(Received for publication, July 15, 1996, and in revised form, September 23, 1996)
From the Mitsubishi Kasei Institute of Life Sciences, 11 Minamiooya, Machida-shi, Tokyo 194, Japan
We show here that amyloid peptide1-42 (A
1-42) may play a key role in
the pathogenesis of the cholinergic dysfunction seen in Alzheimer's
disease (AD), in addition to its putative role in amyloid plaque
formation. A
1-42 freshly solubilized in water (non-aged
A
1-42), which was not neurotoxic without preaggregation, suppressed acetylcholine (ACh) synthesis in cholinergic neurons at very low concentrations (10-100 nM), although
non-aged A
1-40 was ineffective. Non-aged
A
1-42 impaired pyruvate dehydrogenase (PDH) activity by
activating mitochondrial
protein kinase I/glycogen synthase
kinase-3
, as we have already shown in hippocampal neurons (Hoshi,
M., Takashima, A., Noguchi, K., Murayama, M., Sato, M., Kondo, S.,
Saitoh, Y., Ishiguro, K., Hoshino, T., and Imahori, K. (1996)
Proc. Natl. Acad. Sci. U. S. A. 93, 2719-2723). Neither
choline acetyltransferase activity nor choline metabolism was affected.
Therefore, the major cause of reduced ACh synthesis was considered to
be an inadequate supply of acetyl-CoA owing to PDH impairment. Soluble
A
1-42 increases specifically in AD brain (Kuo, Y.-M.,
Emmerling, M. R., Vigo-Pelfrey, C., Kasunic, T. C., Kirkpatrick, J. B.,
Murdoch, G. H., Ball, M. J., and Roher, A. E. (1996) J. Biol. Chem. 271, 4077-4081). This increase in soluble
A
1-42 may disturb cholinergic function, leading to the
deterioration of memory and cognitive function that is characteristic
of AD.
Alzheimer's disease (AD)1 is a progressive dementia. One of the most consistent abnormalities in AD brain is a severe loss of basal forebrain cholinergic neurons and cortical cholinergic innervations (3-11), together with other pathological features (amyloid plaques and neurofibrillary tangles) (12). The degree of cognitive dysfunction in AD patients is significantly correlated with decline in choline acetyltransferase (ChAT) activity, a cholinergic marker, and loss of cholinergic neurons (7, 13-15). Evidence from behavioral, pharmacological, neurochemical, and lesion studies further supports a role of these cholinergic deficits in the memory disturbances in normal aging and AD (16, 17). Cholinergic therapies based on acetylcholine esterase (AChE) inhibition have produced small, but well-attested improvements in AD patients (16). These results suggest a critical involvement of cholinergic deficiency in the impaired learning and memory in AD. However, the basis for this cholinergic deficiency is unclear, and whether amyloid plaques or neurofibrillary tangles are related to the cholinergic deficits is so far unknown.
Amyloid plaques are mainly composed of insoluble aggregates of amyloid
peptides (A
s). In our previous work, aggregated A
25-35, an active portion of A
, was shown to
suppress pyruvate dehydrogenase (PDH) activity in hippocampal neuronal cultures and to result in energy failure, which probably contributes to
neuronal death, together with abnormal phosphorylation of
(1).
Since PDH provides acetyl-CoA for acetylcholine (ACh) synthesis in
cholinergic neurons, the above finding suggests the direct involvement
of A
in the cholinergic abnormality in AD by inhibiting PDH
activity. Among A
species generated in vivo from amyloid
precursor protein (APP), A
1-40 is a major product of
soluble A
s and is constitutively secreted in culture medium and
cerebrospinal fluid, while A
1-42 is produced in a
lesser amount (18, 19). Accumulating evidence indicates that increased
production of A
1-42 is the primary event that leads to
the formation of amyloid plaques in AD. Therefore, we focused on the
effect of A
1-42 on cholinergic function using a primary
neuronal culture as an in vitro model system. We report here
that A
1-42 freshly solubilized in water impairs ACh
synthesis in cholinergic neurons without affecting neuronal survival.
Recently, the pool of oligomeric water-soluble A
1-42
was found to be uniquely elevated in AD compared with normal brain (2).
Taken together, the findings suggest a possible role of soluble
A
1-42 in AD pathogenesis in addition to its
contribution to amyloid plaque formation.
Synthetic A1-40 peptide was
synthesized and purified as described previously (20). Synthetic
A
1-42 peptide was obtained from Bachem (Torrance, CA).
Lyophilized peptides were stored at
20 °C. Non-aged A
s were
prepared by freshly dissolving the lyophilized peptides in autoclaved
MilliQ water to 200 µM just before use. Aged
A
1-42 was produced by dissolving the peptide in
autoclaved MilliQ water to 350 µM and incubating at
37 °C for 7 days as previously reported (21).
