| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Experimental Genetics Group,*
Center for Human
Genetics, Flemish Institute for Biotechnology; and the Department of
Pathology,
University Hospitals Leuven,
K. U. Leuven, Gasthuisberg, Leuven, Belgium
 |
Abstract |
|---|
 Top Abstract IntroductionExperimental ProceduresResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionExperimental ProceduresResultsDiscussionReferences |
|---|
Epidemiological data do not provide evidence for a direct role of ApoE in central nervous system disorders, and its mechanism of action within the central nervous system is not clear. Originally, it was thought that ApoE present in neurons originated only from endocytosis.20-22 Numerous cell culture experiments demonstrated receptor-mediated uptake of ApoE as well as effects on neurite outgrowth and cell-morphology.23-29 However, more and more evidence indicates that intraneuronal ApoE might also originate from synthesis by neurons. In situ hybridization of brain sections of transgenic mice expressing genomic fragments containing the entire human ApoE locus, including its promoter sequences, revealed expression in neurons, besides astrocytes.30-32 Human neuroblastoma cells were shown to synthesize both ApoE mRNA and protein.33,34 Moreover, in situ hybridization of human brain sections conclusively demonstrated ApoE mRNA in neurons.35 Taken together these results suggest that besides its well-known synthesis by glia cells and its role in lipid transport and metabolism, not only uptake but also synthesis of ApoE by neurons might be important. Interestingly, the brain regions expressing ApoE in neurons seemed most vulnerable to the development of neurofibrillary pathology,36 raising the possibility that the neuronal expression pattern of human ApoE might be important in the pathogenesis of AD.
To experimentally test this hypothesis, we have generated transgenic mice overexpressing human ApoE4 in neurons or astrocytes of the central nervous system. Initial characterization of these mice showed that the microtubule-associated protein Tau was hyperphosphorylated in the brain of transgenic mice expressing ApoE4 in neurons, but not in mice expressing human ApoE4 in nonneuronal cells.37 Here we present evidence that transgenic mice expressing human ApoE4 in neurons, suffer from axonal degeneration and gliosis in brain and spinal cord. Phenotypically, this resulted in reduced sensorimotor capacities and neurogenic muscle atrophy. Impairment of axonal transport was documented immunohistochemically by accumulation of synaptophysin and ultrastructurally by accumulation of vesicles and mitochondria. In sharp contrast, none of these features were found in mice expressing human ApoE4 in astrocytes.
|
Experimental Procedures |
|---|
|
TopAbstractIntroductionExperimental ProceduresResultsDiscussionReferences |
|---|
All transgenic mice contained a 5.5-kb genomic fragment of the human ApoE4 gene, including the four exons and three introns, controlled by the Thy1, platelet-derived growth factor (PDGF), glial fibrillary acidic protein (GFAP), or phosphoglycerate kinase promoters. The transgene was expressed in neurons when under control of the Thy1, PDGF, or phosphoglycerate kinase promoters, and in glial cells only, when under control of the GFAP promoter. Offspring of transgenic mice were genotyped for the presence of the transgene as described.37
Collection of Cerebrospinal Fluid (CSF) and Neuronal and Glial Cell Culture
One µl of CSF, collected from 3-month-old mice, was used for Western blotting as previously described.38
Primary hippocampal neuronal cultures of 1-day-old Thy1-ApoE4 transgenic and wild-type mice and primary cortical glial cultures of 3-day-old GFAP-ApoE4 transgenic and wild-type mice were prepared as described.39 Hippocampal neurons were grown in the presence of wild-type astrocyte feeder layers for 18 days. Cortical glial cultures of GFAP-ApoE4 transgenic mice were cultured for 11 days, until confluent. Thereafter the medium was removed and the glial cells were grown in serum-free Dulbecco’s modified Eagle’s medium/F12 medium containing N2 supplement (Life Technologies, Inc., Merelbeke, Belgium) for an additional 72 hours. Medium was collected from both neuron and glial cultures, and 20 µl of unconcentrated medium was analyzed by Western blotting. Cells were washed twice with phosphate-buffered saline (PBS), scraped from the dishes, and homogenized in 50 µl (neurons) or 100 µl (glia) of buffer. Ten µl of the homogenates were analyzed by Western blotting.
Behavioral Testing
All mice used for behavioral testing were kept under standard light-dark cycle conditions and housed with four mice per cage, having free access to food and water. The open field arena was 1-m diameter with walls 35 cm high, painted black, and lighted diffusely (110 lux). Mice were environmentally adapted to the room 1 hour before testing. Behavioral tests were performed between 13.00 and 17.00. Mice were placed individually on one side of the arena and observed for 10 minutes continuously, but analyzed as two observation blocks of 5 minutes. Total horizontal activity was measured using a computer automated tracking system (PolyTrack, Columbus Instruments, San Diego, CA).
