Memory loss in old rats is associated with brain
mitochondrial decay and RNA兾DNA oxidation:
Partial reversal by feeding acetyl-
, Elizabeth Head
, Afshin M. Gharib*
, Wenjun Yuan*, Russell T. Ingersoll*, Tory M. Hagen
Carl W. Cotman
, and Bruce N. Ames*
*Division of Biochemistry and Molecular Biology, University of California, Berkeley, CA 94720;
Children’s Hospital Oakland Research Institute, 5700 Martin
Luther King, Jr., Way, Oakland, CA 94609;
Institute for Brain Aging and Dementia, University of California, Irvine, CA 92697-4540; and
Biochemistry and Biophysics, Linus Pauling Institute, Oregon State University, Corvallis, OR 97331
Contributed by Bruce N. Ames, December 29, 2001
Accumulation of oxidative damage to mitochondria, protein, and
nucleic acid in the brain may lead to neuronal and cognitive
dysfunction. The effects on cognitive function, brain mitochondrial
structure, and biomarkers of oxidative damage were studied after
feeding old rats two mitochondrial metabolites, acetyl-L-carnitine
(ALCAR) [0.5% or 0.2% (wt兾vol) in drinking water], and兾or R-
lipoic acid (LA) [0.2% or 0.1% (wt兾wt) in diet]. Spatial memory was
assessed by using the Morris water maze; temporal memory was
tested by using the peak procedure (a time-discrimination proce-
dure). Dietary supplementation with ALCAR and兾or LA improved
memory, the combination being the most effective for two differ-
ent tests of spatial memory (P < 0.05; P < 0.01) and for temporal
memory (P < 0.05). Immunohistochemical analysis showed that
oxidative damage to nucleic acids (8-hydroxyguanosine and
8-hydroxy-2ⴕ-deoxyguanosine) increased with age in the hip-
pocampus, a region important for memory. Oxidative damage to
nucleic acids occurred predominantly in RNA. Dietary administra-
tion of ALCAR and兾or LA signiﬁcantly reduced the extent of
oxidized RNA, the combination being the most effective. Electron
microscopic studies in the hippocampus showed that ALCAR
and兾or LA reversed age-associated mitochondrial structural decay.
These results suggest that feeding ALCAR and LA to old rats
improves performance on memory tasks by lowering oxidative
damage and improving mitochondrial function.
emory, i.e., performance on memory tasks, declines with
age in animals. In the case of age-related human neuro-
degenerative diseases, such as Alzheimer’s disease (AD), the
deficit can be severe (1–4). Memory loss is accompanied but not
necessarily caused by accumulation of oxidative damage to lipids,
proteins, and nucleic acids, and by mitochondrial decay, all of
which can disrupt neuronal function (5–10).
-lipoic acid (LA) is a coenzyme that is involved in carbo-
hydrate utilization necessary for the production of ATP in
mitochondria; it is reduced in mitochondria to dihydrolipoic acid
(DHLA), a potent antioxidant (11, 12). LA improves long-term
memory in aged female NMRI mice (13).
L-Carnitine is a betaine required for the transport of long-
chain fatty acids into the mitochondria for
production, and for the removal of excess short- and medium-
chain fatty acids (14, 15). Acetyl-
L-carnitine (ALCAR) is more
widely used than
L-carnitine in animal and clinical studies
because it enters cells and crosses the blood–brain barrier more
efficiently (16). ALCAR improves cognitive function and neu-
ronal bioenergetic mechanisms in rats with both acute and
long-term treatments (17–23).
Several clinical studies report the beneficial effects of ALCAR
or LA:ALCAR administration in a small group of patients with
AD that resulted in improved spatial orientation and short-term
memory (24, 25). LA administration in patients with AD for
approximately 1 year also resulted in mild cognitive improve-
ments and stabilization of global neuropsychological test scores
(26). Thus, as both ALCAR and LA improve mitochondrial
decay, their combination may be complementary in decreasing
oxidative damage to neurons and cognitive dysfunction.
As our understanding of the importance of mitochondrial decay
in aging advances (27–29), the importance of improving mitochon-
drial function by dietary interventions of mitochondrial metabolites
such as ALCAR or LA becomes clearer (30–33). Feeding 0.15–
0.5% ALCAR to old rats elevated the levels of carnitine in plasma
and brain to that of young rats (34) and 0.1–0.2% LA (T.M.H.,
unpublished data) was as effective in improving mitochondrial
function in the liver as the higher doses originally used (30–33). We
have examined the effects of these lower doses of ALCAR, LA, and
their combination on spatial memory by using the Morris water
maze, on temporal memory by using the peak procedure, decay in
mitochondrial structure in the hippocampus, and oxidative damage
to nucleic acids in the hippocampus and cortex.
