Neurobiologyof Aging, Vol. 13, pp. 9-23. e PergamonPress plc, 1991. Printed in the U.S.A. 0197-4580/92 $5.00 ÷ .

00

Progressive Decline in Spatial Learning

and Integrity of Forebrain Cholinergic
Neurons in Rats During Aging

W A L T E R F I S C H E R , 1 KAREN S. C H E N , * F R E D H. G A G E * A N D A N D E R S B J I 3 R K L U N D

Department o f Medical Cell Research, Universi O, of Lund, Sweden

and *Department o f Neurosciences, UCSD, La Jolla, CA

R e c e i v e d 3 D e c e m b e r 1990; A c c e p t e d 19 July 1991

FISCHER, W., K. S. CHEN, F. H. GAGE AND A. BJORKLUND. Progressive decline in spatial learning and integriO' of
forebrain cholinergic neurons in rats during aging. NEUROBIOL AGING 13(1) 9-23, 1992.--Rats distributed over five different
age groups, 3, 12, 18, 24 and 30 months of age, were screened for their spatial learning and memory ability in the Morris water
maze, and the degree of place navigational impairments was correlated with morphological changes in the four major forebrain
cholinergic cell groups (medial septum, MS; vertical limb of the diagonal band of Broca, VDB; nucleus basalis magnocellularis,
NBM; and striatum) using choline acetyltransferase (CHAT) and nerve growth factor receptor (NGFr) histochemistry. Impaired
place navigation developed progressively with age, such that 8% of the 12-month-old rats, 45% of the 18-month-old, 53% of the
24-month-old, and over 90% of the 30-month-old rats were behaviorally impaired. Significant reductions in the number of CHAT/
NGFr-positive cell bodies, amounting to between 19 and 45%, were observed in all four cell groups, and the remaining cells were
reduced in size (6-24% reduction in cross-sectional area in the oldest age groups). Although the morphological changes were less
severe and tended to develop later than the behavioral impairments, there was overall a significant correlation between water maze
performance and ChAT/NGFr-positive cell counts, and to a lesser degree also cell size in all four cell groups. These changes were
also highly correlated with age. The highest correlations were seen in MS, VDB and NBM, which are known to play a role in
spatial memory performance in young rats. The results indicate that degenerative and/or atrophic changes in the forebrain cholin-
ergic system and decline in spatial learning ability are parallel processes during aging. Although the magnitude of the morphologi-
cal changes does not appear to be substantial enough, by itself, to explain the severe spatial learning impairments that develop in
the oldest animals, the present data are consistent with the view that impaired function in the forebrain cholinergic system can
contribute to age-dependent cognitive decline in rodents.

Aging Spatial memory Forebrain cholinergic neurons Morris water maze

THE cholinergic neurons of the basal forebrain nuclei show on the morphological changes in the basal forebrain cholinergic
atrophic or degenerative changes in aged rodents (2, 3, 7, 14- system have so far been performed in rodents.
16, 18, 25, 26, 28). In addition, an extensive literature has In a previous study (15), we observed a correlation between
demonstrated age-dependent decline in various parameters of impaired place navigation in the Morris water maze task, on one
learning and memory ability [for review see (30)], and there are hand, and degenerative changes in forebrain cholinergic neurons,
several studies reporting concomitant changes in neurochemical on the other, among individual rats of the same age (i.e., 24-
indices of forebrain cholinergic neuron function and in parame- month-old female Sprague-Dawley rats). Similar results have
ters of impaired learning and memory in aged rats (7, 22, 23, been obtained by Koh et al. (33) in 32-35-month-old male Long
29, 32, 33, 37, 50, 51). As concluded by Decker (11) in a Evans rats and by Gallagher et al. (23) in a biochemical study
comprehensive review, current data are indeed consistent with in 28-29-month-old male Long Evans rats. These observations
the view that the functioning of the forebrain cholinergic system raise the question whether atrophic changes, observed morpho-
is impaired not only in Alzheimer's disease but also to a lesser metrically in the various cholinergic forebrain nuclei, develop in
extent during normal aging. However, the extent to which changes parallel with the appearance of spatial learning impairments dur-
in the cholinergic forebrain system are important for the devel- ing aging. The present study was designed to address this ques-
opment of behavioral or cognitive impairments during normal tion, using Sprague-Dawley rats of a range of different ages (3,
aging remains unclear. Coleman et al. (10) have recently em- 12, 18, 24 and 30 months of age). The rats were screened for
phasized the importance of using multiple time points in order their place navigational ability in the Morris water maze, and
to characterize age-related changes. No such longitudinal studies morphological changes in the four principal forebrain cholinergic

1Requests for reprints should be addressed to Dr. Walter Fischer, Department of Medical Cell Research, University of Lund, Biskopsgatan 5, 223
62 Lund, Sweden.
ll) F I S C H E R , CHt~N. GA(iii: A N i ) B,I()RK~.i'Nir

TABLE I
SWIM SPEED (DM/sl IN THE MORRIS WATER MAZE FOR EACH DA'~ IN THE SCREENING TESI

Age Day t Day 2 Day 3 I)ay 4 Da\ 5

12 Months 2.9 ± 0.2 3.(! *: 0.2 2.q z_ 0.1 2.9 ± 0.2 32 ::: 0.2 !!
3 Months 2.3 ± 0.3 2.6 :~ 0.2 2.6 ± 0.1 2.7 ± 0.1 3,,4 ~ t).2 ~
18 Months 2.6 ± 0.2 3.0 ± 0.2 2.9 +_ 0.2 2.9 ± 0.3 2.~ = (}3 ~5
3 Months 2.4 + 0.2 3.2 :~ t).2 2.6 + 0.3 2.7 ± 0.2 2¢, :_- 0.3
24 Months 2.6 -+ 0.2 2.6 ~: 0.1 3.(i+_ 0.2 2.7 _+. 0.1 2t~ :_ 0.1 ~;;
3 Months 3.0 = 0.4 2.6 0. I 2.8 + 0.1 2.8 -+ 0.1 27 :: (),1 {?
30 Months 2.2 ~ 0.1 2.5 :: O.l 24 ± 0.1 2.5 +_ 0. l 3.(~ :z 1).3 ~×
3 Months 2.4 -+ 0.1 2.3 ~ 0.2 2.2 +_ 0.2 2.4 ± 0.1 3.2 ±: 0.3 ¢,

Each batch of aged rats tested in the water maze were swum together with 3-month-old animals, which served as
a reference group in order to determine whether the aged animals were "nonimpaired," or "impaired" in the task.
No significant differences in swim speed could be detected between the aged and the 3-month-old control groups
using either 2-way or one-way ANOVA tests.

n u c l e i - - m e d i a l s e p t u m (MS), the vertical limb o f the diagonal ers at 10-12 m o n t h s o f age and were h o u s e d for an additional 6
band o f B r o c a ( V D B ) , n u c l e u s basalis m a g n o c e l l u l a r i s ( N B M ) , ( n = 5 5 ) , 12 ( n = 6 3 ) or 18 m o n t h s ( n = 3 8 ) before the b e g i n n i n g
and s t r i a t u m - - w e r e s u b s e q u e n t l y analyzed m o r p h o m e t r i c a l l y , us- o f the experiment. Y o u n g female control rats were b o u g h t at 2
ing choline acetyltransferase (CHAT) or nerve g r o w t h factor re- m o n t h s o f age and h o u s e d for another 1 m o n t h (n = 34) or 10
ceptor (NGRr) h i s t o c h e m i s t r y . m o n t h s ( n = 12) before the b e g i n n i n g o f the behavioral testing.
The aged rats were h o u s e d 5 a n i m a l s per cage and the y o u n g
were h o u s e d 8 per cage in a 12:12-hour light-dark cycle. B e h a v -
Animals ioral testing w a s p e r f o r m e d during the daylight period. T h e vari-
ous s u b g r o u p s o f a n i m a l s were processed for i m m u n o h i s t o c h e m -
A total o f 202 female S p r a g u e - D a w l e y rats ( A L A B , Stock- istry 4 w e e k s after the b e g i n n i n g o f the behavioral testing. T h u s
h o l m ) divided over 5 age groups were used: 3, 12, 16-18, 2 2 - 2 4 the m o r p h o m e t r i c analysis was p e r f o r m e d on a n i m a l s 4 m o n t h s
and 2 8 - 3 0 m o n t h s o f age. Rats were obtained as retired breed- o f age ( " 3 - m o n t h g r o u p " ; n = 16), 13 m o n t h s ( " 12-month g r o u p " :

