- Email: [email protected]
A fostering study of the effects of prenatal cocaine exposure: II. Offspring behavioral measures
Neurotoxicologyand Teratology,Vol. 14, pp. 423-432, 1992
0892-0362/92$5.00 + .00 Copyright©1992PergamonPressLtd.
Printedin the U.S.A. All fightsreserved.
A Fostering Study of the Effects of Prenatal Cocaine Exposure: II. Offspring Behavioral Measures G R E G O R Y A . G O O D W I N , C H A R L E S J. H E Y S E R , C A R O L E A. M O O D Y , L A V A N Y A R A J A C H A N D R A N , V I C T O R A. M O L I N A , H. M O O R E A R N O L D , D A V I D L. M C K I N Z I E , N O R M A N E. S P E A R A N D L I N D A P A T I A S P E A R 1
Department of Psychology, Center for Developmental Psychobiology, State o f University o f New York at Binghamton, Binghamton, N Y 13902 Received 7 A p r i l 1992; Accepted 8 September 1992 GOODWIN, G. A., C. J. HEYSER, C. A. MOODY, L. RAJACHANDRAN, V. A. MOLINA, H. M. ARNOLD, D. L. MCKINZIE, N. E. SPEAR AND L. P. SPEAR. A fostering study of the effects of prenatal cocaineexposure: II. Offspring behavioral measures. NEUROTOX TERATOL 14(6) 423-432, 1992.--The impact of rearing condition was assessed in Spragne-Dawley dams given 40 mg/kg cocaine (C40) or saline (LC control) subcutaneously (SC) from gestational days 8-20 and their offspring. Treated pups reared by their biological dams (LC/LC; C40/C40), treated pups reared by surrogate dams (FOS/LC; FOS/C40), and foster pups raised by treated dams (LC/FOS; C40/FOS) were examined. On postnatal day 7 (P7), pups received either 0 (unpaired) 2, 3, or 4 pairings of an odor and footshock and were tested for their aversion to this odor. Foster and LC pups, regardless of rearing condition, exhibited significant odor aversions following either 2, 3, or 4 training trials. In contrast, C40 pups reared by surrogate dams required 4 trials to acquire the aversion, and C40 pups reared by their own dams did not exhibit conditioning even after 4 trials. At PI7, no differences were seen among the groups in the aversion formed to an auditory or an olfactory stimulus that was paired with footshock. At P60, shockelicited aggression among pairs of siblings was examined. Regardless of prenatal exposure condition, offspring reared by dams given cocaine showed a decreased latency to the first aggressivecontact, an effect that was evident without any alteration in shock sensitivity. Together these data suggest that being reared by a dam previously exposed to cocaine has an impact on offspring behavioral function apart from the effects of prenatal cocaine exposure per se. The implications of the data regarding the cognitive performance of pups exposed prenatally to cocaine are also discussed.
PRIOR work has shown that dams exposed to cocaine during pregnancy differ only slightly from control dams on measures of maternal behavior. Dams exposed during pregnancy to cocaine were not observed to differ from control dams in pup retrieval latency, nest building behavior, time spent in the nest, or time spent suckling pups (11,22). However, cocaineexposed dams rearing their own pups exhibited an increase in the number of aggressive attacks against a female intruder relative to all other groups of dams, an effect that was not evident in cocaine-exposed dams rearing normal pups (11). Other aspects of maternal behavior (e.g., maternal licking) were not examined in these studies, and hence it is not clear whether cocaine-induced disruptions in maternal behavior are restricted to alterations in aggressive behavior, or whether other aspects of maternal behavior that were not assessed might also be affected. Alterations in maternal behavior can have a notable impact
on later offspring physiological and behavioral function (e.g., see 2,3,18,20). To avoid this potential consequence of insultinduced alterations in maternal behavior, researchers in the area of developmental toxicology have frequently used a surrogate fostering procedure. Such fostering essentially eliminates the possibility that offspring neurobehavioral effects are a result of insult-induced alterations in the dams' behavior per se (e.g., see 12 for discussion). However, surrogate fostering is costly in terms of animal usage, and the necessity for routine fostering in developmental toxicological studies has been questioned (e.g., see 30). Thus, in studies with any particular drug agent, it would appear prudent to assess whether potential drng-induced disruptions in maternal behavior have a sufficiently salient effect on later offspring neurobehavioral function to warrant the routine use of surrogate fostering in studies with that substance. Although a few studies have examined the effects of gesta-
t Requests for reprints should be addressed to Linda Patia Spear, Department of Psychology, Box 6000, State University of New York, Binghamton, NY 13902-6000. 423
424
GOODWIN ET AL.
tional cocaine exposure on subsequent maternal behavior during the lactational period (11,22), no studies to date have systematically examined the influence of rearing condition on the neurobehavioral function of offspring exposed prenatally to cocaine. The approach taken to examine this issue in the present study was to compare the behavioral outcome of fostered and nonfostered offspring exposed prenatally to cocaine as well as control offspring subject to the same rearing conditions. In addition, the impact o f potential drug-induced alterations in maternal behavior on subsequent offspring behavioral function was also examined in normal offspring being reared by cocaine-exposed or control dams. The drug exposure regime used in this study has previously been shown to produce slight alterations in maternal behavior (11) as well as neurobehavioral teratogenlc effects in both fostered (4,9, 10,17,21,23,25,26) and nonfostered (25) offspring. Several behavioral assessments were chosen to examine off. spring subject to the various prenatal treatment and rearing conditions. Among the measures used was an olfactory/footshock classical conditioning task that previously has been shown to be sensitive to gestational cocaine exposure early in life (e.g., 9), and hence likely to reveal effects of rearing condition. With this task, the number of trials given during training was manipulated to assess the hypothesis that an increase in the amount of training may allow for the expression of conditioning in cocaine-exposed offspring. Two other classical conditioning tasks were examined: auditory/footshock conditioning in 17-day-old animals and olfactory/footshock conditioning in PI8 animals. In the auditory/footshock conditioning task, animals were trained under two different contexts to assess the potential influence o f varying the saliency of the experimental context on conditioning following gestational cocaine exposure. Finally, because of recent informal reports that children prenatally exposed to cocaine exhibit "unremorseful aggressiveness" in certain situations (8), aggressive behavior was assessed in adult offspring using two different testing p r o c e d u r e s - a n intruder aggression test and shockelicited aggression. For the latter response measure, it was also necessary to assess shock sensitivity thresholds in order to determine whether any potential group differences in the amount of shock-elicited aggression were related to alterations in sensitivity to footshock. METHOD
Subjects, Breeding, and Chronic Drug Treatments The breeding and gestational drug treatment procedures of the rats used in this study are described in greater detail elsewhere (11). Briefly, offspring were derived from SpragueDawley (Charles River) rats bred in our laboratory. Animals were housed in a temperature-controlled colony room with a 12L : 12D cycle with light onset at 0700 h. Following a 2 week acclimation period, dams were placed individually with an adult male at 1700 h and removed the following morning at approximately 0900 h. Detection of a copulatory plug marked day 1 o f gestation (El) at which point dams assigned to the foster (FOS) group continued to be maintained on ad lib access to purina lab chow pellets and water, whereas dams assigned to the prenatal treatment groups were given ad lib access to powdered lab chow and water. On E8 the latter dams were assigned to either cocaine (C40) or lab chow (LC) treatment groups. The C40 and LC dams received a daily SC injection of cocaine hydrochioride (40 mg/kg/3 cc) or saline (3 cc/ kg), respectively, between 1000 and 1200 h from E8-E20. The
date of birth of the pups was designated as postnatal day 0 (P0). On P I , each litter was culled to 8-10 pups, with an attempt to equate sex ratios within each litter whenever possible. Offspring from litters containing less than 8 pups were not examined. Rearing condition. Six experimental groups were established. These included the traditional fostering groups in which C40 and LC pups were fostered by litter to untreated foster dams (FOS/C40; FOS/LC), as well as nonfostered conditions where each dam was allowed to rear her own litter (C40/C40; LC/LC). In addition, pups from foster dams were fostered by litter to C40 and LC dams (C40/FOS; LC/FOS) to examine whether rearing condition influences subsequent behavior of the foster pups.
Assignment of Offspring to Experiments Offspring derived from 10-12 litters within each of the 6 experimental groups were tested. For testing during the preweanling period, assignment to the test conditions was counterbalanced by sex. At P7-8, two males and two females from each litter were tested for odor/footshock conditioning. At P17-18, two different males and females from each litter were tested in the tone/footshock conditioning procedure with the remaining two pups per litter being given odor/footshock conditioning at P18. At P21, males were weaned into hanging cages for testing at P60-70. Of these, one male was housed individually and was used as the resident in a resident-intruder test of aggression while the remaining four males were housed in pairs. One of these males was used as the intruder in a resident-intruder test, two were used in a test of shock-induced fighting, and the remaining male was used to determine shock sensitivity thresholds. For litters in which there were fewer than 10 pups or an unequal sex distribution, slight readjustments of the assignment regimen were necessary. However, under no circumstance was more than one pup per litter placed into any condition or were pups included in more than one condition in any experiment conducted during the preweanling or postweanling period.
General Method All testing was conducted during the light portion of the L : D cycle. Each test was conducted by individuals who were unaware of prenatal history or rearing condition. Following all testing conducted during the preweanling period, pups were returned to the home cage, with pups exposed to odors being held in a holding cage for approximately 1 h following testing to allow odors to dissipate prior to return.
P7-8 Olfactory Conditioning Four pups/litter were used for this procedure with 8-10 pups from each experimental group being assigned to each conditioning group: 2 trials, 3 trials, 4 trials, or explicitly unpaired. Apparatus and procedure. On P8, pups were removed from the dam 30 rain prior to conditioning and placed in a holding cage, maintained at 32°-34°C by heating pads. Conditioning was conducted in two separate Plexigias chambers (9.0 × 15.0 x 10.0 cm) resting on a grid flooring. The grid floor was constructed of 0.1 cm diameter stainless steel rods spaced 0.7 cm apart. Odorants were applied to the surface of absorbent cotton and placed below the grid flooring in both the conditioning and testing apparatus. Each pup in the paired groups (2, 3, or 4 trials) was placed individually in a condition-
A FOSTERING STUDY OF THE EFFECTS: II ing chamber. After an initial 20 s exposure to the C S - (1.0 cc banana extract, Virginia Dare), the pup was removed and placed in a similar chamber for a 20 s exposure to the CS + (0.75 cc lemon oil, Humco Laboratory). Two brief (3 s) footshocks (0.5 mA, measured directly from grids; Grason Stadler shock generator) were administered, the first beginning 7 s after being placed in the chamber, and the second 7 s after termination of the first shock. After termination o f the last footshock, each pup was removed from the grid and returned to the holding cage. This procedure was repeated once for pups in the 2 trial group, twice for the 3 trial group, and three times for the 4 trial group, with each trial being separated by a 1-min intertrial interval (ITI). After completion of the final pairing, each pup was returned to a holding cage for 3 min before testing. Each pup in the explicitly unpaired control group (UP) received 4 trials o f footshock exposure (a total of 8 footshocks, equivalent to the 4 trial paired group) as previously described in the absence of any odor. After completion of the last footshock, the pup was returned to a holding cage for 20 min. Each pup then received a 20-s exposure to the C S (banana) followed by a 20-s exposure to the C S + (lemon). After a l-min ITI, these odor exposures were repeated for 3 more trials (a total of 4 exposures to each odor). Following the last odor exposure, the pups were returned to a holding cage for 3 min before testing. Testing was conducted in a (20.0 × 10.0 × 9.0 cm) rectangular Plexiglas chamber divided into two equal areas. One side of the chamber was scented with the C S + odor (0.75 cc lemon) and the other side was scented with a novel odor (0.75 cc orange, Spectrum Chemical). Each pup was placed individuaily on the midline of the wire grating with the animal's head oriented either directly away or toward the experimenter, with the orientation varying randomly among animals. The amount of time spent on the lemon side was recorded for a test duration of 3 min.
