Lippincott—Raven Publishers, Philadelphia
© 1997 International Society for Neurochemistry
4-Hydroxynonenal-Derived Advanced Lipid Peroxidation End
Products Are Increased in Alzheimer’ s Disease
Lawrence M. Sayre, Dawn A. Zelasko, *peggy L. R. Harris, *George Perry,
Robert G. Salomon, and *Mark A. Smith
Departments of Chemistry and *pathology, Case Western Reserve University, Cleveland, Ohio, U.S.A.
Abstract: Recent studies have demonstrated oxidative
damage is one of the salient features of Alzheimer’s dis-
ease (AD). In these studies, glycoxidation adduction to
and direct oxidation of amino acid side chains have been
demonstrated in the lesions and neurons of AD. To ad-
dress whether lipid damage may also play an important
pathogenic role, we raised rabbit antisera specificfor the
lysine-derived pyrrole adducts formed by lipid peroxida-
tion-derived 4-hydroxynonenal (HNE). These antibodies
were used in immunocytochemical evaluation of brain
tissue from AD and age-matched control patients. HNE-
pyrrole immunoreactivity not only was identified in about
half of all neurofibrillary tangles, but was also evident in
neurons lacking neurofibrillary tangles in the AD cases.
In contrast, few senile plaques were labeled, and then
only the dystrophic neurites were weakly stained,
whereas the amyIoid-~deposits were unlabeled. Age-
matched controls showed only background HNE-pyrrole
immunoreactivity in hippocampal or cortical neurons. In
addition to providing further evidence that oxidative
stress-related protein modification is a pervasive factor
in AD, the known neurotoxicity of HNE suggests that lipid
peroxidation may also play a role in the neuronal death
in ADthatunderlies cognitive deficits. KeyWords: Alzhei-
mer’s disease—Hydroxynonenal— Neurofibrillary tan-
gles—Oxidative stress— ~- protein.
J. Neurochem. 68, 2092—2097 (1997).
Several lines of evidence implicate oxidative stress
in neurodegeneration (Bowling and Beal, 1995), par-
ticularly in Alzheimer’s disease (AD) (Smith et al.,
1995, 1996; Good et al., 1996). Some of these studies
have focused on demonstrating markers of oxidative
stress on the fibrous structures that characterize AD:
neurofibrillary tangles (NFTs) and senile plaques
(SPs). These markers include advanced glycation end
products (Ledesma et a!., 1994; Smith et al., 1994a;
Vitek et al., 1994; Yan et al., 1995), nitrotyrosine
(Good et al., 1996; Smith et al., 1997), free carbonyls
(Smith et al., 1996), and heme oxygenase-1 (Smith
et a!., 1994b), an enzyme active in the cellular re-
sponse to oxidative stress. Our laboratories have pro-
moted the theory that because these markers appear
at the earliest recognizable stage of the pathological
lesions, the formation of paired helical fi1ament-~ and
diffuse SPs, they are not merely artifacts of oxidative
modification of long-lived structures, but instead con-
tribute to disease progression.
Oxidative damage to lipids, i.e., lipid peroxidation,
results in the production of several reactive aldehydes
capable of modifying protein, including malondialde-
hyde and 4-hydroxynonenal (HNE), the most preva-
lent cytotoxic aldehyde yet identified (Esterbauer et
al., 1991). Malondialdehyde has been detected immu-
nochemically in AD (Yan et al., 1994), and attention
has been called to an effect of HNE on apolipoprotein
E in vitro (Montine et al., 1996b) and on cytoskeletal
proteins in cell culture (Montine et al., 1996a), but
therehave been no publishedreports on a role ofHNE
in AD in vivo.
HNE reacts with protein nucleophiles to give Mi-
chael adducts, although this is reversible in the case
of lysine (Nadkarni and Sayre, 1995), a prominent
amino acid in r, the primary constituent in NFTs. For
this reason the commonly used HNE—Michael adduct
antibodies (Uchida et al., 1995) may be inappropriate
for characterizing HNE modification of lysine-rich
proteins. Recently, we reported that the reaction of
HNE with lysine c-amino groups can proceed alterna-
tively through an initial Schiffbase, ultimately produc-
ing, among several other as yet unidentified advanced
lipid peroxidation end products (ALE), a 2-pentylpyr-
role (Sayre et al., 1993) that can be detected immuno-
chemically on HNE-treated proteins (Sayre et al.,
1996). Here, we report increased levels of HNE-pyr-
role adducts in AD.
