Current Medicinal Chemistry, 2003, 10, 2651-2659
2651
Amyloid β-Peptide [1-42]-Associated Free Radical-Induced Oxidative Stress
And Neurodegeneration in Alzheimer’s Disease Brain: Mechanisms and
Consequences
D. Allan Butterfield*
Department of Chemistry, Center of Membrane Sciences, and Sanders-Brown Center on
Aging, University of Kentucky, Lexington, KY 40506-0055 USA
Abstract: In addition to synapse loss, neurofibrillary tangles, and neurodegeneration,
oxidative stress and amyloid β-peptide [Aβ] deposition are hallmarks of Alzheimer’s
disease [AD] brain. Our laboratory coupled these two characteristics of AD into a
comprehensive model to account for the synapse loss and neurodegeneration in AD
brain. This model combines much of the extant studies on AD and is based on oxidative
stress associated with amyloid β-peptide. This review presents evidence in support of
this model and provides insight into the molecular basis of this devastating dementing
disorder.
Keywords: Amyloid β-peptide; oxidative stress; neurodegeneration; protein oxidation; lipid peroxidation; free radicals;
proteomics.
presents the studies of this model and the insights into AD
that the model provides.
INTRODUCTION
Three major pathological indices characterize Alzheimer’s
disease (AD) brain: senile plaques (SP), neurofibrillary
tangles, and synapse loss [1,2], which are thought to
underlie the principal clinical manifestations of AD, i.e.,
progressive loss of memory, cognition, speech, and
normative behavior. The AD brain is under intense oxidative
stress as manifested by increased protein oxidation, lipid
peroxidation, free radical formation, advanced glycation
endproducts, and DNA/RNA oxidation, among other
biomarkers [3,4].
Accumulating research suggests that the main component
of SP, amyloid β-peptide (1-42) [Aβ(1-42)], is central to the
pathogenesis of AD [5]. The 42-amino acid Aβ(1-42) is
formed by proteolytic cleavage of a transmembrane protein,
amyloid precursor protein (APP) by two proteases, βsecretase and γ-secretase (Figure 1). Mutations in the genes
for APP, and presenilin-1, and presenilin-2 cause the
familial form of the disease, and an overproduction of Aβ(142) is observed [5]. A major risk factor for AD is the
presence of allele 4 of apolipoprotein E [5]. The amino acid
sequence of Aβ(1-42) (Table I) shows the purported redox
metal ion binding sites (H6, 13,14) and the single
methionine at residue 35, both of which are discussed in
detail below.
The oxidative stress under which AD brain exists and the
central role of Aβ(1-42) in the pathogenesis of the disorder
were coupled into a model for neurodegeneration in AD
brain: the Aβ(1-42) – associated free radical oxidative stress
model for neurodegeneration in AD [4,6-10]. This review
*Address correspondence to this author at the Department of Chemistry;
Center of Membrane Sciences; and Sanders-Brown Center on Aging,
University of Kentucky, Lexington, KY 40506 USA; Phone: [859] 2573184; Fax: [859] 257-5876; E-mail: [email protected]
0929-8673/03 $41.00+.00
Table I.
Amyloid β -Peptide (1-42) Primary Structure [Note
Methionine at Residue 35]
D-A-E-F-R 5-H-D-S-G-Y 10-E-V-H-H-Q 15-K-L-V-F-F 20-A-E-D-VG 25 -S-N-K-G-A 30 -I-I-G-L-M 35 -V-G-G-V-V 40 -I-A
NEURODEGENERATION IN AD BRAIN: THE
β ( 1-42) – ASSOCIATED FREE
MODEL OF Aβ
RADICAL OXIDATIVE STRESS
The model (Figure 2) is consistent with the concept that
Aβ(1-42) is central to the oxidative stress and neurotoxicity
in AD brain. Irrespective of the underlying basis of why
elevated levels of Aβ(1-42) arise (inheritance of defective
genes, genetic predisposition, or sporadically), the increased
levels of Aβ(1-42) are hypothesized to insert as a small
aggregate into the neuronal membrane. Lipid peroxidation,
resulting from free radicals associated with Aβ(1-42), occurs
from which direct damage to transmembrane proteins can
result. Alternatively, reactive lipid peroxidation products,
especially 4-hydroxy-2-nonenal (HNE) or 2-propen-1-al
(acrolein), can bind by Michael addition to protein cysteine,
histidine, or lysine residues, thereby altering their
conformation and function [11-14]. As a consequence, Ca2+
dyshomeostasis occurs, leading eventually and inexorably to
cell death. Mitochondrial dysfunction and induction of nitric
oxide synthesase (iNOS) by Aβ(1-42) are posited.
Superoxide radical and NO, that result from such processes,
respectivly, react at a diffusion-controlled rate to form
peroxynitrite, a neurotoxic species [15]. Elevated calcium
ions cause free radical formation by swelling of
mitochondria and endoplasmic reticulum, with subsequent
release of superoxide and other radicals. Cytochrome c
release from mitochondria causes activation of the caspase
cascade, thereby initiating apoptosis. In addition, elevated
© 2003 Bentham Science Publishers Ltd.
2652 Current Medicinal Chemistry, 2003, Vol. 10, No. 24
D. Allan Butterfield
Fig. (1). Cleavage of the transmembrane protein APP by β- and γ-secretases to form Aβ-peptide.
Ca2+ induces phospholipase A2 (PLA2) activation, and free
fatty acids, which are capable of aggregating tau protein, the
main constituent in NFT [2], are liberated. Excess
intracellular Ca2+ can result from NMDA receptor activation,
leading to free radicals [16]. Thus, the possibility of
excitotoxic neuronal death, resulting from excess glutamate
on the external side of neurons, especially when coupled
with diminished glutamine synthetase (GS) activity or
dysfunction of the glutamate transporter, is enhanced.
Fibrillar aggregates of Aβ(1-42) can activate microglia [17],
which, in turn, release neuron-damaging free radicals in an
inflammatory response. The model, the central focus of
which is the oxidative stress and subsequent neurotoxicity
induced by Aβ(1-42), is consonant with several mechanisms
for neuronal death in AD brain.
