Neurotoxicology and Teratology, Vol. 17, No. 5, pp.

545-552, 1995
Copyright 0 1995 Elsevier Science Inc.
Pergamon Printed in the USA. All rights reserved
0892.0362/95 $9.50 + .OO

0892-0362(95)00014-3

Prenatal Ethanol Exposure in Rats:

Long-Lasting Effects on Learning

PETER CLAUSING, *’ SHERRY A. FERGUSON,1_ R. R. HOLSON,-f

RICHARD R. ALLEN* AND MERLE G. PAULE$*

Divisions of *Neurotoxicology and TReproductive and Developmental Toxicology,

National Center for Toxicological Research, 3900 NCTR Road, Jefferson, AR 72079-9502
SDepartment of Pharmacology and Toxicology and Department of Pediatrics,
University of Arkansas for Medical Sciences, Little Rock, AR 72205

Received 24 October 1994; Accepted 23 March 1995

CLAUSING, P., S. A. FERGUSON, R. R. HOLSON, R. R. ALLEN AND M. G. PAULE. Prenatal ethanol expo-
sure in rats: Long-lasting effects on learning. NEUROTOXICOL TERATOL 17(5) 545-552, I995.-Pregnant Sprague-
Dawley rats were fed a liquid diet containing either 0% (group C), 18% (group L), or 36% (group H) ethanol-derived calories
(EDC) from gestational day 1 to 20. Male offspring were assessed under a conditioned taste aversion paradigm (PND 35-45),
in a complex maze (PND 68-80). and for operant behavior (temporal response differentiation and motivation to work for
food, PND 140-198). Although conditioned taste aversion was fully acquired by all groups, retention of the conditioned taste
aversion response was impaired in group H animals. Importantly, deficits in the acquisition of timing behavior were found in
group H (group L not tested), confirming that this operant task is quite sensitive in detecting prenatal drug effects and
demonstrating that neurological effects of prenatal ethanol exposure persist into late adulthood.

Prenatal ethanol exposure Conditioned taste aversion Operant behavior Complex maze activity Rat

ETHANOL, the leading known human teratogen (l), affects enabled the evaluation of retention of the learned conditioned
a variety of organ systems in both humans and laboratory avoidance response. Littermates of these conditioned taste
rodents (9,14). Although a tremendous amount of data has aversion subjects were tested in a complex maze between PND
been accumulated concerning such effects, as evidenced by a 69 and 80, an age that was already beyond the age limit dis-
number of recent reviews [e.g., (7,17,43)], studies dealing with cussed above. Few other studies testing prenatally ethanol-
a variety of endpoints within the same experiment are not very exposed rats in a maze have used appetitive paradigms, and
abundant. The effects of prenatal ethanol exposure on the the testing age was PND lo-12 (4), PND 41-45 and 46-77
learning performance of rodents have been repeatedly de- (26), and PND 60+ (31). The same individuals tested in the
scribed [cf. (27)]. In many of these reports the learning tasks complex maze were assessed for operant behavior after PND
employed were either relatively simple (6,23) or only one 140. The number of studies investigating operant behavior in
learning paradigm was examined (8,19). Learning deficits in rats after prenatal ethanol exposure is very limited (13,35,39),
rats prenatally exposed to ethanol often cannot be detected and only in one of them (13) training of the animals started at
when they are assessed at 70 days of age or older (26). Thus, ages greater than 70 days. Thus, the primary aim of the pres-
the present study was designed to investigate learning perfor- ent study was to further examine if adult rats exhibit learning
mance in rats older than those most often used by previous deficits after prenatal ethanol exposure.
investigators. Two earlier studies described effects of prenatal A second goal was to evaluate the sensitivity of a variety of
ethanol exposure on conditioned taste aversion learning; how- behavioral paradigms for detecting subtle functional deficits.
ever, preweanling (33) or 21-30-day-old rats (34) were used in Comprehensive psychopharmacological studies revealed the
those investigations. Here, 39-45-day-old rats were assessed temporal response differentiation task as the most sensitive of
under a conditioned taste aversion paradigm that was based five different tasks in a monkey operant test battery (30). A
on cyclophosphamide as the unconditioned stimulus, which preliminary study of rats employing two of these tasks indi-

’ Requests for reprints should be addressed to Peter Clausing, NCTR, HF”I-132, Jefferson, AR 72079.9502

545
546 CLAUSING ET AL.

cates similar findings (16). Therefore, the present study aimed TABLE 2
at gathering further information on this issue.
CONDITIONED TASTE AVERSION-
EXPERlMENTAL DESIGN
METHOD
Day 0 Day 3 Day 6
General Experimental Design Subgroup (PND 39)’ (PND 42) (PND 45)