P17-18 Auditory Conditioning Four pups per litter from each of the six experimental groups were used as subjects, with 10-11 pups from each experimental group being randomly placed into each conditioning group: Context Control (CC), UP, and two paired condit i o n s - o n e with a 0 s interstimulus interval (0-ISI) between the conditioned stimulus offset and the unconditioned stimulus onset and another group with a 20 s ISI (20-ISI). Apparatus. In this experiment all subjects received equal exposure to two distinct contexts. The first context "CI" was designed to be less salient to the pups than the second context "C2". The C1 context consisted of a sound insulated chamber which contained a grid floor (38.0 × 30.5 × 7.0 cm) and a clear Plexigias conditioning chamber (28.0 × 20.5 × 25.5 cm) which was partitioned into 2 equal sized compartments and was placed on top of the grid floor. The grid floor was constructed of 0.3 cm diameter stainless steel rods which were spaced 1.0 cm apart. A shock generator (Grason Stadler) provided a 1.0 m A scrambled footshock for 0.5 s. A ventilation fan provided background noise (72 dBA). Dim ambient light within the chamber was provided by a 1.5-W (120-v) fight centered on the right wail 34.0 cm above the grid floor. A fine wire mesh lid was placed over the Plexigias chamber. Clean shavings identical to those in the home cage were placed under the grid floor and changed daily. The C2 context was located in a separate room away from CI. The apparatus in this room
425 was identical to that described in CI except that the Plexigias conditioning chamber was black rather than clear. The context also had the following additions: (a) three frosted 7.5 W light bulbs located 29.0 cm above the grid floor flashed on and off (500 msec on/500 msec off) throughout the experimental session; (b) 1.0 cc of banana odor (Virginia Dare) was spread on a 20.0 x 10.0 cm strip of cotton beneath the grid floor. The CS was a 2,000 Hz pulsing pure tone (500 msec on/500 msec off, 80 dBA) which was delivered from three speakers situated 34.0 cm above the grid floor. The cotton and the odor were changed after each experimental session. Since all animals received the CS in C2 during training, testing was conducted in C2. The test chamber was identical in size to one compartment of the C2 Plexiglas conditioning chamber. On one wail of the testing apparatus were the transmitter and receiver for an ultrasonic activity monitoring device that were spaced 4.0 cm apart from one another and 1.5 crn above the grid floor. Ultrasonic waves emitted by the transmitter were picked up by the receiver, with any movement within the chamber resulting in frequency discrepancies detected by the receiver. Each frequency discrepancy resulted in a closure of an electromechanical relay which was translated into one activity count; thus amount of movement by the animal was positively correlated with the number of activity counts (see 19 for details). The sensitivity of the device was adjusted and periodically checked with the aid of a metronome. Procedure. All pups to be used in the experiment from a given litter were removed from the home cage on PI7 or PI8 and placed individually in holding cages. Subjects in the CC condition were placed in C1 for 64 rain during which time they received 16 presentations of the US (a 0.5 s, 1 m A footshock). Following this exposure, the subjects were immediately taken to C2 where they received 16 presentations of the CS (15 s duration pulsing tone) during a 64 min session. Both the CS and the US were presented on a VI-4 min schedule. In order to equate for the experience of the CC animals, subjects in the UP, 0-ISI, and 20-ISI conditions were given exposure to CI for 64 min during which time no stimuli were presented. Animals then were moved immediately to C2 where they received 16 presentations of the tone and the US. In group UP, the tone CS and the footshock US were both presented in a quasi-random order, each on a separate VI-4 min schedule. Group 0-ISI was given 16 paired presentations of the tone a n d shock with tone offset corresponding to shock onset. Group 20-ISI received similar treatment as did the 0-ISI group with the exception of a 20 s gap between CS offset and US onset. Subjects were individually tested 24 h later in C2. Testing consisted first of placing the pups into the test chamber and measuring activity for 1 min. Following this adaptation period, time to reach 200 activity counts was separately recorded for 2 blocks of 100 activity counts each. This served as a measure of activity prior to the presentation of the CS. After these 200 counts were reached, the tone was presented until 200 more activity counts were recorded, with the time taken to reach the criterion number of counts again being separately recorded for 2 blocks of 100 counts. Following testing, pups were returned to their home cages.
P18 Olfactory Conditioning Apparatus and procedure. The two treatment conditions (paired and unpaired) consisted of 9-12 animals from each experimental group. Prior to any experimental manipulation all subjects were placed in individual holding cages.
426 For conditioning, each subject in the paired condition was placed for 30 s in a clear Plexiglas chamber (10.2 x 15.2 x 20.3 cm) containing the C S - odor (1.0 cc imitation almond extract, McCormick) which was applied to cotton underneath the grid floor in the conditioning chamber. Immediately following this exposure, paired subjects were moved to another room and placed in an identical chamber containing the CS + odor (2.0 cc methyl salicylate, Aldrich Chemicals Co.) for 30 s. During this 30 s, paired subjects received two 1.6 mA scrambled footshocks (Grason Stadler shock generator) during seconds 8-10 and 18-20. Subjects in the unpaired control conditions received two identical footshocks during a 30 s placement in a non-odorized clear Plexiglas chamber 20 rain prior to a 30 s exposure to each of the odor stimuli. The unpaired animals received the odor stimuli in the same order and in the same chambers as the paired subjects. Following the conditioning procedures, subjects were given a 60 s odor preference test between the CS + and a novel odor (1.0 cc lemon oil, Humco). Both odors were applied to cotton which was inserted into two openings, 1.3 mm above the floor, on opposite ends of the testing chamber. For this test, subjects were placed on the midline of the test chamber (27.9 x 10.6 x 12.4 cm) facing a side wail, with the orientation of the animals alternated between animals. The time spent on the methyl salicylate side o f the chamber was recorded. Animals were also given a test to assess freezing in the presence of the CS + . For this test animals were placed in a circular test chamber (25.0 o n diameter, 14.0 cm high, with a black posterboard floor) with 2.0 cc o f methyl salicylate placed on two cotton swabs and taped outside of two openings (4.0 x 8.0 o n ) which were covered with wire mesh and located directly opposite of each other at the level of the chamber floor. The total duration of freezing (defined as total lack of movement) was recorded for three min. Both the freezing and preference tests were given immediately and 3 h after conditioning. The order o f these two tests was counterbalanced such that subjects were either given the freezing or preference tests first at both time intervals. Because extinction may occur during testing, and hence exposure to the first testing situation could influence performance on the next test, the data were initially analyzed for order effects. In the case of a significant order effect, the data from only those subjects that received the test in question first during the initial test session were considered.
Adult Testing Shock induced fighting. Two P60-70 rats from the same litter were chosen to undergo shock-induced fighting procedures based on the proximity of their body weights. The pair of rats was placed in a clear Plexigias chamber (25.4 x 25.4 cm) that was placed on a stainless steel grid floor (0.3 cm diameter grids separated by 0.5 cm). They were allowed to habituate to the chamber and testing room for 5 min before the onset of shock. Subjects received 32, 1.6 mA shocks of 0.5 s duration every 15 s (delivered by a Grason Stadler shock generator). Latency to the first aggressive contact between animals was recorded as well as the number of aggressive attacks throughout the 8 min test period. An aggressive contact was defined as a directed movement toward the opponent which resulted in at least one of the following: sparring, upright attack posture, or a supine submissive posture adopted by the attacked rat (5). Submission was defined as rearing on the hindfeet with the head positioned at an upward angle, the forepaws extended toward the attacking animal and the ventral surface of the body continually facing the opponent (29).
GOODWIN ET AL.
Shock threshold measurement. One P60-70 male animal from each litter was subjected to a series of shocks at the following intensities; 0.1, 0.2, 0.3, 0.5, 1.0, 1.2, 1.3, 1.6, and 2.0 m A (0.5 s duration every 15 s). This series of shocks was delivered in ascending and descending order four consecutive times. The lowest shock intensity at which the rat's hindlimbs flinched was denoted as the threshold intensity for each series o f shocks. These thresholds were averaged for each animal across the 8 shock series (4 ascending, 4 descending) to obtain a mean threshold intensity for each rat. Resident~intruder aggression. One P60-70 male rat from each litter, designated as the resident and housed singly from weaning, was confronted once with a weight-matched group housed littermate in the resident's home cage. For testing, the intruder was placed in the cage with the resident and their behavior was videotaped for 10 min. Both the frequency (total incidence per session) and duration (total amount of time) of two offensive and three defensive behaviors were scored from the videotapes. The behaviors were individually scored for both partners and are described in detail elsewhere (29). Briefly, the two offensive behaviors were: offensive upright attack which was defined as a rapid approach toward the partner usually involving biting; and offensive sideways attack which was the same as an upright attack except that the approach was done laterally with. the head oriented toward the partner. The three defensive behaviors were: defensive upright which was recorded when the rat reared on its hindlegs while using its forepaws to defend against attack; lateral defense which involved rearing on hindlegs with the side facing the opponent and the head turned away; and submission which was defined as rearing on hindiegs with the head positioned at an upward angle and the forepaws extended toward the attacking animal and the ventral surface of the body facing the opponent (29). RESULTS
Maternal~Litter Data Dams in the C40 group exhibited a significant reduction in food intake and body weight gain relative to LC dams, although there were no differences between the two groups in gestational length, pup body weight, litter size, or the ratio of male to female pups (see 11 for presentation of these data).
Offspring Behavioral Measures Preliminary analyses of the preweanling conditioning data revealed no sex differences in the data, so these data were collapsed across that variable prior to< further analysis. Data were analyzed using analysis of variance (ANOVA) tests with a factor of Group comprised of the 6 different groups defined by prenatal treatment and rearing conditions, along with other factors as necessary depending on the design o f each specific subexperiment. Tukey's tests were used to determine the locus of significant main effects and interactions. In addition, Dunn's multiple comparisons (13) were conducted to examine the effects of biological rearing versus fostering per se ( L C / L C and C40/CA0 vs. the other groups), the effects of being reared by a previously cocaine exposed dam (C40/FOS and C40/C40 vs. the other groups), and the effects of prenatal cocaine exposure per se (C40/CA0 and FOS/CA0 vs. the other groups). These planned comparisons are discussed in the results only when they revealed effects not evident in the overall ANOVAs. These planned comparisons were not conducted for the conditioning data in that the de-
427
A FOSTERING STUDY OF THE EFFECTS: II sign of these experiments was such that the critical comparisons consisted o f comparisons within each prenatal treatment/ rearing condition group (i.e., comparing paired vs. unpaired animals).
P7-8 Olfactory Conditioning Time spent on the side of the test apparatus that contained the lemon odor (odor previously paired with footshock) was used as the dependent measure in the analysis of first-order conditioning. A 6 x 4 (Group x Condition [2 trials, 3 trials, 4 trials, and unpaired]) A N O V A conducted on these data reveaied significant main effects of Group, F(5, 196) = 11.36, p < 0.05, and Condition, F(3, 196) = 19.24, p < 0.05, along with a significant Group × Condition interaction, F(15, 196) = 2.24, p < 0.05. Tukey's test revealed no significant differences between the paired (2, 3, and 4 trials) and unpaired C40 pups reared by their biological dam (C40/CA0), reflecting a lack of first-order conditioning in this group following either 2, 3, or 4 training trials (Fig. 1). However, while C40 pups reared by foster dams (FOS/CA0) did not express conditioning when given 2 or 3 training trials, those that received 4 training trials spent significantly less time in the presence of the odor previously paired with footshock than their unpaired counterparts. In contrast, paired pups that received either 2, 3, or 4 conditioning trials from all other prenatal treatment/foster groups (FOS/LC, L C / F O S , L C / L C , and C40/FOS) spent significantly less time in the presence o f the odor previously paired with footshock than unpaired animals, demonstrating significant first-order conditioning. Taken together, these data suggest that prenatal cocaine exposure disrupts this instance of conditioning, a deficit that is even more evident when these pups are reared by dams previously exposed to cocaine.
