![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 300, Issue 3, 736-745, March 2002
Department of Pharmacology and Neurosciences, KEIO University School of Medicine, Shinanomachi, Shinjuku-ku, Tokyo Japan
 |
Abstract |
|---|
 Top
 Abstract
 Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Although it has been established that oxidative stress mediates
cytotoxicity by familial Alzheimer's disease (FAD)-linked mutants of
presenilin (PS)1 and that pertussis toxin inhibits cytotoxicity by
FAD-linked N141I-PS2, it has not been determined whether oxidative
stress is involved in cytotoxicity by N141I-PS2 or which pertussis
toxin-sensitive proteins mediate the cytotoxicity. Here we report that
low expression of N141I-PS2 caused neuronal cell death, whereas low
expression of wild-type PS2 did not. Cytotoxicities by low and high
expression of N141I-PS2 occurred through dissimilar mechanisms: the
former cytotoxicity was blocked by a cell-permeable caspase inhibitor,
and the latter was not. Since both mechanisms were sensitive to a
cell-permeable antioxidant, we examined potential sources of reactive
oxygen species in each mechanism, and found that the caspase
inhibitor-sensitive neurotoxicity by N141I-PS2 was likely through NADPH
oxidase and the caspase inhibitor-resistant neurotoxicity by N141I-PS2
through xanthine oxidase. Pertussis toxin greatly suppressed both toxic
mechanisms by N141I-PS2, and only G
o, a neuron-enriched
pertussis toxin-sensitive G protein, was involved in both mechanisms.
We therefore conclude that N141I-PS2 is capable of triggering multiple
neurotoxic mechanisms, which can be inhibited by the combination of
clinically usable inhibitors of NADPH oxidase and xanthine oxidase.
This study thus provides a novel insight into the therapeutic
intervention of PS2 mutant-associated FAD.
| |
Introduction |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Familial
Alzheimer's disease (FAD) is caused by mutations in amyloid precursor
protein (APP), presenilin (PS)1, and PS2 (Shastry and Giblin, 1999
).
However, how these mutant genes cause neural death, the central
abnormality in Alzheimer's disease (AD), has been little understood. A
clue is the finding that expression of FAD mutant APP and PS genes
causes neural cell death in cultures (Guo et al., 1996, 1999; Wolozin
et al., 1996; Yamatsuji et al., 1996; Zhao et al., 1997; Czech et al.,
1998; Luo et al., 1999; Weihl et al., 1999; Hashimoto et al.,
2000b). It has also been found that V642 mutants (V642I/F/G) of
APP and K595N/M596L-APP (NL-APP) induce cytotoxicity through pertussis
toxin (PTX)-sensitive G proteins (Wolozin et al., 1996; Yamatsuji et
al., 1996; Giambarella et al., 1997; Hashimoto et al., 2000b). Wolozin
et al. (1996) found that N141I-PS2, linked with FAD in the Volga German
families, causes PC12 neuronal cell death in a PTX-sensitive manner. It is thus important to clarify the entire array of death signals generated by FAD genes.
The mechanisms underlying mutant APP-induced neurotoxicity have been
considerably defined (Yamatsuji et al., 1996; Giambarella et al., 1997;
Hashimoto et al., 2000b). Also, it has been established that N141I-PS2
causes neuronal cell death (Wolozin et al., 1996; Araki et al., 2000;
Hashimoto et al., 2001a, b). However, little has been known
about the mechanism whereby N141I-PS2 contributes to neuronal cell
death, except for mediation by PTX-sensitive G proteins (Wolozin et
al., 1996). We (Hashimoto et al., 2000b) have recently established a
neuronal cell system in which low levels of single-cell expression of a
cDNA of interest in an ecdysone (EcD)-inducible plasmid are precisely
controlled by treated EcD doses without affecting transfection
efficiency. In this system, it was found that 1) high expression
(
3-fold expression of endogenous APP) of wild-type (wt) APP weakly
induces neuronal cell death resistant to both glutathione-ethyl-ester
(GEE), a well established cell-permeable antioxidant, and Ac-DEVD-CHO
(DEVD), a cell-permeable inhibitor of the apoptosis-mediating caspases;
2) low expression (<3-fold expression of the endogenous protein) of
two FAD-linked mutant APPs (V642I-APP and NL-APP) causes high levels of
GEE/DEVD-sensitive cytotoxicity; 3) low and high expressions of
V642I-APP induce the same cytotoxicity; and 4) high expression of
NL-APP additionally augments the GEE/DEVD-resistant cytotoxicity of
wtAPP. Multiple mechanisms thus underlie neurotoxicity by even two of
the FAD-linked mutants of APP. Using the same system, the present study
was conducted to investigate the molecular mechanisms whereby wild-type
or mutant PS2 kills neuronal cells. Herein we report that distinct sets of different toxic mechanisms underlie cytotoxicity by N141I-PS2 and
wtPS2, and compare them with the toxic mechanisms of APP mutants.
| |
Materials and Methods |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
wtAPP and V642I-APP cDNAs were described previously (Hashimoto
et al., 2000b). wtPS2 and N141I-PS2 cDNAs were kindly provided by Dr.
