Postnatal Behavioral Alterations Resulting from Prenatal Administration of dl -Alphamethylparatyrosine R. B. LYDIARD S. B. SPARBER Department of Pharmacology University of Minnesota Minneapolis, Minnesota
We have previously reported that injection of dl-alphamethylparatyrosine methylester (AMPT) into the yolk sac of fertilized chicken eggs prior to incubation results in dose-related increases in the specific activity of tyrosine hydroxylase (TH) at 29 days postnatally. We now report that injection of 10.0 and 33.3 mg/kg of AMPT prior to incubation results in dose-related increases in acquisition rates of 2 operant responses. Between 1 and 3 weeks of postnatal age, we observed a dose-related decrease in latency to perform a detour response and, at 55-60 days of postnatal age, an increased rate of acquisition of a key peck response in an autoshaping paradigm. We could not discern any significant alteration in several measures of unconditioned behaviors in an open field situation at 3 weeks of postnatal age. Our data support the contention that early catecholamine (CA) depletion can produce long-term alterations in postnatal behavior.
The postnatal behavioral effects of the perturbation of the normal course of prenatal development have been well documented over the past decade (Joffe, 1969; Kornetsky, 1970; Sparber, 1972). The embryonic insult may be in the form of an overabundance or deficiency of an endogenous compound (Hanson & Simonsen, 1971; Young, 1964), a foreign compound (Davis & Lin, 1972; Ordy, Samorasjki, Collins, & Rolsten, 1966; Rosenthal & Sparber, 1972; Werboff & Kesner, 1963), or environmental factors such as maternal stress (Smith, Heseltine, & Corson, 1971). The data from controlled manipulations of these types have borne out the contention that in the absence of any gross physical malformations, behavioral abnormalities can be shown in what would otherwise be considered a normal animal. Many studies of these types have attempted only to answer the question of whether or not behavioral abnormalities are produced, and very little investigation into the possible neurochemical mechanisms of the alteration has been attempted. The purpose of this study was to explore the postnatal behavioral effects of administration of d1-alphamethylparatyrosine methylester (AMPT) into the yolk sac of Reprint requests should be sent to S . B. Sparber, Department of Pharmacology, 105 Millard Hall, University of Minnesota, Minneapolis, Minnesota 55455, U.S.A. Received for publication 5 April 1976 Revised for publication 3 August 1976 Developmental Psychobiology, lO(4): 305-314 (1977) @ 1977 by John Wiley & Sons, Inc.
305
306
LYDIARD AND SPARBER
fertilized chicken eggs. Previous neurochemical studies in chickens performed in our laboratory have shown that embryonic depletion of brain catecholamines (,CA) with reserpine or AMPT produces specific and dose-related effects postnatally on brain tyrosine hydroxylase (Lydiard, Fossuin, & Sparber, 1975; Lydiard & Sparber, 1974). This enzyme catalyzes the first step in the biosynthesis of CA and is believed to be the primary biological control point of CA synthesis (Weiner, 1970). The 2 operant behavioral measures which were employed in these studies, detour learning (Scholes, 1965) and autoshaping (Brown & Jenkins, 1968), have been shown to be sensitive to manipulation of prenatal development in this species (Rosenthal, 1973; Rosenthal & Sparber, 1972: Sparber & Shideman, 1968). In addition to these operants, the chicks were also observed in an open field apparatus.
Materials and Methods Animals Fertilized White Leghorn chicken eggs (Gallus domesticus) were obtained from a local hatchery, candled to remove any defective eggs, randomized according to a table of random numbers, and distributed into appropriate treatment groups. The dl-alphamethylparatyrosine (AMPT), obtained from Aldrich Biochemicals, Milwaukee, Wisc., was injected into the yolk sac (10.0 and 33.3 mg/kg of egg) by previously published methods (McClaughlin, Marliac, Verrett, Mutchler, & Fitzhugh, 1963; Sparber & Shideman, 1968). The volume of fluid never exceeded 10 pl. Vehicle (isotonic saline) injected and untreated eggs were included to assess what effect the injection procedure might have on hatchability and behavior. Also included was a group of chicks injected intraperitoneally, within 24 hr after hatching (Day I ) , with the absolute dose equivalent to the high dose (about 2 mg AMPT methylester per chick). This allowed us to control for the possibility that an unabsorbed bolus of drug remained in the yolk sac and was taken up into the abdomen with the yolk near the time of hatching, with subsequent absorption of the drug postnatally. Eggs were incubated in a forced-air, rotating incubator for 18% days arid transferred to a forced-air hatcher until hatching occurred and hatched chicks were housed in community-type brooders (incubator Model 100, hatcher Model C. Huniidaire Incubator Co., New Madison, Ohio). Behavioral measurements were made at the appropriate postnatal ages.
Detour Learning The apparatus employed is described in detail elsewhere (Scholes, 1965; Sparber & Shideman, 1969). Various treatment groups (n = 14) were exposed to the detour learning situation as follows: At 7, 8, 10, and 17 days of postnatal age, 2 food-deprived chicks were placed near the light (also a source of warmth) with food (Purina Startena) on the social side of the apparatus and the experimental animals, deprived of food for 18 hr prior to testing, were placed, one at a time, on the test side of the Plexiglas partition facing the reinforcing food, warmth, and broodmates. The appropriate response was to turn away from the reinforcers and detour through the opaque tunnel.
CA DEPLETION AND POSTNATAL BEHAVIOR
307
If no such response was made within 4 min, the chick was guided to the tunnel with a wooden probe and was gently forced to make the appropriate response. Chicks were allowed to remain for 15 sec on the social side once access had been gained. Response latency, a maximum of 240 sec for no response, was measured from the time at which the subject was placed in isolation on the test side t o emergence on the social side. White noise (Grason-Stadler Model 9 0 1B; Grason-Stadler Corp., W. Concord, Mass.) was used to eliminate extraneous auditory stimuli. Each subject was given 1 trial daily and all chicks were given free access to food (Purina Startena) at the end of each day's experimental session. They were again deprived of food for 18 hr before the beginning of the next experimental session.