50
Time

Tto find platform (see)

tO0
Distance to platform

.
(din)

•
O
[]
A
3 mo
!2 mo
18 n l o
24 mo
(n=:~.l)
tn=12)
{i1:=55)
{n=63)
40 80

30 60

20 40

i0 30

o o
day 1 2 3 4 5 day 1 2 3 4 5

FIG. 1. Performance in the Morris water maze task as measured by the time it took the rats to find the hidden platform
(escape latency; left panel) and distance swum before finding the platform (fight panel). Values give means--_ SEM. As
measured by the reduction in escape latency and distance, all groups showed significant learning on the task (p<0.001,
2-way ANOVA). On test day 1 none of the animal groups significantly differed from each other. On the last three days in
the test week the 24-month and the 30-month groups significantly differed in performance from the 3-month group
(*p<0.001; one-way ANOVA, post hoc Fisher PLSD).
SPATIAL MEMORY AND CHOLINERGIC NEURONS IN AGING 11

Time to find platform n = 6), 17-19 months ( " 18-month group"; n = 13), 23-25 months
60 ("24-month group"; n = 14) and 29-31 months of age ("30-
55
: o month group"; n = 4).
50
In the 18-month group, about 17% of the animals had died
z before the beginning of the water maze tests, in the 24-month
45
group about 30% and in the 30-month group about 50% had died
v eoe
40 prior to testing. Some of the animals in the 18- and 30-month
--~---
35 groups were taken from cannulation experiments where the ani-
30
| i mals, after the end of the water maze test, had received vehicle
25 ~ .~. infusion in the lateral ventricle on the fight side (16). The 24-
c~
month group was used as a control group in a parallel experi-
20
__IL__ ment (18). A separate group of 3-month-old young controls
15 . . . . . . . o • ( n = 7 ) was processed together with the 24-month-old rats. For
10 eooo the computerized image analysis the morphometric results ob-
5 . . . . coo• tained from the left hemisphere are presented, except for the 24-
88 e month-old animals where the results were obtained from the
0
A 3 mo 12 mo 18 mo 24 mo 30 mo right hemisphere.

Behavioral Testing
Platform crossings
12"

11 m •
The Morris water maze task (43) was used to evaluate the
10 o• • o•
spatial learning and memory performance of the animals. Our
- -oo-- •co olo
water maze consists of a circular tank, 120 cm in diameter and
9
45 cm in height, filled with water to a depth of 30 cm. Dried
8 - --z z- - - o ~ - - --••ooze- •

milk powder was added to make the water opaque. An escape

7 zzzz= oe •go•o• go•
platform was placed 1 cm beneath the water surface and kept in
6 a constant position in the pool during all trials. The room in
_ _ o ~ _ oo eeoeo eoseo
5 which the pool was situated was rich in extra-maze cues (e.g.,
4 • - - o - - - ~ - ~ oo pictures, window, plants, experimenter's desk). Four equally
3 spaced points around the edge of the pool were used as start
2 oo ooO _ .lqlql _ - ..... positions. The screening of the rats in the pretest was performed
1
in 2 blocks of 4 trials on each day for 5 consecutive days. After
0
the last swim trial on the fifth day the platform was removed,
and the swim path as well as the distance swum in each of the 4
B 3 m• 12 m • 18 m • 24 m • 30 m • quadrants were recorded over 60 s ("spatial probe trial"). The
number of times the rats swam through the area where the plat-
form had been situated were measured as "platform crossings"
as well as the distance swum in the platform quadrant (the third
% in platform quadrant
quadrant clockwise) as percent of the distance swum in all 4
75
quadrants (%Q3).
! If the latency of an aged animal to find the hidden platform
--i--. ~ (escape latency) during the 2 last days of testing exceeded the
a. -&-- -~-- ,~ " value of 2 standard deviations from the mean performance of
50 'w I *II" .o the 3-month-old group, the rat was denoted "cognitively im-
paired" in the morphometric analysis (21). Animals that were
found to perform within 1 standard deviation of the mean per-
oo¢0 aBa
• I __.'_.__ .8• •I. formance of the young rats were denoted " n o n i m p a i r e d . " The
|$ ooo
eeoc various batches of 3-month-old animals that were swum in par-
oooe ooe - -
25 allel with the 12-, 18-, 24- and 30-month groups (see Table 1)
• m
•
g
__~__
did not differ in performance between each other and these
3-month-old animals were put into one large group comprising
oe
34 animals altogether.

C 3 m• 12 m• 18 m• 24 m• 30 m• Immunohistochemisto'
The rats were killed for histochemistry by transcardial perfu-
FIG. 2. Scatter-plots of the performance of the different age groups on sion with 4% phosphate-buffered paraformaldehyde under chlo-
the escape latency parameter (mean of day 4 + 5 ; panel A), and of the ral hydrate anesthesia. The brains were postfixed in the same
platform crossing parameter (panel B) and percent of the total distance solution for about 20 h and then transferred to a 20% sucrose
swum in the platform quadrant (panel C) in the final spatial probe trial
solution (in PBS) for about 12 hours. They were then sectioned
when the platform had been removed from the tank. The horizontal lines
indicate the mean value of the entire group, the broken lines indicate 1 coronally in a freezing microtome at 40 ~ m throughout the fore-
standard deviation of the mean performance and the dotted line in the brain. Every fifth section from the brains of the 30-month group
3-month-old group in panel A denotes 2 standard deviations of the mean (including the 3-month rats that were tested in parallel), and ev-
performance. Open circles denote those animals that are represented in ery fourth section from the 3-, 12- and 18-month groups was
Fig. 3. used for immunostaining with a double-labelling protocol for the
t: FISCHER, (:HEN. GAG[- ~ N I ) BJORKi ,{ 7,~!

3 MO Esc. lat. 7.5

Cross. 9
% Q3 53

f - \
[
12 MO Esc. lat. 6
Cross. 8
% Q3 54

18 M O Esc. lat. 7 Esc. lat. 23

Cross. 9 'I
Cross. 2
% Q3 54 % Q3 18

24 MO Esc. lat.
Cross.
% Q3
8
7
58
/i
I

~ x\

, ESC. lat.
Cross.
% Q3
53
0
21

30 M O / Esc. lat. 8.5 " ~ - ~ \ Esc. lat. 57

ir Cross. 4 / ~ , ~ , i Cross. 1
%Q3 50 '~/, ~ . ,.,~ %Q3 21

NON-IMPAIRED IMPAIRED

FIG. 3. Swim paths of animals from the different age groups plotted in the final spatial probe trial
(when the platform had been removed) during 60 s swim time. The nonimpaired animals (left panel)
show a focused search over the former platform site (4-9 crossings; 50-58% in Q3), whereas the
animals that show impairment in spatial memory performance (right panel) searched randomly over
the pool and only by chance swum over the former platform site (0-2 crossings and 18-21% in Q3).
The thicker line indicates the initial swim path of the animal after having been put into the pool
starting at "3 o'clock" and the asterisks indicate the end of the swim path after having swum for 60
s. The escape latency figures (in seconds) are the means of the escape latencies during the 4th and
5th day of testing.