P. St-George Hyslop (University of Toronto, Canada) and Dr. L. D'Adamio (Albert Einstein College of Medicine, NY), respectively. These PS2 cDNAs were subcloned into pIND (Invitrogen, Carlsbad, CA), an
EcD-inducible plasmid, with sequence confirmation. The pIND-encoded
wtPS2 or N141I-PS2 was designated as pIND-wtPS2 or pIND-N141I-PS2,
respectively. The cDNAs encoding PTX-resistant G protein subunits
(PTXr-G: PTXr-Gi1,
PTXr-Gi2, PTXr-Gi3, and PTXr-Go), described previously (Taussig et
al., 1992), were kindly provided by Drs. R. Taussig (University of
Michigan, MI) and T. Kozasa (University of Texas, Southwestern Medical
Center, TX). EGFP cDNA (pEGFP-N1; CLONTECH, Palo Alto, CA) was also
subcloned to pIND (pIND-EGFP). GEE, apocynin
(4-hydroxy-3-methoxyacetophenone; APO), and diphenyleneiodonium (DPI)
were from Sigma-Aldrich (St. Louis, MO), and PTX from
Calbiochem-Novabiochem (San Diego, CA). Ac-DEVD-CHO was from Peptide
Institute (Osaka, Japan). Ponasterone (Invitrogen) was employed as EcD.
(+) and (
)BOF4272 were provided by Otsuka Pharmaceutical Factory
(Naruto, Japan).
F11 cells were grown in Ham's F-12 plus 18% FBS and antibiotics. F11
cells, the hybrid cells of a rat embryonic day 13 primary cultured
neuron and a mouse neuroblastoma, are one of the best models for
primary cultured neurons, as described previously (Yamatsuji et al.,
1996). F11 cells over-expressing both EcR and RXR (F11/EcR cells) were
established using the coexpression vector pVgRXR (Invitrogen) and
Zeocin selection. For transient transfection of the pIND plasmids, F11/EcR cells were seeded at 7 × 104
cells/well in a 6-well plate and cultured in Ham's F-12 plus 18% FBS
for 12 to 16 h, and were transfected with EcD-inducible pIND
plasmids [1 µg pIND plasmids, 2 µl LipofectAMINE, and 4 µl PLUS
reagent (Invitrogen)] in the absence of serum for 3 h. After subsequent incubation with Ham's F-12 plus 18% FBS for 12 to 16 h, cells were cultured with or without inhibitors in Ham's F-12 plus
10% FBS for 2 h, and EcD was then added to the media (without medium change). Cell mortality was measured by Trypan blue exclusion assay at 72 h after the onset of EcD treatment. Fluorescence of the cells transfected with pIND-EGFP was assessed by transfecting this
plasmid (1 µg of plasmid, 2 µl of LipofectAMINE, and 4 µl of PLUS
reagent) in the absence of serum for 3 h. After subsequent incubation with Ham's F-12 plus 10% FBS for 69 h, fluorescence intensity was assessed by measuring the fluorescence intensity of
randomly chosen cells in each transfection with standardization by the
cell area and calculating the mean ± S.D. of these values for
each transfection. Immunoblot analysis of expressed PS2 constructs was
performed using anti-PS2 antibody 2192 (1:500 dilution; Cell Signaling
Technology, Beverly, MA), which can detect low endogenous levels of PS2
holoprotein at 54 kDa, and the expression was quantified, essentially
according to the method described previously (Hashimoto et al., 2000b).
Mouse primary neuronal cultures, in which >98% of cells were neurons,
were prepared as described previously (Sudo et al., 2000).
Trypan blue exclusion assay was described in detail previously
(Hashimoto et al., 2000b). The basal death rates with or without empty
pIND vector transfection and with or without EcD treatment indicated
the actual fraction of dead cells, but not artificial cell death
occurring after detaching cells, as in situ staining of Trypan
blue-positive cells indicated the presence of similar fractions of
basally occurring cell death (Hashimoto et al., 2000b; Sudo et al.,
2000). It has been demonstrated that the cell mortality assessed by
this assay is quantitatively confirmed by cell viability assay (Wako
Pure Chemicals, Osaka, Japan), using WST-8
[2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt] (Hashimoto et al., 2000a).
All of the experiments described in this study were repeated at least three times with independent transfections and treatments, each of which yielded essentially the same result. Statistical analysis was performed with Student's t test, in which p < 0.05 was assessed as significant.
| |
Results |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Dose-Dependent Expression of the pIND Construct by EcD.
F11/EcR cells were transfected with pIND-EGFP cDNA with subsequent
treatment with various concentrations of EcD. The transfection efficiency was very constant at 65 to 70%, as reported previously (Hashimoto et al., 2000b). The results, shown in the left panel of Fig.
1A, indicate that the cellular expression
of the construct was augmented in a linear relationship with the
treated doses of EcD. Thus, as expected, the pIND construct was
expressed in an EcD-dose-dependent and -proportional manner. It was
also indicated that neither treatment with 100 µM DEVD, 1 mM GEE, nor
1 µg/ml PTX affected EcD-dependent expression of pIND-EGFP (Fig. 1A,
middle panel).
|
), 20 and 40 µM EcD caused ~2 and
~3.5-fold expression of wtPS2 relative to that of wtPS2 induced by 10 µM EcD. This is consistent with our earlier study (Hashimoto et al.,
2000b) that when wtAPP cDNA in pIND was transfected and expressed by
EcD in the same F11/EcR cells, there was a strict linear relationship
between the EcD dose and expressed wtAPP.
