Acetyl-1-carnitine 2: Effects on learning and memory performance of aged rats in simple and complex mazes

Neurobiologyof Aging,Vol. 11, pp. 499-506. Pergamon Press plc, 1990. Printed in the U.S.A.

0197-4580/90 $3.00 + .00

Acetyl-l-Carnitine 2: Effects on Learning and Memory Performance of Aged Rats in Simple and Complex Mazes C. A. B A R N E S , *1 A. L. M A R K O W S K A , t D. K. I N G R A M , : ~ H. K A M E T A N I , : ~ E. L. S P A N G L E R , ~ : V. J. L E M K E N ~ : A N D D. S. O L T O N t

*Department of Psychology, Campus Box 345, University of Colorado, Boulder, CO 80309 ¢'Neuromnemonics Laboratory, Department of Psychology, Johns Hopkins University, Baltimore, MD 21218 SMolecular Physiology and Genetics Section, Laboratory of Cellular and Molecular Biology Gerontology Research Center, National Institute on Aging Francis Scott Key Medical Center, Baltimore, MD 21214 R e c e i v e d 19 A p r i l 1989; A c c e p t e d 19 M a r c h 1990 BARNES, C. A., A. L. MARKOWSKA, D. K. INGRAM, H. KAMETANI, E. L. SPANGLER, V. J. LEMKEN AND D. S. OLTON. Acetyl-l-carnitine2: Effects on learningand memoryperformance of agedrats in simpleand complexmazes. NEUROBIOL AGING 11(5) 499-506, 1990.--Acetyl-l-camitine (AC) was administered via drinking water for six months to one group (OLD-AC) of male F-344 rats beginning at 16 months of age, while another group (OLD-CON) of rats was given water only during that period. The rats were maintained on this treatment throughout behavioral testing, which began at 22 months of age. Performance of the OLD-AC and OLD-CON rats was compared to that of young control (YG-CON) rats on the following set of tasks: spontaneous alternation in the arms of a T-maze, two-choice simultaneous discrimination in the stem of a T-maze, rewarded alternation in the arms of a T-maze, spatial discrimination and reversal on a circular platform, spatial working memory in the radial 8-arm maze, long-term memory in the 14-unit T-maze, and for preference of the light or dark chamber of a two-compartment box. AC improved the long-term memory performance in the split-stem T-maze and on the circular platform but had no discernable effects on performance of aged rats in the other tasks. Possible reasons for the selectivity of this agent's action on behavior are suggested. Acetyl-l-carnitine 14-Unit T-maze

Aging

Learning and memory

T-maze

LEARNING and memory processes do not remain static throughout the life span of mammals. The cognitive changes that occur in the elderly with diseases such as Alzheimer's or multi-infarct dementias [e.g., (1)] are quite pronounced; however, changes are also observed, in relatively subtle ways, in very healthy older humans [e.g., (11)] and other animals [e.g., (7)]. It is therefore of clinical interest that long-term administration of acetyl-l-carnitine (AC) (37) has been reported to have beneficial effects on cognitive and neural function in both aged human and nonhuman species (2, 9, 10, 12, 15, 18, 19, 35). As outlined in more detail in Markowska et al. (29), A C ' s influence on behavior in older organisms may act through neural mechanisms that include cholinomimetic activity (14, 20, 40), increased glucose utilization (15), or free radical scavenging action (16). The behavioral changes that have been observed in old rats treated with AC included improved performance in a simple two-choice discrimination task and a more complex temporal discrimination problem (18), and improved accuracy in a spatial version of the Morris water task (19). Because the rodent has been a particularly successful model of age-related learning and memory changes that can be found

Radial 8-arm maze

Circular platform

in normal aged humans (4, 25, 31), we have used old rats in the present studies to examine more extensively the ability of AC to attenuate cognitive impairments over a range of behavioral situations. As outlined in more detail in Markowska et al. (29), three laboratories participated in these studies (sites at the Gerontology Research Center, GRC; Johns Hopkins University, JHU; and the University of Colorado, UCO). The behavioral tasks selected for the present set of experiments were chosen from among those where cognitive declines with age were expected [as in the circular platform (3,5), the radial maze (6, 8, 13, 17, 26, 41), and the 14-unit T-maze (21, 24, 25, 30)]. Specifically, the tasks included here were spontaneous alternation in the arms of a T-maze, a two-choice simultaneous discrimination in the stem of a T-maze, a rewarded alternation in the arms of a T-maze, a spatial discrimination and reversal on a circular platform, a spatial working memory test in the radial 8-arm maze, and a long-term memory test in the 14-unit T-maze. In addition, the preference that each rat had for the light or dark chamber of a two-compartment box was measured.

1Requests for reprints should be addressed to Dr. C. A. Barnes, Department of Psychology, University of Arizona, Tucson, AZ 85721. 499

BARNES ET AL.

500

TABLE 1 NUMBER OF ANIMALS USED, AND THE ORDER IN WHICH THEY WERE TESTED FOR EACH TREATMENT GROUP AT EACH TESTING SITE Apparatus Testing Sites GRC

JHU

UCO

Animals

T-Maze

YG-CON OLD-CON OLD-AC YG-CON OLD-CON OLD-AC OLD-CON OLD-AC

8[#1] 8 13

Split-Stem T-Maze

Circular Platform

Radial Maze

14-Unit T-Maze

Light-Dark Compartments

8[#2] 8 13 8[#2] 9 14

8[#1] 12 14 14[#1] 13

8[#3] 9 14 13[#2] 12

Abbreviations: GRC=Gerontology Research Center; JHU=Johns Hopkins University; UCO= University of Colorado; OLD=Old Control Rats; YG-CON= Young Control Rats; OLD-AC =Old Acetyl-l-carnitine-Treated Rats; [#1]=First task on which rats were tested at a particular site; [#2] =Second task on which rats were tested at a particular site; [#3] =Third task on which rats were tested at a particular site.