NGF-receptor (NGFr) and choline acetyltransferase (CHAT) [see 0.025% diaminobenzidine, 0.01% H202, 0.04% nickel chloride
(6)]. The tissue sections from the 24-month group (and the par- in Tris-buffered saline (pH 7.45). The nickel salt intensification
allel tested 3-month-old controls) were stained for the NGFr resulted in a black reaction product.
alone. To label for the NGFr the tissue sections were incubated To double label for ChAT on the same sections, residual
with a monoclonal antibody (192-IgG) specific for NGFr [(52), peroxidase enzyme activity had to be blocked by incubation of
courtesy of E. Johnson] with a mixture of 0.25% Triton X-100 the sections in 0.6% n202 solution for 45 rain. In addition, free
and 1% serum in 0.1 M Tris-buffered saline. They were incu- biotinylated binding sites were blocked by the sequential appli-
bated with the primary antibody for 16 h at room temperature. cation of excess avidin and biotin (Vector). Following these two
The sections were then incubated with a rat adsorbed horse anti- blocking procedures, the tissue sections were incubated for 18 h
mouse biotinylated IgG (Vector) for 3 h. Subsequently, the sec- at room temperature with the polyclonal ChAT antibody raised
tions were incubated in complexes of avidin and biotinylated in rabbit (courtesy of L. Hersh). Sections were then incubated
peroxidase (Vector), and finally the sections were reacted in for 4 - 6 h in affinity-purified biotinylated donkey anti-rabbit IgG
S P A T I A L M E M O R Y A N D C H O L I N E R G I C N E U R O N S IN A G I N G 13

TABLE 2
NUMBERS (CORRECTED COUNTS) AND SIZES OF DOUBLE-LABELLED (DL; NGFr + CHAT)
AND SINGLE-LABELLED(SL; CHAT) NEURONS OF THE FOREBRAIN

Cell Number Cell Size (p.m2)

MS VDB Striatum NBM NBM MS VDB Striatum NBM NBM Escape
Age /DLI (DL) (SL) (DL) (SL) (DL) (DL) (SL) (DL) (SL) n Latency Crossings

3 Months* 1224±73 2030_+100 2256_+120 1844-+61 348-+55 196-+6 210_+7 226_+7 279_+8 202-+10 n=16 8_+1 8±1
12Months* 1283-+118 2226-+149 2167-+108 1652_+76 282+61 173-+9 190-+8 198-+10 262_+8 208±7 n=6 8±2 6_+1
18 Months* 1142-+ 120 2147+ 187 1893±250 1888_
+100 283_+49 191±12 195-+5 240_+70 285+7 180_+10 n=6 7_+1 7 +1
-
18Monthst 1129+130 2097±220 1566_+58§ 1676_+134 293-+57 192-+4 193+9 237_+20 291_+11 185-+11 n=7 16-+2.5 5+1
24 Months* 1162+74 1708+68 162-+ 3~ 170+-4 n=7 7.5_+0.5 7-+0.5
24 Months+ 988± 135 1415_+73~- 159-+ 3~: 165 + 10 n=7 48_+2.5 0.7±0.3
30 Months+ 892+97§ 1392_+138§ 967-+84§ 1333-+82§ 91 -+29§ 168-+9§ 189 + 10§ 189+-6§ 243_+12§ 154+6§ n=4 30+4 0.7-+0.5

*Nonimpaired in the Morris water maze spatial learning task

tlmpaired in the Morris water maze spatial learning task
+Significant difference (p<0.05, one-way ANOVA, post hoc Fisher PLSD) compared to the parallely processed young group.
§Significant difference (0<0.05, two-tailed Student's t-test) compared to the young (n = 16) control group.
The numbers represent cell numbers and sizes on the left side in all groups except for the 24-month group, where the values represent measures from the right side. In
the 24-month group only neurons in the MS and VDB were included in the measurement. In this group the cells in the MS and VDB were stained for the NGFr alone. The
statistics for these aged animals were performed with a separate group of young (n = 7) animals (1240-+27 counts of NGFr-stained cells in the MS and 1969-+ 118 in the
VDB and 172 ± 3 ~m2 in cell size of neurons in the MS and 173 _+2 p.m2 in the VDB). All other comparisonswere performed with the 3-month group (n = 16) included.

(preadsorbed against rat, h u m a n , and m o u s e immunoglobulins; involved in the estimation of absolute cell n u m b e r s in microtome
A m e r s h a m ) . Next, the sections were incubated for 2 h in com- sections [where shrinkage of the tissue occurs during processing;
plexes of avidin and biotinylated peroxidase (Vector) and reacted see (28)], we present our data both in terms of corrected and
for 25 min in a solution consisting of 89.5 ml of PBS (pH 7.4), uncorrected counts using the Abercrombie (1) formula. In the
10 txl of 1% a m m o n i u m carbonate, and 500 txl of 0.1% one- correction procedure the setting of the microtome (40 Ixm) was
naphthol (dissolved in ethanol). After a brief w a s h , the sections used as section thickness. The thickness of the mounted sections
were incubated for 10 min in a 0.1% buffered pyronin B solu- after processing was 11.8---0.7 txm in the 3-month-old animals,
tion to give a violet reaction product. 1 1 . 6 ± 0 . 9 ixm in the 12-month group, 12.6_+0.9 txm in the 18-
month group, 1 2 . 9 ± 0 . 8 in the 24-month group and 10.5_+0.8
Image Analysis txm in the 30-month group.
Computerized image analysis (IBAS II instrument, Zeiss-
M o r p h o m e t r y was performed on animals that were either Kontron) was used for the morphometric analysis of size and
nonimpaired or impaired in the water maze task selected on the n u m b e r of double-labelled ( N G F r + ChAT-positive) and single-
basis of the escape latency parameter given above, thereby ex- labelled (CHAT or NGFr-positive) neurons. No single-labelled
cluding animals that performed between 1 and 2 SD of the per- NGFr-positive cells could be detected in the areas measured.
formance of the group of 3-month-old animals that were tested Medial septum (MS) and the vertical limb of the diagonal band
in parallel with each age group. Due to the uncertainties that are of Broca (VDB) were measured from the level of the genu of

TABLE 3
REGRESSION COEFFICIENTS(R) AND STATISTICAL SIGNIFICANCESIN THE LINEAR REGRESSION ANALYSIS
PERFORMED BETWEEN BEHAVIORALAND MORPHOMETRIC MEASURES AND AGE

Escape Latency (s) Platform Crossings Age

Area vs. Cell Counts vs. Cell Size vs. Cell Counts vs. Cell Size vs. Cell Counts vs. Cell Size

MS r= -.61, p<0.0001 r= -.33, p<0.01 r= .43, p<0.001 r = .31, p<0.02 r= - .51, p<0.0001 r= -.41, p<0.001
VDB r= - .58, p<0.0001 r= - .30, p<0.05 r= .47, p<0.0005 r = .32, p<0.02 r= - .41, p<0.001 r= - .47, p<0.0001
Striatum r= - .38, p<0.02 r= -.34, p<0.05 r= .21, n.s. r=.26, n.s. r= -.38, p<0.01 r= -.21, n.s.
NBM r= -.61, p<0.0001 r= - .26, n.s. r= .40, p<0.02 r = .38, p<0.02 r= - .63, p<0.0001 r= - .53, p<0.0005
NBM (SL) r= -.58, p<0.0001 r= -.24, n.s. r= .37, p<0.05 r=.12, n.s. r= -.40, p<0.01 r= -.61, p<0.0001

The analysis was performed on the 53 rats from the 3-, 12-, 18-, 24- or 30-month groups included in Table 2. The areas included were the medial
septum (MS), the vertical limb of the diagonal band of Broca (VDB), the striatum and the nucleus basalis magnocellularis (NBM). In all groups
except the 24-month group, the double-labelling procedure for ChAT and NGFr was used. In the 24-month-old animals, only the MS and the VDB
were included in the morphometric analysis. Single-labelled (CHAT) neurons were found in the striatum and in the NBM (SL).
t-t FISCHER, ('HEN, (;AG[:, ~ N |) B.!ORK I ,~ N I "