Expression of N141I-PS2 holoprotein was not linearly augmented by EcD,
and reached saturation by 10 µM EcD (blue circles in the right
panel of Fig. 1A). Considering the present result that N141I-PS2
induction by 10 µM EcD caused DEVD-sensitive cell death (see below)
and from the literature that PS2 holoprotein is cleaved by
DEVD-sensitive caspases (Vito et al., 1997), we reasoned that in this
system, N141I-PS2 was induced by EcD similarly to the induction of
wtPS2, but degraded through mechanisms involving caspase activation by
expressed N141I-PS2 itself, resulting in certain balanced expression.
In accordance with this idea, in the presence of 100 µM DEVD, an
established cell-permeable inhibitor of caspases, expression of
N141I-PS2 became linear in relationship to EcD concentrations on the
line similar to that of wtPS2 expression (red squares in the right
panel of Fig. 1A). Under the same conditions, expression of wtPS2 was
not affected by DEVD (data not shown). These data suggest that
N141I-PS2 in pIND, like wtPS2 in the same vector, was induced in
proportion to the EcD concentration, and that N141I-PS2, but not wtPS2,
was degraded by activating DEVD-sensitive mechanisms. This is again
consistent with our earlier study that 1) in the same system,
expression of V642I-APP or NL-APP in pIND, both of which can activate
DEVD-sensitive caspases, is not linearly augmented and reaches
saturation by 10 µM EcD; 2) their expression is linearly augmented
by EcD in the presence of DEVD; and 3) wtAPP is expressed
proportionally to the EcD concentration (Hashimoto et al., 2000b).
Combined with the quantitative validity of the employed
EcD/EcR-inducible expression system (No et al., 1996) and the observed
linearity of the EcD-induced EGFP expression in pIND-EGFP-transfected
cells (Fig. 1A, left panel), these data support that the PS2 constructs
(wtPS2 and N141I-PS2) were induced by EcD dose dependently.
In this study, we thus defined low expression as that elicited by 10 µM EcD and high expression as that elicited by 40 µM EcD, as in our
earlier study for inducible APP constructs (Hashimoto et al., 2000b).
In this system, the low expression level of PS2 holoproteins (wtPS2 and
N141I-PS2) was comparable to the endogenous expression level of the PS2
holoprotein in mouse primary cultured neurons (Fig. 1B). Although
anti-PS2 antibody 2192 was against the C terminus of PS2 and can
recognize both the holoprotein and the C-terminal fragment (CTF) of
PS2, we were not able to clearly specify the CTF (deduced size, 23 kDa)
of expressed PS2 proteins because the 2192 antibody was the only one
available that could detect endogenous levels of PS2 (no other
antibodies tested could detect PS2 holoprotein expression by 40 µM
EcD in the present system), and F11/EcR cells express quite a few
transfection-independent proteins recognized by this antibody at 20 to
25 kDa. However, this antibody clearly detected the 54-kDa holoprotein
of PS2. In addition, comparison with the immunoblot result from primary neurons suggests that the strong 23-kDa immunoreactive band was apparently the CTF of PS2. It was thus likely that in this system, the
quantity of the CTF of PS2 was not critically affected by transfected
induction of PS2 constructs (wtPS2 and N141I-PS2); only the quantity of
the holoproteins was controlled by the induction. Also, the observed
linear relationship between EcD and expressed PS2 holoproteins (wtPS2
in the presence or absence of DEVD and N141I-PS2 in the presence of
DEVD), as linear as that between EcD and EGFP in the same vector,
suggests that the cleavage of full-length PS2 was not affected by EcD.
Effect of Low and High Expression of N141I-PS2 on Neuronal Cell
Death.
When N141I-PS2 cDNA was induced by EcD in transfected
F11/EcR cells, cell death robustly occurred in a manner dependent on the concentration of EcD (Fig. 2). In
pIND-N141I-PS2-transfected cells, 10 µM EcD caused death in nearly
60% of cells in a saturated manner after 72 h. In contrast, the
vehicle ethanol (EtOH) caused no increase in cell mortality in either
case. Treatment of nontransfected F11/EcR cells or vector-transfected
F11/EcR cells with or without EcD resulted in low cell mortality
(~10%) for 72 h, which was estimated to be the basal death rate
of these cells. The top panel of Fig. 2 indicates the stability of low
death rates under the negative conditions. In all dose-response
experiments shown in each figure in this study, we performed the
experiments measuring cell mortality in the presence or absence of 40 µM EcD without transfection and in the presence or absence of 40 µM
EcD with empty pIND transfection, both of which were constantly as low as the basal cell mortality in the absence of EcD with pIND-PS construct transfection (Fig. 2, top panel). The low basal death rates
after lipofection and small variations in the stimulated death rates
were mainly attributed to the short lipofection period and subsequent
serum treatment in the present protocol.
|
), it followed that expression of
N141I-PS2 by 10 µM EcD induced death in most of the transfected cells after 72 h. Expression of wtPS2 resulted in different
dose-response curves for death (Fig. 2, black filled circles). EcD
caused little death in pIND-wtPS2-transfected cells at 
10 µM, and
augmented the death rates dose dependently only at 20 µM. Although
EtOH caused no increase in cell mortality in either case (Fig. 2, black open circles), expression of wtPS2 by 30 µM EcD exerted significant cytotoxicity, but to lesser degrees than the cytotoxicity by the same
concentrations of EcD in cells transfected with pIND-N141I-PS2. Thus,
low expression of N141I-PS2 robustly caused cell death, whereas wtPS2
was toxic at only high expression.
Effect of Ac-DEVD-CHO on Cytotoxicity by N141I-PS2.