Pregnant Sprague-Dawley rats were obtained from the

SAC SAC? + CYS SAC SAC
breeding colony of the National Center for Toxicological Re-
“Conditioned”
search and were kept under a 12:12 h lighting regime with
light on from 0600 to 1800 h. They received a liquid diet CON SAC + Saline SAC SAC
(BIOSERV@, Frenchtown, NJ) containing O%, 18%, or 36% “Control”
ethanol-derived calories (EDC) from gestational day (CD) 1 CY SAC + CY HK Hz0
to 20. For dams in the high-dose group (H, n = 13) this diet “Residual”
was provided ad lib, whereas dams in the low-dose (L, n =
11) and control (C, n = 15) groups were yoked to a 36% EDC *After adaptation to restricted water supply, which was grad-
animal of approximately the same body weight and pair-fed ually introduced from postnatal day (PND) 32 to PND 35.
with isocaloric amounts of the diet. Consumption of the liquid TSaccharin (0.15%) in drinking water.
diet was recorded to the nearest 0.1 g by weighing the liquid SCyclophosphamide injection (50 mg/kg, IP).
diet bottles. Maternal body weight was recorded daily from
CD 1 to 19. Blood alcohol concentration was determined by
gas chromatography (flame ionization detection) from two different subgroups [control (CON), conditioned (SAC), and
additional dams from each of groups L and H. BIood was residual (CY)] were established for each prenata1 treatment
collected from the orbita plexus on CD 14 and 18, 3 h after group (C, L, H). From PND 32 to 35 water restriction was
lights off (2100 h), a time at which peak blood alcohol concen- gradually introduced (4 h, 2 h, 1 h, 30 min per day) and
trations can be obtained (28,38). kept at 30 min of water availability per day thereafter. Fluid
Pups were weighed on postnatal day (PND) 0 and culled to consumption was recorded by weighing the water bottles each
eight/litter maintaining an equal number of each sex when day before and after the session. As outlined in Table 2, a
possible. Body weights were obtained on PND 5, 10, and 15. 0.15% saccharin solution (SAC and CON groups) or tap water
The pups remained with their natural mothers, and were (CY group) was provided for 30 min on PND 39, correspond-
weaned and ear-punched for identification on PND 21. After ing to day 0 (conditioning day), followed by IP injections of
weaning, rats were housed in same-sex groups of three to four 50 mg cyclophosphamide/kg (SAC and CY groups) or saline
per cage unless otherwise indicated. Cross-fostering was con- (CON groups) 20 min after removing the drinking bottles.
sidered unnecessary for the present experiment, because post- Cyclophosphamide at this dose causes moderate gastrointesti-
natal maternal influences induced by prenatal treatment with nal distress in rats, and when paired with saccharin produces
ethanol in a liquid diet have been described as “not sufficiently an aversion to it such that conditioned animals (SAC groups)
strong modifying variables on offspring development or be- exhibit a decreased intake of saccharin solution on reexposure
havior to be a source of experimental concern” (41). Offspring (20). On PND 42 (day 3 of experiment) and PND 45 (day 6 of
allocation to the different studies reported here is illustrated experiment), CON and SAC animals were reexposed to the
in Table 1. The litter was considered the experimental unit, conditioning stimulus (saccharin).
and therefore efforts were made to incorporate only one indi-
vidual from each litter into each experimental group. We also Complex Maze
sought to have littermates or, if possible, the same subjects
Activity and learning performance were assessed in male
serve in each of the different studies.
offspring at the ages listed in Table 1. Subjects were individu-
ally housed from PND 56. The apparatus and procedure were
Conditioned Taste Aversion
similar to those described previously (22) except that water
On PND 28, 30 male offspring from each treatment group instead of food served as the reinforcer and each rat remained
(3 per dam, 90 rats total) were individually housed and allo- in the maze without interference for the full length of the
cated to the conditioned taste aversion study (Table 1). Three 15-min test session. The apparatus was an array of 24 inter-

TABLE 1
OFFSPRING ALLOCATION AND TESTING AGE FOR EACH EXPERIMENT

Experimem CWlp Offspring From Dam # Testing Age (PND)

Conditioned taste aversion C 1, 2, 3, 5, 6, 9, 11, 12, 17, 19 35-45

L 4, 5, 8.9, 11, 12, 14, IS, 17, 18
H 1, 4, 5, 9, 10, II, 14, 16, 18, 19
Complex maze C 1.2, 3,5,6,9, 12, 14, 19.20 69-80
L 4, 5, 8, 9, 11, 12, 14, 15, 17, 18
H 4, 10, 11, 12, 14, 15, 15, 16, 18, 19
Operant test battery C Same individuals as in complex maLe testing 140-192
H
PRENATAL ETHANOL AND LEARNING IN RATS 541

