Science of the Total Environment 445–446 (2013) 210–218

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Effects of short-term exposure to environmentally relevant concentrations of different

pharmaceutical mixtures on the immune response of the pond snail Lymnaea stagnalis
M. Gust a, b,⁎, M. Fortier c, J. Garric a, M. Fournier c, F. Gagné b, c
a
IRSTEA, UR MAEP, Laboratoire d'écotoxicologie, 3 bis quai Chauveau, 69009 Lyon, France
b
Emerging Methods Section, Aquatic Contaminants Research Division, Science and Technology, Environment Canada, 105 McGill St., Montreal, Quebec, Canada H2Y 2E7
c
INRS-Institut Armand-Frappier, 531 des Prairies Blvd., Laval, Quebec, Canada H7V 1B7

H I G H L I G H T S

► Pharmaceutical mixtures impaired immune response in Lymnaea stagnalis.

► Immunocompetence was differently affected by the therapeutic class mixtures.
► TLR4, AIF, CAT and GR genes expressions seemed highly modulated by pharmaceuticals.
► Global mixture immunotoxic effects were close to those of antibiotic mixture.
► Effluent immunotoxic responses were similar to those of the global mixture.

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

Article history: Pharmaceuticals are pollutants of potential concern in the aquatic environment where they are commonly
Received 24 September 2012 introduced as complex mixtures via municipal effluents. Many reports underline the effects of pharmaceuticals
Received in revised form 13 December 2012 on immune system of non target species. Four drug mixtures were tested, and regrouped pharmaceuticals by
Accepted 13 December 2012
main therapeutic use: psychiatric (venlafaxine, carbamazepine, diazepam), antibiotic (ciprofloxacine, erythro-
Available online 16 January 2013
mycin, novobiocin, oxytetracycline, sulfamethoxazole, trimethoprim), hypolipemic (atorvastatin, gemfibrozil,
Keywords:
benzafibrate) and antihypertensive (atenolol, furosemide, hydrochlorothiazide, lisinopril). Their effects were
Gastropods then compared with a treated municipal effluent known for its contamination, and its effects on the immune
Immunocompetence response of Lymnaea stagnalis. Adult L. stagnalis were exposed for 3 days to an environmentally relevant concen-
Gene expression tration of the four mixtures individually and as a global mixture. Effects on immunocompetence (hemocyte
Pharmaceuticals viability and count, ROS and thiol levels, phagocytosis) and gene expression were related to the immune
Effluent response and oxidative stress: catalase (CAT), superoxide dismutase (SOD), glutathione reductase (GR),
Selenium-dependent glutathione peroxidase (SeGPx), two isoforms of the nitric oxide synthetase gene
(NOS1 and NOS2), molluscan defensive molecule (MDM), Toll-like receptor 4 (TLR4), allograft inflammato-
ry factor-1 (AIF) and heat-shock protein 70 (HSP70). Immunocompetence was differently affected by the
therapeutic class mixtures compared to the global mixture, which increased hemocyte count, ROS levels
and phagocytosis, and decreased intracellular thiol levels. TLR4 gene expression was the most strongly
increased, especially by psychiatric mixture (19-fold), while AIF-1, GR and CAT genes were downregulated.
A decision tree analysis revealed that the immunotoxic responses caused by the municipal effluent were
comparable to those obtained with the global pharmaceutical mixture, and the latter shared similarity
with the antibiotic mixture. This suggests that pharmaceutical mixtures in municipal effluents represent
a risk for gastropods at the immunocompetence levels and the antibiotic group could represent a model
therapeutic class for municipal effluent toxicity studies in L. stagnalis.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction organisms are more frequently exposed to a variety of compounds in the

Current risk assessment methods focus on the intrinsic toxic proper-
ties of single chemicals, which are less environmentally realistic. Indeed,
⁎ Corresponding author at: IRSTEA, UR MAEP, Laboratoire d'écotoxicologie, 3 bis quai
Chauveau CP 220 69336 Lyon cedex 09, France. Tel.: +33 472208905; fax: +33 478477875.
E-mail address: marion.gust@irstea.fr (M. Gust).
environment (Fent et al., 2006). Pharmaceuticals can exert adverse
effects on non-target organisms in many ways at very low concentra-
tions, and studies seem necessary to better identify their potential effects
(Contardo-Jara et al., 2011). In this respect, municipal effluents are the
most important source of many domestic and some industrial pollutants,
such as pharmaceuticals, in the environment (Kolpin et al., 2002). These
mixtures are made up of diverse drug classes, with diverse mode of

