Comparative Biochemistry and Physiology Part C 134 (2003) 365374

Effects of exposure to 17a-ethinylestradiol during early

development on sexual differentiation and induction of vitellogenin
in zebrafish (Danio rerio)
Gessbob, Leif Norrgrenb, Gitte I. Petersena
Lene Andersenc,*, Henrik Holbechc, Asa
b

a
DHIWater & Environment, Agern Alle 11, Hrsholm DK-2970, Denmark
Swedish University of Agricultural Sciences, Faculty of Veterinary Medicine, Department of Pathology, Uppsala 750 07, Sweden
c
University of Southern Denmark, Institute of Biology, Campusvej 55, Odense, M DK-5230, Denmark

Received 16 June 2002; received in revised form 16 December 2002; accepted 10 January 2003

Abstract
To determine the critical stage of zebrafish development where exposure to xenoestrogens can affect sex ratio and
vitellogenin induction, zebrafish (Danio rerio) were exposed to 17a-ethinylestradiol (actual concentration 15.4"1.4 ng
EE2yl) during early development: from fertilisation to hatch; hatch to 10 days post hatch (dph); 1020 dph; 2030
dph; 2040 dph; 2060 dph; fertilisation to 25 dph; or hatch to 60 dph. Vitellogenin was measured in whole body
homogenate 30 dph by ELISA and sex ratio was determined 60 dph by histological examination of the gonads. All
exposure periods significantly induced vitellogenin synthesis and affected the sex differentiation leading to development
of ovo-testis or complete feminisation of the exposed fish depending on exposure period. Complete sex reversal was
obtained in groups exposed from 20 to 60 dph and hatch to 60 dph. The half-life for degradation of vitellogenin was
calculated. Juvenile zebrafish were exposed to 15.4"1.4 ng EE2yl (actual concentration) from fertilisation to 25 dph
and transferred to clean water, after which weekly measurements of vitellogenin concentration in whole body homogenate
were performed until day 46 post hatch. The half-life of vitellogenin was 2.4 days.
2003 Elsevier Science Inc. All rights reserved.
Keywords: Zebrafish; Danio rerio; Vitellogenin; Endocrine disrupters; Sex reversal; Ovotestis

1. Introduction
During the last decades it has become more and
more obvious that a number of organic pollutants
cause adverse health effects to animals consequent
to changes in endocrine functions (Gillesby and
Zacharewski, 1998). Such chemicals are known as
endocrine disrupters and most attention has been
paid to compounds having estrogenic effects
although compounds with anti-estrogenic, andro*Corresponding author. Tel.: q45-6550-2760; fax: q456593-0457.
E-mail address: lene@biology.sdu.dk (L. Andersen).

genic and anti-androgenic effects are known as

well. Effects of endocrine disrupters in fish include
reduced fertility (decreased sperm number and
quality, or egg number), induction of the synthesis
of the yolk precursor protein vitellogenin (VTG)
in males and juveniles and effects on the development of the gonads (Billard et al., 1981; Ander et al., 2000).
sen et al., 2001; Orn
Endocrine disrupters are effective at sub-lethal
concentrations (Stahlschmidt-Allner et al., 1997)
and it is, therefore, not possible to detect effects
of such compounds using existing standard tests
with endpoints such as mortality. Great effort has,

1532-0456/03/$ - see front matter 2003 Elsevier Science Inc. All rights reserved.
PII: S 1 5 3 2 - 0 4 5 6 0 3 . 0 0 0 0 6 - 1

