Pharmacology Biochemistry and Behavior, Vol. 64 No. 4, pp. 687–694, 1999
© 1999 Elsevier Science Inc.
Printed in the USA. All rights reserved
0091-3057/99/$–see front matter

PII S0091-3057(99)00136-7

687

Perinatal Exposure to the Estrogenic Pollutant
Bisphenol A Affects Behavior in Male and
Female Rats

F. FARABOLLINI,*

1

S. PORRINI* AND F. DESSÌ-FULGHERI*†

*Institute of Human Physiology, University of Siena, Siena, Italy; and

†

Department of Animal Biology and Genetics, University of Florence, Florence, Italy

FARABOLLINI F., S. PORRINI, AND F. DESSÌ-FULGHERI.

Perinatal exposure to the estrogenic pollutant Bisphenol
A affects behavior in male and female rats.

PHARMACOL BIOCHEM BEHAV

64

(4) 687–694, 1999.—Bisphenol A (BPA)
is an environmental estrogen with potentially aversive effects on public health. In rats, we studied the effects of perinatal ex-
posure to BPA on nonsocial behaviors partly influenced by gonadal hormones. BPA was administered orally to one group of
mother rats at a concentration within the range of human exposure from 10 days before mating until the weaning of the pups.
In a second group, BPA was given at a higher dosage during a critical period for brain organization, i.e., from day 14 of gesta-
tion until day 6 after birth. The offspring of the treated mothers were tested in the holeboard and the elevated plus-maze at 85
days of age. Various aspects of nonsocial behavior were affected by BPA, differently in males and females, confirming that
exposure to a weak environmental estrogen in the period of sexual differentiation of the brain can influence adult behavior.
However, contrary to our expectation, a clear masculinization of females was not observed. In general, the factor analysis in-
dicated that in treated males both the motivation to explore and anxiety are reduced, while in females, motor activity and mo-
tivation to explore are depressed. Because there were no substantial differences between the two modalities of BPA adminis-
tration, we suggest that the prolonged treatment with the low dosage compensates for the higher dosage given during a
shorter steroid-sensitive period. This may be a cause of concern for public health, given the greater incidence of prolonged
exposure of humans to low concentrations released into the environment. © 1999 Elsevier Science Inc.
Bisphenol A Estrogens Environmental estrogens Nonsocial behavior

Brain organization Holeboard Elevated plus-maze Sex differences

THERE is increasing concern about the potential negative
impact on public health of a class of environmental chemicals
with estrogenic activity: the potential effects of these environ-
mental estrogens include breast cancer (7,37), falling sperms
counts (3), and a variety of reproductive abnormalities
(5,13,14). Nevertheless, reliable direct evidence of the harm-
ful effects of these substances at the concentrations found in
the environment is lacking or controversial. As it was recently
stressed (8), much debate is generated by the difficulty of ex-
trapolating in vitro effects to an in vivo situation and by the
use of the breast and uterus as models to assess the estrogenic
activity of the substances.
Given the well-known ability of estrogens to affect sexual
differentiation of the brain during a critical period, i.e., peri-
natal life (16), it is likely that behavioral systems organized
under the influence of gonadal hormones can be affected by
precocious exposure to environmental estrogens. Behavior,
as the final point of a cascade of events, may reveal subtle ef-
fects not easily detectable at each single step of the process.
In addition to reproductive behavior, a variety of behav-
ioral activities in rodents are organized and possibly sexually
differentiated under the influence of perinatal gonadal hor-
mones (1,22,23). Social and nonsocial behaviors are masculin-
ized in female rats by neonatal androgenization of females
[see (36)]. This could be mediated by the aromatization of tes-
tosterone into estradiol.
Bisphenol A (BPA) is a particularly important environmen-
tal estrogen. Not only is it widespread in the environment, but it
is commonly ingested by humans, being released by polycar-
bonate plastics, the lining of food cans, and dental sealants
(2,19,27). BPA has a weak estrogenic activity in vitro and in
vivo (18,31), but recent evidence indicates that this substance is
able to interact with the estrogen receptor alpha (ERa) in a
unique manner, somewhat different from estradiol (12).

1

Requests for reprints should be addressed to Professor Francesca Farabollini, Istituto di Fisiologia Umana, Università degli Studi di Siena,
Via Aldo Moro, 53100 Siena, Italy.

688 FARABOLLINI, PORRINI AND DESSÌ-FULGHERI
BPA is able to affect in mice the development of reproduc-
tive organs and their function in the male offspring of mothers
fed with this chemical during pregnancy (32). The same authors
report that exposure during fetal life to other estrogenic chemi-
cals can influence the development of territorial behavior (33).
In the present work, we study the effects of precocious ex-
posure to BPA, at concentrations within the range of human
exposure and not patently teratogenic (2,24,27), on forms of
nonsocial behavior sensitive to the action of perinatal gonadal
hormones (21,35). To this purpose, we administered BPA
orally to female rats by two modalities: (a) a low dosage (40

!

g/kg) from before mating with stud males until weaning of
the pups, to analyze the behavioral effects in offspring of
chronic exposure to BPA; and (b) a high dosage (400

!

g/kg)
from day 14 of gestation until day 6 after delivery, to analyze
the effects in the offspring of acute exposure during a period
that is critical for brain organization.

