Mechanisms o f Ageing and Development, 9 (1979) 143-162
143
©Elsevier Sequoia S.A., Lausanne - Printed in the Netherlands
A G E P I G M E N T S , C E L L LOSS A N D H I P P O C A M P A L F U N C T I O N *
K. R. BRIZZEE and J. M. ORDY Delta Regional Primate Research Center, Tulane University, Covington, La. 70433 [U.S.A.)
(Received January 12, 1978) SUMMARY The specific aims of this study were to perform direct correlational analyses of age differences in learning, short-term memory and arousal in relation to cell loss and lipofuscin increase in the hippocampus CA1 zone and in visual area 17 of the Fisher 344 rat. The following tentative conclusions can be made from the results presented in this study: (1) significant age differences in 2 and 6 hour passive-avoidance retention or memory between mature and old rats were related to non-significant age differences in days to criterion learning, starting latencies, running distance and time in original approach learning, and (2) significant age differences in 2 and 6 hour retention of old, compared to mature rats were correlated significantly with loss of neurons, and very significantly with increases in intraneuronal llpofuscin in the hippocampus CA1 zone and in visual area 17.
INTRODUCTION Experiments with normal human subjects have indicated that age declines in learning and short-term memory may have been overestimated in earlier cross-sectional studies, but that they do represent important sources of performance decrements with age, even in carefully controlled longitudinal studies [1--4]. Other studies have implicated age differences in autonomic nervous system arousal as an important factor in the behavioral decrements of the elderly [5-8]. Although definitions and operational criteria for "normal" aging as distinct from neuropathology remain to be determined [9], comparisons of behavioral and neuropathological differences between normal "nondemented" and "demented" elderly subjects have shown that the most common symptoms of aging in both groups include loss of short-term memory, disturbance in spatial orientation, and alteration in affect, mood, or arousal [10--12]. Until recently, it has been generally assumed that the age dependent cognitive impairments of both normal subjects and those afflicted with senile dementia, organic brain syndrome, and other brain diseases of the elderly may be related to functional and structural changes, including cell loss, in the neocortex [13]. However, more recent studies have reported that senile *Basedon a paper presented at a specialgroup of Symposiaentitled, "Frontiers in AgingResearch", arranged by the program committee of the Biological Seiene~ Section of the Gerontological Society, San Francisco Meeting, November 18-22, 1977.
144 plaques, neurofibriUary tangles, granulovacuolar degeneration, and possibly cell loss may occur in the limbic system, particularly the hippocampus, long before similar morphological changes become observable in the neocortex of man during senescence [12]. Based on the same histological evidence, other neuropathological studies have confirmed that the limbic system may be considerably more vulnerable than the cerebral cortex to these microscopic lesions in presenile and senile dementia and Alzheimer's disease [ 14]. Although not concerned with such functional aspects as learning, memory, or arousal directly, morphological studies have reported age-related cell loss from the cerebral cortex of man [15-17], rhesus monkey [18] and the rat [19]. However, whereas the loss of neurons from the cerebral cortex and limbic system of man has not been determined definitively with quantitative procedures [9], the accumulation of lipofuscin pigment in the cortex and hippocampus of man [20], rhesus monkey [21] and rat [22] has been one of the more consistent morphological findings associated with aging, and the possible loss of neurons [9, 21 ]. The two regions of the brain which have also received increasing attention in animal studies of age-related differences in learning, memory, and arousal are the cortex and limbic systems [23-25]. However, age decrements in short-term memory have been difficult to establish in man and animals since memory declines may be related to reduced sensory or learning capacity, motivation or arousal, loss of motor coordination, or some combination of these factors [26-28]. In animal studies, passive-avoidance tests have been devised to determine to what extent age differences in learning, memory, and arousal are important sources or determinants of age decrements in performance [29]. The possible role of the hippocampus in learning and memory has attracted the attention of an increasing number of investigators [30--33]. It has been established that rats rely on a constellation of visual, auditory, olfactory, and somatosensory cues for creating a "short-term spatial memory map" in the hippocampus [34]. Rats with hippocampal lesions commit a significantly greater number of errors in incorrect "response perseveration", and therefore appear less flexible in short-term memory consolidation [30]. In studies of aging, age differences in short-term 2 and 6 hour memory have been reported among young, mature and old Fisher 344 rats [35, 36]. Compared to 16 month middleaged rats, 34 month senescent hooded rats have been reported to show short-term retention deficits on a complex spatial discrimination maze [37]. In neurophysiological studies, significant age-differences have been reported in synaptic potentials of pyramidal neurons in the hippocampus of old Fisher 344 rats [25], and in synaptic responses recorded from perforant path granule cells of the hippocampus in old hooded rats [37]. Morphological studies have reported neuronal degeneration and astroglial hypertrophy [25], and loss of axosomatic and axodendritic synapses in the hippocampus of old rats [38, 39]. Direct correlational studies of age differences in learning, memory, and arousal in relation to cell loss and lipofuscin accumulation in the hippocampus and cortex of man are not feasible, and they also have not been reported in the Fisher 344 rat previously. A preliminary report of age differences in short-term memory and cell loss from the visual" cortex of the Fisher 344 rat has been published [36]. The aims of this study were: (1)
145 compare the performance of 5 young, 5 adult, and 5 old Fisher 344 male rats on days to criterion for learning the location of food reinforcement in the goal box of a straight runway; (2) examine age differences in 2- and 6-hour short-term retention or memory of the shock-motivated passive-avoidance of the previously learned location of food reinforcement in the goal box; (3) establish age differences in behavioral arousal as determined in a pain-elicited aggression test; (4) evaluate age differences in neuron and glia cell populations in the hippocampus CA1 area and the visual cortex area; (5) establish age differences in intraneuronal lipofuscin accumulation in the same areas of the hippocampus and visual cortex; and (6) compute the intercorrelations of age with learning, Short-term memory, pain-elicited aggression, neuronal cell populations and intraneuronal lipofuscin in the hippocampal CA1 area and in visual cortex area 17.