connected acrylic arms, two of which were designated as goal a microcomputer that administered the behavioral tasks and
arms. There were seven choice points between the two goal recorded the responses.
arms. Tap water (0.25 ml/reinforcer) was available in these Procedure. Beginning on PND 128, rats were gradually
two arms (HOME and GOAL). Subjects were gradually water food deprived to SO-85% of their free-feeding weight and
deprived prior to the first session, as described above, and maintained at this weight throughout the remainder of the
were allowed 30 mm of ad lib water after each daily session. study. Mean + SE body weights prior to food deprivation
Sessions were 15 min in length and occurred on each of 5 were 470 k 17 and 482 + 17 g for the ethanol-treated and
consecutive days. To accommodate the total sample size, each control groups, respectively. During behavioral assessments,
treatment group (C, L, H) was split into two subgroups that mean body weights were 387 -t 11 and 407 f 16 g for the
were tested over 2 consecutive weeks (i.e., at 10 weeks of age ethanol-treated and control groups, respectively. Beginning
for the first, and at 11 weeks of age for the second subgroup) on PND 140, 50-min test sessions were conducted Monday-
and the sequence of testing the animals alternated between Friday at the same time each day.
the different dose groups. The procedure required subjects to Temporal response differentiation task. Initially, the tem-
alternate between GOAL and HOME arms to obtain rein- poral response differentiation or time estimation task (50-min
forcers; that is, a reinforcer in the HOME arm could be ob- training sessions) was the only task presented: subjects were
tained only after a reinforcer had been obtained in the GOAL trained to press the left of two retractable levers (right lever
arm, and vice versa. The number of reinforcers obtainable per retracted, left extended) to obtain food reinforcers. The dura-
test session was not limited. All animals designated for maze tion that the lever had to be held in the depressed position was
testing were moved from the animal room to the maze room 2 then gradually increased over several sessions from zero to a
h prior to testing, which started at the beginning of the dark minimum of 10 s; the maximum hold duration was 14 s. Thus,
phase (1800 h). The location of the animal rack was close to under the final requirements of the task, subjects had to hold
the “bottom” area of the maze [see Fig. 1 in (22)], a circum- down and then release the lever within a 4-s window of oppor-
stance that is important to remember for interpretation of the tunity. Lever holds of less than 10 s or more than 14 s ended
results. A small desk lamp provided minimal illumination in the current tria1 but had no other programmed consequences.
the maze room. A computer connected to the maze recorded Subjects could initiate the next trial immediately. After attain-
a total of 37 different parameters during each session. For ing a time estimation task training criterion of holding the
convenience, only those variables that were significantly lever down for a minimum of 7 s for at least 40 occurrences in
affected or that are important for the interpretation of the a single session, the maximum session length was reduced
results will be reported here. These 10 variables are: total from 50 to 40 min (the task terminated in less than 40 min if
number of arm entries (TOTACT); total number of water subjects obtained 120 reinforcers).
reinforcers dispensed (REWARDS); number of errorless runs Motivation task. Ten-minute motivation tasks immediately
from start arm to goal arm (GOALRUNS); number of error- followed the temporal response differentiation task in the
less runs from goal arm to start arm (HOMERUNS); fre- same behavioral session. Motivation task training was initi-
quency (FREEZES) and duration in seconds (FREEZTIME) ated when the session time for the temporal response differen-
of episodes where subjects remained in an arm longer than tiation task was reduced to 40 min (see above). Motivation
60 s; frequency of U-turns in the maze or turning back prior task training level 1 required performance under a fixed ratio
to reaching a dead end (REVERSES); number of entries into (FR) 1 schedule of reinforcement (every lever press regardless
arms on the direct route between goals (ENTRIES); time (in of hold duration was reinforced). Every 2 lever presses (FR2)
seconds) spent in the bottom region of the maze (BOTTOM- were reinforced in the second training level, every 5 presses
TIME); and the probability of a wrong (left) turn at the first (FR5) in level 3, and every 10 (FRlO) in level 4. Each subject
decision point when leaving the start arm area (GOAL- had to earn a minimum of 10 reinforcers under each training
ERROR4). level to progress to the next level. Following completion of
these four training levels (within four to five sessions for all
Assessment of Operant Behavior subjects), the full motivation task schedule was introduced
wherein subjects were required to respond under a progressive
Eight male rats from each of groups C and H that had also ratio 1 + 1 schedule. Initially, one lever press resulted in rein-
been used in the complex maze experiment (see Table 1) were forcer delivery. The number of responses required for each
used in this part of the investigation. The two operant tasks subsequent reinforcer was then increased by 1 such that the
used, temporal response differentiation and progressive ratio, second reinforcer required two lever presses, the third required
were identical to those previously used in this laboratory (16). three, and so on. Under this schedule, subjects rarely earned
Performance in the temporal response differentiation task is more than 30 reinforcers during any IO-min session.
thought to model aspects of time estimation, whereas perfor- Endpoints. Endpoints measured in each task included per-
mance in the progressive ratio task is thought to model aspects cent task completed (percent of a predetermined criterion),
of motivation (in the following referred to as “motivation response rate (lever presses/second), mean duration of lever
task”). hold (seconds, temporal response differentiation task only),
Apparatus. All operant sessions were conducted in one of accuracy (percent correct responses, temporal response differ-
two identical operant behavior chambers with internal mea- entiation task only), and breakpoint (number of lever presses
surements of 24.8 x 22.9 x 21.0 cm. A front panel was made for the last reinforcer earned, motivation task only). A
instrumented with two retractable response levers, each more detailed description of these endpoints can be found
positioned below a stimulus light and separated by a rein- elsewhere (16).
forcer trough. Each chamber was housed inside a sound-at-
tenuated box. Reinforcer (45mg dustless precision food pel- Statistical Analysis
let, BIOSERV@) delivery was accompanied by the noise of the
pellet dispenser and illumination of a light above the rein- Operant test battery. Number of sessions during acquisi-
forcer trough. Each chamber and panel were controlled by tion of temporal response differentiation and motivation task
548 CLAUSING ET AL.

