- Email: [email protected]
Absence of learning and memory deficits in the vasopressin-deficient rat (Brattleboro strain)
Behavioural Brain Research, 6 (1982) 1 13
1
Elsevier Biomedical Press
Research Papers ABSENCE OF LEARNING AND MEMORY DEFICITS IN THE VASOPRESSIN-DEFICIENT RAT (BRATTLEBORO STRAIN)
ROBERT J. CAREY* and MYRON MILLER
Departments of Medicine and Psychology, Veterans Administration Medical Center and State University of New York, Upstate Medical Center, Syracuse, NY 13210 (U.S.A.) (Received October 15th, 1981) (Revised version received March 4th, 1982) (Accepted March 8th, 1982)
Key words." vasopressin - learning - memory - avoidance - Brattleboro rat
SUMMARY
Active and passive avoidance behavior was compared between vasopressin (ADH)-deficient rats of the Brattleboro strain and normal Long-Evans rats. In retention of a passive avoidance task across three levels of footshock (0.5, 1.0 and 1.5 mA) the ADH-deficient rats exhibited a slight but overall superior passive avoidance performance. In a one-way avoidance task, diabetes insipidus (D.I.) and normal rats had closely comparable performances; whereas, in a shuttle-box avoidance task the D.I. rats made significantly more avoidance responses than the Long-Evans rats. Flinch-jump testing indicated that the D.I. rats had a small but statistically significant lowered jump threshold. These findings, which add to an increasingly conflicting literature regarding the role of ADH in learning and memory processes, indicate that ADH per se is not critical for normal learning and memory. Possibly, variations in breeding procedures for the D.I. rat have resulted in variations in behavioral reactivity which interact with learning tasks to enhance or impair performance.
INTRODUCTION
A number of peptides have increasingly become recognized as having important influences on brain function. Vasopressin (ADH), a neuropeptide * Address for correspondence: VA Medical Center, 800 Irving Avenue, Syracuse, NY 13210 (U.S.A.).
0166-4328/82/0000-0000/$02.75 © 1982 Elsevier Biomedical Press
synthesized in hypothalamic magnacellular nuclei and stored in the neural lobe of the pituitary from which it is released to regulate water balance, has recently been implicated in several complex brain processes. Thus, it has been claimed that vasopressin acts on the brain as a neuromodulator and exerts an important influence on learning and memory capability [ 10-12]. A valuable animal model for studying the possible brain-behavior influences of vasopressin is the Brattleboro strain of the Long-Evans rat which, in the homozygous state, has a specific and complete hereditary deficiency in ability to synthesize ADH with resultant diabetes insipidus (D.I.) [7, 9]. Several studies have tested the learning capability of the homozygous D.I. Brattleboro rat with apparent contradictory outcomes. Celestian et al. [5] first reported that the D.I. rat not only maintained performance of a shuttlebox avoidance response as effectively as normal Long-Evans rats, but the D.I. rat exhibited an even stronger retention for this acquired response. This observation was further supported by Miller et al. [8] who observed that, although D.I. rats had a deficiency in avoidance acquisition, there was no defect in extinction performance. In contrast, Bohus et al. [3] studied D.I. rats trained in a two-day shuttlebox avoidance task and observed that the animals extinguished significantly faster than either heterozygous controls or normal animals of the Wistar strain. In addition, De Wied et al. [ 12] reported that homozygous D.I. rats have a pronounced retention deficit as compared to rats heterozygous for D.I. when tested by a passive avoidance task. Since these seemingly contradictory observations are based upon utilization of different test procedures, the present study was undertaken to re-examine the performance of the D.I. rat in a series of learning tasks which incorporated the testing paradigms used in previous studies. EXPERIMENT 1
The initial study compared the passive-avoidance performance of normal Long-Evans rats with that of rats of the Brattleboro strain homozygous for ADH deficiency. Method Animals All rats were bred and raised in our animal colony. The homozygous D.I. rats were the product of mating homozygous D.I. male rats with heterozygous females. The experimental animals were housed individually in metabolic cages with food and water provided ad libitum. Thirty-six normal (287 g, 60-80 days old) and 36 D.I. (266 g, 90-160 days old) rats were used. Verification of hormonal status was accomplished by measurement of urine osmolality by means of an Advanced Instruments Company freezing point depression osmometer (Model 3-D).
Apparatus A conventional passive-avoidance apparatus was used with both compartments made of plexiglass. The open platform was a translucent open enclosure 9 cm wide, 24 cm long, with 30 cm sides and was separated by a black plexiglass guillotine door from the completely enclosed black plexiglass chamber which was 30 cm long, 24 cm wide and 36 cm high. The black chamber had a grid floor of parallel bars 0.5 cm in diameter, spaced 1.5 cm apart. The grid floor was connected to an LVE DC constant current shocker with a scrambler.
Procedure Before the start of experimentation, all rats were handled for 5-10 min daily for 5 successive days. The normal and D.I. rats were each subdivided into 3 equal groups which were matched for body weight. One normal and one D.I. group was assigned to each of 3 shock levels: 0.5, 1.0 and 1.5 mA respectively. Passive avoidance testing was conducted over three successive days. The first day, each rat was placed on the translucent platform and after 5 sec the guillotine door to the black chamber was opened and the time recorded until the rat's 4 paws were inside the chamber at which time the guillotine door was gently lowered. On the second day, the rats were given 3 trials spaced approximately 5 min apart. Immediately after the rat entered the black chamber on the third trial, footshock was delivered for 5 sec. On the third day of testing, the rats were given two non-shock trials separated by approximately 30 min. If after 120 sec the rat failed to enter the chamber, it was removed from the translucent platform and returned to its home cage. Three days after completion of the testing, urine osmolalities were measured in each rat.