Primary cultures from rat septum regions,
which include septal and basal forebrain cholinergic neurons, were
prepared and plated (1). After 3 days, the medium was changed to
serum-free neurobasal medium with B27 and L-glutamate
supplements (Life Technologies, Inc.). Two days later, non-aged A
was added to the medium.
Intracellular ACh was determined using
an HPLC-electrochemical detector system (1, 22). Rate of ACh synthesis
from [2-14C]pyruvate was measured by the modified method
of Gibson et al. (23). Cultures were incubated with various
concentrations of non-aged A1-42 for 12 h, and
[2-14C]pyruvate (1 µCi/ml) was added to the medium.
After a 40-min incubation, cultures were washed with PBS, extracted
with 5 mM Tris-Cl (pH 7.5), 0.5% Triton X-100, and scraped
off. Supernatants were recovered by centrifugation, and ACh was
extracted by adding 14 mM sodium phosphate buffer (pH 7.4)
containing sodium tetraphenylborate and acetonitrile, followed by
toluene-based scintillant. Incorporation of radioactivity from
[2-14C]pyruvate into ACh was proportional to time, and
the rate of ACh synthesis was calculated as cpm/min. ChAT activity was
assayed as described (1). ATP content in the culture was measured by Luciferase reaction using a Promega kit (FF2000 and FF2040) with a
luminometer, Lumat LB9501/16 (Berthold Japan, Tokyo). AChE activity was
determined in terms of production of [3H]acetate from
[acetyl-3H]choline iodide as described (24).
Choline uptake by the culture was measured as described (25). Cultures
were incubated with various concentrations of non-aged
A
1-42 for 12 h, and [methyl-3H]choline chloride (3 µCi/ml) was
added to the medium. Forty min later, cultures were washed four times
with PBS, treated with trichloroacetic acid (5% w/v), and scraped off.
Supernatants were obtained by centrifugation, neutralized with 1 M Tris buffer (pH 7.8), and counted for radioactivity in
toluene-based scintillant. PDH activity was measured as production of
[14C]acetyl-CoA from [2-14C]pyruvate with
an excess amount of ChAT, which converted the produced
[14C]acetyl-CoA into [14C]ACh. After
non-aged A
1-42 treatment, cultures were washed with
PBS, scraped off in 10 mM potassium phosphate buffer (pH 7.4), 1 mM EGTA, 1 mM EDTA, and 0.1% Triton
X-100, frozen, and thawed. This homogenate was incubated at 37 °C
for 40 min in the standard reaction buffer for PDH activity (1) (50 mM potassium phosphate buffer (pH 8.0), 2.5 mM
NAD, 0.2 mM thiamin pyrophosphate, 0.13 mM CoA,
and 2.6 mM L-cysteine) containing 1 mM choline chloride, ChAT (Sigma, C-3388,
0.05 unit/assay), and 1 mM [2-14C]pyruvate
(0.5 µCi/assay). The reaction was terminated by adding 14 mM sodium phosphate buffer (pH 7.4) containing sodium
tetraphenylborate and acetonitrile, followed by toluene-based
scintillant.
AChE staining was performed as described
(26). Briefly, after aged A1-42 or non-aged
A
1-42 treatments at 10 µM for 24 h,
cultures were washed with PBS, fixed for 30 min with 4%
paraformaldehyde at room temperature, washed with PBS, and then
incubated for 3 days at 4 °C in 50 mM acetate buffer (pH 5.0) containing 4 mM acetylthiocholine iodide, 2 mM copper sulfate, 10 mM glycine, and 10 mg/ml
gelatin. Nonspecific cholinesterases were inhibited by addition of 0.2 mM ethopropagine in the above medium. Then, the gelatin was
dissolved by incubation at 37 °C. Cultures were washed with water,
exposed for 1 min to 1.25% Na2S, washed with water, and
exposed for 1 min to 1% AgNO3.
After non-aged A1-42 treatment,
mitochondrial pellets were prepared from the culture (1) and disrupted
in a buffer (10 mM Tris-Cl (pH 7.4), 1 mM EGTA,
1 mM EDTA, 50 mM NaCl, 50 mM NaF, 50 mM
-glycerophosphate, 0.5 mg/ml benzamidine, 1 µg/ml antipain and leupeptin, 30 nM okadaic acid, 1 mM phenylmethylsulfonyl fluoride, 1 mM
vanadate, and 0.5% Tween 20). TPKI/GSK-3
activity was determined by
an immunoprecipitation assay using modified phosphoglycogen synthase
peptide as a substrate (1).