METHOD

Animals All old rats were virgin male F-344s, obtained from the National Institute on Aging's colony at Harlan at 16 months of age [see Markowska et al. (29) for more detail]. Briefly, each of the three participant sites received 30 such animals; 15 were given 75 mg/kg acetyl-l-carnitine in their drinking water for 6 months, the other 15 were given plain distilled water before testing began. During this period all animals were maintained on full food. In some of the experiments young rats were also tested (at the GRC 3-month-old, JHU 4-month-old F-344s), and were obtained from Harlan approximately 2 weeks before the beginning of training. Table 1 shows the number of old (OLD-CON) and young (YG-CON) control animals as well as old drug (OLD-AC) animals that were tested on each behavioral task and the site where they were tested, for each apparatus. Table 1 also gives the order in which the tasks were administered at a given test site. For tasks requiring food motivation, each rat's food intake was restricted to bring the rat to 80% of its ad lib weight.

Apparatus Spontaneous alternation. The apparatus used for spontaneous alternation (36) was a T-maze made of transparent Plexiglas, with a stainless-steel grid floor (Fig. 1A). It had a central arm (65 × 15 × 15 cm) and two goal arms (22.5 × 15 × 15 cm). An identical black Plexiglas box (25 × 15 × 15 cm) equipped with a guillotine door was placed at the end of all three arms; these were used as both start and goal boxes. The T-maze was surrounded by plywood walls, painted gray, and was illuminated by a 15-W dim, red light which was located 30 cm above the start box. Split-stem T-maze. The apparatus (23) was constructed of wood and had a stem divided into two halves (see Fig. 1B). The starting platform (18 cm long) was separated from the stem of the T (68 × 18 cm) by a guillotine door. At the end of the stem near the arms, an opaque curtain, 14 cm high and 7 cm wide, was suspended from a frame. Rats easily passed underneath the curtain, but it was effective in preventing them from seeing which side of the stem was open or closed. A clear Plexiglas barrier (14 cm high and 8 cm wide) was placed behind one curtain to block access to the arm on the side of the stem. A food cup was located

1 cm from the end of each arm (36 × 9 cm). The maze was located in a small, moderately lit room with cues mounted on the walls. Circular platform. The apparatus (3) for this task was a wooden, white-painted circular platform 1.2 m in diameter, around the circumference of which were 18 equally spaced holes 9 cm in diameter (see Fig. 1C). The platform was elevated 90 cm above the floor. A dark escape box (45 × I 1 cm) was located underneath one of the holes to allow the rat to escape from the platform surface. The surface of the platform could be rotated on each trial for control of possible odor cues. A cylindrical starting chamber could be raised by a pulley system to a position 2 m above the maze surface. Three 250-W floodlights were directed towards the platform surface so that the rat would be highly motivated to find escape. This task was duplicated at JHU and UCO, using as identical apparati and procedures as possible. Radial 8-arm maze. A semiautomated, black Plexiglas 8-arm radial maze (33) had arms 58 cm by 5.5 cm radiating from a round 19.5-cm central platform (see Fig. 1D). The maze was elevated 79 cm above the floor. The arms were lined with 4-mm edges, and a small food cup was at the end of each arm. Halfway down the arms was a hinge that allowed the arms to be raised and lowered. The arm end with the food cup was held rigidly horizontal, while the end of the arm that joined the center was mobile. The rat could thus be confined to the central platform until the arms were raised to allow access to reward. The room was dimly lit and was surrounded by many distal objects that could serve as cues. 14-Unit T-maze. This apparatus (24,39) was constructed of translucent Plexiglas with a stainless steel grid floor that could deliver 0.8 mA scrambled, constant current footshock (see Fig. 1E). The start and goal boxes were interchangeable. Guillotine doors were used to divide the maze into five distinct segments of approximately equivalent area. Infrared photocells were located throughout the maze and were wired to a microprocessor that recorded movements of the rat through the maze. The maze was enclosed by four walls painted flat gray. Four speakers were located under corners of the maze and were wired to a white noise generator (40 dB). The maze was located in a large room containing overhead fluorescent lighting. To facilitate learning of the maze, each rat was pretrained in a straight runway to actively avoid shock. This runway was 2 m long and was constructed of translucent Plexiglas. The floor had

ACETYL-1-CARNITINE: LEARNING AND M E M O R Y

501

A

B

I

[o

t

PLEXIOLAS GOAL BOX

GOAL BOX

,~

PLEXlGL&S BARRIER ~ CURTAINS

~'1-~--"

!

.....

i%, ARM CHOICE POINT

!