MATERIALS AND METHODS
Subjects The subjects for this study were 5 young (11 months), 5 adult (17 month) and 5 old (29 months) Fisher 344 male rats. According to data on the maximum life span of the Fisher 344 rat, the median or 50% mortality occurs at 22 months, and a maximum survival extends to 39 months in this strain [40]. It should be noted that the age groups of 1 I, 17 and 29-months used in the present studies include the periods of development, maturity and senescence. The 29-month age group was located within the interval from the median to the known maximum life span of the Fisher 344 rat. Learning~pproach, passive-avoidance memory tests The learning and memory tests were carried out in a straight runway measuring 132 cm with start and goal boxes of 21 cm located on each end of the runway [36, 41]. For the 6-day learning or approach phase, each rat was run 3 trials a day to a criterion of obtaining food from the goal compartment on all 3 trials for 3 consecutive days. On the 7th day, when rats attempted to obtain the food pellet, they completed an electric circuit between the metal food cup and metal plate on the floor, and this resulted in a 2.5 mA shock. Two and 6 hours after shock, the rats were tested for retention or short-term memory of approach to the food reward in the goal box. Dependent measures for passiveavoidance or short-term memory retention were (a) start box latency, (b) total running time (with an upper limit of 2 min) towards the goal box, and (c) the distance traversed from the start to the goal box which was 132 cm, divided into 20 units (U) with 6.6 cm/U. Pain-elicited aggression, behavioral arousal tests The 5 rats within each of the 3 age groups were paired with an age matched control and tested for pain-elicited aggression and behavioral arousal in a 30 X 32 X 30 cm Plexiglas chamber with a grid floor. The electric shock (2.0 mA scrambled) was delivered through the grid floor on a random schedule with a mean of 15 shocks per minute during
146 a 5 minute test session. The dependent variables recorded in this test were the number of fights, and the cumulative time in seconds of posturing in each fight during the test session.
Morphometric evaluations of neurons and glia in hippocampus CA1 and visual cortex area 17 At the end of the behavioral testing, all 15 rats were perfused with a formalin, acetic acid and methyl alcohol solution (FAM). Serial transverse sections of 20 and 4/am were obtained from each rat hippocarnpus CA1 zone and visual cortex area 17, respectively. The 20 /am sections were stained with toluidine blue. Differential cell counts of neurons and glia in the hippocampus were obtained by a human observer. The counts were made by focusing a grid reticule through the full depth of the 20/am toluidine bluestained sections. Neurons and glia were counted within a 110 /am segment of CA1. In the hippocampus, differential neuron and glia counts were made in 3 adjacent but separate sites in the central region of the pyramidal layer in the CAI zone. In the differential cell counts, all nucleoli of neuron nuclei, and all glial nuclei were counted separately in accordance with the differentiating morphological criteria. Total neuron and glia counts for each rat hippocampus CA1 zone were based on 12 reticule counts in four sections. For the morphometric evaluations of cell-packing density in area 17 of the visual 6ortex, histological sections of 4/am were obtained from the widest zone of area 17 [42]. The histological sections were stained with methyl green which is selective for DNA, Differential neuron and glia counts were performed on 4 adjacent 4/am sections with an automated image analyzer (Leitz TAS) [36]. Differential neuron and glia counts were based on differences in optical density of neuronal and glial nuclear DNA. Cell counts were expressed as neuron and glia cell packing density of visual cortex, area 17 (cells/ mma/20U) X 10-a. Quantitative histological evaluation of lipofuscin by fluorescence For the histological quantification of lipofuscin, 4 /am sections from the hip. pocampus and visual cortex of each of the 15 rats were stained with pyronin Y and examined under blue-light fluorescence (approximate wavelength 420 nm), with an incident-light fluorescence microscope, equipped with an Osram XBO high pressure xenon lamp. All sections were evaluated with a 100X achromatic off immersion objective. The histological quantification of lipofuscin was carried out with the use of a measuring point ocular (25 cross hairs randomly distributed) which was focused to the midplane of each 4/am section (depth 2/am). In the hippocampus, pigment counts were made with the measuring point ocular at each of 3 adjacent sites in the CA1 zone in each of 4 sections. "Hits" were recorded selectively on neuron perikarya, in neurons in which the nueleolus was identifiable, in autofluorescent 0ipofuscin) bodies in such perikarya, and on all other structures including neuron nuclei, vascular dements and neuropil. From the "hits" recorded, the mean proportion of the perikaryon occupied by lipofusein was calculated and expressed as the percentage of the volume of the perikaryon.
147 In the visual cortex, area 17, the lipofuscin quantification procedure was essentially the same except that the measuring point ocular was focused at 20 successive, equally spaced depth levels from the most superficial to the deepest laminae of the cortex. A complete sampling of all 20 levels in one column constituted a single "traverse". Four such traverses were made in each of 4 sections for each of the 15 rat brains.
Body and organ weights Although the major focus in the study was on learning, short-term memory and arousal in relation to cell loss and lipofuscin in the brain, body, pituitary and adrenal, weights expressed in grams were also compared as part of the morphological evaluations. Since the rats were perfused for the morphometric evaluations, the organ weights represent total perfused weight.
Statistical evaluations In the cross-sectional experimental design, age represented the independent variable. Learning, short-term memory, pain-elicited aggression or behavioral arousal, the morphometric cell counts and intraneuronal lipofuscin in the hippoeampus and visual cortex, constituted the quantitative dependent variables. The 3 age groups were selected to cover development, maturity, and senescence. Following analysis of variance (ANOVA) of differences among groups, in the specific comparisons of age differences between the 11-month young and the 17-month mature age groups, 2-tared t-tests were then used since no assumptions were made concerning the direction of age differences. Since age deficits were predicted in the 29-month senescent group, one-tailed t-tests were used for evaluating the significance of age differences between the 17-month and 29month senescent age group. Intercorrelations of age with learning, 2 and 6 hour memory, pain-elicited aggression, neuronal cell populations and intraneuronal lipofuscin in CA1 and area 17 were computed with an "all possible multiple regression analysis" [43]. RESULTS
Approach-learning, passive avoidance-memory Regarding age differences in approach-learning and passive avoidance 2 and 6 hour memory, the 11-month group reached the learning criterion significantly faster in a mean of 1.8 as opposed to 4.2 days by the 17 month mature group ( P < 0.04). The 29-month senescent group reached the learning criterion in a mean of 3.4 days. This mean did not differ significantly from the days to criterion means of the other 2 age groups. There were no other age differences in starting latency, running time or running distance in the approach-learning phase in the runway test. The results of the approach-learning phase are illustrated in Fig. 1. On day 7, the rats received shock when they attempted to obtain food from the cup in the goal box. They were returned at 2 and 6 hours for passive-avoidance memory. Differences in starting latency, running distance and time were used as measures of retention.