Autoshaping Brown and Jenluns (1968) described acquisition of a key peck response in the pigeon by the autoshaping method. The nature of the paradigm is such that acquisition of both the key peck response and discrimination can be studied with essentially no interaction with the experimenter. The apparatus used was a 2-key operant conditioning chamber described by Sparber (1970). One key was inoperable at all times; the other translucent plastic key served as the operandum and was located 10 cm above the mesh floor. A red light was used to transilluminate the key. A modified Gerbrands pigeon grain feeder (Ralph Gerbrands Co., Arlington, Mass.) was placed in the floor below the center of chamber manipulanda panel. When the food hopper operated, a chicken had 3-sec access to food through a 2.5 x 3.8 crn hole in the mesh floor. A feeder light just above the aperture came on during operation of the food hopper. A 12-W houselight was on at all times. A small exhaust fan provided ventilation and masking noise during the experimental sessions. Data collection and reinforcement schedules were controlled by standard electromechanical relay-type equipment.
Procedure At 55 days of postnatal age, the 3 groups of naive subjects, which were injected and hatched as described above, included those injected with 10.0 and 33.3 mg AMPT/kg of egg prior to incubation and vehicle-injected control chickens. Eight subjects were randomly assigned to each group. Animals were deprived of food for 16-18 hr before each session. Each subject was placed in the operant chamber on 5 consecutive days. A session consisted of 25 pairings of an 8-sec red key-light period could be monitored. White noise (Grason-Stadler Model 901 B, Crason-Stadler Corp., W. Concord, Mass.) was used to mask external auditory stimuli during test periods. Each chick was placed in the center square and monitored on 3 different days for 120-sec duration each day. Measures taken were latency to leave the center square, lines crossed, and chirps emitted during the 2-min test period.
Statistics Group differences in detour learning were determined using Student’s t-test following a 2-way analysis of variance for repeated measures. Group differences f o r autoshaping behavior was determined by a Kruskal-Wallis analysis of variance by rank.
Detour Learning No statistical difference existed between the uninjected and vehicle-injected (saline) controls, indicating that the injection procedure had no postnatal behavioral
CA DEPLETION AND POSTNATAL BEHAVIOR
309
effect; accordingly, the untreated control group was omitted so that the vehicleinjected, 10.0, and 33.3 mg AMPTikg of egg groups (n = 14 each) could be appropriately compared statistically (3 x 4 factorial 2-way analysis of variance with repeated measures). The analysis showed that within treatment groups across trials the latency to respond decreased significantly ( F = 40.34; df = 3/117; p < .Ol); of greater interest was the finding of a significant difference in the rates at which the 3 groups decreased the latency of the detour response across trials ( F = 2.98; d f = 6/117; p < .01). On the 1st exposure to the detour learning task (Day 7 postnatally), a significant difference existed between the low dose (10.0 mg AMPT/kg of egg prior to incubation) and controls, indicating that exposure to this low dose of AMPT resulted in lowering the operant level of the detour response in this group (see Fig. 2). Only 1 of the subjects in the low-dose group made the detour response correctly within 4 min. The high-dose group (33.3 mg AMPT/kg of egg prior to incubation) showed a tendency toward lower average latencies than control on the 1st trial, but the difference was not statistically significant (F > .05). By the 2nd trial, however, the high-dose group performed the appropriate response with a significantly lower latency than control. A rapid acquisition of the detour response between Trials 1 and 2 exhibited by the low-dose group was observed. Although the difference between the low dose and control mean latencies was not significant, the change from 1 out of 14 responses on Trial 1 to 12 out of 14 successful detour responses in this group by Trial 2 is noteworthy. By Trial 3, the response latencies of these groups were approaching asymptotic levels and no further differences between groups were detected. Chicks given the absolute dose equivalent of the high-dose group within 24 hr after hatching were never different from control. 240
-
220 200 180-
z
8 160W
140-
V
2 120W
c
a IOOJ
5
80-
0 t- 60-
w n
40
-
20 O'
;
1
2
I
I
3
4
I 1
T R I ALS
Fig. 2. Effect of prenatal administration of 10.0 and 33.3 mg AMF'T/kg of egg prior t o incubation on detour response latencies during trials at 7, 8, 10, and 17 days postnatally. Mean response latencies are plotted along the ordinate and trials along the abcissa. Circles represent the mean latencies of the vehicle-injected (saline) control group (n = 14); squares and triangles represent mean latencies of groups treated with 10.0 and 33.3 mg/kg of AMPT, respectively (n = 14 each). * p < .05; * * p < .01; t-test after 2-way analysis of variance for repeated measures (see Results section).
310
LYDIARD AND SPARBER
Open Field Due to the unexpected finding of an apparent facilitation of the acquisition of the detour learning, we raised the question of whether the operant level of locomotor activity and/or exploration was greater in the drug-treated chicks. To test this possibility we placed these same chicks in the open-field apparatus (as described in Methods) 4 days after the termination of the detour experiment. No group difference in latency t o leave the center square, lines crossed, or chirps emitted was observed during any of the 3 exposures.
Autoshaping The cumulation of mean total key peck responses showed an increase in the total response output in the high-dose group (see Fig. 3 ) emerging by the 3rd session and reaching statistical significance by Session 5 (Kruskal-Wallis analysis of variance by ranks: ( H ' = 4.37, df 2 2 , p < .OOS). No difference between saline controls and the low-dose groups was seen with respect to total response output. I n addition to more total responses, the high-dose group emitted a greater number of S D (during lit-key period) responses relative t o controls (see Fig. 4 ) by Session 5 (Kruskal-Wallis analysis of variance by ranks; H ' = 8.77, df = 2 , p < .OS). Again, 110 difference between the saline-treated and the low-dose group was noted with respect to 5'" responses. No difference in the mean cumulated S" responses was observed between groups during any session, indicating that the total mean cumulative response differences observed were due t o appropriate discrimination in the high-dose group.
