Oxidative and apoptotic effects of fluoxetine and its metabolite norfluoxetine in Daphnia magna

D. magna specimens were purchased from Carolina Biological Supply Company (Burlington, NC, USA) and were acclimatised to laboratory conditions in 15 L of hard water (American Society for Testing and Materials, ASTM) in plastic containers for a minimum of ten generations (17). The physicochemical properties of hard water were as follows: pH 8.00, 20.20±0.35 °C, total alkalinity 120.42±7.77 mg/L CaCO₃, total hardness 138.51±8.52 mg/L CaCO₃, and light intensity 10.53 μmol/m₂/s. The light-to-dark photoperiod was 16:8. Animals were fed every other day with a 1.5 mL/L mixture of trout chow, yeast, and alfalfa (18). Before the toxicity experiments, the fifth-generation mothers from one mother were used as the source of experimental organisms. Less than 24 h-old second-generation offspring of these mothers were used in the 24 h K₂Cr₂O₇ EC₅₀ susceptibility test. The half maximal effective concentration (EC₅₀) was 0.972 mg/L (95 % confidence interval was 0.904–1.042) and was in the range proposed by the Organisation for Economic Co-operation and Development (OECD) (19).

Fluoxetine hydrochloride (Supelco PHR1394) and its human metabolite norfluoxetine hydrochloride (Supelco 40724) were obtained from Supelco, Inc. (Bellefonte, PA, USA). Serial dilution method was used to prepare the exposure media. The efficiency of our dilution method was confirmed by liquid chromatography with tandem mass spectrometry (LC-MS/MS). 100 μL of water sample was injected into the LC-MS/MS device (LCMS-8040, Shimadzu, Kyoto, Japan) equipped with Inertsil^® ODS-4 colon (3 μm, 2.1 x 50 mm, GL Sciences, Japan). Flow rate, colon temperature, and mobile phases A and B were 0.4 mL/min., 40 °C, and 0.2 mmol/L for ammonium formate + 0.004 % formic acid and methanol, respectively. Parent and daughter ions and retention time for Flx were 310.2, 148.1, and 8.033 min, respectively. Parent and daughter ions and retention time for NorFlx were 296.1, 134.1, and 8.011 min, respectively. As our device and method could not detect environmental concentrations, we spiked environmental solutions with 1 μg/L Flx or NorFlx and calculated the difference between 1 μg/L and spiked solutions. Method accuracy expressed in terms of recovery was ≥89.78 % for Flx and ≥93.00 % for NorFlx. r² values of standard calibration curve prepared for both chemicals was 0.999 (Table 1).

For the chronic toxicity experiments we used third-generation offspring younger than 24 h divided in seven groups and kept in beakers containing 800 mL of ASTM hard water: control group, Flx group (0.091 μg/L as environmentally realistic concentration), NorFlx group (0.011 μg/L as environmentally realistic concentration), Flx+NorFlx group (mixture of their environmentally realistic concentrations), 10 x Flx group (0.91 μg/L), 10 x NorFlx group (0.11 μg/L), and 10 x Flx+10 x NorFlx group (a mixture of their 10-fold environmental concentrations). Twenty-one animals were placed in each beaker, and four beakers for each group were used as replicates. The experiments lasted 21 days. Water was completely changed every third day. During the change, 1.5 mL/L of food was pipetted into beakers. Less than 24 h-old nestlings in each beaker were counted and collected into 50 mL Falcon tubes (containing 50 % ethyl alcohol) at 24 h intervals starting from the eighth day. Collected animals were stored at +4 °C before microscopy.

For subacute, four-day exposure experiments we used the fourth-generation offspring of 500 less than 24 h-old animals divided into two plastic 15 L containers filled with ASTM hard water. The animals were fed 1.5 mL/L trout chow/yeast/alfalfa mixture for 17 days. On day 17, 21 animals per group were placed into a beaker containing 800 mL of ASTM hard water. The groups included control, 100 x Flx (9.1 μg/L), 100 x NorFlx (1.1 μg/L), and 100 x Flx+100 x NorFlx (a mixture of their 100-fold environmental concentrations). Five beakers for each group were used as replicates. 1.5 mL/L of food was pipetted into each beaker. After 48 h, water was changed, and the animals stopped receiving food. At the end of 96 h exposure period, animals were 21 days old.

At the end of subacute and chronic experiments, one animal per beaker was placed into a microplate well and the microplate stored at -80 °C until 5,6-dichloro-2-[(E)-3-(5,6-dichloro-1,3-diethylbenzimidazol-3-ium-2-yl)prop-2-enylidene]-1,3-diethylbenzimidazole;iodide (JC-1; CAS No: 3520-43-2) staining for MMP analysis. The remaining animals from each beaker were carefully collected on a filter, washed with clean ASTM hard water, blotted dry, placed into one micro tube, weighed (all these steps were done on ice as much as possible), and stored at -80 °C until biochemical analysis. The procedure was repeated with four beakers as replicates until we acquired an adequate sample for Western blotting of cytochrome c and ERK1/2.