POINT

~ CHOICE GUILLOTINE

RUN

DOOR -,9 __ PLEXlGLAS START BOX

P"'t

i FORCED RUN

oI

STEM CHOICE ~ GUILLOTINE DOOR

•

STARTING I~.ATFORM

E r [

FIG. 1. Schematic diagrams of five of the apparati used in the present experiments. (A) T-maze apparatus used for the spontaneous alternation task. (B) Split-stern T-maze apparatus. The left diagram shows the stem choice point and the arm choice point for the two types of discriminations measured. On the fight, the top and bottom diagrams show the configuration of the maze for the forced run and subsequent choice run, respectively. The dashed line shows the correct response. (C) Circular platform apparatus, with the filled circle indicating the correct location of the hole that resided over the escape box (indicated by the dashed rectangle). (D) Radial 8-arm maze, shown here with all 8 arms available to the rat. A food cup was located at the end of each arm. (E) 14-unit T-maze apparatus, with " S " indicating the location of the start box, and " G " the location of the goal box. The arrows indicate the correct path to the goal, and the lines in the alleyways indicate the location of guillotine doors that were closed to prevent animals from backtracking through the maze.

diagonal, stainless steel rods that delivered the same shock parameters described above. This apparatus also had exchangeable start and goal boxes and resided in the same large room as the 14-unit T-maze. Light-dark preference. Constructed of wood, the apparatus had two compartments (35 x 50 x 40 cm each). One chamber was white and lighted and the other was black and darkened. The lid of the darkened compartment was made of wood and painted black, while the lid of the lighted compartment was made of clear plastic.

Procedure Spontaneous alternation. Each rat was adapted to the maze room for 60 min before beginning the spontaneous alternation procedure. Two tests were given to each rat for 3 consecutive days. The interval between tests in each day was 4 hours. Each test had two trials. For the first of each test, a rat was placed into the start box, and the guillotine door by the start box was opened slowly. After the rat entered the goal box, the guillotine door by the goal box was closed; 60 sec later the rat was returned to its home cage. Each rat was given a maximum of 300 sec to make a choice. One hour after the first trial, the rat was again placed into the start box for a second trial. If the rat turned in the direction opposite to that observed on the first trial, the response was recorded as an alternation. Split-stem T-maze. Each rat was shaped for several days to run to the end of the arms for a food reward. During the first day, food was spread throughout the maze and each rat was allowed 10 minutes to explore. During each successive day, the amount of

food was progressively restricted so that by the end of shaping, food was present only in the food cups. During each choice trial, the rat was required to perform two discriminations, each of which placed different demands on memory (23). At the stem choice point, a two-choice simultaneous discrimination solution was required that measured trial-independent (long-term reference) memory, because the same side of the stem was correct for every trial. At the a n n choice point, a delayed alternation solution was required that measured trial-dependent (working) memory, because the a n n that was correct (and had food) changed from trial to trial. Each session was composed of one force trial, when the rat was forced to enter one ann, and 10 choice trials, when the rat could enter either arm. The two discriminations are described in detail below. For the stem discrimination, long-term reference memory was assessed by blocking one side of the stem throughout training. Consequently the rat was required to go down the correct open side to get to the arm choice point. A response was recorded for the stem discrimination when the rat placed its forepaws beyond the start of the stem partition. If the rat chose the open side of the stem, a correct stem response was recorded and the rat continued to the arms. If the rat chose the blocked side of the stem, an incorrect stem response was recorded; the rat was allowed to turn around and choose the open side. The side that was blocked remained constant for any particular rat during all trials but varied among rats. For the arm discrimination, working memory was assessed by requiring the rat to alternate body turns after an initial trial where

BARNES ET AL.

502

the turn was forced. For the first, force trial (Fig. IB), a piece of food was placed in the end of the one arm that was open, and the arm of the maze that did not contain the food was blocked. For all subsequent choice trials (see Fig. 1B), food was placed in the arm not entered during the previous trial. If the rat entered the arm with food, a correct arm response was recorded; the rat was allowed to eat the food and was returned to the start platform. If the rat entered the arm without food, an incorrect response was recorded, and the rat was returned to the starting position without receiving any food. This procedure was continued for 20 days. Circular plaO~orm. On the first day, each animal was initially placed next to the hole over the escape tunnel. Each rat went down into this hole immediately and was allowed to stay there for 1 minute. The rat was then returned to its holding cage while a randomly selected hole on the platform surface was positioned over the fixed goal location. For all subsequent trials the start chamber was lowered to the center of the platform, and the rat was placed into it facing a randomly selected direction. After 15 seconds, the start chamber was raised, the trial began, and the rat was free to explore the surface of the platform. A response was defined as the rat placing its nose in a hole. An error was a response to a hole not leading to the goal box. When the rat found the correct location of the goal box, it was allowed to enter it, remain there for 30 seconds, and was then removed. If the rat did not find the goal box within 4 minutes, it was placed next to the correct hole and allowed to descend into the box for 30 seconds. For Trials 1 though 16 (acquisition), the goal box remained in the same location in the room (trials were given once per day). For Trials 17-21 (probe trials) the goal box was moved 6 holes from the original location. Radial 8-arm maze. The procedure for this task was divided into three phases. In all phases, only one trial was given per day. At the beginning of each trial, each food cup was filled with chocolate milk. In Phase 1, each rat was given access to all arms simultaneously and was given a maximum of 10 minutes to find the reward on each arm. An error was scored if the rat entered an arm without milk. When a given rat's performance on a trial reached 3 errors or less, or the rat had been given 14 trials, Phase 2 began. The rat was initially forced to enter a randomly selected set of 4 baited arms. When the animal obtained food from each of these arms, it was confined for 10 seconds on the central platform. All 8 arms were then raised. The optimal strategy for the rat was to remember the 4 arms that had already been chosen and to enter the other four arms, which had food. When a given rat's performance on 2 or 3 consecutive days averaged 5 errors or less, or 20 trials had been given, Phase 3 began. The third part of the experiment involved extending the 10-second delay interval to 5 minutes, and this procedure continued for a 14-day period. Because of the number of hours per day required to test all of the old rats and the large amount of baseline data from this maze in young animals, no young rats were included in this portion of the study. 14-unit T-maze. Each rat received pretraining in a one-way active avoidance task in the straight runway before being tested in the 14-unit T-maze. The pretraining involved 10-second periods where the rat was required to move from the start box to the goal box at the end of the runway to avoid shock. Ten such trials were given daily with an intertrial interval of 2 min, for three consecutive days. Rats meeting a criterion of 8 correct avoidance responses in 10 consecutive trials on the third day of training were then transferred to training on the 14-unit T-maze. There were no differences in the number of OLD-CON and OLD-AC rats reaching this initial criterion. On the 14-unit T-maze, the rat was required to move through