EXPERIMENTAL SUPPORT OF THE MODEL
β (1-42)
Protein Oxidation Caused by Aβ
Protein oxidation, assessed by immunochemical
detection of the hydrazone formed by reaction of 2,4dinitrophenylhydrazine with carbonyl functional groups
[14,18], is prominent in AD brain [19-21]. Free radical
induced peptide backbone scission, oxidation of selective
amino acid side chains, and covalent modification of
proteins by reactive alkenals such as HNE and acrolein are
the three major ways of introducing carbonyl groups into
proteins following oxidative stress [14].
Consistent with a free radical process involving Aβ(142), the chain-breaking antioxidant vitamin E blocks free
radical formation, protein oxidation, lipid peroxidation, and
cell death in hippocampal cultures treated with Aβ(1-42) [2224]. Protein oxidation also was assessed in neurons treated
with Aβ using histofluorescence methods, and propyl
gallate, also a free radial scavenger, blocked this effect [25].
Supporting the role of free radicals associated with Aβinduced protein oxidation, vitamin E inhibited the Aβinduced loss of enzyme activity of creatine kinase (CK, BB
isoform) [26]. The activity of CK is lower in AD brain than
in control brain [19,20]. Oxidation of GS by Aβ was
indexed by analysis of the real time kinetics of spin label
incorporation into this protein that contains 88 sulfhydryl
groups [27]. The protein-specific spin label was covalently
bound to oxidized GS at a considerably slower rate than in
unoxidized GS. Similarly, highly purified GS isolated from
AD brain incorporated the spin label at a slower rate than GS
from control brain, consistent with the GS being oxidized in
AD. Oxidation of both CK and GS in AD brain was
confirmed directly using proteomics (see below). Addition
of Aβ to GS lowered the rate of incorporation of the spin
label into the protein, suggesting that this peptide caused
oxidation of GS [27]. In AD brain, diminution of GS
activity relative to that in control brain is found [19,20].
β
Lipid Peroxidation Caused by Aβ
Lipid peroxidation induced by Aβ, and blocked by
vitamin E, was shown using electron paramagnetic resonance
(EPR) in conjunction with a paramagnetic, lipid-specific
spin label [28-30]. Overexpression of bcl-2, an anti-apototic
protein, bcl-2 prevented Aβ-induced lipid peroxidation [30].
The reactive alkenals HNE and acrolein, found in excess
in AD brain [31,32], react with protein-resident cysteine,
histidine, or lysine residues by Michael addition (Figure 3)
[11,14]. Transmembrane proteins provide proximal targets
for attack by HNE or acrolein [11,12], which are formed in
the lipid bilayer. Direct evidence for HNE formation
Mechanisms and Consequences
Current Medicinal Chemistry, 2003, Vol. 10, No. 24
2653
Fig. (2). Model of Aβ-associated free radical oxidative stress model for neurotoxicity in AD brain. See text.
following Aβ(1-42) addition to neurons or synaptosomes
has been reported [33,34]. Other lipid peroxidation products,
F2-isoprostanes, also are formed following addition of Aβ(142) to cells [35]. Both Aβ and HNE bind to and alter the
P
CH2
H
SH
OH
Acrolein
O
H
O
HNE
Fig. (3). Michael addition of HNE and acrolein to form covalent
adducts on proteins.
structure of synaptosomal membrane proteins [12,36], and
both Aβ and HNE alter the function of ion-motive ATPases
[37-39], transporters for glucose [40], glutamate [34,40], and
polyamines [41], and various G-protein coupled signal
transduction pathways [42]. Both HNE and acrolein alter the
conformation of membrane proteins, which likely is the
cause of the altered function of transport proteins induced by
these reactive products of lipid peroxidation [12,13].
The glutamate transporter Glt-1 [EAAT2] in AD brain is
covalently adducted by HNE in excess compared to that in
control brain [34], with this oxidative modification likely
explaining the reported loss of Glt-1 activity in AD [43].
The same result as in AD brain could be induced by addition
of Aβ(1-42) to synaptosomal preparations [34], consistent
with the idea that Aβ(1-42)-induced lipid peroxidation
2654 Current Medicinal Chemistry, 2003, Vol. 10, No. 24
results in HNE-induced modification of this glutamate
transporter in AD brain. Moreover, in AD brain excess
extracellular glutamate and resultant excitotoxicity, as
sequelae to oxidative modification and loss of activity of
Glt-1, and coupled to Aβ-induced inhibition of GS activity
[19,20,27,44,45], provide one mechanism for
neurodegeneration in AD. Since glutamate-stimulated
NMDA receptor activation leads to elevated intracellular
Ca 2+ accumulation [16], this glutamate-based excitotoxic
mechanism is a means to account for increased intracellular
Ca2+ following Aβ addition to neurons, consistent with the
model (Figure 2).
Elevated in AD brain [32], the reactive alkenal, acrolein,
similar to HNE, inserts a carbonyl functionality to proteins
upon binding (Figure 3). Elevation of the endogenous
antioxidant glutathione (GSH) in rodents eliminated the
acrolein-induced increased carbonyl content in subsequently
isolated synaptosomes [13]. This result is consistent with
the notion that one strategy to potentially combat Aβinduced oxidative stress and its sequelae is to increase the
endogenous antioxidant content of neurons (particularly
GSH [46]) or increase the level of enzymes that inhibit
oxidative stress, e.g., heme oxygenase [47].
Of the three principal alleles of the lipoprotein,
apolipoprotein E, apoE4 is a risk factor for AD [5]. Aβ(1-42)
was added to synaptosomes from knock-in mice having
human Apo E2, Apo E3, or Apo E4 [48]. In these ApoE
knock-in mice, the mouse promoter is present, but the
human gene replaces the mouse gene for Apo E;
consequently, the normal level and location of Apo E is
present, but of the human form, not the mouse form. Of the
markers of oxidative stress determined, among which were
DCF fluorescence (to measure ROS), protein carbonyls (to
index protein oxidation), and TBARS (a crude measure of
lipid peroxidation), all were more sensitive to Aβ(1-42) in
synaptosomes from human Apo E4 mice than in those from
human Apo E2 or Apo E3 animals [48]. Although there are
other interpretations of these results, particularly those
involving clearance, one interpretation of these findings is
that Apo E4 may be a risk factor for AD because it is unable
to handle the oxidative stress associated with Aβ(1-42)
relative to that of Apo E2 or Apo E3. This idea may be
associated with the lack of SH groups in Apo E4, one
consequence of which is decreased free radical scavenging
potential, or to altered binding of Aβ(1-42) by ApoE4,
thereby providing increased opportunity for Aβ(1-42)induced oxidative stress.