I
2
3
4
5
SESSIONS
Fig. 3. Effect of prenatal administration of 10.0 and 33.3 mg AMPT/kg of egg, prior to incubation, o n mean cumulative reinforced responses in 55-60 day-old chickens. Circles, squares. and triangles represent groups on vehicle (saline), 10.0, and 33.3 mg AMPT/kg of egg, respectively (8 animals in each group). ' p < .005 (see Results section).
CA DEPLETION AND POSTNATAL BEHAVIOR
3 11
u)
5
W
50-
a >
W Y
400
a
0
4 30-
u a W
5> 2 0 _1
3
5
0
z a
10-
W
I SESSIONS
Fig. 4. Effect of prenatal administration of 10.0 and 33.3 mg AMPT/kg of egg, prior to incubation, on mean cumulative reinforced responses in 55-60 day-old chickens. Circles, squares, and triangles represent groups on vehicle (saline), 10.0, and 33.3 rng AMFT/kg of egg, respectively (8 animals in each group). * p < .05 (see Results section).
Discussion Sparber (1 974) observed sex-related alterations in the variability of discrimination responding in the albino rat after exposure to AMPT in utero. Loizou (1971) studied the developmental differences in CA turnover after neonatal AMPT administration in the rat, but no behavioral tests were performed in that study. In a more recent study, Smith, Cooper, and Breese (1973) found hyperactivity and increased acquisition of a shuttle-box avoidance test in 42-day-old rats in which central norepinephrine stores were depleted neonatally by 6-hydroxydopamine (6-OH-DA). Due to the species difference and the vastly different mechanisms of action of 6-OH-DA and AMPT, a rigorous comparison of these results and the data we present in this study are of limited value. In addition, the hyperactivity reported by Smith et al. may indeed have been the primary drug effect produced and increased avoidance behavior acquisition may have been secondary to this phenomenon. In contrast to other studies of behavioral teratology using the chick as a model (Rosenthal, 1973; Rosenthal & Sparber, 1972; Sparber & Shideman, 1968), no interference with compatible postnatal behavior was observed in an operant situation in our studies. We have, in fact, observed an apparent facilitation in the acquisition of 2 operant responses: the detour learning task and the key peck in the autoshaping experiments. No differences in any of several unconditioned behaviors were noted in the open-field studies. In the detour learning test, group differences, when observed, usually emerge within the first few trials, most probably due to the inherent simplicity of the response. As stated above, a decrement in acquisition or performance is usually
3 12
LYDIARD AND SPARBER
observed as a result of embryonic insult from drug administration. Sparber and Shideman (1969) showed that exposure t o reserpine during embryogenesis produced a decrease in detour responding postnatally in chickens of the same developmental age as those used in the present studies. Reserpine shares the CA depleting action of AMPT. but does so by interfering with storage of CA (Sulser & Sanders-Bush, 1971) as opposed to the inhibition of CA synthesis caused by AMPT (Nagatsu, Levitt. & Udenfriend, 1964). The paradoxically opposite behavioral effects seen after prenatal exposure t o these compounds are possibly explained by some of the differences in their pharmacological actions on developing embryos. After exposure t o reserpine duririg development, CA stores in whole brain of lo-, 15-. and 20-day-old embryos and in 3-day-old chicks are depleted significantly, and are elevated by approximately 50% above control at 29 days after hatching (Lydiard & Sparber, 1974). Although both drugs cause an equivalent increase in the specific activity of the tyrosine hydroxylase enzyme at 29 days postnatally, n o increase in brain CA at 29 days after hatching is observed after AMPT. In addition, the period of CA depletion after administration of 10.0 or 33.3 nig AMPT/kg of egg prior t o incubation is terminated by Day 20 of embryogenesis (Lydiard et al., 1975). Also, whereas AMPT affects rather specifically CA synthesis, reserpine has profound effects on storage of a number of transmitter substances in the brain, such as serotonin, histamine, and acetylcholine (Beani, Ledda, Bianchi, & Baldi, 1966; Sjoerdsma, Waalkes, & Weissbach, 1958). In light of the wider variety of neurochemical actions of reserpine, the failure t o observe similar behavioral results due to a single common action is not surprising. One could easily postulate that an interaction of reserpine with several amine systems simultaneously (or a disturbance in homeostasis between these amine systems) could produce the net decrease observed in detour performance observed by Sparber and Shideman in their studies. Whether a direct relationship exists between the effects of AMEYT administered prenatally on postnatal tyrosine hydroxylase and postnatal behavior is unclear. Administration of 10.0 and 33.3 rng AMPT/kg of egg prior to incubation resulted in respective increases in whole brain TH of about 15 and 3096, respectively (Lydiard et al.. 1975). A similar dose-related, step-wise increment in acquisition of the detour and autoshaping keypeck responses was not observed. The low dose (10.0 mg AMPT) was not different from vehicle-injected control at any time during the behavioral studies, except upon the 1st exposure t o the detour learning situation (Fig. 2). If a relationship exists hetween tyrosine hydroxylase and the behaviors examined in these studies, possibly a threshold level of increased enzyme activity must be reached before a measurable behavioral effect is realized. Of course, no relationship may exist between the activity of whole brain tyrosine hydroxylase and the behaviors observed in these studies, but the coexistence of differences in neurocheniistry and behavior caused by the same compound remains a provocative suggestion of some interrelation between these 2 phenomena. The apparent facilitation of the detour response acquisition after prenatal exposure to AMPT could have been secondary t o increased exploratory behavior, or some nonspecific increase in locomotion as a drug effect. The negative findings from the open-field studies, which were run in close temporal proximity t o the detour testing, suggest this was not the case. However, because the open-field testing was not performed a t the same age period in which the facilitation o f the detour responding
CA DEPLETION AND POSTNATAL BEHAVIOR
313
was observed, we cannot exclude the possibility that this was due to a drug-induced developmental increase in exploration, locomotion, or response to a novel situation which had disappeared by the time the open-field testing was performed. The observed facilitation, however, in the autoshaping acquisition study in the high-dose group in 55-60 day-old chickens, almost a month later, further supports the possibility of a developmentally induced facilitation of learning. Certainly, several types of behavioral testing done at varying ages would be required to confirm that a general facilitation of learning exists after prenatal exposure to AMPT. The detour learning requires locomotion and exploration to a much greater extent than does the autoshaping procedure, which requires more exact responses (key pecking) and attending to the discriminative stimuli. In addition to the variety of responses which were involved in these tests, an extended time period was involved from drug administration to the last behavioral testing (nearly 90 days). Whether a true facilitation of learning occurred is not important; our experiments have shown, at both doses, divergence from control behavior. Perseverance in a laboratory situation, resulting in performance enhancement, may be incompatible with survival and/or normal development in the natural habitat. Until proven otherwise, we consider this to be a deleterious effect resulting from prenatal drug exposure.