TABLE 2 MEAN PERCENTALTERNATIONRESPONSEON THE T-MAZE Treatment Group

% Alternation __-SEM

Young Control Old Control Old Acetyl-l-carnitine

64.6 ___ 10.2 63.0 ± 7.8 42.2 __- 6.3

each segment within 10 seconds to avoid scrambled foot shock. To prevent retracing of paths towards the start box, the guillotine doors were lowered each time the rat moved into a new maze segment. When the rat entered the goal box it was confined there for 2 minutes before the next trial. During this interval, the maze was raised by a set of motorized pulleys, and the grid floor was mopped with an ethanol solution to mask possible odor cues. Ten trials were given per day on five days (with 2 days intervening between day 2 and day 3). Light-dark preference. For this test, the rat was placed into the white, lighted compartment of the two-compartment box. The latency to enter the black, darkened compartment, the total time spent in the black compartment, and the number of visits to the black compartment were recorded during 3 min. RESULTS

Spontaneous Alternation The mean and standard error of the mean (SEM) of the percent of trials that the animals alternated their responses can be seen in Table 2. One-way ANOVA revealed no significant difference among groups, F(2,27)= 2.47, p>0.5.

Split-Stem T-Maze In the stem discrimination, YG-CON rats had a mean choice accuracy of 60% during the first block of two test sessions. They steadily improved choice accuracy and took a mean of 8 (-+ 1) sessions to reach the criterion of 90% correct responses during a single day. OLD-CON rats began testing with a similar level of choice accuracy, but took longer to reach criterion, a mean of 13 (--_ 1) sessions. OLD-AC rats also began testing with a similar level of choice accuracy; they reached criterion more rapidly than OLD-CON rats, but not as rapidly as YG-CON rats, taking a mean of 10 (---1) sessions. A one-way ANOVA showed significant differences among groups [main effect of group, F(2,28)=3.5, p<0.05]. Post hoc analysis showed an aging effect (OLD-CON performed worse than YG-CON, p<0,03), but not an AC effect (OLD-AC did not perform better than OLD-CON, p>0.05). Both an aging effect and an AC effect, however, were apparent in the percentage of rats reaching criterion after 10 sessions: YG-CON, 88; OLD-CON, 33; OLD-AC, 79. Fisher's exact probability test showed that OLD-CON rats took more trials than YG-CON rats (p<0.04; an aging effect), and that OLD-AC rats took fewer trials than OLD-CON rats (p<0.05; an AC effect). For the arm discrimination, the two-way ANOVA on the number of correct responses (groups x blocks of 2 sessions) showed significant differences among groups [main effect of groups, F(2,26)= 14.6, p<0.0001], and a significant effect of trials (F = 2.5, p<0.01). Post hoc analysis showed an aging effect (YG-CON rats performed significantly better than both the OLDCON, p<0.001, and the OLD-AC, p<0.001 rats) but not an AC effect (see Fig. 2B).

ACETYL-I-CARNITINE: LEARNING AND M E M O R Y

503

STEM DISCRIMINATION

A

10

1110 if)

90

I

O

nrr uJ

Z

'~

8 6

Z <[

711

uJ

=S n,.

~-

o

611

o

a OLD-CON • OLD-AC

Z

<

511

I

1st

411

I

last

A []

YG- CON OLD-CON

•

OLD- AC

I

1st day

blocks of 3 trials 1

B

5

10

m w m z

~_ O w E nO

611 511 a, YG-CON cl OLD-CON • OLD-AC

Z

-,

:E

411

11

I

1

5

P R O B E TRIAL

FIG. 3. Results from the circular platform task (means of Trials 1-3, 14-16, and 17, and standard error of the means). All three groups of rats significantly improved their performance from the beginning of training (first 3 acquisition trials) to the end of training (last 3 acquisition trials). There was also an aging effect (YG-CON rats made fewer errors on the final acquisition trials than did either the OLD-CON or the OLD-AC rats), and an AC effect (OLD-AC rats made fewer errors on the final acquisition trials than did the OLD-CON rats). Furthermore, on the first probe trial, when the location of the goal was changed to a new place, both the YG-CON and the OLD-AC animals significantly increased the number of errors that they made, because they had searched the spatial location learned on the acquisition trials. The OLD-CON animals, on the other hand, did not show this behavior pattern.