After our publication of Aβ-induced lipid peroxidation
[28], others have reported that Aβ causes lipid peroxidation
and that various antioxidants are effective in blocking this
effect (extensively reviewed in [7-9]). In contrast to the wide
agreement of studies from many different laboratories, a few
reports doubt the ability of Aβ to induce lipid peroxidation
or lipid-soluble antioxidants to prevent this oxidation [4951]. However, methodological differences, such as using 13-day old neuronal cultures to test Aβ, that because the full
complement of important receptors and other factors require
considerably longer to be expressed, may not reflect a good
model for adult neuronal brain cells. The preponderance of
the evidence supports a role for Aβ -induced lipid
peroxidation that is also extensive in AD brain.
D. Allan Butterfield
Peroxynitrite
3-Nitrotyrosine, a marker for the presence of
peroxynitrite, formed from NO and superoxide radical [15],
is elevated in AD brain [52,53]. In the model for
neurodegeneration in AD brain (Figure 2), Aβ(1-42) (or its
downstream effects) is hypothesized to cause mitochondrial
dysfunction. This, in turn, leads to subsequent superoxide
release. Also, i-NOS induction, with subsequent formation
of NO, is stimulated by Aβ(1-42) or its effects. Peroxynitrite
then leads to 3-nitrotyrosine formation. Peroxynitrite added
to synaptosomes caused increased protein oxidation, indexed
by increased protein carbonyls, that could be inhibited by
glutathione [54]. Glutathione also prevented the extensive
peroxynitrite-induced neuronal death in hippocampal
neuronal culture [54].
The Aβ(1-42)-associated free radical oxidative stress
model for neurodegeneration in AD brain predicts that
elevated intracellular calcium ions cause numerous
destructive cellular processes to occur (proteolysis,
phospholipid breakdown, mitochondrial swelling, apoptosis,
and endonuclease activation, among others). When Aβ is
added to neurons or glial cells in culture excess intracellular
calcium is observed [55], consistent with the notion that
Aβ(1-42) may be responsible for similar events in AD brain.
β (1-42): Role in PeptideMethionine Residue 35 of Aβ
Induced Oxidative Stress and Neurotoxicity
Considerable evidence exists suggesting that the single
methionine residue at position 35 in Aβ(1-42) (Table I),
contributes to the oxidative stress and neurotoxic properties
of this peptide. For example, if a single atom of Aβ(1-42) is
changed, namely a carbon atom for the sulfur atom of the
thioether of Met-35, the resulting norleucine (Nle) derivative
of Aβ (1-42) with the same length and relative
hydrophobicity as the native peptide, is, in contrast, no
longer oxidative nor neurotoxic [56]. Further, C. elegans
transgenic for human Aβ(1-42) show increased protein
oxidation, a finding that is abrogated if the codon for
methionine is replaced by a codon of a different amino acid
[56], consistent with the notion that both in vitro and in
vivo the single methionine residue of Aβ(1-42) is critically
important to the oxidative stress and neurotoxic properties of
the peptide. This notion is supported by several different
observations. If Aβ(1-42) with the sulfur atom of Met35
already oxidized (methionine sulfoxide) [Aβ(1-42)M35SO]
is added to neuronal cultures, no neuronal cell death nor
protein oxidation occurs [57]. This and other studies led to
our idea that the reactive form of methionine in Aβ(1-42) is
the sulfur radical cation (Figure 4). This species is capable of
abstracting a labile H atom from a carbon atom adjacent to
unsaturated carbon atoms on phospholipid acyl chains to
form a carbon-centered radical that, in turn, will rapidly bind
paramagnetic molecular oxygen to form a peroxyl free
radical. The peroxyl radical can partake in subsequent chain
reactions forming free radicals that can damage membrane
components directly or yield reactive aldehydes, such as
HNE or acrolein. The protonated sulfur atom of Aβ(1-42)
has a pKa of - 5; consequently, this thio acid will
immediately lose the H+ to a nearby base, resulting in
methionine, i.e., the process is catalytic [57] (Figure 4).
Mechanisms and Consequences
Current Medicinal Chemistry, 2003, Vol. 10, No. 24
R
Methionine in
Aβ (1-42)
β (1-42) Neurotoxicity: Cu2+ and Methionine
Aβ
O
O
H
N
Oxidation
-e
R
R
H
N
R
S. ..
S. .:
LH or PH
-H
O
H
N
R
R
2655
.
.
-L or -P
Protein Carbonyls
Lipid Peroxidation
and Its Sequelae
SH
..
Fig. (4). Oxidation of methionine to form a sulfuramyl free
radical on the S atom and its subsequent reaction with labile H
atoms on lipids (L-H) or proteins (P-H). The resulting lipid
radical (L . ) binds paramagnetic oxygen, leading to a chain
reaction to form lipid peroxidation and lipid peroxidation
products, such as HNE. Protein radicals (P.) also bind molecular
oxygen leading to protein carbonyls. The acid formed on the S
atom of Met in residue 35 of Aβ1-42) has a pKa of minus 5;
thus, any base would immediately bind to this H+ , leading to
methionine in the Aβ(1-42) peptide, i.e., the process is proposed
to be catalytic. The lack of such chemistry in the absence of the
S-atom of methionine (norleucine) in Aβ(1-42) or absence of a
helical secondary structure around Met-35 is consistent with
this proposed mechanism. See text.
Aβ(1-42) in solution adopts an α-helical structure for the
carboxy-terminal portion of the peptide, including the
methionine at residue 35 [58,59]. One of the properties of
helices is to have an i-i+4 interaction. Accordingly, NMR
data show that the peptide carbonyl of Ile31 is within a van
der Waals distance from the sulfur atom of Met35. We
hypothesized that this interaction, based on studies of model
thioethers [60], alters the electronic environment around the
sulfur atom of Met35, making the sulfur prone to oxidation.