We next
investigated whether DEVD affects cytotoxicity by N141I-PS2. The
dose-response relationship for EcD-stimulated cytotoxicity in
pIND-N141I-PS2-transfected cells was examined in the presence or
absence of 100 µM DEVD. Mortality of cells transfected with or
without pIND treated with or without EcD in the presence of 100 µM
DEVD was ~10% (Fig. 3, left top
panel). The curve of N141I-PS2-induced cytotoxicity in the presence of
DEVD was intermediate between the cytotoxicity curve of N141I-PS2 and
that of wtPS2. Cytotoxicity by low expression of N141I-PS2 was blocked
by DEVD, whereas cytotoxicity by high expression of N141I-PS2 was
DEVD-resistant (Fig. 3, left panel). These data indicate that
cytotoxicity by N141I-PS2 consists of two components in addition to
wtPS2 cytotoxicity: DEVD-sensitive cytotoxicity by low expression of
N141I-PS2 and DEVD-resistant cytotoxicity by its high expression.
|
Effect of GEE on Cytotoxicity by N141I-PS2. The sensitivity of N141I-PS2-induced cytotoxicity was different for GEE. The dose-response curve of N141I-PS2 cytotoxicity was altered sharply downward by 1 mM GEE and became virtually identical to the dose-response curve of wtPS2 cytotoxicity (Fig. 3, right panel). This result suggests that N141I mutation-specific cytotoxicity was overlaid on wtPS2 cytotoxicity and was sensitive to GEE. Combining these data with the aforementioned DEVD data further suggests that in addition to the GEE/DEVD-resistant mechanism of wtPS2 cytotoxicity, N141I-PS2 triggered two distinct cytotoxic mechanisms, both different from the mechanism for wtPS2-induced cytotoxicity. One, activated by low expression of N141I-PS2, was GEE/DEVD-sensitive. Another, activated by its high expression, was GEE-sensitive/DEVD-resistant.
Effect of PTX on Cell Death by N141I-PS2.
We next investigated
whether PTX, an inhibitor of heterotrimeric G proteins, affects
cytotoxicity by N141I-PS2, as reported by Wolozin et al. (1996), and if
so, which mechanism for cell death, by low or high expression, is
inhibited by PTX. When F11/EcR cells transfected with pIND-N141I-PS2
were treated with EcD in the presence of 1 µg/ml PTX, EcD-stimulated
cytotoxicity was drastically suppressed, even below the dose-response
curve of wtPS2 cytotoxicity (Fig. 4, left
panel). These data suggest that cytotoxicity by N141I-PS2 (both low and
high expression) is PTX-sensitive, and that cytotoxicity by wtPS2 would
also be sensitive to PTX.
|
Effect of Specific Xanthine Oxidase Inhibitors on Cell Death by
N141I-PS2.
Not only to confirm the GEE-sensitive cytotoxicity by
N141I-PS2, but also to specify the underlying mechanism, we examined the effect of oxypurinol (OXYP), a cell-permeable xanthine oxidase inhibitor (XOI). Xanthine oxidase (XO) has been established to be a
source of the reactive oxygen species (ROS) involved in various signaling pathways. The left panel of Fig.
5 indicates that the dose-response curve
of N141I-PS2 cytotoxicity in the presence of 100 µM OXYP was
virtually identical to its dose-response curve in the absence of OXYP.
This result demonstrates that GEE/DEVD-sensitive cytotoxicity by
N141I-PS2 was resistant to OXYP, suggesting that the GEE target source
of ROS in this mechanism was different from XO. Likewise, 100 µM OXYP
had no effect on GEE/DEVD-sensitive cytotoxicity by V642I-APP under the
same conditions (data not shown), consistent with the notion that
N141I-PS2 and V642I-APP share the same mechanism for the
GEE/DEVD-sensitive cytotoxicity.
|
), it was unlikely that
OXYP suppressed this GEE-sensitive mechanism by high expression of
N141I-PS2 through such a scavenging effect, because the same
concentration of OXYP did not affect other GEE-sensitive mechanisms
activated by V642I-APP (data not shown) or low expression of N141I-PS2
(Fig. 5, left panel).
To confirm that the source of ROS involved in the
GEE-sensitive/DEVD-resistant cytotoxicity by N141I-PS2 is XO, we
examined the effect of ()BOF4272, a specific inhibitor of XO, which
has no direct scavenging effect on superoxide anions (Suzuki et al., 1998). We also used (+)BOF4272, an inactive enantiomer as a negative control that does not block XO (Suzuki et al., 1998). In the presence of 100 µM DEVD, the dose-response curve of N141I-PS2 cytotoxicity was
suppressed by 100 nM ()BOF4272 down to the dose-response curve of
wtPS2 cytotoxicity (Fig. 5, right panel). The same concentration of
(+)BOF4272 had no effect under the same conditions. These data demonstrate that GEE-sensitive/DEVD-resistant death by high expression of N141I-PS2 is sensitive to XOIs.
Effect of Specific NADPH Oxidase Inhibitors on Cell Death by
N141I-PS2.