148
AOIIIfAI~ OlIMOITHS
IITMONTHS it 29MONTHS
LIARNING CmTIRION lAYS
START LATENCY (SIC) s 6 De,VII
:f
"°f ~o
'oo°
|llll!C)ltlilt'fll
11.4~ • IOItSOAYII
400
Fig. 1. Results of approach learning for 11, 17, and 29 month rats on (1) days-to-criterion of learning; and (2) starting latencies. Group means of sums of seconds for 6 days (top). Running time group mean of sums in seconds for 6 days; running distance cm for 6 days (bottom). t.ATENC't Ttlt~
I 1 ~ OI~'IrMCE lt.acml • ! 0
•~ 2o'XX "F,,I 02
$
02
g
02
g
Fig. 2. Passive-avoidance 2 and 6 hour retention-memory. Group means of start latency, running time and distance at 0, 2 and 6 hours. In comparisons of age differences in two and six hour passive-avoidance retention or short-term memory, the mean running time of the 29-month senescent group from the start towards the goal box was markedly faster at 2 hours and significantly faster at 6 hours compared to the mean of the 17-month group (P < 0.01). The mean running distance from the start towards a goal box at 2 hours after shock was also significantly greater at 18.8 units o f the 29-month senescent, as opposed to the mean of 15.0 units o f the 17-month group (P < 0.04). The age differences in starting latency were not significant among the 3 age groups. Figure 2 illustrates the age differences in 2 and 6 hour passive-avoidance retention or short-term memory.
149
Pain-elicited aggression, behavioral arousal The rats within each of the 3 age groups were paired with an age matched control and tested for pain-elicited aggression and behavioral arousal. The mean number of 0.6 fights in the 29-month group during the 1st test session was significantly less than the 6.6 mean of fights per session of the 11-month age group. Similarly, the mean of the cumulative posturing time of 3.4 s per session of the 29-month group was significantly less than the 88.8 mean of the 11-month age group. The number of fights and the cumulative posturing time per session of the 17-month group did not differ significantly from the 11 or 29-month age groups. Figure 3 illustrates the number of fights and cumulative posturing during the 1st sessions in the pain-elicited aggression and behavioral arousal tests.
Pooture
St©. I001
Fights
No. 6
6O
4
4O
~o:~ "%<.ore %',, ~ l k , ~ 4;'" "
,d, ".
Fig. 3. Results of pain-elicited aggression, behavioral arousal. Number of fights, and group means of cumulative posture time in seconds.
Body and organ weights Regarding age differences in body and organ weights, the mean body weight of 326 grams of the 29-month old senescent group was significantly less than the mean of 364 grams of the 17-month mature group, as well as the 356 gram mean of the ll-month group (P < 0.05). Brain weights of 2.15, 2.19 and 2.16 grams respectively did not differ significantly among the 3 age groups. The mean pituitary weight of 1.4 grams of the 11month young group was significantly lower than the mean of 3.2 of the 17-month group, as well as the 3.2 grams of the 29-month old group (P < 0.01). Both the left and right adrenals of the 11-month group were significantly lower than both adrenals in the 17month and 29-month old age groups (P < 0.05). Figure 4 illustrates age differences in body and organ weights.
Hippocampus CA1; age differences in ceU populations and intraneuronal lipofuscin There is still considerable controversy concerning the nomenclature, topographic boundaries or zones within the mammalian hippocampal formation. It is composed of the subiculum, Ammon's horn, and the dentate gyrus. Based primarily on the anatomical appearance of pyramidal cells, Ammon's horn has been subdivided into zones CA1, CA2,
150 9
300 200 I00
.05
i
~,
,03
OE
2
,02
,04
I
OOdy
BrQin
Pltuito~ry
Adr(m41t
Fig. 4. Body and organ weightsof the Fisher 344 male rat at 11, 17, and 29 months.
CA3 and CA4 [44]. Figure 5 represents a low-power photomicrograph showing the rat hippocampus. Differential counts of neurons, gila, and intraneuronal lipofusein were performed in the CA1 zone in the area located between the arrows (Fig. 5). In the morphometric evaluations of cell populations in the hippocampus CA1 zone, the neuronal cell mean of 24.75 cells for a segment of 110 ~na of CA1 zone in the hippocarnpus of the 29-month old group was significantly less than the mean of 28.63 of the 1 l-month group (P < 0.03). The neuron population mean of 25.79 of the 17-month ~roup did not differ significantly from the population means of the other two age groups. The age differences in the combined glia, endothelial cell and pericyte populations among the three groups were not significant. Regarding age-differences in intraneuronal lipofuscin in the hippocampus CA1 zone, the mean percentage of neuronal perikaryon volume occupied by lipofuscin of the 1 l-month old group was 13.62;in the 17-month group it was 26.59, and in the 29-month group it was 35.58%. The progressive age increases in intraneuronal lipofuscin from 11 to 17, and from 17 to 29 months were all ldghly significant (P
Visual area 17; age differences in cell packing density and intraneuronal lipofuscin In the morphometric evaluations of age differences in cell populations in the visual cortex, the neuron packing density mean of 77.0 (ceUs/mm3/2OU) X l0 -s of the 29month group was significantly lower than the mean of 93.8 (cells/mm3/2OU) X 10-s of the 17-month age group (P < 0.02) (Fig. 8). The differences in packing density of the combined gila, endothelial cell and pericyte populations among the 3 age groups were not significant. Regarding intraneuronal lipofuscin accumulation in visual area 17, the mean percentage of neuronal perikarya content of llpofuscin was 18.38 in the 11-month, 24.17
Fig. 5. Photomicrograph of transverse section of hippocampus showing the portion of the CA1 zone (arrows) in which the cell counts in the stratum pyramid&e and depth measurements of the stratum oriens and stratum pyramidale were performed. Toluidine blue. X 35.