ARM DISCRIMINATION 9O

w E

ACQUISITION

10

BLOCKS OF T W O S E S S I O N S FIG. 2. Results from the split-stem T-maze discrimination task (means in blocks of two sessions and standard error of the means). (A) For the stem discrimination, AC improved performance; more OLD-AC rats reached criterion in 10 test sessions than did OLD-CON rats. (B) For the arm discrimination component of the task, both the OLD-CON and OLD-AC rats performed significantly worse than the YG-CON rats.

did the OLD-CON rats on Trials 14-16, p < 0 . 0 4 5 4 ; and see Fig. 3). When the position of the goal tunnel was changed (probe trials) on Trial 17 to a location different from the original location to which the animals were trained, both the YG-CON and the OLD-AC rats made substantially more errors, as would be expected if they were searching in the vicinity of the initial goal location (see Fig. 3). The OLD-CON rats, however, did not increase the number of errors made on Trial 17. A one-way A N O V A on the ratio of errors made on Trials 17 versus Trials 14-16 showed a significant main effect of group, F ( 2 , 5 7 ) = 7 . 7 , p < 0 . 0 0 5 . Post hoc analyses showed an aging effect (OLD-CON rats did not increase their errors on Trial 17 as did the YG-CON rats, p < 0 . 0 0 5 ) and an AC effect (OLD-AC rats increased their errors on Trial 17 unlike the OLD-CON rats, p < 0 . 0 1 ) .

Circular Platform For the first several days of testing, the rats in each group made a large number of errors, ranging from about 3 to over 20 on a single trial. By the end of 16 trials, all groups made fewer errors, with some of the YG-CON rats showing perfect accuracy (no errors). A two-factor repeated measures A N O V A on the combined data from JHU and UCO for errors during the first and last block of 3 acquisition trials yielded a main effect of group, F ( 2 , 5 8 ) = 6.225, p = 0 . 0 0 3 6 , and a main effect of trials ( F = 4 1 . 7 0 5 , p < 0 . 0 0 0 1 ) . The post hoc tests revealed a significant effect of trials in each group of rats (YG-CON, p = 0 . 0 0 0 4 ; OLD-CON, p = 0.0138; OLD-AC, p < 0 . 0 0 0 1 ) , an aging effect. (YG-CON rats made fewer errors at the end of training than did either the OLD-CON, p = 0 . 0 0 0 5 , or the OLD-AC, p = 0 . 0 0 1 6 , rats) and an effect of AC (OLD-AC rats made significantly fewer errors than

Radial 8-Arm Maze In Phase 1 of this task, where all 8 arms were simultaneously available to the animal, the mean trials to criterion (see Fig. 4) for the OLD-CON rats was 14.4 (6.8 S.E.M.) and for the OLD-AC group was 14.8 (6.0 S.E.M.). In Phase 2, where a 10-second delay was imposed between the 4th and 5th choices for the animals, again, there was no AC effect in the mean trials to criterion (OLD-CON = 16.8 --- 5.0; OLD-AC = 17.6 --- 3.8). In Phase 3 (5-min delay), the most challenging part of the training schedule, the OLD-AC and OLD-CON rats still did not differ in their performance, with the main effect of group not reaching statistical significance, F ( 1 , 2 3 ) = 1.23, p = 0 . 2 7 9 , although there was a significant effect of trials ( F = 1.78, p = 0 . 0 4 6 ) , indicating that they had improved over time.

BARNES ET AL

504

NO DELAY PHASE1

10 SEC DELAY PHASE 2

25.

20

5 MIN DELAY PHASE 3

n, I-

.6

o_ 15

O 3¢ ¢J

"" ~,~\\~'"~\~; E ¢J

10

_u

0

OLD-CON

~.,

N

~ O O

OLD-AC

OLD-CON

OLD-AC

,x.\\\\'~ .0

~x -~\ -, \ ,,

OLD-CON

OLD-AC

FIG. 4. Results from the radial 8-arm maze task (means and standard error of the means for trials to criterion in Phase 1 and 2, and for correct choices final trial Phase 3). The delay between the 4th and 5th choice increased with testing: Phase 1, no delay; Phase 2, 10-sec delay; Phase 3, 5-min delay. AC did not improve choice accuracy in any part of this task.

14-Unit T-Maze

Light-Dark Preference

A 3 (group) by 5 (block) ANOVA performed on the number of errors over blocks of 6 trials showed a significant main effects of group, F(2,26)=47.73, p<0.0001, and of block (F=13.08; p<0.0001), with no significant interaction. Post hoc tests showed an effect of age (YG-CON group performed better than both the OLD-CON and OLD-AC groups, p<0.0001), but no significant AC effect (OLD-AC did not perform better than OLD-CON, p>0.05). As well as the number of errors (incorrect entries into alleyways), the % alternation errors (the % of incorrect versus correct alley entries that could be attributed to alternating body turns), the run time (total latency to complete the trial), the number of shock episodes per trial, and the total shock duration per trial yielded a similar pattern (see Fig. 5): the YG-CON rats showed statistically better performance scores on each of these measures compared with the OLD-CON and OLD-AC rats.

There was no significant difference between groups in the latency to first entry (p>0.05). However, there was a significant group effect in the measure of amount of time in the dark compartment, F(2,26)= 4.6, p<0.025. Post hoc analysis showed an aging effect (OLD-CON and OLD-AC spent less time in the dark than YG-CON, p < 0 . 0 2 and p < 0 . 0 1 , respectively) but not an AC effect (OLD-AC did not differ from OLD-CON, p>0.05). A one-way ANOVA was performed on the three measures of preference. As shown in Table 3, post hoc tests revealed no significant age effect in latency to enter the dark compartment; however, the YG-CON rats spent a greater amount of time in the dark when compared to both OLD-CON and OLD-AC groups. As a result, both of the OLD-CON and OLD-AC groups had at least one more entry into the dark compartment when compared to the YG-CON group. There was no significant AC effect on any of these measures.