Before biochemical analysis, a sample of 20 animals from each microtube were homogenised with chilled 100 mmol/L KCl and 1 mmol/L EDTA (containing pH 7.4 100 mmol/L potassium phosphate buffer on ice at a ratio of 1:4 w/v in a homogeniser equipped with stainless steel shaft (HS30T, WiseStir, Daihan, Korea). The homogenate was then centrifuged at 10000 g and +4 °C for 10 min. The obtained supernatant was used to measure the activity of glutathione peroxidase (GPx) and cholinesterase (ChE) and to measure thiobarbituric acid-reactive substances (TBARS) and total proteins. Spectrophotometric analysis was done in a microplate reader (EON^TM, BioTek, Winooski, VT, USA).

Glutathione peroxidase (GPx) protects the cell against oxidative stress by reducing hydrogen peroxide (H₂O₂) and organic peroxides via glutathione consumption (20). It was analysed according to the method described by Beutler (21). Briefly, 25 μL of supernatant was incubated with 10 μL of 1 mol/L pH 8.0 Tris buffer, 25 μL of 0.1 mol/L glutathione (GSH), 100 μL of 10 U/mL glutathione reductase (GR), and 100 μL of 2 mmol/L reduced β-nicotinamide adenine dinucleotide phosphate (NADPH) at +37 °C for 10 min. The reaction was started with the addition of 7 mmol/L t-butyl hydroperoxide. Absorbance change was measured at 340 nm for 2 min using the 6.22 L/mmol/cm extinction coefficient.

Cholinesterases (ChE) are proposed as biomarkers for apoptosis (and neurotoxicity) because of increased expression during apoptotic conditions (22). Cholinesterase activity was analysed according to the method of Ellman et al. (23). Briefly, 100 μL of supernatant were added to a 125 μL mixture of 0.1 mol/L potassium phosphate buffer (pH 7.2), 10 mmol/L 5,5’-dithiobis-(2-nitrobenzoic acid) (DTNB), and 10 mmol/L sodium bicarbonate and incubated at room temperature for 10 min. The reaction was started with the addition of 100 μL of acetylthiocholine iodide. Absorbance change was measured at 405 nm for 5 min using the 13.6 L/mmol/cm extinction coefficient.

TBARS is a universal measure of reactive oxygen species (ROS) damage on unsaturated fatty acids (24). Its levels were measured according to the method described by Wills (25). 25 μL of supernatant was mixed with 25 μL of 10 % trichloroacetic acid (TCA). After centrifugation at 500 g for 10 min, 37.5 μL of supernatant were mixed with 37.5 μL of 0.65 % thiobarbituric acid and incubated at 100 °C for 10 min. The absorbance was measured at 535 nm and converted to concentration values using a standard calibration curve prepared with 1,1’3,3’-tetra methoxy propane.

Total protein levels were measured with a modified Lowry method (26). 10 μL of supernatant were mixed with 45 μL of Lowry reagent and incubated at room temperature for 10 min. Then, 45 μL of Folin & Ciocâlteu’s phenol reagent were added, and the mixture incubated at room temperature for another 30 min. The absorbance was measured at 750 nm and converted to concentration values using a standard calibration curve prepared with bovine serum albumin (BSA).

JC-1 probe was used as qualitative measure of change in mitochondrial membrane potential (27). Membrane depolarisation is a common biomarker of apoptotic cell death (28, 29). At low potential, the membrane turns fluorescent green, but at higher potential it turns fluorescent red. Each D. magna sample was incubated in its microplate well with 6 mmol/L JC-1 in dark at room temperature for 90 min. Fluorescence was measured by a confocal laser microscope (LSM700, Carl Zeiss, Oberkochen, Germany).