Primary septal cultures have been established from
rat septum regions that include septal and basal forebrain cholinergic neurons (1). A is normally secreted in vivo as a soluble
form, which aggregates into insoluble amyloid fibrils to form amyloid plaques (21, 27). Synthetic A
required preincubation for 2-7 days
at 37 °C to produce insoluble amyloid fibrils, which were cytotoxic
to septal cultures (Fig. 1, B and
F), as is consistent with the previous report (21). Here, we
use synthetic A
1-40 and A
1-42 peptides
solubilized freshly in distilled water, which we designate as non-aged
A
1-40 and non-aged A
1-42, respectively.
Non-aged A
1-40 did not present any toxicity to septal
cultures. Since soluble A
1-42 is known to aggregate into insoluble, toxic amyloid fibrils more rapidly than
A
1-40 (28), we have examined the cytotoxicity of
non-aged A
1-42 at concentrations up to 10 µM. Septal cultures treated with non-aged A
1-42 at 10 µM for 24 h (Fig.
1C) were morphologically indistinguishable from the control
(Fig. 1A). No difference was observed between the control
culture and the culture treated with non-aged A
1-42 for
48 h (data not shown). The cellular ATP level was stable after
non-aged A
1-42 treatment for 12 h (Fig.
1D), indicating that non-aged A
1-42 at
concentrations up to 10 µM did not disturb the energy
metabolism in the culture. Histochemical staining for AChE, a marker
enzyme of cholinergic neurons, revealed that cholinergic neurons were
not affected by non-aged A
1-42 treatment at 10 µM for 24 h (Fig. 1, E and G).
ChAT activity, another marker of cholinergic neurons, was not decreased
by exposure of the neurons to non-aged A
1-42 for
12 h (Fig. 1H). Thus, non-aged A
1-42
was not cytotoxic to cholinergic neurons in culture. These results
suggest that neither non-aged A
1-42 nor non-aged
A
1-40 used in our experiments formed toxic amyloid
fibrils.
Non-aged A
Primary septal
cultures were treated with either non-aged A1-40 or
non-aged A
1-42 for 12 h, and the intracellular ACh
level was determined by means of HPLC with an electrochemical detector
system (1, 22). Non-aged A
1-42 at a saturating concentration of 100 nM reduced the intracellular ACh level
to ~40% of the control after a 12-h treatment, whereas non-aged
A
1-40 was ineffective (Fig.
2A). Since neither non-aged
A
1-40 nor non-aged A
1-42 was
neurotoxic, cholinergic neuronal loss cannot be the cause of the
reduced ACh level. Therefore, the decline in the intracellular ACh
level results either from reduced ACh synthesis or from accelerated ACh
release. Non-aged A
1-42 at 10 nM caused a
significant reduction in the intracellular ACh level (Fig.
2A). A corresponding reduction was induced in the synthesis
rate of ACh from pyruvate by treatment with non-aged A
1-42 at 10 nM for 12 h (Fig.
2B). Thus, reduced ACh synthesis is primarily responsible
for the decline in the intracellular ACh level after non-aged
A
1-42 treatment. Treatment of non-aged
A
1-42 at zero time affected neither the intracellular ACh level nor the ACh synthesis rate. The results suggest
A
1-42 remaining in the cell lysate has no direct effect
on the assay procedures, but non-aged A
1-42 works on
the living neurons to disturb the ACh synthesis.
Suppression of ACh Synthesis Can Be Attributed to Reduced Mitochondrial PDH Activity, Probably via TPKI/GSK-3
ACh is synthesized from choline
and acetyl-CoA by ChAT, which is not rate-limiting under normal
conditions (29). Treatment with non-aged A1-42 for
12 h did not change ChAT activity (Figs. 1H and
3A). Therefore, a reduced supply of either choline or
acetyl-CoA accounts for the decline in ACh synthesis.
Neurons obtain choline mainly from the diet using a specialized
transporter system that also salvages free choline generated from ACh
released intrasynaptically by AChE (30). Treatment with non-aged
A1-42 did not interfere with this choline uptake
system, since we found no statistically significant change in the
choline uptake in the case of non-aged A
1-42 treatment (Fig. 3A). In accordance with AChE
cytochemistry (Fig. 1, E and G), AChE activity
was stable after non-aged A
1-42 treatment for 12 h
(Fig. 3A). These results suggest that non-aged
A
1-42 did not influence choline turnover in cholinergic
neurons. This implies that a reduced supply of acetyl-CoA from pyruvate
must be critical (Fig. 2B).
In adult mammalian brain, acetyl-CoA is mainly produced by oxidative
decarboxylation of pyruvate by mitochondrial PDH complex. Non-aged
A1-42 at a saturating concentration of 100 nM reduced PDH activity to ~30% of the control (Fig.
