Chemosphere 74 (2008) 125–130

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Uptake and depuration of the anti-depressant fluoxetine by the Japanese

medaka (Oryzias latipes)
Gordon Paterson *, Chris D. Metcalfe
Worsfold Water Quality Centre, Trent University, 1600 West Bank Drive, Peterborough, Ontario, Canada K9J 7B8

a r t i c l e i n f o a b s t r a c t

Article history: The selective serotonin reuptake inhibitor (SSRI) class of anti-depressants is among the most widely pre-
Received 11 April 2008 scribed groups of pharmaceuticals. Consequently, aquatic ecosystems impacted by municipal wastewater
Received in revised form 7 August 2008 discharges are predicted to receive substantial annual loadings of these compounds. Although SSRIs have
Accepted 14 August 2008
been detected in fish tissues, little is known of their uptake and depuration in freshwater fish species. In
Available online 8 October 2008
this study, Japanese medaka (Oryzias latipes) were exposed to fluoxetine at a nominal concentration of
0.64 lg L1 for 7 d and subsequently allowed to depurate in clean water over a 21 d period. Fluoxetine
Keywords:
uptake by medaka was observed within the first 5 h of exposure and the biologically active metabolite,
Fluoxetine
Norfluoxetine
norfluoxetine, was also detected in medaka tissues during this timeframe. A maximum fluoxetine con-
Uptake centration was measured in medaka by the third day of the uptake phase, yielding an uptake rate con-
Depuration stant (k1) of 5.9 ± 0.5 (d1). During the depuration phase of the experiment, a half life of 9.4 ± 1.1 d
Japanese medaka was determined for fluoxetine. Using these data, bioconcentration factor (BCF) values of 74 and 80 were
estimated for fluoxetine and a pseudo-BCF (the ratio of the concentration of norfluoxetine in medaka and
the aqueous fluoxetine concentration) of 117 was calculated for norfluoxetine. These results indicate
longer persistence and greater potential for the bioaccumulation of fluoxetine and norfluoxetine in fish
tissues than would be predicted from prior half life estimates derived using mammalian species.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction be surprising that there is considerable variation in the reporting

Anti-depressant pharmaceuticals and their metabolites from
the class of selective serotonin reuptake inhibitors (SSRIs) have
been detected in water and municipal wastewater in North Amer-
ica and in northern Europe (Weston et al., 2001; Kolpin et al., 2002;
Metcalfe et al., 2003; Zorita et al., 2007; Schultz and Furlong,
2008). Fluoxetine, which is the active ingredient in ProzacTM, was
one of the first SSRIs developed for clinical use as an anti-depres-
sant, and it remains one of the most widely prescribed pharmaceu-
ticals in North America (Preskorn, 1996; NDC Health, 2008).
Fluoxetine persists in water for a relatively long time relative to
other pharmaceuticals (Johnson et al., 2005; Kwon and Armbrust,
2006). Octanol water partition coefficients (log Kow) for fluoxetine
have been variously reported as 1.57 at pH 7 by Brooks et al.
(2003), 1.8 by Huggett et al. (2004), 4.05 by Christensen et al.
(2007), and 4.51 by Lienert et al. (2007), indicating at least moder-
ate potential for bioaccumulation. Attempts to extend the concept
of Kow from hydrophobic organic compounds to more polar or
charged molecules have included determination of octanol water
distribution ratio (Dow) at a given pH and with a given ionic
strength (Scherrer, 1984). Given these constraints it should not
* Corresponding author. Tel.: +1 705 748 1011x7567; fax: +1 705 748 1569.
E-mail address: gordonpaterson@trentu.ca (G. Paterson).
log Kow values for pharmaceuticals. A recent study demonstrated
that the liposome/water distribution ratio (Dlip–wat) may be a more
useful descriptor than Dow for predicting pharmaceutical bioaccu-
mulation and toxicity as there is very little ionic strength depen-
dence for the partitioning of compounds across liposome
membranes (Nakamura et al., 2008).
In mammals, fluoxetine is biotransformed to norfluoxetine
(Fig. 1) through the activity of cytochrome P450 enzymes (Hiemke
and Härtter, 2000). Pharmacokinetics data for SSRIs in mammalian
models and in humans indicate that fluoxetine has a relatively long
half life of 1–4 d, in comparison to values of 61 d for other SSRIs
(Hiemke and Härtter, 2000). However, little is currently known
regarding the pharmacokinetics of fluoxetine and other SSRIs in
fish. SSRIs are known to induce biological effects in fish, including
delays in reproductive and physiological development (Foran et al.,
2004), decreased aggressiveness (Perrault et al., 2003), and inhibi-
tion of feeding responses (Stanley et al., 2007).