301

U}

20

u) 50

250

40

20C

~ 30

~ 150

rOr nUJ

z

4

200

U)

3 O

~100

20

"'

2

100

v

Z

,<

IM

o "I"

so

:i

0' yG OLD OLD CON CON AC

0

1 ~ OI.D OI.D (~1C(~ AC

0

1

0 YG OLD OLD CON CON AC

OLD O U )

COIk CON AC

¥G C~J) (X.D

CON CON

AC

FIG. 5. Results from the 14-unit T-maze task (per trial means and standard error of the means). For all of the measures obtained from this task, the OLD-CON rats performed worse than the YG-CON rats, however, the OLD-AC rats did not perform better than the OLD-CON rats.

ACETYL-1-CARNITINE: LEARNING AND MEMORY

TABLE 3 PERFORMANCEMEASURESON LIGHT-DARKPREFERENCETASK Units of Measure (Mean - S.E.M.)

YG-CON

OLD-CON

OLD-AC

6.9 --+ 2.6

11.4 --+ 6.2

7.2 _ 2.0

Total time in dark (sec)

168.1 _.+ 4.7*

132.4 -+ 11.6

133.8 --+ 8.4

Number of entries into dark compartment

1.3 -+ 0.1"

Latency of first entry into dark (sec)

2.6 - 0,5

2.8 + 0.5

*p<0.05 as compared to both aged groups by two-tailed t-test.

DISCUSSION Acetyl-l-carnitine (AC) partially reduced the magnitude of the age-related impairment of performance in two tasks. In the circular platform task, old rats treated with AC (OLD-AC) showed greater accuracy in locating the spatial position of the hidden goal box by the end of acquisition than did old control (OLD-CON) rats. Furthermore, the results obtained on this task were similar at both of the testing sites that employed the circular platform (JHU and UCO). When the goal location was changed on Trial 17 to a new place (the probe trials), the OLD-AC rats, like the young control (YG-CON) rats, returned more often to the previously correct location of the goal than did the OLD-CON rats. This pattern of results indicates that the OLD-AC rats learned the discrimination better than did the OLD-CON rats, and showed the behavior pattern typical of young rats. It should be noted, however, that the OLD-AC rats did not improve to levels equivalent to that of the YG-CON rats in this task, but to a point intermediate between the young and old controls. The other main effect of AC on the old animals in these experiments included an improved performance in the two-choice discrimination in the stem of the T-maze, which was the trial-independent or longer-term reference memory component of the split-stem T-maze task. The OLD-AC rats reached criterion performance faster than did the OLD-CON rats, and their performance did not differ on this measure from the YG-CON rats. Taken together, these results suggest that AC tended to improve memory performance in tasks in which the event to be remembered was required to be retained across days. In the circular platform task, the spatial location of the escape box was the same for all acquisition trials. In the two-choice simultaneous discrimination task in the split-stem T-maze, the side of the maze that allowed access to reward was the same on all trials. This simple description of the results, however, does not account for the finding that the AC-treated animals did not perform better than did the OLD-CON rats on the 14-unit T-maze. This task is also a test of long-term reference memory. It is possible, therefore, that the strategies used by the rats for optimal performance in the enclosed 14-unit T-maze are different from those used in the open elevated split-stem T-maze and circular platform (25). AC may produce its effects via neural mechanisms underlying only the latter cognitive operations. Treatment with AC did not influence performance of the old rats in several other aspects of the behavioral test battery, Nonrewarded spontaneous alternation performance was not affected, nor was rewarded alternation performance on the split-stem T-maze. This latter finding, involving no effect of AC on the trial-dependent or working memory component in the T-maze, is consistent with a lack of effect in the spatial working memory