For the forced swim test, mice were allowed to swim freely in a water-tank (105 cm in diameter, 35 cm high), with walls and floor painted white. The tank was filled to a height of 20 cm with opaque water kept at 31 ± 2°C. Total swimming distance was measured with the computerized tracking system and average swimming speed was calculated.
Mice were subjected to a series of sensorimotor tasks to assess muscle strength and equilibrium. The grid suspension test measured the ability of mice to hang upside-down from a wire grid (10 x 18 cm with 1-mm wires spaced 10 mm apart) for a maximum of 1 minute in two separate trials. The prehensile reflex was measured as the ability of an animal to remain suspended by the forepaws grasped around an elevated horizontal bar (2 mm in diameter) and is considered to be a measure of muscle strength. The traction capacity was scored as the number of hindlimbs that the animal raised to reach the wire. This test is considered to measure equilibrium and muscle tone as well as muscle strength.40 Sensorimotor tests were performed as described.40
Data were statistically analyzed by one-way analysis of variance and by Student’s t-test.
All behavioral tests were performed with the genetic status of the mice unknown to the experimenter.
Histology and Immunohistochemistry
Anesthetized mice were perfused with saline followed by 4% paraformaldehyde in PBS. The brains were postfixed for 16 hours in 4% paraformaldehyde in PBS at 4°C and washed in PBS buffer, dehydrated, and embedded in paraffin for sectioning (5 to 7 µm). Alternatively, vibratome sections of 50 µm were cut and transferred to microtiter wells in PBS buffer containing 0.1% sodium azide at 4°C.
Spinal cord was removed and sectioned in four parts (cervical, thoracal, lumbal, and sacral). Muscle biopsies were taken from the quadriceps. Tissues were postfixed overnight in 4% paraformaldehyde in PBS at 4°C, embedded in paraffin, and sectioned (5 to 7 µm).
Paraffin sections were routinely stained with hematoxylin and eosin (H&E) and Bielchowsky’s silver impregnation. Immunohistochemical staining of paraffin sections was performed according to standard procedures with commercially available antibodies to the following antigens: GFAP (1:500; DAKO, Glostrup, Denmark), synaptophysin (1:500, DAKO), ubiquitin (1:200, DAKO), AT8 (Innogenetics, Gent, Belgium), AT180 (Innogenetics), and SMI31 (1:1,000; Affinity, Nottingham, UK). Vibratome sections were treated with 0.6% (w/v) hydrogen peroxide to quench the endogenous peroxidase activity. After rinsing, sections were incubated in Tris-buffered saline, containing 10% goat serum and 0.2% Triton X-100, for 2 hours before incubation with specific antibodies. Secondary antibody was biotinylated goat anti-rabbit or goat anti-mouse IgG followed by StreptABComplex/HRP (DAKO). Final staining was developed with 0.075% 3,3-diaminobenzidine and 0.01% (w/v) hydrogen peroxide. Sections were mounted on gelatin-coated glass slides, counterstained with hematoxylin, dehydrated, and mounted. Sections were stained with commercially available antibodies to ApoE (1:10,000, DAKO) and NF200 (1:500, Sigma, St Louis, MO).
Hippocampi used for ultrastructural examination, were excised from vibratome sections from mice perfused with 4% paraformaldehyde in PBS and were postfixed in 1.25% glutaraldehyde in PBS. Sciatic nerves were fixed in 1.25% glutaraldehyde in PBS. Semithin and ultrathin sections were prepared in a routine manner after embedding in epon and postfixation with osmium tetroxide.
Quantification of Neurons in the Ventral Horns of the Spinal Cord
Three wild-type FVB mice and four Thy1-ApoE4 transgenic mice that were 12 to 14 months old were transcardially perfused with 4% paraformaldehyde in PBS. Spinal cords were immersion-fixed overnight and embedded in paraffin. Microtome sections (6 µm) of the cervical and thoracolumbal regions were submitted to standard cresyl-violet staining. Composite images from a 3-charge-coupled device color video camera were collected and assembled with appropriate software (AIS/C 4.0; Imaging Research, St. Catharine’s, Ontario, Canada) and the number of neurons in the right and left ventral horn in nine series of three successive sections was quantified. Counting of the neurons was performed with the genetic status of the mice unknown to the experimenter.