FIG. 4. A section through the midportion of the nucleus basalis magnocellularis in m~ aged (30-month-oldt rat. 'fhe technique to~ doui-Jic-hd,~i
immunohistochemistry gives the opportunity to detect two different antigens on the same tissue section using tx~o different chromogens: The cc[is it:
pink/violet color (chromogen is Pyronin B) are labelled for choline acetyltransferase (CHAT). The other cell population wa~, al~(~ labelk-d tc~r lh~,
marker, but, in addition, for the nerve growth factor receptor (NGFr: chromogen is diaminobenzidine); this combined labelling f,~)cedure pr~duc'ed a
brownish color within the cells.

the corpus callosum, rostrally, to the level of the crossing of the ble-labelled neurons ( N G F r + CHAT) had a brownish appearance
anterior commissure, caudally. The boundary between MS and and the single-labelled neurons (CHAT) appeared violet, these
VDB was defined as the anterior commissure, and the lateral neuronal populations could be easily distinguished by the ob-
border of the VDB was set at the medial border of the olfactory server and then analyzed in the image analysis program
tubercle. The area included in the striatum covered the entire All values given represent cell numbers on the left side li.e.,
cross-sectional area of the head of the caudate-putamen down to the side opposite to the vehicle infusion in those cases where
the level of the anterior commissure, starting rostrally at the the rats had been part of an infusion experin3entL The number
level of the genu of the corpus callosum and ending caudally at of sections measured in the 3-, 12-, 18- and 24-month group
the level of the crossing of the anterior commissure. Nucleus was about 8 for MS, VDB and striatum, and about 12 ff~r NBM
basalis magnocellularis (NBM) was defined as the immunola- (every fourth section stained); in the 30-month group, the num-
belled neurons located in the globus pallidus and the adjoining ber was about 6 for the MS, VDB and the striatum, and about
part of the internal capsule. The NBM neurons were measured 10 for the NBM (every fifth section stained). The mean number
from the level of the caudal septum, rostrally to the level of the of cells measured on each side of each brain was about 230(1 in
tail of the caudate-putamen. Cells were defined (under 10 x ob- the 3- and 12-month-old animals, about 2300 in the nonimpaired
jective magnification) as immunolabelled cell bodies with a di- 18-month-old animals, about 2000 in the cognitively impaired
ameter of > 8 p~m along the minor axis and a cross-sectional area 18-month-old animals, and about 1600 in the 30-month-old ani-
:>70 p.m 2. These criteria were used to avoid excluding any pop- mals, which were all impaired. The number of single-labelled
ulation of small atrophied immunolabelled neurons. Since dou- neurons (NGFr) measured in the MS and VDB in the 24-month
SPATIAL MEMORY AND CHOLINERGIC NEURONS IN AGING 15

i
t '~,+
i

*L ,\

,,e , 3
g "
l

I
j -
!
"i ~' " P ~
• . A
k \

• 0
'2-
.~N "i /
J
I
It t

.,a2 LL
. b ~" •6 " I~

FIG. 5. Photomicrograph of double-labelled neurons (ChAT+NGFr) in the medial septum (MS) in a 3-month-old (a) and a 30-
month-old animal (b). Note the apparent loss and atrophy of immunolabelled neurons in the aged animal (panel b). The scale bar
(in panel a) is 100 Ixm.

group was about 1100 in the 3-month-old animals (tested in par- 30-month-old animals showed significantly less learning on these
allel with the aged rats), about 1000 in the nonimpaired 24- measures compared to the 3-, 12- and 18-month-old animals in
month-old animals and 800 in the impaired aged animals. this task. Thus, on the last three days of testing (day 3 to 5),
the 24-month-old animals had learnt significantly less compared
Statistical Evaluation to the 3-month-old animals (p<0.001, one-way ANOVA, post
hoc Fisher PLSD, Fig. 1) as had the 30-month-old animals
Data given represent means-+ SEM. Two-way ANOVA with which performed significantly worse compared to the 3-month-
post hoc Student's related t-test or one-way ANOVA with post old groups on days 2-5 (p<0.001, Fig. 1). The 30-month-old
hoc Fisher PLSD was used for group analysis and statistical rats tended to perform worse than the other rats on the escape
comparison of the results in the Morris water maze test. In the latency measure (Fig. I A) also on the first day. This was, how-
evaluation of the results of the morphometric analysis one-way ever, due to the fact that the 30-month-old rats swam slower on
ANOVA with post hoc Fisher PLSD or Student's related t-test that day, and the effect was thus not seen on the distance mea-
was used. Statistical significance was accepted at p < 0 . 0 5 . For sure (Fig. 1B). Overall, the swim speed did not differ between
the correlation analysis a statistical program for linear regression any of the age groups (Table 1). Nor did the swim speed differ
(Statview) was used. compared to the groups of 3-month-old animals which were
tested in parallel with each age group (see Table 1).
RESULTS Consistent with previous reports (15,19), a large variability
in performance in the water maze task was observed in the 18-,
Spatial Memor 3' Performance
24- and 30-month-old animals, as illustrated in the scatter-plots
Escape latency and distance swum to platform. All animals in Fig. 2A-C. On the escape latency parameter (mean of the
showed learning of the Morris water maze task over the 5 test last two days of the test week) the 3-month-old animals per-
days as measured by reduction in the time to find the hidden formed with little variability (between 3 and 16 s) as did the
platform (escape latency) and distance swum during the test 12-month-old rats (between 6 and 14 s), except for one rat which
week (p<0.001, 2-way ANOVA, Fig. 1). However, the 24- and had a mean latency of 24 s (Fig. 2A). The 18-month-old ani-
16 FISCHER, CHI-IN. GAGti ,\ >, D BJt)R K Li ?,;i

a m"

-11

Q
h

o,

! e

FIG. 6. Photomicrographs of single-labelled neurons (CHAT) in the striatum of a 3-month-old ta) and a 30-month-old
animal (b). Loss of immunolabelled cells as well as cellular atrophy is evident in this nucleus. The scale bar !in panel
a) is tO0 ~m.

mals performed within a wider range compared to the younger age group. According to this criterion, impaired place naviga-
animal groups (between 3.5 and 35 s) and this variability was tion was found in 8% of the 12-month-old rats, in 45% of the
further pronounced in the 24- and 30-month-old groups, which 18-month-old animals, in 53% of the 24-month-old animals and
showed escape latencies that ranged between 4 and 57 s (24- in over 90% of the animals in the 30-month group. In the mor-
month-old rats) and between 13 and 60 s in the 30-month-old phometric analysis "nonimpaired'" rats denotes animals which
group (Fig. 2). performed within one SD of the parallel tested young animals.
The criterion for behaviorally impaired rats, used in the sub- Only two rats in the 30-month group (5%) conformed to this
sequent morphometric analysis, was an escape latency longer criterion, while 30% of the 24-month-old rats and 33% of the
than the mean + 2SD of the escape latency recorded in the group 18-month-old rats fell within this range.
of young (3-month-old) rats that were tested in parallel with each Place navigation. The heterogeneity in performance of the
SPATIAL MEMORY AND CHOLINERGIC NEURONS IN AGING 17

E
._E

z+-~~,

2~
0 0

"A

2z
~_~
d~
Is FISCHER, CHEN, GAGE AND Bff)RKI_.UNI~

MS striatum
2000 3ooo] °~o~
_5oo l t i 2r~z)

1500
. • 0 0 2000 "~ "1~ o ~ •

1ooo-
• 30~rK)
I
[] [] ~ 0 ~

8
500-
'°1
500
.
0 , ' ' ' 0 , i ,
0 10 20 30 40 50 60 0 10 20 30 40 50 60

esc. lat. day 4+5 (sec) esc. lat. day 4+5 (sec)

VDB NBM

ooo-
'°t
2500

2000'
O

1500 1 o 1500 "1~ + [] u-"'2"~*---~

8 0o •
1000 1 • • • 8 1000

500 500"

0 0 , , , , i

0 10 20 30 40 50 60 0 10 20 30 40 50 60

esc. lat. day 4+5 (sec) esc. lat. day 4+5 (sec)

FIG. 8. Correlation between escape latency (means of day 4 + 5) and FIG. 9. Correlation between escape latency (means of days 4 + 5) and
cell counts in the medial septum (MS) and the vertical limb of the diag- cell counts in the striatum and nucleus basalis magnocellularis (NBM) in
onal band of Broca (VDB) in 3-, 12-, 18- and 30-month-old animals. 3-, 12-, 18- and 30-month-old animals. Statistics and linear regression
Statistics and linear regression coefficients are given in Table 3. coefficients are shown in Table 3.