Results Urine osmolality measurement in the D.I. rats was indicative of complete A D H deficiency with a mean osmolality of 175.0 + S.E. 8.4 mOsm/kg; whereas the normal rats had a urine osmolality of 1950.7 + S.E. 40.6 mOsm/kg. Figure 1 presents the latency intervals for entry into the dark chamber before and after administration of the 3 levels of footshock. Prior to the footshock, the D.I. rats entered the dark chamber faster than the normals. After the shock, they generally took longer than the normals for each of the levels of shock intensity. The latency scores after the footshock were evaluated statistically by means of a two-way ANOVA. For both retention tests, shock level was a highly statistically significant variable with latency increasing as shock intensity increased (F = 21.6, df 2, 66, P < 0.01 and F = 18.2, df2, 66, P < 0.01, respectively). Also, the longer latencies of the D.I., as compared to the normal groups, approached statistical significance (F = 3.73, df 1, 66) for the first retention test
4
120
[ ] Normal [] Diabetes Insipidus
I,T
I00 8O c
o 60 03
4O 20 0
P r e - Shock
0.5
1.0 Foot Shock ( m A m p )
1.5
Fig. 1. Passive avoidance performance following prior exposure to increasing intensity offootshock. The latency interval (seconds) for movement of the rats from an illuminated chamber to a dark chamber is represented by the height of the vertical bars. Each pair of bars at each shock level indicates the values for the first and second retention tests. There is progressive lengthening of the latency interval with increase in shock intensity and the interval for D.I. rats is significantly greater than that for normal rats on the second retention test (P < 0.05, two-way ANOVA).
and reached statistical significance for the second test (F--4.1, df 1, 66, P < 0.05). EXPERIMENT 2
Since the rats used in the first experiment were higher in body weight than those used in the study of De Wied et al., a second experiment was undertaken to compare the passive avoidance of D.I. and normal Long-Evans rats at two levels of body weight.
Subjects Two sets of 16 male D.I. and 16 male normal Long-Evans rats were used. One set of 16 D T and normal rats were equated for 'light' (L) body weight (188 g, 55-60 days old, and 178 g, 75-80 days old, respectively), and the second set of 16 D.I. and normal rats were selected for equivalent 'heavy' (H) body weights (302 g, 90-95 days old, and 298 g, 150-270 days old, respectively).
Method The passive avoidance procedure used in Experiment 1 was employed except that the rats were only tested at the 0.5 mA and 1.0 mA shock intensity.
TABLE I
Mean pre- and post-shock latency scores (sec) on a passive avoidance task for D.I. and normal Long-Evans rats Separate groups of rats at two body weight levels (L = low and H = high) were tested at two shock intensities (0.5 and 1.0 mA).
Group
Body weight (g)
mA
D.I.
L 180
0.5
3.0
43.8
51.0
H 321 L 196
0.5 1.0
1.5 2.8
73.7 84.1
88.7 86.1
H 282
1.0
6.7
113.7
113.9
L 175 H 314
0.5 0.5
7.3 4.4
37.8 36.6
57.5 20.8
L 180 H 281
1.0 1.0
5.9 14.5
66.8 102.1
61.7 97.0
Normal
Pre-shock
Post-shock
Results
Table I shows the effect of body weight on the passive avoidance behavior of D.I. and normal rats. Although there were some overall mean differences among groups at a particular shock level, none of these differences reached statistical significance. The only variables that reliably (P < 0.05) affected performance were shock presentation and shock intensity. EXPERIMENT
3
In the present set of experiments the rats were given extensive handling before testing; whereas handling was not apparent from the report of De Wied et al. Accordingly, the next experiment was undertaken to compare passive avoidance learning in D.I. rats with and without pre-test handling. Animals Two sets of 6 male D.I. and 6 male normal Long-Evans rats matched for body weight (260.8 and 258.3 g, respectively) were used. Method
The same passive-avoidance procedure used in the previous two experiments was employed except that only 1.0 and 1.5 mA shock levels were used.
TABLE I1 Passive avoidance latem3' scores (sec) for handled (H) and not handled (NH) D.I. and normal Long-Evans rats tested at 1.0 and 1.5 mA shock intensities Group
Handling treatment
mA
Pre-shock
Post-shock
DT
H NH H NH
1.0 1.0 1.5 1.5
15.4 8.3 4.9 13.0
120.0 108.4 98.7 107.9
120.0 108.8 120.0 109.5
Normal
H NH H NH
1.0 1.0 1.5 1.5
21.3 21.0 9.0 6.9
101.6 107.4 42.5 103.3
85.4 120.0 b6.3 90.6
Results
Table II presents the results showing the effect of handling on the passive avoidance performance of D.I. and normal L o n g - E v a n s rats. For any specific shock intensity, the presence or absence of handling did not produce a statistically significant effect on passive avoidance performance. The only statistically significant (P < 0.05) effect was the presentation of footshock but not the intensity of the footshock. Although not statistically significant, it is interesting to note that, overall, the D.I. handled rats performed better than the D.I. non-handled rats; whereas the reverse was the case for the normal rats. EXPERIMENT 4 In an effort to assess the performance of D.I. rats in a different avoidance task, the next test compared D T versus normal rats in an active avoidance task. This was accomplished by basically reversing the passive test so that instead of having the rats receive shock on entering the dark chamber, the rats were required to leave the dark chamber to avoid shock. Method Animals
Twenty-four rats were used, 12 normal and 12 D.I. which were matched for body weight (D.I. 209 g, normal 244 g). Apparatus
The test apparatus was the same as used in Experiment 1.
Procedure
After one week of handling, the rats were placed into the dark chamber of the apparatus, the guillotine door was opened and the rats were allowed 1 min to leave the chamber or they were removed. On subsequent days, the rats were placed in the chamber and 10 sec later a 0.5 mA footshock was delivered for 5 sec or less if the rat left the compartment. The latency scores for leaving the dark chamber were recorded with an electronic timer. If the rat failed to leave the chamber during the shock, it was removed immediately at the termination of the shock. Latencies less than 10 sec were scored as avoidance responses. If an animal failed to leave the chamber during footshock, it received a latency score of 15 sec. A total of 9 trials were given with one trial per day. The one trial per day paradigm seemingly required retention of the shock delivery over days for avoidance learning to occur. Thus, in the temporal spacing of trials, the retention aspect of this study was comparable to Experiment 1. Results
Comparison was made of the total number of avoidance responses by each group for the 9 daily trials. The two groups proved to be virtually identical, with the normal group making 4.25 + 2.2 avoidance responses and the D.I. group making 4.16 + 1.4 avoidance responses. A statistical comparison using a t-test for independent measures showed that the differences were not statistically significant (t = 0.13). Also, rate of acquisition and latency scores for the two groups were similar. Table III presents the avoidance responses and latency scores for each group over trials. As is apparent in Table III, the acquisition rate and level were closely comparable for both groups. TABLE III Total active avoidance responses per day and mean response latencies per day for normal and D.I. rats (n = 12 per group) Days 1
2
3
4
5
6
7
8
9
5 7
6 7
6 7
8 7
7 11
9 9
9.1 9.6
8.2 7.5
8.4 7.2
6.7 7.2
(,4) Total avoidance responses per day
Normal D.I.