Annual rates of discharge of pharmaceuticals from municipal
wastewater treatment plants (WWTPs) into receiving waters may
reach kilogram levels (Daughton and Ternes, 1999; Lindberg et al.,
2005). Under conditions of chronic exposure, there is potential for
pharmaceuticals to bioaccumulate in fish and other aquatic organ-
isms. Mimeault et al. (2005) observed bioaccumulation of the
cholesterol reducing drug, gemfibrozil, in the plasma of goldfish to

0045-6535/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chemosphere.2008.08.022
126 G. Paterson, C.D. Metcalfe / Chemosphere 74 (2008) 125–130
Fig. 1. Structure of fluoxetine and its active metabolite, norfluoxetine.
levels equivalent to a bioconcentration factor (BCF) of 89. Combined
concentrations of fluoxetine and its metabolite, norfluoxetine, were
>10 lg kg1 in the tissues of fish collected downstream of a munici-
pal wastewater treatment plant (Brooks et al., 2005). Fluoxetine
and norfluoxetine were detected at concentrations of 1.02 and
1.08 lg kg1, respectively, in a gizzard shad (Dorosoma cepedianum)
collected from an urbanized embayment of western Lake Ontario,
Canada (Chu and Metcalfe, 2007). Aquatic systems with substantial
inputs of municipal wastewater effluent or whose flow is dominated
represent the worse case exposure scenarios for emerging pollu-
tants such as pharmaceuticals and personal care products (Brooks
et al., 2006). Our ability to predict the extent of bioaccumulation
and thresholds for toxicity in aquatic organisms exposed to pharma-
ceuticals in such habitats is dependent on knowledge of the rates of
uptake and elimination of these compounds. In this study, the
aquarium fish species, the Japanese medaka (Oryzias latipes) was ex-
posed to aqueous concentrations of fluoxetine to determine the
kinetics of uptake and depuration for both the parent compound
and the metabolite, norfluoxetine. This research is among the first
to study the kinetics of both the uptake and elimination of a pharma-
ceutical in an aquatic species.
2. Materials and methods
2.1. Exposure
Japanese medaka were obtained from a broodstock that has
been continuously maintained at Trent University over the past
15 years. A total of 46 adult Japanese medaka averaging
0.36 ± 0.02 g in mass were used in the uptake and depuration
experiments. Throughout the exposures, fish were held in dechlo-
rinated tap water in 2  10 L glass aquaria (23 fish per aquarium)
at 25 ± 2 °C under a 16:8 h light dark cycle. Water quality parame-
ters, including pH, alkalinity, and hardness, were monitored
throughout the study (Table 1). Over the duration of the experi-
ment, fish were fed to satiation with live artemia two times daily.
Animals were maintained and handled throughout the experiment
according to the Canadian Council for Animal Care Guidelines.
A stock solution was prepared by dissolving fluoxetine hydro-
chloride (USP, Rockville, MD, USA) in acetone to a concentration
of 640 lg of fluoxetine per mL of acetone. Volumes of stock solu-
tion (10 lL) were added to dechlorinated tap water in the exposure
aquaria to produce a nominal fluoxetine concentration of
0.64 lg L1 and a solvent concentration of 0.014 lmol L1. This
nominal concentration is approximately an order of magnitude
higher than the ng L1 concentrations of fluoxetine that have been
reported in surface waters in North America (Kolpin et al., 2002;
Metcalfe et al., 2003).
Table 1
Ranges and average values for water quality parameters measured over 28 d
experimental duration
Parameter Units Range Average
pH N/A 7.2–7.5 7.4
Alkalinity mg L1 60–75 68
Hardness mg L1 75–90 83
Fluoxetine concentrationa lg L1 0.67–0.18 0.55
Fluoxetine concentration is the average measured over the 24 h renewal period.
N/A, unitless.
a
Fluoxetine concentration measured in water during the 24 h renewal period.
Exposures were conducted under static conditions for 7 d, with
complete daily renewal of the exposure solution. After the 7 d up-
take phase, the medaka were transferred to aquaria with clean
water and allowed to depurate for 21 d, with daily renewal of
the water. A minimum of 4 and a maximum of 7 fish were collected
on each of days 0.2 (i.e. 5 h), 3 and 7 during the uptake phase, and
on days 7, 14, and 21 during the depuration phase. To minimize
potential bias due to fish size, individuals observed to be much
smaller or larger than the average fish at the time of collection
were returned to their respective tanks. Control fish were main-
tained in a separate aquarium and sampled on days 0 and 7 of
the uptake phase and day 21 of the depuration phase (n = 5 per
sampling period). An equivalent volume (10 lL) of the acetone car-
rier solvent was added to the control tank on a daily basis during
the 7 d uptake phase.
In separate, preliminary experiments to determine the concen-
trations of fluoxetine in aquaria over the 24 h renewal period, a
10 lL aliquot of the stock solution was added to a 10 L aquarium
containing 23 medaka. Samples of water (50 mL) were collected
in triplicate for analysis at 0, 6, and 24 h post-addition (Table 1).
Exposure conditions, stock solutions, and reagents throughout
the experimental procedure were identical to those of the preli-
minary procedures.