To further specify the underlying mechanism for
GEE-sensitive death by N141I-PS2, we examined the effect of APO
(acetovanillone or apocynin), a cell-permeable NADPH oxidase inhibitor
(Dodd-O and Pearse, 2000). NADPH oxidase is the major enzyme that
generates superoxide in a number of systems, including a
stress-transducing system in neurons (Noh and Koh, 2000;
http://www.jneurosci.org/cgi/content/full/20/23/RC111), and APO is a
specific inhibitor of this enzyme. The basal death rates of F11/EcR
cells transfected with or without pIND vector in the presence or
absence of EcD were not affected by either 100 µM DEVD or 300 µM
APO (Fig. 6, upper panels), or by a
combination of the two (data not shown). In the presence of 300 µM
APO, the dose-response curve of N141I-PS2 cytotoxicity was greatly
suppressed and became virtually identical to its dose-response curve in
the presence of DEVD (Fig. 6, left lower panel). This result indicates that GEE/DEVD-sensitive cytotoxicity by N141I-PS2 is APO-sensitive. It
is thus highly likely that the GEE target source of ROS in this
mechanism is NADPH oxidase, because NADPH oxidase is the only enzyme
that is inhibited by 300 µM APO, which cannot affect any of the other
superoxide-generating enzymes (t'Hart and Simons, 1992). As was the
case with N141I-PS2 cytotoxicity, GEE/DEVD-sensitive cytotoxicity by
V642I-APP was suppressed by APO to the level of wtAPP cytotoxicity
(Fig. 6, right bottom panel). In contrast, L-NAME, an
inhibitor of NO synthase, had no effect on cytotoxicity by N141I-PS2 or
V642I-APP in the same system (data not shown), confirming the action
specificity of APO. These data lend additional credence to the notion
that N141I-PS2 and V642I-APP share the same mechanism for the
GEE/DEVD-sensitive cytotoxicity.
|
). As shown in Fig. 6 (left bottom panel), the
cytotoxicity curve of N141I-PS2 was suppressed by 100 µM DPI down to
a level equivalent to the cytotoxicity curves of N141I-PS2 in the
presence of APO or DEVD. This finding suggests that treatment with DPI
precisely reproduced the action of APO. This was also the case with DPI
suppression of V642I-APP-induced cell death (Fig. 6, right bottom
panel). DPI suppressed the cytotoxicity curve of V642I-APP to a level
similar to the curve suppressed by APO. These data confirm that NADPH
oxidase is involved in GEE/DEVD-sensitive death by low expression of
N141I-PS2 and by low and high expression of V642I-APP.
It should be emphasized that the suppression of N141I-PS2 cytotoxicity
by APO or DPI was intermediate between the cytotoxicity curve of
N141I-PS2 and that of wtPS2. In contrast, APO or DPI suppressed
V642I-APP cytotoxicity to the level of wtAPP cytotoxicity. This
observation is again consistent with the notion that in addition to the
wild-type construct cytotoxicity, cytotoxicity by N141I-PS2 has two
components and cytotoxicity by V642I-APP has but a single component.
Effect of PTX-Resistant Mutants of G on Cell Death by N141I-PS2
in the Presence of PTX.
As explained above, cytotoxicity by
N141I-PS2 (both low and high expression) was inhibited by PTX. We thus
investigated which PTX-sensitive G proteins are involved. The
PTX-sensitive G proteins potentially involved include
Go, Gi1,
Gi2, and Gi3. It was expected that Go mediates cytotoxicity by low
expression of N141I-PS2, because cytotoxicity by V642I-APP is mediated
by Go (Yamatsuji et al., 1996; Giambarella et
al., 1997), and cytotoxicity by low expression of N141I-PS2 shares a
number of pharmacological characteristics (GEE-sensitive,
DEVD-sensitive, APO/DPI-sensitive, XOI-resistant, and
L-NAME-resistant) with cytotoxicity by V642I-APP. In
contrast, it remained totally unknown which PTX-sensitive proteins
mediate cell death by high expression of N141I-PS2.
subunits (PTXr-G). They have a mutation at the fourth Cys from the
extreme C terminus and can mediate the originally coupled receptor
signals, even in the presence of PTX (Taussig et al., 1992). We
transfected pIND-N141I-PS2 or pIND-V642I-APP with or without each of
the four PTXr-G cDNA into F11/EcR cells, treated the transfected
cells with low or high doses of EcD in the presence of PTX, and
examined whether cell death occurred (Fig.
7). In the presence or absence of PTX,
single transfection with each PTX-resistant mutant cDNA did not
significantly increase cell mortality in the absence of cotransfected
pIND-N141I-PS2 or pIND-V642I-APP (Fig. 7A), as described previously
(Hashimoto et al., 2000a). However, in cells transfected with
pIND-N141I-PS2 or pIND-V642I-APP, 10 µM EcD killed cells
cotransfected with PTX-resistant mutant Go
(Fig. 7A). Under the same conditions, however, 10 µM EcD did not
augment mortality of cells cotransfected with any of the other
PTX-resistant mutants of the Gi family members.
These data demonstrate that low expression of N141I-PS2 or V642I-APP causes neuronal cell death only through Go,
but not other Gi family G proteins.
|
o, but not with any other
PTX-resistant Gi mutants. Each PTXr-G protein
was expressed similarly, and its expression was not affected by PTX
(data not shown), as described previously (Hashimoto et al., 2000a).
These findings provide definite evidence that 1)
Go mediates cytotoxicities by both low and
high expression of N141I-PS2; 2) none of Gi
can mediate cell death by either low or high expression of N141I-PS2;
and 3) cytotoxicity by V642I-APP is mediated by
Go, but not other Gi
family members, as suggested by earlier studies (Nishimoto et al.,
1993; Yamatsuji et al., 1996; Brouillet et al., 1999).
