Chronic prenatal exposure to paroxetine (Paxil) and cognitive

development of mice offspring

H. Dix Christensena,b,*, William F. Rayburna,b,c, Christina L. Gonzaleza

aDepartment of Obstetrics and Gynecology, The University of Oklahoma Health Sciences Center, Oklahoma City, OK USAbDepartment of Pharmacology and Toxicology, The University of Oklahoma Health Sciences Center, Oklahoma City, OK USA

cDepartment of Obstetrics and Gynecology, The University of New Mexico, School of Medicine, Albuquerque, NM USA

Received 20 July 1999; accepted 7 June 2000

Abstract

This study investigated the impact on cognitive development in CD-1 mice from chronic prenatal exposure to the antidepressant

paroxetine. CD-1 mice were given either paroxetine as 30 mg/kg/day or a placebo in food bars for 2 weeks before mating and

throughout gestation. One offspring per gender from each litter was tested on each of the following tasks: tube runway, spatial maze,

passive avoidance chamber, and water straight runway followed by an unforced decision maze. Learning occurred in both genders in all

tasks ( p<0.001) with no significant differences between treatment groups at the final learning session. Juvenile runway was the only

task in which the paroxetine -exposed males demonstrated a learning rate that was slower than the placebo-exposed offspring ( p=0.06).

Post learning sessions did not show any significant treatment differences during the juvenile and adult periods during the water straight

runway, mazes, and avoidance chamber tasks. In conclusion, chronic prenatal exposure in mice of paroxetine did not impact cognition

on select tasks. D 2000 Elsevier Science Inc. All rights reserved.

Keywords: Paroxetine; Pregnancy; Serotonin uptake inhibitor; Depression

The prevalence of depression is 8±10% in women of

reproductive age [12]. Serotonin specific reuptake inhibi-

tor (SSRI) drugs appear to be the best and most widely

prescribed treatment option for depression [12]. For those

with a history of depression, a previous suicide attempt,

or several previous exacerbations of disease, we concur

with Kuller et al. [8] that prescribing medications, such

as SSRI drugs, are warranted throughout pregnancy.

Certain of these drugs, such as fluoxetine (Prozac; Eli

Lilly, Indianapolis IN) and sertraline (Zoloft; Pfizer, New

York, NY), are relatively safe during pregnancy and

lactation [12].

Paroxetine (Paxil, SmithKline Beechman, Pittsburgh,

PA), another SSRI drug specific reuptake inhibitor, has

become widely prescribed to treat persons with depression,

obsessive±compulsive disorders, or panic attacks [9]. Com-

pared with other serotonin inhibitor drugs, paroxetine is

believed to have fewer side effects and improved patient

compliance due to single day dosing [4]. Although there are

no perfectly designed studies, investigations in animals and

humans have not shown any greater risk of malformation

with these medications [1,9,12].

Animal investigators have suggested alterations in

neurologic development with prenatal exposure to high

doses of certain SSRI drugs [2,14]. Information is

lacking about in utero exposure to paroxetine and any

long-term neurobehavioral impairment. We reported that

chronic prenatal exposure to paroxetine led to heightened

responsiveness to select anxiety testing in infant off-

spring and more aggressive behavior in adult males

[11]. The objective of the present investigation was to

determine whether chronic exposure to paroxetine in

utero was associated with any alteration in the cognition

of mice offspring.

* Corresponding author. Department of Pharmacology and Toxicology,

College of Pharmacy, University of Oklahoma Health Sciences Center, P.O.

Box 26901, Oklahoma City, OK 73190, USA. Tel.: +1 - 405 -271 - 6593;

fax: +1- 405 - 271- 7477.

E-mail address: dix - christensen@uokhsc.edu (H.D. Christensen).

0892-0362/00/$ ± see front matter D 2000 Elsevier Science Inc. All rights reserved.

PII: S0 8 9 2 - 0 3 6 2 ( 0 0 ) 0 0 0 99 - 4

Neurotoxicology and Teratology 22 (2000) 733± 739

1. Methods

1.1. Animal care and drug dosing

This proposal was approved by our Institutional Animal

Care and Use Committee. Female and male CD-1 adult

mice (Charles River, New Brunswick, NJ), were transferred

to our laboratory at 8 weeks old and were allowed 1 week of

adjustment to their new environment. The mice were housed

in an American Association for the Accreditation of La-

boratory Animal Care -approved facility in an isolated

animal room with a controlled temperature (72°F), humidity

(50% � 15%), and 12-h light /12-h dark schedule (lights on

at 6 a.m.). The mice were maintained on water and a

nutritionally certified breeding diet (5015 PMI Feeds, St.