2.2. Extraction and analysis
Water samples were extracted for the analysis of fluoxetine and
norfluoxetine as described by Metcalfe et al. (2003). Briefly, sam-
ples were extracted by solid phase extraction (SPE) with Oasis
HLB cartridges (Waters Inc., Milford, MA, USA), and the analytes
were then eluted from the cartridges with methanol. Medaka were
extracted for the analysis of fluoxetine and norfluoxetine using the
methods described previously by Chu and Metcalfe (2007). Briefly,
whole medaka were homogenized and mixed with HydromatrixÒ
medium (Varian Inc., Palo Alto, CA, USA) and then extracted into
methanol by pressurized liquid extraction (PLE) using an ASE 300
accelerated solvent extraction system (Dionex Corp., Sunnyvale,
CA, USA). The PLE extracts were cleaned up using Oasis MCX SPE
cartridges (Waters Inc.), as described by Chu and Metcalfe (2007).
The extracts from water and fish were analyzed by liquid chro-
matography tandem mass spectrometry (LC–MS/MS) using meth-
ods described by Chu and Metcalfe (2007). Liquid chromatography
was performed on an Agilent Series 1100 HPLC system (Agilent
Technologies Canada, Mississauga, ON, Canada) and the analytes
were separated chromatographically on a Genesis C18 column
(150  2.1 mm i.d., 4 lm particle size) purchased from Chromato-
graphic Specialties Inc. (Brockville, ON, Canada). The HPLC system
was coupled to a Q-Trap tandem mass spectrometer (Applied Bio-
systems/MDS Sciex, Concord, ON, Canada) with an atmospheric
pressure chemical ionization (APCI) source operated in positive
ion mode. Detection was by multiple reaction monitoring (MRM)
of two transition ions. Quantification was performed by an isotope
dilution method using fluoxetine-D5 (Sigma–Aldrich, St. Louis,
 G. Paterson, C.D. Metcalfe / Chemosphere 74 (2008) 125–130
MO, USA) as an internal standard for both fluoxetine and norfluoxe-
tine, with a five point calibration curve spanning the range of antic-
ipated analyte concentrations in the samples. A series of external
standards were prepared with different concentrations of the target
analytes and fixed concentrations (50 ng ml1) of the internal
standard.
Lab blanks using the extraction procedures for both water and
fish tissues were prepared for analysis. The limits of quantitation
(LOQs), defined as a 10:1 signal-to-noise ratio were 0.005 and
0.007 lg L1 for the fluoxetine and norfluoxetine, respectively, in
water, and 0.25 and 0.35 lg kg1 wet weight for fluoxetine and
norfluoxetine, respectively, in fish tissues. The limits of detection
(LODs), defined as a 3:1 signal-to-noise ratio, were 0.002 and
0.003 lg L1 for the fluoxetine and norfluoxetine, respectively, in
water, and 0.08 and 0.12 lg kg1 wet weight for fluoxetine and
norfluoxetine, respectively, in fish tissues. Matrix spike recoveries
were determined as outlined by Chu and Metcalfe (2007). Briefly,
water and fish tissue samples were spiked with the fluoxetine-
D5 internal standard prior to extraction. Both the internal standard
and the target analytes are subject to the same matrix effects, thus
potential reduced responses of the target compounds can be
compensated for by analysis of the internal standard (Chu and
Metcalfe, 2007).
2.3. Pharmacokinetics data analysis
1
Average concentrations (lg kg wet weight) measured at each
sampling event (n = 4 or 5) were used to determine the fluoxetine
depuration (k2) rate constant. The general calculations for this rate
constant (d1) and the corresponding fluoxetine half life (t½) fol-
lowed first order kinetics, as outlined in Eqs. (1)–(3) below (Barron
et al., 1998)
ln½C fðtÞ  ¼ aðtÞ þ b; ð1Þ
k2 ¼ jaj; ð2Þ
lnð2Þ
t 1=2 ¼ ; ð3Þ
k2
where [Cf(t)] is the concentration of fluoxetine measured in the fish
at time (t) during the depuration period, with a and b representing
the regression slope coefficient and constants, respectively.
A fluoxetine uptake (k1) rate constant during the exposure per-
iod was estimated by substituting the depuration rate constant cal-
culated during the elimination period into the accumulation model
outlined in Eq. (4) (Barron et al., 1998) and solving for k1
k1
C fðtÞ ¼  C w  ð1  ek2 t Þ; ð4Þ
k2
where Cf(t) is the fluoxetine concentration measured in the fish at
time (t) during the exposure period and Cw is the fluoxetine concen-
tration measured in the water during the exposure period. Two bio-
concentration factor (BCF) estimates were determined for fluoxetine.
The first estimate (BCFa) was calculated using the fluoxetine concen-
trations measured in the fish after 7 d exposure (Cfish) and the aver-
age water concentration (Cwater) (Nakamura et al., 2008):
C fish
BCFa ¼ : ð5Þ
C water
The second fluoxetine bioconcentration factor (BCFb) was calculated
as the ratio of the uptake and depuration rate constants (Mackay
and Fraser, 2000)
k1
BCFb ¼ : ð6Þ
k2
A pseudo-BCF (Nakamura et al., 2008) was also calculated for nor-
fluoxetine after 7 d exposure using the concentration of norfluoxe-
127
tine measured in the fish after 7 d and the average fluoxetine water
concentration as per Eq. (5).