Therefore, two possibilities arose. One was that the observed
Go-mediated cell death by high expression of
N141I-PS2 is only through APO-sensitive cytotoxicity due to
Go-mediated cell death by low expression of
N141I-PS2. The other was that the Go-mediated cell death by high expression of N141I-PS2 is a combination of two
cytotoxic mechanisms: APO- and OXYP-sensitive cytotoxicities. We
therefore analyzed whether the
PTXr-Go-restored cytotoxicity by high
expression of N141I-PS2 consists only of APO-sensitive cytotoxicity or
consists of both APO- and OXYP-sensitive cytotoxicities, in addition to
wtPS2 cytotoxicity. As shown in Fig. 7B, to inhibit the
PTXr-Go-restored cytotoxicity by high
expression of N141I-PS2 to the level of wtPS2 cytotoxicity, both APO
and OXYP were required. This result demonstrates that high expression
of N141I-PS2 causes both APO- and OXYP-sensitive cytotoxicities through
Go. In contrast, the
PTXr-Go-restored cytotoxicity by high
expression of V642I-APP was simply sensitive to APO, which is in good
agreement with the notion that both low and high expressions of
V642I-APP solely cause APO-sensitive cell death through
Go. Therefore, it is highly likely that in
addition to activation of the Go/NADPH oxidase pathway by low expression of N141I-PS2, high expression of N141I-PS2 generates certain additional signals that allow
Go activation to stimulate XO.
| |
Discussion |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
We have herein shown that wtPS2 weakly and N141I-PS2 strongly
induce neuronal cell death, but through distinct sets of toxic mechanisms. This was somewhat unexpected, as the well known analogy with the oncogenic activation of the epidermal growth factor
receptor/c-ErbB to v-ErbB prompted us to speculate that
N141I-PS2-induced cytotoxicity might be based on the quantitatively
amplified function of wtPS2. However, the present observation closely
coincides with more recent information that high expression of wtAPP
weakly induces cell death in a manner resistant to GEE and DEVD, and
low expression of both V642I-APP and NL-APP strongly induces
GEE/DEVD-sensitive cytotoxicity, qualitatively different from the
cytotoxicity by wtAPP (Hashimoto et al., 2000b). This study therefore
indicates, for the first time, that N141I-PS2 causes GEE/DEVD-sensitive
death at low expression and that cell death by high expression of
N141I-PS2 is the combination of three mechanistically different cell
deaths: GEE/DEVD-sensitive death (by low to high expression of
N141I-PS2), GEE-sensitive/DEVD-resistant death (caused only by high
expression of N141I-PS2), and GEE/DEVD-resistant death (by wtPS2).
It should be noted here that transgenic over-expression of N141I-PS2
does not result in global neuronal loss in mice (Oyama et al., 1998).
Because expression of N141I-PS2 can induce cell death in isolated
rodent neuronal cells
both immortalized cell lines and primary neurons
(Wolozin et al., 1996; Araki et al., 2000; the present study)a simple
interpretation for this discrepancy is, as Martin and colleagues
(Fukuchi et al., 1993) argued previously, that factors suppressing
neurotoxicity of N141I-PS2 are present in vivo, although this
hypothesis must be directly examined in transgenic models. It is also
noted that the present study was performed in only one immortalized
neuronal cell line. However, the principal observations that N141I-PS2
causes neuronal cell death and that PTX inhibits N141I-PS2
neurotoxicity have been reported using a different neuronal cell line
(Wolozin et al., 1996) and primary cortical neurons (Araki et al.,
2000), suggesting that observations noted in this study are not limited
to F11/EcR cells.
The provided evidence further indicates that NADPH oxidase is involved
in GEE/DEVD-sensitive cytotoxicity by low expression of N141I-PS2 and
that XO is involved in GEE-sensitive/DEVD-resistant cytotoxicity by
high expression of N141I-PS2, although we could not totally exclude the
possibility that XO-like enzymes, such as aldehyde oxidase (Terao et
al., 2000), may contribute to the latter cytotoxicity. In support of
the established theory, NADPH oxidase is a major source of the
oxygen-radical production not only in neutrophils but in various
tissues, including neuronal cells (Noh and Koh, 2000). XO is another
key enzyme in ROS formation playing a significant role in cell
oxidative stress in neurons (Canas, 1999; Atlante et al., 2000).
Accordingly, we have confirmed the presence of subunits of NADPH
oxidase and XO in F11 cells (data not shown). Combined with the
observed XOI-resistant, APO/DPI-sensitive cytotoxicity by V642I-APP and
with earlier studies on NL-APP-induced cytotoxicity (Hashimoto et al.,
2000b), three conclusions are highly likely. First, the N141I mutation
endows PS2 with the ability to trigger two distinct mechanisms for
cytotoxicity, neither of which wtPS2 can activate. Second, the NADPH
oxidase-mediated cytotoxic mechanism triggered by low expression of
N141I-PS2 is shared with the cytotoxic mechanism by low expression of
V642I-APP and NL-APP. Finally, the cytotoxic mechanism by high
expression of N141I-PS2, mediated by XO, differs from that by high
expression of V642I-APP and NL-APP. It has been shown that cytotoxicity
by low expression of V642I-APP and NL-APP is very probably relevant to
the cause of FAD (Hashimoto et al., 2000b). Therefore, NADPH
oxidase-mediated GEE/DEVD-sensitive cytotoxicity by N141I-PS2, shared
by both APP mutants, would be more important as a cause of FAD than
other forms of cytotoxicity.