0 1
2 0
6 1
(B) Daily mean response latencies (sec)
Normal D.I.
14.4 14.6
13.4 14.3
9.0 13.4
6.4 5.0
4.0 6.4
EXPERIMENT 5
The next experiment was undertaken in order to compare the performance of the D.I. and normal rats on another conventional avoidance task, the two-way shuttle box avoidance.
Method Animals Ten naive rats were used, 5 normal (438 g) and 5 D.I. (376 g). Apparatus A two-compartment automated shuttle box (Lehigh Valley, LVE No. 3770) was used. The floor of each of the two compartments contained a grid, consisting of parallel bars 2 mm in diameter spaced 1.1 cm apart. The entire shuttle box was enclosed within a sound attenuating chamber, LVE No. 1488, ventilated by an electric fan. Rats were viewed through a one-way mirror. The source of footshock was a LVE No. 1531 constant current shocker with scrambler. Procedure Before the start of experimentation, all rats were handled for 5 min on each of 3 consecutive days. On the first and subsequent days of avoidance testing, the rats were allowed to adapt to the shuttle box for a 5 min period before the conditioning commenced. During this 5 min pretest adaptation period, measurements of each rat's spontaneous crossing rate were compiled. The conditioned stimulus (CS) was a light of 40 W, situated in the upper corner of the sound attenuating chamber and was presented for 5 sec. If the rat did not cross within this 5 sec CS period, a 0.5 mA electric footshock (unconditioned stimulus, UCS) was delivered through the appropriate grid floor until the rat crossed to the other side, or for a 10 sec period if it failed to cross. Ten trials were given per day with variable inter-trial intervals of from 25 to 80 sec duration separating the trials in a predetermined sequence in order to reduce the likelihood of temporal conditioning. A total of 90 trials were given over 9 days. Results The shuttlebox performances of the D.I. and normal rats are shown in Fig. 2. As is apparent in Fig. 2, the avoidance performance of the D.I. group was overall superior to that of the normal group. A statistical evaluation with a two-way ANOVA procedure indicated that the groups, trials and the interaction were statistically significant (F = 7.4, df 1, 8, P < 0.05; F = 9.6, df 8, 8, P < 0.05;
9
7 ,-6
~5
/. l
o
~ 4
.,-',
~3
/
Normo,
""
I ! ",.,-.-_i
//
jO
OSS 0
I
I
I
I
I
I
I
I
I
I
2
3
4
5
6
7
8
9
Number of Sessions
Fig. 2. Active avoidance response in a two-way shuttle box system. The avoidance performance of the D.I. group is superior to that of the normal group (P < 0.05, two-way ANOVA).
F = 2.4, df8, 64, P < 0.05 respectively. Also, D.I. rats had more inter-trial crossings, (D.I. 16.6, normal 4.2, P < 0.05). EXPERIMENT 6
Experiment 6 was undertaken to determine if the homozygous D.I. rats are more sensitive or reactive to footshock since such an increased reactivity might account for their enhanced performance on several of the avoidance tasks based on footshock.
Method Animals Twenty rats from the second and fifth experiments were used; 10 normal (422 g) and 10 homozygous for D.I. (379 g). Apparatus The apparatus used was an Evans Flinch-Jump Box, which consisted of a 8 x 14 in. box with a metal grid floor and clear plastic walls and top.
10 Procedure Each rat was placed in the testing apparatus and given 5 min to adapt to the box. First, the rat was given an ascending series (0.1 mA increments starting with a 0.1 mA) of footshocks until a flinch response (crouch-flinch-jerk) was detected. Then, the shock level was reduced (0.1 mA steps) until no response was observed. This procedure was repeated 8 times and the flinch threshold was defined as the mean of the shock levels for the 8 ascending responses and the 8 descending no responses. After completion of the flinch response testing, a similar procedure was followed for the jump response, which was defined as any rear paw lifted off the grid floor in response to shock. Results
The mean flinch thresholds for the normal and D.I. rats were 0.24 + 0.06 mA and 0.20 + 0.04 mA respectively. Statistical comparison of the mean difference by t-test for independent means did not achieve the P < 0.05 level (t = 1.99). For the jump response, the mean shock level for the normal rats was 0.60 + 0.08 mA and for the D.I. group 0.53 + 0.06 mA. Statistical comparison by t-test indicated that this difference was significant at the P < 0.05 level (t = 2.33). DISCUSSION
De Wied et al. [ 12] have reported that D.I. rats have a defect in the ability to acquire a passive avoidance behavior in response to footshock. This result, along with similar findings in rats rendered deficient in A D H by a variety of experimental approaches [ 10, 11 ], has led De Wied to propose that A D H plays a significant role in permitting normal learning and memory processes to occur. However, a mounting body of evidence based on studies of the D.I. rat has raised some question regarding the true physiological importance of A D H in learning and memory processes. Thus, the studies of Celestian et al. [5] and Miller et al. [8] which utilized two-way shuttle box active avoidance tasks, failed to disclose deficits in extinction performance following acquisition of an avoidance response although there was evidence suggestion of impaired capability in acquiring the avoidance response. Similarly, Bailey and Weiss [1, 2], using a passive avoidance procedure, found that the performance of homozygous D.I. rats was much better than that of normal Long-Evans rats. Recently, Brito et al. [4] evaluated the performance capability of D.I. rats in a variety of behavior tasks which included visual and olfactory discrimination learning associated with food motivation, a punishmentmotivated approach-avoidance conflict test and a test of ~timidity'. The D.I. rats were found to have deficits in tasks considered to be based on reference memory
11 (visual and olfactory discrimination learning) but were considered intact in working-memory processes (approach-avoidance conflict). These data indicate that while D.I. rats may differ from normal rats in their learning ability, not all aspects of learning are equally affected. When all of these studies are considered, it is evident that the vasopressin deficiency of the D.I. rat is not accompanied by a generalized impairment of learning and memory capacity. In the present study, D.I. rats in 3 different learning situations based on footshock and involving both active and passive avoidance responses performed as well as or superior to normal rats. These results, while in general agreement with the previous studies of Celestian et al. [5], Miller et al, [8] and Bailey and Weiss [ 1, 2], stand in sharp contrast to the report of De Wied et al. [ 12] that D.I. rats are unable to acquire a passive avoidance response. Since the general passive avoidance paradigm used in the present study and in the De Wied et al. study was similar, experimental variation in some critical test components seems unlikely to account for these discrepant observations. An important general consideration is that alteration of learning and memory processes is inferred from task performance and that task performance can be influenced