L1POFUWN % FmKAR?w 30 I:
OLI4 CELLS/II0 r 30
f
I
4
p-n
t
‘m
kk I *O‘77
IO
--__-_
0
___--0
80
10
20
20
30
20
30
40
__--_
____
40
Fig. 6. Age differences in neurons, lipofuscin and glia in hippocampus CA1 zone.
in the 17-month, and 30.65 in the 29-month age groups. The progressive age increases in intraneuronal lipofuscin content from 11 to 17 and from 17 td 29 months were $l highly significant (P < 0.01) (Fig. 8). The relative amount, and the appearance of lipofbscin in the large pyramidal neurons of lamina V in area 17 of young, middle age and old animals are shown in Fig. 9.
152
'i
J Fig. 7. (A) Photomicrograph of pyronin-stained 4 am section of pyramidal cell layer in CAI zone of hippocampus viewed with blue-light fluorescence. Fine, faintly fluorescent lipofuscin granules (arrows) are seen sparsely scattered throughout the perikarya of several neurons. In pyronin-stained preparations the nucleoli of neurons also exlu"oited a white fluorescence, but are distinguished from bodies by their nuclear location and white color. Lipofuscin bodies exhibit a yellow-orange orescence, with the color becoming more intense with age. × 700. (B) Similar transverse section in middle aged rat, showing more distinct small lipofuscin bodies (arrows) in neuron perikaryon. × 700. (C) Comparable site in CA1 zone of aged rat showing characteristic large accumulation of lipofuscin (arrow) in neuron perikaryon. × 700. Fig. 9. (A) Photomicrograph of pyronin-stained 4 pzn transverse section of pyramidal neuror~ in lamina V of visual cortex (area 17),.viewed with blue-light fluorescence. A few small lipofusein granules are seen in the perikarya of two neurons (arrows). X 700. (B) Similar transv~se section in middle age rat, showing larger iipofuscin body composed of several loosely aggregated lipofuscin granules, characteristic of this age level (arrow). X 700. (C) Comparable site in lamina V of visual cortex in aged rat showing numerous relatively large lipofusein bodies (arrows). Note that every neuron perikarya in the field exhibits lipofuscin. X 700.
153 VISUAL
AREA 17
NEURONS (cef~s/rnm3/2OU)xlO-3 80
60
~ '~
L JPOFUSCIN %/per*kor]mn
GLIA + (cel Is/rnm~'20) x JO"3
40
80
30
60 Ot
20 O~
I0
oj~O
600
40 20
I0
CORTICAL DEPTH lu,m
NS
to
300
•/10 20
• 30
30
30
3O
Fig. 8. Age differences in neurons, lipofuscin, ;Ha, cortical depth among 11, 17, 29 month rats.
Fig. 9. (For legend please see facing page.)
154
Hippocampus CAI; correlation of neuronal population decrease, and intraneuronal lipofuscin increase with age Neuronal cell populations and intraneuronal lipofuscin were determined quantitatively in the same CA1 area in each of the rats of the 3 age groups. A Spearman rank order (the) correlation coefficient was computed between the age related decrease in neurons and increase in intraneuronai lipofuscin content in CA1. The correlation coefficient rho was r = 0.49; n = 15; P < 0.03, indicating a significant inverse or negative relation between loss of neurons and increase in intraneuronal lipofuscin pigment from 11 to 17, and from 17 to 29 months of age in CAl. Expressed as relative percentage of the age differences of neuronal cell populations and lipofuscin, the decrease in neuronal cell population from 11 to 17 months was 7%, from 17 to 29 months it was 8% and from 11 to 29 months it was 14%. The increase in intraneuronal lipofuscin from 11 to 17 months was 100%, from 17 to 29 months it was 35%, and from 11 to 29 months it was 169%.
Visual area 17; correlation of neuronal population decrease, and intraneuronal lipofuscin increase with age In visual area 17, the Spearman rank order correlation coefficient rho was r = 0.67; n = 15; P < 0.01, indicating a highly significant negative or inverse relationship of neuronal cell loss and intraneuronal lipofuscin accumulation from 11 to 17 and from 17 to 29 months of age. The decrease in neuronal cell population from 11 to 17 months was 8%, from 17 to 29 months it was 18%, and from 11 to 29 months it was 25%. The increase in intraneuronal lipofuscin from 11 to 17 months was 33%, from 17 to 29 months it was 25%, and from 11 to 29 months it was 67%. Figure 10 illustrates age differences in neuronal cell population decreases and in intraneuronal lipofuscin increases among the three age groups expressed as percentage of the 11 month age group.
Intercorrelations o f age with behavior and morphology variables The major focus of the study was on age-related differences in learning, memory, and pain-elicited aggression in relation to loss of neurons and increases in lipofuscin in the hippocampus and visual cortex. Consequently, in addition to the separate univariate statistical analyses of age differences on each dependent variable, intercorrelations were computed of age with all 10 of the appropriate behavioral and morphological variables. The 10 variables intercorrelated with variable 1 (age) included: learning, 2 - days to criterion (Days Crit.); memory, 3 - running time (R. T. 2 h); 4 - (R. T. 6 h); 5 - running distance (R. D. 2 h); 6 - (R. D. 2 h); 6 -- R. D. 6h; pain-elicited aggression; 7 - no. of fights, neurons; 8 - hippocampus; 9 - area 17; lipofuscin; 10 - hippocampus; 11 visual area 17 (see Table I). According to the intercorrelation analysis, the running time at 6 hours, used as a measure of short-term memory, was correlated significantly and negatively with age (-0.55, P < 0,05). Also, the number of fights in pain-elicited aggression was significantly and negatively correlated with age (-0.57, P < 0.05). Loss of neurons from the hippocampus CAI zone was also significantly and negatively correlated with age (-0.52, P < 0.05). Loss of neurons from visualarea 17 was correlated negatively and highly significantly with age (-0.73, P < 0.01). It seems important to note that of
155 NEURONS Is/HOWl/.m 30
.~- rho=-049, p(.03 .,11.------ LIPOFUSCIN %/PE RIKARYON 40
eel
~o
~
. . . . . . . . .