This is also the first study describing whether and how antioxidants
attenuate neurotoxicity by N141I-PS2. Guo et al. (1999) found that
death-stimulating action of FAD mutant PS1 is inhibited by
antioxidants. Wolozin et al. (1996) had established neurotoxicity by
N141I-PS2 before neurotoxicity by FAD mutant PS1 was identified. Nonetheless, no study has so far analyzed the antioxidant action on
neural death by FAD mutant PS2. The present study not only addresses
this question, but indicates, in combination with our earlier study
(Hashimoto et al., 2000b), that there are two different antioxidant
targets, NADPH oxidase and XO, in N141I-PS2 cytotoxicity, as discussed
above, and three different FAD genes (V642I-APP, NL-APP, and N141I-PS2)
can induce neurotoxicity through different combinations of at least
three cytotoxic mechanisms, most of which can be inhibited by
antioxidants (Fig. 8).
|
Wolozin et al. (1996) found that PTX inhibits neurotoxicity by
N141I-PS2. Yamatsuji et al. (1996) reported PTX-sensitivity of
V642I-APP neurotoxicity and specified the implicated G protein to be
Go. Nonetheless, the PTX-sensitive G protein
implicated in N141I-PS2 cytotoxicity remained undetermined. Using
PTX-resistant mutants of G proteins, this study has identified the
implicated G protein. Figure 8 summarizes the most likely mechanisms
for cytotoxicity caused by N141I-PS2 and APP mutants. Low expression of
N141I-PS2 induces NADPH oxidase-mediated GEE/DEVD-sensitive cytotoxicity through the PTX-sensitive G protein
Go, as is the case with cytotoxicity by low
expression of V642I-APP and NL-APP. Whereas low to high expression of
V642I-APP solely activates this pathway and low expression of NL-APP
also does so, high expression of NL-APP induces GEE/DEVD-resistant
death (Hashimoto et al., 2000b). The data that
Go-mediated cytotoxicity is shared by low expression of N141I-PS2 and V642I-APP (Fig. 7A) are consistent with
earlier reports that V642I-APP causes cell death through Go, but not Gi as well as
that N141I-PS2 kills neuronal cells in a PTX-sensitive manner (Wolozin
et al., 1996; Yamatsuji et al., 1996; Giambarella et al., 1997).
High expression of N141I-PS2 causes XO-mediated
GEE-sensitive/DEVD-resistant cytotoxicity. This pathway is again
mediated by Go, but not by
Gi, and nevertheless is not activated by high expression of V642I-APP (Fig. 7B). These findings suggest that only
high expression of N141I-PS2 allows activated Go
to stimulate XO. A simple interpretation would be that
Go activated by low expression of N141I-PS2like
that by low expression of V642I-APP and NL-APPonly gives rise to
NADPH oxidase activation, but that highly expressed N141I-PS2 may not
only activate Go but also be able to act like a
chaperone of Go in accession to XO, allowing activated Go to act on this enzyme. In support,
chaperone-like function of PS has been postulated thus far (Gething,
2000). Alternatively, the effect of highly expressed N141I-PS2 on the
action of Go may not be direct. Investigation is
necessary to clarify how N141I-PS2 regulates Go,
because this protein does not belong to a G-protein-coupled receptor
group consisting of a seven-transmembrane structure. It should be noted
that some proteins without such a structure, including GAP-43, the
IGF-II/mannose 6-phosphate receptor, the epidermal growth factor
receptor, and the transmembrane galactosyltransferase, can directly
regulate the activity of specific G proteins (Murayama et al., 1990;
Strittmatter et al., 1990; Gong et al., 1995; Sun et al., 1995). It
should also be noted that the Go/NADPH
oxidase-mediated cytotoxicity by V642I-APP is enhanced by extracellular
A
(Hashimoto et al., 2000b). Because this pathway is shared with
N141I-PS2, it is very likely that extracellular A potentiates the
cytotoxicity by N141I-PS2, pointing to a way in which A contributes
to neurotoxic mechanisms in AD.
Finally, it would be important that both APO and OXYP be clinically
usable. Some XOIs, including OXYP, are already used clinically for
patients with gout. APO, a methoxy-substituted catechol derived from
the root extract of the medicinal herb Picrorhiza kurroa, has been
shown to confer protection in animal models of arthritis and various
causes of lung injury through the inhibition of NADPH oxidase in
polymorphonuclear leukocytes by reacting with thiol groups required for
enzyme assembly (t'Hart et al., 1991; Stolk et al., 1994; Dodd-O and
Pearse, 2000). APO can be provided as a water extract. Therefore, the
present study, which points to a novel possibility that the combination
of APO and XOI could be effective in preventing neurotoxicity by
N141I-PS2, may help to open a new avenue toward the development of
medical prevention for mutation-positive asymptomatic family members of
this type of FAD. Considering the fact that oxidative stress could also play a toxic role in sporadic AD (Albers and Beal, 2000), the proposed
combined antioxidant therapy would provide an effective method to treat
at least certain factions, if not all, of sporadic AD.
| |
Acknowledgments |
|---|
We thank Dr. Mark C. Fishman for F11 neuronal hybrids; Drs.
Masaki Kitajima, Sadakazu Aiso, John T. Potts Jr., and Mr. and Mrs. Y. Tamai for indispensable support; Mr. Yusuke Tomita and Dr. Yoh-ichiro
Abe for cooperation; Ms. Takako Hiraki, Ms. Kazumi Nishihara, and Dr.