Pl 7-18 Auditory Conditioning Following the first minute, time to reach 400 activity counts was recorded for four blocks of 100 counts each. The CS onset occurred on completion of block 2, therefore, the first two blocks are called pre-CS blocks and the second two blocks are called the CS blocks. Total time (in seconds) to reach 100 activity counts for each of the first two pre-CS blocks were transformed using a log conversion. ANOVA was then conducted separately for each of the blocks in a 6 x 4 (Group × Condition [CC, UP, 0-ISI, 20-ISI]) design. The A N O V A of block 1 revealed only a significant main effect of Condition, F(3,208) = 7.23, p < 0.001. Subsequent Tukey's tests revealed that animals in the 0-ISI condition took significantly more time to reach 100 activity counts relative to all other conditions. There were no significant main effects or interactions involving Group in block 1. The A_NOVA of block 2 also revealed a significant main effect of Condition, F(3, 208) = 1.44, p < .001, again with no main effects or interactions involving Group. Tukey's tests revealed that animals in both the 0-ISI and 20-ISI conditions exhibited a significantly longer latency to reach 100 activity counts than animals in the context control condition. To estimate the degree of conditioning in the two CS blocks, suppression ratios were calculated using untransformed data from blocks 2, 3, and 4, with the following formula: suppression ratio for first CS block (SR1) = block 2 / (block 2 + block 3); suppression ratio for second CS block (SR2) = block 2/(block 2 + block 4). The ANOVA of SR1 revealed only a significant main effect of Condition, F(3, 208) = 1.05, p < 0.001, with subsequent Tukey's test revealing that animals in the 0-ISI condition had a significantly lower suppression ratio than all other conditions (see Fig. 2). The ANOVA of SR2 revealed no significant effects. These results indicate significant conditioning to the tone for rats
120-
100
80 Z O
Z O iii
6O
[]
UNPAIRED
[]
2 TRIALS
[]
3 TRIALS
[]
4 TRIALS
iii
(J FOS/LC
LC/FOS
LC/LC
PRENATAL
FOS/C40
C40/FOS
C40/C40
TREATMENT
FIG. 1. Mean time in seconds (+ SEM) spent on the lemon side of the test apparatus by P7-8 pups from the various prenatal treatment groups. (FOS = Foster, LC = Saline injected Controls, C40 = Cocaine Group. Groups are designated by maternal treatment and pup origin [e.g., FOS/C40 pups reared by FOS dams]). *Significantly different from UP control, p < 0.05.
428
GOODWIN ET AL.
300
=. TIME TO REACH FIRST 100 PRE-CS COUNTS t
b. TIME TO REACH SECOND 100 PRE-CS COUNTS
140 • 120 • 200
100 • 80-
.=
.= 100
@
60
k4O 20 0 CC
20-1SI
LIP
©. SUPPRESSION RATIO FOR FIRST CS BLOCK
20-1SI
d. SUPPRESSION RATIO FOR SECOND CS BLOCK
0,4.
O p-
0-1SI
CONDITION
CONDITION
0.5
_o
0.3.
0,4.
I0.30.2. uJ n-
(¢
0.2"
0.1 0.1' 0.0 OC
UP
O*ISI
20-1SI
0.0 CC
CONDITION
UP
0-1SI
20-1SI
CONDITION
FIG. 2. Auditory/footshock conditioning in P17-18 pups with means collapsed across rearing and prenatal treatment groups: (a) Mean time ( + SEM) to reach the first 100 activity counts for the four conditioning groups; Co)Mean time (=l=SEM) to reach the second 100 activity counts for the four conditioning groups; (c) Mean suppression ratio for the first CS+ block (+ SEM) for the four conditioning groups; (d) Mean suppression ratio for the second CS + block (+ SEM) for the four conditioning groups. *Significantlydifferent from UP controls, p < 0.05.
given no delay between the tone and footshock although the degree of conditioning was unaffected by prenatal treatment or rearing condition.
P18 Olfactory Conditioning Odor preference. For the odor preference test, a 6 × 2 × 2 × 2 (Group x Condition [paired vs. unpaired] x Order × Time [immediate vs. 3 h]) ANOVA indicated a significant effect of order of testing, F(1, 97) = 5.60,p < 0.05. Animals receiving the preference test after the freezing test did not exhibit conditioning, suggesting that extinction to the C S + may have taken place as a consequence of prior unreinforced exposure to the CS + in the freezing test. These data are consistent with previous findings showing that an acquired aversion in preweanling animals could be extinguished with as little as 30 s exposure to the C S + (e.g., 1). Because of this significant order effect, only those subjects that received the odor preference test first are reported here. A 6 × 2 × 2 (Group × Condition × Time) ANOVA revealed a significant effect of Condition, F(1, 58) = 4.84, p < 0.05 and Time, F ( I , 59) = 22.33, p < 0.001, with paired subjects spending significantly less time over the CS + than unpaired animals and animals in the immediate test spending significantly less time over the CS + than animals in the 3 h test. Although the Time × Condition interaction did not reach sig-
nificance, F(I, 59) = 3.34, p < 0.07, inspection of Fig. 3A indicates that the main effect of Time is primarily the result of an increase in the amount of time spent over the C S + by the paired animals at the 3-h interval. This suggests that at 3 h, the paired animals were showing less of an aversion to the CS + than during the immediate test as would be expected based on prior studies (16). There was no main effect or interaction involving condition; thus, prenatal treatment and rearing history had no effect on performance of this task. Freeze test. A 6 x 2 x 2 x 2 (Group x Condition x Order x Time) ANOVA revealed only a significant effect of Condition, F(1, 93) = 9.69, p < 0.01 with paired subjects spending more time freezing than unpaired animals. As there was no interaction of Time and Condition, indicating similar amounts of freezing in the immediate and three hour tests, it appeared that the freeze test may be a more sensitive indicator of conditioning than the odor preference test. Again, no effect of Condition was observed, demonstrating that this index of conditioning was also unaffected by prenatal treatment and rearing history.
Adult Testing Shock-inducedfighting. A 6 (Group) one-way ANOVA revealed no significant differences in the mean number of fighting episodes, F(5, 58) = 1.028, p > 0.05 or latency to the
A F O S T E R I N G S T U D Y O F T H E E F F E C T S : II
429
a. ODOR PREFERENCE TEST 60
[] PAIRED [] UNPAIRED
4-
¢n to w-r
50
iz
40
0 ¢n E3
30
z
0 W U~ Z ,,=
20
IMMEDIATE
3HOUR
TEST
b.
FREEZE TEST
30
t
Z m
T
0 Z N W
[] []
PAIRED UNPAIRED
+
1.1.182o re I.i.I.I. m 0 aw
0
~ W
Ul W !¢1¢~ Z
~
IMMEDIATE
3 HOUR
TEST FIG, 3. Odor/footshock conditioning in PI8 pups with means collapsed across rearing and prenatal treatment groups: (a) Mean time ( ± SEM) spent on methyl salicylate side of the apparatus in both the immediate and 3 h tests; Co) Mean time ( ± SEM) spent freezing when presented with the CS + in both the immediate and 3 h tests. *Significantly different from UP controls, p < 0.05.
430
GOODWIN ET AL.
first aggression, F(5, 58) = 1.23, p > 0.05 among the different experimental groups. However, planned comparisons between offspring reared by cocaine dams (C40/C40; C40/FOS) and the remainder of the treatment groups (LC/LC; LC/FOS; FOS/LC; FOS/C40) revealed a significant difference in the latency to the first aggressive contact, with offspring being reared by cocaine-exposed dams, regardless of prenatal treatment history, exhibiting shorter attack latencies than offspring from the other experimental groups (see Fig. 4). No significant differences in the mean number of fighting episodes were observed employing the same planned comparisons. Shock threshold measurement. No differences in pain sensitivity to shock were observed between the different experimental groups, in that a 6 (Group) one-way ANOVA of jump/ flinch scores revealed no significant main effect of Group. Resident/intruder aggression. No indication of aggressive behavior was observed in test subjects from any of the experimental groups. This apparent lack of aggression between the resident and intruder rats could be due to the strain of rats used in this study. Although the intruder aggression test has proved useful in examining aggression in Wistar rats (7,29), Sprague-Dawley rats in general are not a particularly aggressive strain as further documented by the lack of aggression in this test.
per se. On other cognitive tests, however, no effects of prenatal cocaine exposure or rearing condition were observed. Thus, although the primary purpose of this study was to examine the interaction between fostering and prenatal drug treatment as reflected in offspring behavior, the data may also provide valuable indications as to the nature of the cognitive deficits produced by gestational cocaine exposure. Deficits were seen in offspring exposed prenatally to cocaine in an odor/footshock classical conditioning task conducted at P7-8. Whereas foster and LC control offspring under all rearing conditions exhibited a significant aversion to the odor paired with footshock following as few as two training trials, C40 offspring reared by foster dams required four training trials to demonstrate conditioning and C40 pups reared by their own dams did not exhibit conditioning even after four training trials. By varying number of training trials in this experiment, these data provide evidence that the previously reported lack of first order conditioning in P7-8 pups exposed gestationally to cocaine (9,24) does not reflect an absolute inability of these animals to learn a classical conditioning task at this age but rather demonstrates that they may require more training to exhibit significant conditioning relative to control age-mates. From the present data it can be concluded that training parameters are critical in determining whether conditioning deficits will be evident in cocaineexposed offspring. Hence, in future cognitive studies with cocaine-exposed offspring it may be prudent to vary degree of training (e.g., by manipulating number of training trials and/ or US or reinforcement intensity) to provide a sufficiently broad range of associative strength to avoid potential false negatives.
DISCUSSION
In the present study, fostering was observed to influence subsequent offspring neurobehavioral function in two testing situations, with one of these tests also revealing alterations in cognitive function associated with prenatal cocaine exposure
400 "
))
300 A
O UJ u) v
>,. 0 z l.u I,,~
200
_I
1 O0
FOS/LC
LC/FOS
LC/LC
FOS/C40
C40/FOS
C40/C40
GROUP
FIG. 4. Latency (:t:SEM) to the first aggressive contact in P60-70 offspring. *Groups significantly different from the other groups using planned comparisons.