Notes Supported in part by grants from the USPHS (MH-08565) and the Minnesota Medical Foundation (FSW-1-71).
References Beani, L., Ledda, F., Bianchi, C., and Baldi, J. (1966). Reversal by 3,4-dihydroxyphenyldanine of reserpine-induced regional changes in acetylcholine content in guinea pig brain. Biochem. Pharmacol., 15:779-784. Brown, P. L., and Jenkins, H. M. (1968). Autoshaping of the pigeon’s key peck. J. Exp. Anal. Behav., 1 1 : 1-8. Davis, W. M., and Lin, C. H. (1972). Prenatal morphine effects on survival and behavior of rat offspring. Res. Comm. Chem. Path. Pharmacol., 3:205-214. Hall, C. S. (1934). Emotional behavior in the rat. 1. Defecation and urination as measures of individual differences in cmotionality. J. Comp. Psychol. 18:385-403. Hanson, H. M. and Simonsen, M. (1971). Effects of fetal undernourishment on experimental anxiety. Nutr. Rep. Natl., 4:307-314. Hess, E. H. (1959). Imprinting. Science, 130: 133-141. Joffe, J. M. (1969). Prenatal Determinants of Behavior. Oxford: Pergamon. Kornetsky, C. (1970). Psychoactive drugs in the immature organism. Psychopharmacologia, 17: 105-136. Loizu, L. A. (1971). Effect of inhibition of catecholamine synthesis on central catecholaminecontaining neurones in the developing albino rat. Br. J. Pharmacol., 41 :41-48. Lydiard, R. B., Fossom, L. H., and Sparber, S. B. (1975). Postnatal elevation of brain tyrosine hydroxylase activity, without concurrent increases in steady-state catecholamine levels, resulting from dl-alphamethylparatyrosine administration during embryonic development. J. Pharmacol. Exp. Ther., 194:27-36. Lydiard, R. B., and Sparber, S. B. (1974). Evidence for a critical period for postnatal elevation of brain tyrosine hydroxylase activity resulting from reserpine administration during embryonic development. J. Pharmacol. Exp. Ther.. 189:370-379.
314
LYDIARD AND SPARBER
McGlaughlin, J., Marliac, J. P., Verrett, M. J., Mutchler, M. K . , and Fitzhugh, 0. H. (1963). The injection of chemicals into the yolk sac of fertile eggs prior to incubation as a toxicity test. Tox. Appl. Pharmacol., 5:760-771. Nagatsu, T., Levitt, M., and Udenfriend, S. (1964). Tyrosine hydroxylase: The initial step in norepinephrine biosynthesis. J. Biol. Chem., 229:2910-2919. Ordy, J . M., Samorajski, T., Collins, R. L., and Rolsten, C. (1966). Prenatal chlorpromazine effects on liver, survival, and behavior of mice offspring. J. Pharmacol. Exp. Ther., 151: 110-125. Rosenthal, E. (1973). The effect of methylmercury dicyandiamide administration during embryogenesis, upon the development of the young chicken. Unpublished Ph.D. thesis, University of Minnesota. Rosenthal, E., and Sparber, S. B. (1972). Methylmercury dicyandiamide: retardation of detour learning in chicks hatched from injected eggs. Life Sci., 11:883-892. Scholes, N. W. (1965). Detour learning and development in the domestic chick. J. Comp. Physiol. Psychol., 60: 114-116. Sjoerdsma, A,, Waalkes, T. P., and Weissbach, H. (1958). Studies on serotonin and histamine in mast cells. J. Phamacol., 122:69A. Smith, D. J., Heseltine, G. F. D., and Corson, J . A. (1971). Pre-pregnancy and prenatal stress in five consecutive pregnancies: Effects o n female rats and their offspring. Life Sci., 10:1233-1242. Smith, R. D., Cooper, B. R., and Breese, G . R. (1973). Growth and behavioral changes in developing rats treated intracisternally with 6-hydroxydopamine: Evidence for involvement of brain dopamine. J. Pharmacol. Exp. Ther., 185:609-6 19. Sparber, S. B. (1970). Schedule control of concurrent fixed-ratio responding and discrimination acquisition in chickens. Psychon. Sci., 18:135-136. Sparber, S. B. (1972). Effects of drugs on the biochemical and behavioral responses of developing organisms Fed. Proc., 31 :74-80. Sparber, S. B. (1974). Postnatal behavioral effects of in utero exposure t o drugs which modify catecholamines and/or serotonin. A. Vernadakis and N. Weiner (Eds.), Drugs and the Dewloping Brain. New York: Plenum Press. Pp. 81-102. Sparber, S. B., and Shideman, I;. E. (1968). Prenatal administration of reserpine: Effect upon hatching, behavior, and brainstem catecholamines of the young chick. Devel. Psychobiol., I : 236-244. Sparber, S . B., and Shideman, F. E. (1969). Detour learning in the chick: Effect of reserpine administered during embryonic development. Devel. Psychobiol., 256-59. Sulser, I-., and Sanders-Bush, E. (1971). Effect of drugs on amines in the CNS. Ann. Rev. Pharmacol., I I :209-229. Weiner, N. (1970). Regulation of norepinephrine biosynthesis. Ann. Rev. Pharmacol., 10:273-290. Wcrboff, J., and Kesner, R. (1963). Learning deficits of offspring after administration of tranquillizing drugs t o mothers. Nature, 197:106-107. Young, R. D. (1964). Effect of prenatal drugs and neonatal stimualtion on later behavior. J. Comp. Physiol. Psychol., 58:309-311.