2.4. Statistics
All statistical analyses were completed using SYSTAT version
11.0 for windows with a criterion of significance of P < 0.05 (SY-
STAT, 2004). Potential differences in fish mass between each of
the sampling events were compared using a one way analysis of
variance (ANOVA). The general linear module of SYSTAT and a
Scheffe’s post significance test were used to account for unequal
sample sizes in the ANOVA (Zar, 1984). Linear regression analysis
was used to determine if a significant change in the average mass
of fish occurred over the duration of the experiment. The non-lin-
ear regression module of SYSTAT was used to estimate the fluoxe-
tine uptake rate constant (k1) during the exposure period. The
uptake rate constant was estimated by substituting the depuration
rate constant calculated during the elimination period into the
accumulation model outlined in Eq. (4) and solving the model for
k1. Multiple iterations of the regression were completed to achieve
optimal fit to the data. Regressions were not corrected for outliers
as their inclusion in the analyses did not significantly alter the fit to
the data.
3. Results and discussion
No changes in the water quality parameters were observed over
the duration of the experiment (Table 1). The mean measured con-
centration of fluoxetine during the 24 h renewal period was
0.55 ± 0.14 lg L1 in the 0.64 lg L1 nominal treatment. Fluoxetine
concentrations declined slightly during this daily renewal period
and norfluoxetine was not detected in any of the water samples.
Fluoxetine and norfluoxetine were present in whole Japanese
medaka at concentrations that increased over the uptake phase
and declined over the depuration phase (Fig. 2). Average wet weight
concentrations of fluoxetine and norfluoxetine, and fish masses re-
corded at each sampling time are summarized in Table 2. The aver-
age mass of fish increased from 0.37 ± 0.03 to 0.40 ± 0.03 g over the
duration of the experiment. However, linear regression analysis
demonstrated that this change was not significant (P = 0.11). Addi-
tionally, no significant differences in the average mass of the fish
were observed between any of the sampling dates (F6,27 = 2.07;
P = 0.09). No mortalities were observed over the duration of the
experiment.
Accumulation of fluoxetine by medaka was observed within 5 h
(i.e. 0.2 d) of the start of exposure. A maximum fluoxetine concen-
tration in medaka of 49.4 ± 6.4 lg kg1 wet weight was measured
at day 3 of the uptake phase. Norfluoxetine was detected in meda-
ka by day 0.2 (i.e. 5 h) of the uptake phase, with the concentration
averaging approximately 40% of the fluoxetine parent compound
(Fig. 2). Norfluoxetine concentrations remained below those mea-
sured for fluoxetine until day 7, when the concentration of the
metabolite averaged 64.3 ± 8.7 lg kg1 wet weight, relative to
40.8 ± 5.0 lg kg1 for fluoxetine. The BCFa and pseudo-BCF values
for fluoxetine and norfluoxetine after 7 d exposure were deter-
mined to be 74 and 117, respectively. The concentrations of fluox-
etine and norfluoxetine measured in control fish were below the
limits of detection over the duration of the study (i.e. days 0, 7,
and 28) and no control correction of the data was necessary.
Extraction recovery efficiencies were P85% for all samples
analyzed.
Depuration of both fluoxetine and norfluoxetine was observed
after removal of the fish from the exposure solution (Fig. 2). A dep-
uration rate constant of 0.074 ± 0.009 d1 was determined for flu-
oxetine, with a corresponding half life of 9.4 ± 1.1 d (Fig. 3). Using
128 G. Paterson, C.D. Metcalfe / Chemosphere 74 (2008) 125–130

uptake depuration 5
120
a
100 4

Ln[fluoxetine] (µg kg-1)

Fluoxetine
Norfluoxetine
80 k2 = 0.074 ± 0.009
3
R2 = 0.969
60
2

40
1
Concentration (µg kg-1)

20
0
0 0 5 10 15 20 25
0 5 10 15 20 25 30
Time (# of days depuration)

120 Fig. 3. Depuration profile for fluoxetine in Japanese medaka. Solid line represents
b the least squares regressions with the fluoxetine depuration rate constant (k2 ± 1
standard error) and coefficient of determination (R2). Each data point represents the
100 mean concentration ± 1 standard error of all fish collected at each sampling
interval.