0048-9697/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.scitotenv.2012.12.057
 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218
action and unknown subtherapeutic side effects, especially to non-target
organisms. In this study, we focused on four main therapeutic
groups, representing the bulk of drug prescriptions in the province
of Quebec (Canada) and the United States: antibiotics, hypolipemics,
anti-psychotics and anti-hypertensives (http://www.conseiller.ca/
avantages/avantages-sociaux/depenses-en-medicaments-le-quebec-
devance-toujours-le-canada-10916 and http://www.rxlist.com/script/
main/art.asp?articlekey=124463). These therapeutic classes address
the most frequent health issues for the treatment and prevention
of various disorders. Our choice was also influenced by previous pri-
oritization lists (Besse et al., 2008) and concentrations found in the
environment (Supplementary data, Table 1). Accordingly, mixtures
of environmentally relevant concentrations of antihypertensive drugs
(atenolol, furosemide, hydrochlorothiazide and lisonipril), hyprolipemic
drugs (atorvastatin, gemfibrozil and benzafibrate), psychiatric
drugs (venlafaxine, carbamazepine and diazepam) and antibiotics
(ciprofloxacine, erythromycine, novobiocin, oxytetracyclin, sulfameth-
oxazole and trimethoprim) were designed for toxicity investigations.
Many reports underline the effects of pharmaceuticals on the
vertebrate immune system: antibiotics (Labro, 2000), peroxisome
proliferator-activated receptor (PPAR) agonists (Gocke et al., 2009),
statins (Djaldetti et al., 2006; Mausner-Fainberg et al., 2008), antiepileptic
drugs (Beghi and Shorvon, 2011), antidepressant drugs (Basterzi et
al., 2010), angiotensin-converting enzyme (ACE) inhibitor (Kratnov
et al., 2005), etc. In recent studies, the effects of some pharmaceuticals
on the bivalve immune system were observed (Canesi et al., 2007;
Contardo-Jara et al., 2011; Gust et al., 2012) and suggested that the im-
mune system is a major target for drugs in molluscs.
Lymnaea stagnalis is an ecologically relevant freshwater gastropod,
which has been used as a model organism to investigate the effects of
environmental pollutants on immunological defense mechanisms
(Russo and Lagadic, 2000, 2004). One feature of this model is that he-
molymph collection is easy and does not require animal sacrifice.
Moreover, mRNA expression of genes involved in immune response
can be used to assess toxic effects and seem a very promising marker
(Gust et al., 2013). Freshwater snails depend mainly on cell-mediated
cytotoxicity for the elimination of foreign materials, and hemocytes
are the best-characterized circulating cell type involved in mediating
immune functions (van der Knaap et al., 1993). Among the various
functions of hemocytes, phagocytosis is an important parameter of
non-specific immunity (Galloway and Depledge, 2001). Phagocytosis
involves the production of reactive oxygen species (ROS) known as
the respiratory burst. This burst leads to the destruction of the foreign
particles and is considered to be a major defense mechanism in fresh-
water snails (see Russo and Lagadic, 2004). Functions of mollusc hemo-
cytes can be affected by changes in various environmental factors.
Different types of stressors – including pathogens (such as bacteria
and viruses), parasites and xenobiotics – have been identified as hemo-
cyte effectors and the immune response encompasses all stressors
(Galloway and Depledge, 2001; Russo and Lagadic, 2000, 2004;
Russo and Madec, 2007). In aquatic gastropods, the identification of
immunological-based biomarkers might be of great value for moni-
toring water quality, given that individual measurements without
organism sacrifice can be performed.
The rapid elimination of ROS is essential for organism survival,
linking oxidative stress and immune response. This is performed by
the antioxidant system, including superoxide dismutase (SOD), cat-
alase (CAT), selenium-dependent glutathione peroxidase (Se-GPx)
and glutathione reductase (GR). Gluthatione (GSH) is an important
scavenger of free radicals and plays a role in the maintenance of
the redox status of protein sulfhydryl groups. Nitric oxide (NO) is
produced by NO-synthases in hemocytes during phagocytosis and
reacts with hydrogen peroxide to form peroxynitrite, a highly potent
bactericide (Gourdon et al., 2001). Moreover, as it does in verte-
brates, NO may act as an immunomodulator and mediates the effects
of estrogens and opioids on immunity and inflammation (Galloway
211
and Depledge, 2001; Stefano et al., 2000). The allograft inflammatory
factor-1 (AIF-1) is an interferon-like (IFN-γ) that has been cloned in
gastropods (De Zoysa et al., 2010). It was demonstrated that AIF-1 is
a modulator of the immune response during macrophage activation
and plays a significant role, not only in immune defense reactions
but also in different host responses to inflammatory stimuli. The
Toll-like receptor (TLR) family is the best-characterized class of pat-
tern recognition receptors (PRR) and is involved in the detection of
multiple pathogen-associated molecular patterns (PAMPs), includ-
ing lipopolysaccharides (LPS; detected by TLR4) and other biogenic
molecules such as peptides and DNA. Recognition of these molecular
structures is coupled with several effector responses, including phago-
cytosis (Iwasaki and Medzhitov, 2004). The molluscan defense mole-
cule (MDM) presents similarity with hemolin, an Ig family member of
the insect internal defense system (Hoek et al., 1996). Heat-shock pro-
teins (HSPs) are a family of stress proteins that are induced by a variety
of stressors (Das et al., 1995). HSPs are known to be involved in cell
protection and survival (Das et al., 1995).
The purpose of this study was to determine the effects of four en-
vironmentally relevant pharmaceutical mixtures on immunological
parameters in L. stagnalis, using flow-cytometry and high-throughput
gene expression assays (qPCR) encompassing the whole innate
immune response (TLR4, CAT, SOD, GR, SeGPx, NOS and NOSbis,
HSP70, MDM and AIF). These effects were then compared to the global
mixture of all pharmaceuticals, and discriminative endpoints were
looked for. Finally the immune effects of pharmaceuticals mixtures
were compared to a previous exposure performed with a municipal
effluent known for its contamination with various drugs (Gust et al.,
2013), to determine the similarities, and to assess if immunological
endpoints could be used as monitoring tools.
2. Material and method
2.1. Pharmaceuticals mixtures
All pharmaceuticals were purchased from Sigma-Aldrich (Sigma,
Oakville, ON, Canada), except venlafaxine, lisinopril and atorvastat-
in, which were purchased from Toronto Research Chemicals (TRC,
Toronto, ON, Canada). The relative drug concentrations were pre-
pared according to the reported concentrations found in municipal
effluents (Supplementary data, Table 1). An intermediate “median”
concentration was chosen as follows: (i) psychiatric mixture:
venlafaxine (200 ng/L), carbamazepine (200 ng/L) and diazepam
(10 ng/L); (ii) antihypertensive mixture: atenolol (500 ng/L), furo-
semide (300 ng/L), hydrochlorothiazide (300 ng/L) and lisinopril
(50 ng/L); (iii) hypolipemic mixture: atorvastatin (50 ng/L), gem-
fibrozil (100 ng/L) and benzafibrate (100 ng/L); (iv) antibiotic
mixture: ciprofloxacine (100 ng/L), erythromycine (50 ng/L), no-
vobiocin (100 ng/L), oxytetracycline (200 ng/L), sulfamethoxazole
(50 ng/L) and trimethoprim (50 ng/L); (v) global mixture: the
whole pharmaceutical mixture, i.e. exposed to all the classes combined
(Supplementary data, Table 1). All pharmaceuticals were solubilized in
DMSO (dimethylsulfoxyde), and the final concentration of the solvent
was 0.01%. A solvent control was included. The nominal concentrations
were obtained by serial dilutions. No measurements were made of the
chemicals content in the beakers. However, a previous laboratory
study found that the measured concentrations of all drug solutions
used during a 96 h exposure were within 90% of the nominal concentra-