366

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

therefore, been put into this area to develop methods for testing endocrine disrupters containing new
endpoints. At the 1st OECD Expert Consultation
on Testing in Fish in 1998 gonadal development
and VTG induction was, therefore, suggested as
the key endpoints for in vivo evaluation of estrogenic compounds in fish (EDF2, 2000).
One of the fish suggested as a test animal was
the zebrafish, Danio rerio. In this species, males
pass through a stage of juvenile hermaphroditism.
Approximately 10 days post hatch the differentiation of the gonads begins and all fish, irrespective
of their definitive sex, develop ovaries. At approximately day 23 post hatch the ovaries of approximately half of the fish start to degenerate and are
transformed into testes. This process is completed
at approximately 40 days post hatch. In the remaining fish, the development and maturation of ovaries continue (Takahashi, 1977; Uchida et al.,
2002).
Gonadogenesis in fish can be a very complex
and plastic process. Although sex determination is
under genetic control, the final differentiation of
the gonads in fish also depends on endocrine
signals, i.e. estrogens and androgens (Yamamoto,
1969; Arcand-Hoy and Benson, 1998; Campbell
and Hutchinson, 1998). In most gonochoristic fish,
the germ cells of the undifferentiated gonads are
sexually bipotential (Kobayashi et al., 1991). During specific critical periods of early development,
changes in sex hormone levels can, therefore,
affect the final sex independently of the genetic
sex (Donaldson and Hunter, 1982). Due to the
lability of sex differentiation in fish, exposure to
endocrine disrupters during certain critical periods
of early development can lead to sex reversal.
To determine the critical stage where sex differentiation and induction of VTG in zebrafish can
be affected by estrogenic compounds, zebrafish
were exposed to 17a-ethinylestradiol at stages
according to the observations on gonadal development by Takahashi (1977). If exposure to endocrine disrupters during this critical period is
sufficient to manifest effects of endocrine disrupters, this may considerably lower the costs of
future short-term methods for testing chemicals for
an endocrine disrupting effect.
2. Materials and methods
2.1. Test substance and artificial zebrafish medium
The test substance was 17a-ethinylestradiol
(EE2) (ICN Biomedicals Inc., Ohio, USA) was

dissolved in methanol. The final concentration of

methanol did, in accordance with the OECD guideline no. 202 (1992), not exceed 100 ml methanol
ly1 artificial zebrafish medium. Zebrafish were
kept in artificial zebrafish medium, i.e. Millipore
water filtrated through active charcoal and added
salts. The medium was aerated until a stabile pH
(pH 7.8"0.2) was reached (approx. 24 h). The
approximate composition was Ca2q 80.15 mgyl,
Cly 144.53 mgyl, Mg2q 12.18 mgyl, SO2y
48.13
4
mgyl, Naq 17.10 mgyl, HCOy
3 45.40 mgyl and
Kq 3.02 mgyl. Fish sampled for histological examination were fixed in Lilies neutral buffer (composition: 4 gyl NaH2PO4, 6.5 gyl NaHPO4, 4%
formalin).
2.2. Determination of water concentrations of EE2
In order to determine the actual exposure concentration of EE2, weekly water samples (1 l)
where analysed by LC-MS at the University of
Southern Denmark.
2.3. Spawning procedures
Sexually mature zebrafish (Danio rerio) were
bought from a local supplier. Eight females and
16 males were placed in breeding aquaria
(25=35=52 cm). The aquaria were connected to
a recycling system (recycling volume 125 l), and
the water was changed twice a month. The water
temperature was 25"1 8C, and the lightydark
cycle was 14:10 h. Adult fish were fed TetraMin
Hauptfutter (Tetrawerke, Mell, FRG) and Artemia
ssp. nauplii (San Francisco Bay brand).
Before first spawning, the fish were acclimatised
for at least 12 days to artificial zebrafish media.
After acclimatisation, a spawning tray was placed
in the aquarium in the afternoon. The spawning
tray consisted of a stainless steel tray
(18.5=30=7 cm) covered with a stainless steel
grid (mesh size 4.0 mm), to prevent the adults
from eating the eggs. Two to three breeding trees
(untwisted nylon rope) were attached to the upper
part. The spawning of zebrafish is stimulated by
light and a maximum of 4 h after the light was
turned on eggs they were gently collected from
the spawning trays by use of a glass pipette (inner
diameter F4 mm). Under binocular microscope
(1225= magnification) fertilised eggs were collected, and egg batches showing high fertilisation
success were used for further studies.

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

367

Fig. 1. Exposure periods in the study of critical periods in zebrafish. Development stages of zebrafish from Takahashi (1977).