18-, 24- and 30-month-old animals was also evident in the spa- ranged between 36 and 75% and the 12-month-old animals be-
tial probe trial which was performed at the end of the fifth test tween 35 and 57% (Fig. 2C). These two groups did not differ
day. The young animals showed focused search over the former from each other, whereas the 18-, 24- and 30-month-old animals
platform site, as reflected in a high number of platform cross- were significantly impaired compared to the 3-month-old group
ings (7 - 0.4) and a high percentage of their swim path (50 --- 2%) (p<0.0001, one-way ANOVA, post hoc Fisher PLSD; Fig. 2C).
confined to the platform quadrant, whereas the impaired animals In the 18-month group the animals swam between 17 and 68%
swam randomly over the four quadrants of the circular tank and of the total distance in the third quadrant and were significantly
showed low numbers of platform crossings (Fig. 3). The num- worse in their performance compared to the 3-month-old ani-
ber of impaired animals (i.e., rats with fewer platform crossings mals. The 24-month-old animals ranged between 10 and 60%
and percentage values than any of the 30 rats in the 3-month and the 30-month-old animals between 2 and 53% (Fig. 2C).
group) increased from 1-2 in the 12-month group (8-16%) and
6-9 in the 18-month group (10-15%) to 24-26 in the 24-month Image Analysis
group (40%) and 28-30 in the 30-month group (75%).
The 12-, 18-, 24- and 30-month-old animals were signifi- All animals that were taken for image analysis showed clear
cantly impaired on the platform crossing parameter compared to cytoplasmic immunostaining for ChAT and/or NGFr in neurons
the 3-month-old group (p<0.0001, one-way ANOVA, post hoc of the forebrain (Fig. 4). In the MS (Fig. 5) and the VDB, la-
Fisher PLSD). The 12-month-old animals crossed the platform belling for the NGFr was exclusively found on ChAT-labelled
site over a range from 2-10 times (mean 6+-_ 1 crossings; Fig. neurons and virtually all ChAT-positive neurons were clearly
2B). The 18-month-old animals crossed the platform site from NGFr-positive in all age groups. Striatal ChAT-labelled neurons
1-11 times (mean 6---0.3 crossings) and the 24-month-old ani- (Fig. 6) showed no immunolabelling for the NGFr except for in
mals from 0-9 times (mean 4---0.3 crossings), and the 30- the most basal part, where a few scattered neurons could be de-
month-old animals from 0--4 platform crossings (mean 2___0.2 tected. In the NBM (Fig. 7) 75-80% of the ChAT-labelled neu-
crossings; Fig. 2B). On the %Q3 parameter (i.e., % of the swim rons stained for the NGFr. The percentage of ChAT-positive
path spent in the platform quadrant) the 3-month-old animals neurons that was immunoreactive for the NGFr did not change
SPATIAL MEMORY AND CHOLINERGIC NEURONS IN AGING 19

MS striatum
2000" 3000"
In * oo+
a od ~ $ o I: ~221 250o- ~o o
AA 4 0 4 - 0 0
1500" in i-1

° g .,, 8 ~ 2000-
_ g
a ..~" ~ o o • In

• o s
ooo ~ . | • 8 ]500 I n ~
[]
O o

l'" [] [] ~ lOOO"
5OO
500'

r I I I I I I I I I I , * , , I | 0 , , I , I I w , , I w , , , I w

-2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

platform crossings platform crossings

VDB NBM
30110"
0 4- [] O
2500" [] + o , J- 2500
4- o o o e _ . . . . r . - - ~ °o
2000" [] ~- -
00 8
o 0 0 ~ / ° -
., | o 0 o I. 3°=1
1500" I - i - . _ " °o o =8 1500 , ~ , , o ° 4-
• 0
28 lOOO-
• •

500- 500-

0 ~ i
~ | i ~ i i i J | i i i i i i 0 i i i i i v i , i i i i i i i i

-2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 -2-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

platform crossings platform crossings

FIG. 10. Correlation between number of platform crossings (in the spa- FIG. 11. Correlation between number of platform crossings (in the spa-
tial probe trial) and cell counts in the medial septum (MS) and the verti- tial probe trial) and cell counts in the striatum and nucleus basalis mag-
cal limb of the diagonal band of Broca (VDB) in 3-, 12-, 18- and 30- nocellularis (NBM) in 3-, 12-, 18-, 24- and 30-month-old animals.
month-old animals. Statistics and linear regression coefficients are given Statistics and linear regression coefficients are shown in Table 3.
in Table 3.

24-month nor the 18-month groups did the impaired and the
with age. As shown in Figs. 5-7, a clear loss of immunolabelled nonimpaired rats differ from each other (p>0.05).
neurons was observed in the 30-month-old animals compared to Cell size. Significant reductions in cell size compared to the
the 3-month-old animals and this change was equally evident 3-month group could be detected only in the 24- and the 30-
with both immunohistochemical markers. month groups. In the 30-month-old animals the reduction (in
Cell number. As is summarized in Table 2, significant dif- cross-sectional area) amounted to 14% in the MS, 10% in the
ferences in cell number of double-labelled neurons (CHAT+ VDB, 16% in the striatum and 13% in the NBM. The single-
NGFr) could be detected in the 30-month-old and the 24-month- labelled neurons in the NBM were reduced by 24% ( p < 0 . 0 5 ,
old groups compared to the 3-month-old animals. The reduction one-way ANOVA, post hoc Fisher PLSD). In the 24-month
in cell number in the 30-month-old animals (all of which were group a small but significant reduction in size of NGFr-positive
behaviorally impaired) amounted to about 27% (28%) in the MS neurons (6-8%) was observed in the MS, but not in VDB. In
(comparisons between noncorrected cell counts are given within neither the 18-month nor the 24-month groups did the cell size
brackets); 31% (33%) in the VDB; 23% (24%) in the striatum; differ between the impaired and the nonimpaired animals.
28% (30%) of the double-labelled neurons in the NBM; and 45%
(50%) of the single-labelled neurons in the NBM ( p < 0 . 0 5 , two- Correlation Between Morphometric Parameters, Behavioral
tailed Student's t-test). In the 24-month group, a significant re- Impairment, and Age
duction in the number of NGFr-positive neurons 19% (20%) was
evident in the MS and of 27% (28%) in the VDB in those aged As is illustrated in Table 2, 53 rats from five different age
animals that were impaired in the spatial memory task ( p < 0 . 0 5 , groups were taken for morphometric analysis. In the correlation
one-way ANOVA, post hoc Fisher PLSD). In the 18-month analysis we wanted to investigate whether the atrophic changes
group a significant reduction in cell number of 30% (30%) was (cell numbers and cell size) in the four forebrain cholinergic nu-
observed in the striatum in impaired animals ( p < 0 . 0 5 , two-tailed clei were correlated with either behavioral performance or age
Student's unrelated t-test). Otherwise the cell counts did not dif- among these 53 rats. The results are summarized in Table 3 and
fer between the 3-, 12- and 18-month-old rats. In neither the in the scatter-plots in Figs. 8-11. In all four nuclei, cell counts
2o FISCHER, CHEN, GAGE AND BJI)RKLUNI

Impaired (as a measure of spatial memory impairment), calculated ~ro,-~

30 r 100~ the data in Table 2, are plotted against advancing age for the
behaviorally impaired (A) and the behaviorally nonimpaired IB!