tion (Pascoe et al., 2003). This combined with their relative persistence
in the environment indicates that these compounds do not readily
degrade or adsorb to the beakers during the exposure period.
2.2. Exposures
L. stagnalis (Linné, 1758) (Gastropoda, Pulmonata, Basommatophora)
is maintained at the laboratory of Environment Canada (Montreal,
212 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218
Canada). Snails were collected in a clean area (Pointe-aux-Cascades,
Quebec, Canada) and maintained under laboratory conditions for at
least 2 months (14/10 light/dark daily cycles, 20 ± 1 °C) in 30 L
aquaria containing dechlorinated tap water supplemented with
mussel shell as calcium source (calcium carbonate). They were fed
pesticide-free lettuce daily ad libitum, and the aquaria were cleaned
daily by changing of 1/3 volume of the aquaria.
The snails (40 ± 1 mm shell length) were exposed for 3 days to
500 mL of the different mixtures of pharmaceuticals (psychiatric, an-
tihypertensive, antibiotic, hypolipemic), their global mixtures and a
control. Water was entirely renewed daily and beakers cleaned. Snails
were fed pesticide-free lettuce. Exposures were performed in beakers,
with three beakers per conditions, containing three snails per
beakers. Hemolymph was collected on the nine snails and analyzed
in flow-cytometry and RT-qPCR for each experimental condition.
2.3. Hemolymph collection and analysis by flow cytometry
Hemolymph was collected after 3 days of exposure by stimulation
of the foot sole, as described by Sminia (1972). When the snail
retracts into its shell, a drop of hemolymph is extruded through the
hemal pore and is collected using a micropipette. The collected hemo-
lymph was used to measure flow-cytometry and the genomic param-
eters for each snail as described in a previous study (Gust et al., 2013).
Briefly, immediately after the hemolymph collection, hemocyte
counts and viability were evaluated by flow cytometry using a
three-color Guava EasyCyte Plus System cytometer (GuavaTechnologies,
Hayward, CA, USA). The intracellular levels of ROS and reduced thiols
were measured using fluorescent probes. Phagocytosis was also
monitored using a flow cytometric methodology. The results were
expressed as the percentage (%) of hemocytes having engulfed one
and three or more fluorescent beads (respectively, immunocapacity
and immunoefficiency).
2.4. Gene expression
On Day 3, after the hemocyte count, 20,000 hemocytes per snail
were collected, and genomic analysis was performed as previously
described (Gust et al., 2013). Briefly, total RNA was isolated using
the RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany). It was subse-
quently converted to cDNA using the QuantiTect Reverse Transcription
Kit (QIAGEN, Valencia, CA, USA). Primer pairs for elongation factor (EF)
(housekeeping gene), β-tubuline (housekeeping gene), TLR4, AIF-1,
CAT, SOD, SeGPx, GR, NOS1, NOS2 and HSP70 were designed using the
program Primer3 (http://frodo.wi.mit.edu/) over gene sequences avail-
able in GenBank (Supplementary data, Table 2).
Real-time PCR was performed using the iCycler iQ Real-Time De-
tection System (Bio-Rad, Hercules, CA, USA) and the iQ SYBR Green
Supermix (Bio-Rad). The reaction was carried out using the following
thermal profile: 15 s at 95 °C (denaturation) and 1 min at 60 °C
(annealing/extension) for 40 cycles. The expression of two house-
keeping genes (EF and β-Tub) was used for gene normalization.
The stability of mRNA expression of the reference genes was assessed
using GeNorm and Normfinder. mRNA expression of β-tubulin gene
remained unchanged, suggesting that only rearrangements of proteins
occurred during phagocytosis. The quantification cycle (Cq) values
were used to calculate relative expression. Cq values were corrected
when the efficiency was not 100%. The control group corresponded to
the solvent control and served to normalize mRNA expression data of
genes where 1× expression corresponds to that of the control group.
2.5. Comparison with municipal effluent
The observed responses were compared to those obtained under
similar exposures conditions and protocols in a previously published
study (see Gust et al., 2013). Briefly snails were exposed to water
collected upstream (control site) and downstream of the effluent
discharge of Montreal. All measurements and used protocols were
the same as those described in this study.
2.6. Statistical analysis
Normality distribution and homogeneity of variance of the data
were tested. One-way analysis of variance (ANOVA) was performed,
followed by the least-square difference (LSD) test for post hoc com-
parisons of groups, with Bonferroni adjustments. When necessary, a
Kruskall–Wallis test and a Mann–Whitney U-test were performed.
Correlations were assessed using the Pearson product–moment pro-
cedure test. All statistical analyses were performed using Statistica®
7.1 for Windows (StatSoft Inc., Tulsa, OK, USA). Because the snails
were exposed to only one but environmentally realistic concentration
of each mixture, a decision tree analysis was used instead of discrim-
inant function analysis. The CART (classification and regression trees)
algorithm was used to maximize the number of univariate divisions
in the data set to develop a classification tree. The CART procedure
consists in finding a rule (usually in the form of if/then arguments)
that best predicts the observed changes in the biomarkers from the
predictor variable (therapeutic class). Canonical analysis was also
performed to find sub-groups of genes that were related to the observed
immunological changes (phagocytosis, total thiols, ROS and viability).
Significance was set at pb 0.05, and when results are given, the word
significant refers to statistical significance.
3. Results
3.1. Immunocompetence
First, no mortality was observed in any treatment group. No signifi-
cant effect was observed on hemocyte viability after 3 days of exposure
to the various therapeutic classes. Compared to the global mixture
(i.e. the mixture made up from the four therapeutic classes), there
was also no change in cell viability (Fig. 1A). Hemocyte count was
significantly increased (3.1-fold) in the snails exposed to the global
mixture, compared to control snails and the different therapeutic
classes (Fig. 1B). ROS levels were significantly higher in hemocytes of
snails exposed to the hypolipemic (2.9-fold) and global (3.3-fold) mix-
tures, compared to control snails (Fig. 2). Compared to the global mixture,
the psychiatric, antihypertensive and antibiotic mixtures displayed
significantly lower ROS levels. The intracellular levels of thiols were
significantly decreased in snails exposed to the hypolipemic drugs, antibi-
otics and global mixtures, compared to control snails (Fig. 3). Compared
to the global mixture, only the antihypertensive mixture exhibited a
significant decrease.
Phagocytic activity was also impaired by the different drug mixtures.
Immunocapacity (i.e. the proportion of hemocytes that engulfed at least
one bead) was significantly decreased by 25% and 28% in hemocytes of
snails exposed, respectively, to the antihypertensive drugs and the antibi-
otic mixtures compared to the control. However, both immunocapacity
and immunoefficiency (i.e. the proportion of hemocytes that engulfed
at least three beads) were significantly increased in hemocytes exposed
to the global mixture compared to the control (Fig. 4). Compared to the
global mixture, immunocapacity was significantly lower for the psychiat-
ric, antihypertensive and antibiotic mixture. For immunoefficiency, all
therapeutic classes were significantly lower than the global mixture
(i.e. the combination of the four therapeutic classes).
3.2. Gene expression
Exposure to pharmaceutical mixtures impaired immune gene expres-
sion in L. stagnalis (Fig. 5A and B). HSP70, SOD, NOS1, NOS2, SeGPx and
TLR4 mRNA expression were globally increased, while CAT, GR and
AIF-1 mRNA expression were decreased. MDM mRNA expression was
 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218 213