A FOSTERING STUDY OF THE EFFECTS: II Task complexity (and hence, degree of difficulty) may also be an important factor in revealing conditioning deficits in cocaine-treated offspring and this may in turn vary with age. For instance, whereas cocaine-exposed pups given two training trials (with two shock exposures per training trial) did not exhibit odor/footshock conditioning at P7-8 (present results; also ref. 9), C40 pups given similar training at P12 or P21 did exhibit a significant aversion for the odor paired with footshock (9). These findings are reminiscent of those from the present study where C40 offspring at PI8 were observed to develop a significant odor/footshock aversion even though the odor was paired with footshock in only one training trial (during which two brief footshocks were given). Thus, by 1221 days of age, cocaine-exposed offspring appear to be able to learn a simple first order odor/footshock classical conditioning task even when only a limited amount of training is provided. Similarly, the training parameters used in the present study were sufficient to support significant auditory/ footshock conditioning in C40 offspring tested at P17-18. However, when task complexity is increased, conditioning deficits are evident in C40 offspring in this older age range. For instance, deficits in higher order conditioning have been observed in C40 pups at P12 using a sensory preconditioning task (9). Data from the present experiment also suggest that cocaine-exposed offspring are as sensitive to context as control offspring during the preweanling period. In the PI 8 auditory/ footshock conditioning experiment, greater conditioning to the context was observed when the CS (tone) and US (footshock) were paired in that context than when they were not, and there was no effect of prenatal treatment or rearing condition in this regard. Normal adult rats exhibit the opposite effect, akin to overshadowing--poorer conditioning to the context when the CS and US are paired in that context than when they are not (e.g., 14). Normal preweanling rats, however, have been shown in other experiments (15) to exhibit, instead, potentiation of conditioning to the context by pairings of the CS and US, like the present results. This is consistent with the general tendency for less potentiation and more overshadowing observed with development from infancy to adulthood (e.g., 27,28). Given that there were no significant effects of prenatal treatment (or rearing condition) in these analyses, it appeared that prenatal exposure to cocaine does not influence this tendency to show potentiation during the preweanling period, suggesting that C40 animals are as sensitive to context as are control offspring. Another issue addressed by these data is with regard to the relationship between prenatal cocaine exposure and offspring aggressive behavior. Whereas there are clinical reports that children prenatally exposed to cocaine are more aggressive than non-exposed agemates (8), in the present study prenatal exposure to cocaine was not observed to have an influence on aggressive behavior in adulthood with the response measures used. However, planned comparisons revealed that being reared by a previously cocaine-exposed dam was associated with a slightly decreased latency to aggressive attack in adult offspring in a shock-elicited aggression task. This difference with rearing condition was seen in the absence of any alteration in sensitivity to footshock per se, thus it appears to reflect an alteration in the propensity to elicit aggressive behavior in response to this stressful situation. It is not clear what aspect of the maternal behavior of dams previously exposed to cocaine results in increased aggressive behavior in the offspring they reared. However, it is interesting that prior cocaine exposure has been observed to increase maternal ag-
431 gression toward a female intruder without influencing time spent in the nest or pup retrieval latencies (11). This raises the possibility that increased maternal aggression may be reflected later in life by an increase in aggressive behavior in the offspring reared by these dams. However, note that the increase in intruder aggression seen in previously cocaine-exposed dams was only observed when they were rearing their own pups, with dams rearing their biological pups, regardless of prior drug exposure, exhibiting shorter latencies to initiate aggressive attacks (11). Whether dams previously exposed to cocaine might exhibit increased aggressive behavior in other circumstances (e.g., toward the pups during the process of weaning) that could influence the latter propensity for aggression in their offspring is presently unknown. It is also possible that some presently unspecified alteration in maternal behavior apart from the alteration in maternal aggression induced by prior cocaine exposure in the dams may produce this effect in the offspring. Finally it should be mentioned that whatever differences in maternal behavior led to the differences in the latency to attack of the offspring, it is possible that these effects may not be related to the prior cocaine exposure of the dams per se but rather due to their failure to gain normal amounts of weight during pregnancy. Pair-fed controls were not included in this experiment due to the already complex nature of the design necessary to examine the fostering issue. Although the reduction in maternal weight gain associated with cocaine exposure is relatively modest (approximately 10°70) and has not been observed previously to result in neurobehavioral alterations in pair-fed offspring (e.g., 24), it is nevertheless possible that this attenuation in weight gain could contribute to the findings observed. The final issue to be discussed regards the importance of fostering in behavioral teratology studies. In the present study, rearing condition and prenatal cocaine exposure were both observed to have an influence on the neurobehavioral function of offspring. When compared with control offspring, offspring exposed prenatally to cocaine exhibited deficits in odor/footshock conditioning at P7-8, with such deficits being more pronounced when cocaine-exposed offspring were reared by their previously cocaine-exposed dams than when reared by untreated foster dams. On another response measure, shock-elicited aggression; planned comparisons revealed that offspring reared by previously cocaine-exposed dams exhibited a decreased latency to the first aggressive contact in the absence of any alteration in shock sensitivity thresholds; this modest effect was seen regardless of whether the offspring had been exposed gestationally to cocaine. Thus, being reared by a dam previously exposed to cocaine was observed to have an impact on offspring behavioral function apart from the effects of the offspring's prenatal exposure to cocaine. It is not currently possible to generate a true estimate of the prevalence of such indirect maternally-mediated effects of cocaine. Given the limited number of measures examined in offspring in this study, it would be imprudent to attempt to use this measurement sample to generalize to the overall rate of such indirect effects of rearing condition, other than documenting that they can indeed occur. Moreover, although in this study the effect of prenatal cocaine exposure on the offspring was exacerbated in pups reared by their own dams, there is no clear a priori reason to assume that this would always be the case. It is also seemingly possible that alterations in maternal behavior induced by prior cocaine treatment during pregnancy could attenuate, rather than potentiate, manifestations of prenatal cocaine exposure per se on the offspring. This study demonstrates that rearing condition does
432
G O O D W I N E T AL.
exert a detectable influence, albeit modest on offspring neurobehavioral function following prenatal cocaine exposure. Without fostering, it would appear that there is a significant potential for false positives (and possibly even false negatives) when attributing offspring effects to the consequences o f gestational cocaine exposure per se. Whether or not fostering should be included in the design of experiments in developmental toxicology is dependent in part on the research questions being asked, along with the relative weighting o f the costs o f fostering versus the necessary ambiguity associated with interpretation of data derived from studies where pups were reared by their own dams. If the objective is to determine the consequences o f prenatal exposure per se on offspring function, the results o f this work support the importance o f fostering, at least in studies with
cocaine. Given current indications that the majority o f human infants exposed gestationally to cocaine are reared by individuais other than their biological mothers (6), the use o f such fostering may also resemble the clinical situation. If offspring are not fostered, the results o f this study indicate it cannot be tacitly assumed that the observed effects are prenatal in origin, as postnatal maternal influences may exacerbate or induce offspring alterations beyond those produced by the prenatal treatments per se. ACKNOWLEDGEMENT This research was supported by National Institute on Drug Abuse Grants R01 DA14478 and K02 DA00140 to Linda P. Spear. We gratefully acknowledge the assistance of Yolanda Pinzas for technical support.
REFERENCES 1. Arnold, H. M.; Spear, N. E. Order and duration of stimuli are important determinants of reactivation (submitted). 2. Birke, L. I. A.; Sadier, D. Maternal behavior in rats and the effects of neonatal progestins given to the pups. Dev. Psychobiol. 18:467-475; 1985. 3. Cierpial, M. A.; McCarty, R. Hypertension in SHR rats: Contribution of the maternal environment. A. J. Physiol. 235:H980H984; 1987. 4. Clow, D. W.; Hammer, R. P.; Kirstein, C. L; Spear, L. P. Gestational cocaine exposure increases opiate receptor binding in weanling offspring. Dev. Brain Res. 59:179-185; 1991. 5. Cuadra, G. R.; Molina, V. A. Antidepressants reverse the inhibition of shock-induced aggression elicited by a prior inescapable shock. Pharmacol. Biochem. Beh. 40:69-73; 1991. 6. Davis, E. Characteristics of "exposed" fetus and newborn human: Characteristics of the infant. Presentation at Conference entitled: Cocaine in pregnancy: Is the fetus affected? (Sponsored by the National Institute on Drug Abuse, and Office of Substance Abuse Prevention), Washington, DC; 1991. 7. Fernt~ndez-Espejo, E.; Mir, D. Ethological analysis of the male rat's socioagonlstic behavior in a resident-intrnder paradigm. Aggressive Beh. 16:41-55; 1990. 8. Field, T. Behavioral consequences. Presentation at Conference entitled: Cocaine in pregnancy: Is the fetus affected? (Sponsored by the National Institute on Drug Abuse, and Office of Substance Abuse Prevention), Washington, DC; 1991. 9. Heyser, C. J.; Chen, W. J.; Miller, J.; Spear, N. E.; Spear, L. P. Prenatal cocaine exposure induces deficits in Pavlovian conditioning and sensory preconditioning among infant rat pups. Behav. Nenrosci. 104:955-963; 1990. 10. Heyser, C. 3.; Miller, J. S.; Spear, N. E.; Spear, L. P. Prenatal exposure to cocaine disrupts cocaine-induced conditioned place preference in rats. Neurotoxicol. Teratol. 14:57-64; 1992. 11. Heyser, C. J.; Molina, V. A.; Spear, L. P. A fostering study of the effects of prenatal cocaine exposure: I. maternal behaviors. Neurotoxicol. Teratol. 14:415-421; 1992. 12. Hutchings, D. E. Issues of methodology and interpretation in clinical and animal behavioral teratology studies. Neurobehav. Toxicol. Teratol. 7:639-642; 1985. 13. Kirk, R. E. Experimental design: Procedures for the behavioral sciences. Belmont, CA: Brooks/Cole; 1982. 14. Marlin, N. A. Within-compound associations between the context and the conditioned stimulus. Learn. Mot. 13:526-541; 1982. 15. McKinzie, D. L. Ontogenetic differences in the expression of context and CS conditioning as a function of context saliency and CS-UCS interval. Binghamton, NY: State University of New York; 1992. Thesis. 16. Miller, J. S.; Jagielo, J. A.; Spear, N. E. Differential effectivehess of various prior cuing treatments on the reactivation and
maintenance of memory. J. Exp. Psych.: Animal Behav. Process. 17:249-258; 1991. 17. Moody, C. A.; Frambes, N. A.; Spear, L. P. Psychopharmacological responsiveness to the dopamine agonist quinpirole in norreal weanlings and in weanling offspring exposed gestationally to cocaine. Psychopharm. 108:256-262; 1992. 18. Moore, C. L; Power, K. L. Prenatal stress affects mother-infant interaction in norway rats. Dev. Psychobiol. 19:235-245; 1986. 19. Moye, T.; Rudy, J. Ontogenesis of learning: VI., learned and unlearned responses to visual stimulation in the infant hooded rat. Devel. Psychobio. 18:395-409; 1985. 20. Pearson, D. E.; Teicher, M. H.; Shaywitz, B. A.; Cohen, D. J.; Young, J. G.; Anderson, G. M. Environmental influences on body weight and behavior in developing rats after neonatal 6hydroxydopamine. Science 209:715-717; 1980. 21. Scalzo, F. M.; All, S. F.; Frambes, N. A.; Spear, L. P. Weanling rats exposed prenatally to cocaine exhibit an increase in striatal D2 binding associated with an increase ligand affinity. Pharmacol. Biochem. Behav. 37:371-373; 1990. 22. Sobrian,, S. K.; Burton, L. E.; Robinson, N. L.; Ashe, W. K.; James, H.; Stokes, D. L.; Turner, L. M. Nenrobehavioral and immunological effects of prenatal cocaine exposure in rat. Pharmacol. Biochem. Behav. 35:617-629; 1990. 23. Spear, L. P.; Heyser, C. J. Cocaine and the developing nervous system: Laboratory findings. In: Zagnn, I. S.; Slotkin, T. A., eds. Maternal substance abuse and the developing nervous system. Orlando, FL: Academic Press; 1992:155-175. 24. Spear, L. P.; Kirstein, C. L.; Bell, J.; Yoottanasumpun, V.; Greenbanm, R.; O'Shea, J.; Hoffman, H.; Spear, N. E. Effects of prenatal cocaine exposure on behavior during the early postnatal period. Nenrotoxicol. Teratol. 11:57-63; 1989. 25. Spear, L. P.; Kirstein, C. L.; Frambes, N. A. Cocaine effects on the developing central nervous system: Behavioral, psychopharmacologicai, and neurochemical studies. Ann. NY Acad. Sci. 526:290-307; 1989. 26. Spear, L. P.; Kirstein, C. L.; Frambes, N. A.; Moody, C. A. Nenrobehavioral teratogenlcity of gestational cocaine exposure. NIDA Res. Monogr. 95:232-238; 1990. 27. Spear, N. E.; Kucharski, D. Ontogenetic differences in stimulus selection during conditioning. In: R. Kail; N. E. Spear, eds. Comparative perspectives on the development of memory. Hillsdale, N J: Erlbanm. 227-252; 1984. 28. Turkewitz, G.; Mellon, R. C. Dynamic organization of intersensory function. Can. J. Psych. 43:286-301; 1989. 29. Vergnes, M.; Depanlis, A.; Boehrer, A. Parachlorophenylalanine-induced serotonln depletion increases offensive but not defensive aggression in male rats. Physiol. Behav. 36:653-658; 1986. 30. Vorhees, C. V. A fostering/crossfostering analysis of the effects of prenatal ethanol exposure in a liquid diet on offspring development and behavior in rats. Neurotoxicol. Teratol. 11:115-120; 1989.