.2
/ [ 60 ~
"
rats. In the female Sprague-Dawley rats used here, the first be-
havioral and morphometric changes appeared in a subgroup ot
12-month-old rats. By 30 months, over 90% of the surviving rats
were severely impaired in the water maze task, and at thi~ time
10 ~ point all forebrain eholinergic regions showed atrophic change~
40 and reduced ChAT-positive and NGF-positive cell counts.
b~

o; 1

t eo
The morphometrie changes were significantly correlated no~
only with age but also with the escape latency and the platform
crossing measures in the water maze test, which supports the
o
view that age-dependent cholinergic neuronal atrophy and age-
?
- 1 0 " , lo dependent learning and memory impairments develop in paral-
A 12 m o 18 m o 24 rno 30 m o lel. The highest correlations with performance in the water maze
test were seen for ChAT/NGFr-positive cell counts in MS. VDB
and NBM, i.e., three structures which are known to play a role
in spatial memory performance in young rats isee (461 for re-
3o Non-impaired ,oo view]. Cell size, on the other hand, was better correlated with
age than any of the water maze parameters in these structures.
As seen in Fig. 12, the morphological changes were overall
>~e,. z0 less severe and tended to appear later than the place navigational
"4 e. o
1.2 impairments. Thus no significant changes in cell numbers or cell
60 ~
o
size were detectable in the MS, VDB or NBM in 18-month-old
I0 rats, despite the finding that a subgroup of these rats showed a
i 40 ,~ ~ moderate, but clear, behavioral impairment. Only in the 24-
~N
month-old rats, when severe place navigational impairments had
• ~ 0 developed, were there morphological changes manifest in MS
2o g?
and VDB. Consistent with our previous study (I 5) these changes
were most pronounced in the behaviorally impaired animals.
B -to 12 m o 18 m o 34 mo 30 mo
Although these data are consistent with the idea that degen-
erative changes in forebrain cholinergic neurons may contribute
to the cognitive decline in aging rats, the magnitude of cholin-
FIG. 12. Mean reduction in cell counts in the medial septum (MS; filled ergic neuron atrophy and/or degeneration does not appear to be
circles) and the vertical limb of the diagonal band of Broca (VDB; filled substantial enough to explain, by itself, the severely impaired
triangles)--measured as percent of the 3-month-old group--plotted against place navigational performance of the aged rats. Lesions in
the mean reduction in spatial memory performance (platform crossings) young rats which produce less than 80-85% depletion of hippo-
measured as percent of the 3-month-group in impaired (Panel A) and
campal ChAT levels produce only transient impairments (unpub-
nonimpaired (panel B) rats. Data is from Table 2. The figure illustrates
the overall parallel decline in the morphological and behavioral parame- lished observations), and nucleus basalis lesions (induced by
ters with age among impaired animals. quisqualic acid) which produce about 40-50% reductions in cor-
tical ChAT levels have little effect (12). in the most severely
affected aged rats the number of ChAT-positive cells was re-
were significantly correlated with both escape latency (during duced by only 10-30%, and the atrophy (in terms of cross-sec-
day 4 and 5 of the test week) and platform crossings (in the tioned area of the ChAT-positive cells) amounted to no more
spatial probe trial) as well as with age. The correlation coeffi- than 20-25%.
cients were highest for MS, VDB and NBM. In the NBM, cell Indeed, none of the most severely impaired aged rats had
counts showed equally good correlation with both escape latency ChAT/NGFr-positive cell numbers lower than 50% of normal in
and age, whereas in the MS and the VDB (when also 24-month- any of the regions analyzed. These data are generally in agree-
old animals were included) the best correlations were seen with ment with previous observations (2, 3, 7, 25), as well as our
the escape latency parameter. Inspection of the scatter-plots in own earlier findings (14,15), reporting reductions in both cholin-
Figs. 8-11 indicated that rats that performed poorly had overall ergic cell numbers and cell size in the forebrain of aged rats of
cell counts in the lower range of the young 3-month-old group, similar magnitude as those observed here. In aged mice, on the
and that high cell counts (in MS >1200; in VDB >2000; in other hand (28,42) reductions in cell size, but not in cell num-
NBM >1900; in striatum >2200) were seen only among rats bers, were observed with increasing age.
that performed well. Cell size showed overall better correlation It is interesting to compare the present results with those re-
with age than with escape latency or platform crossings in the ported from aged humans. Mann et al. (40), McGeer et al. (39)
water maze test. Again, the highest correlations were seen in the and Jacobs and Butcher (31) have reported a progressive decline
NBM, MS and the VDB. with age in neuronal cell numbers (up to 30-50%) in the nu-
cleus basalis of nondemented individuals over 65 years of age.
DISCUSSION This reduction is similar in magnitude to those seen in the old-
est (30-month-old) rats in the present study. Other investigators
The present results indicate that the atrophic changes in the (9,56), however, have failed to detect changes of this magnitude
forebrain cholinergic system and the decline in spatial learning during normal aging, and it seems clear that substantial cell loss
ability are parallel processes during aging. This is illustrated in in the nucleus basalis area (>40-50%) occurs only in severely
Fig. 12, where the average percentage changes in ChAT/NGFr- demented patients [see, e.g., (4, 31, 40, 39, 55)].
positive cell numbers in MS and VDB and platform crossings The cholinergic forebrain cell loss must therefore be consid-
SPATIAL MEMORY AND CHOLINERGIC NEURONS IN AGING 21

ered to be of modest magnitude in aged rats. The concomitant fected. The selective markers used here did not allow us to ad-
shrinkage, however, indicates that the remaining cholinergic dress this problem. Without concomitant data on other relevant
neurons may be atrophic and thus functionally impaired. Previ- neuronal systems, it is indeed difficult to interpret the parallel
ous biochemical and electrophysiological studies provide sub- changes in cholinergic neuron atrophy and behavior in causative
stantial support for this possibility. Thus Gibson et al. (24) and terms.
Sims et al. (50) have observed decreases in brain acetylcholine Studies in young animals using cholinergic receptor blockers
synthesis of up to 70% in aged mice and rats. Aston-Jones et al. or extensive lesions of the septo-hippocampal or basalo-cortical
(5) observed impaired impulse conduction from nucleus basalis projection systems have indicated that blockade of cholinergic
to cortex in aged rats, and Lamour et al. (34,35) have reported neurotransmission is sufficient to produce impairments in the
altered electrophysiological properties and axonal conduction ve- water maze task (54), but that the impairment that can be
locities in identified septo-hippocampal neurons in aged rats. In achieved in this way does not approach that seen in the severely
addition, electrophysiological studies (8, 27, 37, 38, 48) have affected aged rats. This discrepancy suggests that impairments
demonstrated a diminished responsiveness to acetylcholine in in other systems, besides the cholinergic ones, contribute to the
postsynaptic neurons in the hippocampus in aged rats. This re- age-dependent cognitive decline. Pharmacological experiments
duced postsynaptic cholinergic responsiveness may be corre- (53) and studies on combined cholinergic and serotonergic le-
lated, at least in part, with reduced muscarinic receptor binding sions (44,47) have shown that blockade of the serotonergic sys-
in the cholinergic terminal fields [see (11) for discussion]. Avail- tem greatly potentiates the spatial memory impairment induced
able data thus indicate that the degenerative changes observed by septal lesions or cholinergic receptor blockade. Indeed, young
morphometrically in the cholinergic system are accompanied by rats with combined cholinergic-serotonergic lesions appear to be
age-dependent decrements both in the functioning of the residual as affected as the most severely impaired aged rats.
cholinergic neurons as well as in the postsynaptic responsiveness Taken together, available data thus indicate that cognitive
of their targets. Nevertheless, some parameters of ACh synthe- impairments in aged rats reflect multisystem deficits. It should
sis (e.g., hippocampal ChAT-activity) and ACh release have be pointed out, however, that impaired place navigation in the
been observed to be maintained at relatively normal levels also water maze task may result not only from more specific deficits
in behaviorally severely impaired aged rats (15, 18, 23, 29, 49), in spatial learning and memory ability, but also from more gen-
which suggests that the aging cholinergic system retains the ca- eral performance deficits in the aged animals. While the water
pacity to undergo compensatory changes in response to the on- maze test allowed us to control for general motoric impairments,
going degenerative process. as reflected in the swimming ability and the swim speed, we
The histochemical markers used here, ChAT and NGFr, are cannot exclude that age-dependent impairments, e.g., in vision
dynamic markers in the sense that they may be lost without or endocrine functions may have contributed to the severe per-
death of the neurons themselves. Thus the reduced CHAT- and formance deficits in the oldest animals. However, since trans-
NGFr-positive cell numbers in aged rats do not necessarily mean plants of cholinergic rich cell suspensions, implanted into the
that the neurons have died. Indeed, observations after axotomy hippocampus or cortex, can improve the performance of cogni-
in the septo-hippocampal and vagal system, using prelabelling tively impaired aged rats in both water maze and delayed match-
with retrograde axonal tracers (17, 36, 45) have shown that ing tasks (13, 20, 21) it appears that such nonspecific performance
ChAT immunoreactivity may be lost from axotomized surviving deficits are not the only factors involved.
neurons. Similarly, in the aged animals, ChAT and NGFr may In conclusion, the results show that age-dependent degen-
be down-regulated in some affected neurons to a degree that erative and/or atrophic changes in the forebrain cholinergic nu-
would make them undetectable histochemically. It is notable, clei develop in parallel with spatial learning and memory im-
however, that the two markers were reduced to the same extent. pairments in aged rats. Although the magnitude of the degener-
Contrary to the suggestion by Koh et al. (33), there was no evi- ative changes in the cholinergic system is not substantial enough,
dence for a loss of NGFr from ChAT-positive neurons in any of by itself, to explain the severe spatial learning deficits that de-
the forebrain nuclei; either both markers were lost due to an ac- velop in the oldest animals, the present data are consistent with
tual cell death, or else they were simultaneously down-regulated the view that impaired function in the forebrain cholinergic sys-
in a portion of the affected neurons. tem could contribute to the age-dependent cognitive decline in
All four forebrain regions studied, i.e., MS, VDB, NBM and rodents.
striatum, were affected to a similar degree in the different age
groups (although data from NBM and striatum were lacking ACKNOWLEDGEMENTS
from the 24-month group). This indicates that the age-dependent We thank Alicja Flasch, Ulla Jarl, Agneta Othberg and Donna Okino
atrophic changes affect cholinergic neurons throughout the fore- for excellent technical assistance, Ragnar M~rtensson for preparation of
brain. This observation raises the question whether the underly- the diagrams and Agneta Persson for photographic work. The study was
ing process is a diffuse one, involving many types of neurons in supported by grants from the Swedish MRC (04X-3874), the National
the brain, or whether the cholinergic system is selectively af- Institute of Aging (A06088) and the Greta and Johan Kock Foundation.