RNA Extraction and Northern Blotting
Mouse tissues were mechanically homogenized in Trizol Reagent (Life Technologies, Glasgow, UK) and total RNA was isolated as recommended by the manufacturer. Samples of total RNA (10 µg) were denatured for 10 minutes at 70°C in 50% deionized formamide, 20 mmol/L MOPS, 5 mmol/L sodium acetate, 1 mmol/L ethylenediaminetetraacetic acid, and 2.2 mol/L formaldehyde, pH 8.0, separated by electrophoresis in agarose gels, and the RNA was transferred to Hybond-N membranes (Amersham, Amersham, UK). Membranes were prehybridized for 6 hours at 42°C in 5x sodium chloride/sodium phosphate/ethylenediaminetetraacetic acid, 5x Denhardt’s, 0.5% sodium dodecyl sulfate, 50% deionized formamide, 100 µg/ml denatured salmon sperm DNA, and 50 µg/ml heparin, and subsequently hybridized overnight at 42°C in the same solution supplemented with 10% dextransulfate and the appropriately [32P]-radiolabeled DNA probe (2 x 106 cpm/ml). After washing with 0.3x sodium chloride/sodium phosphate/ethylenediaminetetraacetic acid and 0.5% sodium dodecyl sulfate at 68°C for 1 hour, membranes were exposed to Hyperfilm MP (Amersham) for 3 hours to 7 days.
Protein Extraction and Western Blotting
Spinal cord tissue was processed as described for brain tissue,37 and Western blotting was performed after separation of proteins by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on precast gels (NOVEX, San Diego, CA). Western blotting was performed with monoclonal antibodies AT8 (Innogenetics), AT180 (Innogenetics), PHF-1 (gift from Dr. P. Davies, Albert Einstein College of Medicine, Bronx, NY), TAU5 (Pharmingen, San Diego, CA), ApoE (DAKO), aGFAP (DAKO), and aSYN (DAKO).
Quantitative analysis of Western blots, assessed by densitometric scanning, was normalized, and three to four mice per transgenic strain were analyzed.
|
Results |
|---|
|
TopAbstractIntroductionExperimental ProceduresResultsDiscussionReferences |
|---|
The different transgenic mouse strains were demonstrated to express human ApoE4 in neurons, when under control of the Thy1, PDGF, and PGK gene promoters, and exclusively in astrocytes when under control of the GFAP gene promoter.37
The transgenic human ApoE4 was also expressed in spinal cord, as
demonstrated by Northern blotting of total RNA extracts from spinal
cord of Thy1-ApoE4 (tae-II), PDGF-ApoE4 (pae-II), GFAP-ApoE4 (gae-I),
and PGK-ApoE4 (pgk-I) transgenic mice (Figure 1a)
. The human ApoE probe used did not
react with RNA from wild-type mice or ApoE knockout mice. Western
blotting of extracts of spinal cord identified human ApoE4 protein
(Figure 1b)
. Quantification of relative levels of human ApoE4 protein
in spinal cord demonstrated that the expression levels in the tae-II
and tae-XIII transgenic mice were respectively 3.5- and 12-fold higher
than in the pae-II transgenic mice. In the gae-I transgenic mice, ApoE4
expression levels were 6.7-fold higher than in the pae-II transgenic
mice.
|
. In Thy1-ApoE4
transgenic mice, the ApoE4 protein was present in the neuronal cell
bodies in the gray matter (Figure 1, d and C)
, whereas in GFAP-ApoE4
transgenic mice the ApoE4 protein was found only in astrocytes (Figure 1, d and A)
, as expected.
Human ApoE4 was demonstrated in CSF and in cell culture by Western
blotting. CSF was collected from Thy1-ApoE4, GFAP-ApoE4, and wild-type
mice to demonstrate the secretion of human ApoE4 in vivo
(Figure 1c)
. Medium of primary hippocampal neurons derived from
Thy1-ApoE4 transgenic mice, or primary cortical glial cells derived
from GFAP-ApoE4 transgenic mice contained human ApoE4 demonstrating
secretion from neurons and from astrocytes (Figure 1, f and g)
.
Spontaneous Behavior of ApoE4 Transgenic Mice
During the first 2 months of life, all ApoE4 transgenic mice used
in this study appeared normal. Thereafter, Thy1-ApoE4 mice manifested
loss of body weight, which progressively worsened with aging. When
lifted by the tail, Thy1-ApoE4 mice showed an abnormal extension
reflex. They retracted both forelimbs and hindlimbs (Figure 1, e and B)
, whereas nontransgenic mice extended their legs (Figure 1, e and A)
.
