag-
gregation populate {beta}-structure and bind to microtubules
and polyanions. J Biol Chem 280, 24978-24986.
[26] Wickelgren WO (1977) Physiological and anatomical charac-
teristics of reticulospinalneurones in lamprey. J Physiol 270,
Page 17
664 W. Kim et al. / Secretion and interneuronal movement of tau
[27] Nacharaju P, Lewis J, Easson C, Yen S, Hackett J, Hutton
M, Yen SH (1999) Accelerated filament formation from tau
protein with specific FTDP-17 missense mutations. FEBS Lett
447, 195-199.
[28] Schneider A, Biernat J, von Bergen M, Mandelkow E, Man-
delkow EM (1999) Phosphorylation that detaches tau protein
from microtubules (Ser262, Ser214) also protects it against
aggregation into Alzheimer paired helical filaments. Biochem-
istry 38, 3549-3558.
[29] Tanudji M, Hevi S, Chuck SL (2002) Improperly folded green
fluorescent protein is secreted via a non-classical pathway. J
Cell Sci 115, 3849-3857.
[30] McKee AC, Kowall NW, Kosik KS (1989) Microtubular re-
organization and dendritic growth response in Alzheimer’s
disease. Ann Neurol 26, 652-659.
[31] Hall GF, Lee S, Yao J (2002) Neurofibrillary degeneration
can be arrested in an in vivo cellular model of human tauopa-
thy by application of a compound which inhibits tau filament
formation in vitro. J Mol Neurosci 19, 253-260.
[32] Nickel W (2003) The mystery of nonclassical protein secre-
tion. A current view on cargo proteins and potential export
routes. Eur J Biochem 270, 2109-2119.
[33] Crowther RA, Olesen OF, Smith MJ, Jakes R, Goedert M
(1994) Assembly of Alzheimer-like filaments from full-length
tau protein. FEBS Lett 337, 135-138.
[34] Hasegawa M, Crowther RA, Jakes R, Goedert M (1997)
Alzheimer-like changes in microtubule-associated protein Tau
induced by sulfated glycosaminoglycans. Inhibition of micro-
tubule binding, stimulation of phosphorylation, and filament
assembly depend on the degree of sulfation. J Biol Chem 272,
[35] Perez M, Valpuesta JM, Medina M, Montejo de Garcini E,
Avila J (1996) Polymerization of tau into filaments in the
presence of heparin: the minimal sequence required for tau-tau
interaction. J Neurochem 67, 1183-1190.
[36] Biernat J, Gustke N, Drewes G, Mandelkow EM, Mandelkow
E (1993) Phosphorylation of Ser262 strongly reduces bind-
ing of tau to microtubules: distinction between PHF-like im-
munoreactivity and microtubule binding. Neuron 11, 153-163.
[37] Liu F, Zaidi T, Iqbal K, Grundke-Iqbal I, Merkle RK, Gong
CX (2002) Role of glycosylation in hyperphosphorylation of
tau in Alzheimer’s disease. FEBS Lett 512, 101-106.
[38] Cataldo AM, Barnett JL, Berman SA, Li J, Quarless S, Bursz-
tajn S, Lippa C, Nixon RA (1995) Gene expression and cellular
content of cathepsin D in Alzheimer’s disease brain: evidence
for early up-regulation of the endosomal-lysosomal system.
Neuron 14, 671-680.
[39] Vega IE, Cui L, Propst JA, Hutton ML, Lee G, Yen SH (2005)
Increase in tau tyrosine phosphorylation correlates with the
formation of tau aggregates. Brain Res Mol Brain Res 138,
[40] Tournaviti S, Hannemann S, Terjung S, Kitzing TM, Stegmay-
er C, Ritzerfeld J, Walther P, Grosse R, Nickel W, Fackler OT
(2007) SH4-domain-induced plasma membrane dynamization
promotes bleb-associated cell motility. J Cell Sci 120, 3820-
[41] Freedman SD, Katz MH, Parker EM, Gelrud A (1999) En-
docytosis at the apical plasma membrane of pancreatic aci-
nar cells is regulated by tyrosine kinases. Am J Physiol 276,
[42] Sverdlov M, Shajahan AN, Minshall RD (2007) Tyrosine
phosphorylation-dependence of caveolae-mediated endocyto-
sis. J Cell Mol Med 11, 1239-1250.
[43] Geddes JW (2005) alpha-Synuclein: a potent inducer of tau
pathology. Exp Neurol 192, 244-250.
[44] Giasson BI, Forman MS, Higuchi M, Golbe LI, Graves CL,
Kotzbauer PT, Trojanowski JQ, Lee VM (2003) Initiation and
synergistic fibrillization of tau and alpha-synuclein. Science
300, 636-640.