by non-associative as well as associative processes. When a variable (in this instance ADH level) can seemingly impair or enhance task performance, some non-associative variable would appear to be implicated (i.e. motivational or emotionality factors). Significantly, there was a suggestion in the present study that D.I. rats had a lowered jump (but not flinch) threshold. Although the difference in threshold was slight in comparison with the range of intensities used in the passive avoidance task and would seemingly only enhance performance, the relevance of this threshold difference may be that it points to an altered reactivity to noxious stimulation in the D.I. rat. This is further supported by the observation of Bailey and Weiss [ 1] that, in open field tests, D.I. rats defecate more and are less active than normal Long-Evans rats, characteristics they considered as indicative of increased arousal or fearfulness. Thus, an increased emotionality factor could be exacerbated or attenuated with differences in experimental procedures (i.e. handling) and could interact with the associative aspects of the learning task to enhance or impair learning. While body weight and handling differences exist between the first experiment in the present study and the experiments of De Wied et al. [12], the additional findings in Experiments 2 and 3 indicate that these factors did not contribute importantly to the observation of normative passive avoidance learning by D.I. rats in the present report. Another source of experimental difference between the two studies is that De Wied et al. used heterozygous D.I. rats for the control group, whereas normal Long-Evans rats were used as a control in the present study. If heterozygous D.I. rats are superior to normal Long-Evans rats at learning a passive avoidance task, then this could account for different observations. If this proved to be the case, however, the issue would be to account for
12 the enhanced performance of the heterozygous D.I. rat rather than lbr a deficiency in the homozygous D.I. rat. Another possible important difference between the present report and that of De Wied et al. is in the data presentation. The present study expressed the results in terms of means, whereas De Wied et al. used medians. Since passive avoidance tasks tend to generate bimodal distributions, the use of medians runs the risk of having a single score produce a shift in modal frequency such that the difference between groups can appear as large as the intermodal difference. Since the mean weighs each score equally, such a shift in the distribution is minimized. A difficulty in studying the D.I. rat is that the body weight growth curve is stunted compared to normal rats [6]. Thus, matching D.I. and normal rats on the basis of body weight, as in the De Wied et al. study and in the present experiments, results in differences in age. Since the D.I. rat is lower in body weight than a normal rat for a given age, it is therefore possible that the D.I. rat might behave similarly to a normal rat raised on a restricted feeding regimen (e.g. one which would match the D.I. rat in both age and body weight). Since a reduction in body weight can alter behavioral reactivity, such a secondary effect of D.I. might account for the effects or behavioral performance observed in the present study. This interpretation changes the focus in the vasopressin-deficient rat to general systemic variables which might alter behavioral reactivity rather than to direct CNS effects of vasopressin which modify the processing and storage of information. It certainly would be useful to maintain D.I. rats with exogenouslyadministered ADH to equate growth curves and to then compare their performance both with and without ADH to age-matched normal Long-Evans rats. In a recent Symposium on the Brattleboro rat*, the point was made that, as a consequence of genetic drift, variation in breeding procedures and extent of inbreeding, D.I. animals from different colonies, while having similar deficiency in ADH synthesis, may be markedly different in other characteristics. Thus, observed discrepancies in such parameters as behavior or learning and memory ability may be more a consequence of genetic differences which have arisen among various colonies of D.I. rats than of hormonal environment or experimental variables. Regardless of which pertinent variable may account for the discrepant observations, the comparable, and sometimes superior, performance of the D.I. rat relative to the normal Long-Evans rat in several different types of learning situations in the present study, indicates that A D H does not appear to be necessary for at least some normal learning and memory processes to take place.
* International Symposium on the Brattleboro Rat, Dartmouth College, Hanover, NH, U.S.A., September 4-7th, 1981.
13 ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Ms. Melissa A. Bettinger and Mr. Michael A. Miller for the expert technical assistance which they provided. This work was supported by VA Medical Research Service Funds.
REFERENCES 1 Bailey, W.H. and Weiss, J.M., Effect of ACTH 4-10 on passive avoidance of rats lacking vasopressin (Brattleboro strain), Horm. Behav., 10 (1978) 22-29. 2 Bailey, W.H. and Weiss, J.M., Evaluation of a 'memory deficit' in vasopressin-deficient rats, Brain Res., 162 (1979) 174-178. 3 Bohus, B., Van Wimersma Greidanus, Tj. B. and De Wied, D., Behavioral and endocrine responses of rats with hereditary hypothalamic diabetes insipidus (Brattleboro strain), Physiol. Behav., 14 (1975) 609-615. 4 Brito, G.N.O., Thomas, G.J., Gingold, S.I. and Gash, D.M., Behavioral characteristics of vasopressin-deficient rats (Brattleboro strain), Brain Res. Bull., 6 (1981) 71-75. 5 Celestian, J.F., Carey, R.J. and Miller, M., Unimpaired maintenance of a conditioned avoidance response in the rat with diabetes insipidus, Physiol. Behav., 15 (1975) 707-711. 6 Dlouma, H., Krecek, J. and Zicha, J., Growth and urine osmolarity in young Brattleboro rats, J. Endocr., 75 (1977) 329-330. 7 Miller, M. and Moses, A.M., Radioimmunoassay of urinary antidiuretic hormone with application to study of the Brattleboro rat, Endocrinology, 88 (1969) 1389-1396. 8 Miller, M., Barranda, E.G., Dean, M.C. and Brush, F.R., Does the rat with hereditary hypothalamic diabetes insipidus have impaired avoidance learning and/or performance? Pharmacol. Biochem. Behav., 5 (1976) 35-40. 9 Valtin, H., Hereditary hypothalamic diabetes insipidus in rats (Brattleboro strain): a useful experimental model, Amer. J. Med., 42 (1967) 814-827. 10 De Wied, D., Effects ofpeptide hormones on behavior. In L. Martini and W.F. Ganong (Eds.), Frontiers in Neuroendocrinology, Oxford University Press, 1969, pp. 97-140. 11 De Wied, D., Van Delft, A.M.L., Gispen, W.H., Weijnen, J.A,W.M. and Van Wimersma Greidanus, Tj., The role of pituitary-adrenal system hormones in active avoidance conditioning. In S. Levine (Ed.), Hormones and Behavior, Academic Press, 1972, pp. 135-171. 12 De Wied, D., Bohus, B. and Van Wimersma Greidanus, Tj. B., Memory deficit in rats with hereditary diabetes insipidus, Brain Res., 85 (1975) 152-156.