Consistent with this notion, substitution of proline, a helixbreaking amino acid, for isoleucine in position 31 of Aβ(142) showed that, in contrast to native Aβ(1-42), no protein
oxidation nor neurotoxicity was found [61]. Thus, the
oxidative stress and neurotoxic properties of Aβ(1-42) are
presumed to emanate from both the chemistry of thioethers
and the unique secondary structure of the peptide [61,62].
The model outlined in Figure 2 implies that lipid
peroxidation induced by small aggregates of Aβ(1-42) is an
early event in the oxidative stress and neurotoxic action of
this toxic peptide. Consistent with this implication,
substitution of aspartic acid for glycine in position 37 of
Aβ(1-42) caused the Aβ(1-42)G37D to no longer be toxic
nor oxidative to neurons [63]. We reasoned that the negative
charge on Asp would force the methionine residue out of the
low dielectric environment of the bilayer, thereby preventing
lipid peroxidation. Thus, even though the chemistry
associated with the structure of the peptide and the
involvement of the methionine were theoretically possible,
no target for lipid peroxidation was possible, since the
peptide was outside the bilayer.
Redox metal ions have been invoked in the mechanisms
of oxidative stress and neurotoxicity associated with Aβ(142) [64,65]. The three histidine residues of Aβ(1-42) at
positions 6,13,14 bind Cu 2 + , and the peptide is
hypothesized to reduce this redox ion to Cu(I). Subsequent
chemistry that has Cu+ reacting with molecular oxygen in a
two-electron process is postulated to form H2O 2, which is
suggested to react with Cu(I) to form the damaging hydroxyl
radical [64,65]. To investigate this postulated mechanism,
and the potential role of the interaction between peptidebound Cu+ and Met35, the norleucine derivative of Aβ(142), which, as noted above, lacks the methionine sulfur, but
does contain the three Cu(II)-binding His residues, was
neither oxidative nor neurotoxic, even in the presence of
added Cu(II) [57]. In addition, Tyr, which binds Cu2+ at
least one hundred fold less avidly than does His, was
substituted for the three His residues, and the resulting
Aβ(1-42) modified peptide was still not toxic [57]. Thus,
Cu 2+ conceivably may be important in Aβ(1-42)-induced
oxidative stress and neurotoxicity, but if this is the case, the
above studies suggest that Met is required. Consistent with
this notion, others showed that the shorter fragment, Aβ(128), which still contains the three Cu2+ -binding His
residues, but not methionine, was not toxic and did not
reduce Cu2+ [66]. However, these researchers showed that
addition of Met to the Aβ(1-28) preparation caused the
reduction of Cu2+ to Cu+ and intercalation into the lipid
bilayer [66]. As noted above, substitution of helix-breaking
Ile at residue 31 or synthesis of the norleucine derivative at
residue 35 of Aβ(1-42) abrogates the oxidative stress and
neurotoxic properties of this peptide in both cases, even
though the three His Cu(II)-binding sites are still present in
these modified peptides.
Iron(II) also can undergo Fenton-type chemistry, with
H 2O 2, producing hydroxyl radicals. However, the normal
state of iron is Fe3+, so a reductant is necessary to produce
Fe2+. Ascorbate can serve this function, as can a number of
other agents, including neuromelanin. Hence, in Parkinson’s
disease, excessively high iron content in the substantia
nigra, where neuromelanin is extensively located, may be
one source of the oxidative stress in this neurodegenerative
disorder [67]. As a harder acid, Fe3+ is much less likely to
bind to His than is Cu2+ . Hence, although the chemistry
associated with Cu2+/1+ and Aβ(1-42) noted above is
theoretically possible with Fe3+/2+, the probability of this
chemistry is lower due to the hardness of Fe3+ relative to
Cu2+ and consequent decreased binding affinity for His.
β(1-42) Neurotoxicity and Fibril Formation
Aβ
The presence of neuritic plaques in AD brain that are rich
in aggregated Aβ is consistent with the idea that soluble Aβ
monomers in the AD brain presumably aggregate to form
oligomers or fibrils [68]. Indeed, soluble Aβ oligomers have
been isolated from normal and AD brains [69]. In AD brain,
the percentage of soluble Aβ to total Aβ is substantially
increased relative to control brain, and the proportion of
soluble Aβ(1-42) is increased over the shorter, soluble Aβ(140). Many laboratories have reported that low molecular
weight oligomers of Aβ are neurotoxic [70-76]. The
2656 Current Medicinal Chemistry, 2003, Vol. 10, No. 24
D. Allan Butterfield
neurotoxic peptides, Aβ(1-42), Aβ(1-42)His6,13,14Tyr,
Aβ(1-40) and Aβ(25-35) form fibrillar structures upon
incubation. Many reports show that Aβ(1-42) fibrils are not
required for toxicity [70-76], a finding confirmed by fibrilforming, non-neurotoxic, non-protein-oxidizing, and nonfree radical-forming Aβ (1-42)Met35Nle and Aβ(1 42)M35SO [57]. Thus, studies with Aβ(1-42) suggest that
while the methionine residue of this peptide, and possibly
redox metal ions, are crucial for the peptide-induced
neurotoxicity, fibril formation per se is not required for
neurotoxicity. Rather, reports from many laboratories
suggest that small aggregates of Aβ(1-42) are likely the toxic
species [72-77]. Such small aggregates would be capable of
inserting into the lipid bilayer to induce lipid peroxidation,
while large fibrillar structures would have difficulty doing
so. Consistent with the notion that small aggregates of
Aβ(1-42) are the toxic species of this peptide, we recently
showed in living C . e l e g a n s engineered to have a
temperature-sensitive human Aβ(1-42) expression system
that oxidative stress occurred 24 hours after induction of the
peptide expression, a time that exactly mirrored the
phenotypic expression of the worm (paralysis) and in the
absence of large fibrillar deposits [77].
Oxidatively-Modified Proteins
Identification by Proteomics
in
AD
Brain:
Protein oxidation in AD brain regions where Aβ(1-42) is
present, but not in Aβ(1-42)-poor cerebellum, is significant
[19]. But, which particular proteins are oxidized?