Louis, MO) in polycarbonate cages with hardwood chip

bedding (Sani-chips, PJ Murphy Forest Products, Mont-

ville, NJ) that was changed at least weekly.

The standard deviation of measured endpoints for cogni-

tive tasks has been reported as 30% in control offspring [10].

We considered, in the current study, that a 50% difference in

mean scores on task performance between the paroxetine-

and placebo-exposed groups would be necessary for there to

be a noteworthy effect. Assuming this, 10 pregnant mice

needed to be in each treatment group for there to be a 95%

confidence if no statistical significance was obtained.

Timed mating consisted of one female being matched to

a randomly selected experienced male. Food was removed

from the male's cage, and the female was introduced into the

cage for 2 h (8:00±10:00 a.m.) each day for a maximum of

5 consecutive days. The presence of a copulatory plug

indicated gestational day (GD) 0. The female mice were

then housed individually throughout gestation.

CD-1 mice were given either a placebo or 30 mg/kg/

day of paroxetine in food bars for 2 weeks before mating

and throughout gestation. This dosage of paroxetine

achieves a concentration in serum of approximately 100

ng/ml, which is comparable to the upper therapeutic range

in humans [3±5]. Deliveries were designated as being on

postnatal day (PND) 1. The pups were culled to eight per

litter on PND 5. One offspring per gender from each litter

(paroxetine: n =11; placebo: n =12) was used for each

neurobehavior task. All pups on PND 26 were sorted and

housed by gender pairs until the time for task randomiza-

tion. This sorting created subsets, so that one offspring per

gender from each litter could be tested in the different

cognitive tasks.

1.2. Neurobehavior tasks

Five standard tasks were used to assess motivation,

learning acquisition, and memory.

1.2.1. Tube runway

This motivation task used a clear 1-m tube runway

that was suspended horizontally and contained food pel-

lets (45 mg, Bioserv, Frenchtown, NJ) at one end. The

mice were given the food pellets in their home cage 2 days

before testing. Food was restricted to 5 h during the day of

each task. The time to approach the food was recorded

once daily for 5 consecutive days between PND 34 and 38

to assess learning. The task was repeated 7 days later (PND

45) as juveniles and 5 weeks later (PND 80) as adults to

evaluate memory.

1.2.2. Water straight runway

This motivation task utilized a straight 2.5-m runway

with a climb-out platform at one end. The water was at

room temperature. The mice were placed in the water facing

away from the platform. The runway was 5 cm wide and

had a water depth of 15 cm. The time to reach the platform

for six trials, with a 20-s interval between trials, was

recorded from PND 60 to 64.

1.2.3. Morris spatial maze

The Morris spatial water maze consisted of a coated

galvanized circular pool (110 cm diameter) filled with

room-temperature water and having two starting points,

four external clues, and a clear plastic 11 cm circular

platform. Each mouse was housed alone for 3 days before-

hand. The task was conducted in three modes: (1) session

1: the platform with flag was above the water line, to

confirm each mouse's ability to swim and see, (2) session

2±6: the platform was 0.5 cm below the water mixed with

white tempra paint being used to obscure its location, and

(3) session 7: the platform was removed. Each of sessions

2±6 consisted of four trials with alternating starting points

where the swim time to reach the platform was recorded.

A maximum of 60 s was allowed for each trial with 30 s

between trials. In the seventh session, the number of

crossings over the former site of the platform was recorded

for 60 s. These sessions were conducted between PND 65

and 69.

1.2.4. Cincinnati maze

The Cincinnati maze consisted of a series of nine

interlocking T mazes where the decision point was an

opening during a straight runway [13]. The maze was

constructed of white Plexiglas (25 cm high, 10 cm wide,

200 cm minimum path length). Since mice have a greater

exploratory drive than rats, most consistent results with

mice occur when the maze is run dry. At completion of

the task, the mouse was given a few Bioserv pellets. The

task consisted of nine sessions over 5 days (five in the

morning, four in the afternoon) between PND 97 and

104. The time to complete the task (maximum 300 s) and

the number of correct choices at the nine decision points

were recorded.