2.4. Maintenance of eggs and larvae

Viable eggs were transferred to 1 l-glass beakers
and kept for 2 weeks. The water was renewed
daily, and the temperature was 25.0"1 8C. Fourteen dph, the larvae were transferred to tanks of
stainless steel (21=33=21 cm), equipped with an
overflow pipe; the volume of the test solution was
8 l. During the exposure, a continuous flowthrough system was employed. Artificial zebrafish
medium was fed from the holding tank by a Gilson
HPLC pump (306) through a manometric module
(Gilson model 506), and further through a sample
changer (Gilson sample changer model 223) to
six mixing vessels. To obtain the desired concentrations an appropriate amount of a stock solution
was fed to the mixing vessels by a Gilson 402
diluter controlled by a computer program (developed by Bio-Lab, Riskov, DK). The mixed test
solutions were pumped into the test aquaria (6
mlymin) resulting in an exchange of the test
solution once daily. The stock solution was
changed once weekly. To avoid EE2 adsorption to
faeces and leftovers from feeding the aquaria were
cleaned daily. The larvae were kept in stainless
steel net-containers (mesh size 0.3 mm) (working
volume 3.5 l) to ease the daily cleaning and
decrease the disturbance of the larvae.
Feeding of larvae was initiated 23 days post
hatch, as they became actively free-swimming.

The larvae were fed Tetra AZ 100starter food

for ornamental fish (Tetra Werke, Melle, FRG)
until 2 weeks post hatch. After 2 weeks, the starter
food was substituted with newly hatched Artemia
ssp. nauplii. Larvae were fed Artemia ssp. nauplii
only until 4 weeks post hatch after which they
were fed ground TetraMin Hauptfutter as well.
Larvae were fed twosix times daily.
2.5. Determination of the elimination, half-life, of
VTG
Newly fertilised eggs (maximum 4-h-old) were
exposed to 15.4=1.4 ng EE2yl for 4 weeks (one
replicate of 200 eggs). The control group was
exposed to artificial zebrafish medium without
methanol. At the end of the exposure period (25
dph), the larvae were transferred to 10-l aquarium
containing zebrafish media having not added EE2
and methanol, and kept there until 60 dph. Twice
a week 75% of the water was renewed. Fish were
sampled regularly, i.e. day 25 (ns20), 32 (ns
20), 39 (ns20) and 46 (ns10) post hatch, for
analysis of VTG. For the group exposed to
15.4"1.4 ng EE2yl, the half-life was calculated
using the equation of a first-order process.
2.6. Determination of critical periods
In accordance with Fig. 1, the larvae (one
replicate of 100 eggs per exposure period) were

368

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

exposed to 15.4"1.4 ng EE2yl for different periods from fertilisation (maximum 4-h-old) until
sexual maturity, i.e. 60 dph. The control group was
exposed to artificial zebrafish media without EE2
and methanol. Fish for analysis of induction of
VTG synthesis (ns20) were sampled at 30 dph.
At the end of the experiment, the surviving fish
were fixed in Lilies neutral buffer in for sex
determination. The mortality of all groups was
recorded daily.
2.7. Detection of VTG by ELISA
The concentration of VTG was measured on
whole body homogenate using ELISA as described
in Holbech et al. (2001). Due to the small amount
of body tissue of juvenile zebrafish the method
was slightly modified. The minimum amount of
body tissue for VTG detection using this method
is approximately 25 mg. Thus, in order to obtain
an appropriate amount of supernatant for the ELISA, fish were weighed and pooled (110 fish per
measurement). In each group a total number of 20
fish were used for each measurement of VTG. The
fish were frozen in liquid nitrogen and kept at y
80 8C for later determination of VTG. The fish
were homogenised with a plastic pestle, and 10
times the tissue weight of ice-cold homogenate
buffer (50 mM TrisHCl pH 8.0, 0.02% aprotinin,
0.1 mM PMSF) was added. The homogenate was
centrifuged for 60 min at 50 000=g at 4 8C. The
supernatant was removed and analysed for VTG
by a direct, non-competitive sandwich ELISA.
For the groups in which the exposure had been
terminated before day 30 post hatch, an estimate
of the VTG concentration at the end of exposure
was calculated. Based on the results from the
previous described experiment, elimination of
VTG followed a first-order process. The following
equation was, therefore, used to calculate the VTG
concentration: CendsC25 dphq10(kety2303), where
C25dph is the VTG concentration measured 25 dph,
t is the time from end of exposure to 25 dph in
days, Cend is the VTG concentration at the end of
the exposure period, and ke is the elimination rate
constant calculated from t of VTG.
2.8. Sex determination by histology
Fish were fixed in toto in Lilies neutral buffer
at s tissueybuffer ratio of 1:10. The fish were
dehydrated in graded series of ethanol solutions