80
approximately three times longer than the half life estimates for
60 this pharmaceutical in mammalian species (Hiemke and Härtter,
2000). This may reflect lower capacity for biotransformation and
elimination of fluoxetine in fish, relative to mammals. Relative to
40 other SSRIs (e.g. fluvoxamine, paroxetine, citalopram, and sertra-
line), fluoxetine and norfluoxetine were observed to have the lon-
20 gest half lives in mammals (Hiemke and Härtter, 2000). Catterson
and Preskorn (1996) reported that fluoxetine has the highest vol-
ume of tissue distribution (i.e. Vd = 14–100 L kg1) among this
0
0 5 10 15 20 25 30 group of anti-depressants, which likely contributes to its long half
life in tissues. Interestingly, trapping of fluoxetine in cellular lyso-
Time (days) somes may play a role in the retention of this compound in tissues
Fig. 2. Accumulation and depuration in Japanese medaka of (a) fluoxetine and
(Daniel and Wojcikowski, 1997). These results indicate that fluox-
norfluoxetine and, (b) the sum total concentration of fluoxetine and norfluoxetine. etine and norfluoxetine are relatively recalcitrant in biological tis-
Each data point represents the mean concentration ± 1 standard error of all fish sues, and therefore, may have a higher potential for chronic
collected at each sampling interval. toxicity, relative to other SSRIs. Recent studies have demonstrated
that fluoxetine is more acutely toxic to aquatic species in compar-
ison to fluvoxamine, paroxetine, and citalopram (Christensen et al.,
Table 2 2007; Henry and Black, 2007).
Sample sizes, average fish mass (g ± 1 standard error), and average concentrations of
By the end of the uptake phase of this study, the concentration
fluoxetine and norfluoxetine (lg kg1 wet weight ± 1 standard error) in whole
Japanese medaka (n = 4–7) sampled on days during the uptake and depuration phases of norfluoxetine in medaka exceeded that of fluoxetine. Nakamura
et al. (2008) observed similarly whereby norfluoxetine concentra-
Experimental n Mass (g) Fluoxetine Norfluoxetine
tions measured in Japanese medaka were higher than fluoxetine
day (lg kg1) (lg kg1)
tissue levels after 30 d exposure to fluoxetine solutions of 30 and
Uptake phase
300 lg L1. Brooks et al. (2005) also measured higher norfluoxetine
0 5 0.37 (0.03) 6.7 (0.8) 2.6 (0.9)
1 5 0.35 (0.02) 29.2 (7.5) 23.2 (6.2) concentrations in comparison to fluoxetine in the tissues of fish
3 4 0.35 (0.02) 49.4 (6.4) 35.1 (14.2) collected from an effluent dominated stream. After 7 d exposure
7 7 0.37 (0.03) 40.8 (5.0) 64.3 (8.7) to a fluoxetine concentration of 0.55 lg L1 in the current study,
Depuration phase norfluoxetine concentrations measured in the medaka were
14 4 0.43 (0.05) 22.9 (6.4) 29.3 (12.0) approximately double those measured for fluoxetine. In contrast,
21 4 0.55 (0.10) 16.4 (1.7) 17.1 (2.0) Nakamura et al. (2008) determined a norfluoxetine to fluoxetine
28 5 0.40 (0.03) 8.2 (1.1) 28.1 (2.2)
ratio of 5.3 after 30 d exposure to a 30 lg L1 fluoxetine solution.
These trends suggest that longer term exposures may allow a
greater proportion of the parent compound to be metabolized,
this depuration rate constant and an average daily fluoxetine water thereby underestimating its bioaccumulation potential. The BCF
concentration of 0.55 lg L1, the fluoxetine uptake rate constant estimates of 74 and 80 determined for fluoxetine in this study were
during the exposure was estimated at 5.9 ± 0.5 d1. The ratio of slightly higher than those estimated for this compound by Nakam-
the uptake and depuration rate constants yielded a fluoxetine BCFb ura et al. (2008) at the same pH and in the same fish species. The
estimate of 80 for this pharmaceutical in Japanese medaka. accumulation kinetics observed during the exposure period in
This study is among the first to investigate both the uptake and the current study emphasizes the importance of measuring the
depuration of fluoxetine, or any other pharmaceutical in a fish spe- accumulation profile of the parent compound and its metabolite(s)
cies. The half life of 9.4 d for fluoxetine in Japanese medaka is during the exposure period. Especially for pharmaceuticals such as
 G. Paterson, C.D. Metcalfe / Chemosphere 74 (2008) 125–130
fluoxetine whose metabolite retains a high degree of activity as an
SSRI (Hiemke and Härtter, 2000).