Materials and Methods
Materials. ALCAR (hydrochloride salt) was a gift of Sigma Tau
(Pomezia, Italy), and LA was a gift of Asta Medica (Frankfurt兾
Main, Germany). All other chemicals were reagent grade or the
highest quality available from Sigma.
Animals and Diet. Fischer 344 male rats were obtained from the
National Institute on Aging. Control animals were fed AIN93M
diet from Dyets (Bethlehem, PA) and MilliQ water (pH 5.2). The
rats in the experimental groups were fed either 0.5% or 0.2%
(wt兾vol) ALCAR in MilliQ water (pH was adjusted to 5.2 with 1
N NaOH), 0.2% or 0.1% (wt兾wt) LA in AIN93M diet, or a
combination of (0.5% ALCAR and 0.2% LA) or (0.2% ALCAR
and 0.1% LA). The food consumption was determined by weighing
the diet and measuring the volume of water weekly; the average
daily consumption was then calculated. The weight gain during the
course of the experiment was also measured. We did not find any
significant differences in diet, water consumption, or weight gain
between the unsupplemented old rats (13.4 ⫾ 0.5 g兾day; 18.6 ⫾ 1.19
ml兾day; body weight from 416.1 ⫾ 14.4 to 409.2 ⫾ 10.1 g mean ⫾
SE) and the old supplemented rats (For example, the ALCAR ⫹
LA group 13.1 ⫾ 0.4 g兾day; 18.4 ⫾ 0.9 ml兾day; body weight from
416.0 ⫾ 19.0 to 414.9 ⫾ 9.4 g; mean ⫾ SE). All animals were
Abbreviations: ALCAR, acetyl-L-carnitine; LA, R-
-lipoic acid; oxo8dG, 8-hydroxy-2⬘-
deoxyguanosine; oxo8G, 8-hydroxyguanosine; AD, Alzheimer’s disease.
To whom reprint requests should be addressed. E-mail: firstname.lastname@example.org.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
February 19, 2002
no. 4 www.pnas.org兾cgi兾doi兾10.1073兾pnas.261709299
acclimatized at the Northwest Animal Facilities on the University
of California at Berkeley campus for at least 2 weeks before
treatment. Rats were housed individually and provided with
ALCAR and兾or LA for 7 weeks. The young and old rats were 4.5
and 24.5 months old at the start of the experiment; they were more
than 7 weeks older at the time of death. Death, by approved
protocol, was with an overdose of ether.
Morris Water Maze Test of Spatial Memory. The Morris water maze
task tests spatial memory by requiring rats to find a submerged
platform in a pool of water using external visual cues (35, 36).
The time required for individual rats to find the platform and the
length of the swim path was measured by using a digital camera
and a computer system to record movement (VideoMex-V,
Columbus Instruments, Columbus, OH). Trials (4 consecutive
days, 4 trials per day) were with the same hidden platform
location, but with varied start locations. On day 5, the platform
was removed from the pool (transfer trial, 60 sec), and the time
spent at the actual site where the platform was located was
examined. On day 6, the time required to reach a visible platform
was measured to determine visual function and motor ability.
Peak Procedure Test of Temporal Memory. The peak procedure is a
modified fixed-interval schedule commonly used to study tem-
poral memory (37). Rats were tested in 18 identical boxes that
contained a light source and a speaker (for delivering light or
noise signals) and a lever that dispenses single food pellets (45
mg) when pressed (mix T101, Bioserv, Frenchtown, NJ). The
food supply of the rats was decreased to 85% of the free-feeding
amount. In this test, the animal is rewarded with one pellet only
if the lever is pressed at 40 sec from the signal. In 20% of the tests,
no food was given, and an empty trial and the signal lasted 195
sec plus a geometrically distributed duration that averaged 50
sec. The results are presented as a sum of the two types of tests.
Peak rate, which is the maximum response rate in a given trial
and reflects the rats’ choices of what responses to make and their
motivation, was measured.
Electron Microscopic Observations. A subset of rats from each
experimental condition was perfused transcardially with 2.5%
glutaraldehyde for 2 h. The brain was removed from the skull and
the hippocampus was postfixed in 0.1 M PBS with 1% osmium
tetroxide. The tissues were block-stained with uranyl acetate and
embedded in Epon. Sections were cut at 0.6–0.9-
m thick from
the block, stained with uranyl acetate and lead citrate, and
examined with a JEOL 100 CX electron microscope.