0892-0362/92$5.00 + .00 Copyright©1992PergamonPressLtd.
Printedin the U.S.A. All fightsreserved.
A Fostering Study of the Effects of Prenatal Cocaine Exposure: II. Offspring Behavioral Measures G R E G O R Y A . G O O D W I N , C H A R L E S J. H E Y S E R , C A R O L E A. M O O D Y , L A V A N Y A R A J A C H A N D R A N , V I C T O R A. M O L I N A , H. M O O R E A R N O L D , D A V I D L. M C K I N Z I E , N O R M A N E. S P E A R A N D L I N D A P A T I A S P E A R 1
Department of Psychology, Center for Developmental Psychobiology, State o f University o f New York at Binghamton, Binghamton, N Y 13902 Received 7 A p r i l 1992; Accepted 8 September 1992 GOODWIN, G. A., C. J. HEYSER, C. A. MOODY, L. RAJACHANDRAN, V. A. MOLINA, H. M. ARNOLD, D. L. MCKINZIE, N. E. SPEAR AND L. P. SPEAR. A fostering study of the effects of prenatal cocaineexposure: II. Offspring behavioral measures. NEUROTOX TERATOL 14(6) 423-432, 1992.--The impact of rearing condition was assessed in Spragne-Dawley dams given 40 mg/kg cocaine (C40) or saline (LC control) subcutaneously (SC) from gestational days 8-20 and their offspring. Treated pups reared by their biological dams (LC/LC; C40/C40), treated pups reared by surrogate dams (FOS/LC; FOS/C40), and foster pups raised by treated dams (LC/FOS; C40/FOS) were examined. On postnatal day 7 (P7), pups received either 0 (unpaired) 2, 3, or 4 pairings of an odor and footshock and were tested for their aversion to this odor. Foster and LC pups, regardless of rearing condition, exhibited significant odor aversions following either 2, 3, or 4 training trials. In contrast, C40 pups reared by surrogate dams required 4 trials to acquire the aversion, and C40 pups reared by their own dams did not exhibit conditioning even after 4 trials. At PI7, no differences were seen among the groups in the aversion formed to an auditory or an olfactory stimulus that was paired with footshock. At P60, shockelicited aggression among pairs of siblings was examined. Regardless of prenatal exposure condition, offspring reared by dams given cocaine showed a decreased latency to the first aggressivecontact, an effect that was evident without any alteration in shock sensitivity. Together these data suggest that being reared by a dam previously exposed to cocaine has an impact on offspring behavioral function apart from the effects of prenatal cocaine exposure per se. The implications of the data regarding the cognitive performance of pups exposed prenatally to cocaine are also discussed.
PRIOR work has shown that dams exposed to cocaine during pregnancy differ only slightly from control dams on measures of maternal behavior. Dams exposed during pregnancy to cocaine were not observed to differ from control dams in pup retrieval latency, nest building behavior, time spent in the nest, or time spent suckling pups (11,22). However, cocaineexposed dams rearing their own pups exhibited an increase in the number of aggressive attacks against a female intruder relative to all other groups of dams, an effect that was not evident in cocaine-exposed dams rearing normal pups (11). Other aspects of maternal behavior (e.g., maternal licking) were not examined in these studies, and hence it is not clear whether cocaine-induced disruptions in maternal behavior are restricted to alterations in aggressive behavior, or whether other aspects of maternal behavior that were not assessed might also be affected. Alterations in maternal behavior can have a notable impact
on later offspring physiological and behavioral function (e.g., see 2,3,18,20). To avoid this potential consequence of insultinduced alterations in maternal behavior, researchers in the area of developmental toxicology have frequently used a surrogate fostering procedure. Such fostering essentially eliminates the possibility that offspring neurobehavioral effects are a result of insult-induced alterations in the dams' behavior per se (e.g., see 12 for discussion). However, surrogate fostering is costly in terms of animal usage, and the necessity for routine fostering in developmental toxicological studies has been questioned (e.g., see 30). Thus, in studies with any particular drug agent, it would appear prudent to assess whether potential drng-induced disruptions in maternal behavior have a sufficiently salient effect on later offspring neurobehavioral function to warrant the routine use of surrogate fostering in studies with that substance. Although a few studies have examined the effects of gesta-
t Requests for reprints should be addressed to Linda Patia Spear, Department of Psychology, Box 6000, State University of New York, Binghamton, NY 13902-6000. 423
424
GOODWIN ET AL.
tional cocaine exposure on subsequent maternal behavior during the lactational period (11,22), no studies to date have systematically examined the influence of rearing condition on the neurobehavioral function of offspring exposed prenatally to cocaine. The approach taken to examine this issue in the present study was to compare the behavioral outcome of fostered and nonfostered offspring exposed prenatally to cocaine as well as control offspring subject to the same rearing conditions. In addition, the impact o f potential drug-induced alterations in maternal behavior on subsequent offspring behavioral function was also examined in normal offspring being reared by cocaine-exposed or control dams. The drug exposure regime used in this study has previously been shown to produce slight alterations in maternal behavior (11) as well as neurobehavioral teratogenlc effects in both fostered (4,9, 10,17,21,23,25,26) and nonfostered (25) offspring. Several behavioral assessments were chosen to examine off. spring subject to the various prenatal treatment and rearing conditions. Among the measures used was an olfactory/footshock classical conditioning task that previously has been shown to be sensitive to gestational cocaine exposure early in life (e.g., 9), and hence likely to reveal effects of rearing condition. With this task, the number of trials given during training was manipulated to assess the hypothesis that an increase in the amount of training may allow for the expression of conditioning in cocaine-exposed offspring. Two other classical conditioning tasks were examined: auditory/footshock conditioning in 17-day-old animals and olfactory/footshock conditioning in PI8 animals. In the auditory/footshock conditioning task, animals were trained under two different contexts to assess the potential influence o f varying the saliency of the experimental context on conditioning following gestational cocaine exposure. Finally, because of recent informal reports that children prenatally exposed to cocaine exhibit "unremorseful aggressiveness" in certain situations (8), aggressive behavior was assessed in adult offspring using two different testing p r o c e d u r e s - a n intruder aggression test and shockelicited aggression. For the latter response measure, it was also necessary to assess shock sensitivity thresholds in order to determine whether any potential group differences in the amount of shock-elicited aggression were related to alterations in sensitivity to footshock. METHOD
Subjects, Breeding, and Chronic Drug Treatments The breeding and gestational drug treatment procedures of the rats used in this study are described in greater detail elsewhere (11). Briefly, offspring were derived from SpragueDawley (Charles River) rats bred in our laboratory. Animals were housed in a temperature-controlled colony room with a 12L : 12D cycle with light onset at 0700 h. Following a 2 week acclimation period, dams were placed individually with an adult male at 1700 h and removed the following morning at approximately 0900 h. Detection of a copulatory plug marked day 1 o f gestation (El) at which point dams assigned to the foster (FOS) group continued to be maintained on ad lib access to purina lab chow pellets and water, whereas dams assigned to the prenatal treatment groups were given ad lib access to powdered lab chow and water. On E8 the latter dams were assigned to either cocaine (C40) or lab chow (LC) treatment groups. The C40 and LC dams received a daily SC injection of cocaine hydrochioride (40 mg/kg/3 cc) or saline (3 cc/ kg), respectively, between 1000 and 1200 h from E8-E20. The
date of birth of the pups was designated as postnatal day 0 (P0). On P I , each litter was culled to 8-10 pups, with an attempt to equate sex ratios within each litter whenever possible. Offspring from litters containing less than 8 pups were not examined. Rearing condition. Six experimental groups were established. These included the traditional fostering groups in which C40 and LC pups were fostered by litter to untreated foster dams (FOS/C40; FOS/LC), as well as nonfostered conditions where each dam was allowed to rear her own litter (C40/C40; LC/LC). In addition, pups from foster dams were fostered by litter to C40 and LC dams (C40/FOS; LC/FOS) to examine whether rearing condition influences subsequent behavior of the foster pups.
Assignment of Offspring to Experiments Offspring derived from 10-12 litters within each of the 6 experimental groups were tested. For testing during the preweanling period, assignment to the test conditions was counterbalanced by sex. At P7-8, two males and two females from each litter were tested for odor/footshock conditioning. At P17-18, two different males and females from each litter were tested in the tone/footshock conditioning procedure with the remaining two pups per litter being given odor/footshock conditioning at P18. At P21, males were weaned into hanging cages for testing at P60-70. Of these, one male was housed individually and was used as the resident in a resident-intruder test of aggression while the remaining four males were housed in pairs. One of these males was used as the intruder in a resident-intruder test, two were used in a test of shock-induced fighting, and the remaining male was used to determine shock sensitivity thresholds. For litters in which there were fewer than 10 pups or an unequal sex distribution, slight readjustments of the assignment regimen were necessary. However, under no circumstance was more than one pup per litter placed into any condition or were pups included in more than one condition in any experiment conducted during the preweanling or postweanling period.
General Method All testing was conducted during the light portion of the L : D cycle. Each test was conducted by individuals who were unaware of prenatal history or rearing condition. Following all testing conducted during the preweanling period, pups were returned to the home cage, with pups exposed to odors being held in a holding cage for approximately 1 h following testing to allow odors to dissipate prior to return.
P7-8 Olfactory Conditioning Four pups/litter were used for this procedure with 8-10 pups from each experimental group being assigned to each conditioning group: 2 trials, 3 trials, 4 trials, or explicitly unpaired. Apparatus and procedure. On P8, pups were removed from the dam 30 rain prior to conditioning and placed in a holding cage, maintained at 32°-34°C by heating pads. Conditioning was conducted in two separate Plexigias chambers (9.0 × 15.0 x 10.0 cm) resting on a grid flooring. The grid floor was constructed of 0.1 cm diameter stainless steel rods spaced 0.7 cm apart. Odorants were applied to the surface of absorbent cotton and placed below the grid flooring in both the conditioning and testing apparatus. Each pup in the paired groups (2, 3, or 4 trials) was placed individually in a condition-
A FOSTERING STUDY OF THE EFFECTS: II ing chamber. After an initial 20 s exposure to the C S - (1.0 cc banana extract, Virginia Dare), the pup was removed and placed in a similar chamber for a 20 s exposure to the CS + (0.75 cc lemon oil, Humco Laboratory). Two brief (3 s) footshocks (0.5 mA, measured directly from grids; Grason Stadler shock generator) were administered, the first beginning 7 s after being placed in the chamber, and the second 7 s after termination of the first shock. After termination o f the last footshock, each pup was removed from the grid and returned to the holding cage. This procedure was repeated once for pups in the 2 trial group, twice for the 3 trial group, and three times for the 4 trial group, with each trial being separated by a 1-min intertrial interval (ITI). After completion of the final pairing, each pup was returned to a holding cage for 3 min before testing. Each pup in the explicitly unpaired control group (UP) received 4 trials o f footshock exposure (a total of 8 footshocks, equivalent to the 4 trial paired group) as previously described in the absence of any odor. After completion of the last footshock, the pup was returned to a holding cage for 20 min. Each pup then received a 20-s exposure to the C S (banana) followed by a 20-s exposure to the C S + (lemon). After a l-min ITI, these odor exposures were repeated for 3 more trials (a total of 4 exposures to each odor). Following the last odor exposure, the pups were returned to a holding cage for 3 min before testing. Testing was conducted in a (20.0 × 10.0 × 9.0 cm) rectangular Plexiglas chamber divided into two equal areas. One side of the chamber was scented with the C S + odor (0.75 cc lemon) and the other side was scented with a novel odor (0.75 cc orange, Spectrum Chemical). Each pup was placed individuaily on the midline of the wire grating with the animal's head oriented either directly away or toward the experimenter, with the orientation varying randomly among animals. The amount of time spent on the lemon side was recorded for a test duration of 3 min.