"-~'~
17
II
,o
29
NEURONS ~ rho = - 0.67", p ( 0 0 1 4 - - " LIPOFUSC|N (Cells/mm3/2OO) x I0 -3 % / PERIIC~RYON 120~ 40 Jo0
3o
80
t
•
@N i.=o
"°
_----_NN II
17
AOE
,o
29
MONTHS
Fig. 10. Age differences in neuronal decreases and lipofuscin increases of 29 and 17 month rats expressed as percent of the 11 month group. Top, l~ppocampus CA1 zone;bottom, ~sual cortex area 17. the significant intercorrelations of age with behavior and brain morphology, the accumulation of lipofuscin with age was correlated highly significantly in the hippocampus (0.87, P < 0.01), and in visual area 17 (0.96, P < 0.01). Table I contains the intercorrelations of age with learning, memory, pain-elicited aggression, loss of neurons and increase of lipofuscin in CA1 and area 17.
DISCUSSION The specific aims of this study were to perform direct correlational analyses of age differences in learning, short-term memory and arousal in relation to cell loss and lipofuscin increase in the hippocampus CA1 zone and in visual area 17 of the Fisher 344 rat. The following tentative conclusions can be made from the results presented in this study: (1) significant age differences in 2 and 6 hour passive-avoidance retention or memory between mature and old rats were related to non-significant age differences in days to criterion learning, starting latencies, running distance and time in original approach learning, and (2) significant differences in 2 and 6 hour retention of old, compared to mature rats were correlated significantly with loss of neurons, and very significantly with increases in intraneuronal lipofuscin in the hippocampus CAI zone and in visual area 17. However, in the interpretation of these results, numerous factors need further clarification. Some of these include: (I) the effects of age on learning and memory as determined in the passive-avoidance test, (2) the use of pain-elicited aggression
0.30
-0.49
-0.55*
0.46
0.22
Leaxning 2. Days Crit.
Mem, Time 3. R.T. 2 h
4. R. T. 6 h
Mem. Dis. 5. R. D. 2 h
6. R. D. 6 h
0.87**
0.96**
9. Visual 17
Lipofuscin 10. Hippo.
11. Visual 17
0.18
0.37
1.00
0.30
0.34
0.39
-0.07
-0.48
-0.05
0.14
-0.15
2
-0.42
-0.37
0.36
-0.12
0.16
-0.28
-0.47
0.64**
1.00
0.37
-0.49
3
* Correlations among variables ate significant (P < 0.05). ** Correlations among variables are significant (P < 0.01).
-0.52*
-0.73**
Neurons 8. Hippo,
-0.57*
7. No. Fights
Pain E. Agg.
1.00
1
Variables 1-11
1. Ago
Variables 1-11
-0.46
-0.33
0.33
0.06
0.28
-0.55*
-0.72**
1.00
0.64**
0.18
-0.55*
4
0.46
0.45
0.38
-0.11
0.00
-0.14
0.72**
1.00
-0.72**
-0.47
-0.15
5
0.14
0.22
0.29
0.20
0.17
0.02
-0.21
1.00
0.72**
-0.55*
-0.28
6
INTERCORRELATIONS OF AGE, BEHAVIOR AND MORPHOLOGY VARIABLES
TABLE I
-0.63**
-0.55*
0.52*
0.49
1.00
-0.21
-0.14
0.28
0.16
-0.05
-0.57*
7
-0.56*
-0.50
0.60*
1.00
0.49
0.02
0.00
~ 0.06
-0.12
-0.48
-0.52*
8
-0.70**
-0.54*
1.00
0.60*
0.52*
0.17
-0.11
0.33
0.36
-0.07
-0.73**
9
0.87**
1.00
-0.54*
-0.50
-0.55*
0.20
0.38
-0.33
-0.37
0.39
0.87**
10
1.00
0.87**
-0.70**
-0.56*
-0.63**
0.29
0.45
-0.46
-0.42
0.34
0.96**
11
157
as a measure of arousal, (3) the rate of loss of neurons and increase in intraneuronal lipofuscin across the 3 age groups, (4)the possible role of intraneuronal iipofuscin accumulation in the loss of neurons, (5) the intercorrelations of age with the behavioral and morphological variables, and (6)the possible role of the hippocampus and visual cortex in learning, memory, spatial orientation and arousal. (1). The effects of age on learning and short-term memory observed in this study and also reported in other studies with the Fisher 344 rat [35] appear to be similar to the short-term memory deficits reported in other animals [23], including old monkeys [45], and in man [26, 27]. The results of this study and those reported by others tend to support short-term memory declines independent of learning deficits in senescent rats. However, the findings are not unequivocal. When young and slower-learning old age groups are equated for original learning, significant age differences in memory tend to decrease. In this study, the young 11-month group learned the task significantly faster than the mature 17-month group. The 29-month old group did not differ significantly in days to criterion of learning from the 17-month mature group. Consequently, differences in short-term memory between the mature and old group may not be related to different levels of original learning. However, it has been shown that when noxious stimuli are used even in a relatively simple learning and passive-avoidance short-term memory task, age differences may exist in electric shock thresholds, arousal and reaction time [29]. (2). In this study, pain-elicited aggression was used as a measure of behavioral arousal in response to noxious stimuli in the presence of another rat [41]. Compared to the 11 month young group, there was a highly significant decline in the 29 month old rats in the number of fights and in the duration of "posturing" in the test session. Significant age differences in pain-elicited aggression have been reported during development, but not during aging in the rat [46]. Various limbic system structures have been implicated in the regulation of pain-elicited aggression [47]. According to the intercorrelations of painelicited aggression with age and with the selected morphological variables in the brain, pain-elicited aggression was inversely and significantly correlated with age, positively and significantly with the loss of neurons, and negatively and significantly with the increase in intraneuronal lipofuscin in the hippocampus CA1 zone and visual area 17. Although the hippocampus may be involved in arousal, previous lesion studies have shown that such other limbic system structures as the septal area may play a more direct role in the regulation of pain-elicited aggression in the rat [48]. (3). From the significant age differences in neurons and lipofuscin among the 3 age groups, the rate of loss of neurons and the increases in lipofuscin cannot be determined directly. In the hippocampus CA1 zone, the decrease in neuronal cell population from 11 to 17 months was 7%, from 17 to 29 months it was 8% and from 11 to 29 months it was 14%. The increase in intraneuronal lipofuscin from 11 to 17 months was 100%, from 17 to 29 months it was 35%, and from 11 to 29 months it was 169%. In visual area 17, the decrease in neuronal cell population from 11 to 17 months was 8%, from 17 to 29 months it was 18%, and from 11 to 29 months it was 25%. The increase in intraneuronal lipofuscin from 11 to 17 months was 33%, from 17 to 29 months it was 25%, and from 11 to 29 months it was 67%. From the comparison of age differences in neurons and