(70100%), and embedded in paraffin. To determinate the sex, longitudinal sections (34-mm
thick) of the fish were prepared using a microtome.
Sections were stained by hematoxylin-eosin (HE)
and examined by light microscopy.
2.9. Statistics

x2-Analysis was used to examine differences in

sex ratio and mortality between the control group
and treatments group. To meet the criteria for x2test, intersex individuals were added to the group
of males. ANOVA or when variances were unequal
the non-parametric KruskalWallis ANOVA was
used for detection of differences in VTG content
between groups. Individual differences were
detected by appropriate ad hoc tests (Dunns method and Tukey test).
3. Results
3.1. Determination of water concentrations of EE2
The control group contained no EE2, while the
water concentration of EE2 in the exposure group
was 15.4"1.4 ng EE2yl.
3.2. Determination of the elimination, half-life, of
VTG
The VTG measurements of this experiment evaluated the fate of VTG after termination of the
exposure. At the end of the treatments, i.e. 25 dph,
the VTG concentrations were significantly
increased in fish exposed to 15.4"1.4 ng EE2yl
when compared to the control fish (P-0.05). In
the control group, no significant change in the
VTG concentration from termination of exposure,
2546 dph was found (Fig. 2a). In the group
exposed to EE2 all concentrations of VTG were
significantly different from each other at all times
(Fig. 2b). When the log10 of the mean concentration of VTG in whole body homogenate was
plotted against time after end of exposure, a
straight line resulted (Fig. 3). The metabolismy
elimination of VTG in fish exposed to 15.4"1.4
ng EE2yl can thus be described as a first-order
process. The half-life of the VTG concentration in
fish exposed to 15.4"1.4 ng EE2yl was 2.4 days.

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

369

post hatch, any mortality in these fish, which may

have occurred from the time of sampling to 60
dph is, therefore, assumed to be negligible.
3.3.2. VTG measurements
The VTG concentrations of zebrafish exposed
to 15.4"1.4 ng EE2yl during different periods
from fertilisation to 30 dph were measured 30 dph
in whole body homogenate (Fig. 4). All treatments
increased the VTG concentration significantly
from that of the control group, even when the
exposure only included the embryo stage. Since
elimination andyor metabolism of VTG continue
from the end of exposure to the time of sampling,
high levels of VTG must have occurred in these
fish. To compare the levels of VTG of each group
at the end of exposure and thereby be able to
reveal possible sensitive periods for induction of
VTG, we estimated the original concentrations of
VTG at the end of the exposure period using the
VTG half-life and the equation for a first-order
process (Fig. 5).

Fig. 2. Elimination of VTG in juvenile zebrafish (30 days post

hatch). Notice the linear scale for the y-axis in the control
group and the logarithmic scale for the y-axis in the group
exposed to 15.4"1.4 ng EE2yl. (a) Control group, (b) group
exposed to EE2. VTG concentrations (ng VTGyg fish"S.D.)
measured in whole body homogenate (20 fish were sampled
and 110 fish were pooled to obtain 25-mg tissue). *Significantly different from control group of the same day and all
other groups (P-0.05).

3.3. Determination of critical periods

3.3.1. Mortality
Except for the group of fish exposed from 10
to 20 dph, no significant difference in the mortality
of fish exposed to 15.4"1.4 ng EE2yl for different
periods from fertilisation to 60 days post hatch
was observed compared to the control group (Table
1). The mortality during the embryo stage was
less than 2% in all groups. Mortality was mainly
observed from day 3 to 20 post hatch and reached
its maximum at approximately day 18 post hatch
after which less than 1% of the fish died. As all
fish for VTG analysis were sampled after day 20

3.3.3. Sex ratio

In the control group, the fish had either testes
or ovaries. The ovaries contained oocytes in different developmental stages and sperm in different
developmental stages was found in the testes.
In zebrafish exposed to 15.4"1.4 ng EE2yl for
different periods from fertilisation to 60 dph, ovotestes were found in fish exposed from fertilisation
to hatch (fh), hatch to 10 dph (h10), 10 to 20
dph (1020), 20 to 30 dph (2030) and from
hatch to 25 dph. Fig. 6 shows the sex ratio of all
groups.
The gonads of the individuals having ovo-testis
contained one or a few primary oocytes in close
connection with the testicular tissue. In all groups
exposed at different periods until 30 dph, the
gonads were fully mature at the end of the experiment. The ovaries contained oocytes in different
developmental stages, and sperm was found in the
testes.
A high incidence of ovo-testes (59%) occurred
in fish exposed from 20 to 40 dph and the sex
ratio of this group was significantly different from
the control group (P-0.05). Unlike the previously
described ovo-testes, the oocytes of these ovotestes were atretic and occurred more frequently.
Furthermore, the gonads looked immature with
primordial germ cells represented and as if they