505

problem in the radial maze. This suggests that shorter-lasting memory processes may not be altered by treatment with this agent. Although there were no differences between groups in the latency to enter the dark compartment in the light-dark preference test, the young rats tended to remain in the dark longer than did either of the old groups of rats. This observation may be due to the unusually high incidence of visual pathology in all of the old rats in this experiment [see Markowska et al. (29)], as other studies have suggested that older rats actually show a greater preference for the dark than do young rats (22). The lack of an effect of AC on performance in this preference test, however, indicates that the improved performance of the old rats on the circular platform (that depends on avoidance of bright illumination for motivation) cannot be due to changes in preference for the dark goal box. Furthermore, the levels of performance of the OLD-CON rats on the circular platform and radial maze were equivalent to populations of old F-344 rats that have been tested on the same apparati, but have not had cataracts. The ability to reach similar performance levels in spite of reduced visual acuity is most likely due to the fact that although both tasks rely on distal visual cues for task solution, rats do not use detailed visual information to solve them. Instead they use broad form and intensity gradients to aid navigation. For these reasons it appears unlikely that the visual problems of the old animals would alter our main conclusions concerning the effects of AC on cognitive performance of the old rats. The pattern of results found here may have implications for the mode of action of AC. Because colchicine lesions of the granule cells in the hippocampus produce profound impairments on performance of the circular platform task (28), it is possible that AC may alter some aspect of hippocampal function in the old rats. The results from Ghirardi et al. (19), showing improved performance of AC-treated old rats on the Morris water task, are also consistent with the idea that AC may be acting to improve hippocampal function. For a number of reasons, it is not likely that the cholinergic projection to the hippocampus is responsible for AC's effect on behavior in this situation. First of all, lesions of the medial septal area or the nucleus basalis of Meynert impair the arm discrimination and not the stem discrimination on the split-stem T-maze (23), and the latter was the task affected by AC in the present study. Furthermore, behavior on the radial maze is severely impaired in rats when the cholinergic system is disconnected from the hippocampus [e.g., (27, 32, 34)] and scopolamine has been shown to impair performance of rats on the 14-unit T-maze (38). The performance of AC-treated animals was not affected on either the radial or 14-unit mazes in the present study. In summary, long-term administration of AC tended to improve behavioral performance on tasks in which longer-term memory processes were required. Furthermore it is likely that these behavioral changes were a result of AC's influence on cognitive rather than on performance factors, because AC does not act to improve motor performance of rats that were treated identically to those in the present study [as reported in Markowska et al. (29)], nor does AC change the preference of the old rats for light or dark. The present results are also consistent with reports of improved behavior of old rats treated with AC on a spatial reference memory task in a water pool escape apparatus (19), however, we did not replicate the finding of improved performance on the radial 8-arm maze using this agent (10). This apparent discrepancy may be due to the fact that extramaze cues were available for task solution in the present study, but absent in the latter. While more work will be required to identify the behavioral processes that benefit from AC treatment, they may fall into a category of mechanisms involved in longer-lasting information storage processes.

BARNES ET AL.

506

ACKNOWLEDGEMENTS We thank J. Kaufman and C. Elkins for maintaining the rats, C. Elkins for the radial maze behavioral training, G. Rao for graphics, and B.

Peterson for secretarial assistance. We also thank Z. Khachaturian for his efforts in initiating this project and M. T. Ramacci for helpful discussions concerning all phases of these experiments. Funding was provided by Sigma Tan.

REFERENCES 1. Albert, M. S.; Moss, M. B. Geriatric neuropsychology. New York: The Guilford Press; 1988. 2. Angelucci, L.; Ghirardi, O.; Patacchioli, F. R.; Ramacci, M. T. Disinhibition of the hypothalamo-pituitary-adrenocortical axis as a marker of brain aging in the rat: A model for the study of anti-aging agents. Clin. Neuropharmacol. 9(Suppl. 4):104-106; 1986. 3. Barnes, C. A. Memory deficits associated with senescence: A neurophysiological and behavioral study in the rat. J. Comp. Physiol. Psychol. 93:74-104; 1979. 4. Barnes, C. A. Aging and the physiology of spatial memory. Neurobiol. Aging 9:563-568; 1988. 5. Barnes, C. A.; McNaughton, B. L. An age comparison of the rates of acquisition and forgetting of spatial information in relation to longterm enhancement of hippocampal synapses. J. Behav. Neurosci. 99:1040-1048; 1985. 6. Barnes, C. A.; Nadel, L.; Honig, W. K. Spatial memory in senescent rats. Can. J. Psychol. 34:29-39; 1980. 7. Bartus, R. T.; Dean, R. L. Animal models for age-related memory disturbances. In: Coyle, J. T., ed. Animal models of dementia: A synaptic neurochemical perspective. New York: Alan R. Liss; 1987: 69-79. 8. Beatty, W. W.; Bierley, R. A.; Boy& J. G. Preservation of accurate spatial memory in aged rats. Neurobiol. Aging 6:219-225; 1985. 9. Bonavita, E. Study of the efficacy and tolerability of l-acetylcarnitine therapy in the senile brain. Int. J. Clin. Pharmacol. Ther. Toxicol. 24:511-516; 1986. 10. Capriolli, A.; Ghirardi, O.; Ramacci, M. T.; Angelucci, L. Effect of long-term acetyl-l-carnitine on the old rat performance in the eightarm maze. Psychopharmacology (Berlin) Suppl. to vol. 96; 1988. 11. Crook, T.; Bartus, R. T.; Ferris, S. H.; Whitehouse, P.; Cohen, G. D.; Gerson, S. Age-associated memory impairment: Proposed diagnostic criteria and measures of clinical change. Report of a National Institute of Mental Health Work Group. Dev. Neuropsychol. 2: 261-276; 1986. 12. Cucinotta, D.; Ventura, S.; Passeri, M.; Iannuccelli, M.; Orfalian, Z.; Senin, V.; Pametti, L.; Bonati, P. A. Clinical experience with acetyl-l-carnitine in the treatment of signs and symptoms of senile mental deterioration in the aged. Fifth Capo Boi Conference on Neuroscience 1987:24 (abstract). 13. de Toledo-Morrell, L.; Morrell, F.; Kessler, E. S.; Fleming, S. Spatial memory and hippocampal kindling are impaired as a function of aging. Neurology 31:101; 1981. 14. Falcbetto, S.; Kato, G.; Provini, L. The action of camitines on coxtical neurons. Can. J. Physiol. Pharmacol. 49:1-7; 1971. 15. Fariello, R. G.; Ferraro, T. N.; Golden, G. T.; DeMattei, M.; Phillips, T.; Shug, A. L. Hypotheses on the role of the carnitine system in normal and aging brain. In: Shagass, C., ed. Biological psychiatry. New York: Elsevier Science Pub.; 1986:1337-1339. 16. Fariello, R. G.; Ferraro, T. N.; Golden, G. T.; DeMattei, M. Systemic acetyl-l-carnitine elevates nigral levels of glutathione and GABA. Life Sci. 43:289-292; 1988. 17. Gallagher, M.; Bostock, E.; King, R. Effects of opiate antagonists on spatial memory in young and aged rats. Behav. Neural Biol. 44: 374-385; 1985. 18. Ghirardi, 0.; Milano, S.; Ramacci, M. T.; Angelucci, L. Effect of acetyl-l-carnitine chronic treatment on discrimination models in aged rats. Physiol. Behav. 44:769-773; 1988. 19. Ghirardi, O.; Milano, S.; Ramacci, M. T.; Angelucci, L. Acetyl1-carnitine and spatial learning in senescent rats. Prog. NeuroPsychopharmacol. Biol. Psychiatry; in press. 20. Gibson, G. E.; Shimada, M. Studies on the metabolic pathway of the acetyl group for acetylcholine synthesis. Biochem. Pharmacol. 29:167-174; 1980. 21. Goodrick, C. L. Learning, retention, and extinction of a complex