Immunohistochemical Studies. A subset of rats from each treat-
ment condition was anesthetized with ether and perfused with
4% paraformaldehyde for 1.5–2 h. The brain was removed and
postfixed for preparing paraffin sections. Sections of hippocam-
pus were incubated with monoclonal anti-8-hydroxy-2⬘-
deoxyguanosine兾8-hydroxyguanosine (oxo8dG兾oxo8G; 1:2000;
QED Bioscience, San Diego) and visualized by using standard
immunocytochemical methods. Two independent analyses were
done on each rat. To determine whether DNA or RNA was
oxidatively damaged, sections were pretreated with either 10
l of RNase-free DNase I or 10 mg兾ml of DNase-free
RNase (Roche Molecular Biochemicals) for 3 h prior to incu-
bation with oxo8dG兾oxo8G Ab (38). To quantify the extent of
oxo8G兾oxo8dG immunolabeling, a 525 ⫻ 410
m area of
staining was captured by using a ⫻2.5 photo eyepiece, a Sony
(Tokyo) high-resolution charge-coupled device (CCD) video
camera (XC-77), and the built-in video capture capabilities of a
Macintosh 8100兾 80AV. All sections from a given region were
captured sequentially during one session and were analyzed blind
with respect to treatment condition. Subsequently, public do-
main image analysis software (
IMAGE 1.55, National Institutes of
Health) and gray-scale thresholding were used to separate
positive staining from background and to calculate the percent-
age of area occupied by oxo8G兾oxo8dG immunoreactivity.
Spatial Memory. Rats are proficient swimmers and are motivated
to escape from water. Once animals learn where the hidden
platform is located, they can remember the location and swim
rapidly to it from any starting point. Both time taken to reach the
platform (Fig. 1) and swimming distance traveled (data not
shown) were measured and gave similar results. Fig. 1 A shows
results obtained on day 4. Young rats spent a significantly shorter
time than old rats (P ⬍ 0.001) in finding the hidden platform.
ALCAR or LA seems to shorten the time in old rats, but the
differences were not significant. However, the combination
resulted in significantly shorter times (P ⬍ 0.05) as compared
with old control rats. The tracks of individual rats on successive
trials and days have been shown (34).
A transfer test, in which the platform was removed, was carried
out on day 5. The time spent at the previous platform position
is a measure of search accuracy and spatial memory. Young rats
spent significantly more time at the former platform position
(P ⬍ 0.001) than old rats did. The ALCAR (P ⬍ 0.05) and LA
(P ⬍ 0.05) significantly restored the lost procedural subcompo-
nent of spatial memory and the combination was even more
effective (P ⬍ 0.01; Fig. 1B).
A clearly visible platform was used to measure deficits in vision,
motivation, motor strength, or coordination on day 6 of the training
cycle. The platform protruded 1 cm above the surface of the water.
Young rats required less time to find the visible platform than the
old animals (Fig. 1C). All three supplementation groups showed
improvement, but only the combination treatment group reached
statistical significance (Fig. 1C).
Fig. 1. Morris water maze test in relation to age and treatment. (A) Time on
day 4 taken to ﬁnd the hidden platform. (B) Time spent at the former platform
position in the transfer test. (C) Time to ﬁnd the visible platform. Data are
mean ⫾ SEM of 9 rats in young and old, 5 in LA (0.1%), and 6 in ALCAR (0.2%)
and ALCAR ⫹ LA groups. Higher doses, 0.2% LA and兾or 0.5% ALCAR, showed
similar results (data not shown). Statistical differences were examined with
two-tailed Student t test.
, P ⬍ 0.001 vs. young rats; #, P ⬍ 0.05 and ##, P ⬍
0.01 vs. old rats.
Liu et al. PNAS
February 19, 2002
Temporal Memory. The response rate to a sound (Fig 2A) and to
a light (Fig. 2B) signal is the same, indicating that the rats
responded similarly to both signals. Results from the last 10 days
of testing were used, where responses had reached asymptotic
Peak rate (Fig. 2 C and D) of young animals was significantly
higher than that of all other groups: young compared with old
(P ⫽ 0.001); young compared with old ⫹ ALCAR (P ⫽ 0.004);
young compared with old ⫹ LA (P ⫽ 0.043); and young
compared with old ⫹ ALCAR ⫹ LA (P ⫽ 0.046). Although
ALCAR does not show any significant effect (comparing the
old ⫹ ALCAR group to the old control rats), LA seems to
slightly increase peak rate. The old ⫹ ALCAR ⫹ LA treatment
showed a more significant increase (P ⫽ 0.033) in peak rate in
old animals than treatment with LA alone.
Ultrastructural Observations of Neuronal Mitochondria. Electron
microscope observations of hippocampal neuronal mitochondria
indicate that structural abnormalities develop with age. Com-
pared with young rats, old rats showed some disruption and loss
of cristae in about half of the mitochondria in the dentate gyrus
area, indicating structural decay. Animals treated with 0.5%
ALCAR and兾or 0.2% LA showed less structural disruption and
loss of cristae. In addition, old rats had more lipofuscin in the
cytoplasm of granule cells of the dentate gyrus, and the com-
bined treatment rats also seemed to have less lipofuscin. How-
ever, these results were obtained from one or two animals per
group. Clearly, further quantitative studies with more animals
and more fields are needed to confirm these observations.