P17-18 Auditory Conditioning Four pups per litter from each of the six experimental groups were used as subjects, with 10-11 pups from each experimental group being randomly placed into each conditioning group: Context Control (CC), UP, and two paired condit i o n s - o n e with a 0 s interstimulus interval (0-ISI) between the conditioned stimulus offset and the unconditioned stimulus onset and another group with a 20 s ISI (20-ISI). Apparatus. In this experiment all subjects received equal exposure to two distinct contexts. The first context "CI" was designed to be less salient to the pups than the second context "C2". The C1 context consisted of a sound insulated chamber which contained a grid floor (38.0 × 30.5 × 7.0 cm) and a clear Plexigias conditioning chamber (28.0 × 20.5 × 25.5 cm) which was partitioned into 2 equal sized compartments and was placed on top of the grid floor. The grid floor was constructed of 0.3 cm diameter stainless steel rods which were spaced 1.0 cm apart. A shock generator (Grason Stadler) provided a 1.0 m A scrambled footshock for 0.5 s. A ventilation fan provided background noise (72 dBA). Dim ambient light within the chamber was provided by a 1.5-W (120-v) fight centered on the right wail 34.0 cm above the grid floor. A fine wire mesh lid was placed over the Plexigias chamber. Clean shavings identical to those in the home cage were placed under the grid floor and changed daily. The C2 context was located in a separate room away from CI. The apparatus in this room
425 was identical to that described in CI except that the Plexigias conditioning chamber was black rather than clear. The context also had the following additions: (a) three frosted 7.5 W light bulbs located 29.0 cm above the grid floor flashed on and off (500 msec on/500 msec off) throughout the experimental session; (b) 1.0 cc of banana odor (Virginia Dare) was spread on a 20.0 x 10.0 cm strip of cotton beneath the grid floor. The CS was a 2,000 Hz pulsing pure tone (500 msec on/500 msec off, 80 dBA) which was delivered from three speakers situated 34.0 cm above the grid floor. The cotton and the odor were changed after each experimental session. Since all animals received the CS in C2 during training, testing was conducted in C2. The test chamber was identical in size to one compartment of the C2 Plexiglas conditioning chamber. On one wail of the testing apparatus were the transmitter and receiver for an ultrasonic activity monitoring device that were spaced 4.0 cm apart from one another and 1.5 crn above the grid floor. Ultrasonic waves emitted by the transmitter were picked up by the receiver, with any movement within the chamber resulting in frequency discrepancies detected by the receiver. Each frequency discrepancy resulted in a closure of an electromechanical relay which was translated into one activity count; thus amount of movement by the animal was positively correlated with the number of activity counts (see 19 for details). The sensitivity of the device was adjusted and periodically checked with the aid of a metronome. Procedure. All pups to be used in the experiment from a given litter were removed from the home cage on PI7 or PI8 and placed individually in holding cages. Subjects in the CC condition were placed in C1 for 64 rain during which time they received 16 presentations of the US (a 0.5 s, 1 m A footshock). Following this exposure, the subjects were immediately taken to C2 where they received 16 presentations of the CS (15 s duration pulsing tone) during a 64 min session. Both the CS and the US were presented on a VI-4 min schedule. In order to equate for the experience of the CC animals, subjects in the UP, 0-ISI, and 20-ISI conditions were given exposure to CI for 64 min during which time no stimuli were presented. Animals then were moved immediately to C2 where they received 16 presentations of the tone and the US. In group UP, the tone CS and the footshock US were both presented in a quasi-random order, each on a separate VI-4 min schedule. Group 0-ISI was given 16 paired presentations of the tone a n d shock with tone offset corresponding to shock onset. Group 20-ISI received similar treatment as did the 0-ISI group with the exception of a 20 s gap between CS offset and US onset. Subjects were individually tested 24 h later in C2. Testing consisted first of placing the pups into the test chamber and measuring activity for 1 min. Following this adaptation period, time to reach 200 activity counts was separately recorded for 2 blocks of 100 activity counts each. This served as a measure of activity prior to the presentation of the CS. After these 200 counts were reached, the tone was presented until 200 more activity counts were recorded, with the time taken to reach the criterion number of counts again being separately recorded for 2 blocks of 100 counts. Following testing, pups were returned to their home cages.
P18 Olfactory Conditioning Apparatus and procedure. The two treatment conditions (paired and unpaired) consisted of 9-12 animals from each experimental group. Prior to any experimental manipulation all subjects were placed in individual holding cages.
426 For conditioning, each subject in the paired condition was placed for 30 s in a clear Plexiglas chamber (10.2 x 15.2 x 20.3 cm) containing the C S - odor (1.0 cc imitation almond extract, McCormick) which was applied to cotton underneath the grid floor in the conditioning chamber. Immediately following this exposure, paired subjects were moved to another room and placed in an identical chamber containing the CS + odor (2.0 cc methyl salicylate, Aldrich Chemicals Co.) for 30 s. During this 30 s, paired subjects received two 1.6 mA scrambled footshocks (Grason Stadler shock generator) during seconds 8-10 and 18-20. Subjects in the unpaired control conditions received two identical footshocks during a 30 s placement in a non-odorized clear Plexiglas chamber 20 rain prior to a 30 s exposure to each of the odor stimuli. The unpaired animals received the odor stimuli in the same order and in the same chambers as the paired subjects. Following the conditioning procedures, subjects were given a 60 s odor preference test between the CS + and a novel odor (1.0 cc lemon oil, Humco). Both odors were applied to cotton which was inserted into two openings, 1.3 mm above the floor, on opposite ends of the testing chamber. For this test, subjects were placed on the midline of the test chamber (27.9 x 10.6 x 12.4 cm) facing a side wail, with the orientation of the animals alternated between animals. The time spent on the methyl salicylate side o f the chamber was recorded. Animals were also given a test to assess freezing in the presence of the CS + . For this test animals were placed in a circular test chamber (25.0 o n diameter, 14.0 cm high, with a black posterboard floor) with 2.0 cc o f methyl salicylate placed on two cotton swabs and taped outside of two openings (4.0 x 8.0 o n ) which were covered with wire mesh and located directly opposite of each other at the level of the chamber floor. The total duration of freezing (defined as total lack of movement) was recorded for three min. Both the freezing and preference tests were given immediately and 3 h after conditioning. The order o f these two tests was counterbalanced such that subjects were either given the freezing or preference tests first at both time intervals. Because extinction may occur during testing, and hence exposure to the first testing situation could influence performance on the next test, the data were initially analyzed for order effects. In the case of a significant order effect, the data from only those subjects that received the test in question first during the initial test session were considered.
Adult Testing Shock induced fighting. Two P60-70 rats from the same litter were chosen to undergo shock-induced fighting procedures based on the proximity of their body weights. The pair of rats was placed in a clear Plexigias chamber (25.4 x 25.4 cm) that was placed on a stainless steel grid floor (0.3 cm diameter grids separated by 0.5 cm). They were allowed to habituate to the chamber and testing room for 5 min before the onset of shock. Subjects received 32, 1.6 mA shocks of 0.5 s duration every 15 s (delivered by a Grason Stadler shock generator). Latency to the first aggressive contact between animals was recorded as well as the number of aggressive attacks throughout the 8 min test period. An aggressive contact was defined as a directed movement toward the opponent which resulted in at least one of the following: sparring, upright attack posture, or a supine submissive posture adopted by the attacked rat (5). Submission was defined as rearing on the hindfeet with the head positioned at an upward angle, the forepaws extended toward the attacking animal and the ventral surface of the body continually facing the opponent (29).
GOODWIN ET AL.
Shock threshold measurement. One P60-70 male animal from each litter was subjected to a series of shocks at the following intensities; 0.1, 0.2, 0.3, 0.5, 1.0, 1.2, 1.3, 1.6, and 2.0 m A (0.5 s duration every 15 s). This series of shocks was delivered in ascending and descending order four consecutive times. The lowest shock intensity at which the rat's hindlimbs flinched was denoted as the threshold intensity for each series o f shocks. These thresholds were averaged for each animal across the 8 shock series (4 ascending, 4 descending) to obtain a mean threshold intensity for each rat. Resident~intruder aggression. One P60-70 male rat from each litter, designated as the resident and housed singly from weaning, was confronted once with a weight-matched group housed littermate in the resident's home cage. For testing, the intruder was placed in the cage with the resident and their behavior was videotaped for 10 min. Both the frequency (total incidence per session) and duration (total amount of time) of two offensive and three defensive behaviors were scored from the videotapes. The behaviors were individually scored for both partners and are described in detail elsewhere (29). Briefly, the two offensive behaviors were: offensive upright attack which was defined as a rapid approach toward the partner usually involving biting; and offensive sideways attack which was the same as an upright attack except that the approach was done laterally with. the head oriented toward the partner. The three defensive behaviors were: defensive upright which was recorded when the rat reared on its hindlegs while using its forepaws to defend against attack; lateral defense which involved rearing on hindlegs with the side facing the opponent and the head turned away; and submission which was defined as rearing on hindiegs with the head positioned at an upward angle and the forepaws extended toward the attacking animal and the ventral surface of the body facing the opponent (29). RESULTS
Maternal~Litter Data Dams in the C40 group exhibited a significant reduction in food intake and body weight gain relative to LC dams, although there were no differences between the two groups in gestational length, pup body weight, litter size, or the ratio of male to female pups (see 11 for presentation of these data).
Offspring Behavioral Measures Preliminary analyses of the preweanling conditioning data revealed no sex differences in the data, so these data were collapsed across that variable prior to< further analysis. Data were analyzed using analysis of variance (ANOVA) tests with a factor of Group comprised of the 6 different groups defined by prenatal treatment and rearing conditions, along with other factors as necessary depending on the design o f each specific subexperiment. Tukey's tests were used to determine the locus of significant main effects and interactions. In addition, Dunn's multiple comparisons (13) were conducted to examine the effects of biological rearing versus fostering per se ( L C / L C and C40/CA0 vs. the other groups), the effects of being reared by a previously cocaine exposed dam (C40/FOS and C40/C40 vs. the other groups), and the effects of prenatal cocaine exposure per se (C40/CA0 and FOS/CA0 vs. the other groups). These planned comparisons are discussed in the results only when they revealed effects not evident in the overall ANOVAs. These planned comparisons were not conducted for the conditioning data in that the de-
427
A FOSTERING STUDY OF THE EFFECTS: II sign of these experiments was such that the critical comparisons consisted o f comparisons within each prenatal treatment/ rearing condition group (i.e., comparing paired vs. unpaired animals).
P7-8 Olfactory Conditioning Time spent on the side of the test apparatus that contained the lemon odor (odor previously paired with footshock) was used as the dependent measure in the analysis of first-order conditioning. A 6 x 4 (Group x Condition [2 trials, 3 trials, 4 trials, and unpaired]) A N O V A conducted on these data reveaied significant main effects of Group, F(5, 196) = 11.36, p < 0.05, and Condition, F(3, 196) = 19.24, p < 0.05, along with a significant Group × Condition interaction, F(15, 196) = 2.24, p < 0.05. Tukey's test revealed no significant differences between the paired (2, 3, and 4 trials) and unpaired C40 pups reared by their biological dam (C40/CA0), reflecting a lack of first-order conditioning in this group following either 2, 3, or 4 training trials (Fig. 1). However, while C40 pups reared by foster dams (FOS/CA0) did not express conditioning when given 2 or 3 training trials, those that received 4 training trials spent significantly less time in the presence of the odor previously paired with footshock than their unpaired counterparts. In contrast, paired pups that received either 2, 3, or 4 conditioning trials from all other prenatal treatment/foster groups (FOS/LC, L C / F O S , L C / L C , and C40/FOS) spent significantly less time in the presence o f the odor previously paired with footshock than unpaired animals, demonstrating significant first-order conditioning. Taken together, these data suggest that prenatal cocaine exposure disrupts this instance of conditioning, a deficit that is even more evident when these pups are reared by dams previously exposed to cocaine.