Identification of specifically oxidized proteins may give
insight into loss of protein function and potential
mechanisms of neuronal death in AD brain. As an initial
step toward this end, immunochemical methods coupled to
two-dimensional electrophoresis revealed that CK (BB
isoform) and β-tubulin were specifically oxidized in AD
brain [78,79]. Though reliable, immunochemical detection
of specifically oxidized proteins is labor-intensive, requires
knowledge of the molecular weight and isoelectric point of
unknown proteins, and requires availability of specific
antibodies. Newer, data-based, automated methods of
unknown protein identity are necessary to make the aim of
identifying specifically oxidized proteins in AD brain
feasible. We have begun this process using proteomics [8084]. This mass spectrometric analysis of oxidized protein
spots from a two-dimensional matrix, for the first time, has
been used to identify oxidized proteins from AD brain. CK
(BB isoform), GS, and ubiquitin carboxyl-terminal
hydrolase L-1 (UCH), dihydropyrimindase related protein 2
(DRP-2), and α-enolase were found to be specifically
oxidized in AD brain [80,81]. Proteomic screening of AD
brain to identify proteins posttranslationally modified as
assessed by 3-nitrotyrosine, also revealed specific proteins:
α-enolase, triosephosphate isomerase, and neuropolypeptide
h3 [82].
The proteomic identification of oxidized CK in AD brain
[80] confirms the immunochemical identification of this
oxidized protein [78,79]. The low activity of CK in AD [19]
likely reflects this oxidative modification. The other energyrelated enzymes, α-enolase and triosephosphate isomerase,
identified as oxidatively modified in AD brain may, like
CK, have decreased activity in AD brain as a consequence of
its oxidation. Thus, the oxidation of these three energyrelated enzymes may contribute to the known decreased
energy utilization in AD brain.
GS activity also is decreased in AD brain [19], highly
likely a consequence of its oxidation. In recollection of the
oxidative modification and low activity of GLT-1 noted
above [34,43], the oxidative modification and low activity
of GS may contribute to an excitotoxic mechanism for
neuronal death in AD brain. The recent success of
memantine, a NMDA receptor antagonist, in late-stage AD
patients [85] is consistent with this suggestion.
As proteins become damaged or oxidized, ubiquitin can
be bound to them, marking them for degradation by the 26S
proteasome [86,87]. If oxidized UCH, like CK and GS, has
low activity in AD brain, then recycling of ubiquitin will be
inhibited. Increased protein ubiquitination, decreased activity
of the proteasome, and accumulation of aggregated and
damaged proteins would be predicted, and all are observed in
AD brain. Additional studies in AD and rodent modes
thereof, currently underway in our laboratory, will determine
if this process is operative in AD.
Elongation and directionality of the dendritic growth
cone is promoted by the protein collapsin, which in turn is
regulated or influenced by DRP-2. Normally, this protein is
expressed during development, but under conditions of
neuronal loss, as in AD brain, its expression may be
increased in adult brain as well. Should the function of
DRP-2 be diminished as a consequence of its oxidation,
shortened dendritic lengths and decreased interneuronal
connections would be predicted. Shortened dendritic lengths
are observed in AD brain [88], and decreased network
connections in neuronal cultures treated with Aβ a r e
observed [26]. One could speculate that decreased cellular
contact could affect neuronal signaling and processes
associated with memory, the most prominent clinical
presentation in AD.
Neuropolypeptide
h3,
also
known
as
phosphatidylethanolamine binding protein [PEBP] or
hippocampal cholinergic neurostimulating peptide [HCNP],
is oxidatively modified in AD brain [82]. Conceivably,
oxidation of HCNP contributes to the known deficits of the
cholinergic system in AD brain. In its role as PEBP,
oxidation of this protein could, in principle, affect
phospholipid asymmetry, loss of which would permit
exposure of phosphatidylserine (PS) on the external lamellae
of the neuronal lipid bilayer. Such an orientation of PS is a
signal for apoptosis, which conceivably could contribute to
neuronal death in AD brain.
Our proteomics analysis of posttranslationally modified
proteins in AD brain continues and is expected to reveal new
insights into potential mechanisms of synapse loss and
neuronal death in AD brain.
CONCLUSION
Consistent with the Aβ-associated free radical oxidative
stress model for neurodegeneration in AD brain [4,6-10],
Aβ(1-42), in mechanisms that are inhibited by free radical
antoxidants, induces ROS formation, protein oxidation,
lipid peroxidation, and cell death in hippocampal neuronal
Mechanisms and Consequences
cultures. Since Aβ(1-42) is likely central to the pathogenesis
of AD, the oxidative stress and neurotoxic properties of this
peptide presumably contribute to the oxidative stress,
synapse loss, and neurodegeneration found in AD brain (see
other recent reviews [89,90]). The oxidative stress and
neurotoxic properties of Aβ(1-42) couple the centrality of
Aβ(1-42) to the pathogenesis of AD with the oxidative stress
under which AD brain exists [3,6-10]. The single
methionine of Aβ(1-42) at position 35 is critical to the
oxidative stress and neurotoxic properties of the peptide
[56,57,61-63], and may act mechanistically via the sulfur
radical cation or a redox metal complex. The reactive
alkenal, HNE, formed by Aβ(1-42) addition to neurons,
reacts by Michael addition with crucial proteins in neuronal
membranes, thereby changing their structure and function
[11,12,34]. One such protein is GLT-1, a glial-resident
glutamate transporter. Increased HNE binding to GLT-1 is
found in AD brain [34], probably accounting for decreased
activity of this transporter [43], and Aβ(1-42) leads to the
same effect [34]. When coupled to loss of GS activity in AD
brain [19,20] or by Aβ [27,45], GLT-1 inhibition secondary
to oxidative modification increases the possibility of
excitotoxic neuronal death. Increased levels of glutathione
protect brain membrane preparations against these reactive
alkenals, suggesting that elevation of this endogenous
antioxidant may be a promising therapeutic strategy for AD
[13,46,54,91,92]. New insights into potential mechanisms
of synapse loss and neurodegeneration in AD brain, based on
identification of specifically oxidatively modified proteins
using proteomics, have been gained [80-84].
Extensive support for the amyloid β-peptide-associated
free radical oxidative stress model for neurodegeneration in
AD brain has been obtained and summarized here. These
studies suggest that one therapeutic strategy for AD is to
block oxidative stress associated with Aβ(1-42) using both
exogenous and endogenous, brain-accessible antioxidants
[23,46], taking into account the need to preserve normal free
radical processes necessary for brain function. Studies along
these lines are underway in many laboratories, including
ours.