1
Elsevier Biomedical Press
Research Papers ABSENCE OF LEARNING AND MEMORY DEFICITS IN THE VASOPRESSIN-DEFICIENT RAT (BRATTLEBORO STRAIN)
ROBERT J. CAREY* and MYRON MILLER
Departments of Medicine and Psychology, Veterans Administration Medical Center and State University of New York, Upstate Medical Center, Syracuse, NY 13210 (U.S.A.) (Received October 15th, 1981) (Revised version received March 4th, 1982) (Accepted March 8th, 1982)
Key words." vasopressin - learning - memory - avoidance - Brattleboro rat
SUMMARY
Active and passive avoidance behavior was compared between vasopressin (ADH)-deficient rats of the Brattleboro strain and normal Long-Evans rats. In retention of a passive avoidance task across three levels of footshock (0.5, 1.0 and 1.5 mA) the ADH-deficient rats exhibited a slight but overall superior passive avoidance performance. In a one-way avoidance task, diabetes insipidus (D.I.) and normal rats had closely comparable performances; whereas, in a shuttle-box avoidance task the D.I. rats made significantly more avoidance responses than the Long-Evans rats. Flinch-jump testing indicated that the D.I. rats had a small but statistically significant lowered jump threshold. These findings, which add to an increasingly conflicting literature regarding the role of ADH in learning and memory processes, indicate that ADH per se is not critical for normal learning and memory. Possibly, variations in breeding procedures for the D.I. rat have resulted in variations in behavioral reactivity which interact with learning tasks to enhance or impair performance.
INTRODUCTION
A number of peptides have increasingly become recognized as having important influences on brain function. Vasopressin (ADH), a neuropeptide * Address for correspondence: VA Medical Center, 800 Irving Avenue, Syracuse, NY 13210 (U.S.A.).
0166-4328/82/0000-0000/$02.75 © 1982 Elsevier Biomedical Press
synthesized in hypothalamic magnacellular nuclei and stored in the neural lobe of the pituitary from which it is released to regulate water balance, has recently been implicated in several complex brain processes. Thus, it has been claimed that vasopressin acts on the brain as a neuromodulator and exerts an important influence on learning and memory capability [ 10-12]. A valuable animal model for studying the possible brain-behavior influences of vasopressin is the Brattleboro strain of the Long-Evans rat which, in the homozygous state, has a specific and complete hereditary deficiency in ability to synthesize ADH with resultant diabetes insipidus (D.I.) [7, 9]. Several studies have tested the learning capability of the homozygous D.I. Brattleboro rat with apparent contradictory outcomes. Celestian et al. [5] first reported that the D.I. rat not only maintained performance of a shuttlebox avoidance response as effectively as normal Long-Evans rats, but the D.I. rat exhibited an even stronger retention for this acquired response. This observation was further supported by Miller et al. [8] who observed that, although D.I. rats had a deficiency in avoidance acquisition, there was no defect in extinction performance. In contrast, Bohus et al. [3] studied D.I. rats trained in a two-day shuttlebox avoidance task and observed that the animals extinguished significantly faster than either heterozygous controls or normal animals of the Wistar strain. In addition, De Wied et al. [ 12] reported that homozygous D.I. rats have a pronounced retention deficit as compared to rats heterozygous for D.I. when tested by a passive avoidance task. Since these seemingly contradictory observations are based upon utilization of different test procedures, the present study was undertaken to re-examine the performance of the D.I. rat in a series of learning tasks which incorporated the testing paradigms used in previous studies. EXPERIMENT 1
The initial study compared the passive-avoidance performance of normal Long-Evans rats with that of rats of the Brattleboro strain homozygous for ADH deficiency. Method Animals All rats were bred and raised in our animal colony. The homozygous D.I. rats were the product of mating homozygous D.I. male rats with heterozygous females. The experimental animals were housed individually in metabolic cages with food and water provided ad libitum. Thirty-six normal (287 g, 60-80 days old) and 36 D.I. (266 g, 90-160 days old) rats were used. Verification of hormonal status was accomplished by measurement of urine osmolality by means of an Advanced Instruments Company freezing point depression osmometer (Model 3-D).
Apparatus A conventional passive-avoidance apparatus was used with both compartments made of plexiglass. The open platform was a translucent open enclosure 9 cm wide, 24 cm long, with 30 cm sides and was separated by a black plexiglass guillotine door from the completely enclosed black plexiglass chamber which was 30 cm long, 24 cm wide and 36 cm high. The black chamber had a grid floor of parallel bars 0.5 cm in diameter, spaced 1.5 cm apart. The grid floor was connected to an LVE DC constant current shocker with a scrambler.
Procedure Before the start of experimentation, all rats were handled for 5-10 min daily for 5 successive days. The normal and D.I. rats were each subdivided into 3 equal groups which were matched for body weight. One normal and one D.I. group was assigned to each of 3 shock levels: 0.5, 1.0 and 1.5 mA respectively. Passive avoidance testing was conducted over three successive days. The first day, each rat was placed on the translucent platform and after 5 sec the guillotine door to the black chamber was opened and the time recorded until the rat's 4 paws were inside the chamber at which time the guillotine door was gently lowered. On the second day, the rats were given 3 trials spaced approximately 5 min apart. Immediately after the rat entered the black chamber on the third trial, footshock was delivered for 5 sec. On the third day of testing, the rats were given two non-shock trials separated by approximately 30 min. If after 120 sec the rat failed to enter the chamber, it was removed from the translucent platform and returned to its home cage. Three days after completion of the testing, urine osmolalities were measured in each rat.