Resubmitted manuscript received January 15, 1997; acceptedJan-
uary 15, 1997.
Address correspondence and reprint requests to Dr. M. A. Smith
at Institute of Pathology, Case Western Reserve University, Cleve-
land, OH 44106, U.S.A.
Abbreviations used: AD, Alzheimer’s disease; ALE, advanced
lipid peroxidation end products; apoE, apolipoprotein E; BSA, bo-
vine serumalbumin; HNE, 4-hydroxynonenal; KLH, keyhole limpet
hemocyanin; NFT, neurofibrillary tangle; ON, 4-oxononanal; PPC,
6-(2-pentylpyrrol-l-yl)caproic acid; SP, senile plaque.
LIPID PEROXIDATION IN ALZHEIMER ‘S DISEASE2093
MATERIALS AND METHODS RESULTS
Brain tissue was obtained postmortem from diagnosed
patients with AD (Khatchaturian, 1985) ranging from 68 to
94 years old and nondemented cases (53—81 years old).
Hippocampus (including the surrounding subiculum, ento-
rhinal cortex, and adjacent temporal neocortex), frontal cor-
tex, andcerebellum was fixed in methacarn(methanol/chlo-
roformlacetic acid, 6:3:1 by volume) for 16 h, dehydrated
through graded ethanol followed by xylene, and embedded
in paraffin. Sections 6 /.tm thick were cut and mounted on
silane (Sigma)-coated standard glass microscope slides for
immunocytochemical use. Two different antisera to HNE-
pyrrole were used (Sayre et al., 1996): One was raised to
keyhole limpet hemocyanin (KLH) modifiedby 4-oxonona-
nal (ON), which generates as the major product the same
2-pentylpyrrole on lysine c-amino groups formed (in low
yield)on directexposure to HNE butwhich also gives some
side products; the other was raised to KLH conjugated with
which generates the pure pyrrole epitope but with a longer
tether than when the pyrrole is formed from HNE. Thus, the
coincidence of labeling by both antisera was deemed to be
most convincing of specific immunocytochemical recogni-
tion of the HNE-derived pyrrole.
Following deparaffinization with xylene, sections were
hydrated through graded ethanol. Endogenous peroxidase
activity in the tissue was eliminated by a 30-mm incubation
with 3% H2O2 in methanol, and nonspecific binding sites
were blocked in a 30-mm incubation with 10% normal goat
serum in Tris-buffered saline (150 mM Tris-HC1 and 150
mM NaC1, pH 7.6). HNE-pyrrole antisera were used at a
dilution from 1:10 to 1:1,000; we found a 1:100 dilution was
optimal, and this dilution was used throughout for immuno-
staining with theexception ofthe immunoadsorption experi-
ment (see below). The immunostaining procedure followed
was theperoxidase—antiperoxidase procedure, and the stain-
ing was developed under a microscope by using 3,3 ‘-diami-
nobenzidine at 0.75 mg/ml in 0.015% H2O2 and, 50 mM
Tris-HC1, pH 7.6. The sections were dehydrated through
ethanol and xylene solutions and then mounted in Permount
(Fisher). The location of NFTs and SPs was determined
following immunostaining by counterstaining the sections
with Congo red and viewing them under crossed polarized
beled NFTs, ON-KLH/PPC-KLH-immunolabeled neurons
with NFTs, and ON-KLH/PPC-KLH-immunoiabeled neu-
rons without NFTs was determined for an area of 1 mmfrom the subiculum—Somrner’s sector for each ofthe cases.
The data were used to calculate the percentage of NFTs
labeled by anti-HNE-pyrrole as well as the percentage of
anti-HNE-pyrrole-inimunolabeled neurons that contained
NFTs (Table 1). Antiserum to r (1:1,000) (Perry et al.,
1991) was also used on adjacent serial sections to verifythe
validity of the Congo red staining to detect NFTs and SPs.
The specificity ofboth HNE-pyrrole antibodies was con-
firmed by omitting the primary antibody or by performing
an adsorption.The two antisera toHNE-pyrrole werediluted
1:150 and then incubated with the bovine serum albumin
(BSA)-derived antigens (ON-BSA or PPC-BSA) (Sayre et
al., 1996) at 0.1 mg/mi for 16 h at 4°C before immunocyto-
chemistry. The results of adsorbed immunostaining were
compared with those for adjacent sections labeled with a
1:150 dilution of unadsorbed antibody.