158 lipofuscin among the 3 age groups, it is apparent that the loss of neurons from 11 to 17 and 29 months of age may occur at a significantly slower rate than the accumulation of intraneuronal lipofuscin. However, several older age groups, with a much larger number of subjects per age group are essential to determine whether the rate of loss of neurons is linear or exponential from the brain of the rat with increasing old age. In one previous study, changes in neuron number and size, glia number, and in cortical depth of the medial occipital cortex were compared in young, adult and so-called aging Long-Evans hooded rats [49]. However, since the oldest age group in the study was only 650 days and the maximum life span has not been established, the significant age differences in cell number and depth of the occipital cortex that were reported from 4l to 650 days can be attributed more to "development rather than to aging". Other studies have reported no cell loss from the cerebral cortex of 550 day old rats [19]. Some studies have reported loss of dendritic spines from neurons in the visual cortex of 29 month old rats [24]. As yet, it is apparent that definitive studies of cell loss from the cerebral cortex of very old or "senescent" rats of a specific strain remain to be determined. Also, previous studies have compared cell loss from the neocortex in Long-Evans hooded, the Sprague-Dawley or Charles River albino, or the Fisher 344 albino rat of middle-aged 22-26 months, rather than of 28-36 month "aged" rats [9]. In another study, cell loss from the total brain of the mouse was estimated across the entire life span. It was reported that the loss of neurons from the mouse brain during very old age or senescence is exponential, and that the exponential loss of neurons with age coincides with the age specific increase of mortality, as described by the Gompertz Plot [50]. (4). Regarding the possible role of intraneuronal lipofuscin accumulation in the loss of neurons, it is important to note that the age related decreases in neurons and increases of intraneuronal lipofuscin were significantly and inversely correlated in the hippocampus CAI zone and in visual area 17. However, two paradoxically opposing hypotheses have been proposed concerning the possible role of intraneuronal lipofuscin accumulation in the loss of neurons [9]. It has been proposed that the accumulation of intraneuronal lipofuscin may ultimately result in the loss of neurons by interfering with cell organdies. In contrast to this detrimental view, it has also been proposed that the accumulation of intraneuronal lipofuscin may be beneficial since it may inhibit the loss of neurons. Lipofuscin may provide "sinks" for active but damaging enzymes involved in the formation of lipofuscin from lysosomes, or it may immobilize toxic products of metabolism [51 ]. (5). According to the intercorrelations of age with the behavioral and morphological variables, it seems apparent that the statistical correlations of intraneuronal lipofuscin and loss of neurons with age in the CA1 zone and in area 17 were considerably greater than the correlations of 2 and 6 hour short-term memory and the pain-elicited aggression changes with age. It appears that the cellular changes in the brain as reflected in the morphological variables may be correlated more significantly with age than the behavioral variables. However, based on statistical and methodological considerations, it should be noted that the relatively small number of 5 rats in each of the three cross-sectional age groups, as well as the restricted range of age levels within each age group, make the inter-
159 pretation of non-significant behavioral and morphological age differences difficult in multivariate evaluations of aging [52]. (6). In earlier studies, learning was generally associated with the neocortex, whereas arousal, motivation and emotion were related to changes in the autonomic nervous system and the secretions of the endocrine organs. In many recent studies, learning, motivation and emotion have been "linked" as an integrated adaptive process through the limbic system, particularly the hippocampus [30, 33]. Consequently, the possible role of the hippocampus in learning, memory, arousal and spatial orientation has attracted the attention of an increasing number of investigators who have become concerned with the possible structural and functional changes in the hippocampus during aging. However, many more studies that have not been concerned with aging have provided considerable evidence on the possible role of the hippocampus in learning, memory, arousal and spatial orientation [31, 32]. Evidence on the role of the hippocampus in the regulation of behavior is mainly indirect. Anatomically, the hippocampus is not directly linked with primary sensory or motor areas of the central nervous system [33]. As a major part of the limbic system, the hippocampus efferent pathways may exert influence on behavioral attention and arousal through the hypothalamus, the autonomic nervous system, and the endocrine glands. It has been proposed that within the hippocampus, the CA1 zone may form the emotional motor outflow pathway and that this pathway may be involved in temporary storage of information [33]. If the CAI zone of the hippocampus does regulate states of attention, arousal or emotion, it may modify the motivational and/or reward properties of environmental stimuli and thus play an integral role in the "emotional modulation" or consolidation of recent into long term memory [30]. According to the most prevalent views, the hippocampus may play some "modulatory" role over learning, memory, and arousal by some control over correct response emission, or incorrect response suppression [33]. Until recently, it was also widely accepted that the greatest redundancy of Golgi type II neurons, and the most marked morphological and chemical "plasticity" of the brain occurred in the cerebral cortex [53]. However, some rat studies have shown that the effects of environmental complexity may be even greater on the morphology of the hippocampus than of the visual cortex [54, 55]. In both human and animal studies, a key question has been whether the hippocampus plays a fundamental role in the process of committing recent memory into permanent memory [33]. In animal studies, passive avoidance tests have been used to examine the effects of unilateral and bilateral hippocampal lesions on short-term retention [30]. According to one study, rats with lesions in the medial hippocampus tended to perseverate in approach responses in a straight-alley runway despite electric shock received earlier at the food cup [56]. In another study, rats with bilateral hippocampal lesions perseverated in approach responses to the goal box in a straight runway during extinction trials [57]. In the present study, the 29 month old rats approached the food cup faster than the middle age group at 2 hours, and they also approached the food cup to a greater extent at 6 hours despite an earlier electric shock at the food cup. Since the passive-avoidance deficits in hippocampectomized rats have been attributed to response perseveration [58], the results of this study may be interpreted to suggest that
160 the significant age differences in short-term passive-avoidance memory may be related at least in part to significant cell loss and/or increased lipofuscin accumulation in the hippocampus CA1 zone and visual cortex area 17 of the rat brain. This interpretation is consistent with recent findings that have shown retention or memory deficits by 34 m o n t h senescent rats on a complex spatial orientation test [37].