In healthy cells cytochrome c is found between mitochondrial membranes and is mainly responsible for oxidative phosphorylation. However, its release into the cytosol triggers the intrinsic pathway of apoptosis through the formation of mitochondrial permeability transition pore (MPTP) (30). The other two apoptosis-related proteins we studied – extracellular signal-regulated kinases 1 and 2 (ERK1/2) – act together to increase the activity of anti-apoptotic proteins or their transcription factors to prevent apoptosis. However, they also behave as an apoptotic factor through yet unknown mechanisms (31). For Western blotting of cytochrome c and ERK1/2, 20 D. magna animals were pooled from four beakers per group (totalling 80 animals) and homogenised as one replicate with radioimmunoprecipitation assay (RIPA) lysis buffer. Protein levels were measured following a modified Lowry method (26), in which samples were mixed with lithium dodecyl sulphate buffer at a ratio of 3:1 v/v and denatured at +100 °C for 10 min. 100 μg of protein-containing samples were loaded onto 10 % polyacrylamide gels. Protein bands were constituted for about 1 h and transferred on polyvinylidene fluoride (PVDF) membranes with an iBlot 2 transfer device (Invitrogen, Waltham, MA, USA). The transfer lasted 10 min. A signal enhancer solution was used to better visualise the bands. PVDF membrane was saturated with 5 % BSA to prevent non-specific binding. The membranes were then treated with primary antibodies (cytochrome c and ERK1/2, diluted to 1/1000 concentration) at +4 °C for one night. After washing with PBS-Tris solution (PBS-T), the membranes were treated with secondary antibodies (diluted to 1/6000 concentration, anti-rabbit) at room temperature for another hour. Re-washed with PBS-T solution, the membranes were then treated with a solution of 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium (BCIP/NBT) to visualise and take images of the bands. Their intensity was measured with ImageJ v1.46r software (National Institutes of Health, Sacaton, AZ, USA).

For morphometric analysis nestlings were placed in a concave microscope slide, which contained lactic acid solution. They were photographed under a binocular microscope (BX53, Olympus, Tokyo, Japan) equipped with a Canon EOS1200D camera (Canon Inc., Tokyo, Japan) with 40x magnification. Carapace length was measured between the anterior head and caudal spine basement (32) using the Micam v. 2.0 imaging software (http://www.science4all.nl/?Microscopy_and_Photography). The caudal spine was not included in this measurement, as it constituted another toxicity parameter. Maximum carapace width was measured laterally from the maximum width of carapace ends.

Table 2 shows neurotoxic and oxidative effects of chronic and subacute exposure to Flx, NorFlx, and their combinations. Cholinesterase activity was particularly high at 100-fold environmental concentrations, but NorFlx at environmental concentration and at ten times that also increased ChE activity, while the increase with Flx and Flx+NorFlx was not statistically significant. Li and Tan (33) also reported a rise in ChE in D. magna by triazophos and chlorpyrifos and interpreted it as an adaptive de novo synthesis to maintain homeostasis via discontinuation of acetylcholine-induced ion channel activity (34).

Considering that ChE activity is increased by apoptosis inducers in vitro and in vivo (22, 35), ChE can serve as a biomarker of apoptosis. The observed increase in ChE activity in our study may be the consequence of increased acetylcholine release, which we did not measure, but which was also reported elsewhere (36) and may be related to the known effects of serotonin. Namely, serotonin increases intracellular Ca²⁺ levels (37) and, in turn, stimulates acetylcholine release (38). Furthermore, consecutive synergistic action of serotonin and acetylcholine may intensify the Ca²⁺ uptake (39, 40), which can increase oxidative stress (41), as evidenced by other studies (42, 43). Ca²⁺ at high concentrations can also cause mitochondrial damage. ROS production and consecutive induction of apoptosis can be triggered via net intracellular Ca²⁺ uptake or release from endoplasmic reticulum stores (41). According to Yardimci et al. (44), increased ChE activity may be an adaptive mechanism against acetylcholine-induced and Ca²⁺-mediated ROS production, which they observed in the kidney of imidacloprid-exposed rats. As for the oxidative effect observed in our study through higher lipid peroxidation levels at Flx and NorFlx concentrations higher than environmental, Yang & Dettbarn (45) reported a connection between high lipid peroxidation and increased acetylcholine levels caused by anticholinesterase effects of diisopropyl fluorophosphate.

At 100-fold environmental concentrations Flx, NorFlx, and their combinations decreased total protein levels, but this effect was not observed in chronic toxicity experiments involving environmental and 10-fold environmental concentrations (Table 2). Concentration- and duration-dependent decrease in total protein levels was also observed in D. magna exposed to herbicide tebuconazole (46). Selective serotonin re-uptake inhibitors (SSRIs) increase aerobic metabolism with oxygen consumption (47), which points to increased ROS production (29) and suggests that increased protein catabolism to compensate for increased energy demands may be an important biomarker of toxic stress, as suggested in a study by Villarroel et al. (48).

In contrast to our findings, Campos et al. (47) reported no significant changes in total protein and lipid content in adult D. magna exposed to much higher Flx concentrations (80 μg/L) for 24 h. Instead, they reported increased carbohydrate and oxygen consumption, which may point to higher ROS levels associated with increased aerobic metabolism (29). As several studies suggest higher protein carbonyl levels in oxidative stress (49), which ends with proteasomal degradation of oxidised proteins (50), the lowering of total protein levels in our subacute experiment may display the oxidative damage on tissue proteins.