1.2.5. Passive avoidance

The apparatus consisted of a light and a dark chamber

connected by an opening. The mouse was placed in the

H.D. Christensen et al. / Neurotoxicology and Teratology 22 (2000) 733±739734

light chamber, and the exploratory time to enter the dark

chamber was recorded. On the next day, after the mouse

entered the dark chamber, a single electric shock (0.4 mA

for 0.5 s) was delivered on the grid floor. The time to

enter from the light chamber into the dark chamber was

recorded during four sessions: exploratory (no shock),

shock after entering the chamber (pre - shock), 24 h

post -shock, and 5 days post -shock. The task was per-

formed on PND 100 to 105.

1.3. Statistical analysis

Data for each gender in the litter were presented as a

mean � SEM. Comparisons between offspring in the parox-

etine- and control groups were conducted using either the

paired t - test or Welch's t - test or by analysis of variance for

repeated measures over time. InStat 2.05 Graphpad Soft-

ware and Statistix 4.1 Analytical Software (Tallahassee, FL)

were used. A p -value of <0.05 was considered to be

statistically significant.

2. Results

The adult mice ingested 30.0 � 0.9 mg/kg/day during

the preconception period and 22.9 � 0.7 mg/kg/day at

GD 11. Maternal weight in both groups increased as

gestation advanced. The mean duration of gestation was

19.0 days for both groups. The combined number of live

pups per litter (11.3 � 0.3) and the combined percentage of

male pups (52%) did not differ between the paroxetine-

and the placebo-exposed offspring. None of the pups

displayed any signs of acute central nervous stimulation,

including tremors and convulsions. There was no differ-

ence observed in dam±pup interactions between the

paroxetine- and placebo-exposed groups. Each of the

offspring was observed to not be impaired in locomotor

activity or in swimming.

Learning to reach the feeder was rapid while performing

the tube runway task. The first trial by the placebo group

during the juvenile period required 37 � 12 s on the first day

(exploratory time), 5.0 � 0.5 s on the second day, and

3.0 � 0.3 s by the fifth day. As shown in Fig. 1, these

exploratory and acquisition times were not statistically

significant from the two groups of exposed offspring,

regardless of gender. Later performances of the tube runway

task indicated a retention of memory. The times to perform

the task were not statistically different ( p = ns) between the

paroxetine-exposed and the placebo-exposed offspring at

PND 46 (3.5 � 0.5 s versus 3.3 � 0.5 s) and at PND 80

(7.3 � 1.2 s versus 5.9 � 0.7 s) post acquisition.

Each adult CD-1 mouse swam adequately and learned

the water runway task. Fig. 2 shows the average time to

completion of each session. The time to reach the plat-

Fig. 1. Tube runway task. Performance of the tube runway by male and female offspring exposed prenatally to paroxetine or to a placebo. Offspring were tested

daily between PND 34 and 38 for learning acquisition, and on PND 45 and 80 for memory retention. Mean�SEM. No differences in the times to reach the

feeder were found between the two treatment groups, regardless of gender.

H.D. Christensen et al. / Neurotoxicology and Teratology 22 (2000) 733±739 735

form by the placebo group was 36 � 6 s initially and

declined rapidly to 16 � 2 s by the second trial and

13 � 2 s during the fifth session. The times for each of

the five sessions were not significantly different between

the paroxetine and the corresponding placebo groups,

regardless of gender.

Results of the Morris spatial maze are shown in Fig. 3.

Each mouse was able to reach the elevated platform with the

flag. The times to reach the platform were not significantly

different between the paroxetine- and the placebo-exposed

offspring, regardless of gender. These times decreased with

each of the next five sessions with the submerged platform.

The seventh session, during which the platform was re-

moved, resulted in a similar number of crossings between

treatment groups. The number of crossings over the site of

the removed platform by the placebo-exposed offspring was

Fig. 3. Morris spatial maze. Performance of the spatial maze by adult male and female offspring exposed prenatally to paroxetine or to a placebo. Times to

reach the platform were recorded when it was placed above the water level (session 1, PND 65) and below the water level (sessions 2 to 6, PND 66 to 69).

Mean � SEM. No significant differences were present between the two treatment groups, regardless of gender.

Fig. 2. Water runway task. Performance of the straight water runway task on PND 60 to 64 by adult male and female offspring exposed to paroxetine or to a

placebo. Mean � SEM. No significant differences in times for each session were found between the two treatment groups, regardless of gender.