Two related antisera to HNE-pyrrole (Sayre et al.,
1996) were used to determine sites of HNE modifica-
tion in tissue from the hippocampus (including subicu-
lum, entorhinal cortex, and temporal neocortex) in
cases of AD (n = 19; 68—94 years old; postmortem
interval, 2—22.5 h) and from nondementedcontrols (n
= 4; 53—81 years old; postmortem interval, 4—21 h)
(Table 1). The two antisera to HNE-pynole gave es-
sentially the same results and recognized neurons in
AD but not control cases (Fig. 1). HNE-pyrrole was
detected on NFTs (Fig. 1) and the surrounding cyto-
plasm in AD cases, but less than half of the neurons
labeled by HNE-pyrrole contained NFTs (Fig. 1A and
C and Table 1). Moreover, although intraneuronal
NFTs (Fig. 1A and C) were labeled, as were occa-
sional extracellularNFTs (Fig. 1C), many of the latter
were unlabeled. These aspects account for a great deal
of the case-to-case variation that was observed (Table
1), because the ratio of intracellular versus extracellu-
lar NFTs varied among the cases. Moreover, compari-
sonwith adjacentsections immunolabeledwith an anti-
serum to i indicated, as did the Congo red staining,
that many butnot all of the NFTs recognizedby antise-
rum to r were also labeled by anti-HNE-pyrrole. In
contrast to NFTs, only a fewpercent of the SPs in the
cases of AD showed any staining, and, when noted,
the amyloid-~core was unstained, whereas the sur-
rounding dystrophic neurites were stained just above
background levels (Fig. ID).
Although some case-to-case variability was noted
(Table 1), the pattern of staining was similar in all the
AD and control cases studied. There was no striking
correlation between staining intensity in AD or control
cases and (a) age, (b) postmortem interval, (c) clinical
duration of disease, (d) percentage of neurons with
NFTs, or (e) apolipoprotein Egenotype. Findings were
similar for the frontal cortex, where neuronal staining
was only seen in the AD cases, whereas in cerebellum
there was no difference between AD and control cases.
In control cases (Fig. 1B), the intensity of immuno-
labeling was at background levels, comparable or
lower in intensity to the immunoadsorbed antiserum
(compare Fig. lB with 2B), and showed no relation-
ship to age. In all cases, irrespective of age and disease
presence, relatively strong anti-HNE-pyrrole immuno-
labeling was seen on major blood vessels, consistent
with active steady-state lipid peroxidation occurring in
the basement membrane of this tissue.
In cases where the two HNE-pyrrole antisera (ON-
KLH- and PPC-KLH-derived) were compared on se-
rial sections, the pattern of staining was essentially the
same. Additional evidence ofthe specificity of labeling
by the two antisera forHNE-pyrrole was demonstrated
by (a) omission oftheprimary antisera and (b) adsorp-
tion ofeach ofthe primary antisera withboth the corre-
sponding and noncorresponding BSA-derived antigens
(Fig. 2). Adsorption of staining by the ON-KLH-de-
.1. Neurochem., Vol. 68, No. 5, 1997
2Q94L. M. SAYRE ET AL.
FIG. 1. Sections from the brain of (A) a patient with AD and (B) a nondemented control immunostained with the ON-KLH-derived
antiserum. Many NFTs are labeled (arrows); as well, neurons lacking NFTs (arrowheads) are labeled in AD (A), whereas no specific
staining is seen in the control case (B). Inset: High-magnification view of a neuron containing a NFT recognized by the ON-KLH-
derived antiserum. C: Another high-magnification view shows the three types of neurons recognized by the ON-KLH antiserum—
neurons containing NFTs(large arrowhead), extracellular NFTs (arrow), and a neuron lacking an NFT (small arrowhead). D: Few SPs
were recognized by PPC-KLH antiserum, and, when stained, theamyloid-f3 core (*)was unstained whereas thesurrounding dystrophic
neurites were only weakly immunolabeled. Bars = 50 pm in A, B, and 0, 25 ~sm in C, and 25 ~sm in inset in A.
rived antibody was complete using eitherBSA-denved
antigen, whereas adsorption of staining by the PPC-
KLH-derived antibody was best achieved with the cor-
responding PPC-BSA antigen. The staining intensity
and adsorption results are consistent with our previous
antibody ELISA characterizations showing that the
PPC-KLH antibody recognizes a more extensive epi-
tope (Sayre et a!., 1996).