Results Urine osmolality measurement in the D.I. rats was indicative of complete A D H deficiency with a mean osmolality of 175.0 + S.E. 8.4 mOsm/kg; whereas the normal rats had a urine osmolality of 1950.7 + S.E. 40.6 mOsm/kg. Figure 1 presents the latency intervals for entry into the dark chamber before and after administration of the 3 levels of footshock. Prior to the footshock, the D.I. rats entered the dark chamber faster than the normals. After the shock, they generally took longer than the normals for each of the levels of shock intensity. The latency scores after the footshock were evaluated statistically by means of a two-way ANOVA. For both retention tests, shock level was a highly statistically significant variable with latency increasing as shock intensity increased (F = 21.6, df 2, 66, P < 0.01 and F = 18.2, df2, 66, P < 0.01, respectively). Also, the longer latencies of the D.I., as compared to the normal groups, approached statistical significance (F = 3.73, df 1, 66) for the first retention test
4
120
[ ] Normal [] Diabetes Insipidus
I,T
I00 8O c
o 60 03
4O 20 0
P r e - Shock
0.5
1.0 Foot Shock ( m A m p )
1.5
Fig. 1. Passive avoidance performance following prior exposure to increasing intensity offootshock. The latency interval (seconds) for movement of the rats from an illuminated chamber to a dark chamber is represented by the height of the vertical bars. Each pair of bars at each shock level indicates the values for the first and second retention tests. There is progressive lengthening of the latency interval with increase in shock intensity and the interval for D.I. rats is significantly greater than that for normal rats on the second retention test (P < 0.05, two-way ANOVA).
and reached statistical significance for the second test (F--4.1, df 1, 66, P < 0.05). EXPERIMENT 2
Since the rats used in the first experiment were higher in body weight than those used in the study of De Wied et al., a second experiment was undertaken to compare the passive avoidance of D.I. and normal Long-Evans rats at two levels of body weight.
Subjects Two sets of 16 male D.I. and 16 male normal Long-Evans rats were used. One set of 16 D T and normal rats were equated for 'light' (L) body weight (188 g, 55-60 days old, and 178 g, 75-80 days old, respectively), and the second set of 16 D.I. and normal rats were selected for equivalent 'heavy' (H) body weights (302 g, 90-95 days old, and 298 g, 150-270 days old, respectively).
Method The passive avoidance procedure used in Experiment 1 was employed except that the rats were only tested at the 0.5 mA and 1.0 mA shock intensity.
TABLE I
Mean pre- and post-shock latency scores (sec) on a passive avoidance task for D.I. and normal Long-Evans rats Separate groups of rats at two body weight levels (L = low and H = high) were tested at two shock intensities (0.5 and 1.0 mA).
Group
Body weight (g)
mA
D.I.
L 180
0.5
3.0
43.8
51.0
H 321 L 196
0.5 1.0
1.5 2.8
73.7 84.1
88.7 86.1
H 282
1.0
6.7
113.7
113.9
L 175 H 314
0.5 0.5
7.3 4.4
37.8 36.6
57.5 20.8
L 180 H 281
1.0 1.0
5.9 14.5
66.8 102.1
61.7 97.0
Normal
Pre-shock
Post-shock
Results
Table I shows the effect of body weight on the passive avoidance behavior of D.I. and normal rats. Although there were some overall mean differences among groups at a particular shock level, none of these differences reached statistical significance. The only variables that reliably (P < 0.05) affected performance were shock presentation and shock intensity. EXPERIMENT
3
In the present set of experiments the rats were given extensive handling before testing; whereas handling was not apparent from the report of De Wied et al. Accordingly, the next experiment was undertaken to compare passive avoidance learning in D.I. rats with and without pre-test handling. Animals Two sets of 6 male D.I. and 6 male normal Long-Evans rats matched for body weight (260.8 and 258.3 g, respectively) were used. Method
The same passive-avoidance procedure used in the previous two experiments was employed except that only 1.0 and 1.5 mA shock levels were used.
TABLE I1 Passive avoidance latem3' scores (sec) for handled (H) and not handled (NH) D.I. and normal Long-Evans rats tested at 1.0 and 1.5 mA shock intensities Group
Handling treatment
mA
Pre-shock
Post-shock
DT
H NH H NH
1.0 1.0 1.5 1.5
15.4 8.3 4.9 13.0
120.0 108.4 98.7 107.9
120.0 108.8 120.0 109.5
Normal
H NH H NH
1.0 1.0 1.5 1.5
21.3 21.0 9.0 6.9
101.6 107.4 42.5 103.3
85.4 120.0 b6.3 90.6
Results
Table II presents the results showing the effect of handling on the passive avoidance performance of D.I. and normal L o n g - E v a n s rats. For any specific shock intensity, the presence or absence of handling did not produce a statistically significant effect on passive avoidance performance. The only statistically significant (P < 0.05) effect was the presentation of footshock but not the intensity of the footshock. Although not statistically significant, it is interesting to note that, overall, the D.I. handled rats performed better than the D.I. non-handled rats; whereas the reverse was the case for the normal rats. EXPERIMENT 4 In an effort to assess the performance of D.I. rats in a different avoidance task, the next test compared D T versus normal rats in an active avoidance task. This was accomplished by basically reversing the passive test so that instead of having the rats receive shock on entering the dark chamber, the rats were required to leave the dark chamber to avoid shock. Method Animals
Twenty-four rats were used, 12 normal and 12 D.I. which were matched for body weight (D.I. 209 g, normal 244 g). Apparatus
The test apparatus was the same as used in Experiment 1.
Procedure
After one week of handling, the rats were placed into the dark chamber of the apparatus, the guillotine door was opened and the rats were allowed 1 min to leave the chamber or they were removed. On subsequent days, the rats were placed in the chamber and 10 sec later a 0.5 mA footshock was delivered for 5 sec or less if the rat left the compartment. The latency scores for leaving the dark chamber were recorded with an electronic timer. If the rat failed to leave the chamber during the shock, it was removed immediately at the termination of the shock. Latencies less than 10 sec were scored as avoidance responses. If an animal failed to leave the chamber during footshock, it received a latency score of 15 sec. A total of 9 trials were given with one trial per day. The one trial per day paradigm seemingly required retention of the shock delivery over days for avoidance learning to occur. Thus, in the temporal spacing of trials, the retention aspect of this study was comparable to Experiment 1. Results
Comparison was made of the total number of avoidance responses by each group for the 9 daily trials. The two groups proved to be virtually identical, with the normal group making 4.25 + 2.2 avoidance responses and the D.I. group making 4.16 + 1.4 avoidance responses. A statistical comparison using a t-test for independent measures showed that the differences were not statistically significant (t = 0.13). Also, rate of acquisition and latency scores for the two groups were similar. Table III presents the avoidance responses and latency scores for each group over trials. As is apparent in Table III, the acquisition rate and level were closely comparable for both groups. TABLE III Total active avoidance responses per day and mean response latencies per day for normal and D.I. rats (n = 12 per group) Days 1
2
3
4
5
6
7
8
9
5 7
6 7
6 7
8 7
7 11
9 9
9.1 9.6
8.2 7.5
8.4 7.2
6.7 7.2
(,4) Total avoidance responses per day
Normal D.I.
0 1
2 0
6 1
(B) Daily mean response latencies (sec)
Normal D.I.