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

370

Fig. 3. Elimination of VTG after end of exposure to 15.4"1.4 ng EE2yl, measured in whole body homogenate (ng VTGyg fish"S.D.)
(20 fish were sampled and 110 fish were pooled to obtain 25-mg tissue). Logarithm to the mean concentration of VTG plotted against
days after end of exposure. Zero equals 25 dph, 1s26 dph, etc. Half-life was 2.4 days.
Table 1
Mortality (%) 60 dph of zebrafish, Danio rerio, exposed to 15.4"1.4 ng EE2yl during different periods from fertilisation to 60 dph.
Nsnumber of eggs exposed from beginning of the experiment. *sSignificantly different from the control group (P-0.05)
Experimental
period

Control
ns100

Fert. to
hatch
ns100

Hatch to
10 dph
ns114

10 to
20 dph
ns100

20 to
30 dph
ns100

Hatch to
25 dph
ns100

20 to 40
dph
ns100

20 to 60
dph
ns100

Hatch to
60 dph
ns100

Mortality (%)

47

59

50

63*

53

46

53

54

53

Fig. 4. VTG concentration (ng VTGyg fish"S.D.) of 30-day old zebrafish (post fertilisation) after exposure to 15.4"1.4 ng EE2yl
during different periods until 30 dph. Twenty fish were sampled and 110 fish were pooled to obtain 25-mg tissue. f To hsfrom
fertilisation to hatch, h to 10sfrom hatch to 10 dph, etc. *Significantly different from control group (P-0.05). **Significantly different
from all groups (P-0.05).

were being inygoing through a developmental

stage.
In the group exposed from 20 to 60 dph and
from hatch to 60 dph no intersex was found, and

all individuals developed into females. The ovaries

of the latter group were small and often contained
few primary oocytes situated in close connection
to the much enlarged liver. In this group, 27

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

371

Fig. 5. Estimated concentration of VTG (ng VTGyg fish"S.D.) at the end of exposure to 15.4"1.4 ng EE2yl. Twenty fish were
sampled and 110 fish were pooled to obtain 25-mg tissue. f To hsfrom fertilisation to hatch, h to 10sfrom hatch to 10 dph, etc.
*Significantly different from control group (P-0.05). **Significantly different from all groups (P-0.05).

animals were fixed for examination of the gonads,

but the sex of 44% of the animals could not be
determined. Occurrence of differentiated or undifferentiated gonadal tissue was not found in these
animals, and they may have still been juvenile.

4. Discussion
4.1. Determination of water concentrations of EE2
The actual concentration of EE2 in the exposure
water was lower than the nominal concentration.

Fig. 6. Sex ratio of zebrafish exposed to 15.4"1.4 ng EE2yl during different periods from fertilisation to 60 days post hatch. There
was only one replicate. f To hsfrom fertilisation to hatch, h to 10sfrom hatch to 10 dph, etc. *Sex ratio significantly different from
control group (P-0.05).

372

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

Reduction of the nominal concentration is a problem when working with lipophilic compounds such
as EE2. The fish were, therefore, kept in either
glass jars or aquaria made of stainless steel to
reduce the absorption of EE2 to the surfaces. Even
though the aquaria were cleaned daily and the test
solution was changed daily, microbial growth as
well as absorption of EE2 to faeces and food
leftovers could not be totally avoided. Nimrod and
Benson (1998) observed that after only 4 h of
immersion, juvenile Japanese medaka (Oryzias
latipes) had absorbed approximately half of the
concentration of estradiol from the water.
4.2. Determination of critical periods
4.2.1. VTG
The effects of EE2 on VTG induction were
measured at 30 dph, and all exposure periods
caused a significant increase in the VTG concentration. For the groups where the exposure was
terminated before day 30 post hatch, the VTG
concentration at the end of exposure was estimated
taking the elimination and metabolism of VTG
into account.
According to these estimates, the highest amount
of VTG is induced after exposure during the
embryo stage. This may be due to a large production of VTG by the embryo. However, EE2 is a
lipophilic compound wKows4.12 (Syracuse
Research Corporation, 1999)x and is likely to
accumulate in the lipid fraction of the yolk sac.
The yolk sac may act as a toxicant sink releasing
EE2 when the yolk is metabolised. The VTG may,
therefore, not be synthesised during the embryo
stage, but rather during a later stage of development. Since bioaccumulation of EE2 in the embryo
stage or in larval tissue has not been taken into
account when calculating the estimates, these may
be biased. In the doseresponse experiment, fish
were exposed from fertilisation to 25 dph, i.e.
including the embryo stage, while in the study of
critical periods fish were exposed from hatch to
25 dph. When the embryo stage was included in
the exposure, the concentration of VTG measured
25 dph was approximately one and a half times as
high in fish exposed from fertilisation to 25 dph
(4.33"2.42=106 ng VTGyg fish) (Fig. 4) compared to the estimated concentration in fish
exposed from hatch to 25 dph (2.68"0.93=106
ng VTGyg fish) (Fig. 5). Thus, whatever causes
the high concentration of VTG found after expo-