HNE, ahighly reactive productoflipid peroxidation,
modifies proteins through formation of Michael ad-
ducts on His and Cys side chains and through more
complex reactions with Lys, leading in part to stable
adducts we refer to as ALE (Sayre et al., 1993, 1996;
Nadkarni and Sayre, 1995). Based on the high lysine
content of ~- and neurofilaments, we thought that lysine
adducts of HNE generated as a result ofoxidative dam-
age to membranes (Praprotnik et al., 1996) might be
associated with NFTs. Here, using two distinct antisera
with the immunogenic commonality of recognition of
an HNE-derived 2-pentylpyrrole, we demonstrate that
levels of HNE-lysine adducts are increased in AD not
only in NFT-containing neurons but also in “appar-
ently” normal pyramidal neurons in the AD hippocam-
paltissue sections. The coincidentoverlap ofimmunola-
beling by theON-KLH and PPC-KLHantisera indicates
that we are localizing the HNE-derived pynole, an ad-
vanced lipid peroxidation end product, rather than other
HNE products that might also form. It is of note that
J. Neurochem.. Vol. 68, No. 5, 1997
LIPID PEROXIDATION IN ALZHEIMER ‘S DISEASE2095
FIG. 2. Adjacent serial sections from the brain of a case of AD immunolabeled with (A) the PPC-KLH antiserum (1:150) and (B) the
same antiserum preadsorbed with PPC-BSA. Note neuronal reactivity isadsorbed by theantigen. A landmark blood vessel isindicated
(*). Bar = 50 ~sm.
HNE pyrrole-positive hippocampal neurons were spe-
cificfor AD, as similarlylabeled neurons were not found
in sections taken from control cases. These findings
were reproduced in another region, the frontal cortex,
which, like the hippocampal formation, is affected by
AD pathology. In marked contrast, and supporting the
specificity oflipoperoxidative modifications with brain
regional susceptibility to pathology, there were no such
differences in neuronal labeling between cerebellar tis-
sue taken from AD or control patients.
TABLE 1. Details of the cases of AD and controls (C) used in the present study
and the immunolabeling pattern notedfor each with the ON-KLH antiserum
neurons with NFTs
The percentage of HNE-positive neurons that alsocontained NFTs varied considerably, as did the percentage ofNFTs that werealso HNE-
positive. The latter is likely a reflection ofthe case-to-case heterogeneity in the proportion ofintracellular versus extracellular NFFs because
the latter were, for the most part, negative for HNE. N/A, not available;
—, negative; +, labeled; ++, intensely labeled; APOE, apoE gene.
J. Neurochem., Vol. 68, No. 5, 1997
2096L. M. SAYRE ET AL.
Quantification of the numbers ofHNE-pyrrole—pos-
itive NFTs revealed that although the vast majority of
intracellular NFTs were labeled, the converse was true
for extracellular NFTs. These findings were verified
by both Congo red histochemistry and r immunocyto-
chemistry. Therefore, a greatdeal of case-to-case vari-
ability was found in the percentage of anti-HNE-pyr-
role—labeled lesions reflecting the heterogeneous pro-
portions of intracellular versus extracellular NFTs.
Nonetheless, the apparent distinction between intracel-
lular and extracellular NFTs indicates that, once ex-
posed to the extracellular milieu, the HNE-pyrrole is
either further modified or, instead, is removed.
It is important that our evidence showing oxidative
stress-induced modifications in non—NFT-containing
neurons, i.e., the fact that less than half of the HNE-
pyrrole—positive neurons contain NFTs, parallels other
specific markers ofoxidative damage, including perox-
ynitrite-mediated damage (Smith et a!., 1997) and free
carbonyls (Smith et al., 1996), as well as antioxidant
protectants (M. A. Smith et al., submitted for publica-
tion), which all show a similar, if not identical, neu-
ronal and pathological distribution. Whether these
changes correlatewithearly cytoskeletal changes, spe-
cific neuronal subpopulations, or specific regions of
oxidative stress willrequire further analysis, but, none-
theless, these findings do indicate that oxidative stress-
related events, including the lipoperoxidative HNE-
type modifications reported here, could be important
initiators as well as contributors to cytoskeletal pathol-
ogy and neuronal degeneration.
The fact that SP amyloid does not exhibit markers
for direct oxidation or, as shown here, for HNE-de-
rived-pyrrole is interesting in light of the positive im-
munoreactivity with markers of glycation, i.e., pyrra-
line and pentosidine. Althoughthis apparent distinction
could result from technical aspects of immunocyto-
chemistry, i.e., antigen accessibility, it more likelyre-
flects important pathophysiologic differences in the
chemical type ofoxidative stress-inducedmodification,
i.e., degree and/or stability of modification, between
intracellular and extracellular proteins.