95 A

Cell viability (% viable cells)

90

85

80
Control Psychiatric Hypolipemic Antihypertensive Antibiotic Mixture

900000 B
b
800000
a
Hemocyte density (cells/mL)

700000

600000

500000 a
a
400000
a
300000

200000 a

100000

0
Control Psychiatric Hypolipemic Antihypertensive Antibiotic Mixture

Fig. 1. Hemocyte viability (A) and count (B) in L. stagnalis (mean value and standard deviation) as a function of mixture composition (ab b; Pb 0.05, LSD-tests). N=3 triplicates of 3 snails.

not modulated by the exposure. The psychiatric mixture increased TLR4 (3.3-fold) mRNA expression and reduced GR (0.3-fold) mRNA expres-
(19.1-fold), HSP70 (2.6-fold) and SeGPx (2.1-fold) mRNA expression sion. The global mixture increased TLR4 (2.3-fold) mRNA expression
and decreased AIF (0.4-fold), CAT (0.3-fold) and GR (0.3-fold) mRNA and decreased AIF-1 (0.3-fold) and CAT (0.2-fold) mRNA expression
expression compared to the control. The hypolipemic mixture increased compared to the controls.
NOS1 (2.9-fold) mRNA expression and decreased TLR4 (0.4-fold) mRNA The expression of immune-related genes was discriminately mod-
expression. The antihypertensive mixture increased NOS1 (4.1-fold) ified compared to the global mixture, indicating additive, antagonist
and NOS2 (7.4-fold), SOD (2.2-fold) and TLR4 (9.1-fold) mRNA expres- and synergic responses. Indeed, compared to the global mixture,
sion compared to the controls. The antibiotic mixture increased TLR4 TLR4 expression was significantly higher for the psychiatric and
ROS levels (median green fluorescence

2000
b
1800
induced by DCFDA probe)

1600 b
1400
1200
1000
800 a a
600 a
a
400
200
0
Control Psychiatric Hypolipemic Antihypertensive Antibiotic Mixture

pharmaceuticals mixtures

Fig. 2. ROS production in L. stagnalis hemocytes (mean value and standard deviation) as a function of mixture composition (a b b; P b 0.05). N = 3 triplicates of 3 snails.
214 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218

Thiols levels (green fluorescence induced by CMFDA probe

compared to negative control, expressed in % of control)
140

a
120 a
a
100 b
b
b
80

60

40

20

0
Control Psychiatric Hypolipemic Antihypertensive Antibiotic Mixture

pharmaceuticals mixtures

Fig. 3. Thiol production in L. stagnalis hemocytes (mean value and standard deviation) as a function of mixture composition (a b b; P b 0.05, LSD-tests). N = 3 triplicates of 3 snails.

antihypertensive mixtures, while HSP70 gene expression was signifi- 5.1-fold from the controls, suppression of both AIF-1 (b0.6-fold of the
cantly higher in the psychiatric, antihypertensive and hypolipidemic controls) and low GR gene expressions (b 0.38-fold of the controls). Psy-
drug mixtures, SeGPx gene expression was significantly higher in choactive drugs led to the following changes: high TLR4 expression
the psychiatric mixture, CAT gene expression was significantly higher (above 5.1-fold) and low AIF-1 expression (b0.6-fold of the controls).
in the antihypertensive, antibiotic and hypolipidemic drug classes, Lipid regulators were identified as compounds that downregulate the
NOS1 gene expression was significantly higher in the antihyperten- TLR4 gene (b 0.6-fold of the controls). Antihypertensive drugs are com-
sive and hypolipidemic classes, NOS2 gene expression was signifi- pounds able to increase TLR4 gene expression (above 5.1-fold) and
cantly higher in the antihypertensive mixture, and AIF-1 gene increase AIF-1 gene expression to above 0.6-fold from the controls.
expression was significantly higher in antihypertensive and hypo- Finally, the combined pharmaceutical mixture was similar to the antibi-
lipidemic classes. In contrast, GR gene expression was significantly otic mixture but with GR expression at above 0.38-fold. Based on this
lower in the psychiatric and antibiotic classes. Finally, SOD gene classification algorithm, the municipal effluents response pattern
expression was significantly higher in the antihypertensive and (Gust et al., 2013) was closely related to the global drug mixture
hypolipidemic classes and lower in the psychiatric group, compared group. We repeated the decision tree analysis to associate the combined
to the global mixture. drug mixture responses with a given therapeutic class (not shown). The
analysis (Fig. 6A) that revealed the response pattern of the combined
3.3. Decision tree analysis drug mixture was more closely related to the antibiotics class, i.e. com-
pounds leading to TLR4 (5.1-fold > response > 0.6-fold), low AIF-1
In an attempt to understand the impacts of the various therapeutic (b 0.6-fold of the controls) and low GR expressions (b 0.38-fold of
classes on immunocompetence and gene expression data, a decision the controls). The analysis also allowed us to find the most important
tree analysis was performed to identify the “signature” endpoints that biomarkers (defined as >80% importance) involved in therapeutic
can discriminate between the drug classes (Fig. 6A and B). The analysis drug effects discrimination (Fig. 6B). These were GR, TLR4, AIF-1, and
revealed that each therapeutic mixture had a distinctive response pat- CAT, then NOS2 and HSP70 and finally immunocapacity. The TLR4,
tern. Antibiotic drugs led to the following changes in the response AIF-1, CAT and GR genes displayed an importance of >95%. This sug-
pattern: hemocyte expressing TLR4 expression between 0.6- and gests that these therapeutic classes involve genes for non-self
Phagocytosis (% of population which

Immunocapacity (1 bead or more)

45
Immunoefficiency (3 beads or more)
c
40 bc
engulfed a bead or more)

35 b
b
30
a a
25

20 b
ab
15 ab ab
a a
10

0
Control Psychiatric Hypolipemic Antihypertensive Antibiotic Mixture
pharmaceuticals mixtures

Fig. 4. Immunocapacity, i.e. one bead and more engulfed and immunoefficiency, i.e. three beads and more engulfed in L. stagnalis hemocytes (mean value and standard deviation) as
a function of mixture composition (a b b b c; P b 0.05, LSD-tests). N = 3 triplicates of 3 snails.
 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218 215

Gene expression compared to control (1x)

Psychiatric Hypolipemic *
20 Antihypertensive Antibiotic
Mixture

15

*
10 *

*
* *
5
* * *
*

0
HSP NOS bis NOS SOD TLR4

B
Gene expression compared to control (1x)

4 Psychiatric Hypolipemic
Antihypertensive Antibiotic *
Mixture

1
* * * * *
* *

0
AIF CAT GR MDM SeGPx

Fig. 5. Expression level of immune-related genes HSP70, NOS, NOSbis, SOD and TLR4 (A) and AIF, CAT, GR, MDM, and SeGPx (B) in L. stagnalis after 3 days exposure (mean value and
standard deviation). Control samples were defined as 1× expression level for each analyzed gene in control snails. Asterisks indicate significant differences at the p b 0.05 level
between control and mixtures (Mann–Whitney U-test). N = 3 triplicates of 3 snails.