REFERENCES
1. Abercrombie, U. Estimation of nuclear populations from microtome formation in cortical areas in Alzheimer's disease. Neuroscience 14:
sections. Anat. Rec. 94:239-247; 1946. 1-14; 1985.
2. Altavista, M. C.; Bentivoglio, A. R.; Crociani, P.; Rossi, P.; AI- 5. Aston-Jones, G.; Rogers, J.; Shaver, R. D.; Dinan, T. G.; Moss,
banese, A. Age-dependent loss of cholinergic neurons in the basal D. E. Age-impaired impulse flow from nucleus basalis to cortex.
ganglia of rats. Brain Res. 455:177-181; 1988. Nature 318:462-464; 1985.
3. Altavista, M. C.; Rossi, P.; Bentivoglio, A. R.; Crociani, P.; AI- 6. Batchelor, P. E.; Armstrong, D. M.; Blaker, S. N.; Gage, F. H.
banese, A. Ageing is associated with diffuse loss of forebrain Nerve growth factor receptor and choline acetyltransferase colocal-
cholinergic neurons. Brain Res. 508:51-59; 1990. ization in neurons within the rat forebrain: response to fimbria-fornix
4. Arendt, T.; Bigl, V.; Tennstedt, A.; Arendt, A. Neuronal loss in transection. J. Comp. Neurol. 284:187-204; 1989.
different parts of the nucleus basalis is related to neuritic plaque 7. Biegon, A.; Greenberger, V.; Segal, M. Quantitative histochemistry
22 FISCHER. (?HEN, GAG[i AND B J O R K L U N [ )