responding differed for each subject; thus, individual test ses- TABLE 3
sions could not be used as a repeated measure in an analysis
AVERAGF PUP WEIGHT/LITTER
of variance (ANOVA). Individual subject means were calcu-
lated at acquisition and full schedule performance for each Group c Group L Group H
behavioral endpoint in the temporal response differentiation PND* Sex (n = 15) (n = 13) (n = IS)
and motivation tasks. That is, for each endpoint, there were
two data points for each subject. These means were then sub- 0 M 5.88 t 0.13 6.05 k 0.20 5.68 k 0.18
jected to a one-way repeated-measures ANOVA with treat- 5 10.09 5 0.28 10.73 + 0.32 9.06 f 0.25’
ment as a between-subjects factor and phase (acquisition or 10 19.41 i 0.43 19.92 t 0.54 17.00 k 0.63’.’
full schedule) as a within-subjects factor. Where there was a 15 29.92 k 0.79 30.06 ? 0.95 27.32 f 0.96’.’
significant treatment by phase interaction, the ethanol-treated 0 F 5.73 i 0.15 5.84 * 0.17 5.35 a 0.14
group was compared to control performance at each phase 5 9.88 k 0.36 10.08 t 0.34 8.46 2 0.26’
using a Bonferroni correction. This relatively conservative ap- 10 18.32 k 0.53 18.93 f 0.66 16.12 k 0.62’,’
proach of statistical testing was used to alIow for the increased 15 28.51 k 0.78 28.71 + 1.01 27.11 + 0.60’
sample size due to the multiple testing of the same individuals
over time. Under full schedule data were additionally analyzed Values are mean +- SEM; n = number of litters.
by multiple linear regression using the generalized linear *Postnatal day.
model procedure (y = 01 + /3, * group + & * session + pi * ’ J Significantly different from group C and/or 1~(/-, < 0.05).
group * session).
Other endpoints. With the exception of interday compari-
sons in the conditioned taste aversion study, data for all other 6 was two times higher than that for group C-SAC (p 5 0.05).
endpoints were analyzed by one-way ANOVA with post hoc Also, saccharin consumption in group H-SAC increased sig-
analyses by Fisher’s least significant difference test, if signifi- nificantly on day 6 compared to day 3 (p = 0.0031, paired
cant differences were detected by ANOVA. Interday fluid t-test). Both results indicate an impaired retention of the ac-
consumption in the conditioned taste aversion study was ana- quired conditioned taste aversion response in rats with high
lyzed by paired t-test. prenatal ethanol exposure. On day 3 a slight but significant
reduction of saccharin consumption was recorded in all CON-
RESUI.TS
groups (Fig. 2). The lack of any decrease in fluid consumption
in the CY groups 3 days after injection (Fig. 3) illustrates that
Maternal and Pup Data 50 mg/kg cyclophosphamide per se did not adversely affect
the animal’s well-being for an extended period of time.
The average maternal weight gain from CD I to 19 was
51.7 f 2.6 g, 49.6 + 2.8 g, and 63.8 ? 5.5 g for C, L, and
Complex Maze
H groups, respectively. As revealed by ANOVA and post hoc
LSD test, the (ad lib-fed) H dose group dams gained signifi- The maze performance of group H subjects was not signifi-
cantly more weight than the dams of the two other groups, cantly different from that of group C animals except for a
F(2, 36) = 3.77, p = 0.033. Groups H and L consumed an
average of 13.18 and 6.47 g ethanol/kg/day, respectively.
Blood alcohol concentrations averaged from blood taken on
CD 14 and 18 were 38.7 + 9.2 mg/dl for group L and 126.2
f 9.6 mg/dl for group H.
One dam in group C and two in group H failed to deliver
or raise their litters successfully and therefore were excluded
from the study. For the remaining individuals the average
number of pups born per dam was 14.1 & 0.7, 13.5 + 0.8,
and 13.1 * 0.5 for groups C, L, and H, respectively.
ANOVA failed to reveal significant effects of prenatal treat-
ment on number of pups per litter, F(2, 33) = 0.950, p =
0.397. PND 0 body weight was lower in group H, but the
difference was not significant [male pups: F(2, 32) = 1.165, p
= 0.325; female pups: F(2, 32) = 2.662, p = 0.0851. This
difference became larger and statistically significant on PND
5, 10, and 15 (Table 3).

Conditioned Taste Aversion

The results of the conditioned taste aversion study are pre- FIG. 1. Conditioned taste aver-sion - conditioned (SAC-) groups.
sented in Figs. l-3. All prenatal treatment groups exhibited a Saccharin fluid consumption (mean + SEM) of male rats prenatally
clear taste aversion to saccharin after having had it paired with exposed to 0% (C, open bars), 18% (L, dashed bars), or 36% (H,
a cyclophosphamide injection 3 days earlier (group SAC, Fig. solid bars) of ethanol-derived calories (EDC). On day 0 animals were
given access to a 0.15% saccharin drinking solution paired 20 min
1). On day 3 there were no significant differences in this re-
later with a 50 mg/kg injection of cyclophosphamide, IP. On days 3
sponse between prenatal treatment groups C, L, and H. On
and 6 the animals were again given access to the saccharin drinking
day 6, groups C still had maximal conditioned taste aversion solution. On days 3 and day 6 all groups had significantly decreased
response, whereas animals in group H exhibited a significant consumption compared to day 0 (*p < 0.05). #Significantly different
increase and group L the tendency to an increase in saccharin from saccharin consumption of group H on day 3 (p = 0.003 I), and
consumption. Saccharin consumption in group H-SAC on day significantly different from the 0% EDC group on day 6 (p = 0.047).
PRENATAL ETHANOL AND LEARNING IN RATS 549

TABLE 4
COMPLEX MAZE PERFORMANCE COLLAPSED
OVER THE 5 DAY TEST PERIOD

Endpoint Control Low Dose High Dose

TOTACT 275.2 + 11.9 193.4 + 8.85*t 263.3 k 13.9

REWARDS 22.0 f 2.11 12.5 + 1.24*t 22.1 + 2.45
GOALRUNS 6.76 + 0.96 2.86 + 0.55*t 6.18 k 0.96
HOMERUNS 5.28 & 0.88 2.56 k 0.46*$ 4.74 k 0.78
REVERSES 35.4 + 2.33 23.0 + 1.63*t 34.7 k 2.59
ENTRIES 178.5 + 12.7 109.6 + 8.35*t 170.0 & 14.4
FREEZES 0.30 * 0.10 0.26 & 0.08 0.06 k 0.030
FREEZTIME 35.7 * 14.7 25.1 * 8.09 4.40 k 2.520
BOTTOMTIME 88.8 + 5.10 115.6 + 8.02* 107.2 * 7.57
GOALERROR 0.54 f 0.02 0.64 + 0.02* 0.57 * 0.02

Values are mean + SEM; n = 10 per group.