We have previously reported that injection of dl-alphamethylparatyrosine methylester (AMPT) into the yolk sac of fertilized chicken eggs prior to incubation results in dose-related increases in the specific activity of tyrosine hydroxylase (TH) at 29 days postnatally. We now report that injection of 10.0 and 33.3 mg/kg of AMPT prior to incubation results in dose-related increases in acquisition rates of 2 operant responses. Between 1 and 3 weeks of postnatal age, we observed a dose-related decrease in latency to perform a detour response and, at 55-60 days of postnatal age, an increased rate of acquisition of a key peck response in an autoshaping paradigm. We could not discern any significant alteration in several measures of unconditioned behaviors in an open field situation at 3 weeks of postnatal age. Our data support the contention that early catecholamine (CA) depletion can produce long-term alterations in postnatal behavior.
The postnatal behavioral effects of the perturbation of the normal course of prenatal development have been well documented over the past decade (Joffe, 1969; Kornetsky, 1970; Sparber, 1972). The embryonic insult may be in the form of an overabundance or deficiency of an endogenous compound (Hanson & Simonsen, 1971; Young, 1964), a foreign compound (Davis & Lin, 1972; Ordy, Samorasjki, Collins, & Rolsten, 1966; Rosenthal & Sparber, 1972; Werboff & Kesner, 1963), or environmental factors such as maternal stress (Smith, Heseltine, & Corson, 1971). The data from controlled manipulations of these types have borne out the contention that in the absence of any gross physical malformations, behavioral abnormalities can be shown in what would otherwise be considered a normal animal. Many studies of these types have attempted only to answer the question of whether or not behavioral abnormalities are produced, and very little investigation into the possible neurochemical mechanisms of the alteration has been attempted. The purpose of this study was to explore the postnatal behavioral effects of administration of d1-alphamethylparatyrosine methylester (AMPT) into the yolk sac of Reprint requests should be sent to S . B. Sparber, Department of Pharmacology, 105 Millard Hall, University of Minnesota, Minneapolis, Minnesota 55455, U.S.A. Received for publication 5 April 1976 Revised for publication 3 August 1976 Developmental Psychobiology, lO(4): 305-314 (1977) @ 1977 by John Wiley & Sons, Inc.
305
306
LYDIARD AND SPARBER
fertilized chicken eggs. Previous neurochemical studies in chickens performed in our laboratory have shown that embryonic depletion of brain catecholamines (,CA) with reserpine or AMPT produces specific and dose-related effects postnatally on brain tyrosine hydroxylase (Lydiard, Fossuin, & Sparber, 1975; Lydiard & Sparber, 1974). This enzyme catalyzes the first step in the biosynthesis of CA and is believed to be the primary biological control point of CA synthesis (Weiner, 1970). The 2 operant behavioral measures which were employed in these studies, detour learning (Scholes, 1965) and autoshaping (Brown & Jenkins, 1968), have been shown to be sensitive to manipulation of prenatal development in this species (Rosenthal, 1973; Rosenthal & Sparber, 1972: Sparber & Shideman, 1968). In addition to these operants, the chicks were also observed in an open field apparatus.
Materials and Methods Animals Fertilized White Leghorn chicken eggs (Gallus domesticus) were obtained from a local hatchery, candled to remove any defective eggs, randomized according to a table of random numbers, and distributed into appropriate treatment groups. The dl-alphamethylparatyrosine (AMPT), obtained from Aldrich Biochemicals, Milwaukee, Wisc., was injected into the yolk sac (10.0 and 33.3 mg/kg of egg) by previously published methods (McClaughlin, Marliac, Verrett, Mutchler, & Fitzhugh, 1963; Sparber & Shideman, 1968). The volume of fluid never exceeded 10 pl. Vehicle (isotonic saline) injected and untreated eggs were included to assess what effect the injection procedure might have on hatchability and behavior. Also included was a group of chicks injected intraperitoneally, within 24 hr after hatching (Day I ) , with the absolute dose equivalent to the high dose (about 2 mg AMPT methylester per chick). This allowed us to control for the possibility that an unabsorbed bolus of drug remained in the yolk sac and was taken up into the abdomen with the yolk near the time of hatching, with subsequent absorption of the drug postnatally. Eggs were incubated in a forced-air, rotating incubator for 18% days arid transferred to a forced-air hatcher until hatching occurred and hatched chicks were housed in community-type brooders (incubator Model 100, hatcher Model C. Huniidaire Incubator Co., New Madison, Ohio). Behavioral measurements were made at the appropriate postnatal ages.
Detour Learning The apparatus employed is described in detail elsewhere (Scholes, 1965; Sparber & Shideman, 1969). Various treatment groups (n = 14) were exposed to the detour learning situation as follows: At 7, 8, 10, and 17 days of postnatal age, 2 food-deprived chicks were placed near the light (also a source of warmth) with food (Purina Startena) on the social side of the apparatus and the experimental animals, deprived of food for 18 hr prior to testing, were placed, one at a time, on the test side of the Plexiglas partition facing the reinforcing food, warmth, and broodmates. The appropriate response was to turn away from the reinforcers and detour through the opaque tunnel.
CA DEPLETION AND POSTNATAL BEHAVIOR
307
If no such response was made within 4 min, the chick was guided to the tunnel with a wooden probe and was gently forced to make the appropriate response. Chicks were allowed to remain for 15 sec on the social side once access had been gained. Response latency, a maximum of 240 sec for no response, was measured from the time at which the subject was placed in isolation on the test side t o emergence on the social side. White noise (Grason-Stadler Model 9 0 1B; Grason-Stadler Corp., W. Concord, Mass.) was used to eliminate extraneous auditory stimuli. Each subject was given 1 trial daily and all chicks were given free access to food (Purina Startena) at the end of each day's experimental session. They were again deprived of food for 18 hr before the beginning of the next experimental session.