ACKNOWLEDGEMENTS This work was supported in part by NIH 1 R01 HD09942 and NIH RR00164. Sincere thanks are expressed to Drs. J. W. May, UNO, J. W. Hansche, Tulane for their assistance in this research, and to Drs. M. O. Nahmmacher and W. Hunn for technical support in the use of TAS, Leitz, Inc., Rockleigh, N. J.
REFERENCES 1 J. Botwinick and M. Storandt, Memory, Related Functions and Age, Charles C. Thomas, Springfield, Illinois, 1974. 2 D, Arenberg and E. A. Robertson-Tchabo, Learning and aging, in J. E. Birren and K. W. Schaie (eds.), Handbook o f the Psychology o f Aging, Van Nostrand-Reinhold, New York, 1977. 3 J. Botwinick, Intellectual abilities, in J. E. Birren and K. W. Schaie (eds.), Handbook o f the Psychology o f Aging, Van Nostrand-Reinhold, New York, 1977, pp. 580--605. 4 F. I. M. Craik, Age differences in human memory, in J. E. Birren and K. W. Schaie (eds.), Handbook o f the Psychology o f Aging, Van Nostrand-Reinhold, New York, 1977, pp. 384-420. 5 D. S, Woodruff, A physiological perspective of the psychology of aging, in D. S. Woodruff and J. E. Bitten (eds.), Aging: Scientific Perspectives and Social Issues, Van Nostrand, New York, 1975, pp. 179-200. 6 C. Eisdorfer and F. Wilkie, Stress, disease, aging and behavior, in J. E. Birren and K. W, Schaie (eds.), Handbook o f the Psychology o f Aging, Van Nostrand-Reinhold, New York, 1977. 7 V. V. Frolkis, Aging of the autonomic nervous system and trends in its age changes, in J. E. Birren and K. W. Schaie (eds.), Handbook o f the Psychology o f Aging, Van Nostrand-Reinhold, New York, 1977, pp. 177-189. 8 G. R. Marsh and L. W. Thompson, Psychophysiology of aging, in J. E. Bitten and K. W. Schaie (eds.), Handbook of the Psychology of Aging, Van Nostrand-Reinhold, New York, 1977, pp. 219-248. 9 K. R. Brizzee, Gross morphometdc analyses and quantitative histology of the aging brain, in J. M. Ordy and K. R. Brizzee (eds.), Neurobiology o f Aging, Plenum, New York, 1975, pp. 401-424. 10 B. E. Tomlinson, G. Blessed and M. Roth, Observations on the brains of non-demented old people, J. Neurol. Sci., 9 (1968) 331-356. 11 B. E. Tomlinson, G. Blessed and M. Roth, Observations on the brains of demented old people, J. Neurol. Sci., 11 (1970)205-242. 12 B. E. Tomlinson, Morphological changes and dementia in old age, in W. L. Smith and M. Kinsbourne (eds.), Aging and Dementia, Spectrum, New York, 1977, pp. 25-56. 13 G. Blessed, B. E. T0mfinson and M. Roth, The association between quantitative measures of dementia and of senile changes in the cerebral grey matter of elderly subjects, Br. J. Psychiatry, 114 (1968) 797-811. 14 M.W. Hooper and F. S. Vogel, The limbic system in Alzheimcr's disease, Am. J. Pathol., 85 (1976) 1-14. 15 G. Henderson, B. E. Tomlinson and D. Weightman, Cell counts in the human cerebral cortex using a traditional and an automatic method, J. Neurol. Sci., 25 (1975) 129-144.
161 16 H. Brody, An examination of cerebral cortex and brain stem aging, in R. D. Terry and S. Gershon (eds.), Aging: Neurobiology of Aging, Raven Press, New York, 1976, pp. 177-182. 17 B. W. TomIinson and G. Henderson, Some quantitative cerebral findings in normal and demented old people, in R. D. Terry and S. Gershon (eds.), Aging: Neurobiology of Aging, Raven Press, New York, 1976, pp. 183-204. 18 K. R. Brizzee, J. M. Ordy, J. Hansche and B. Kaack, Quantitative assessment of changes in neuron and glia cell packing density and lipofuscin accumulation with age in the cerebral cortex of a nonhuman primate (Macaca mulatta), in R~ D. Terry and S. Gershon (eds.), Aging: Neurobiology of Aging, Raven Press, New York, 1976, pp. 229-244. 19 K. R. Brizzee, Quantitative histological studies on aging changes in cerebral cortex of rhesus monkey and albino rat with notes on effects of prolonged low-dose ionizing irradiation in the rat, in D. H. Ford (ed.), Progr. Brain Res., 40 (1973) 141-160. 20 R. L. Friede, Topographic Brain Chemistry, Academic Press, New York, 1966. 21 K. R. Brizzee, J. M. Ordy and B. Kaack, Early appearance and regional differences in intraneuronal and extraneurnnal lipofuscin accumulation with age in the brain of a nonhuman primate, Z Gerontol., 29 (1974) 366-381. 22 K. R. Brizzee, P. A. Cancilla, N. Sherwood and P. S. Timiras, The amount and distribution of pigments in neurons and glia of the cerebral cortex. Autofluorescent and ultrastructural studies, or. Gerontol., 24 (1969) 127-135. 23 J. Walker and C. Hertzog, Aging, brain function and behavior, in D. S. Woodruff and J. E. Birren (eds.), Aging: Scientific Perspectives and Social Issues, Van Nostrand, New York, 1975, pp. 152-178. 24 M. L. Feldman, Aging changes in morphology of cortical dendrites, in R. D. Terry and S. Gershon (eds.), Aging: Neurobiology of Aging, Raven Press, New York, 1976, pp. 211-228. 25 P. W. Landfield, G. Rose, L. Sandles, T. C. Wohlstadter and G. Lynch, Patterns of astroglial hypertrophy and neuronal degeneration in the hippocampus of aged, memory-deficient rats, £ GerontoL, 32 (1977) 3-12. 