recognition (TLR4), cytokines (AIF-1) and oxidative stress (CAT, GR). immune effects of single pharmaceuticals of this therapeutic class. A
Canonical analysis revealed a significant correlation of immunocompe- much higher concentration of carbamazepine (14 mg/L) was required
tence (Rc = 0.62; p = 0.03) with the latter group of genes (AIF-1, CAT to induce in vitro phagocytosis in freshwater mussels after 24 h of
and GR genes), while the former group (TLR4, HSP, SeGPx) was not exposure (Gagne et al., 2006). In mussels exposed to 0.1–10 μg/L of
significantly related (Rc = 0.52; p > 0.05). carbamazepine for 7 days, a significant decrease in lysosomal mem-
brane permeability in hemocytes was observed, with increased gluta-
4. Discussion thione S-transferase and catalase activities in the digestive gland and
mantle (Martin-Diaz et al., 2009).
To the best of our knowledge, this is the first report on the immune However, the psychiatric mixture leads to significant changes at
effects of environmentally relevant concentrations of pharmaceutical the gene expression level. TLR4, HSP70 and SeGPx gene expression
mixtures on L. stagnalis. We found studies dealing with pesticides was upregulated, while AIF-1, CAT and GR gene expression was
(Russo and Lagadic, 2004; Russo et al., 2007) and with surface waters downregulated. The gene expression inductions suggest that the
contaminated by municipal effluents (Gust et al., 2013). This study re- psychoactive group leads to glutathione-dependent peroxidase ac-
vealed that all therapeutic drug classes were able to affect the immune tivity (SeGPx) and protection response towards protein denatur-
system of snails in one way or another. As in previous studies (Dietrich ation (HSP70). However, catalase and GR gene expression were
et al., 2010; Gust et al., 2012; Quinn et al., 2009), hemocyte viability was also reduced in hemocytes, which suggest decreased ROS handling
little modified by pharmaceutical exposures. and decreased inflammation. The psychoactive group also leads to
strong TLR4 expression in the hemocytes of snails. This could indi-
4.1. Discriminative effects of the therapeutic mixtures cate either a strong inflammation signal, owing to the presence of
biogenic molecules, or a compensation mechanism against the loss of
4.1.1. Psychiatric mixture Toll-like receptor signaling, perhaps through the blockade of Toll-like
In the present study, the psychiatric mixture – containing venlafaxine receptors. Fluoxetine and citalopram were found to exhibit strong
(200 ng/L), carbamazepine (200 ng/L) and diazepam (10 ng/L) – did not anti-inflammatory effects in vertebrate models (humans and mice)
directly impair immunocompetence parameters in the pond snails, as and inhibited Toll-like receptors (Sacre et al., 2010). Indeed, fluoxetine
defined by the following endpoints: hemocyte density, viability, phagocy- and citalopram treatment inhibited the signaling of TLRs 3, 7, 8, and 9,
tosis, ROS and thiol levels. This is consistent with previous studies on the providing a potential mechanism for their anti-inflammatory action. If
216 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218

Control
Glutathione peroxidases are involved in the elimination of lipid hydro-
A 1
Control
Psychoactive peroxides where the oxidized glutathione is reduced by GR in the pres-
Hypylipodemics
Anti-hypertensives
ence of NADPH. This is consistent with the positive correlation between
Antibiotics GR and SeGPx (r=0.38), suggesting that these drugs have a local mem-
9 45 Mixture
TLR < 0.6
brane effect in hemocytes. Venlafaxine was reported to induce HSP70
Control
2 3 levels in rat glioma cells (Yu et al., 2010), while carbamazepine was
able to induce HSP70 mRNA levels in zebra mussel gills (Contardo-Jara
et al., 2011). This suggests that carbamazepine has the ability to induce
Hypolipidemic
27 18 protein-folding changes, which is consistent with the capacity of the
Control TLR<=5,0617 Psychoactives drug to stabilize the inactive state of sodium channels in membranes.
4 5
Residual analysis of TLR4 gene expression, corrected for HSP70 changes,
revealed a 2.5-fold reduction in TLR4 gene expression for the psychiatric
group (containing carbamazepine) only, thus supporting the hypothesis
18 AIF < 0.6 9 9 AIF < 0.6 9
Antibiotics Control that this drug could affect protein folding or conformation. Interestingly,
6 7 8 9 the psychiatric mixture downregulated AIF-1 mRNA, which is IFN-γ-like.
It has been proposed that AIF-1 plays a significant role, not only in
immune defense reactions but also in different host responses to inflam-
Psychoactives Antihypertensives
9
GSR < 0.38
9 matory stimuli (De Zoysa et al., 2010), as suggested by the correlation
between AIF-1 expression and CAT and SOD. Venlafaxine was also
10 11
shown to downregulate AIF-1 mRNA expression in human lymphocytes,
Mixture
which is in keeping with the anti-inflammatory properties of SSRIs
Antibiotics
(Kalman et al., 2005), and diazepam inhibits human T-cell function
through peripheral benzodiazepine receptors, also decreasing IFN-γ pro-
duction (Beghi and Shorvon, 2011).
100 B
4.1.2. Antibiotic mixture
Relative importance (%)