sure during the embryo stage, including this stage

in a treatment clearly affects the induction of VTG
synthesis at later stages.
The estimated VTG contents of the groups
exposed from hatch to 10 dph and from 10 to 20
dph were significantly lower than the other groups
(Fig. 5). In addition, no significant difference in
the concentration of VTG was found whether the
fish had been exposed from hatch to 30 dph or
from 20 to 30 dph (Fig. 4). This may indicate that
the enzyme system of the liver may not be fully
developed until approximately 20 dph.
Unexposed male adult zebrafish have VTG concentrations at approximately 702 ng VTGyg fish,
measured in whole body homogenate (Petersen et
al., 2000), while the VTG concentration of nonexposed juvenile zebrafish was 224"78 and
247"103 ng VTGyg fish, measured 25 and 30
dph, respectively. After exposure to 20 ng EE2yl
for 30 days, the VTG concentration in adult males
was approximately 2.27=106 ng VTGyg fish
(Petersen et al., 2000). In the present experiment,
the level of VTG in juveniles measured after
exposure from hatch to 30 dph to 15.4"1.4 ng
EE2yl was 2.26"1.49=106 ng VTGyg fish (Fig.
4). Thus, even though the normal levels of VTG
in juveniles are approximately three times less than
levels of adult males, it is still possible for the
juvenile zebrafish to induce levels similar to adult
males. This suggests that the potential of the liver
to synthesise VTG is present in juveniles, but in
juveniles the natural E2 concentration is normally
lower than in males.
4.2.2. Sex ratio
Intersexyovo-testes do not normally appear in
adult zebrafish (Takahashi, 1977; Yamazaki, 1976;

Orn
et al., 2000). Individuals with ovo-testes
appeared in all groups exposed for different periods from fertilisation to 30 dph (Fig. 6). The ovotestes of the intersex animals consisted of fully
mature testicular tissue with sperm present as well
as onefour oocytes. Ovo-testis may appear when
a treatment with exogenous sex hormones does
not completely override the endogenous endocrine
signals (Johnstone et al., 1978; Pandian and Sheela, 1995).
In particular, the group exposed from 20 to 40
dph showed a high frequency of fish with ovotestes (59%). The ovo-testes of this group looked
differently than those previously described. The
exposure period corresponds to the stage at which

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

the transformation process occurs and the ovotestes looked like a gonad going through this
developmental stage as described by Takahashi
(1977) and Uchida et al. (2002). Thus, it is
possible that exposure to EE2 during this particular
stage somehow delays the onset of the differentiation of testes and that the differentiation of the
gonads of these animals may not have been fully
completed at the time of killing. The transformation of ovaries into testes can be delayed in fish
having undergone retarded growth during development (Takahashi, 1977). Large amounts of energy are required in VTG synthesis. As the VTG
content was high in this group, energy normally
used for transformation of ovaries into testes may
instead have been used for VTG synthesis slowing
down the development of the gonads.
As described previously, EE2 has a feminising
effect on the differentiation of the gonads when
et al.,
exposed from fertilisation to adulthood (Orn
2000). Similar results were obtained in the present
study, where exposure from hatch to 60 dph
completely feminised all fish.
If the fish were exposed to EE2 from the onset
of differentiation to sexual maturity, i.e. exposure
from 20 to 60 dph, feminisation was complete.
These findings suggest that the critical stage at
which EE2 has feminising effect on the differentiation of the gonads of zebrafish is from 20 to 60
dph. Furthermore, these results suggest that the
transformation process of ovaries into testes may
be caused by a decrease in the plasma level of
estrogens and in order to override the endogenous
signals estrogens must be applied until the fish
have become sexually mature. In cichlid fish, the
presence of estrogen induces the female phenotype
while the absence of estrogen leads to development
of testes (Hackmann and Reinboth, 1974).
In the two groups containing only females, the
development of ovaries was severely depressed.
Furthermore, in the group exposed from hatch to
60 dph, it was not possible to determine the sex
of 44% of the individuals, which may have been
juvenile. This was not observed in any of the other
groups. In an experiment by Petersen et al. (2000),
exposure of zebrafish to 100 ng EE2yl showed
that the longer the exposure period the higher was
the proportions of fish that could not be sex
determined. The synthesis of VTG in these groups
is high and as mentioned previously, this may
represent a considerable investment of energy and
a perversion of the metabolism that normally