Current Medicinal Chemistry, 2003, Vol. 10, No. 24
DRP-2
Dihydropyrimidinase-related protein 2
EAAT2 =
Excitatory amino acid tranporter, type 2
EPR
=
Electron paramagnetic resonance
GS
=
Glutamine synthetase
GSH
=
Glutathione
HCNP
=
Hippocampal cholinergic neurostimulating
peptide
HNE
=
4-Hydroxy-2-trans-nonenal
iNOS
=
Inducible nitric oxide synthase
NFT
=
Neurofibrillary tangles
Nle
=
Norleucine
PEBP
=
Phosphatidylethanolamine binding protein
PLA2
=
Phospholipase A2
PS
=
Phosphatidyl serine
SP
=
Senile plaques
UCH
=
Ubiquitin carboxy-terminal hydrolyase L-1
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
ACKNOWLEDGEMENTS
[14]
This work was supported in part by grants from NIH
(AG-05119; AG-10836). The author thanks his past and
present graduate students, postdoctoral fellows, and faculty
colleagues, who have contributed to the results reviewed
here. Ms. Mollie Fraim and Dr. Alessandra Castegna are
acknowledged and thanked for their help in preparing this
manuscript as is Ms. Marcia Butterfield for technical
assistance.
=
2657
[15]
[16]
[17]
[18]
[19]
ABBREVIATIONS
[20]
acrolein =
2-Propen-1-al
[21]
AD
=
Alzheimer’s disease
APP
=
Amyloid precursor protein
CK
=
Creatine kinase
DCF
=
Dicholorfluoroscein
[22]
[23]
[24]
Katzman, R. Neurobiol. Aging, 2000, 21, S1.
Katzman, R.; Saitoh, T. FASEB J., 1991, 4, 278-286.
Markesbery, W.R. Free Rad. Biol. Med., 1997, 23, 134-147.
Butterfield, D.A.; Drake, J.; Pocernich, C.; Castegna, A. Trends
Molecular Med., 2001, 7, 548-554.
Selkoe, D.J. J. Biol. Chem., 1996, 271, 18295-18298.
Butterfield, D.A. Chem. Res. Toxicol., 1997, 10, 495-506.
Varadarajan, S.; Yatin, S.; Aksenova, M.; Butterfield, D.A. J.
Struct. Biol., 2000, 130, 184-208.
Butterfield, D.A.; Lauderback, C.M. Free Radical Biol. Med.,
2002, 32, 1050-1060.
Butterfield, D.A.; Castegna, A.; Lauderback, C.M.; Drake, J.
Neurobiology Aging., 2002, 23, 665-664.
Butterfield, D.A. Free Rad. Res., 2002, 36, 1307-1313.
Esterbauer, H.; Schaur, R.J.; Zollner. Free Rad. Biol. Med., 1991,
11, 81-128.
Subramaniam, R.; Roediger, F.; Jordan, B.; Mattson, M.P.; Keller,
J.N.; Waeg, G.; Butterfield, D.A. J. Neurochem., 1997, 69, 11611169.
Pocernich, C.B.; Cardin, A.L.; Racine, C.L.; Lauderback, C.M.;
Butterfield, D.A. Neurochem. International, 2001, 39, 141-149.
Butterfield, D.A.; Stadtman, E.R. Adv. Cell Aging Gerontol., 1997,
2, 161-191.
Beckman, J.S. Chem. Res. Toxicol., 1996, 9, 836-844.
Lafon-Cazal, M.; Pietri, S.; Culcasi, M.; Bochaert, J. Nature,1993,
364, 535-537.
Xie, Z.; Wei, M.; Morgan, T.E.; Fabrizio, P.; Han, D.; Finch, C.E.;
Longo, V.D. J. Neurosci., 2002, 22, 3484-3492.
Butterfield, D.A.; Yatin, S.M.; Varadarajan, S.; Koppal, T. Meth.
Enzymol., 1999, 309, 746-768.
Hensley, K.; Hall, N.; Subramaniam, R.; Cole, P.; Harris, M.;
Aksenov, M.; Aksenova, M.; Gabbita, S.P.; Wu, J.F.; Carney, J.M.;
Lovell, M.; Markesbery, W.R.; Butterfield, D.A. J. Neurochem.,
1995, 65, 2146-2156.
Smith, C.D.; Carney, J.M.; Starke-Reed, P.E.; Oliver, C.N.;
Stadtman, E.R.; Floyd, R.A.; Markesbery, W.R. Proc. Nat. Acad.
Sci. USA, 1991, 88, 10540-10543.
Smith, M.A.; Richey-Harris, P.; Sayre, L.M.; Beckman, J.S.; Perry
G. J. Neurosci., 1997, 17, 2653-2657.
Butterfield, D.A.; Koppal, T.; Subramaniam, R.; Yatin, S. Rev.
Neurosci., 1999, 10, 141-149.
Butterfield, D.A.; Castegna, A.; Drake, J.; Scapagnini, G.;
Calabrese, V. Nutr. Neurosci., 2002, 5, 229-239.
Yatin, S.M.; Varadarajan, S.; Butterfield, D.A. J. Alzheimer's Dis.,
2000, 2, 123-131.
2658 Current Medicinal Chemistry, 2003, Vol. 10, No. 24
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
Harris, M.E.; Hensley, K.; Butterfield, D.A.; Leedle, R.A.;
Carney, J.M. Exp. Neurol., 1995, 131, 193-202.
Yatin, S.M.; Aksenov, M.; Butterfield, D.A. Neurochem. Res.,
1999, 24, 427-435.
Butterfield, D.A.; Hensley, K.; Cole, P.; Subramaniam, R.;
Aksenov, M.; Aksenova, M.; Bummer, P.M.; Haley, B.E.; Carney,
J.M. J. Neurochem., 1997, 68, 2451-2457.
Butterfield, D.A.; Hensley, K.; Harris, M.; Mattson, M.; Carney, J.
Biochem. Biophys. Res. Commun., 1994, 200, 710-715.
Koppal, T.; Subramaniam, R.; Drake, J.; Prasad, M.R.; Butterfield,
D.A. Brain Res., 1998, 786, 270-273.