26 J. Botwinick, Behavioral processes, in S. Gershon and A. Raskin (eds.), Aging: Genesis and Treat. ment of Psychological Disorders in the Elderly, Raven Press, New York, 1975, pp. 1-18. 27 D. A. Waish, Age differences in learning and memory, in D. S. Woodruff and J. E. Birren (eds.), Aging." Scientific Perspectives and Social lssues, Van Nostrand, New York, 1975, pp. 125-151. 28 M. F. Elias and P. K. Elias, Motivation and activity, in J. E. Birren and K. W. Schaie (eds.), Handbook of the Psychology of Aging, Van Nostrand-Reinhold, New York, 1977, pp. 357-383. 29 R. L. Sprott and K. Staveness, Avoidance learning, behavior genetics and aging: a critical review and comment on methodology, Exp. AgingRes., 1 (1975) 145-168. 30 R. L. lsaacson, Memory processes and the hippocampus, in R. L. Isaacson and K. H. Pribram (eds.), The Hippocampus, Vol. 2, Physiology and Behavior, Plenum, New York, 1975, pp. 3 1 3 - 3 3 7 . 31 R. L. lsaacson and K. H. Pribram (eds,), The Hippocampus, Vol. 2, Structure and Development, Plenum, New York, 1975. 32 R. L. Isaacson and K. H. Pribram (eds.), The Hippocampus, Vol.2, Physiology and Behavior, Plenum, New York, 1975. 33 S. D. lversen, Do hippocampal lesions produce amnesia in animals?, lnt. Rev. Neurobiol., 19 (1976) 1--49. 34 D. S. Olton, Spatial memory, ScL Am., 236 (1977) 82-98. 35 P. E. Gold and J. L. McGaugh, Changes in learning and memory during aging, in J. M. Ordy and K. R. Brizzee (eds.), Neurobiology of Aging, Plenum, New York, 1975, pp. 145-158. 36 J.M. Ordy, K. R. Brizzee, B. Kaaek and J. Hansche, Age differences in short-term memory and cell loss in the cortex of the rat, Gerontology, 24 (1978) 2 7 6 - 2 8 5 . 37 C. A. Barnes, Synaptic modifiability in middle-aged and senescent rats, Soc. Neurosci. Abstr., 3 (285) (1977). 38 M. Hasan and P. Glees, Ultrastructural age changes in hippocampal neurons, synapses and neuroglia, Exp. GerontoL, 8 (1973) 75-83. 39 W. Bondareff and Y. Geinisman, Loss of synapses in the dentate gyrus of the senescent rat, Am. £ Anat., 145 (1976) 129-136. 40 J. Chesky and M. Rockstein, Survival data for a colony of male Fischer rats, Gerontologist, 15 (1975) 29.
162 41 J. H. McLean, R. Kostrezwa and J. G. May, Behavioral and biochemical effect of neonatal treatment of rat with 6-hydroxydopa, PharmacoL Biochem. Behavior, 4 (1976) 601-607. 42 W. J. S. Krieg, Connections of the cerebral cortex. I. The albino rat. A topography of the cortical areas, J. Comp. Neurol., 84 (1946) 221-275. 43 W. J. Hansche and W. P. Dunlap, A Fortran IV program for computing all possible multiple regressions, Behavior Res, Meth. Instrum., 8 (1976) 384-393. 44 J. B. Angevine, Time of neuron origin in the hippocampal region: An autoradiographic study in the mouse, Exp. Neurol. (Suppl.), (1965) 2. 45 D. L. Medin, P. O'Neil, E. Smeltz and R. T. Davis, Age differences in retention of concurrent discrimination problems in monkeys, J. Gerontol., 28 (1973) 63-67. 46 R. E. Ulrich, Pain-aggression, in G. A. Kimble (ed.), Foundations of Conclitioning and Learning, Appleton Century Crofts, New York, 1967, pp. 600--622. 47 B.E. Eleftheriou (ed.), The Neurobiology of the Amygdala, Plenum, New York, 1972. 48 E. B. Wetzel, R. L. Conner and S. Levine, Septal lesions, Psychonomic Sci., 9 (1967) 133-134. 49 M. C. Diamond, R. E. Johnson and M. W. Gold, Changes in neuron number and size and glia number in the young, adult and aging rat medial occipital cortex, Behavioral Biol., 20 (1977) 409-418. 50 H. A. Johnson and S. Erner, Neuron survival in the aging mouse, Exp. Gerontol., 7 (1972) 111-117. 51 A. N. Siakotos and D. Armstrong, Age pigment: A biochemical indicator of intraeellular aging, in J. M. Ordy and K. R. Brizzee (eds.),Neurobiology o f Aging, Plenum, New York, 1975, pp. 369-400. 52 J. M. Ordy and O. A. Schjeide, Univariate and multivariate models for evaluating long-term changes in neurobiological development, maturity and aging, in D. H. Ford (ed.), Progr. Brain Res., 40 (1973) 25-52. 53 M. C. Diamond, Anatomical brain changes induced by environment, in L. Petrinovich and J. L. McGaugh (eds.), Knowing, Thinking and Believing, Plenum Press, New York, 1976, pp. 215-239. 54 R. N. Walsh, O. E. Budtz-Olsen, J. E. Penny and R. A. Cummins, The effects of environmental complexity on the histology of the rat hippocampus, J. Com. Neurol., 137 (1969) 361-368. 55 M. C. Diamond, R. E. Johnson and C. lngham, Morphological changes in the young, adult and aging rat cerebral cortex, hippocampus and dieneephalon, Behavioral BioL, 14 (1975) 163-174. 56 D. Kimura, Effects of selective hippocampal damage on avoidance behavior in the rat, Can. J. Psychol, 12 (1958) 213-218. 57 L. E. Jarmrd and R. L. lsaacson, Hippocampal ablation in rats: effects of intertrial interval, Nature, 207 (1965) 109-110. 58 D. P. Kimble, R. J. Kirkby and D. G~ Stein, Response perseveration interpretation of passiveavoidance deficits in hippoeampectomized rats, J. Comp. Physiol. Psychol., 61 (1966) 141-143,