80 Exposure to the antibiotic mixture decreased phagocytosis activity

60
40
20
0
Viability P1 ROS AIF GSH SeGPX NOS2 SOD
Hem # P2 SH CAT MDM Hsp NOs1 TLR
Predictive effects
Fig. 6. Decision tree analysis of the immunocompetence responses. Both the
immuncompetence and gene expression biomarkers were used in this analysis.
The decision tree (A) reveals how the therapeutic classes are discriminated at the
immune level, and the most important biomarkers (b75% importance) are shown
(B). Hem#: hemocyte count, P1: immunocapacity, P2: immunoefficiency, SH: thiol
levels.
this holds true, then the serotonergic mix of the psychoactive drugs
could also block TLR4 signaling in snails and this could be related to
the anti-inflammatory effects. Other drug interactions with this recep-
tor can also occur, as induction of TLR4 expression has often been de-
scribed in vertebrates, particularly with carbamazepine (Hutchinson
et al., 2010a, 2010b). Based on the endpoints related to oxidative stress,
it appears that cells were in a resting phase rather than inflammation:
there was no change in SOD, NOS1 or 2, thiols or ROS, with low expres-
sion in CAT, AIF-1 and GR. In addition, TLR4 gene expression was nega-
tively correlated with total thiols and catalase gene expression, which is
consistent with reduced inflammation condition in hemocytes. In-
creased TRL4 gene expression could also result from increased exposure
to commensal bacteria and the activation of inflammatory cells, which
suggests that bacteria growth might have occurred in mussels (Chen
et al., 2010). However, we did not observe any signs of such growth in
the exposure vessels and TLR4 expression was not related to phagocyto-
sis and inflammation markers. However, overexpression of the TLR4
gene was also associated with increased HSP70 and SeGPx expression.
and intracellular thiol levels. It also upregulated TLR4 expression and
downregulated GR expression. The observed response pattern of this
therapeutic class closely resembled those of the mixture and the
effluents. The reduced number of bacteria triggered by the large spec-
trum antibiotic mixture could have reduced phagocytosis activity in
and of itself, while maintaining immune alertness with a high expres-
sion of TLR4. The reduced levels of GSH (or thiols) could be the result
of xenobiotic conjugation in snails or oxidative stress that was in-
duced by these antibiotics. In bivalves, in vitro exposure of hemocytes
to the same antibiotic mixture decreased thiol levels but increased
phagocytosis and ROS levels in hemocytes (Gust et al., 2012). Differences
between the immune effects of drugs on molluscs, regardless of whether
the exposure was in vitro or in vivo, have already been described. Indeed,
in vivo exposures may induce changes in hemolymph concentration of
different endogenous factors (estrogens, neuro-immune modulators,
cytokine-like proteins, etc.) that have been shown to affect the immune
response (Canesi et al., 2007).
4.1.3. Antihypertensive mixture
The antihypertensive mixture decreased phagocytosis and
upregulated TLR4, NOS-1, NOS-2 and SOD expression, compared to
the control. L. stagnalis possesses two NOS genes: an ancestral NOS
was duplicated where one copy retained its original function, whereas
an internal DNA inversion occurred in the other. The inversion resulted
in the creation of new regulatory elements required for the termination
and activation of transcription. Consequently, the duplicated gene was
split, and two new and independently expressed genes were created
(Korneev and O'Shea, 2002). Previous studies showed differential ex-
pression of both NOS genes (Korneev and O'Shea, 2005). This is also
the case in this study, where both NOS gene expressions are modified
differently by the tested mixtures. For example, the hypolipemic mixture
induced an upregulation of the NOS-1 mRNA, while the antihypertensive
mixture induced an upregulation of both NOS mRNA expressions. NO
production, following stimulation by cytokines, is thought to be impor-
tant in host defense and inflammatory mechanisms (Culic et al., 2001),
suggesting that the antihypertensive mixture has inflammatory proper-
ties. It was found that the release of NO could downregulate immunocyte
activity in mussels (Stefano et al., 2004). Thus, the decrease in phagocytic
 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218
activity observed with the antihypertensive mixture could be caused by
increased NO production.
4.1.4. Hypolipemic mixture
The hypolipemic mixture increased intracellular ROS levels and
decreased thiol levels. As in bivalves, fibrates affect the immune sys-
tem of L. stagnalis. In an in vitro study with Elliptio complanata, Gagne
et al. (2006) measured an induction of phagocytosis by both drugs
benzafibrate and gemfibrozil. They also increased phagocytic activity
in vivo in Mytilus galloprovincialis, but the addition of the two compounds
did not result in significantly different effects (Canesi et al., 2007). The
hypolipemic mixture also upregulated NOS-1 mRNA expression. It is
the only mixture that did not upregulate TLR4 mRNA expression, but
instead downregulated it. The reduction of thiol levels in hemocytes,
associated with elevated ROS levels, and NOS expression suggest that
the oxidative burst could have detrimental effects (i.e. inflammation)
in hemocytes. Benzafibrate (above 360 ng/L) upregulated HSP70,
SOD and CAT gene expression in the gills and digestive gland of
D. polymorpha after 7 days of exposure (Contardo-Jara et al., 2011).
Lower exposure concentrations could explain the lack of induction of
these genes in the present study, hence the importance of taking into
account the actual levels of drugs in municipal effluent.
4.1.5. Effect of the global mixture and municipal effluent
Generally, the more complex the mixture the more additive is the
overall toxicity (Cleuvers, 2003; Quinn et al., 2009). This kind of
non-linear additive effect can also be suspected for other molecules
occurring in municipal effluent. The effects on immunocompetence
were more significant in snails exposed to the global mixture than in
snails exposed to the “individual” therapeutic class mixtures. Moreover,
AIF-1 and GR were the endpoints capable of discriminating the various
therapeutic classes, the global mixture and the municipal effluent.
These markers hold promise as potential biomarkers of pharmaceutical
products in complex mixtures.
Most studies dealing with pharmaceutical mixtures investigated
the cumulative effects of the given specific mixtures but not between
specific therapeutic classes. Our data suggests that the complex mixture
of the four different therapeutic classes led to different responses in
snails, but the separation of the therapeutic classes could be achieved
using three genes: TLR4, AIF-1 and GR activities. TLR4 gene expression
is involved in non-self recognition and inflammation reaction, while
GR expression is involved in the maintenance of reduced glutathione
in cells, and AIF-1 is IFNγ-like. This suggests that pharmaceutical com-