14.4 14.6
13.4 14.3
9.0 13.4
6.4 5.0
4.0 6.4
EXPERIMENT 5
The next experiment was undertaken in order to compare the performance of the D.I. and normal rats on another conventional avoidance task, the two-way shuttle box avoidance.
Method Animals Ten naive rats were used, 5 normal (438 g) and 5 D.I. (376 g). Apparatus A two-compartment automated shuttle box (Lehigh Valley, LVE No. 3770) was used. The floor of each of the two compartments contained a grid, consisting of parallel bars 2 mm in diameter spaced 1.1 cm apart. The entire shuttle box was enclosed within a sound attenuating chamber, LVE No. 1488, ventilated by an electric fan. Rats were viewed through a one-way mirror. The source of footshock was a LVE No. 1531 constant current shocker with scrambler. Procedure Before the start of experimentation, all rats were handled for 5 min on each of 3 consecutive days. On the first and subsequent days of avoidance testing, the rats were allowed to adapt to the shuttle box for a 5 min period before the conditioning commenced. During this 5 min pretest adaptation period, measurements of each rat's spontaneous crossing rate were compiled. The conditioned stimulus (CS) was a light of 40 W, situated in the upper corner of the sound attenuating chamber and was presented for 5 sec. If the rat did not cross within this 5 sec CS period, a 0.5 mA electric footshock (unconditioned stimulus, UCS) was delivered through the appropriate grid floor until the rat crossed to the other side, or for a 10 sec period if it failed to cross. Ten trials were given per day with variable inter-trial intervals of from 25 to 80 sec duration separating the trials in a predetermined sequence in order to reduce the likelihood of temporal conditioning. A total of 90 trials were given over 9 days. Results The shuttlebox performances of the D.I. and normal rats are shown in Fig. 2. As is apparent in Fig. 2, the avoidance performance of the D.I. group was overall superior to that of the normal group. A statistical evaluation with a two-way ANOVA procedure indicated that the groups, trials and the interaction were statistically significant (F = 7.4, df 1, 8, P < 0.05; F = 9.6, df 8, 8, P < 0.05;
9
7 ,-6
~5
/. l
o
~ 4
.,-',
~3
/
Normo,
""
I ! ",.,-.-_i
//
jO
OSS 0
I
I
I
I
I
I
I
I
I
I
2
3
4
5
6
7
8
9
Number of Sessions
Fig. 2. Active avoidance response in a two-way shuttle box system. The avoidance performance of the D.I. group is superior to that of the normal group (P < 0.05, two-way ANOVA).
F = 2.4, df8, 64, P < 0.05 respectively. Also, D.I. rats had more inter-trial crossings, (D.I. 16.6, normal 4.2, P < 0.05). EXPERIMENT 6
Experiment 6 was undertaken to determine if the homozygous D.I. rats are more sensitive or reactive to footshock since such an increased reactivity might account for their enhanced performance on several of the avoidance tasks based on footshock.
Method Animals Twenty rats from the second and fifth experiments were used; 10 normal (422 g) and 10 homozygous for D.I. (379 g). Apparatus The apparatus used was an Evans Flinch-Jump Box, which consisted of a 8 x 14 in. box with a metal grid floor and clear plastic walls and top.
10 Procedure Each rat was placed in the testing apparatus and given 5 min to adapt to the box. First, the rat was given an ascending series (0.1 mA increments starting with a 0.1 mA) of footshocks until a flinch response (crouch-flinch-jerk) was detected. Then, the shock level was reduced (0.1 mA steps) until no response was observed. This procedure was repeated 8 times and the flinch threshold was defined as the mean of the shock levels for the 8 ascending responses and the 8 descending no responses. After completion of the flinch response testing, a similar procedure was followed for the jump response, which was defined as any rear paw lifted off the grid floor in response to shock. Results
The mean flinch thresholds for the normal and D.I. rats were 0.24 + 0.06 mA and 0.20 + 0.04 mA respectively. Statistical comparison of the mean difference by t-test for independent means did not achieve the P < 0.05 level (t = 1.99). For the jump response, the mean shock level for the normal rats was 0.60 + 0.08 mA and for the D.I. group 0.53 + 0.06 mA. Statistical comparison by t-test indicated that this difference was significant at the P < 0.05 level (t = 2.33). DISCUSSION
De Wied et al. [ 12] have reported that D.I. rats have a defect in the ability to acquire a passive avoidance behavior in response to footshock. This result, along with similar findings in rats rendered deficient in A D H by a variety of experimental approaches [ 10, 11 ], has led De Wied to propose that A D H plays a significant role in permitting normal learning and memory processes to occur. However, a mounting body of evidence based on studies of the D.I. rat has raised some question regarding the true physiological importance of A D H in learning and memory processes. Thus, the studies of Celestian et al. [5] and Miller et al. [8] which utilized two-way shuttle box active avoidance tasks, failed to disclose deficits in extinction performance following acquisition of an avoidance response although there was evidence suggestion of impaired capability in acquiring the avoidance response. Similarly, Bailey and Weiss [1, 2], using a passive avoidance procedure, found that the performance of homozygous D.I. rats was much better than that of normal Long-Evans rats. Recently, Brito et al. [4] evaluated the performance capability of D.I. rats in a variety of behavior tasks which included visual and olfactory discrimination learning associated with food motivation, a punishmentmotivated approach-avoidance conflict test and a test of ~timidity'. The D.I. rats were found to have deficits in tasks considered to be based on reference memory
11 (visual and olfactory discrimination learning) but were considered intact in working-memory processes (approach-avoidance conflict). These data indicate that while D.I. rats may differ from normal rats in their learning ability, not all aspects of learning are equally affected. When all of these studies are considered, it is evident that the vasopressin deficiency of the D.I. rat is not accompanied by a generalized impairment of learning and memory capacity. In the present study, D.I. rats in 3 different learning situations based on footshock and involving both active and passive avoidance responses performed as well as or superior to normal rats. These results, while in general agreement with the previous studies of Celestian et al. [5], Miller et al, [8] and Bailey and Weiss [ 1, 2], stand in sharp contrast to the report of De Wied et al. [ 12] that D.I. rats are unable to acquire a passive avoidance response. Since the general passive avoidance paradigm used in the present study and in the De Wied et al. study was similar, experimental variation in some critical test components seems unlikely to account for these discrepant observations. An important general consideration is that alteration of learning and memory processes is inferred from task performance and that task performance can be influenced by non-associative as well