Fluoxetine is a racemic mixture of two enantiomers with the
S-enantiomer being approximately 1.5 times more potent than
the R-enantiomer as an SSRI (Hiemke and Härtter, 2000). It has
been demonstrated that the fluoxetine S-enantiomer has a higher
potency in standard toxicity tests using aquatic species (Stanley
et al., 2007). For the norfluoxetine metabolite, the S-enantiomer
is approximately 20 times more potent as an SSRI than the R-enan-
tiomer and, under prolonged exposure conditions, the concentra-
tion of racemic norfluoxetine typically exceeds that of racemic
fluoxetine (Hiemke and Härtter, 2000). Further, blood concentra-
tions of the norfluoxetine S-enantiomer are higher than the less
potent R-enantiomer (Hiemke and Härtter, 2000). Norfluoxetine
has been quantified in sewage samples (Zorita et al., 2007) and
effluent from a WWTP has been demonstrated to be elevated in
the fluoxetine S-enantiomer relative to the incoming waste water
(MacLeod et al., 2007). The results of this study and those of Brooks
et al. (2005) and Nakamura et al. (2008) demonstrate that nor-
fluoxetine is the predominant compound measured in fish tissues
following exposure to fluoxetine. Norfluoxetine is estimated to
have a longer half life than fluoxetine (Hiemke and Härtter,
2000; Nakamura et al., 2008) and such toxicokinetics dictate that
this compound will exhibit a greater potential for bioaccumulation
than the parent compound. This is confirmed by the higher pseu-
do-BCF values estimated for norfluoxetine relative to fluoxetine
in this study and by Nakamura et al. (2008). Such results indicate
that a better understanding of the environmental fate and tissue
distribution of chiral pharmaceuticals such as fluoxetine and nor-
fluoxetine is required to fully evaluate the risks posed to aquatic
species exposed to these chemicals.
Fluoxetine is a weak base, and estimates of the octanol–water
partition coefficient (log Kow) for this compound range from 1.25
to 4.30 over a pH range from 2 to 11 (Brooks et al., 2003), indicat-
ing relatively low capacity for bioaccumulation in fish. log Kow is
the most widely used descriptor of hydrophobicity, which has been
used to predict bioaccumulation potential (Nendza and Russom,
1991) and toxicity (Hermens and Verhaar, 1995). However, the
log Kow data reported in the literature may vary by orders magni-
tude for the same pharmaceutical. This suggests that there is lim-
ited utility of log Kow data for estimating the bioaccumulation
potential of more polar and ionic pharmaceuticals. Using the fluox-
etine log Kow estimate of 1.57 at pH 7 calculated by Brooks et al.
(2003) and predictive models for ionic compounds, generate fluox-
etine BCFs much lower than determined in this study and those of
Nakamura et al. (2008) (Mackay and Fraser, 2000; Brooks et al.,
2003). More recently, Nakamura et al. (2008) demonstrated that
the liposome/water distribution coefficient (Dlip–wat) is more ro-
bust to changes in pH than the octanol water distribution coeffi-
cient (Dow). Additionally, the bioaccumulation potential of
fluoxetine is positively correlated with pH over the range typical
of surface waters receiving wastewater inputs of pharmaceuticals
(Nakamura et al., 2008). This indicates that Dlip–wat is likely a better
descriptor for predicting the bioaccumulation and toxicity (Avdeef
et al., 1998; Escher et al., 2000) of ionizable compounds such as
pharmaceuticals, as there is very little ionic strength dependence
for the partitioning of compounds across the liposome membrane.
There are models available for estimating the Dlip–wat from pKa and
Kow values and their use has proven important for understanding
the ecological hazards posed by many pharmaceuticals (Lienert
et al., 2007). Significantly, the results of this study indicate that
the bioaccumulation potential of fluoxetine in fish is markedly
greater than would be predicted from its log Kow value alone.
Aquatic species near discharges from WWTPs will be exposed to
a mixture of pharmaceuticals. Acute toxicity tests with alga and
invertebrate test species have shown that the toxic effects of SSRIs
129
are additive (Christensen et al., 2007; Henry and Black, 2007). The
results reported here indicate that testing of other SSRIs for their
pharmacokinetics and toxicity in fish is warranted; either singly
or in mixtures. In addition, SSRIs have been shown in mammalian
models to inhibit the oxidative metabolism of other drugs
(Hemeryck et al., 2000). This introduces the potential for drug–
drug interactions in fish and other aquatic organisms exposed to
mixtures of pharmaceuticals.
Overall, these data indicate that some pharmaceuticals can
accumulate to part per billion (i.e. lg kg1 wet weight) levels in
fish, and that these compounds can persist in tissues for a period
of several days. Fluoxetine concentrations ranging from 0.05 to
0.1 lg L1 have been detected in surface waters receiving inputs
from municipal wastewater effluent (Metcalfe et al., 2003). Using
the model develop by Huggett et al. (2003) for assessing the risk
posed to fish species due to chronic exposure to human pharma-
ceuticals, these data indicate that fluoxetine does not meet pro-
posed safety factor limits and further research is warranted
regarding the potential aquatic toxicity of this compound (Huggett
et al., 2003). In this study, whole medaka were analyzed and no
data were generated for individual fish tissues. Brooks et al.
(2005) observed higher concentrations of fluoxetine and nor-
fluoxetine in the brain and liver tissue of exposed fish, relative to
muscle tissue. Both fluoxetine and norfluoxetine may have ex-
tended residence times in the liver and brain tissues of fish, leading
to an increased potential for biological effects, especially in aquatic
systems with effluent dominated flow (Brooks et al., 2006). Under
these considerations, further studies focused on the distribution
and persistence of SSRIs in various fish tissues, including the brain,
liver, and also plasma levels would prove useful for predicting toxic
effects in fish (Huggett et al., 2004).