373

would have led to development and maturation of

the gonads. This suggests that exposure to EE2
can delay development, which may be the case in
the group exposed from 20 to 40 dph, or even
inhibit gonadal development as seen in the group
exposed from hatch to 60 dph. In rainbow trout
(Oncorhynchus mykiss), exposure to 2 ng EE2yl
for 3 weeks reduced the gonad-somatic index with
40% (Jobling et al., 1996).
The critical period during which it is possible
to successfully reverse the sex of an individual
may be difficult to establish, as that period may
not be restricted to one particular stage during
development (Pandian and Sheela, 1995). In a
mixed sex population of coho salmon (O. kisutch),
bimodal distributions of sex steroids occur during
yolk sac absorption and during the development
of the gonads, which suggest that in this species
the critical period is not restricted to one single
stage (Feist et al., 1990). The slight increase in
females in the group exposed during the period at
which differentiation of ovaries occurs in all individuals (1020 dph) together with a complete
feminisation of the fish exposed from 20 to 60
dph suggest that the period where zebrafish can
be affected by exogenous estrogens may not be
restricted to only one stage during ontogenesis
(Fig. 6).
In conclusion, even relatively short periods of
exposure to a nominal concentration of 15.4"1.4
ng EE2yl during early life stages induce synthesis
of VTG and affect the sex differentiation of zebrafish leading to either development of ovo-testis or
complete feminisation of the population depending
on the stage of exposure. As EE2 is a potent
endocrine disrupter (Piferrer and Donaldson, 1992;
et al., 2000), the effect of less potent endoOrn
crine disrupters may not be as pronounced. Even
though effect of exposure to EE2 was found during
all exposure stages, the critical period at which the
sex differentiation of zebrafish can be significantly
affected by EE2 is from 20 to 60 dph. However,
Donaldson and Hunter (1982) have previously
stated that an inverse relationship between the
exposure concentration and the duration of the
treatment occurs. Petersen et al. (2000) showed
that exposure of zebrafish to 100 ng EE2yl during
the embryo stage caused a significant feminisation.
To refine the critical period it will, therefore, be
recommendable to repeat the present experiments
with a dose-dependent design.

374

L. Andersen et al. / Comparative Biochemistry and Physiology Part C 134 (2003) 365374

Acknowledgments
This project was financed by the Nordic Council
of Ministers.
References