Oxidative Damage to Nucleic Acids. Various regions of the brain
were stained with an Ab that recognizes oxidized DNA or RNA
Fig. 2. Peak procedure test related to age and treatment. Response rate functions plotted separately for sound-signaled trials (A) and light-signaled trials (B)
obtained during the last 10 days of the test. (C) Peak rate over the 20 days of peak procedure testing. Each data point averages 2 days of testing. (D) Peak rate
of the last 10 days averaged. Data are mean ⫾ SEM of 6 in young, 7 in old, 4 in LA (0.2%), and 5 in ALCAR (0.5%) and ALCAR ⫹ LA groups. Treatment with lower
doses, 0.1% LA and兾or 0.2% ALCAR, showed similar results (data not shown). Statistical differences were examined with two-tailed Student t test.
, P ⬍ 0.05
, P ⬍ 0.001 vs. young rats; #, P ⬍ 0.05 vs. old unsupplemented rats.
www.pnas.org兾cgi兾doi兾10.1073兾pnas.261709299 Liu et al.
(oxo8dG or oxo8G; ref. 39). Fig. 3A shows representative images
of the CA1 region of the hippocampus. Fig. 3C shows that old
rats without treatment showed significantly higher immunore-
activity than young rats in areas CA1, CA4, cerebral cortex, and
in the white matter. Both ALCAR and LA reduced immunore-
activity, but only LA showed a significant effect in the CA4
region. The combination showed a significant effect on lowering
immunoreactivity in CA1, CA3, CA4, and dentate granule cells
in old rats.
Fig. 3B illustrates that pretreatment of sections, including area
CA1 with RNase but not DNase, virtually eliminated the im-
munoreactivity, indicating that the predominant damage to
neuronal nucleic acids is to RNA (oxo8G). In CA1, RNase
pretreatment reduced the immunoreactivity by 92%, whereas
DNase, rather than reducing, enhanced (168%) the immuno-
reactivity (Fig. 3B).
Old rats have increased mitochondrial dysfunction and oxidative
damage, which is associated with cognitive deficits in both spatial
and temporal memory. Spatial memory relies on intact hip-
pocampal function. Temporal memory may also be associated
Fig. 3. Immunostaining relative to age and treatment for oxidized nucleic acids in neurons. (A) Representative photographs of oxo8G immunoreactivity in area
CA1 of the hippocampus and adjacent white matter, from individual rats selected from young, old, old ⫹ ALCAR (0.5%), old ⫹ LA (0.2%), and old ⫹ ALCAR
(0.5%) ⫹ LA (0.2%) groups. (B) CA1 sections pretreated with either DNase or RNase before incubation with Ab. (C) Extent of immunoreactivity to oxo8G in the
hippocampus [CA1, CA3, CA4, dentate gyrus (DG), cerebral cortex (CX), and white matter (WM) in rat brain]. [Bar ⫽ 50
m.] Values are mean ⫾ SEM of 5 animals
for young and old groups, 3 for old ⫹ ALCAR and old ⫹ LA groups, and 2 for the old ⫹ ALCAR ⫹ LA group. The Mann–Whitney U test was used to compare values.
, P ⬍ 0.005 vs. young rats; #, P ⬍ 0.05 vs. old control rats.
Liu et al. PNAS
February 19, 2002
with the hippocampus, although it may be more closely associ-
ated with the striatum and cerebellum. The dietary administra-
tion of a combination of ALCAR and LA to old rats improves
mitochondrial function in liver (40). The purpose of this study
was to determine whether it also improves cognition.
Spatial memory was assayed in the Morris water maze. The
Morris water maze has been used extensively to measure cog-
nitive deficits in spatial memory in lesion studies (41–47) and in
aging (48–52). Old rats showed decreased spatial memory
compared with young rats; ALCAR and兾or LA restored some of
this function, the combination being more effective than each
compound alone. We also observed significant age effects in the
transfer test, which measures search accuracy and is considered
a procedural (habitual) subcomponent of spatial memory (36).
ALCAR and兾or LA significantly restored performance in this
test, the combination being more effective (P ⬍ 0.01) and not
significantly different from that of young rats (Fig. 1B). Schenk
and Morris (36) have shown that after a retrohippocampal
lesion, the procedural component of spatial memory can be
partially recovered after training. We also observed significant
age effects on the latencies of animals in finding a visible
platform, which is a control procedure used to detect sensory
motor deficits or motivational differences that impair water
maze performance. The dietary interventions have similar ef-
fects on the visible platform test as those observed during the
hidden platform tests (Fig. 1C).