During movement in the cage, Thy1-ApoE4 mice showed difficulties with
walking and climbing. Transgenic mice of the highest expressing
Thy1-ApoE4 strain (tae-XIII) developed the most severe phenotype, and
homozygous mice developed the symptoms earlier than their heterozygous
littermates, indicating that the symptoms were related to the
expression level of the transgene. Thy1-ApoE4 mice died prematurely,
which was also more pronounced in the higher expressing
strain.37
In PDGF-ApoE4 transgenic mice that were >16 months old, an abnormal extension reflex was observed in a limited number of mice, together with a reduction in body weight. In contrast, abnormal behavior or decreased body weight was never observed in GFAP-ApoE4 transgenic mice, even when >24 months old.
Experimental Testing of Behavior
Spontaneous locomotor activity was measured blindly, in a
computerized open field test. Mice used in this test were between 12
and 16 months old. Thy1-ApoE4 transgenic mice were the least active:
they traveled only 9.5% of the distance compared to age-matched
wild-type mice (P < 0,001) (Figure 2E)
. The PDGF-ApoE4 and GFAP-ApoE4
transgenic mice were normally active in this test (Figure 2E)
. Neither
the computerized analysis nor the visual inspection of the recorded
locomotion paths revealed anomalies, such as wall-bumping, circling, or
meandering paths, except for one GFAP-ApoE4 mouse, which was excluded
from the analysis. This particular mouse was circling around its tail,
a phenotype occasionally observed in wild-type mice of the Fvb strain
and considered irrelevant to the transgenic phenotype.
|
. This reduction further increased with age, as 12- to 16-month-old
Thy1-ApoE4 transgenic mice showed a 92% reduction in wire hang time
(P < 0.001) (Figure 2B)
. All other 12- to
16-month-old transgenic mice analyzed scored normally in the grid
suspension test (Figure 2B)
.
The decrease in muscle strength in the Thy1-ApoE4 transgenic mice was
closely related to the decreased body weight (Figure 2, A and C)
. At 3
months the body weight of Thy1-ApoE4 transgenic mice was reduced by
7.5% relative to wild-type mice (Figure 2H)
, although the difference
was not significant. At 12 to 16 months however, the body weight of the
Thy1-ApoE4 transgenic mice was significantly reduced by 30%
(P < 0.001) (Figure 2A)
. In addition,
PDGF-ApoE4 transgenic mice of 2 years showed a 20% reduction in body
weight (P < 0.01), that was not observed in 12-
to 16-month-old mice.
When the 12- to 16-month-old mice were subjected to a forced swimming test, no differences in swimming speed were detected between PDGF-ApoE4, GFAP-ApoE4, and wild-type mice. The Thy1-ApoE4 transgenic mice were not tested because preliminary testing revealed that they were not able to swim and would drown during the 2 minutes forced swimming test.
Motor impairment in 3-month-old Thy1-ApoE4 transgenic mice was examined
by additional tests. The prehensile reflex, a measure for the muscle
strength of the forepaws,40
was reduced with 39%
(P < 0.02) (Figure 2F)
in Thy1-ApoE4 transgenic
mice compared to wild-type littermates. Traction capacity, a measure
for the muscle tone and strength of the hindlimbs,40
was
reduced by 40% (P < 0.01) (Figure 2D)
in
Thy1-ApoE4 transgenic mice compared to wild-type littermates. These
results demonstrate that both forelimbs and hindlimbs were affected.
Rod walking, an index for equilibrium and psychomotor integration, was
clearly disturbed in Thy1- ApoE4 transgenic mice. Approximately
45% of the 3-month-old Thy1-ApoE4 transgenic mice fell almost
immediately from the rod, whereas wild-type mice always completed the
task.
Taken together, these results illustrate the sensorimotor defects in mice expressing human ApoE4 in neurons. These defects progressed with age, and in individual mice the decrease in muscle strength was correlated with a decrease in body weight.
Histology and Ultrastructural Examination
Brains and spinal cords of 20 Thy1-ApoE4 transgenic mice, aged 3
to 23 months, were analyzed to determine the neuropathological cause of
their phenotype. Routine H&E staining of brain sections showed
hypertrophy of astrocytes. Immunostaining for GFAP demonstrated
widespread astrogliosis in the hippocampus, neocortex, and amygdala of
Thy1-ApoE4 mice, aged 8 months or older (Figure 3A)
.37
Immunohistochemical
staining for ubiquitin revealed small rounded or irregular inclusions
in the neuropil of the stratum oriens of the hippocampus37
(Figure 3C)
and in the hippocampal fimbria (Figure 3E)
.