[45] Wang XF, Dong CF, Zhang J, Wan YZ, Li F, Huang YX, Han
L, Shan B, Gao C, Han J, Dong XP (2008) Human tau protein
forms complex with PrP and some GSS- and fCJD-related PrP
mutants possess stronger binding activities with tau in vitro.
Mol Cell Biochem 310, 49-55.
[46] Nickel W (2007) Unconventional secretion: an extracellular
trap for export of fibroblast growth factor 2. J Cell Sci 120,
[47] Gomez-Ramos A, Diaz-Hernandez M, Rubio A, Miras-
Portugal MT, Avila J (2008) Extracellular tau promotes in-
tracellular calcium increase through M1 and M3 muscarinic
receptors in neuronal cells. Mol Cell Neurosci 37, 673-681.
[48] Gomez-Ramos A, Diaz-Hernandez M, Cuadros R, Hernandez
F, Avila J (2006) Extracellular tau is toxic to neuronal cells.
FEBS Lett 580, 4842-4850.
[49] Armstrong RA, Cairns NJ, Lantos PL (1999) Clustering of
cerebral cortical lesions in patients with corticobasal degener-
ation. Neurosci Lett 268, 5-8.
[50] Braak H, Braak E (1991) Neuropathological stageing of
Alzheimer-related changes. Acta Neuropathol 82, 239-259.
[51] Armstrong RA, Cairns NJ, Lantos PL (1998) Clustering of
Pick bodies in patients with Pick’s disease. Neurosci Lett 242,
[52] Andreasen N, Minthon L, Vanmechelen E, Vanderstichele H,
Davidsson P, Winblad B, Blennow K (1999) Cerebrospinal
fluid tau and Abeta42 as predictors of development of
Alzheimer’s disease in patients with mild cognitive impair-
ment. Neurosci Lett 273, 5-8.
[53] Borroni B, Gardoni F, Parnetti L, Magnob L, Malinverno M,
Saggese E, Calabresi P, Spillantini MG, Padovani A, Di Luca
M (2008) Pattern of tau form in CSF is alterered in progressive
supranuclear palsy. Neurobiology of Aging 30, 34-40.
[54] Urakami K, Wada K, Arai H, Sasaki H, Kanai M, Shoji M,
Ishizu H, Kashihara K, Yamamoto M, Tsuchiya-Ikemoto K,
Morimatsu M, Takashima H, Nakagawa M, Kurokawa K,
Maruyama H, Kaseda Y, Nakamura S, Hasegawa K, Oono H,
Hikasa C, Ikeda K, Yamagata K, Wakutani Y, Takeshima T,
Nakashima K (2001) Diagnostic significance of tau protein
in cerebrospinal fluid from patients with corticobasal degen-
eration or progressive supranuclear palsy. J Neurol Sci 183,
Page 18
  • Source
    • "A nontransgenic lower vertebrate tauopathy model (the lamprey ABC model) has been used to express full-length wild type and mutant human Tau isoforms in identified neurons , thus allowing localization of toxic Tau sources. Thanks to this model system, Tau was found to be secreted before the onset of neuronal degeneration and to be transferred among neurons, thus spreading in a disease-specific pattern to the brain and playing a major role in pathogenesis [144, 145]. Association of Tau with exosomes suggests that extracellular vesicles is at least one of the routes for active interneuronal transfer of toxic protein [140]. "
    [Show abstract] [Hide abstract] ABSTRACT: Extracellular vesicles are involved in a great variety of physiological events occurring in the nervous system, such as cross talk among neurons and glial cells in synapse development and function, integrated neuronal plasticity, neuronal-glial metabolic exchanges, and synthesis and dynamic renewal of myelin. Many of these EV-mediated processes depend on the exchange of proteins, mRNAs, and noncoding RNAs, including miRNAs, which occurs among glial and neuronal cells. In addition, production and exchange of EVs can be modified under pathological conditions, such as brain cancer and neurodegeneration. Like other cancer cells, brain tumours can use EVs to secrete factors, which allow escaping from immune surveillance, and to transfer molecules into the surrounding cells, thus transforming their phenotype. Moreover, EVs can function as a way to discard material dangerous to cancer cells, such as differentiation-inducing proteins, and even drugs. Intriguingly, EVs seem to be also involved in spreading through the brain of aggregated proteins, such as prions and aggregated tau protein. Finally, EVs can carry useful biomarkers for the early diagnosis of diseases. Herein we summarize possible roles of EVs in brain physiological functions and discuss their involvement in the horizontal spreading, from cell to cell, of both cancer and neurodegenerative pathologies.