Acknowledgements
This work was supported by grants from the Natural Sciences
and Engineering Research Council (NSERC) of Canada through the
Discovery Grants program, and by the Ontario Ministry of the Envi-
ronment through the Best in Science program. Thanks to Tracy
Metcalfe, Patrick Connelly, and Maria Figueroa for their technical
assistance and to Shaogang Chu for analysis of the extracts.
References
Avdeef, A., Box, K.J., Comer, J.E.A., Hibbert, C., Yam, K.Y., 1998. PH-metric log P 10.
Determination of liposomal membrane–water partition coefficients of ionizable
drugs. Pharm. Res. 15, 209–215.
Barron, G.B., Stehly, G.R., Hayton, W.L., 1998. Pharmacokinetic modeling in aquatic
animals I. Models and concepts. Aquat. Toxicol. 18, 61–86.
Brooks, B.W., Foran, C.M., Richards, S.M., Weston, J., Turner, P.K., Stanley, J.K.,
Solomon, K.R., Slattery, M., La Point, T.W., 2003. Aquatic ecotoxicology of
fluoxetine. Toxicol. Lett. 142, 169–183.
Brooks, B.W., Chambliss, C.K., Stanley, J.K., Ramirez, A., Banks, K.E., Johnson, R.D.,
Lewis, R.J., 2005. Determination of select anti-depressants in fish from an
effluent dominated stream. Environ. Toxicol. Chem. 24, 464–469.
Brooks, B.W., Riley, T.M., Taylor, R.D., 2006. Water quality of effluent dominated
ecosystems: ecotoxicological, hydrological, and management considerations.
Hydrobiologia 556, 365–379.
Catterson, M.L., Preskorn, S.H., 1996. Pharmacokinetics of selective serotonin
reuptake inhibitors: clinical relevance. Pharmacol. Toxicol. 78, 203–208.
Christensen, A.M., Faaborg-Andersen, S., Ingerslev, F., Baun, A., 2007. Mixture and
single-substance toxicity of selective serotonin reuptake inhibitors toward
algae and crustaceans. Environ. Toxicol. Chem. 26, 85–91.
Chu, S., Metcalfe, C.D., 2007. Analysis of paroxetine, fluoxetine and norfluoxetine in
fish tissues using pressurized liquid extraction, mixed mode SPE cleanup and
LC–MS/MS. J. Chromatogr. A 1163, 112–118.
Daniel, W.A., Wojcikowski, J., 1997. Contribution of lysosomal trapping to the total
tissue uptake of psychotropic drugs. Pharmacol. Toxicol. 80, 62–68.
Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in
the environment: agents of subtle change? Environ. Health Persp. 107, 907–
937.
Escher, B.I., Schwarzenbach, R.P., Westall, J.W.C., 2000. Evaluation of liposome–
water partitioning of organic acids and bases: I. Development of sorption
model. Environ. Sci. Technol. 34, 3954–3961.
130 G. Paterson, C.D. Metcalfe / Chemosphere 74 (2008) 125–130
Foran, C.M., Weston, J., Slattery, M., Brooks, B.W., Huggett, D.B., 2004. Reproductive
assessment of Japanese medaka (Oryzias latipes) following a four-week
fluoxetine (SSRI) exposure. Arch. Environ. Contam. Toxicol. 46, 511–517.
Hemeryck, A., Eng, C de V., Belpaire, F.M., 2000. Effect of selective serotonin
reuptake inhibitors on the oxidative metabolism of propafenone: in vitro
studies using human liver microsomes. J. Clin. Psychopharmacol. 20, 428–
434.
Henry, T.B., Black, M.C., 2007. Mixture and single-substance acute toxicity of
selective serotonin reuptake inhibitors in Ceriodaphina dubia. Environ. Toxicol.
Chem. 26, 1751–1755.
Hermens, J.L.M., Verhaar, H.J.M., 1995. QSARs in environmental toxicology and
chemistry. In: Hansch, C., Fujita, T. (Eds.), Classical and Three-Dimensional
QSAR in Agrochemistry, vol. 606. American Chemical Society, Washington, DC,
USA, pp. 130–140.
Hiemke, C., Härtter, S., 2000. Pharmacokinetics of selective serotonin reuptake
inhibitors. Pharmacol. Therap. 85, 11–28.
Huggett, D.B., Cook, J.C., Ericson, J.F., Williams, R.T., 2003. A theoretical model for
utilizing mammalian pharmacology and safety data to prioritize potential
impacts of human pharmaceuticals to fish. Hum. Ecol. Risk Assess. 9, 1789–
1799.
Huggett, D.B., Ericson, J.F., Cook, J.C., Williams, R.T., 2004. Plasma concentrations of
human pharmaceuticals as predictors of pharmacological responses in fish. In:
Kümmereer, K. (Ed.), Pharmaceuticals in the Environment: Sources, Fate, Effects
and Risks. Springer, Berlin, Germany, pp. 373–385.