FIG. 2. Conditioned taste aversion-control (CON-) groups. On day *Significantly different from group C (p < 0.01).
0 animals were given access to a 0.15% saccharin drinking solution tsignificantly different from group H (p < 0.01).
paired 20 min later with an IP injection of saline. On days 3 and 6 the isignificantly different from group H (p < 0.05).
animals were again given access to the saccharin drinking solution. OSignificantly different from group C (p < 0.05).
On day 3 all prenatal treatment groups had a significantly lower fluid
consumption compared to day 0 and to day 6 (*p < 0.05). #Signifi-
cantly increased compared to day 0 (p = 0.03). For further details see
Fig. 1 legend. Operant Behavior
There was a significant difference in the slope of the regres-
sion lines fit to the two time estimation endpoints: response
rate and mean duration of lever press. As depicted in Fig. 4,
decreased number of FREEZES and a decreased FREEZ-
subjects exposed prenatally to ethanol did not exhibit the typi-
TIME. In contrast, group L was unchanged in FREEZES and
cal decrease in response rate in the temporal response differen-
FREEZTIME, but exhibited significant differences in a num-
tiation task (seen in controls) during continued training under
ber of other endpoints. Table 4 summarizes these findings that
the time estimation full schedule. This is illustrated by a nega-
are presented as averages for all 5 maze testing days. Analysis
tive slope (b = - 0.0012 f 0.0002) that was significantly dif-
of data for individual days (not shown) generally revealed
ferent from zero (p I 0.0001) in group C, contrasted by
significant differences on 2 or 3 consecutive days. Decreases
virtually no slope (b = -0.0001 + 0.0003, p = 0.6552) in
in group L were observed for TOTACT, REWARDS, both
group H. Also, the slope estimates of the two groups differed
HOMERUNS and GOALRUNS, REVERSES, and ENTRIES
significantly from each other (p = 0.003). Although initially
(these decreases were often significantly different from group
starting full schedule time estimation at a lower response rate
H also; cf. Table 4). Interestingly, BOTTOMTIME and
than group C subjects, H subjects did not continue to improve
GOALERROR were significantly increased.
their performance with repeated testing. The slope of the
mean duration of lever press vs. session line was significantly
less steep (p = 0.047) for group H subjects (b = 0.0743 +
0.0213) than for control subjects (b = 0.1343 + 0.0205) (Fig.
5). None of the other endpoints of the temporal response
differentiation task (percent task completed, response accu-
racy) nor any of the parameters of the motivation task were
significantly different between groups (data not shown). The
subjects’ performance in these two tasks was comparable to
previous results (16).

DISCUSSION

Offspring of Sprague-Dawley rats exposed to ethanol

through a liquid diet from gestational day O-20 exhibited an
impaired retention of a (fully learned) conditioned taste aver-
sion task on PND 45, and poor performance in an operant
behavior task when tested subjects were older than 5 months.
Average daily ethanol consumption and peak blood alco-
hol concentrations were within the range that is usually re-
ported for liquid diet administration to pregnant rats at con-
FIG. 3. Conditioned taste aversion-residual (CY-) groups. On day
0 animals received a single IP injection of 50 mg/kg cyclophospha- centrations of 18% and 36% EDC (14,26). Gestational weight
mide. Tap water was provided for drinking on all days. On day 3 two gain of dams receiving 35-40% EDC in their liquid diet was
and on day 6 all prenatal treatment groups had a significantly higher sometimes lower (21,25), often equal to [e.g., (19,45)], and
water consumption than on day 0. *Significantly increased compared sometimes higher (12,44) than that of pair-fed (0% EDC) con-
to day 0 (JJ < 0.05). For further details see Fig. 1 legend. trol dams. Recently, it has been suggested that the type of
550 CLAUSING ET AL.

Control High Dose

-T’75
5 10 io 25 5 10 15 20 25
SESSION SESSION

FIG. 4. Operant behavior-temporal response differentiation task, response rate. The

response rate decreases significantly over time in the control group (slope = -0.0012
+ 0.0002, p 5 O.OOOl), whereas it remained unchanged in group H (slope = - 0.0001
& 0.0003, p = 0.6552). Also, the two slopes were significantly different from each
other (p = 0.003). The regression line is depicted with 95% mean confidence curve.

liquid diet may interact with the effect of ethanol on gesta- after conditioning, animals from the high ethanol group con-
tional weight gain. In a comparative study it was found that sumed significantly more saccharin than the control group.
ethanol administered via Sustacal-based diets reduced mater- Furthermore, animals in the high ethanol group significantly
nal weight gain, whereas the weight of ethanol-treated dams increased their saccharin consumption during their second ac-
fed a Lieber & DeCarli formula did not differ from that of cess to this drinking solution. Thus, the findings of Riley et al.
pair-fed controls (24). In accord with some previous reports (33,34), suggesting that prenatal ethanol exposure interferes
(see above), our 35% EDC dams had a higher weight gain, with conditioned taste aversion learning, are basically con-
although the statistical significance of this finding must be firmed, although not simply replicated. Both the earlier stud-
considered incidental. Reduced offspring body weight, a char- ies used Long-Evans rats (33,34), in contrast to the Sprague-
acteristic feature of prenatal ethanol exposure (18), was also Dawley strain used here. Strain differences in the conditioned
observed in this study. The fact that the body weight reduction taste aversion paradigm were investigated, and a Sprague-
was only 3-5% (and not significant) in neonates, and most Dawley strain was described that had a greater initial aversion
pronounced (12-14070 reduction) on PND 5 and 10, indicates and a more sustained response than Long-Evans rats (3). This
that suckling deficits (1 l), possibly induced by impaired olfac- observation may provide an explanation as to why the effects
tion (5), may have played a role in this postnatal growth retar- of prenatal ethanol treatment became obvious only during the
dation. second presentation of saccharin in the present experiment.
In the present investigation exposure to 18% or 36% EDC During the first presentation of saccharin after conditioning
from CD O-20 did not affect the ability of subjects to develop the effect of prenatal ethanol may have been masked by the
a taste aversion response 3 days after conditioning. However, strong initial aversive response reported for Sprague-Dawley
on the second postconditioning access to saccharin 6 days rats. One of the two earlier studies describing effects of expo-