Orn,
Andersen, L., Petersen, G.I., Gessbo, A.,
S., Holbech, H.,
Bjerregaard, P., et al., 2001. Zebrafish (Danio rerio) and
roach (Rutilus rutilus)two species suitable for evaluating
effects of endocrine disrupting chemicals? Aquat. Ecosys.
Health Manage. 4, 275282.
Arcand-Hoy, L.D., Benson, W.H, 1998. Fish reproduction: an
ecologically relevant indicator of endocrine disruption. Environ. Toxicol. Chem. 17, 4957.
Billard, R., Breton, B., Richard, M., 1981. On the inhibitory
effect of some steroids on spermatogenesis in adult rainbow
trout (Salmo gairdneri). Can. J. Zool. 59, 14791487.
Campbell, P.M., Hutchinson, T.H., 1998. Wildlife and endocrine disrupters: requirements for hazard identification.
Environ. Toxicol. Chem. 17, 127135.
Donaldson, E.M., Hunter, G.A., 1982. Sex control in fish with
particular reference to salmonids. Can. J. Fish. Aquat. Sci.
39, 99110.
EDF2. Draft report from the 2nd OECD Expert Consultation
on Endocrine Disrupters Testing in Fish. Tokyo, 15th16th
March, 2000.
Feist, G., Schreck, C.B., Fitzpatrick, M.S., Redding, J.M.,
1990. Sex steroid profiles of coho salmon (Oncorhynchus
kisutch) during early development and sexual differentiation.
Gen. Comp. Endocrinol. 80, 299313.
Gillesby, B.E., Zacharewski, T.R., 1998. Exoestrogens: mechanisms of action and strategies for identification and assessment. Environ. Toxicol. Chem. 17, 314.
Hackmann, E., Reinboth, R., 1974. Delimitation of the critical
stage of hormone-influenced sex differentiation in Hemihaplochromis multicolor (Hilgendorf) (Cichlidae). Gen. Comp.
Endocrinol. 22, 4253.
Holbech, H., Bjerregaard, P., Andersen, L., Petersen, G.I.,
Korsgaard, B., Pedersen, K.L., 2001. Development of an
ELISA for vitellogenin in whole body homogenate of
zebrafish (Danio rerio). Comp. Biochem. Physiol. C 130,
119131.
Jobling, S., Sheahan, D., Osborne, J.A., Matthiessen, P.,
Sumpter, J.P., 1996. Inhibition of testicular growth in rainbow trout (Oncorhynchus mykiss) exposed to estrogenic

alkylphenolic chemicals. Environ. Toxicol. Chem. 15,

194202.
Johnstone, R., Simpson, T.H., Youngson, A.F., 1978. Sex
reversal in salmonid culture. Aquaculture 13, 115134.
Kobayashi, M., Aida, K., Stacey, N.E., 1991. Induction of
testis development by implantation of 11-ketotestosterone
in female goldfish. Zool. Sci. 8, 389393.
Nimrod, A.C., Benson, W.H., 1998. Reproduction and development of Japanese medaka following an early life stage
exposure to xenoestrogens. Aquat. Toxicol. 44, 141156.

Steinholz, A.,
Norrgren, L., 2000. ZebraOrn,
S., Gessbo, A.,
fish (Danio rerio)a candidate to evaluate endocrine dis
rupting chemicals. In: Andersen, L., Bengtsson, B.-E., Bjork,
Holbech, H., Hylland, K.et al. (Eds.), Zebra
M., Gessbo, A.,
Fish for Testing Endocrine Disrupting Chemicals. TemaNord
555, Copenhagen, pp. 4762.
Pandian, T.J., Sheela, S.G., 1995. Hormonal induction of sex
reversal in fish. Aquaculture 138, 122.
Petersen, G.I., Andersen, L., Holbech, H., Pedersen, K.L.,
2000. Baseline test studies and development of ELISA. In:

Andersen, L., Bengtsson, B.-E., Bjork,

M., Gessbo, A.,
Holbech, H., Hylland, K.et al. (Eds.), Zebra Fish for Testing
Endocrine Disrupting Chemicals. TemaNord 555, Copenhagen, pp. 3546.
Piferrer, F., Donaldson, E.M., 1992. The comparative effectiveness of the natural and a synthetic estrogen for direct
feminization of chinook salmon (Oncorhynchus tshawytscha). Aquaculture 106, 183193.

Stahlschmidt-Allner, P., Allner, B., Rombke,

J., Knacker, T.,
1997. Endocrine disrupters in the aquatic environment.
Environ. Sci Pollut. Res. 4, 155162.
Syracuse
Research
Corporation,
1999.
http:yy
esc_plaza.syrres.comyinterkowylogkow.htm.
Takahashi, H., 1977. Juvenile hermaphroditism in the zebrafish, Brachydanio rerio. Bull. Fac. Fish. Hokkaido Univ. 28,
5765.
Uchida, D., Yamashita, M., Kitano, T., Iguchi, T., 2002. Oocyte
apoptosis during the transition from ovary-like tissue to
testis during sex differentiation of juvenile zebrafish. J. Exp.
Zool. 205, 711718.
Yamamoto, T., 1969. Sex differentiation. In: Hoar, W., Randall,
D. (Eds.), Fish Physiology. Academic Press, New York, pp.
117158.
Yamazaki, F., 1976. Application of hormones in fish culture.
J. Fish. Res. Board Can. 33, 948958.