Ubiquitin-positive inclusions were also seen in the neocortex and the
amygdala, and in the white matter fiber tracts. The inclusions also
stained with SMI31, a monoclonal antibody that strongly reacts with a
phosphorylated epitope on neurofilament-high, and with NF200, a
phosphorylation-independent monoclonal antibody specific for
neurofilament-high (Figure 3, H and G)
. From sections stained for
neurofilament it became clear that the inclusions in the neuropil were
dilated axons (Figure 3G)
. In addition, brain sections stained for
synaptophysin showed accumulation of synaptophysin in axons in brain
white matter fiber tracts, mainly in the hippocampal fimbria, internal
and external capsule, corticospinal tract, and corpus callosum (Figure 4, B–D)
.
|
|
. White matter fiber tracts
displayed dilated axons staining for ubiquitin (Figure 5, H and I)
and
synaptophysin (Figure 4F)
. The white matter abnormalities were most
obvious in the posterior column (Figure 4F
and Figure 5, H and I
). Mild
gliosis was present in the spinal cord gray and white matter (Figure 5B)
.
|
. The dilation of the
axons varied, with the most dilated axons having diameters almost equal
to the cell body of neurons. Some dilated axons showed degenerative
features and were filled with numerous dense and multivesicular bodies
and electron dense amorphous material (Figure 7, C and D)
. Other
dilated axons were well-preserved and demonstrated accumulated
mitochondria and vesicles (Figure 7, F and G)
. Dilated axons often had
thinned myelin sheaths only 2 to 3 lamellae thick (Figure 7, A and C)
and sometimes the myelin sheath was almost completely absent or
retracted (Figure 7E)
. On ultrastructural examination the axonal
pathology was more prominent than suggested by the immunohistochemical
staining, because many scattered nondilated axons were observed to
contain an increased number of mitochondria and vesicles, or dense
bodies, which were not present in the age-matched nontransgenic mice
(Figure 7A)
. The perikarya and dendrites present in the ultrathin
sections appeared normal.
|
; A, D, and F), indicative of
Wallerian degeneration. The damaged axons were atrophic and often
detached from their myelin sheaths (Figure 8D)
. Rare axons showed
accumulation of electron dense material (Figure 8, B and E)
. The nerves
of the Thy1-ApoE4 transgenic mice showed rare clusters of small axons
with thin myelin sheets, indicative of regeneration (Figure 6A)
.
|
|
, 18-month-old Thy1-ApoE4 transgenic mice (tae-II)
showed obvious grouping of atrophic fibers, diagnostic for neurogenic
muscle atrophy (Figure 6C)
.
As muscle atrophy was evident in motorically affected Thy1-ApoE4
transgenic mice, neuronal loss in the ventral horn of the spinal cord
was examined. No significant reduction in the number of neurons was
measured. In wild-type (n = 3) and Thy1-ApoE4
transgenic mice (n = 4), the average number of
neurons was 55.0 ± 4.1 (mean ± SEM) and 55.2 ± 3.7
(mean ± SEM) per ventral horn, respectively. Rare pale neurons
with loss of Nissl substance were observed in the gray matter of the
spinal cord (Figure 9c)
.
|
Western blotting demonstrated hyperphosphorylation of protein Tau
in spinal cord of Thy1-ApoE4 transgenic mice. Three monoclonal
antibodies (AT8, AT180, and PHF1) directed against known phosphorylated
epitopes on the Tau protein, demonstrated increased protein Tau
phosphorylation in spinal cord in Thy1-ApoE4 transgenic mice, compared
to wild-type littermates (Figure 9a)
. Immunoblotting with TAU5, a
phosphate-independent monoclonal antibody, demonstrated similar
protein Tau levels in transgenic and wild-type mice (Figure 9a)
.
Immunohistochemical staining for protein Tau on spinal cord with the
phosphate-dependent monoclonal antibody AT8 revealed strong reactivity
in axons and weak somatodendritic staining in the neuronal cell bodies
in 12- to 16-month-old wild-type Fvb mice. In 12- to 16-month-old
Thy1-ApoE4 transgenic mice, a similar staining pattern was observed,
except for some rare neurons with increased AT8 immunoreactivity
(Figure 9b, A and C)
. The phosphate-dependent monoclonal antibody AT180
revealed strong somatodendritic staining, which was not increased, in
the Thy1-ApoE4 transgenic mice. These results are in agreement with
results obtained by Western blotting and with the increased AT8
immunoreactivity in brain.37
Neither immunohistochemistry
nor silver staining revealed neurofibrillary tangles.