Glutathione peroxidase as first-line defence against oxidative stress also increased at both 10 and 100 times Flx and NorFlx environmental concentrations. At environmental levels, however, its basal activity seems to have been sufficient to counter ROS and prevent oxidative damage (Table 2). A case in point could by the anticholinesterase insecticide chlorpyrifos study (49), which increased cytosolic ChE and GPx activities and H₂O₂ levels but not lipid peroxidation. In our study, the concentration-dependent increase in GPx activity and lipid peroxidation levels should be viewed through redox balance, which tipped toward oxidation at higher Flx and NorFlx concentrations. Similar concentration-dependent changes in D. magna was reported with organochlorine insecticide endosulfan (51) and antibiotic chloramphenicol (52).

Mitochondrial membrane depolarisation is a common marker of cell death and physiological condition of cells and tissues (28, 53). In our study, mitochondrial membrane depolarisation was observed in all the Flx, NorFlx, and combination groups in both chronic and subacute exposure (Figures 1 and 2) and especially with NorFlx. This suggests that Flx and NorFlx exert their toxicity via mitochondrial mechanisms directly or indirectly. Similar mechanisms have been observed in other studies in different models, including D. magna (42, 54, 55).

Figure 1

Figure 2

Mitochondrial membrane depolarisation is an important mechanism for the release of cytochrome c from the mitochondria (28), and cytochrome c release induces mitochondrial apoptosis in mammalian cells (56). In our study, cytochrome c protein levels were generally increased, except for the Flx and 10 x Flx+10 x NorFlx combination groups (Figures 3 and 4). However, increased cytochrome c protein content alone does not provide strong evidence of the mitochondrial release and apoptosis in our study, but in combination with increased lipid peroxidation and mitochondrial membrane depolarisation it does give grounds for such an assumption. We believe that increased cytochrome c levels in our study are related to intensive stress conditions and higher mitochondrial damage induced by Flx and NorFlx. Such a result was also found in the liver of mercury chloride-exposed rats (57).

Figure 3

Either compound or their combination did not affect ERK1/2 protein levels at environmental concentrations but did at higher concentrations, except for 100 times higher NorFlx alone (Figures 3 and 4). ERK1/2 is activated by a toxicant and regulates response to cellular damage (58). It promotes cell survival against apoptotic mechanisms, but in certain condition it also favours apoptosis (31). This may explain different responses to NorFlx in our study and calls for further investigation. We believe that increase in ERK1/2 protein levels may be an adaptive response to apoptotic conditions.

Figure 4

Table 3 shows a significant increase in carapace length and width in the 2^nd and 3^rd generation offspring of the 10 x Flx+10 x NorFlx and 10 x Flx groups, respectively. Width also increased in the 10 x NorFlx group. Prey animals experience a variety of changes to avoid their predators. They may include toxin production (physiological changes), late breeding (life-cycle changes), and production of spicules and sheathing (morphological changes) (59). Flx alone and in combination with its metabolite may have triggered this avoidance behaviour at 10-fold environmental concentrations. Decreased carapace length, width, and spine length seen in other groups since the 2^nd generation offspring may be related to lower energy production, as reported elsewhere (60, 61). In a study by Pery et al. (62) Flx also decreased carapace length of less than 24 h-old 3^rd generation D. magna, but the effect was observed at higher concentrations than those used in our study (62). Campos et al. (63) suggest that decreased body size in offspring is a result of poor adaptation to toxic conditions and we believe that long-term exposure to SSRIs may impair adaptation of non-target populations even at low concentrations.

NorFlx alone and in combination with Flx significantly increased brood production at their environmental concentrations until the 5^th generation offspring (Table 4). A similar increase in the first four generations was also observed with Flx and its combination with NorFlx at 10-fold concentrations. Campos et al. (47) suggested that brood production of D. magna in their study may have been increased due to higher serotonin availability. Serotonin affects the neurohormones, which play a role in reproduction, growth, pubescence, larval development, immune functions, metabolism, behaviour, and colour physiology in crustaceans (64). We therefore believe that Flx alone or in combination with NorFlx may have increased brood rate by blocking serotonin reuptake in our study.

However, in the 5^th generation offspring brood production dropped in all groups (Table 4), which may be related to the bioaccumulation of Flx and NorFlx in D. magna (14). A similar drop in brood production rate was seen at higher Flx concentrations and longer exposure in other studies (62, 63). In fish models higher Flx and NorFlx concentrations were reported to decrease egg production (3, 65).

In conclusion, D. magna can serve as a mechanistic model for understanding toxic effects in non-target organisms. However, more studies are needed to explain their cellular mechanisms as clearly as those of the vertebrates. Flx, NorFlx, and their combinations induced morphological and/or biochemical responses, especially at 10-fold and 100-fold environmental concentrations, which suggests that their increasing presence in aquatic environments will have even worse effects on non-target organisms than they have now. However, our study has left some questions unanswered, and future research of serotonin, acetylcholine, Ca²⁺, and ROS levels should address them, including the mechanistic relationships between them and the parameters studied here.