The presence of oxidative posttranslational modifi-
cations can arise as a consequence of the long half-
life ofproteins, for example, demonstrated here by the
association of lipid peroxidation adducts with cerebral
blood vessels in all cases studied, but it is of impor-
tance that we noted no correlations between the degree
of HNE modification and age, postmortem interval,
clinical duration of disease, or percentage of neurons
with NFTs. Therefore, HNE modification is a conse-
quence of the disease and, furthermore, is an early
consequence ofthe pathogenic disease process. In light
of the current focus on the apolipoprotein E apoE4
genotype as arisk factorin AD (Strittmatterand Roses,
1995), we have been investigating whether the pres-
ence oftheoxidative stressmarkers wedetectimmuno-
chemically mightcorrelatewith apoE genotype. In this
study, for those cases where genotype information was
known (Table 1) we found no obvious correlation of
HNE-pyrrole immunostainingwith any particular apoE
allele, including apoE4. Thislatter finding is incontrast
to a related study (T. J. Montine et al., personal com-
munication) where (using a different polyclonal anti-
body) an association of HNE-pyrrole immunoreactiv-
ity with apoE4 was found among a limited number of
AD cases examined. Clarification of a possible link
between HNE modification and apoE genotype will
require further study.
In conclusion, we show not only that NFTs, but also
neurons in cases ofAD, bear immunochemical markers
for an advanced lipid peroxidation end product. These
findings reinforce previous evidence that AD is associ-
ated with pervasive oxidative stress. Moreover, it is
intriguing to consider the role HNE may play in AD
pathogenesis because recent studies show that HNE
rapidly modifies cytoskeletal proteins, including r and
apoE, and, significantly, results in neurotoxicity (Mon-
tine et al., 1996a,b). Ofconsequence is that HNEleads
to a 40-fold induction of the mRNA for heme oxy-
genase-1 (Basu-Modak et al., 1996), a protein pre-
viously found to be induced in NFT-containing neu-
rons in AD and that, in itself, might contribute to neu-
rodegeneration (Smith et a!., 1994b; R. B. Petersen
and M. A. Smith, unpublished data). Therefore, taken
together with the findings reported here, these findings
make it tempting to consider a major role for HNE in
the degeneration of neurons in AD.
Acknowledgment: This work was supported by grants
HL 53315 and AG09287 from the National Institutes of
Health, the American Health Assistance Foundation, pilot
grants from the CaseWesternReserve University Alzheimer
and Claude Pepper Centers, a summer fellowship to D.A.Z.
fromthe Howard HughesMedical Institute, and aRuth Salta
Student Research Fellowship to D.A.Z. We thank Sandra
Siedlak and Guozhang Xu for technical assistance.
Basu-Modak S.,LuscherP., andTyrrell R. M. (1996) Lipidmetabo-
lite involvement in theactivation ofthehumanhemeoxygenase-
1 gene. Free Radic. Biol. Med. 20, 887—897.
Bowling A. C. and Beal M. F. (1995) Bioenergetic and oxidative
stress in neurodegenerative diseases. Life Sci. 56, 1151—1171.
Esterbauer H., Schaur R. J., and Zoilner H. (1991) Chemistry and
biochemistry of4-hydroxynonenal, malonaldehyde and related
aldehydes. Free Radic. Biol. Med. 11, 81—128.
GoodP. F., WernerP., HsuA.,Olanow C. W., andPen D.P. (1996)
Evidence for neuronaloxidative damagein Alzheimer’s disease.
Am. J. Pathol. 149, 21—28.
Khachaturian Z. S. (1985) Diagnosis ofAlzheimer’s disease. Arch.
Neurol. 42, 1097—1105.
Ledesma M. 0., Bonay P., Colaco C., and Avila J. (1994) Analysis
ofmicrotubuie-associated protein tau giycation in paired helical
filaments. J. Biol. Chem. 269, 21614—21619.
Montine T. J., Amamath V., Martin M. E., Strittmatter W. J., and
Graham 0. G. (1996a)E-4-Hydroxy-2-nonenal is cytotoxic and
cross-links cytoskeletal proteins in P19 neuroglial cultures. Am.
J. Pathol. 148, 89—93.
Montine T. J., Huang 0. Y., Valentine W. M., Amamath V., Saun-
ders A., WeisgraberK. H., Graham D. G., and StrittmatterW. J.