Phosphorylation of the 200-kd neurofilament subunit (NF-H) remained unchanged in sensorimotor affected Thy1-ApoE4 transgenic mice, compared to wild-type littermates, as analyzed by antibodies SMI31 and NF200 (results not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionExperimental ProceduresResultsDiscussionReferences |
|---|
The present study demonstrated that expression of human ApoE4 in neurons resulted in axonopathy and in disruption of axonal transport, which was accompanied by reduced sensorimotor capacities. The sensorimotor impairment increased with age and neuronal expression level. In contrast, these symptoms were not observed in transgenic mice expressing human ApoE4 in astrocytes.
Examination of brain and spinal cord revealed axonal dilations staining for ubiquitin and neurofilament in the neuropil of the gray matter and in white matter tracts of affected ApoE4 transgenic mice. Moreover, in numerous axons we observed accumulation of synaptophysin, a protein associated with precursors of synaptic vesicles and transported by fast axonal transport. In addition, ultrastructural analysis revealed accumulation of mitochondria and vesicles in dilated axons. Astrogliosis, a general indicator of CNS injury, was observed in neocortex, hippocampus, amygdala, and in spinal cord. This is not surprising because we observed widespread axonopathy. Ultrathin sections revealed degenerated axons corresponding to the immunohistochemically observed accumulation of ubiquitinated proteins. Taken together, these results suggest that the neuronal expression of human ApoE4 contributed to a disruption of axonal transport and to axon degeneration. Surprisingly, no neuronal loss was established, although some neurons with loss of Nissl substance and increased AT8 immunoreactivity were observed. In this context, the observed impairment of axonal transport in AD neurons, their loss of polarity and subsequent degeneration,41,42 and an increased vulnerability to neuronal death in those brain regions that express ApoE in neurons,35,36 is of special interest. Although great care should be taken in extrapolating results from mouse models to humans, it is possible to speculate that the axonal degeneration we observe in our mice represents some form of neuronal degeneration.
Tau proteins are microtubule-associated proteins that help to regulate the dynamics of the microtubular network, for which the degree of their phosphorylation is important.43 In AD, Tau is abnormally hyperphosphorylated and aggregates into paired helical filaments.44 Furthermore, Tau in AD is dislocated mainly to the somatodendritic compartment of neurons,45 whereas protein Tau concentration was reported to be elevated in AD brain.46 Overexpression of protein Tau inhibits kinesin-dependent trafficking of vesicles, mitochondria, and endoplasmic reticulum in neuro-2a cells,47 and causes prominent axonopathy in htau-40 transgenic mice.48 These data strengthen the importance of proper control by posttranslational modification of the microtule-associated protein Tau, its localization, and intracellular concentration, in vivo. We have shown that expression of ApoE4 in neurons results in hyperphosphorylation of protein Tau in brain37 and in spinal cord (this study). As the Tau protein is considered to be a major component of the short cross-bridges between microtubules,49 and as in vitro data have demonstrated that hyperphosphorylation of protein Tau reduced its affinity for microtubules,50,51 hyperphosphorylation of protein Tau is likely to influence cross-bridge formation or stability of microtubules and thereby affect normal axonal transport. In the Thy1-ApoE4 transgenic mice it is possible that because of increased protein Tau phosphorylation, the dynamics of the cytoskeleton are impaired, which could contribute to disruption of the axonal transport and result in the degeneration of axons. In this context, the so-called dying back of axons, beginning from the most distal regions, which would lead to the degeneration of neurons, is of special interest.45,52
At this moment the mechanism whereby expression of human ApoE4 in neurons is leading to protein Tau hyperphosphorylation is not known. Recently, it was shown that extracellular reelin binds to the very low density lipoprotein receptor (VLDLR) and ApoER2 lipoprotein receptors, thereby inducing tyrosine phosphorylation of Disabled-1, which activates tyrosine kinases. The binding was competed for by ApoE-lipoproteins. In addition, mice lacking both VLDLR and ApoER2 showed increased protein Tau phosphorylation, as did reelin knockout mice.53,54 One can speculate that co-expression of human ApoE4 with ApoER2 and VLDLR in neurons could result in aberrant triggering of intracellular signal transduction pathways leading to protein Tau phosphorylation. On the other hand, ApoE particles secreted by neurons might differ from those secreted by astrocytes, which might result in a differential interaction with ApoE-lipoprotein receptors such as VLDLR and ApoER2. Obviously, this hypothesis does not exclude that other mechanisms might be involved.