    Full-text · Article · Nov 2015
  • Source
    • "The cell-to-cell propagation ability is out of the question not only for A and tau as it has been shown in a neural graft transplanted mouse model (Meyer-Luehmann et al., 2003) but also in several animal models where seeding and the propagation ability of those proteins were assessed and the propagation from injection site through neuronal projections was observed (Gotz et al., 2001; Walker et al., 2002; Bolmont et al., 2007; Clavaguera et al., 2009 Clavaguera et al., , 2013 Sydow and Mandelkow, 2010; Hurtado et al., 2010; de Calignon et al., 2012; Lasagna-Reeves et al., 2012; Iba et al., 2013). And even in an unusual animal model, tau spreading was shown along human tau expressing Lamprey neurons (Kim et al., 2010). In particular, some studies focus on propagation mechanisms paying special attention, for instance, to the temporal progression of A plaque accumulation in animal models showing that it starts with intracellular deposits in certain brain areas that slowly progress to extracellular inclusions possibly derived from 'leaking' neurons that eventually spread to most of the brain (Leon et al., 2010 ). "
    [Show abstract] [Hide abstract] ABSTRACT: Prion diseases or Transmissible Spongiform Encephalopathies (TSEs) are a group of fatal neurodegenerative disorders affecting several mammalian species being Creutzfeldt-Jacob Disease (CJD) the most representative in human beings, scrapie in ovine, Bovine Spongiform Encephalopathy (BSE) in bovine and Chronic Wasting Disease (CWD) in cervids. As stated by the "protein-only hypothesis", the causal agent of TSEs is a self-propagating aberrant form of the prion protein (PrP) that through a misfolding event acquires a β-sheet rich conformation known as PrP(Sc) (from scrapie). This isoform is neurotoxic, aggregation prone and induces misfolding of native cellular PrP. Compelling evidence indicates that disease-specific protein misfolding in amyloid deposits could be shared by other disorders showing aberrant protein aggregates such as Alzheimer's Disease (AD), Parkinson's Disease (PD), Amyotrophic lateral sclerosis (ALS) and systemic Amyloid A amyloidosis (AA amyloidosis). Evidences of shared mechanisms of the proteins related to each disease with prions will be reviewed through the available in vivo models. Taking prion research as reference, typical prion-like features such as seeding and propagation ability, neurotoxic species causing disease, infectivity, transmission barrier and strain evidences will be analyzed for other protein-related diseases. Thus, prion-like features of amyloid β peptide and tau present in AD, α-synuclein in PD, SOD-1, TDP-43 and others in ALS and serum α-amyloid (SAA) in systemic AA amyloidosis will be reviewed through models available for each disease. Copyright © 2015. Published by Elsevier B.V.
    Full-text · Article · Apr 2015 · Virus Research
  • Source
    • "It is unclear whether the release of pathogenic tau from neurons is related to the normal pathway of physiological release or unconventional secretory mechanisms [102, 103]. Experimentally, tau fibrils have been demonstrated to transfer between cells in cell culture [81, 104, 105] and transynaptically in vivo [41, 42, 105, 106]. Transynaptic spread in mouse models and both anterograde and retrograde trafficking of misfolded tau in neurons [87] strongly indicates the importance of synaptic connectivity in tauopathy pathogenesis. "
    [Show abstract] [Hide abstract] ABSTRACT: Tauopathies encompass a broad family of neurodegenerative diseases, including Alzheimer's disease, which are characterized by the fibrillization of the microtubule-associated tau protein. The normal function of tau is to stabilize and promote the assembly of microtubules in neuronal axons. Sequestration of tau into amyloid fibrils results in destabilization of the microtubule network and may contribute to disease progression. As tau is an intracellular protein and proteins do not passively cross cell membranes, tau fibril formation has been assumed to occur spontaneously within individual cells. However, recent evidence suggests that tau shares several characteristics with prions, which propagate through the brain by protein-protein interactions in the interstitial space; these characteristics include conformational templating of native tau into disease-associated fibrils and intercellular fibril propagation. Tau adopts diverse fibril structures, or strains, which have been shown to self-propagate in the presence of monomeric recombinant tau protein. Exogenous tau fibrils induce misfolding of native tau in both cell culture and animal models, causing strain-dependent cellular dysfunction and differential patterns of neuropathology. Tau fibers have also been found recently in patient samples or models of several diseases not formerly identified as tauopathies, including chronic traumatic encephalopathy, Parkinson's disease, and Huntington's disease, suggesting a common underlying mechanism for neurodegenerative diseases. The possibility that tauopathies and other neurodegenerative diseases involve prion-like mechanisms has implications for studies designed to understand disease pathogenesis and for the development of therapies, which may be devised to impede tau strain propagation and intercellular transmission, allowing clearance of tau fibrils and potentially halting or reversing disease progression.
    Full-text · Chapter · Jan 2015
Show more