22.

23.

24.

25. 26.

27.

28.

29.

30. 31. 32. 33.

34.

35.

36.

37.

38.

39. 40. 41.

maze habit for mature-young and senescent Wistar albino rats. J. Gerontol. 23:298-304; 1968. Goodrick, C. L. Variables affecting free exploration responses of male and female Wistar rats as a function of age. Dev. Psychol. 4:440--446; 1971. Helpler, D. J.; Olton, D. S.; Wenk, G. L.; Coyle, J. T. Lesions in nucleus basalis magnocellularis and medial septal area of rats produce qualitatively similar memory impairments. J. Neurosci. 5:866-873; 1985. Ingram, D. K. Analysis of age-related impairments in learning and memory in rodent models. In: Olton, D. S.; Gamzu, E.; Corkin, S., eds. Memory dysfunctions: An integration of animal and human research from preclinical and clinical perspectives. Ann. NY Acad. Sci. 444:313-331; 1985. Ingram, D. K. Complex maze learning in rodents as a model of age-related memory impairment. Neurobiol. Aging 9:475--485; 1988. Ingram, D. K.; London, E.; Goodrick, C. L. Age and neurochemical correlates of radial maze performance in rats. Neurobiol. Aging 2:41--47; 1981. Jarrard, L. E. Selective hippocampal lesions: Differential effects on performance by rats of a spatial task with preoperative versus postoperative training. J. Comp. Physiol. Psychol. 93:119-127; 1978. McNanghton, B. L.; Barnes, C. A.; Meltzer, J.; Sutherland, R. J. Hippocampal granule cells are necessary for normal spatial learning but not for spatially-selective pyramidal cell discharge. Exp. Brain Res. 76:485--496; 1989. Markowska, A. L.; Ingram, D. K.; Barnes, C. A.; Spangler, E. L.; Lemken, V. J.; Kametani, H.; Yee, W.; Olton, D. S. Acetyll-carnitine 1: Effects on mortality, pathology and sensory motor performance in aging rats. Neurobiol. Aging 11:491-498; 1990. Michel, M. E.; Klein, A. W. Performance differences in a complex maze between young and aged rats. Age 1:13-16; 1978. Olton, D. S.; Markowska, A. L. Within-subjects, parametric manipulations to investigate aging. Neurobiol. Aging 9:469-474; 1988. Olton, D. S.; Papas, B. C. Spatial memory and hippocampal function. Neuropsychologia 17:669-682; 1979. Olton, D. S.; Samuelson, R. J. Remembrance of places passed: Spatial memory in rats. J. Exp. Psychol. [Anita. Behav.] 2:97-116; 1976. Olton, D. S.; Wenk, G. Dementia: Animal models of the cognitive impairments produced by degeneration of the basal forebrain cholinergic system. In: Meltzer, H. Y., ed. Psychopharmacology: The third generation of progress. New York: Raven Press; 1987:941-953. Ramacci, M. T.; Patacchioli, F. R.; Ghirardi, D.; Ferraris, L.; Angelucci, L. Experimental pharmacological aspects on brain aging. In: Lenzi, S.; Descovici, G. C., eds. Atberosclerosis and cardiovascular disease. London: MTP Press; 1987:478-480. Roberts, W. W.; Dember, W. N.; Brodwick, M. Alternation and exploration in rats with hippocampal lesions. J. Comp. Physiol. Psychol. 55:695-700; 1962. Shug, A. L.; Schmidt, M. J.; Golden, G. T.; Fariello, R. G. The distribution and role of carnitine in the mammalian brain. Life Sci. 31:2869-2874; 1982. Spangler, E.; Rigby, P.; Ingram, D. K. Scopolamine impairs learning performance of rats in a 14-unit T-maze. Pharmacol. Biochem. Behav. 25:673-679; 1986. Stone, C. The age factor in animal learning. I. Rats in the problem box and the maze. Genet. Psychol. Mono. 5:1-130; 1929. Tempesta, E.; Janiri, L.; Salera, P. The effects of microiontophoretically applied D-L, acetylcarnitine on single neurons in the rat brainstem. Neuropharmacology 21:1207-1210; 1982. Wallace, J. E.; Krauter, E. E.; Campbell, B. A. Animal models of declining memory in the aged: Short-term and spatial memory in the aged rat. J. Gerontol. 35:355-363; 1980.