pounds could impede the ability of organisms to defend against foreign
agents. On the one hand, the antibiotics lead to decreased AIF-1, GR and
TLR4 expression, which could constitute a signature effect for this
therapeutic class. On the other, the psychoactive and antihyperten-
sive classes stimulated the expression of AIF and TLR4 expression
in hemocytes. As explained above, serotonergic drugs have strong
anti-inflammatory properties and block Toll-like receptor signaling.
For antihypertensives, the β-blockers would likely act through the
AMPc and protein kinase A signaling pathway but TLR 4 appears to
be more dependent on the protein kinase C pathway (Grkovich et al.,
2009). Hence, high TLR4 expression is not related to the mode of action
of some antihypertensive drugs. It was shown that LPS stimulation of
TLR4 decreased the absorption of carbonates (HCO3−) in kidney epithelial
cells, which is consistent with the mode of action of diuretics which
limits the reabsorption of sodium and favors the elimination of salts
and water (Good et al., 2009). However, whether the loss of carbonates
is from hemocytes and whether this can influence TLR4 gene expression
remains to be confirmed in snail hemocytes. AIF-1 gene expression
displayed non-linear additivity in its response to the four mixtures ex-
amined (compared to the global mixture) and was downregulated by
the antibiotic and psychiatric drug groups. Hence, this gene product
could represent a highly sensitive marker to detect for pharmaceutical
217
products, especially from the antibiotic and psychiatric classes, in com-
plex mixtures such as municipal effluents.
In vivo exposure of L. stagnalis to pharmaceuticals had effects on
its immune system that were similar to exposure to urban effluents
(Gust et al., 2013). The contamination in pharmaceutical products of
the examined municipal effluent has been extensively described in
previous studies (Gagne et al., 2006; Lajeunesse et al., 2008; Segura
et al., 2007). However no measurements were available in the tested
effluent, limiting the comparison. Both exposures led to increased cellu-
lar count, ROS levels and phagocytosis, and decreased thiol levels. How-
ever, no effect was observed on hemocyte viability with pharmaceutical
exposure, while urban effluent drastically decreased it. Thus, even if the
bacterial load of urban effluent greatly explained the observed effects
on immunocompetence, the pharmaceuticals contained in the effluent
were also certainly at play.
4.1.6. Decision analysis tree
Based on the decision tree analysis, the municipal effluent had a re-
sponse pattern that was close to the global mixture and, in turn, the global
mixture was closely associated with the antibiotic mixture. Antibiotics
represent a major group of contaminants in municipal effluent, with the
combined concentration sometimes reaching well above 10 μg/L. In a pre-
vious study, antibiotics were found to closely resemble the immunotoxic
effects of primary-treated municipal effluent (after the removal of micro-
organisms by filtration) in freshwater mussels (Gust et al., 2012).
However, both the effluent and pharmaceutical mixture upregulated
TLR4 gene expression, the psychiatric pharmaceuticals being the most
potent upregulators. The mechanisms involved might be different. On
the one hand, the presence of bacteria (especially Gram-negative ones)
in the effluent would likely increase TLR4 expression. On the other
hand, the presence of SSRIs was shown to be able to block TLR4 signaling
in cells. More research will be needed to understand the relationships
between TLR4 gene expression, activation and cell signaling.
5. Conclusion
Overall, our results support the hypothesis that gastropods are or-
ganisms that are sensitive to the effects of pharmaceuticals in the envi-
ronment. The effects of urban effluent on the immune response of snails
were found to be closely related to those of the global pharmaceutical
mixture, which suggests that this mixture exhibits similar modes of
action to the overall toxicity of urban effluent. More studies are needed
to better characterize the consequences of decrease immune response
on organisms and populations. Many applications could be derived
from the use of immune response in snails in ecotoxicology. The mea-
sured parameters could be used to track the upstream movement of ex-
posed organisms in a stream, or to help identifying a group of chemicals
potentially triggering the overall effects. This information could also
be used for a target orientated improvement of used and removal of
chemicals from wastewater.
Acknowledgments
Marion Gust held a fellowship from La Région Rhône-Alpes. This
project was funded by the municipal effluent research program of
Environment Canada. The manuscript was copyedited by Patricia Potvin
of Environment Canada.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.scitotenv.2012.12.057.
218 M. Gust et al. / Science of the Total Environment 445–446 (2013) 210–218
References
Basterzi AD, Yazici K, Buturak V, Cimen B, Yazici A, Eskandari G, et al. Effects of
venlafaxine and fluoxetine on lymphocyte subsets in patients with major depressive
disorder: a flow cytometric analysis. Prog Neuropsychopharmacol Biol Psychiatry
2010;34:70–5.
Beghi E, Shorvon S. Antiepileptic drugs and the immune system. Epilepsia 2011;52:
40–4.
Besse JP, Kausch-Barreto C, Garric J. Exposure assessment of pharmaceuticals and their
metabolites in the aquatic environment: application to the French situation and
preliminary prioritization. Hum Ecol Risk Assess 2008;14:665–95.
Canesi L, Lorusso LC, Ciacci C, Betti M, Regoli F, Poiana G, et al. Effects of blood lipid lower-
ing pharmaceuticals (bezafibrate and gemfibrozil) on immune and digestive gland
functions of the bivalve mollusc, Mytilus galloprovincialis. Chemosphere 2007;69:
994-1002.
Chen LW, Chang WJ, Chen PH, Hsu CM. Commensal microflora induce host defense and
decrease bacterial translocation in burn mice through toll-like receptor 4. J Biomed
Sci 2010;17.
Cleuvers M. Aquatic ecotoxicity of pharmaceuticals including the assessment of combi-
nation effects. Toxicol Lett 2003;142:185–94.
Contardo-Jara V, Lorenz C, Pflugmacher S, Nuetzmann G, Kloas W, Wiegand C. Expo-
sure to human pharmaceuticals Carbamazepine, Ibuprofen and Bezafibrate causes
molecular effects in Dreissena polymorpha. Aquat Toxicol 2011;105:428–37.
Culic O, Erakovic V, Parnham MJ. Anti-inflammatory effects of macrolide antibiotics.
Eur J Pharmacol 2001;429:209–29.
Das DK, Maulik N, Moraru II. Gene expression in acute myocardial stress-induction by
hypoxia, ischemia, reperfusion, hyperthermia and oxidative stress. J Mol Cell Cardiol
1995;27:181–93.
De Zoysa M, Nikapitiya C, Kim Y, Oh C, Kang DH, Whang I, et al. Allograft inflammatory
factor-1 in disk abalone (Haliotis discus discus): molecular cloning, transcriptional
regulation against immune challenge and tissue injury. Fish Shellfish Immunol
2010;29:319–26.
Dietrich S, Dammel S, Ploessl F, Bracher F, Laforsch C. Effects of a pharmaceutical mixture
at environmentally relevant concentrations on the amphipod Gammarus fossarum.