Pl 7-18 Auditory Conditioning Following the first minute, time to reach 400 activity counts was recorded for four blocks of 100 counts each. The CS onset occurred on completion of block 2, therefore, the first two blocks are called pre-CS blocks and the second two blocks are called the CS blocks. Total time (in seconds) to reach 100 activity counts for each of the first two pre-CS blocks were transformed using a log conversion. ANOVA was then conducted separately for each of the blocks in a 6 x 4 (Group × Condition [CC, UP, 0-ISI, 20-ISI]) design. The A N O V A of block 1 revealed only a significant main effect of Condition, F(3,208) = 7.23, p < 0.001. Subsequent Tukey's tests revealed that animals in the 0-ISI condition took significantly more time to reach 100 activity counts relative to all other conditions. There were no significant main effects or interactions involving Group in block 1. The A_NOVA of block 2 also revealed a significant main effect of Condition, F(3, 208) = 1.44, p < .001, again with no main effects or interactions involving Group. Tukey's tests revealed that animals in both the 0-ISI and 20-ISI conditions exhibited a significantly longer latency to reach 100 activity counts than animals in the context control condition. To estimate the degree of conditioning in the two CS blocks, suppression ratios were calculated using untransformed data from blocks 2, 3, and 4, with the following formula: suppression ratio for first CS block (SR1) = block 2 / (block 2 + block 3); suppression ratio for second CS block (SR2) = block 2/(block 2 + block 4). The ANOVA of SR1 revealed only a significant main effect of Condition, F(3, 208) = 1.05, p < 0.001, with subsequent Tukey's test revealing that animals in the 0-ISI condition had a significantly lower suppression ratio than all other conditions (see Fig. 2). The ANOVA of SR2 revealed no significant effects. These results indicate significant conditioning to the tone for rats
120-
100
80 Z O
Z O iii
6O
[]
UNPAIRED
[]
2 TRIALS
[]
3 TRIALS
[]
4 TRIALS
iii
(J FOS/LC
LC/FOS
LC/LC
PRENATAL
FOS/C40
C40/FOS
C40/C40
TREATMENT
FIG. 1. Mean time in seconds (+ SEM) spent on the lemon side of the test apparatus by P7-8 pups from the various prenatal treatment groups. (FOS = Foster, LC = Saline injected Controls, C40 = Cocaine Group. Groups are designated by maternal treatment and pup origin [e.g., FOS/C40 pups reared by FOS dams]). *Significantly different from UP control, p < 0.05.
428
GOODWIN ET AL.
300
=. TIME TO REACH FIRST 100 PRE-CS COUNTS t
b. TIME TO REACH SECOND 100 PRE-CS COUNTS
140 • 120 • 200
100 • 80-
.=
.= 100
@
60
k4O 20 0 CC
20-1SI
LIP
©. SUPPRESSION RATIO FOR FIRST CS BLOCK
20-1SI
d. SUPPRESSION RATIO FOR SECOND CS BLOCK
0,4.
O p-
0-1SI
CONDITION
CONDITION
0.5
_o
0.3.
0,4.
I0.30.2. uJ n-
(¢
0.2"
0.1 0.1' 0.0 OC
UP
O*ISI
20-1SI
0.0 CC
CONDITION
UP
0-1SI
20-1SI
CONDITION
FIG. 2. Auditory/footshock conditioning in P17-18 pups with means collapsed across rearing and prenatal treatment groups: (a) Mean time ( + SEM) to reach the first 100 activity counts for the four conditioning groups; Co)Mean time (=l=SEM) to reach the second 100 activity counts for the four conditioning groups; (c) Mean suppression ratio for the first CS+ block (+ SEM) for the four conditioning groups; (d) Mean suppression ratio for the second CS + block (+ SEM) for the four conditioning groups. *Significantlydifferent from UP controls, p < 0.05.
given no delay between the tone and footshock although the degree of conditioning was unaffected by prenatal treatment or rearing condition.
P18 Olfactory Conditioning Odor preference. For the odor preference test, a 6 × 2 × 2 × 2 (Group x Condition [paired vs. unpaired] x Order × Time [immediate vs. 3 h]) ANOVA indicated a significant effect of order of testing, F(1, 97) = 5.60,p < 0.05. Animals receiving the preference test after the freezing test did not exhibit conditioning, suggesting that extinction to the C S + may have taken place as a consequence of prior unreinforced exposure to the CS + in the freezing test. These data are consistent with previous findings showing that an acquired aversion in preweanling animals could be extinguished with as little as 30 s exposure to the C S + (e.g., 1). Because of this significant order effect, only those subjects that received the odor preference test first are reported here. A 6 × 2 × 2 (Group × Condition × Time) ANOVA revealed a significant effect of Condition, F(1, 58) = 4.84, p < 0.05 and Time, F ( I , 59) = 22.33, p < 0.001, with paired subjects spending significantly less time over the CS + than unpaired animals and animals in the immediate test spending significantly less time over the CS + than animals in the 3 h test. Although the Time × Condition interaction did not reach sig-
nificance, F(I, 59) = 3.34, p < 0.07, inspection of Fig. 3A indicates that the main effect of Time is primarily the result of an increase in the amount of time spent over the C S + by the paired animals at the 3-h interval. This suggests that at 3 h, the paired animals were showing less of an aversion to the CS + than during the immediate test as would be expected based on prior studies (16). There was no main effect or interaction involving condition; thus, prenatal treatment and rearing history had no effect on performance of this task. Freeze test. A 6 x 2 x 2 x 2 (Group x Condition x Order x Time) ANOVA revealed only a significant effect of Condition, F(1, 93) = 9.69, p < 0.01 with paired subjects spending more time freezing than unpaired animals. As there was no interaction of Time and Condition, indicating similar amounts of freezing in the immediate and three hour tests, it appeared that the freeze test may be a more sensitive indicator of conditioning than the odor preference test. Again, no effect of Condition was observed, demonstrating that this index of conditioning was also unaffected by prenatal treatment and rearing history.
Adult Testing Shock-inducedfighting. A 6 (Group) one-way ANOVA revealed no significant differences in the mean number of fighting episodes, F(5, 58) = 1.028, p > 0.05 or latency to the
A F O S T E R I N G S T U D Y O F T H E E F F E C T S : II
429
a. ODOR PREFERENCE TEST 60
[] PAIRED [] UNPAIRED
4-
¢n to w-r
50
iz
40
0 ¢n E3
30
z
0 W U~ Z ,,=
20
IMMEDIATE
3HOUR
TEST
b.
FREEZE TEST
30
t
Z m
T
0 Z N W
[] []
PAIRED UNPAIRED
+
1.1.182o re I.i.I.I. m 0 aw
0
~ W
Ul W !¢1¢~ Z
~
IMMEDIATE
3 HOUR
TEST FIG, 3. Odor/footshock conditioning in PI8 pups with means collapsed across rearing and prenatal treatment groups: (a) Mean time ( ± SEM) spent on methyl salicylate side of the apparatus in both the immediate and 3 h tests; Co) Mean time ( ± SEM) spent freezing when presented with the CS + in both the immediate and 3 h tests. *Significantly different from UP controls, p < 0.05.
430
GOODWIN ET AL.
first aggression, F(5, 58) = 1.23, p > 0.05 among the different experimental groups. However, planned comparisons between offspring reared by cocaine dams (C40/C40; C40/FOS) and the remainder of the treatment groups (LC/LC; LC/FOS; FOS/LC; FOS/C40) revealed a significant difference in the latency to the first aggressive contact, with offspring being reared by cocaine-exposed dams, regardless of prenatal treatment history, exhibiting shorter attack latencies than offspring from the other experimental groups (see Fig. 4). No significant differences in the mean number of fighting episodes were observed employing the same planned comparisons. Shock threshold measurement. No differences in pain sensitivity to shock were observed between the different experimental groups, in that a 6 (Group) one-way ANOVA of jump/ flinch scores revealed no significant main effect of Group. Resident/intruder aggression. No indication of aggressive behavior was observed in test subjects from any of the experimental groups. This apparent lack of aggression between the resident and intruder rats could be due to the strain of rats used in this study. Although the intruder aggression test has proved useful in examining aggression in Wistar rats (7,29), Sprague-Dawley rats in general are not a particularly aggressive strain as further documented by the lack of aggression in this test.
per se. On other cognitive tests, however, no effects of prenatal cocaine exposure or rearing condition were observed. Thus, although the primary purpose of this study was to examine the interaction between fostering and prenatal drug treatment as reflected in offspring behavior, the data may also provide valuable indications as to the nature of the cognitive deficits produced by gestational cocaine exposure. Deficits were seen in offspring exposed prenatally to cocaine in an odor/footshock classical conditioning task conducted at P7-8. Whereas foster and LC control offspring under all rearing conditions exhibited a significant aversion to the odor paired with footshock following as few as two training trials, C40 offspring reared by foster dams required four training trials to demonstrate conditioning and C40 pups reared by their own dams did not exhibit conditioning even after four training trials. By varying number of training trials in this experiment, these data provide evidence that the previously reported lack of first order conditioning in P7-8 pups exposed gestationally to cocaine (9,24) does not reflect an absolute inability of these animals to learn a classical conditioning task at this age but rather demonstrates that they may require more training to exhibit significant conditioning relative to control age-mates. From the present data it can be concluded that training parameters are critical in determining whether conditioning deficits will be evident in cocaineexposed offspring. Hence, in future cognitive studies with cocaine-exposed offspring it may be prudent to vary degree of training (e.g., by manipulating number of training trials and/ or US or reinforcement intensity) to provide a sufficiently broad range of associative strength to avoid potential false negatives.
DISCUSSION
In the present study, fostering was observed to influence subsequent offspring neurobehavioral function in two testing situations, with one of these tests also revealing alterations in cognitive function associated with prenatal cocaine exposure
400 "
))
300 A
O UJ u) v
>,. 0 z l.u I,,~
200
_I
1 O0
FOS/LC
LC/FOS
LC/LC
FOS/C40
C40/FOS
C40/C40
GROUP
FIG. 4. Latency (:t:SEM) to the first aggressive contact in P60-70 offspring. *Groups significantly different from the other groups using planned comparisons.