Many physiological changes occur with age and can have
major consequences on cognitive performance (53). We ob-
served age and treatment effects on several noncognitive factors,
such as motivation and locomotor activity, which can potentially
contribute to the cognitive results. The age-associated decline in
the Morris water maze test, therefore, should not be considered
solely a test of cognition, but also as revealing a general decline
in other systems as a result of aging. Old animals are known to
be less sensitive to pain and possibly to temperature, which may
affect their motivation to find the hidden platform. ALCAR and
LA reduce mitochondrial dysfunction in peripheral systems (31,
33, 54, 55), including sensory systems such as hearing (56).
Therefore, improvements shown here in test performances
attributable to ALCAR, LA, or their combination, including the
visible platform test and ambulatory activity (see ref. 40), suggest
that reversing mitochondrial decay might reverse age-associated
declines in nervous, cardiovascular, visual, and auditory systems,
as well as general effects on motivation and physical strength.
Temporal memory, as assayed by the peak procedure, mea-
sures the function of the internal clock, learning processes,
attention, and exploratory behavior. The combination of LA
with ALCAR showed a significant improvement on peak rate
(P ⬍ 0.05). The peak procedure is a time-discrimination proce-
dure, which resembles a discrete-trials fixed-interval schedule
with catch trials; it has been used to study the timing abilities of
animals (37). Several studies have shown that old rats have
deficits in time perception (57–61). One advantage of the peak
procedure is that it allows for comparison of performance by
using different types of signals and sensory modalities. The
similarity of performance with light and sound signals suggests
that the deficits are the results of deficits in cognition as rats of
different ages do not differ in their sensitivity to light and sound
at the two levels of light and sound used in this study. Peak rate
reflects changes in a response learning mechanism. Old rats had
lower peak rates, suggesting that old animals have difficulty
learning the relevant response. The combination of ALCAR and
LA seems to have a complementary effect on improving the peak
Not all of the old rats tested had cognitive deficits; this resulted
in a large SD and the need for larger numbers of rats to achieve
statistical significance. In future experiments it would be useful
to separate cognitively impaired from unimpaired old rats to
show more pronounced effects in old rats that receive treatment
The current study also has tested the hypothesis that cognitive
improvements in response to ALCAR and兾or LA interventions
are linked to reductions in oxidative damage in old brain. To
measure oxidative damage to nucleic acids, we used an Ab that
detects both oxidized DNA and RNA (39). RNase pretreatment
decreased immunoreactivity extensively, whereas DNase had a
smaller effect. This result suggests that the oxidized nucleic acid
in the aged rat brain is predominantly RNA, which is consistent
with studies in human brains with AD (38). It is clear that more
than 90% immunoreactivity is from RNA, suggesting that RNA
oxidation is a significant biomarker of aging in rat brain. The
mechanism of the DNase enhancement of immunoreactivity
remains unclear; the digestion of DNA may have unmasked
binding sites allowing greater access of the mAb to the RNA.
Cytoplasmic punctate staining is consistent with either cellular
Rna or mtDNA兾RNA. RNA being the predominant oxidized
nucleic acid is consistent with the lack of staining of nuclear
DNA. The type of RNA oxidized and its subcellular localization
remain to be determined, particularly with respect to mitochon-
dria, the most likely oxidant target and the one that is improved
by ALCAR and兾or LA. RNA oxidation increased significantly
as a function of age in rats in areas CA1 and CA4 in the
hippocampus, in cortical neurons, and in white matter in the
frontoparietal cortex. Feeding old rats LA significantly reduced
the levels of oxidized RNA in CA4. The combination of ALCAR
and LA was effective in significantly reducing oxidative RNA
damage in neurons in CA1, CA3, CA4, and dentate gyrus of the
hippocampus to levels not significantly different from young
Poorer performance on memory tasks by old rats could
involve, in part, oxidative damage to RNA, with errors in
translation (62) compromising protein synthesis critical for the
formation of new memories (63, 64). Although oxidative damage
to RNA has been shown to be more extensive than damage to
DNA in urine and plasma (39), oxidized RNA has not been a
focus of interest as an oxidative damage marker for brain aging
or cognition, except in some patients with AD sample studies.
Neuronal RNA oxidation is a prominent feature of vulnerable
neurons in AD, Down’s syndrome, and Parkinson’s disease, all
of which are diseases associated with severe cognitive deficits
(38, 65, 66). Neuronal RNA oxidation may thus contribute to
memory decline and serve as a sensitive marker for intervention
studies. However, oxidant-induced enzyme dysfunction is also an
important contributor to neuronal decay and aging (67–69).