H.D. Christensen et al. / Neurotoxicology and Teratology 22 (2000) 733±739736

6.3 � 0.7 for the males and 6.5 � 0.6 for the females. For this

reason, the data were not graphed.

The unforced decision task, using the dry Cincinnati

maze, required more time than the spatial maze. Few

mice in either treatment group were able to initially

complete the task in 300 s. Fig. 4A displays the correct

number of choices made by the mice during sessions 1, 5,

and 9. No differences between treatment groups were

found when the task was performed, regardless of gender.

A gradual decrease in the time to complete the task and a

greater number of correct choices became evident (Fig.

4B) during the subsequent eight sessions. Performance of

the final task, passive avoidance, revealed that all mice

responded to the shock stimulus by delaying entry into

the dark chamber. This delay was still evident at 24 h and

at 5 days after the shock (Fig. 5). No differences in the

Fig. 4. Cincinnati decision maze. Performance of the decision (dry Cincinnati) maze during sessions 1, 5, and 9 by adult male and female offspring exposed to

paroxetine or to a placebo. The times to complete the task were recorded in an unforced direction (PND 97 to 104). A maximum of 300 s was allowed for each

session, and the maximum number of correct choices was 9. Mean � SEM. No significant differences in the correct number of choices (A) and in the

completion time of the task (B) were present between the two treatment groups, regardless of gender.

H.D. Christensen et al. / Neurotoxicology and Teratology 22 (2000) 733±739 737

entry times were observed between the paroxetine- and

placebo-exposed offspring, regardless of gender.

3. Discussion

This investigation was undertaken to explore the effect

of chronic in utero exposure to the highly potent SSRI

drug paroxetine on the cognitive development of offspring.

Studies in nongravid adult rats have demonstrated that this

drug temporarily lowers both 5-HT and 5 hydroxyindo-

leacetic acid in all brain regions [3]. Any centrally acting

drug given prenatally may interfere with developmental

processes in the fetal brain, leading to biochemical altera-

tions or to subtle disturbances at the cellular or subcellular

level. Most receptor systems in the brain develop during

fetal life, and inadequate or precocious activation may,

in theory, result in a long-term disorganization of cogni-

tive processes.

Much of the early work with these serotonin-modulating

drugs involved toxic doses and was confined to perinatal

survival and to gross physical markers at early postnatal

ages. Like similar studies using rodents, we found in the

present investigation using a 30 mg/kg /day dose that

paroxetine did not impair fertility or reduce litter size

[1,6,7]. A standard battery of cognitive tasks was examined

in the current investigation to screen for any deficiencies.

Offspring exposed to paroxetine performed motivation,

learning, and memory tasks in manners that were indistin-

guishable from the placebo-controlled group. Tasks were

performed at the same juvenile and adult periods under

identical conditions.

We were unable to locate any other studies in rodents

examining effects from paroxetine on learning and memory

after prenatal exposure. In a similar study examining multi-

ple noncognitive behavior tasks, we observed no statistical

differences between mice offspring exposed in utero to

paroxetine or to a placebo in many early development task

(geotaxis, homing, social play) and in exploratory activities

throughout development [11]. Performances during a de-

pression task (forced swim) and anxiety tasks (elevated plus

maze) were indistinguishable between the two treatment

groups, regardless of gender. Offspring exposed to parox-

etine had a 15±25% increase in separation vocalization

( p<0.05) and a significant increase in male aggression

( p<0.03) during cage changing.

Despite extrapolation variations between mammalian

species, it was reassuring to find that there was no major

impact on cognition among CD-1 mice offspring exposed

prenatally to paroxetine at dose levels used in this investiga-

Fig. 5. Passive avoidance chamber. Performance of the passive avoidance task on PND 100 to 105 by adult male and female offspring exposed prenatally to

paroxetine or to a placebo. Times to enter the dark chamber from the light chamber were recorded during the initial exploratory session, second exploratory

session (preshock) followed by shock therapy, 24 h post shock, and 5 days post shock. Mean � SEM. No significant differences were found between the two

treatment groups, regardless of gender.

H.D. Christensen et al. / Neurotoxicology and Teratology 22 (2000) 733±739738

tion. We may conclude, after performing several examina-

tions on juvenile and adult offspring, that prenatal exposure

to paroxetine was not associated with an enhancement or a

decrement in cognitive testing. Findings from the current

study provide additional information to the prescribing

clinician and to the scientist who is interested in under-

standing more about prenatal effects of serotonin-modulat-

ing drugs on the developing brain.