A FOSTERING STUDY OF THE EFFECTS: II Task complexity (and hence, degree of difficulty) may also be an important factor in revealing conditioning deficits in cocaine-treated offspring and this may in turn vary with age. For instance, whereas cocaine-exposed pups given two training trials (with two shock exposures per training trial) did not exhibit odor/footshock conditioning at P7-8 (present results; also ref. 9), C40 pups given similar training at P12 or P21 did exhibit a significant aversion for the odor paired with footshock (9). These findings are reminiscent of those from the present study where C40 offspring at PI8 were observed to develop a significant odor/footshock aversion even though the odor was paired with footshock in only one training trial (during which two brief footshocks were given). Thus, by 1221 days of age, cocaine-exposed offspring appear to be able to learn a simple first order odor/footshock classical conditioning task even when only a limited amount of training is provided. Similarly, the training parameters used in the present study were sufficient to support significant auditory/ footshock conditioning in C40 offspring tested at P17-18. However, when task complexity is increased, conditioning deficits are evident in C40 offspring in this older age range. For instance, deficits in higher order conditioning have been observed in C40 pups at P12 using a sensory preconditioning task (9). Data from the present experiment also suggest that cocaine-exposed offspring are as sensitive to context as control offspring during the preweanling period. In the PI 8 auditory/ footshock conditioning experiment, greater conditioning to the context was observed when the CS (tone) and US (footshock) were paired in that context than when they were not, and there was no effect of prenatal treatment or rearing condition in this regard. Normal adult rats exhibit the opposite effect, akin to overshadowing--poorer conditioning to the context when the CS and US are paired in that context than when they are not (e.g., 14). Normal preweanling rats, however, have been shown in other experiments (15) to exhibit, instead, potentiation of conditioning to the context by pairings of the CS and US, like the present results. This is consistent with the general tendency for less potentiation and more overshadowing observed with development from infancy to adulthood (e.g., 27,28). Given that there were no significant effects of prenatal treatment (or rearing condition) in these analyses, it appeared that prenatal exposure to cocaine does not influence this tendency to show potentiation during the preweanling period, suggesting that C40 animals are as sensitive to context as are control offspring. Another issue addressed by these data is with regard to the relationship between prenatal cocaine exposure and offspring aggressive behavior. Whereas there are clinical reports that children prenatally exposed to cocaine are more aggressive than non-exposed agemates (8), in the present study prenatal exposure to cocaine was not observed to have an influence on aggressive behavior in adulthood with the response measures used. However, planned comparisons revealed that being reared by a previously cocaine-exposed dam was associated with a slightly decreased latency to aggressive attack in adult offspring in a shock-elicited aggression task. This difference with rearing condition was seen in the absence of any alteration in sensitivity to footshock per se, thus it appears to reflect an alteration in the propensity to elicit aggressive behavior in response to this stressful situation. It is not clear what aspect of the maternal behavior of dams previously exposed to cocaine results in increased aggressive behavior in the offspring they reared. However, it is interesting that prior cocaine exposure has been observed to increase maternal ag-
431 gression toward a female intruder without influencing time spent in the nest or pup retrieval latencies (11). This raises the possibility that increased maternal aggression may be reflected later in life by an increase in aggressive behavior in the offspring reared by these dams. However, note that the increase in intruder aggression seen in previously cocaine-exposed dams was only observed when they were rearing their own pups, with dams rearing their biological pups, regardless of prior drug exposure, exhibiting shorter latencies to initiate aggressive attacks (11). Whether dams previously exposed to cocaine might exhibit increased aggressive behavior in other circumstances (e.g., toward the pups during the process of weaning) that could influence the latter propensity for aggression in their offspring is presently unknown. It is also possible that some presently unspecified alteration in maternal behavior apart from the alteration in maternal aggression induced by prior cocaine exposure in the dams may produce this effect in the offspring. Finally it should be mentioned that whatever differences in maternal behavior led to the differences in the latency to attack of the offspring, it is possible that these effects may not be related to the prior cocaine exposure of the dams per se but rather due to their failure to gain normal amounts of weight during pregnancy. Pair-fed controls were not included in this experiment due to the already complex nature of the design necessary to examine the fostering issue. Although the reduction in maternal weight gain associated with cocaine exposure is relatively modest (approximately 10°70) and has not been observed previously to result in neurobehavioral alterations in pair-fed offspring (e.g., 24), it is nevertheless possible that this attenuation in weight gain could contribute to the findings observed. The final issue to be discussed regards the importance of fostering in behavioral teratology studies. In the present study, rearing condition and prenatal cocaine exposure were both observed to have an influence on the neurobehavioral function of offspring. When compared with control offspring, offspring exposed prenatally to cocaine exhibited deficits in odor/footshock conditioning at P7-8, with such deficits being more pronounced when cocaine-exposed offspring were reared by their previously cocaine-exposed dams than when reared by untreated foster dams. On another response measure, shock-elicited aggression; planned comparisons revealed that offspring reared by previously cocaine-exposed dams exhibited a decreased latency to the first aggressive contact in the absence of any alteration in shock sensitivity thresholds; this modest effect was seen regardless of whether the offspring had been exposed gestationally to cocaine. Thus, being reared by a dam previously exposed to cocaine was observed to have an impact on offspring behavioral function apart from the effects of the offspring's prenatal exposure to cocaine. It is not currently possible to generate a true estimate of the prevalence of such indirect maternally-mediated effects of cocaine. Given the limited number of measures examined in offspring in this study, it would be imprudent to attempt to use this measurement sample to generalize to the overall rate of such indirect effects of rearing condition, other than documenting that they can indeed occur. Moreover, although in this study the effect of prenatal cocaine exposure on the offspring was exacerbated in pups reared by their own dams, there is no clear a priori reason to assume that this would always be the case. It is also seemingly possible that alterations in maternal behavior induced by prior cocaine treatment during pregnancy could attenuate, rather than potentiate, manifestations of prenatal cocaine exposure per se on the offspring. This study demonstrates that rearing condition does
432
G O O D W I N E T AL.
exert a detectable influence, albeit modest on offspring neurobehavioral function following prenatal cocaine exposure. Without fostering, it would appear that there is a significant potential for false positives (and possibly even false negatives) when attributing offspring effects to the consequences o f gestational cocaine exposure per se. Whether or not fostering should be included in the design of experiments in developmental toxicology is dependent in part on the research questions being asked, along with the relative weighting o f the costs o f fostering versus the necessary ambiguity associated with interpretation of data derived from studies where pups were reared by their own dams. If the objective is to determine the consequences o f prenatal exposure per se on offspring function, the results o f this work support the importance o f fostering, at least in studies with
cocaine. Given current indications that the majority o f human infants exposed gestationally to cocaine are reared by individuais other than their biological mothers (6), the use o f such fostering may also resemble the clinical situation. If offspring are not fostered, the results o f this study indicate it cannot be tacitly assumed that the observed effects are prenatal in origin, as postnatal maternal influences may exacerbate or induce offspring alterations beyond those produced by the prenatal treatments per se. ACKNOWLEDGEMENT This research was supported by National Institute on Drug Abuse Grants R01 DA14478 and K02 DA00140 to Linda P. Spear. We gratefully acknowledge the assistance of Yolanda Pinzas for technical support.
REFERENCES 1. Arnold, H. M.; Spear, N. E. Order and duration of stimuli are important determinants of reactivation (submitted). 2. Birke, L. I. A.; Sadier, D. Maternal behavior in rats and the effects of neonatal progestins given to the pups. Dev. Psychobiol. 18:467-475; 1985. 3. Cierpial, M. A.; McCarty, R. Hypertension in SHR rats: Contribution of the maternal environment. A. J. Physiol. 235:H980H984; 1987. 4. Clow, D. W.; Hammer, R. P.; Kirstein, C. L; Spear, L. P. Gestational cocaine exposure increases opiate receptor binding in weanling offspring. Dev. Brain Res. 59:179-185; 1991. 5. Cuadra, G. R.; Molina, V. A. Antidepressants reverse the inhibition of shock-induced aggression elicited by a prior inescapable shock. Pharmacol. Biochem. Beh. 40:69-73; 1991. 6. Davis, E. Characteristics of "exposed" fetus and newborn human: Characteristics of the infant. Presentation at Conference entitled: Cocaine in pregnancy: Is the fetus affected? (Sponsored by the National Institute on Drug Abuse, and Office of Substance Abuse Prevention), Washington, DC; 1991. 7. Fernt~ndez-Espejo, E.; Mir, D. Ethological analysis of the male rat's socioagonlstic behavior in a resident-intrnder paradigm. Aggressive Beh. 16:41-55; 1990. 8. Field, T. Behavioral consequences. Presentation at Conference entitled: Cocaine in pregnancy: Is the fetus affected? (Sponsored by the National Institute on Drug Abuse, and Office of Substance Abuse Prevention), Washington, DC; 1991. 9. Heyser, C. J.; Chen, W. J.; Miller, J.; Spear, N. E.; Spear, L. P. Prenatal cocaine exposure induces deficits in Pavlovian conditioning and sensory preconditioning among infant rat pups. Behav. Nenrosci. 104:955-963; 1990. 10. Heyser, C. 3.; Miller, J. S.; Spear, N. E.; Spear, L. P. Prenatal exposure to cocaine disrupts cocaine-induced conditioned place preference in rats. Neurotoxicol. Teratol. 14:57-64; 1992. 11. Heyser, C. J.; Molina, V. A.; Spear, L. P. A fostering study of the effects of prenatal cocaine exposure: I. maternal behaviors. Neurotoxicol. Teratol. 14:415-421; 1992. 12. Hutchings, D. E. Issues of methodology and interpretation in clinical and animal behavioral teratology studies. Neurobehav. Toxicol. Teratol. 7:639-642; 1985. 13. Kirk, R. E. Experimental design: Procedures for the behavioral sciences. Belmont, CA: Brooks/Cole; 1982. 14. Marlin, N. A. Within-compound associations between the context and the conditioned stimulus. Learn. Mot. 13:526-541; 1982. 15. McKinzie, D. L. Ontogenetic differences in the expression of context and CS conditioning as a function of context saliency and CS-UCS interval. Binghamton, NY: State University of New York; 1992. Thesis. 16. Miller, J. S.; Jagielo, J. A.; Spear, N. E. Differential effectivehess of various prior cuing treatments on the reactivation and
maintenance of memory. J. Exp. Psych.: Animal Behav. Process. 17:249-258; 1991. 17. Moody, C. A.; Frambes, N. A.; Spear, L. P. Psychopharmacological responsiveness to the dopamine agonist quinpirole in norreal weanlings and in weanling offspring exposed gestationally to cocaine. Psychopharm. 108:256-262; 1992. 18. Moore, C. L; Power, K. L. Prenatal stress affects mother-infant interaction in norway rats. Dev. Psychobiol. 19:235-245; 1986. 19. Moye, T.; Rudy, J. Ontogenesis of learning: VI., learned and unlearned responses to visual stimulation in the infant hooded rat. Devel. Psychobio. 18:395-409; 1985. 20. Pearson, D. E.; Teicher, M. H.; Shaywitz, B. A.; Cohen, D. J.; Young, J. G.; Anderson, G. M. Environmental influences on body weight and behavior in developing rats after neonatal 6hydroxydopamine. Science 209:715-717; 1980. 21. Scalzo, F. M.; All, S. F.; Frambes, N. A.; Spear, L. P. Weanling rats exposed prenatally to cocaine exhibit an increase in striatal D2 binding associated with an increase ligand affinity. Pharmacol. Biochem. Behav. 37:371-373; 1990. 22. Sobrian,, S. K.; Burton, L. E.; Robinson, N. L.; Ashe, W. K.; James, H.; Stokes, D. L.; Turner, L. M. Nenrobehavioral and immunological effects of prenatal cocaine exposure in rat. Pharmacol. Biochem. Behav. 35:617-629; 1990. 23. Spear, L. P.; Heyser, C. J. Cocaine and the developing nervous system: Laboratory findings. In: Zagnn, I. S.; Slotkin, T. A., eds. Maternal substance abuse and the developing nervous system. Orlando, FL: Academic Press; 1992:155-175. 24. Spear, L. P.; Kirstein, C. L.; Bell, J.; Yoottanasumpun, V.; Greenbanm, R.; O'Shea, J.; Hoffman, H.; Spear, N. E. Effects of prenatal cocaine exposure on behavior during the early postnatal period. Nenrotoxicol. Teratol. 11:57-63; 1989. 25. Spear, L. P.; Kirstein, C. L.; Frambes, N. A. Cocaine effects on the developing central nervous system: Behavioral, psychopharmacologicai, and neurochemical studies. Ann. NY Acad. Sci. 526:290-307; 1989. 26. Spear, L. P.; Kirstein, C. L.; Frambes, N. A.; Moody, C. A. Nenrobehavioral teratogenlcity of gestational cocaine exposure. NIDA Res. Monogr. 95:232-238; 1990. 27. Spear, N. E.; Kucharski, D. Ontogenetic differences in stimulus selection during conditioning. In: R. Kail; N. E. Spear, eds. Comparative perspectives on the development of memory. Hillsdale, N J: Erlbanm. 227-252; 1984. 28. Turkewitz, G.; Mellon, R. C. Dynamic organization of intersensory function. Can. J. Psych. 43:286-301; 1989. 29. Vergnes, M.; Depanlis, A.; Boehrer, A. Parachlorophenylalanine-induced serotonln depletion increases offensive but not defensive aggression in male rats. Physiol. Behav. 36:653-658; 1986. 30. Vorhees, C. V. A fostering/crossfostering analysis of the effects of prenatal ethanol exposure in a liquid diet on offspring development and behavior in rats. Neurotoxicol. Teratol. 11:115-120; 1989.