The improving effects on performance on memory tasks by
ALCAR and兾or LA on hippocampal mitochondria are sup-
ported by morphological observations. There seems to be a loss
of mitochondrial cristae with age. Evidence that ALCAR re-
versed this loss with a dose-dependent response has been
presented (34). Similar to ALCAR, LA also reduced age-
dependent cristae loss in the dentate granule cells of the
hippocampus. Because ALCAR alone showed a virtually com-
plete reversal of the cristae loss, we cannot say whether the
combination has an improving effect or not, but it produced at
least as large a reduction as the ALCAR or LA alone.
The loss of memory with age seems to be caused in good part
by oxidative mitochondrial decay in neurons. (i) The effective-
ness of the mitochondrial metabolites ALCAR and LA suggests
that mitochondrial decay is involved. (ii) The oxidation of
RNA兾DNA in neurons is likely to be mitochondrial (70). (iii)
Neuronal mitochondria show structural decay with age.
The cognition-improving effect of ALCAR may also be
caused in part by the donation of an acetyl group for the synthesis
of the neurotransmitter acetylcholine through choline acetyl-
transferase and carnitine acetyltransferase (17, 71, 72). Low
acetylcholine levels in certain brain regions are associated with
www.pnas.org兾cgi兾doi兾10.1073兾pnas.261709299 Liu et al.
age-related cognitive dysfunction, including AD (73). Because of
the profound effects of calorie restriction, we have compared
dietary intakes carefully and have found no significant differ-
ences in food and water consumption or in body weight (see
Materials and Methods).
In conclusion, feeding old rats ALCAR and兾or LA improved
performance on memory tasks, reduced brain mitochondrial
structure decay, and reduced oxidative damage in the brain. The
combination of ALCAR and LA showed a greater effect than
ALCAR or LA alone. These results suggest that feeding a
combination of mitochondrial metabolites to old animals may
prevent mitochondrial decay in neurons and restore cognitive
dysfunction. These results also suggest that consumption of high
levels of mitochondrial metabolites may be an efficient inter-
vention in humans for delaying brain aging and age-associated
We are indebted to Seth Roberts for his stimulation and advice in using
the peak procedure; Judith Campisi, John Nides, and Seth Roberts for
critical reading of the manuscript; and the Electron Microscope Lab at
the University of California at Berkeley for the electron microscopic
studies. We thank M. Nistor at the Institute for Brain, Aging, and
Dementia for technical assistance. This work was supported by grants
from the Ellison Foundation, the National Institute on Aging, the
Wheeler Fund of the Dean of Biology, and the National Institute of
Environmental Health Sciences Center Grant ES01896 (to B.N.A.), and
by National Institute of Aging Grant AG12694 (to C.W.C.).
1. Grady, C. L., McIntosh, A. R., Horwitz, B., Maisog, J. M., Ungerleider, L. G.,
Mentis, M. J., Pietrini, P., Schapiro, M. B. & Haxby, J. V. (1995) Science 269,
2. Wallace, J. E., Krauter, E. E. & Campbell, B. A. (1980) J. Gerontol. 35, 355–363.
3. Rahner-Welsch, S., Frolich, L., Stoll, S. & Hoyer, S. (1995) Neurosci. Lett. 194,
4. Head, E., Mehta, R., Hartley, J., Kameka, M., Cummings, B. J., Cotman, C. W.,
Ruehl, W. W. & Milgram, N. W. (1995) Behav. Neurosci. 109, 851–858.
5. Forster, M. J., Dubey, A., Dawson, K. M., Stutts, W. A., Lal, H. & Sohal, R. S.
(1996) Proc. Natl. Acad. Sci. USA 93, 4765–4769.
6. Carney, J. M. & Carney, A. M. (1994) Life Sci. 55, 2097–2103.
7. Carney, J. M., Starke, R. P., Oliver, C. N., Landum, R. W., Cheng, M. S., Wu,
J. F. & Floyd, R. A. (1991) Proc. Natl. Acad. Sci. USA 88, 3633–3636.
8. Dandekar, T. (1997) Redox Rep. 3, 71–73.
9. Shukitt-Hale, B., Smith, D. E., Meydani, M. & Joseph, J. A. (1999) Exp.
Gerontol. 34, 797–808.
10. Guerrero, A. L., Dorado-Martinez, C., Rodriguez, A., Pedroza-Rios, K.,
Borgonio-Perez, G. & Rivas-Arancibia, S. (1999) NeuroReport 10, 1689–1692.
11. Packer, L., Roy, S. & Sen, C. K. (1997) Adv. Pharmacol. 38, 79–101.
12. Packer, L., Tritschler, H. J. & Wessel, K. (1997) Free Radical Biol. Med. 22,
13. Stoll, S., Hartmann, H., Cohen, S. A. & Muller, W. E. (1993) Pharmacol.
Biochem. Behav. 46, 799–805.