Acknowledgments

Presented at the Annual Meeting of Neurobehavior

Teratology Society, Keystone, CO, 1999. Financial assis-

tance from the John W. Records Perinatal Research Fund.

References

[1] T.M. Cabrera, G. Battuglia, Delayed decrease in brain 5 - hydroxytryp-

tamine receptor density and function in male rat progeny following

prenatal fluoxetine, J. Pharmacol. Exp. Ther. 269 (1994) 637±645.

[2] A.M. Caccias, A. Codegoni, C. Fracasso, S. Garattini, The effects of

single and repeated anorectic doses of 5 - hydroxytryptamine uptake

inhibitors on indole levels in rat brain, Br. J. Pharmacol. 110 (1993)

355± 359.

[3] H. Christensen, T. Kupiec, J. Jacobsen, J. Stewart, C. Gonzalez, W.

Rayburn, Tissue concentration from consumption of paroxetine (Pax-

il) in mice, Teratology 20 (1998) 365.

[4] K.L. Dechant, S.P. Clissold, Paroxetine: a review of its pharmacody-

namic and pharmacokinetic properties, and therapeutic potential in

depressive illness, Drugs 41 (2) (1991) 225± 253.

[5] R.E. Haddock, A.M. Johnson, P.F. Langley, D.R. Nelson, J.A. Pope,

D.R. Thomas, F.R. Woods, Metabolic pathway of paroxetine in ani-

mals and man and the comparative pharmacological properties of its

metabolites, Acta Psychiatr. Scand. (Suppl. 350) (1980) 24±26.

[6] W. Inman, K. Kubota, G. Pearce, L. Wilton, PEM report number 6,

Pharmacoepidemiol. Drug Saf. 2 (1993) 393±422.

[7] N.A. Kulin, A. Pastuszak, S.R. Sage, B. Schick-Boschetto, G. Spivey,

M. Feldkamp, K. Ormond, D. Matsui, A. Stein-Schechman, L. Cook,

J. Brochu, M. Rieder, G. Koren, Pregnancy outcome following ma-

ternal use of the new selective serotonin reuptake inhibitors, JAMA

279 (1998) 609± 610.

[8] J. Kuller, V. Katz, M. McMahon, S. Wells, R. Bashford, Pharmaco-

logic treatment of psychiatric disease in pregnancy: fetal and neonatal

effects, Obstet. Gynecol. 87 (1996) 789±794.

[9] Paxil [package insert], SmithKline Beecham Pharmaceuticals, Missis-

sauga, Ontario, 1995.

[10] W.F. Rayburn, H.D. Christensen, C. Gonzalez, A placebo- controlled

comparison between antenatal betamethasone and dexamethasone for

fetal maturation: Effects on long- term neurobehavior in mouse off-

spring, Am. J. Obstet. Gynecol. 176 (1) (1997) 842± 885.

[11] W.F. Rayburn, C.L. Gonzalez, H.D. Christensen, T.C. Kiepiec, S.A.

Jacobsen, J.D. Stewart, Effect of antenatal exposure to paroxetine

(Paxil) on growth and physical immaturation of mice offspring, J.

Matern. -Fetal Med., in press.

[12] L.A. Schmidt, B.D. Greenberg, G.B. Holzman, J. Schulkin, Treatment

of depression by obstetrician± gynecologists: A survey study, Obstet.

Gynecol. 90 (1997) 296± 300.

[13] C.V. Vorhees, W.P. Weisenburger, K.D. Acuff - Smith, O.R. Mink, An

analysis of factors influencing complex water maze learning in rats:

Effects of task complexity, path order and escape assistance on per-

formance following prenatal exposure to phenytoin, Neurotoxicol.

Teratol. 13 (1991) 213±222.

[14] C.V. Vorhees, K.D. Acuff - Smith, M.A. Schilling, J.E. Fisher, M.S.

Buelke - Sam, J. Buelke- Sam, A developmental neurotoxicity evalua-

tion of the effects of prenatal exposure to fluoxetine in rats, Fundam.

Appl. Toxicol. 23 (1994) 194± 205.

H.D. Christensen et al. / Neurotoxicology and Teratology 22 (2000) 733±739 739