3B), which corresponds well with the
dose-dependent reductions of the intracellular ACh level
(Fig. 2A) and ACh synthesis rate (Fig. 2B).
Non-aged A
1-42 treatment for 3 h was enough to
cause maximal inhibition of PDH activity in the culture.
Correspondingly, the intracellular ACh reached the lowest level after
non-aged A
1-42 exposure for 3 h. Thus, it appears
that non-aged A
1-42 suppressed ACh synthesis mainly by
inhibiting PDH activity and so reducing the availability of acetyl-CoA.
PDH activity is regulated by phosphorylation (31, 32). We have
previously found that PDH is phosphorylated and inactivated by
TPKI/GSK-3 in vitro as well as in
A
25-35-treated hippocampal neurons (1). In septal
cultures, 10 nM A
1-42 was enough to
activate mitochondrial TPKI/GSK-3
by 2-fold (Fig. 3B).
These results suggest that TPKI/GSK-3
is the main mediator of the
inactivation of PDH by non-aged A
1-42 in cholinergic
neurons, as previously found in hippocampal neurons.
A1-42, a minor species in normal APP metabolism,
is highly aggregable (28), acts as a seed for amyloid fibrils (33), and
seems to deposit initially in amyloid plaques in AD (34) and Down's
syndrome (DS) (35). APP mutation linked to familial AD (FAD) or an
increased APP gene dosage in DS brain leads to a specific increase in
A
1-42 production (19, 36). Recently, Younkin and
colleagues reported that mutations in FAD genes, including presenilin 1 (37) and presenilin 2 (38), result
in increased levels of A
1-42 in fibroblasts and plasma
in some FAD kindreds (39). These studies indicate a critical and
central role for A
1-42 in amyloid plaque formation in
AD. The findings presented here reveal a new aspect of
A
1-42 function, implying a role in inducing cholinergic
dysfunction, besides amyloid plaque formation.
In the present work, we used A1-42 freshly dissolved in
water (non-aged A
1-42), which exhibited no
neurotoxicity and was confirmed to be in an oligomeric state by laser
light-scattering analysis (data not shown). Therefore, the non-aged
A
1-42 used in our experiments probably does not form
toxic amyloid fibrils, but does impair ACh synthesis in cholinergic
neurons at very low concentrations (10-100 nM). Soluble
A
1-42 production is uniquely elevated in brains of AD
(2) and DS patients before amyloid plaque formation (36), suggesting a
role of soluble A
1-42 in the pathogenesis. We
hypothesize that soluble A
1-42 produced at an early
stage of AD starts affecting cholinergic neurons by suppressing ACh
synthesis, causes a reduction in ACh release, modulates synaptic
connections, and finally results in cholinergic deficits, which may
induce progressive loss of memory and cognitive function in AD
patients. Thus, nontoxic, soluble A
1-42 may play a
primary role in the cholinergic dysfunction of AD patients by
suppressing ACh synthesis, besides contributing to amyloid plaque
formation.
TPKI/GSK-3 is present both in cytoplasm and in mitochondria in
neurons (40). We found that in cholinergic neurons non-aged A
1-42 at 10 nM activates mitochondrial
TPKI/GSK-3
, which mediates the reduction in ACh synthesis via PDH
inactivation, as we had previously shown in hippocampal neurons (1).
One critical difference between the cholinergic and hippocampal systems lies in the energy metabolism. In the case of hippocampal neurons, activation of mitochondrial TPKI/GSK-3
by aggregated
A
25-35 inactivates PDH, which causes the energy
depletion (1). Hippocampal neurons may die as a result of insufficient
energy supply, as well as abnormally phosphorylated
that impairs
axonal transport (41). However, in the case of cholinergic neurons,
inactivation of PDH activity by non-aged A
1-42 does not
interfere with the energy metabolism and the neurons survive. We
speculate that the anaplerotic pathway between glycolytic breakdown and
the tricarboxylic acid cycle functions in cholinergic neurons, but
aggregated A
inhibits this pathway in hippocampal neurons. Whether
the difference in the energy metabolism depends on the A
species or
on the neuronal species needs to be clarified.
Based on the present study, we suggest that A1-42
participates directly in the induction of cholinergic deficiency, as
well as amyloid plaque formation in AD. Application of the findings
presented here as a hypothetical in vitro model for AD pathogenesis should open up many possibilities for further research.
We thank Dr. Y. Kasai and R. Kuwahara (Mitsubishi Kasei Institute of Life Sciences) for their help with the HPLC electrochemical detection system and Dr. T. Hama (Mitsubishi Kasei Institute of Life Sciences) for her help with AChE staining.