14. Bieber, L. L. (1988) Annu. Rev. Biochem. 57, 261–283.
15. Rebouche, C. J. (1992) FASEB J. 6, 3379–3386.
16. Kidd, P. M. (1999) Alt. Med. Rev. 4, 144–161.
17. Ando, S., Tadenuma, T., Tanaka, Y., Fukui, F., Kobayashi, S., Ohashi, Y. &
Kawabata, T. (2001) J. Neurosci. Res. 66, 266–271.
18. Barnes, C. A., Markowska, A. L., Ingram, D. K., Kametani, H., Spangler, E. L.,
Lemken, V. J. & Olton, D. S. (1990) Neurobiol. Aging 11, 499–506.
19. Caprioli, A., Ghirardi, O., Ramacci, M. T. & Angelucci, L. (1990) Prog.
Neuropsychopharmacol. Biol. Psychiatry 14, 359–369.
20. Bertoni-Freddari, C., Fattoretti, P., Casoli, T., Spagna, C. & Casell, U. (1994)
Brain Res. 656, 359–366.
21. Ghirardi, O., Caprioli, A., Milano, S., Giuliani, A., Ramacci, M. T. &
Angelucci, L. (1992) Physiol. Behav. 52, 185–187.
22. Ghirardi, O., Giuliani, A., Caprioli, A., Ramacci, M. T. & Angelucci, L. (1992)
J. Neurosci. Res. 31, 375–379.
23. Taglialatela, G., Caprioli, A., Giuliani, A. & Ghirardi, O. (1996) Exp. Gerontol.
24. Bonavita, E. (1986) Int. J. Clin. Pharmacol. Ther. Toxicol. 24, 511–516.
25. Rai, G., Wright, G., Scott, L., Beston, B., Rest, J. & Exton-Smith, A. N. (1990)
Curr. Med. Res. Opin. 11, 638–647.
26. Hager, K., Marahrens, A., Kenklies, M., Riederer, P. & Munch, G. (2001) Arch.
Gerontol. Geriatr. 32, 275–282.
27. Ames, B. N., Shigenaga, M. K. & Hagen, T. M. (1993) Proc. Natl. Acad. Sci.
USA 90, 7915–7922.
28. Harman, D. (1999) J. Anti-Aging Med. 2, 15–36.
29. Shigenaga, M. K., Hagen, T. M. & Ames, B. N. (1994) Proc. Natl. Acad. Sci.
USA 91, 10771–10778.
30. Hagen, T. M., Wehr, C. M. & Ames, B. N. (1998) Ann. N.Y. Acad. Sci. 854,
31. Hagen, T. M., Ingersoll, R. T., Wehr, C. M., Lykkesfeldt, J., Vinarsky, V.,
Bartholomew, J. C., Song, M. H. & Ames, B. N. (1998) Proc. Natl. Acad. Sci.
USA 95, 9562–9566.
32. Hagen, T. M., Ingersoll, R. T., Lykkesfeldt, J., Liu, J., Wehr, C. M., Vinarsky,
V., Bartholomew, J. C. & Ames, A. B. (1999) FASEB J. 13, 411–418.
33. Hagen, T. M., Vinarsky, V., Wehr, C. M. & Ames, B. N. (2000) Antioxid. Redox
Signal 2, 473–483.
34. Liu, J., Atamna, H., Kuratsyne, H. & Ames, B. N. (2002) Ann. N.Y. Acad. Sci.,
959, in press.
35. Morris, R. (1984) J. Neurosci. Methods 11, 47–60.
36. Schenk, F. & Morris, R. G. (1985) Exp. Brain Res. 58, 11–28.
37. Roberts, S. (1981) J. Exp. Psychol. Anim. Behav. Processes 7, 242–268.
38. Nunomura, A., Perry, G., Pappolla, M. A., Wade, R., Hirai, K., Chiba, S. &
Smith, M. A. (1999) J. Neurosci. 19, 1959–1964.
39. Park, E. M., Shigenaga, M. K., Degan, P., Korn, T. S., Kitzler, J. W., Wehr,
C. M., Kolachana, P. & Ames, B. N. (1992) Proc. Natl. Acad. Sci. USA 89,
40. Hagen, T. M., Liu, J., Lykkesfeldt, J., Wehr, C. M., Ingersoll, R. T., Vinarsky, V.,
Bartholomew, J. C. & Ames, B. N. (2002) Proc. Natl. Acad. Sci. USA 99,
41. Brandeis, R., Brandys, Y. & Yehuda, S. (1989) Int. J. Neurosci. 48, 29– 69.
42. Dalm, S., Grootendorst, J., de Kloet, E. R. & Oitzl, M. S. (2000) Behav. Res.
Methods Instrum. Comput. 32, 134 –139.