Dovie Wylie for expert technical assistance; Drs. R. Taussig and T. Kozasa for the cDNAs encoding GPT; Dr. M. Suematsu for indispensable
cooperation and advice; and Otsuka Pharmaceutical Factory for providing
() and (+)BOF4272. We are especially indebted to Drs. E. Garattini
and T. Nishino for XO cDNA and indispensable discussion; Drs. P. St-George Hyslop and L. D'Adamio for PS2 constructs; Dr. Zongjun Shao
for indispensable cooperation; and all members of the Department of
Pharmacology and Neurosciences at KEIO University School of Medicine
for essential assistance.
| |
Footnotes |
|---|
Accepted for publication November 6, 2001.
Received for publication August 20, 2001.
This work was supported in part by grants from Tomy Medico, Noevir, Japan Foundation for Neuroscience and Mental Health (Y. H.), Ono Medical Research Foundation, Keio Gijuku Academic Development Funds (Y. H. and Y. K.), and the Ministry of Education, Culture, Sports, Science, and Technology of Japan.
Address correspondence to: Dr. Ikuo Nishimoto, Department of Pharmacology and Neurosciences, KEIO University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. E-mail: nisimoto{at}sc.itc.keio.ac.jp
| |
Abbreviations |
|---|
FAD, familial Alzheimer's disease;
PS, presenilin;
APP, amyloid precursor protein;
NL-APP, K595N/M596L-APP;
PTX, pertussis toxin;
GEE, glutathione-ethyl-ester;
DEVD, Ac-DEVD-CHO;
APO, apocynin;
DPI, diphenyleneiodonium;
XO, xanthine oxidase;
XOI, XO
inhibitor;
ROS, reactive oxygen species;
OXYP, oxypurinol;
CTF, C-terminal fragment;
EGFP, enhanced green fluorescent protein;
wt, wild-type;
FBS, fetal bovine serum;
EcD, ecdysone;
EtOH, ethanol;
PTXr-G, PTX-resistant mutant of G protein subunit
Gi1, Gi2, Gi3, or
Go;
L-NAME, N
-nitro-L-arginine methyl
ester.
| |
References |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
complex mediated apoptosis by familial Alzheimer's disease mutant of APP.
EMBO (Eur Mol Biol Organ) J
16:
4897-4907[CrossRef][Medline].-1,4-galactosyltransferase on the surface of sperm.
Science (Wash DC)
269:
1718-1721-peptide.
Neuroreport
8:
379-383[Medline]..
Proc Natl Acad Sci USA
98:
6336-6341-amyloid precursor protein.
Mol Cell Neurosci
16:
708-723[CrossRef][Medline].
-conotoxin-sensitive calcium current by distinct G proteins.
Neuron
8:
799-809[CrossRef][Medline].This article has been cited by other articles:
|
|
 |
|
  K. Kanekura, Y. Hashimoto, Y. Kita, J. Sasabe, S. Aiso, I. Nishimoto, and M. Matsuoka A Rac1/Phosphatidylinositol 3-Kinase/Akt3 Anti-apoptotic Pathway, Triggered by AlsinLF, the Product of the ALS2 Gene, Antagonizes Cu/Zn-superoxide Dismutase (SOD1) Mutant-induced Motoneuronal Cell Death J. Biol. Chem., February 11, 2005; 280(6): 4532 - 4543. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Kanekura, Y. Hashimoto, T. Niikura, S. Aiso, M. Matsuoka, and I. Nishimoto Alsin, the Product of ALS2 Gene, Suppresses SOD1 Mutant Neurotoxicity through RhoGEF Domain by Interacting with SOD1 Mutants J. Biol. Chem., April 30, 2004; 279(18): 19247 - 19256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Hashimoto, O. Tsuji, K. Kanekura, S. Aiso, T. Niikura, M. Matsuoka, and I. Nishimoto The Gtx Homeodomain Transcription Factor Exerts Neuroprotection Using Its Homeodomain J. Biol. Chem., April 16, 2004; 279(16): 16767 - 16777. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Abe, K. Kouyama, T. Tomita, Y. Tomita, N. Ban, M. Nawa, M. Matsuoka, T. Niikura, S. Aiso, Y. Kita, et al. Analysis of Neurons Created from Wild-Type and Alzheimer's Mutation Knock-In Embryonic Stem Cells by a Highly Efficient Differentiation Protocol J. Neurosci., September 17, 2003; 23(24): 8513 - 8525. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Hashimoto, T. Niikura, T. Chiba, E. Tsukamoto, H. Kadowaki, H. Nishitoh, Y. Yamagishi, M. Ishizaka, M. Yamada, M. Nawa, et al. The Cytoplasmic Domain of Alzheimer's Amyloid-{beta} Protein Precursor Causes Sustained Apoptosis Signal-Regulating Kinase 1/c-Jun NH2-Terminal Kinase-Mediated Neurotoxic Signal via Dimerization J. Pharmacol. Exp. Ther., September 1, 2003; 306(3): 889 - 902. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
, significant versus basal cell death rates
in the upper panel and cell death rates by pIND-N141I-PS2 transfection
in the absence of EcD. 
, significant and not significant
increases, respectively, versus cell death rates by pIND-FAD gene
transfection in the presence of indicated concentrations of EcD in the
presence of PTX without transfection of PTXr-G