Control High Dose

i 10 15 50 -25 . 5 lb 15 20 25
SESSION SESSION

FIG. 5. Operant behavior-temporal response differentiation task, mean duration

of lever press. The mean duration of lever press was increasing significantly faster
over time (p = 0.047) in the control group (slope = 0.1343 ? 0.0205) compared to
group H (slope = 0.0743 ? 0.0213). The regression line is depicted with 95% mean
confidence curve.
PRENATAL ETHANOL AND LEARNING IN RATS 551

sure to ethanol in utero (34) tested the subjects at approxi- ethanol exposure on learning become less prominent or disap-
mately the same age. However, the experimental design of pear with age. As reviewed by Means et al. (26), three reports
this investigation resembled more a passive avoidance than have described deficits in acquisition when subjects were
a conditioned aversion task. The aversive stimulus (lithium tested before PND 40; however, no effect was observed when
chloride) was not matched with a well-tasting fluid (saccharin subjects were tested later (at PND 80-90). In addition, eight
solution) presented later. Instead, lithium chloride was pre- studies were cited that report deficits in acquisition using ani-
sented directly in the drinking water on consecutive days and mals assessed before PND 69 as opposed to six studies that
the decrease in fluid consumption was recorded. In our experi- were cited in which acquisition of different learning tasks was
ment animals were older (beyond PND 38) and effects of unaffected in animals assessed after PND 70. In a more recent
prenatal treatment on saccharin consumption were not dem- review that specifically focussed on the long-term behavioral
onstrated on first reexposure to saccharin. Adding a retention effects of prenatal ethanol exposure in rats it was concluded
session (i.e., a second access to saccharin) to the “classical” that, “it appears that task complexity is a factor in these age-
conditioned taste aversion paradigm proved to be more sensi- related findings” (32). The temporal response differentiation
tive for detecting the effects of prenatal ethanol exposure. task is one that is comparatively difficult to train and it is also
Finally, it should be mentioned that subjects in the control very sensitive to drug effects (30). It is different from the more
group (injected with saline and exposed to saccharin 3 days traditional differential reinforcement of low response rate
later) exhibited a slight, but significant, decrease in fluid con- (DRL) schedules, in that it requires maintenance of a time-
sumption, indicating that the IP injection procedure itself had measured response rather than the time-measured withholding
a certain aversive conditioning effect. This finding has been of that response. In accordance with others (13), who tested
observed by others also (15). Taken together, adverse effects rats prenatally exposed to ethanol under a DRL schedule be-
of in utero ethanol exposure on conditioned taste aversion ginning at PND 80, our treated subjects had a lower average
learning were confirmed, although a number of differences response rate than control animals. However, the most impor-
exist in design and outcome between the present experiment tant aspect of the time estimation performance in the present
and those previously published. This indicates that the effects study was that control animals significantly improved their
of prenatal ethanol on learning are demonstrable in a variety performance (decreased their response rate) over successive
of conditions and suggest that such effects are robust. sessions whereas animals in the high dose group did not. A
In contrast to the deficits shown in retention of the condi- few other studies have employed operant tasks for testing rats
tioned taste aversion response for group H animals, litter- prenatally exposed to ethanol at ages of 30-t days (35) or
mates of these subjects did not differ from control animals 60+ days (39), showing significant learning impairments. In
in the complex maze performance. Group L rats, however, another report rats were assessed at PND 145 + or 168 + using
exhibited significant decreases in several activity-related maze a side preference test, requiring the animals to alternate
parameters. Furthermore, these individuals spent significantly presses on levers to the left and right of a center food trough
more time in the BOTTOM area and more frequently entered (45). Again, this procedure revealed no effects during the ac-
the maze arm leading into this area (significant increase in quisition with such old animals, but during later, more com-
GOALERROR4). The cage rack housing the animals to be plicated testing conditions prenatally ethanol exposed animals
tested that day was positioned close to the BOTTOM area, had a poorer performance.
thus providing visual, auditory, and olfactory cues concerning In summary, rats exposed to ethanol in utero (group H),
the presence of other rats. Therefore, increases in BOTTOM- which were not different from control animals in maze perfor-
TIME and GOALERROR compared to control animals may mance from PND 70-82, exhibited effects in a temporal re-
indicate a change in response to these stimuli rather than a sponse differentiation task when assessed after PND 140. This
learning deficit in these animals. On the other hand, the de- finding reinforces the hypothesis that long-term effects of pre-
crease in FREEZES and FREEZTIME in group H suggests natal ethanol exposure can be detected when sophisticated
that these subjects may be less emotional than the rats from behavioral techniques are used. The rodent operant test bat-
groups C and L. Deficits (8,19,31,36,37), no effect (2,26,40), tery used in this laboratory has proven to be a useful tool for
and even enhanced maze performance (10,29,42) have all been this goal.
reported for rats prenatally exposed to ethanol. Due to varia-
tions in numerous aspects of the maze experimental designs ACKNOWLEDGEMENTS
(schedules and levels of prenatal exposure, testing age, maze
characteristics, etc.), it is difficult to draw any generalized
This work was supported by grant #Cl 114/1-l of the Deutsche
conclusions from these studies. The lack of dose-related ef-
Forschungsgemeinschaft, F.R.G. (P. Clausing) and by NCTR experi-
fects in the present complex maze experiment does not appear ment #6796. S. A. Ferguson was supported through a postdoctoral
to provide information that will help to clarify the conse- appointment to the Oak Ridge Institute for Science Education. The
quences of prenatal ethanol exposure for maze learning situa- authors wish to thank C. M. Fogle for excellent technical support, L.
tions. Rushing for the determination of blood alcohol levels, and E. Allen
Evidence has accrued suggesting that the effects of in utero for the generation of Figs. 4 and 5.