43. Hatfield, T. & McGaugh, J. L. (1999) Neurobiol. Learn. Mem. 71, 232–239.
44. Lai, H., Carino, M. A. & Ushijima, I. (1998) Bioelectromagnetics 19, 117–122.
45. Moghaddam, M. & Bures, J. (1997) Neurobiol. Learn. Mem. 68, 239–251.
46. McAlonan, G. M., Dawson, G. R., Wilkinson, L. O., Robbins, T. W. & Everitt,
B. J. (1995) Eur. J. Neurosci. 7, 1034–1049.
47. Kant, G. J., Yen, M. H., D’Angelo, P. C., Brown, A. J. & Eggleston, T. (1988)
Pharmacol. Biochem. Behav. 31, 487– 491.
48. Joseph, J. A., Shukitt-Hale, B., Denisova, N. A., Bielinski, D., Martin, A.,
McEwen, J. J. & Bickford, P. C. (1999) J. Neurosci. 19, 8114–8121.
49. Shukitt-Hale, B., Mouzakis, G. & Joseph, J. A. (1998) Exp. Gerontol. 33, 615–624.
50. Socci, D. J., Crandall, B. M. & Arendash, G. W. (1995) Brain Res. 693, 88–94.
51. Sack, C. A., Socci, D. J., Crandall, B. M. & Arendash, G. W. (1996) Neurosci.
Lett. 205, 181–184.
52. Issa, A. M., Rowe, W., Gauthier, S. & Meaney, M. J. (1990) J. Neurosci. 10,
53. Andrews, J. S. (1996) Brain Res. Cogn. Brain Res. 3, 251–267.
54. Arivazhagan, P., Pramanathan, K. & Panneerselvam, C. (2001) J. Nutr.
Biochem. 12, 2–6.
55. Paradies, G., Ruggiero, F. M., Petrosillo, G., Gadaleta, M. N. & Quagliariello,
E. (1995) Mech. Aging Dev. 84, 103–112.
56. Seidman, M. D., Khan, M. J., Bai, U., Shirwany, N. & Quirk, W. S. (2000) Am. J.
Otol. 21, 161–167.
57. Campbell, B. A. & Haroutunian, V. (1981) J. Gerontol. 36, 338–341.
58. Soffie, M. & Lejeune, H. (1991) Neurobiol. Aging 12, 25–30.
59. Meck, W. H., Church, R. M. & Wenk, G. L. (1986) Eur. J. Pharmacol. 130,
60. Lejeune, H., Maquet, P., Bonnet, M., Casini, L., Ferrara, A., Macar, F.,
Pouthas, V., Timsit-Berthier, M. & Vidal, F. (1997) Neurosci. Lett. 235, 21–24.
61. Lejeune, H., Ferrara, A., Simons, F. & Wearden, J. H. (1997) J. Exp. Psychol.
Anim. Behav. Processes 23, 211–231.
62. Rhee, Y., Valentine, M. R. & Termini, J. (1995) Nucleic Acids Res. 23,
63. Meiri, N. & Rosenblum, K. (1998) Brain Res. 789, 48–55.
64. Wells, D. G. & Fallon, J. R. (2000) Cell. Mol. Life. Sci. 57, 1335–1339.
65. Nunomura, A., Perry, G., Hirai, K., Aliev, G., Takeda, A., Chiba, S. & Smith,
M. A. (1999) Ann. N.Y. Acad. Sci. 893, 362–364.
66. Nunomura, A., Perry, G., Zhang, J., Montine, T. J., Takeda, A., Chiba, S. and
Smith, M. A. (1999) J. Anti-Aging Med. 2, 227–230.
67. Fucci, L., Oliver, C. N., Coon, M. J. & Stadtman, E. R. (1983) Proc. Natl. Acad.
Sci. USA 80, 1521–1525.
68. Stadtman, E. R. & Levine, R. L. (2000) Ann. N.Y. Acad. Sci. 899, 191–208.
69. Liu, J., Killilea, D. W. & Ames, B. N. (2002) Proc. Natl. Acad. Sci. USA 99,
70. Hirai, K., Aliev, G., Nunomura, A., Fujioka, H., Russell, R. L., Atwood, C. S.,
Johnson, A. B., Kress, Y., Vinters, H. V., Tabaton, M., et al. (2001) J. Neurosci.
71. Goodman, D. R. & Harbison, R. D. (1981) Biochem. Pharmacol. 30, 1521–1528.
72. White, H. L. & Scates, P. W. (1990) Neurochem. Res. 15, 597–601.
73. Coyle, J. T., Price, D. L. & DeLong, M. R. (1983) Science 219, 1184–1190.
Liu et al. PNAS
February 19, 2002