J. Neurochem., Vol. 68, No. 5, 1997
LIPID PEROXIDATION IN ALZHEIMER ‘S DISEASE2097 Download full-text
(1996b) Crosshnking of apolipoprotein Eby products of lipid
peroxidation. J. Neuropathol. Exp. Neurol. 55, 202—210.
Nadkami 0. V. and Sayre L. M. (1995) Structuraldefinition ofearly
lysine and histidine adduction chemistry of4-hydroxynonenal.
Chem. Res. Toxicol. 8, 284—291.
Perry G., Kawai M., Tabaton M., Onorato M., Mulvihill P., Richey
P., Morandi A., Connolly J. A., and Gambetti P. (1991) Neu-
nopil threads ofAlzheimer’s disease show a marked alteration
ofthe normal cytoskeleton. J. Neurosci. 11, 1748—1755.
Pnaprotnik 0., Smith M. A., Richey P. L., Vinters H. V., and Perry
G. (1996) Plasma membrane fragility in dystrophic neurites in
senile plaques of Alzheimer’s disease: an index of oxidative
stress. Acta Neuropathol. (Ben.) 91, 1—5.
Sayre L. M., Arora P. K., Iyer R. S., and Salomon R. 0. (1993)
Pyrrole formation from4-hydroxynonenal and primary amines.
Chem. Res. Toxicol. 6, 19—22.
Sayre L. M., Sha W., Xu 0., Kaur K.,Nadkarni D., Subbanagounder
G., and Salomon R. 0. (1996) Immunochemical evidence sup-
porting 2-pentylpyrrole formation on proteins exposed to 4-
hydroxy-2-nonenai. Chem. Res. Toxicol. 9, 1194—1201.
Smith M. A., Taneda S., Richey P. L., Miyata S., Yan S.-D., Stern
D., Sayre L. M., Monnier V. M., and Perry G. (1994a) Ad-
vanced Maillard reaction products are associated with Alzhei-
mer disease pathology. Proc. Nati. Acad. Sci. USA 91, 5710—
Smith M. A., Kutty R. K., RicheyP. L., Yan S.-D., Stern0., Chader
0. J., Wiggert B., Petersen R. B., and Perry 0. (1994b) Heme
oxygenase-l is associated with the neurofibrillary pathology of
Aizheimer’s disease. Am. J. Pathol. 145, 42—47.
Smith M. A., Sayre L. M., Monnier V. M., and Perry 0. (1995)
Radical AGEing in Aizheimer’s disease. Trends Neurosci. 18,
Smith M. A., Perry G., Richey P. L., Sayre L. M., Anderson V. E.,
Beal M. F., and KowaliN. (1996) Oxidative damagein Alzhei-
mer’s. Nature 382, 120—121.
Smith M. A., Harris P.L. R., Sayre L. M., Beckman J. S., and Perry
0. (1997) Widespread peroxynitnite mediated damage in Aiz-
heimer disease. J. Neurosci. (in press).
Strittmatter W. J. and Roses A. D. (1995) Apolipoprotein E and
Alzheimer disease. Proc. Nat!. Acad. Sci. USA92, 4725—4727.
Uchida K., Itakura K., Kawakishi S., Hiai H., Toyokuni S., and
Stadtman E. R. (1995) Characterization of epitopesrecognized
by 4-hydroxy-2-nonenal specific antibodies.Arch. Biochem.Bi-
ophys. 324, 241—248.
Vitek M. P., Bhattcharya K., Glendening J. M., Stopa E., Vlassara
H., Bucala R., Manogue K., and Cerami A. (1994) Advanced
glycation end products contribute to amyloidosis in Alzheimer
disease. Proc. Nat!. Acad. Sci. USA 91, 4766—4770.
Yan S.-D., Chen X., Schmidt A.-M., Brett J., Godman 0., Zou
Y.-S., Scott C. W., Caputo C., Frappier T., Smith M. A., Perry
G., Yen S-H., and Stern D. (1994) Glycated tau protein in
Alzheimerdisease: a mechanismfor induction ofoxidant stress.
Proc. Nat!. Acad. Sci. USA 91, 7787—779 1.
Yan S.D., Yan S. F., Chen X., Fu J., Chen M., Kuppusamy P.,
Smith M. A., Perry G., Godman 0. C., Nawroth P., Zweier
J. L., and Stern D. (1995) Non-enzymatically glycated tau in
Alzheimer’ s disease induces neuronal oxidant stress resulting
in cytokine gene expression and release ofamyloid /3-peptide.
Nat. Med. 1, 693—699.
J. Neurochem., Vol. 68. No. 5, 1997