Johnson, D.J., Sanderson, H., Brain, R.A., Wilson, C.J., Bestari, K.T., Solomon, K.R.,
2005. Exposure assessment and microcosm fate of selective serotonin reuptake
inhibitors. Reg. Toxicol. Pharmacol. 42, 313–323.
Kolpin, D., Meyer, F., Thurman, M.T., Zaugg, S.D., Barber, L.B., Buxton, H.T., 2002.
Pharmaceuticals, hormones, and other organic wastewater contaminants in US
streams, 1999–2000: a national reconnaissance. Environ. Sci. Technol. 36,
1202–1211.
Kwon, J.W., Armbrust, K.L., 2006. Laboratory persistence and fate of fluoxetine in
aquatic environments. Environ. Toxicol. Chem. 25, 2561–2568.
Lienert, J., Güdel, K., Escher, B.I., 2007. Screening method for ecotoxicological hazard
assessment of 42 pharmaceuticals considering human metabolism and
excretory routes. Environ. Sci. Technol. 41, 4471–4478.
Lindberg, R.H., Wennberg, P., Johannson, M.I., Tyslind, M., Andersson, B.A.V., 2005.
Screening of human antibiotic substances and determination of weekly mass
flows in five sewage treatment plants in Sweden. Environ. Sci. Technol. 39,
3421–3429.
Mackay, D., Fraser, A., 2000. Bioaccumulation of persistent organic chemicals:
mechanisms and models. Environ. Pollut. 110, 375–391.
MacLeod, S.L., Sudhir, P., Wong, C.S., 2007. Stereoisomer analysis of wastewater
derived b-blockers, selective serotonin re-uptake inhibitors, and salbutamol by
high-performance liquid chromatography-tandem mass spectrometry. J.
Chromatogr. A 1170, 23–33.
Metcalfe, C.D., Miao, X.S., Koenig, B.G., Struger, J., 2003. Distribution of acidic and
neutral drugs in surface waters near sewage treatment plants in the lower Great
Lakes, Canada. Environ. Toxicol. Chem. 22, 2881–2889.
Mimeault, C., Woodhouse, A.J., Miao, X-S., Metcalfe, C.D., Moon, T.W., Trudeau, V.L.,
2005. The human lipid regulator, gemfibrozil bioconcentrates and reduces
testosterone in the goldfish, Carassius auratus. Aquatic. Toxicol. 73, 44–54.
Nakamura, Y., Yamamoto, H., Sekizawa, J., Kondo, T., Hirai, N., Tatarakazo, N., 2008.
The effects of pH in Japanese medaka (Oryzias latipes): acute toxicity in fish
larvae and bioaccumulation in juvenile fish. Chemosphere 70, 865–873.
NDC Health, 2008. The internet drug index (the top 200 prescriptions of 2006).
<http://www.rxlist.com/top200.htm> (cited March 28 2008).
Nendza, M., Russom, C., 1991. QSAR modeling of the ERL-D fathead minnow acute
toxicity database. Xenobiotica 21, 147–170.
Perrault, H.A.N., Semsar, K., Godwin, J., 2003. Fluoxetine treatment decreases
territorial aggression in a coral reef fish. Physiol. Behav. 79, 719–724.
Preskorn, S.H., 1996. Clinical pharmacology of selective serotonin reuptake
inhibitors. Professional Communications Inc., Caddo, OK, USA.
Scherrer, R.A., 1984. The treatment of ionizable compounds in quantitative
structure–activity studies with special consideration to ion partitioning. In:
Magee, P.S., Kohn, G.K., Menn, J.J. (Eds.), Pesticide Synthesis through Rational
Approaches, vol. 255. ACS Symposium Series, Washington, DC, pp. 225–246.
Schultz, M., Furlong, E.T., 2008. Trace analysis of antidepressant pharmaceuticals
and their select degradates in aquatic matrixes by LC/ESI/MS/MS. Anal. Chem.
80, 1756–1762.
Stanley, J.K., Ramirez, A.J., Chambliss, C.K., Brooks, B.W., 2007. Enantiospecific
sublethal effects of the antidepressant fluoxetine to a model aquatic vertebrate
and invertebrate. Chemosphere 69, 9–16.
SYSTAT, 2004. SYSTAT for Windows version 11.0. SYSTAT Software Inc. Richmond,
CA.
Weston, J., Huggett, D.B., Rimoldi, J., Foran, C.M., Slattery, M., 2001. Determination of
fluoxetine (ProzacTM) and norfluoxetine in the aquatic environment. In: Proc.
22nd Annual Meeting of the Society of Environmental Toxicology and
Chemistry, Baltimore, MD, USA, p. 353.
Zar, J., 1984. Biostatistical Analysis, second ed. Prentice Hall Inc., Engle Wood Cliffs,
NJ.
Zorita, S., Mårtensson, L., Mathiasson, L., 2007. Hollow-fibre supported liquid
membrane extraction for the determination of fluoxetine and norfluoxetine
concentration at ultra trace level in sewage samples. J. Sep. Sci. 30, 2513–2521.