REFERENCES
1. Abel, E. L. Fetal alcohol syndrome in families. Neurotoxicol. 5. Barron, S.; Gagnon, W. A.; Mattson, S. N.; Ketch, L. E.;
Teratol. lO:l-2; 1988. Meyer, L. S.; Riley. E. P. The effects of prenatal alcohol expo-
2. Abel, E. L. Prenatal effects of alcohol on adult learning in rats. sure on odor-associative learning in rats. Neurotoxicol. Teratol.
Pharmacol. Biochem. Behav. 10:239-243; 1979. 10:33; 1988.
3. Ader, R. “Strain” differences in illness-induced taste aversion. 6. Becker, H. C.; Randall, C. L. Effects of ethanol exposure in
Bull. Psychon. Sot. 1:253-254; 1973. C57BL mice on locomotor activity and passive avoidance behav-
4. Anandam, N.; Stern, J. M. Alcohol in utero: Effects on prewean- ior. Psychopharmacology (Berlin) 97:40-44; 1989.
ling appetitive learning. Neurobehav. Toxicol. 2: 199-205; 1980. 7. Blakley, P. M. Experimental teratology of ethanol. In: Kalter,
552 CLAUSING ET AL.

H., ed. Issues and reviews in teratology, vol. 4. New York: 27. Meyer, L. S.; Riley, E. P. Behavioral teratology of alcohol. In:
Plenum Press; 1986:237-282. Riley, E. P.; Vorhees, C. V., eds. Handbook of behavioral tera-
8. Blanchard, B. A.; Riley, E. P.; Hannigan, .I. H. Deficits on a tology. New York: Plenum Press; 1986:101-140.
spatial navigation task following prenatal exposure to ethanol. 28. Morrisett, R. A.; Martin, D.; Wilson, W. A.; Savage, D. D.;
Neurotoxicol. Teratol. 9:253-258; 1987. Swartzwelder, H. S. Prenatal exposure to ethanol decreases the
9. Burd, L. Fetal alcohol syndrome: Diagnosis and syndromal vari- sensitivity of the adult rat hippocampus to N-methyl-o-aspartate.
ability. Physiol. Behav. 46:39-43; 1989. Alcohol 6:415-420; 1989.
IO. Caul, W. F.; Osborne, G. L.; Fernandez, K.; Henderson, G. I. 29. Osborne, G. L.; Caul, W. F.; Fernandez, K. Behavioral effects
Open-field and avoidance performance of rats as a function of of prenatal ethanol exposure and differential early experience in
prenatal ethanol treatment. Addict. Behav. 4:311-322; 1979. rats. Pharmacol. Biochem. Behav. 12:393-401; 1980.
11. Chen, J.-S.; Driscoll, C. D.; Riley, E. P. Ontogeny of suckling 30. Paule, M. Cl. Analysis of brain function using a battery of sched-
behavior in rats prenatally exposed to alcohol. Teratology 26: ule-controlled operant behaviors. In: Weiss, B.; O’Donoghue, J.,
145-153; 1982. eds. Neurobehavioral toxicity: Analysis and interpretation. New
12. Church, M. W.; Holloway, J. A. Effects of prenatal ethanol expo- York: Raven Press; 1994:331-338.
sure on the postnatal development of the brainstem auditory evoked 3 1. Reyes, E.; Wolfe, 3 .; Savage, D. D. The effects of prenatal alco-
potential in the rat. Alcohol. Clin. Exp. Res. 8:258-265; 1984. hol exposure on radial maze performance in adult rats. Physiol.
13. Driscoll, C. D.; Chen, J. S.; Riley, E. P. Operant DRL perfor- Behav. 46:45-48; 1989.
mance in rats following prenatal alcohol exposure. Neurobehav. 32. Riley, E. P. The long-term behavioral effects of prenatal alcohol
Toxicol. 2:207-21 I; 1980. exposure in rats. Alcohol. Clin. Exp. Res. 14:670-673; 1990.
14. Driscoll, C. D.; Streissguth, A. P.; Riley, E. P. Prenatal ethanol 33. Riley, E. P.; Barron, S.; Driscoll, C. D.; Chen, J.-S. Taste aver-
exposure: Comparability of effects in humans and animal models. sion learning in preweanling rats exposed to ethanol prenatally.
Neurotoxicol. Teratol. 12:231-237; 1990. Teratology 29:325-331; 1984.
15. Dyck, G. D.; Jam, L.; Osachuk, T. A. Cl.; Falk, J.; Labinsky, 34. Riley, E. P.; Lochry, E. A.; Shapiro, N. R. Lack of response
J.; Greenberg, A. H. The Pavlovian conditioning of IL-induced inhibition in rats prenatally exposed to ethanol. Psychopharma-
glucocorticoid secretion. Brain Behav. Immun. 4:93-104; 1990. cology (Berlin) 62:47-52; 1979.
16. Ferguson, S. A.; Holson, R. R.; Paule, M. G. Effects of metha- 35. Riley, E. P.; Shapiro, N. R.; Lochry, E. A.; Broida, J. Fixed-