Mar Freshw Res 2010;61:196–203.
Djaldetti M, Salman H, Bergman M, Bessler H. Effect of pravastatin, simvastatin and
atorvastatin on the phagocytic activity of mouse peritoneal macrophages. Exp
Mol Pathol 2006;80:160–4.
Fent K, Weston AA, Caminada D. Ecotoxicology of human pharmaceuticals. Aquat
Toxicol 2006;76:122–59.
Gagne F, Blaise C, Fournier M, Hansen PD. Effects of selected pharmaceutical products
on phagocytic activity in Elliptio complanata mussels. Comp Biochem Physiol C
Toxicol Pharmacol 2006;143:179–86.
Galloway TS, Depledge MH. Immunotoxicity in invertebrates: measurement and
ecotoxicological relevance. Ecotoxicology 2001;10:5-23.
Gocke AR, Hussain RZ, Yang Y, Peng H, Weiner J, Ben L-H, et al. Transcriptional modulation
of the immune response by peroxisome proliferator-activated receptor-alpha
agonists in autoimmune disease. J Immunol 2009;182:4479–87.
Good DW, George T, Watts BA. Lipopolysaccharide directly alters renal tubule transport
through distinct TLR4-dependent pathways in basolateral and apical membranes.
Am J Physiol Renal Physiol 2009;297:F866–74.
Gourdon I, Guerin MC, Torreilles J, Roch P. Nitric oxide generation by hemocytes of the
mussel Mytilus galloprovincialis. Nitric Oxide Biol Chem 2001;5:1–6.
Grkovich A, Armando A, Quehenberger O, Dennis EA. TLR-4 mediated group IVA phos-
pholipase A(2) activation is phosphatidic acid phosphohydrolase 1 and protein
kinase C dependent. Biochim Biophys Acta Mol Cell Biol Lipids 2009;1791:975–82.
Gust M, Gélinas M, Fortier M, Fournier M, Gagné F. In vitro immunotoxicity of environ-
mentally representative antibiotics to the freshwater mussel Elliptio complanata.
Environ Pollut 2012;169:50–8.
Gust M, Fortier M, Garric J, Fournier M, Gagné F. Immunotoxicity of surface waters con-
taminated by municipal effluents to the snail Lymnaea stagnalis. Aquat Toxicol
2013;126:392–403.
Hoek RM, Smit AB, Frings H, Vink JM, deJongBrink M, Geraerts WPM. A new
Ig-superfamily member, molluscan defence molecule (MDM) from Lymnaea
stagnalis, is down-regulated during parasitosis. Eur J Immunol 1996;26:939–44.
Hutchinson MR, Loram LC, Zhang Y, Shridhar M, Rezvani N, Berkelhammer D, et al.
Evidence that tricyclic small molecules may possess Toll-Like Receptor and
myeloid-differenciation-protein-2 activity. Neuroscience 2010a;168:551–63.
Hutchinson MR, Zhang Y, Shridhar M, Evans JH, Buchanan MM, Zhao TX, et al. Evidence
that opioids may have toll-like receptor 4 and MD-2 effects. Brain Behav Immun
2010b;24:83–95.
Iwasaki A, Medzhitov R. Toll-like receptor control of the adaptive immune responses.
Nat Immunol 2004;5:987–95.
Kalman J, Palotas A, Juhasz A, Rimanoczy A, Hugyecz M, Kovacs Z, et al. Impact of
venlafaxine on gene expression profiles in lymphocytes of the elderly with major
depression — Evolution of antidepressants and the role of the “neuro-immune”
system. Neurochem Res 2005;30:1429–38.
Kolpin DW, Furlong ET, Meyer MT, Thurman EM, Zaugg SD, Barber LB, et al. Pharma-
ceuticals, hormones, and other organic wastewater contaminants in U.S. streams,
1999–2000: A national reconnaissance. Environ Sci Technol 2002;36:1202–11.
Korneev S, O'Shea M. Evolution of nitric oxide synthase regulatory genes by DNA inver-
sion. Mol Biol Evol 2002;19:1228–33.
Korneev S, O'Shea M. Natural antisense RNAs in the nervous system. Rev Neurosci 2005;16:
213–22.
Kratnov AE, Popov SA, Kratnov AA, Demyankova YO. Lisinopril-induced changes in
antioxidant defense in patients with acute coronary syndrome and concomitant
diabetes mellitus type 2. Ter Arkh 2005;77:14–8.
Labro MT. Interference of antibacterial agents with phagocyte functions: immunomodulation
or “immuno-fairy tales”? Clin Microbiol Rev 2000;13:615–50.
Lajeunesse A, Gagnon C, Sauve S. Determination of basic antidepressants and their
n-desmethyl metabolites in raw sewage and wastewater using solid-phase extraction
and liquid chromatography–tandem mass spectrometry. Anal Chem 2008;80(14):
5325–33.
Martin-Diaz L, Franzellitti S, Buratti S, Valbonesi P, Capuzzo A, Fabbri E. Effects of envi-
ronmental concentrations of the antiepilectic drug carbamazepine on biomarkers
and cAMP-mediated cell signaling in the mussel Mytilus galloprovincialis. Aquat Toxicol
2009;94:177–85.
Mausner-Fainberg K, Luboshits G, Mor A, Maysel-Auslender S, Rubinstein A, Keren G,
et al. The effect of HMG-CoA reductase inhibitors on naturally occurring CD4(+)
CD25(+) T cells. Atherosclerosis 2008;197:829–39.
Pascoe D, Karntanut W, Muller CT. Do pharmaceuticals affect freshwater invertebrates?
A study with the cnidarian Hydravulgaris. Chemosphere 2003;51:521–8.
Quinn B, Gagne F, Blaise C. Evaluation of the acute, chronic and teratogenic effects of a
mixture of eleven pharmaceuticals on the cnidarian, Hydra attenuata. Sci Total Environ
2009;407:1072–9.
Russo J, Lagadic L. Effects of parasitism and pesticide exposure on characteristics and func-
tions of hemocyte populations in the freshwater snail Lymnaea palustris (Gastropoda,
Pulmonata). Cell Biol Toxicol 2000;16:15–30.
Russo J, Lagadic L. Effects of environmental concentrations of atrazine on hemocyte
density and phagocytic activity in the pond snail Lymnaea stagnalis (Gastropoda,
Pulmonata). Environ Pollut 2004;127:303–11.
Russo J, Madec L. Haemocyte apoptosis as a general cellular immune response of the
snail, Lymnaea stagnalis, to a toxicant. Cell Tissue Res 2007;328:431–41.
Russo J, Lefeuvre-Orfila L, Lagadic L. Hemocyte-specific responses to the peroxidizing
herbicide fomesafen in the pond snail Lymnaea stagnalis (Gastropoda, Pulmonata).
Environ Pollut 2007;146:420–7.
Sacre S, Medghalchi M, Gregory B, Brennan F, Williams R. Fluoxetine and citalopram
exhibit potent antiinflammatory activity in human and murine models of rheuma-
toid arthritis and inhibit Toll-like receptors. Arthritis Rheum 2010;62:683–93.
Segura PA, Gagnon C, Sauve S. A fully automated on-line preconcentration and liquid
chromatography-tandem mas spectrometry method for the analysis of anti-infectives
in wastewaters. Anal Chim Acta 2007;604:147–57.
Sminia T. Structure and function of blood and connective tissue cells of the fresh water
pulmonate Lymnaea stagnalis studied by electron microscopy and enzyme histo-
chemistry. Z Zellforsch Mikrosk Anat 1972;130:497–526.
Stefano GB, Cadet P, Breton C, Goumon Y, Prevot V, Dessaint JP, et al. Estradiol-stimulated
nitric oxide release in human granulocytes is dependent on intracellular calcium
transients: evidence of a cell surface estrogen receptor. Blood 2000;95:3951–8.
Stefano GB, Mantione K, Jones D, Zhu W, Casares F, Cadet P. Immunocytes modulate
ganglionic nitric oxide release which later affects their activity level. Neuroendocrinol
Lett 2004;25:57–61.
van der Knaap WPW, Adema CM, Sminia T. Invertebrate blood cells: morphological and
functional aspects of the haemocytes in the pond snail Lymnaea stagnalis. Comp
Hematol Int 1993;3:20–6.
Yu J, Roh S, Lee JS, Yang BH, Choi MR, Chai YG, et al. The effects of venlafaxine and dexa-
methasone on the expression of HSP70 in rat C6 glioma cells. Psychiatry Investig
2010;7:43–8.