METHOD

Subjects

We used 84 Sprague–Dawley rats (42 females and 42
males) that were 85 days old at the time of testing, the off-
spring of the treated mothers, and born and bred in the De-
partment of Animal Biology, University of Florence. Rats
were housed in groups of six per cage (42

"

26

"

15 cm), all
of the same sex and treatment, in a room adjacent to the test-
ing rooms, under a reversed light–dark cycle (lights off from
0900 to 1900 h). Food and water were freely available.
During the fetal and neonatal period, the animals were ex-
posed to Bisphenol A, which was administered to their moth-
ers, as subsequently reported.
Experimental procedures followed the regulations of the
European Communities Council Directive 86⁄609/EEC.

Treatment Procedure

Two dosages of Bisphenol A (FLUKA Ltd.), 40 and 400

!

g/kg body weight, were administered daily to two groups of
mothers during pregnancy and lactation, for periods of differ-
ent duration. Thirty-one females in reproductive age were ran-
domly allocated to three groups: low dosage (

n



#

11), receiv-
ing 40

!

g/kg Bisphenol A, administered from day 10 preceding
conception until the weaning of the pups; high dosage (

n



#

11), receiving arachis oil from day 10 preceding conception un-
til day 13 of gestation, followed by 400

!

g 7 kg Bisphenol A
until day 6 after delivery, and then arachis oil again until the
weaning of the pups; control (

n



#

9), receiving arachis oil from
day 10 preceding conception until the weaning of the pups.
The substance, dissolved in arachis oil at the concentration of
5.32 and 53.2

!

g/ml, for the low and high dosages, respectively,
was administered orally by micropipette, in a variable volume
according to body weight. Because the animals were trained to
receive the oil, the procedure was not stressful.
The litters were weaned on day 21. For each treatment
group, the animals were randomized and housed in mixed-sex
groups of three males and three females. The three treatment
groups consisted of: low dosage, 30 pups (15 males and 15 fe-
males), high dosage, 24 pups (12 males and 12 females), and
control, 30 pups (15 females and 15 males). Three weeks be-
fore testing, the animals were housed as groups of six per
cage, of the same sex and treatment.

Behavioral Testing

Behavioral testing started at day 85 after birth and was
completed in 3 days. Animals belonging to the same treat-
ment and sex were tested in separate sessions. Animals of the
same cage were tested in sequence. Each subject performed
the two tests in sequence, the holeboard (5 min) being fol-
lowed immediately by the elevated plus-maze (5 min). The
tests were carried out during the dark phase under dim red
light, combined with a low, indirect white light. All animals
were unfamiliar with the two apparatuses. Both apparatuses
were thoroughly cleaned at the end of each test. All sessions
were recorded with a videocamera (Sony HI8) positioned
above each apparatus. We subsequently scored the video-
tapes with software that allowed analysis of the frequency and
duration of the behavioral parameters (Observer software,
Noldus Information Technology).
The holeboard was a black Perspex box (63

"

63

"

43 cm)
with four holes (3.8 cm diameter) equally spaced on the floor.
The floor was divided into 36 squares. Frequency and dura-
tion of head dipping, self-grooming, and rearing were re-
corded, while locomotor activity was recorded as the number
of squares crossed. In addition, the percentage of internal
crosses [no. internal/total crosses

"

100] was calculated. The
number of boluses was recorded at the end of the test. The
holeboard apparatus provides separate measures of motor ac-
tivity and exploration (6,10).
The elevated plus-maze was made of black Perspex, and
consisted of 2 open arms (50

"

10 cm) and 2 closed arms (of
the same size but with 34 cm high walls) extending from a cen-
tral square platform (10

"

10). The arms were arranged so
that those of the same type were opposite each other. The ap-
paratus was elevated 60 cm above the floor. Each animal was
placed in the central square facing an open arm. We then re-
corded standard spatiotemporal measures: the numbers of en-
tries into the open and closed arms (arm entry defined as

3

⁄

4

of
body into an arm), and the time spent in both types of arms.
In addition to these parameters, the time spent in the central
square was also recorded. We also calculated the percentage
of time spent in each of the three sections of the plus-maze
(time/300 s

"

100), the number of total entries (open

$

closed
entries), and the percentage of open entries (open entries/to-
tal entries

"

100).
In addition to these traditional spatiotemporal measures,
giving separate measures of anxiety and locomotion (15,20),
we also recorded some behavioral parameters that have re-
cently been included in an enlarged version of the elevated
plus-maze test in mice (28) and rats (6) to supplement and
strengthen the traditional measures. In our experiments, these
were confined to the frequency of head dips (protruding the
head over the edge of the apparatus) and of stretched-attend
posture (stretching the body forward without movement of
the paws and then returning it to the original position). The
frequency and duration of self-grooming and rearing were
also scored.

Statistical Analysis

The holeboard and elevated plus-maze data were tested by
two-way ANOVA with the factors sex (males and females)
and treatment (low dosage, high dosage, control). One-way
ANOVA (treatment) was later applied to males and females.
Post hoc analysis was used for comparisons between groups
(Fisher LSD test; level of significance required 0.05).
A factor analysis was applied to the data from both tests
for all groups of animals; the principal components method