Bruce-Keller, A.J.; Begley, J.G.; Fu, W.; Butterfield, D.A.;
Bredesen, D.E.; Hutchins, J.B.; Hensley, K.; Mattson, M.P. J.
Neurochem., 1998, 70, 31-39.
Markesbery, W.R.; Lovell, M.A. Neurobiol. Aging, 1998, 19, 3336.
Lovell, M.A.; Xie, C.; Markesbery, W.R. Neurobiol. Aging, 2001,
22, 187-194.
Mark, R.J.; Lovell, M.A.; Markesbery, W.R.; Uchida, K.; Mattson,
M.P. J. Neurochem., 1997, 68, 255-264.
Lauderback, C.M.; Hackett, J.M.; Huang, F.F.; Keller, J.N.;
Szweda, L.I.; Markesbery, W.R.; Butterfield, D.A. J .
Neurochem., 2001, 78, 413-416.
Mark, R.J.; Fuson, K.S.; May, P.C. J. Neurochem., 1999, 72, 11461153.
Subramaniam, R.; Koppal, T.; Green, M.; Yatin, S.; Jordan, B.;
Butterfield, D.A. Neurochem. Res., 1998, 23, 1403-1410.
Mark, R.J.; Hensley, K.; Butterfield, D.A.; Mattson, M.P. J.
Neurosci., 1995, 15, 6239-6249.
Keller, J.N.; Mattson, M.P. Rev. Neurosci., 1998, 9, 105-116.
Mark, R.J.; Blanc, E.M.; Mattson, M.P. Mol. Neurobiol., 1996, 12,
211-224.
Keller, J.N.; Pang, Z.; Geddes, J.W.; Begley, J.G.; Germeer, A.;
Waeg, G.; Mattson, M.P. J. Neurochem., 1997, 69, 273-284.
Yatin, S.M.; Yatin, M.; Varadarajan, S.; Ain, K.B.; Butterfield,
D.A. J. Neurosci. Res., 2001, 63, 395-401.
Kelly, J.F.; Furukawa, K.; Barger, S.W.; Rengen, M.R.; Mark,
R.J.; Blanc, E.M.; Roth, G.S.; Mattson, M.P. Proc. Nat. Acad. Sci.
USA, 1996, 93, 6753-6758.
Masliah, E.; Alford, M.; DeTeresa, R.; Mallory, M.; Hansen, L.
Ann. Neurol., 1996, 40, 759-766.
Hensley, K.; Carney, J.; Mattson, M.; Aksenova, M.; Harris, M.;
Wu, J.F; Floyd, R.; Butterfield, D.A. Proc. Nat. Acad. Sci. USA,
1994, 91, 3270-3274.
Aksenov, M.Y.; Aksenova, M.V.; Harris, M.E.; Hensley, K.;
Butterfield, D.A.; Carney, J.M. J. Neurochem., 1995, 65, 18991902.
Butterfield, D.A.; Pocernich, C.B.; Drake, J. Drug Discovery Res.,
2002, 56, 428-437.
Butterfield, D.A.; Castegna, A.; Pocernich, C.B.; Drake, J.;
Scapagnini, G.; Calabrese, V. J. Nutr. Biochem., 2002, 13, 444461.
Lauderback, C.M.; Kanski, J.; Hackett, J.M.; Maeda, N.; Kindy,
M.S.; Butterfield, D.A. Brain Research, 2002, 924, 90-97.
Yao, Z.; Drieu, K.; Szweda, L.I.; Papadopoulos, V. Brain Res.,
1999, 847, 203-210.
Lockhart, B.P.; Benicourt, C.; Junien, J.L.; Privat, A. J. Neurosci.
Res., 1994, 39, 494-505.
Pike, C.J.; Ramezan-Arab, N.; Cotman, C.W. J. Neurochem.,
1997, 69, 1601-1611.
Hensley, K.; Maidt, M.L.; Yu, Z.; Sang, H.; Markesbery, W.R.;
Floyd, R.A. J. Neurosci., 1998, 18, 8126-8132.
Beal, M.F. Free Rad. Biol. Med., 2002, 32, 797-803.
Koppal, T.; Drake, J.; Yatin, S.; Jordan, B.; Varadarajan, S.;
Bettenhausen, L.; Butterfield, D.A. J. Neurochem., 1999, 72, 310317.
Mattson, M.P.; Barger, S.W.; Cheng, B.; Lieberburg, I.; SmithSwintosky, V.L.; Ryel, R.E. Trends Neurosci., 1993, 16, 409-414.
Yatin, S.M.; Link, C.D.; Butterfield, D.A. Neurobiol. Aging, 1999,
20, 325-330.
Varadarajan, S.; Kanski, J.; Aksenova, M.; Lauderback, C.;
Butterfield, D.A. J. Am. Chem. Soc., 2001, 123, 5625-5631.
Shao, H.; Jao, S-C.; Ma, K.; Zagorski, M.G. J. Mol. Biol., 1999,
285, 775-773.
Coles, M.; Bicknell, W.; Watson, A.A.; Fairlie, D.P.; Craik, D.J.
Biochemistry, 1998, 37, 11064-11077.
D. Allan Butterfield
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
[82]
[83]
[84]
[85]
[86]
[87]
[88]
Pogocki, D.; Schoneich, C. Chem. Res. Toxicol., 2002, 15, 408418.
Kanski, J.; Aksenova, M.; Schoneich, C.; Butterfield, D.A. Free
Rad. Biol. Med., 2002, 32, 1205-1211.
Butterfield, D.A.; Kanski, J. Peptides, 2002, 23, 1299-1309.
Kanski, J.; Aksenova, M.; Butterfield, D.A. Neurotox. Res., 2002,
4, 219-223.
Huang, X.; Atwood, C.S.; Hartshorn, M.A.; Multhaup, G.;
Goldstein, L.E.; Scarpa, R.C.; Cuajungco, M.P.; Gray, D.N.; Lim,
J.; Moir, R.D.; Tanzi, R.E.; Bush, A.I. Biochemistry, 1999, 38,
7609-7616.