Recently it was shown that ApoE normally resides within the secretory pathway and that direct expression of ApoE in the cytoplasm of neuro-2a cells is toxic.55 The human ApoE4 protein in our transgenic mice is not expected to escape from the secretory pathway. Although we cannot exclude leakage into the cytosol, and hence a direct cytotoxic effect, this is not considered likely. Indeed, the transgene contains the signal sequence of ApoE, which directs it to the endoplasmic reticulum, and the ApoE4 protein is secreted from neurons derived from Thy1-ApoE4 transgenic mice and from astrocytes derived from GFAP-ApoE4 transgenic mice. In addition, human ApoE4 was found in the CSF of both GFAP-ApoE4 and Thy1-ApoE4 transgenic mice.
Synthetic peptides, as well as a 22-kd thrombin-cleavage fragment derived from ApoE, have been demonstrated to exhibit cytotoxicity56-59 and cause neurite degeneration.60 This process was suggested to be receptor-mediated.58 We did not observe 22-kd ApoE fragments in the ApoE4 transgenic mice with the method that we used, which argues against a role for these fragments in the pathology of the Thy1-ApoE4 transgenic mice.
Neurodegeneration of the central nervous system and peripheral sensory nerve defects have been reported in mice lacking ApoE.61,62 Because endogenous mouse ApoE levels remained unchanged in Thy1-ApoE4 transgenic mice, the observed axon degeneration in the brain and in spinal cord and Wallerian degeneration in sciatic nerve, cannot be attributed to a down-regulation or absence of endogenous ApoE.
Other transgenic mice, generated in our laboratory, that express high levels of unrelated proteins, under the control of the mouse Thy1 gene promoter, did not show axonopathy and remained healthy throughout life. Therefore we exclude the possibility of a general neurotoxic effect by overexpressing high levels of any protein in the central nervous system, under control of the mouse Thy1 gene promoter. In addition, PDGF-ApoE4 transgenic mice, expressing human ApoE4 in neurons at a lower level, also showed signs of axonopathy, at age 2 years or older. The GFAP-ApoE4 transgenic mice on the other hand convincingly demonstrate that even very high levels of human ApoE4 per se, are not toxic or detrimental for normal functioning of the mouse brain.
Other transgenic mice, expressing human ApoE in neurons have been
generated.30-32,61-63
Effects on learning and
exploratory behavior were demonstrated in female NSE-ApoE4 transgenic
mice at the age of 6 months, but no axonopathy was
reported.63
In the PDGF-ApoE4 transgenic mice, we observed
a similar, but less severe phenotype as in the Thy1-ApoE4 transgenic
mice, but only when they aged to 
2 years. It would be interesting to
investigate whether axonopathy and motor impairments are also present
in old NSE-ApoE63
and human ApoE mice.30-32
In conclusion, we demonstrate that expression of human ApoE4 in neurons, in vivo, resulted in axonopathy and impairment of axonal transport, as opposed to expression in nonneuronal cells. We further demonstrated that this is not because of a general toxic effect of human ApoE4 in mouse brain. The current mouse model offers an excellent opportunity to investigate the mechanism by which ApoE could be involved in protein Tau phosphorylation and in neurodegeneration.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the ‘Fonds voor Wetenschappelijk Onderzoek’ (FWO-Vlaanderen), the Interuniversity-Network for Fundamental Research (IUAP), by the special Action Program for Biotechnology of the Flemish government (VLAB/IWT, COT-008), by the Rooms-fund, by Janssen Research Foundation, and by K. U. Leuven Research and Development.
Both I. T. and J. V. D. contributed equally to this work.
Accepted for publication July 20, 2000.