Autoshaping Brown and Jenluns (1968) described acquisition of a key peck response in the pigeon by the autoshaping method. The nature of the paradigm is such that acquisition of both the key peck response and discrimination can be studied with essentially no interaction with the experimenter. The apparatus used was a 2-key operant conditioning chamber described by Sparber (1970). One key was inoperable at all times; the other translucent plastic key served as the operandum and was located 10 cm above the mesh floor. A red light was used to transilluminate the key. A modified Gerbrands pigeon grain feeder (Ralph Gerbrands Co., Arlington, Mass.) was placed in the floor below the center of chamber manipulanda panel. When the food hopper operated, a chicken had 3-sec access to food through a 2.5 x 3.8 crn hole in the mesh floor. A feeder light just above the aperture came on during operation of the food hopper. A 12-W houselight was on at all times. A small exhaust fan provided ventilation and masking noise during the experimental sessions. Data collection and reinforcement schedules were controlled by standard electromechanical relay-type equipment.
Procedure At 55 days of postnatal age, the 3 groups of naive subjects, which were injected and hatched as described above, included those injected with 10.0 and 33.3 mg AMPT/kg of egg prior to incubation and vehicle-injected control chickens. Eight subjects were randomly assigned to each group. Animals were deprived of food for 16-18 hr before each session. Each subject was placed in the operant chamber on 5 consecutive days. A session consisted of 25 pairings of an 8-sec red key-light period could be monitored. White noise (Grason-Stadler Model 901 B, Crason-Stadler Corp., W. Concord, Mass.) was used to mask external auditory stimuli during test periods. Each chick was placed in the center square and monitored on 3 different days for 120-sec duration each day. Measures taken were latency to leave the center square, lines crossed, and chirps emitted during the 2-min test period.
Statistics Group differences in detour learning were determined using Student’s t-test following a 2-way analysis of variance for repeated measures. Group differences f o r autoshaping behavior was determined by a Kruskal-Wallis analysis of variance by rank.
Detour Learning No statistical difference existed between the uninjected and vehicle-injected (saline) controls, indicating that the injection procedure had no postnatal behavioral
CA DEPLETION AND POSTNATAL BEHAVIOR
309
effect; accordingly, the untreated control group was omitted so that the vehicleinjected, 10.0, and 33.3 mg AMPTikg of egg groups (n = 14 each) could be appropriately compared statistically (3 x 4 factorial 2-way analysis of variance with repeated measures). The analysis showed that within treatment groups across trials the latency to respond decreased significantly ( F = 40.34; df = 3/117; p < .Ol); of greater interest was the finding of a significant difference in the rates at which the 3 groups decreased the latency of the detour response across trials ( F = 2.98; d f = 6/117; p < .01). On the 1st exposure to the detour learning task (Day 7 postnatally), a significant difference existed between the low dose (10.0 mg AMPT/kg of egg prior to incubation) and controls, indicating that exposure to this low dose of AMPT resulted in lowering the operant level of the detour response in this group (see Fig. 2). Only 1 of the subjects in the low-dose group made the detour response correctly within 4 min. The high-dose group (33.3 mg AMPT/kg of egg prior to incubation) showed a tendency toward lower average latencies than control on the 1st trial, but the difference was not statistically significant (F > .05). By the 2nd trial, however, the high-dose group performed the appropriate response with a significantly lower latency than control. A rapid acquisition of the detour response between Trials 1 and 2 exhibited by the low-dose group was observed. Although the difference between the low dose and control mean latencies was not significant, the change from 1 out of 14 responses on Trial 1 to 12 out of 14 successful detour responses in this group by Trial 2 is noteworthy. By Trial 3, the response latencies of these groups were approaching asymptotic levels and no further differences between groups were detected. Chicks given the absolute dose equivalent of the high-dose group within 24 hr after hatching were never different from control. 240
-
220 200 180-
z
8 160W
140-
V
2 120W
c
a IOOJ
5
80-
0 t- 60-
w n
40
-
20 O'
;
1
2
I
I
3
4
I 1
T R I ALS
Fig. 2. Effect of prenatal administration of 10.0 and 33.3 mg AMF'T/kg of egg prior t o incubation on detour response latencies during trials at 7, 8, 10, and 17 days postnatally. Mean response latencies are plotted along the ordinate and trials along the abcissa. Circles represent the mean latencies of the vehicle-injected (saline) control group (n = 14); squares and triangles represent mean latencies of groups treated with 10.0 and 33.3 mg/kg of AMPT, respectively (n = 14 each). * p < .05; * * p < .01; t-test after 2-way analysis of variance for repeated measures (see Results section).
310
LYDIARD AND SPARBER
Open Field Due to the unexpected finding of an apparent facilitation of the acquisition of the detour learning, we raised the question of whether the operant level of locomotor activity and/or exploration was greater in the drug-treated chicks. To test this possibility we placed these same chicks in the open-field apparatus (as described in Methods) 4 days after the termination of the detour experiment. No group difference in latency t o leave the center square, lines crossed, or chirps emitted was observed during any of the 3 exposures.
Autoshaping The cumulation of mean total key peck responses showed an increase in the total response output in the high-dose group (see Fig. 3 ) emerging by the 3rd session and reaching statistical significance by Session 5 (Kruskal-Wallis analysis of variance by ranks: ( H ' = 4.37, df 2 2 , p < .OOS). No difference between saline controls and the low-dose groups was seen with respect to total response output. I n addition to more total responses, the high-dose group emitted a greater number of S D (during lit-key period) responses relative t o controls (see Fig. 4 ) by Session 5 (Kruskal-Wallis analysis of variance by ranks; H ' = 8.77, df = 2 , p < .OS). Again, 110 difference between the saline-treated and the low-dose group was noted with respect to 5'" responses. No difference in the mean cumulated S" responses was observed between groups during any session, indicating that the total mean cumulative response differences observed were due t o appropriate discrimination in the high-dose group.
I
2
3
4
5
SESSIONS
Fig. 3. Effect of prenatal administration of 10.0 and 33.3 mg AMPT/kg of egg, prior to incubation, o n mean cumulative reinforced responses in 55-60 day-old chickens. Circles, squares. and triangles represent groups on vehicle (saline), 10.0, and 33.3 mg AMPT/kg of egg, respectively (8 animals in each group). ' p < .005 (see Results section).