Huang, X.; Cuajungco, M.P.; Atwood, C.S.; Hartshorn, M.A.;
Tyndall, J.D.; Hanson, G.R.; Stokes, K.C.; Leopold, M.; Multhaup,
G.; Goldstein, L.E.; Scarpa, R.C.; Saunders, A.J.; Lim, J.; Moir,
R.D.; Glabe, C.; Bowden, E.F.; Masters, C.L.; Fairlie, D.P.; Tanzi,
R.E.; Bush, A.I. J. Biol. Chem., 1999, 274, 37111-37116.
Curtain, C.C.; Ali, F.; Volitakis, I.; Cherny, R.A.; Norton, R.S.;
Beyreuther, K.; Barrow, C.J.; Masters, C.L.; Bush, A.I.; Barnham,
K.J. J. Biol. Chem., 2001, 276, 20466-20473.
Butterfield, D.A.; Kanski, J. Mech. Ageing Develop., 2001, 122,
945-962
Podlisny, M.B.; Walsh, D.M.; Amarante, P.; Ostaszewski, B.L.;
Stimson, E.R.; Maggio, J.E.; Teplow, D.B.; Selkoe, D.J.
Biochemistry, 1998, 37, 3602-3611.
Kuo, Y.M.; Emmerling, M.R.; Vigo-Pelfrey, C.; Kasunic, T.C.;
Kirkpatrick, J.B.; Murdoch, G.H.; Ball, M.J.; Roher, A.E. J. Biol.
Chem., 1996, 271, 4077-81.
Walsh, D.M.; Hartley, D.M.; Kusumoto, Y.; Fezoui, Y.; Condron,
M.M.; Lomakin, A.; Benedek, G.B.; Selkoe, D.J.; Teplow, D.B. J.
Biol. Chem., 1999, 274, 25945-25952.
Roher, A.E.; Chaney, M.O.; Kuo, Y.M.; Webster, S.D.; Stine,
W.B.; Haverkamp, L.J.; Woods, A.S.; Cotter, R.J.; Tuohy, J.M.;
Krafft, G.A.; Bonnell, B.S.; Emmerling, M.R. J. Biol. Chem., 1996,
271, 20631-20635.
Lambert, M.P.; Barlow, A.K.; Chromy, B.A.; Edwards, C.; Freed,
R.; Liosatos, M.; Morgan, T.E.; Rozovsky, I.; Trommer, B.; Viola,
K.L.; Wals, P.; Zhang, C.; Finch, C.E.; Krafft, G.A.; Klein, W.L.
Proc. Natl. Acad. Sci. USA, 1998, 95, 6448-6453.
Aksenov, M.Y.; Aksenova, M.V.; Butterfield, D.A.; Hensley, K.;
Vigo-Pelfrey, C.; Carney, J.M. J. Neurochem., 1996, 66, 20502056.
Walsh, D.M.; Hartley, D.M.; Kusumoto, Y.; Fezoui, Y.; Condron,
M.M.; Lomakin, A.; Benedek, G.B.; Selkoe, D.J.; Teplow, D.B. J.
Biol. Chem., 1999, 274, 25945-25952.
Walsh, D.M.; Klyubin, I.; Fadeeva, J.V.; Cullen, W.K.; Anwyl, R.;
Wolfe, M.S.; Rowan, M.J.; Selkoe, D.J. Nature, 2002, 416, 535539.
Oda, T.; Wals, P.; Osterburg, H.H; Johnson, S.A.; Pasinetti, G.M.;
Morgan, T.E.; Rozovsky, I.; Stine, W.B.; Snyder, S.W.; Holzman,
T.F.; Krafft, G.A.; Finch, C.E. Exp. Neurol., 1995, 136, 22-31.
Drake, J.; Link, C.D.; Butterfield, D.A. Neurobiol. Aging, 2003,
24, 415-420.
Aksenov, M.; Aksenova, M.; Butterfield, D.A.; Markesbery, W.R.
J. Neurochem., 2000, 74, 2520-2527.
Aksenov, M.Y.; Aksenova, M.V.; Butterfield, D.A.; Geddes, J.W.;
Markesbery, W.R. Neuroscience, 2001, 103, 373-383.
Castegna, A.; Aksenov, M.; Aksenova, M.; Thongboonkerd, V.;
Klein, J.B.; Pierce, W.M.; Booze, R.; Markesbery, W.R.;
Butterfield, D.A. Free Rad. Biol. Med., 2002, 33, 562-571.
Castegna, A.; Aksenov, M.; Aksenova, M.; Thongboonkerd, V.;
Klein, J.B.; Pierce, W.M.; Booze, R.; Markesbery, W.R.;
Butterfield, D.A. J. Neurochem., 2002, 82, 1524-1532.
Castegna, A.; Thongboonkerd, V.; Klein, J.B.; Lynn, B.;
Markesbery, W.R.; Butterfield, D.A. J. Neurochem., 2003, 85,
1394-1401.
Butterfield, D.A.; Castegna, A. Amino Acids 2003; in press.
Butterfield, D.A.; Boyd-Kimball, D.; Castegna, A. J. Neurochem.,
2003, 86, 1313-1327.
Reisberg, B.; Doody, R.; Stoffler, A.; Schmitt, F.; Ferris, S.;
Mobius, H.J. New Engl. J. Med., 2003, 348, 1333-1341.
Shringarpure, R.; Davies, K.J. Free Rad. Biol. Med., 2002, 32,
1084-1089.
Lee, M.; Hyun, D.H.; Marshall, K.A.; Ellerby, L.M.; Bredesen,
D.E.; Jenner, P.; Halliwell, B. Free Rad. Biol. Med., 2001, 31,
1550-1559.
Coleman, P.D.; Flood, D.G. Neurobiol. Aging, 1987, 8, 521-545.
Mechanisms and Consequences
[89]
[90]
Sayre, L.M.; Smith, M.A.; Perry, G. Curr. Med. Chem., 2001, 8,
721-738.
Butterfield, D.A.; Howard, B.J.; LaFontaine, M.A. Curr. Med.
Chem., 2001, 8, 815-828.
Current Medicinal Chemistry, 2003, Vol. 10, No. 24
[91]
[92]
2659
Drake, J.; Kanski, J.; Varadarajan, S.; Tsoras, M.; Butterfield,
D.A. J. Neurosci. Res., 2002, 68, 776-784.
Drake, J.; Akenova, M.; Butterfield, D.A. J. Neurosci. Res., 2003,
in press.