|
References |
|---|
|
TopAbstractIntroductionExperimental ProceduresResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  I. Cuchillo-Ibanez, A. Seereeram, H. L. Byers, K.-Y. Leung, M. A. Ward, B. H. Anderton, and D. P. Hanger Phosphorylation of tau regulates its axonal transport by controlling its binding to kinesin FASEB J, September 1, 2008; 22(9): 3186 - 3195. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Brodbeck, M. E. Balestra, A. M. Saunders, A. D. Roses, R. W. Mahley, and Y. Huang Rosiglitazone increases dendritic spine density and rescues spine loss caused by apolipoprotein E4 in primary cortical neurons PNAS, January 29, 2008; 105(4): 1343 - 1346. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. Xu, A. Bernardo, D. Walker, T. Kanegawa, R. W. Mahley, and Y. Huang Profile and Regulation of Apolipoprotein E (ApoE) Expression in the CNS in Mice with Targeting of Green Fluorescent Protein Gene to the ApoE Locus. J. Neurosci., May 10, 2006; 26(19): 4985 - 4994. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. W. Mahley, K. H. Weisgraber, and Y. Huang Inaugural Article: Apolipoprotein E4: A causative factor and therapeutic target in neuropathology, including Alzheimer's disease PNAS, April 11, 2006; 103(15): 5644 - 5651. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Huang Apolipoprotein E and Alzheimer disease Neurology, January 24, 2006; 66(1_suppl_1): S79 - S85. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Van Dooren, D. Muyllaert, P. Borghgraef, A. Cresens, H. Devijver, I. Van der Auwera, S. Wera, I. Dewachter, and F. Van Leuven Neuronal or Glial Expression of Human Apolipoprotein E4 Affects Parenchymal and Vascular Amyloid Pathology Differentially in Different Brain Regions of Double- and Triple-Transgenic Mice Am. J. Pathol., January 1, 2006; 168(1): 245 - 260. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. M. Harris, W. J. Brecht, Q. Xu, R. W. Mahley, and Y. Huang Increased tau Phosphorylation in Apolipoprotein E4 Transgenic Mice Is Associated with Activation of Extracellular Signal-regulated Kinase: MODULATION BY ZINC J. Biol. Chem., October 22, 2004; 279(43): 44795 - 44801. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. J. Brecht, F. M. Harris, S. Chang, I. Tesseur, G.-Q. Yu, Q. Xu, J. Dee Fish, T. Wyss-Coray, M. Buttini, L. Mucke, et al. Neuron-Specific Apolipoprotein E4 Proteolysis Is Associated with Increased Tau Phosphorylation in Brains of Transgenic Mice J. Neurosci., March 10, 2004; 24(10): 2527 - 2534. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. M. Harris, I. Tesseur, W. J. Brecht, Q. Xu, K. Mullendorff, S. Chang, T. Wyss-Coray, R. W. Mahley, and Y. Huang Astroglial Regulation of Apolipoprotein E Expression in Neuronal Cells: IMPLICATIONS FOR ALZHEIMER'S DISEASE J. Biol. Chem., January 30, 2004; 279(5): 3862 - 3868. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Aoki, T. Uchihara, N. Sanjo, A. Nakamura, K. Ikeda, K. Tsuchiya, and Y. Wakayama Increased Expression of Neuronal Apolipoprotein E in Human Brain With Cerebral Infarction Stroke, April 1, 2003; 34(4): 875 - 880. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Enzinger, S. Ropele, S. Strasser-Fuchs, P. Kapeller, H. Schmidt, B. Poltrum, R. Schmidt, H.-P. Hartung, and F. Fazekas Lower Levels of N-Acetylaspartate in Multiple Sclerosis Patients With the Apolipoprotein E {epsilon}4 Allele Arch Neurol, January 1, 2003; 60(1): 65 - 70. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Utton, J. Connell, A. A. Asuni, M. van Slegtenhorst, M. Hutton, R. de Silva, A. J. Lees, C. C. J. Miller, and B. H. Anderton The Slow Axonal Transport of the Microtubule-Associated Protein Tau and the Transport Rates of Different Isoforms and Mutants in Cultured Neurons J. Neurosci., August 1, 2002; 22(15): 6394 - 6400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. S. B. Goldstein Kinesin molecular motors: Transport pathways, receptors, and human disease PNAS, June 19, 2001; 98(13): 6999 - 7003. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, n =
15) and wild-type (wt
, n = 17) littermates.
D: Traction capacity in Thy1-ApoE4
(tae, n =
15) and wild-type (wt,
n = 17) mice, 3 months old.
E: Spontaneous locomotor activity, measured as distance
traveled in an open field area for 10 minutes, and analyzed in two
consecutive rounds of 5 minutes each (
= 1'
to 5'; 
= 5' to 10'). Thy1-ApoE4
(tae, n =
6), GFAP-ApoE4 (gae,
n = 5), PDGF-ApoE4
(pae, n =
6), and PGK-ApoE4 (pgk,
n = 7) transgenic mice and
wild-type (wt, n =
8) mice, 12 to 16 months old, were used.
F: Prehensile reflex in Thy1-ApoE4
(tae, n =
15) and wild-type (wt,
n = 17) mice, 3 months old.
G: Wire hang time for Thy1-ApoE4
(tae, n =
10) transgenic mice and wild-type
(wt, n =
13) mice, 3 months old, measured in the grid
suspension test. H: Body weight of Thy1-ApoE4
(tae, n =
10) transgenic mice and wild-type
(wt, n =
13) mice of 3 months. All results are expressed
as means (±SEM).