of brain acetylcholinesterase and learning rate in the aged rat. Neu- Aging 2:41-47; 1981.
robiol. Aging 7:215-217; 1986. 30. Ingram, D. Analysis of age-related impairments in learning and
8. Buzsaki, G.; Bickford, R. G.; Armstrong, D. M.; Ponomareff, G.; memory in rodent models. Ann. NY Acad. Sci. 44:312-331 : 1985.
Chen, K. S.; Ruiz, R.; Thai, L. J.; Gage, F. H. Electric activity in 31. Jacobs, R. W.; Butcher, L. L. Pathology of the basal forebrain m
the neocortex of freely moving young and aged rats. Neuroscience Alzheimer's disease and other dementias. In: Scheibel, A. B.,
26:735-744; 1988. Wechsler, A. F., eds. The biological substrates of Alzheimer'~ dis
9. Chui, H. C.; Bondareff, W.; Zarow, C.; Slager, U. Stability of ease. New York: Academic Press; 1986:87-100.
neuronal number in the human nucleus basalis of Meynert with age. 32. Koh, S.; Loy, R. Age-related loss of nerve growth lactor sensLtivit,,
Neurobiol. Aging. 5:83-88; 1984. in rat basal forebrain neurons~ Brain Res. 440:396-401; 1988.
10. Coleman, P.; Finch, C.; Joseph, J. The need tot multiple time 33. Koh, S.; Chang, P.; Collier, T. J.: Loy, R Los~, of NGF receptor
points in aging studies. Neurobiol. Aging 11 : 1-2; 1990. immunoreactivity in basal forebrain neurons of age rats: correlation
11. Decker, M. W. The effects of aging on hippocampal and cortical with spatial memory impairment. Brain Res. 498:397-404; 1989.
projections of the forebrain cholinergic system. Brain Res. Rev. 12: 34. Lamour, Y.; Dutar, P.; Jobert, A. Septo-hippocampal neurons: al-
423-438; 1987. tered properties in the aged rat. Brain Res. 416:277-282; 1897~
12. Dunnet, S. B.; Whishaw, I. Q.; Jones. G. H.; Bunch, S. T. Behav- 35. Lamour, Y.; Bassant, M. H.; Jobert, A.; Joly. M. Septo-hippocam-
ioural, biochemical and histochemical effects of different neurotoxic pal neurons in the aged rat: Relation between their electrophysiologi-
amino acids injected into nucleus basalis magnocellularis of rats. cal and pharmacological properties and behavioral performances.
Neuroscience 20:653-669; 1987. Neurobiol. Aging 10:181-186; 1989.
13. Dunnet, S. B.; Badman, F.; Rogers, D, C.; Evenden, J. L.; lversen, 36. Lams, B. E.; Isacson, O.: Sofroniew, M. V Loss of transmitter-
S. D. Cholinergic grafts in the neocortex or hippocampus of aged associated enzyme staining fi~llowing axotomy does not indicate
rats: reduction of delay-dependent deficits in the delayed nonmatch- death of brainstem cholinergic neurons. Brain Res. 475:401-406',
ing to position task. Exp. Neurol. 102:57-64; 1988. 1988.
14. Fischer, W.; Wictorin, K.; Bj6rklund, A.; Williams, L. R.; Varon, 37. Lippa, A. S.; Pelham, R. W.; Beer, B.; Critchett, D. J.; Dean, R.
S.: Gage, F. H. Amelioration of cholinergic neuron atrophy and L.; Bartus, R. T. Brain cholinergic dysfunction and memory in aged
spatial memory impairment in aged rats by nerve growth factor. rats. Neurobiol. Aging 1: 13-19: 1980.
Nature 329:65-68; 1987. 38. Lippa, A. S.; Loullis, C. C.: Rotrosen, J.; Cordasco, D. M.:
15. Fischer, W.; Gage, F. H.; Bjrrklund, A. Degenerative changes in Critchett, D. J.; Joseph, J. A. Conformational changes in muscar~
forebrain cholinergic nuclei correlate with cognitive impairments in inic receptors may produce diminished cholinergic neurotransmis-
aged rats. Eur. J. Neurosci. 1:34-45; 1989. sion and memory deficits in aged rats. Neurobiot. Aging. 6:317
16. Fischer, W.; Chen, K. S.; Gage, F. H.; BjOrklund, A. NGF im- 323:1985.
proves spatial memory in aged rodents as a function of age. J. Neu- 39. McGeer, P. L.; McGeer, E. G.: Suzuki, J.; Dolman, C. E.; Nagai,
rosci. 11(7):1889-1906; 1991. T. Aging, Alzbeimer's disease, and the cholinergic system of the
17. Fischer, W.; Bjrrklund, A. Loss of ACHE- and NGFr-labelling pre- basal forebrain. Neurology 34:741-745; 1984.
cedes neuronal death of axotomized basal forebrain neurons: rever- 40. Mann, D. M. A.: Yates, P. O.: Marcyniuk, B. Changes in nerve
sal by intraventricular NGF-infusion. Exp. Neurol. 113:93-108; 1991. cells of the nucleus basalis of Meynert in Alzbeimer's disease and
18. Fischer, W.; Nilsson, O. G.; Bj6rklund, A. In vivo acetylcholine their relationship to ageing and to the accumulation of lipofuscin
release is unaltered in the hippocampus of cognitively impaired aged pigment. Mech. Ageing Dev. 25:189-204; 1984.
rats with degenerative changes in the basal forebrain. Brain Res. 41. Markram, H.; Segal, M. Regional changes in NGF receptor immu-
556:44-52; 1991. nohistochemical labeling in the septum of the aged rat. Neurobiol.
19. Gage, F. H.; Kelly, P. A. T.; BjtJrklund, A. Regional changes in Aging. 11:481-484; 1990.
brain glucose metabolism reflect cognitive impairments in aged rats. 42. Mesutam, M. M.; Mufson, E. J.; Rogers, J. Age-related shrinkage
J. Neurosci. 4:2856-2865; 1984. of cortically projecting cholinergic neurons: a selective effect. Ann.
20. Gage, F. H.; BjtJrklund, A.; Stenevi, U.; Dunnet, S. B.; Kelly, P. Neurol. 22:31-36; 1987.
A. T. Intrahippocampal septal grafts ameliorate learning deficits in 43. Morris, R. G. M. Developments of a water-maze procedure lot
aged rats. Science 225:533-536; 1984. studying spatial learning in the rat. J. Neurosci. Methods 11:47-60;
21. Gage, F. H.; Bjrrklund, A. Cholinergic septal grafts into the hippo- 1984.
campal formation improve spatial learning and memory in aged rats 44. Nilsson, O. G.; Strecker, R. E.; Daszuta, A.; Bjrrklund, A. Com-
by an atropine-sensitive mechanism. J. Neurosci. 6:2837-2847; 1986. bined cholinergic and serotonergic denervation of the forebrain pro-
22. Gallagher, M.; Pelleymounter, M. A. An age-related spatial learn- duces severe deficits in spatial learning in the rat. Brain Res. 67:
ing deficit: choline uptake distinguishes "impaired" and "unim- 195-215; 1988.
paired" rats. Neurobiol. Aging 9:363-369; 1988. 45. O'Brien, T. S.; Svendsen, C. N.; Isacson, O.; Sofroniew, M. V.
23. Galtagher, M.; Burwell, R. D.; Kodsi, M. H.; McKinney, M.; Loss of True blue labelling from the medial septum following
Southerland, S.; Vella-Roundtree, L.; Lewis, M. H. Markers for transection of the fimbria-fornix: evidence for the death of cholin-
biogenic amines in the aged rat brain: Relationship to decline in ergic and noncholinergic neurons. Brain Res. 508:249-256; 1990.
spatial learning ability. Neurobiol. Aging. 11:507-514; 1990. 46. Olton, D. S.; Wenk, G. L. Dementia: animal models of the cogni-
24. Gibson, G. E.; Peterson, C.; Jenden, D. J. Brain acetylcholine syn- tive impairments produced by degeneration of the basal forebrain
thesis declines with senescence. Science 213:674--676; 1981. cholinergic system. In: Meltzer, H. Y., ed. Psychopharmacology:
25. Gilad, G. M.; Rabey, J. M.; Tizabi, Y.; Gilad, V. H. Age-depen- the third generation of progress. New York: Raven Press; 1987:941-
dent loss and compensatory changes of septohippocampal cholin- 953.
ergic neurons in two rat strains differing in longevity and response 47. Richter-Levin, G.; Segal, M. Spatial performance is severely ina-
to stress. Brain Res. 436:311-322; 1987. paired in rats with combined reduction of serotonergic and cholin-
26. Gomez-Pinilla, F.; Cotman, C. W.; Nieto-Sampedro, M. NGF-re- ergic transmission. Brain Res. 477:404---407; 1989.
ceptor immunoreactivity in aged rat brain. Brain Res. 479:255-262; 48. Segal, M. Changes in neurotransmitter actions in the aged rat hip-
1989. pocampus. Neurobiol. Aging. 3:121-132; 1982.
27. Haigler, H. J.; Cahill, L.; Crager, M.; Charles, E. Acetylcholine, 49. Sherman, K. A.; Kuster, J. E.; Dean, R. L.; Bartus, R. T.; Fried-
aging and anatomy: differential effects in the hippocampus. Brain. man, E. Presynaptic cholinergic mechanisms in brains of aged rats
Res. 362:157-160; 1985. with memory impairments. Neurobiol. Aging 2:99-104; t981.
28. Hornberger, J. C.; Buell, S. J.; Flood, D. G.; McNeill, T. H.; 50. Sims, N. R.; Marek, K. L.; Bowen, D. M.; Davison, A. N. Pro-
Coleman, P. D. Stability of numbers but not size of mouse fore- duction of [V~C] acetylcholine and [t'*C] carbon dioxide from [U -14]
brain cholinergic neurons to 53 months. Neurobiol. Aging 6:269- glucose in tissue prisms from aging rat brain. J. Neurochem. 38:
275; 1985. 488-492; 1982.
29. Ingram, D. K.; London, E. D.; Goodrick, C. L. Age and neuro- 51. Strong, R.; Hicks, P.; Hsu, L.; Bartus, R. T.; Enna, S. J. Age-
chemical correlates of radial maze performance in rats. Neurobiol. related alterations in the rodent brain cholinergic system and behav-
S P A T I A L M E M O R Y A N D C H O L I N E R G I C N E U R O N S IN A G I N G 23

ior. Neurobiol. Aging 1:59453; 1980. blockade. Brain Res. Bull. 23:347-358; 1989.
52. Taniuchi, M.; Schweizer, J. B.; Johnson, E. M. Nerve growth fac- 55. Whitehouse, P. J.; Price, D. L.; Struble, R. G.; Clark, A. W.;
tor receptor molecules in rat brain. Proc. Natl. Acad. Sci. 83:1950- Coyle, J. T.; De Long, M. R. Alzheimer's disease and senile de-
1954; 1986. mentia: loss of neurons in the basal forebrain. Science 215:1237-
53. Vanderwolf, C. H. Near-total loss of 'learning' and 'memory' as a 1239; 1982.
result of combined cholinergic and serotonergic blockade in the rat. 56. Whitehouse, P. J.; Parhad, I. M.; Hedreen, J. C.; Clark, A. W.;
Behav. Brain Res. 23:43-57; 1987. White, C. L., III; Struble, R. G.; Price, D. L. Integrity of the nu-
54. Whishaw, I. Q. Dissociating performance and learning deficits on cleus basalis of Meynert in normal aging. Neurology. 33(Suppl. 2):
spatial navigation tasks in rats subjected to cholinergic muscarinic 159; 1983.