as associative processes. When a variable (in this instance ADH level) can seemingly impair or enhance task performance, some non-associative variable would appear to be implicated (i.e. motivational or emotionality factors). Significantly, there was a suggestion in the present study that D.I. rats had a lowered jump (but not flinch) threshold. Although the difference in threshold was slight in comparison with the range of intensities used in the passive avoidance task and would seemingly only enhance performance, the relevance of this threshold difference may be that it points to an altered reactivity to noxious stimulation in the D.I. rat. This is further supported by the observation of Bailey and Weiss [ 1] that, in open field tests, D.I. rats defecate more and are less active than normal Long-Evans rats, characteristics they considered as indicative of increased arousal or fearfulness. Thus, an increased emotionality factor could be exacerbated or attenuated with differences in experimental procedures (i.e. handling) and could interact with the associative aspects of the learning task to enhance or impair learning. While body weight and handling differences exist between the first experiment in the present study and the experiments of De Wied et al. [12], the additional findings in Experiments 2 and 3 indicate that these factors did not contribute importantly to the observation of normative passive avoidance learning by D.I. rats in the present report. Another source of experimental difference between the two studies is that De Wied et al. used heterozygous D.I. rats for the control group, whereas normal Long-Evans rats were used as a control in the present study. If heterozygous D.I. rats are superior to normal Long-Evans rats at learning a passive avoidance task, then this could account for different observations. If this proved to be the case, however, the issue would be to account for
12 the enhanced performance of the heterozygous D.I. rat rather than lbr a deficiency in the homozygous D.I. rat. Another possible important difference between the present report and that of De Wied et al. is in the data presentation. The present study expressed the results in terms of means, whereas De Wied et al. used medians. Since passive avoidance tasks tend to generate bimodal distributions, the use of medians runs the risk of having a single score produce a shift in modal frequency such that the difference between groups can appear as large as the intermodal difference. Since the mean weighs each score equally, such a shift in the distribution is minimized. A difficulty in studying the D.I. rat is that the body weight growth curve is stunted compared to normal rats [6]. Thus, matching D.I. and normal rats on the basis of body weight, as in the De Wied et al. study and in the present experiments, results in differences in age. Since the D.I. rat is lower in body weight than a normal rat for a given age, it is therefore possible that the D.I. rat might behave similarly to a normal rat raised on a restricted feeding regimen (e.g. one which would match the D.I. rat in both age and body weight). Since a reduction in body weight can alter behavioral reactivity, such a secondary effect of D.I. might account for the effects or behavioral performance observed in the present study. This interpretation changes the focus in the vasopressin-deficient rat to general systemic variables which might alter behavioral reactivity rather than to direct CNS effects of vasopressin which modify the processing and storage of information. It certainly would be useful to maintain D.I. rats with exogenouslyadministered ADH to equate growth curves and to then compare their performance both with and without ADH to age-matched normal Long-Evans rats. In a recent Symposium on the Brattleboro rat*, the point was made that, as a consequence of genetic drift, variation in breeding procedures and extent of inbreeding, D.I. animals from different colonies, while having similar deficiency in ADH synthesis, may be markedly different in other characteristics. Thus, observed discrepancies in such parameters as behavior or learning and memory ability may be more a consequence of genetic differences which have arisen among various colonies of D.I. rats than of hormonal environment or experimental variables. Regardless of which pertinent variable may account for the discrepant observations, the comparable, and sometimes superior, performance of the D.I. rat relative to the normal Long-Evans rat in several different types of learning situations in the present study, indicates that A D H does not appear to be necessary for at least some normal learning and memory processes to take place.
* International Symposium on the Brattleboro Rat, Dartmouth College, Hanover, NH, U.S.A., September 4-7th, 1981.
13 ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Ms. Melissa A. Bettinger and Mr. Michael A. Miller for the expert technical assistance which they provided. This work was supported by VA Medical Research Service Funds.
REFERENCES 1 Bailey, W.H. and Weiss, J.M., Effect of ACTH 4-10 on passive avoidance of rats lacking vasopressin (Brattleboro strain), Horm. Behav., 10 (1978) 22-29. 2 Bailey, W.H. and Weiss, J.M., Evaluation of a 'memory deficit' in vasopressin-deficient rats, Brain Res., 162 (1979) 174-178. 3 Bohus, B., Van Wimersma Greidanus, Tj. B. and De Wied, D., Behavioral and endocrine responses of rats with hereditary hypothalamic diabetes insipidus (Brattleboro strain), Physiol. Behav., 14 (1975) 609-615. 4 Brito, G.N.O., Thomas, G.J., Gingold, S.I. and Gash, D.M., Behavioral characteristics of vasopressin-deficient rats (Brattleboro strain), Brain Res. Bull., 6 (1981) 71-75. 5 Celestian, J.F., Carey, R.J. and Miller, M., Unimpaired maintenance of a conditioned avoidance response in the rat with diabetes insipidus, Physiol. Behav., 15 (1975) 707-711. 6 Dlouma, H., Krecek, J. and Zicha, J., Growth and urine osmolarity in young Brattleboro rats, J. Endocr., 75 (1977) 329-330. 7 Miller, M. and Moses, A.M., Radioimmunoassay of urinary antidiuretic hormone with application to study of the Brattleboro rat, Endocrinology, 88 (1969) 1389-1396. 8 Miller, M., Barranda, E.G., Dean, M.C. and Brush, F.R., Does the rat with hereditary hypothalamic diabetes insipidus have impaired avoidance learning and/or performance? Pharmacol. Biochem. Behav., 5 (1976) 35-40. 9 Valtin, H., Hereditary hypothalamic diabetes insipidus in rats (Brattleboro strain): a useful experimental model, Amer. J. Med., 42 (1967) 814-827. 10 De Wied, D., Effects ofpeptide hormones on behavior. In L. Martini and W.F. Ganong (Eds.), Frontiers in Neuroendocrinology, Oxford University Press, 1969, pp. 97-140. 11 De Wied, D., Van Delft, A.M.L., Gispen, W.H., Weijnen, J.A,W.M. and Van Wimersma Greidanus, Tj., The role of pituitary-adrenal system hormones in active avoidance conditioning. In S. Levine (Ed.), Hormones and Behavior, Academic Press, 1972, pp. 135-171. 12 De Wied, D., Bohus, B. and Van Wimersma Greidanus, Tj. B., Memory deficit in rats with hereditary diabetes insipidus, Brain Res., 85 (1975) 152-156.