zoxymethanol-induced micrencephaly on temporal response dif- ratio performance and subsequent extinction in rats prenatally
ferentiation and progressive ratio responding in rats. Behav. Neu- exposed to ethanol. Physiol. Psycho]. 8:47-50; 1980.
ral Biol. 62:77-81; 1994. 36. Shaywitz, B. A.; Klopper, J. H.; Gordon, J. W. A syndrome
17. Galanter, M., ed. Children of alcoholics. New York: Plenum resembling minimal brain dysfunction (MBD) in rat pups born to
Press; 1991. alcoholic mothers. Pediatr. Res. 10:451; 1976.
18. Gavin, C. E.; Kates, B.; Gerken, L. A.; Rodier, P. M. Patterns 37. Shaywitz, B. A.; Griffieth, G. G.; Warshaw, J. B. Hyperactivity
of growth deficiency in rats exposed in-utero to undernutrition, and cognitive deficits indeveloping rat pups born to alcoholic
ethanol, or the neuroteratogen methylazoxymethanol (MAM). mothers: An experimental model of the expanded fetal alcohol
Teratology49:113-121; 1994. syndrome (EFAS). Neurobehav. Toxicol. 1:113-122; 1979.
19. Gianoulakis, C. Rats exposed to alcohol exhibit impairment in 38. Vavrousek-Jakuba, E. M.; Baker, R. A.; Shoemaker, W. J. Ef-
spatial navigation test. Behav. Brain Res. 36:217-228; 1990. feet of ethanol on maternal and offspring characteristics: Corn-
20. Grote, F. W., Jr.; Brown, R. T. Rapid learning of passive avoid- parison of three liquid diet formulations fed during gestation.
ance by weanling rats: Conditioned taste aversion. Psychonom. Alcoholism: Clin. Exp. Res. 15:129-135; 1991.
Sci. 3:163-165; 1971. 39. Vigliecca, N. S.; Fulginiti; S.; Minetti, S. A. Acute ethanol expo-
21. Hannigan, J. H. The ontogeny of SCH 23390.induced catalepsy sure during pregnancy in rats: Effects upon a multiple learning
in male and female rats exposed to ethanol in utero. Alcohol 7: task. Alcohol 6:363-368; 1989.
I l-16; 1990. 40. Vorhees, C. V. Maternal age as a factor in determining the repro-
22. Holson, R. R.; Ali, S. F.; Scallet, A. C.; Slikker, W., Jr.; Paule, ductive and behavioral outcome of rats prenatally exposed to
M. G. Benzodiazepine-like behavioral effects following with- ethanol. Neurotoxicol. Teratol. 10:23-34; 1988.
drawal from chronic delta-9-tetrahydrocannabinol administration 41. Vorhees, C. V. A fostering/cross fostering analysis of the effects
in rats. Neurotoxicology 10:605-620; 1989. of prenatal ethanol exposure in a liquid diet on offspring develop-
23. LaFiette, M. H.; Carlos, R.; Riley, E. P. Effects of prenatal ment and behavior in rats. Neurotoxicol. Teratol. 11:115-120;
alcohol exposure on serial pattern performance in the rat. Neuro- 1989.
toxicol. Teratol. 16:41-46; 1994. 42. Vorhees, C. V.; Fernandez, K. Effects of short-term prenatal
24. Mankes, R. F.; Battles, A. H.; LeFevre, R.; Van der Hoeven, T.; alcohol exposure on maze activity, and olfactory orientation per-
Click, S. D. Preferential alcoholic embryopathy: Effects of liquid formance in rats. Neurobehav. Toxicol. Teratol. 8:23-28; 1986.
diets. Lab Anim. Sci. 42:561-566; 1992. 43. West, J. R., ed. Alcohol and brain development. New York:
25. McCivern, R. F.; Raum, W. J.; Handa, R. J.; Sokol, R. Z. Oxford University Press; 1986.
Comparison of two weeks versus one week of prenatal ethanol 44. Zidenberg-Cherr, S.; Rosenbaum, J.; Keen, C. L. Influence of
exposure in the rat on gonadal organ weights, sperm count, and ethanol consumption on maternal-fetal transfer of zinc in preg-
onset of puberty. Neurotoxicol. Teratol. 14:351-358; 1992. nant rats on day 14 of pregnancy. J. Nutr. 118:865-870; 1988.
26. Means, L. W.; Russ, R. D.; Medlin, C. W.; Gray, S. L. Prenatal 45. Zimmerberg, B.; Mattson, S.; Riley, E. P. Impaired alternation
ethanol exposure in rats does not alter maze exploration or impair test performance in adult rats following prenatal alcohol expo-
visual discrimination with or without distracting stimuli. Neuro- sure. Pharmacol. Biochem. Behav. 32:293-299; 1989.
behav. Toxicol. Teratol. 8:1-S: 1986.