CA DEPLETION AND POSTNATAL BEHAVIOR
3 11
u)
5
W
50-
a >
W Y
400
a
0
4 30-
u a W
5> 2 0 _1
3
5
0
z a
10-
W
I SESSIONS
Fig. 4. Effect of prenatal administration of 10.0 and 33.3 mg AMPT/kg of egg, prior to incubation, on mean cumulative reinforced responses in 55-60 day-old chickens. Circles, squares, and triangles represent groups on vehicle (saline), 10.0, and 33.3 rng AMFT/kg of egg, respectively (8 animals in each group). * p < .05 (see Results section).
Discussion Sparber (1 974) observed sex-related alterations in the variability of discrimination responding in the albino rat after exposure to AMPT in utero. Loizou (1971) studied the developmental differences in CA turnover after neonatal AMPT administration in the rat, but no behavioral tests were performed in that study. In a more recent study, Smith, Cooper, and Breese (1973) found hyperactivity and increased acquisition of a shuttle-box avoidance test in 42-day-old rats in which central norepinephrine stores were depleted neonatally by 6-hydroxydopamine (6-OH-DA). Due to the species difference and the vastly different mechanisms of action of 6-OH-DA and AMPT, a rigorous comparison of these results and the data we present in this study are of limited value. In addition, the hyperactivity reported by Smith et al. may indeed have been the primary drug effect produced and increased avoidance behavior acquisition may have been secondary to this phenomenon. In contrast to other studies of behavioral teratology using the chick as a model (Rosenthal, 1973; Rosenthal & Sparber, 1972; Sparber & Shideman, 1968), no interference with compatible postnatal behavior was observed in an operant situation in our studies. We have, in fact, observed an apparent facilitation in the acquisition of 2 operant responses: the detour learning task and the key peck in the autoshaping experiments. No differences in any of several unconditioned behaviors were noted in the open-field studies. In the detour learning test, group differences, when observed, usually emerge within the first few trials, most probably due to the inherent simplicity of the response. As stated above, a decrement in acquisition or performance is usually
3 12
LYDIARD AND SPARBER
observed as a result of embryonic insult from drug administration. Sparber and Shideman (1969) showed that exposure t o reserpine during embryogenesis produced a decrease in detour responding postnatally in chickens of the same developmental age as those used in the present studies. Reserpine shares the CA depleting action of AMPT. but does so by interfering with storage of CA (Sulser & Sanders-Bush, 1971) as opposed to the inhibition of CA synthesis caused by AMPT (Nagatsu, Levitt. & Udenfriend, 1964). The paradoxically opposite behavioral effects seen after prenatal exposure t o these compounds are possibly explained by some of the differences in their pharmacological actions on developing embryos. After exposure t o reserpine duririg development, CA stores in whole brain of lo-, 15-. and 20-day-old embryos and in 3-day-old chicks are depleted significantly, and are elevated by approximately 50% above control at 29 days after hatching (Lydiard & Sparber, 1974). Although both drugs cause an equivalent increase in the specific activity of the tyrosine hydroxylase enzyme at 29 days postnatally, n o increase in brain CA at 29 days after hatching is observed after AMPT. In addition, the period of CA depletion after administration of 10.0 or 33.3 nig AMPT/kg of egg prior t o incubation is terminated by Day 20 of embryogenesis (Lydiard et al., 1975). Also, whereas AMPT affects rather specifically CA synthesis, reserpine has profound effects on storage of a number of transmitter substances in the brain, such as serotonin, histamine, and acetylcholine (Beani, Ledda, Bianchi, & Baldi, 1966; Sjoerdsma, Waalkes, & Weissbach, 1958). In light of the wider variety of neurochemical actions of reserpine, the failure t o observe similar behavioral results due to a single common action is not surprising. One could easily postulate that an interaction of reserpine with several amine systems simultaneously (or a disturbance in homeostasis between these amine systems) could produce the net decrease observed in detour performance observed by Sparber and Shideman in their studies. Whether a direct relationship exists between the effects of AMEYT administered prenatally on postnatal tyrosine hydroxylase and postnatal behavior is unclear. Administration of 10.0 and 33.3 rng AMPT/kg of egg prior to incubation resulted in respective increases in whole brain TH of about 15 and 3096, respectively (Lydiard et al.. 1975). A similar dose-related, step-wise increment in acquisition of the detour and autoshaping keypeck responses was not observed. The low dose (10.0 mg AMPT) was not different from vehicle-injected control at any time during the behavioral studies, except upon the 1st exposure t o the detour learning situation (Fig. 2). If a relationship exists hetween tyrosine hydroxylase and the behaviors examined in these studies, possibly a threshold level of increased enzyme activity must be reached before a measurable behavioral effect is realized. Of course, no relationship may exist between the activity of whole brain tyrosine hydroxylase and the behaviors observed in these studies, but the coexistence of differences in neurocheniistry and behavior caused by the same compound remains a provocative suggestion of some interrelation between these 2 phenomena. The apparent facilitation of the detour response acquisition after prenatal exposure to AMPT could have been secondary t o increased exploratory behavior, or some nonspecific increase in locomotion as a drug effect. The negative findings from the open-field studies, which were run in close temporal proximity t o the detour testing, suggest this was not the case. However, because the open-field testing was not performed a t the same age period in which the facilitation o f the detour responding
CA DEPLETION AND POSTNATAL BEHAVIOR
313
was observed, we cannot exclude the possibility that this was due to a drug-induced developmental increase in exploration, locomotion, or response to a novel situation which had disappeared by the time the open-field testing was performed. The observed facilitation, however, in the autoshaping acquisition study in the high-dose group in 55-60 day-old chickens, almost a month later, further supports the possibility of a developmentally induced facilitation of learning. Certainly, several types of behavioral testing done at varying ages would be required to confirm that a general facilitation of learning exists after prenatal exposure to AMPT. The detour learning requires locomotion and exploration to a much greater extent than does the autoshaping procedure, which requires more exact responses (key pecking) and attending to the discriminative stimuli. In addition to the variety of responses which were involved in these tests, an extended time period was involved from drug administration to the last behavioral testing (nearly 90 days). Whether a true facilitation of learning occurred is not important; our experiments have shown, at both doses, divergence from control behavior. Perseverance in a laboratory situation, resulting in performance enhancement, may be incompatible with survival and/or normal development in the natural habitat. Until proven otherwise, we consider this to be a deleterious effect resulting from prenatal drug exposure.
Notes Supported in part by grants from the USPHS (MH-08565) and the Minnesota Medical Foundation (FSW-1-71).
References Beani, L., Ledda, F., Bianchi, C., and Baldi, J. (1966). Reversal by 3,4-dihydroxyphenyldanine of reserpine-induced regional changes in acetylcholine content in guinea pig brain. Biochem. Pharmacol., 15:779-784. Brown, P. L., and Jenkins, H. M. (1968). Autoshaping of the pigeon’s key peck. J. Exp. Anal. Behav., 1 1 : 1-8. Davis, W. M., and Lin, C. H. (1972). Prenatal morphine effects on survival and behavior of rat offspring. Res. Comm. Chem. Path. Pharmacol., 3:205-214. Hall, C. S. (1934). Emotional behavior in the rat. 1. Defecation and urination as measures of individual differences in cmotionality. J. Comp. Psychol. 18:385-403. Hanson, H. M. and Simonsen, M. (1971). Effects of fetal undernourishment on experimental anxiety. Nutr. Rep. Natl., 4:307-314. Hess, E. H. (1959). Imprinting. Science, 130: 133-141. Joffe, J. M. (1969). Prenatal Determinants of Behavior. Oxford: Pergamon. Kornetsky, C. (1970). Psychoactive drugs in the immature organism. Psychopharmacologia, 17: 105-136. Loizu, L. A. (1971). Effect of inhibition of catecholamine synthesis on central catecholaminecontaining neurones in the developing albino rat. Br. J. Pharmacol., 41 :41-48. Lydiard, R. B., Fossom, L. H., and Sparber, S. B. (1975). Postnatal elevation of brain tyrosine hydroxylase activity, without concurrent increases in steady-state catecholamine levels, resulting from dl-alphamethylparatyrosine administration during embryonic development. J. Pharmacol. Exp. Ther., 194:27-36. Lydiard, R. B., and Sparber, S. B. (1974). Evidence for a critical period for postnatal elevation of brain tyrosine hydroxylase activity resulting from reserpine administration during embryonic development. J. Pharmacol. Exp. Ther.. 189:370-379.
314
LYDIARD AND SPARBER
McGlaughlin, J., Marliac, J. P., Verrett, M. J., Mutchler, M. K . , and Fitzhugh, 0. H. (1963). The injection of chemicals into the yolk sac of fertile eggs prior to incubation as a toxicity test. Tox. Appl. Pharmacol., 5:760-771. Nagatsu, T., Levitt, M., and Udenfriend, S. (1964). Tyrosine hydroxylase: The initial step in norepinephrine biosynthesis. J. Biol. Chem., 229:2910-2919. Ordy, J . M., Samorajski, T., Collins, R. L., and Rolsten, C. (1966). Prenatal chlorpromazine effects on liver, survival, and behavior of mice offspring. J. Pharmacol. Exp. Ther., 151: 110-125. Rosenthal, E. (1973). The effect of methylmercury dicyandiamide administration during embryogenesis, upon the development of the young chicken. Unpublished Ph.D. thesis, University of Minnesota. Rosenthal, E., and Sparber, S. B. (1972). Methylmercury dicyandiamide: retardation of detour learning in chicks hatched from injected eggs. Life Sci., 11:883-892. Scholes, N. W. (1965). Detour learning and development in the domestic chick. J. Comp. Physiol. Psychol., 60: 114-116. Sjoerdsma, A,, Waalkes, T. P., and Weissbach, H. (1958). Studies on serotonin and histamine in mast cells. J. Phamacol., 122:69A. Smith, D. J., Heseltine, G. F. D., and Corson, J . A. (1971). Pre-pregnancy and prenatal stress in five consecutive pregnancies: Effects o n female rats and their offspring. Life Sci., 10:1233-1242. Smith, R. D., Cooper, B. R., and Breese, G . R. (1973). Growth and behavioral changes in developing rats treated intracisternally with 6-hydroxydopamine: Evidence for involvement of brain dopamine. J. Pharmacol. Exp. Ther., 185:609-6 19. Sparber, S. B. (1970). Schedule control of concurrent fixed-ratio responding and discrimination acquisition in chickens. Psychon. Sci., 18:135-136. Sparber, S. B. (1972). Effects of drugs on the biochemical and behavioral responses of developing organisms Fed. Proc., 31 :74-80. Sparber, S. B. (1974). Postnatal behavioral effects of in utero exposure t o drugs which modify catecholamines and/or serotonin. A. Vernadakis and N. Weiner (Eds.), Drugs and the Dewloping Brain. New York: Plenum Press. Pp. 81-102. Sparber, S. B., and Shideman, I;. E. (1968). Prenatal administration of reserpine: Effect upon hatching, behavior, and brainstem catecholamines of the young chick. Devel. Psychobiol., I : 236-244. Sparber, S . B., and Shideman, F. E. (1969). Detour learning in the chick: Effect of reserpine administered during embryonic development. Devel. Psychobiol., 256-59. Sulser, I-., and Sanders-Bush, E. (1971). Effect of drugs on amines in the CNS. Ann. Rev. Pharmacol., I I :209-229. Weiner, N. (1970). Regulation of norepinephrine biosynthesis. Ann. Rev. Pharmacol., 10:273-290. Wcrboff, J., and Kesner, R. (1963). Learning deficits of offspring after administration of tranquillizing drugs t o mothers. Nature, 197:106-107. Young, R. D. (1964). Effect of prenatal drugs and neonatal stimualtion on later behavior. J. Comp. Physiol. Psychol., 58:309-311.
