Fluoxetine Accumulation and Metabolism as Exposure Biomarker to Better Understand Biological Efects in Gastropods ^†

Introduction

Fluoxetine is a widely used antidepressant frequently found in wastewater treatment plant effluent and in aquatic ecosystems at concentrations below µg/L [1]. Numerous studies have explored the sensitivity of invertebrates to waterborne fluoxetine showing marked dif- ferences of sensitivity. In a previous study, the New-Zealand mudsnail, Potamopyrgus antipo- darum, and the European valve snail, Valvata piscinalis, were exposed to waterborne fluoxe- tine (3.7-100 µg/L) during 42 days. While effects were observed on the mudsnail repro- duction and F1, the valve snail was not affect- ed. Several assumptions were proposed to explain these results including dissimilar metabolic capacities or bioavailability of fluox- etine for the snails [2]. Indeed, recently, fluoxe- tine was measured in tissues of mussels Eliptio complanata exposed downstream of a municipal wastewater, assorted with reproduc- tive effects under laboratory experiments [3]. Both fluoxetine and norfluoxetine (the active metabolite) have been measured in tissues of fishes sampled or caged in streams down- stream of urban and industrialized effluent discharges [4,5,6,7]. Thus the aim of this study was to asses if differential sensitivity of gastropods snails to fluoxetine could be explained by its bioaccumulation and metabolism into norflu- oxetine.

Materials and Methods

P. antipodarum and V. piscinalis were exposed in the same beakers to fluoxetine as previously described and concentrations were chosen in accordance with the biological effects observed in these species.2 However, due to lack of abundance of V. piscinalis in our culture, exposure of the both species was not rigorously similar. The nominal exposure con- centrations were 11, 33 and 100 µg/L for P. antipodarum plus a control, and 33, 100 µg/L plus a control for V. piscinalis. Six replicates per concentration were prepared and exposed for 7 (V. piscinalis) and 14 days (P. antipo- darum). Organisms were fed ad libitum. Adult snails were used at the start of the experiment. On days 7 and 14, 3 P. antipodarum were col- lected per condition for fluoxetine and norflu- oxetine analysis and 3 V. piscinalis on day 7. Due to the lipophilic nature of fluoxetine [8,9], tri- acylglycerol (TG) content of snails was meas- ured [10], as proxy of total lipid content, TG being the principal lipid class in mollusks [11].

Fluoxetine was measured in water on days 7 and 14 as published elsewhere [10], as well as flu- oxetine and norfluoxetine in snails. Briefly, an easy and quick extraction similar to Quick Easy Cheap Effective Rugged Safe (QuEChERS) procedure [12] in a miniature ver- sion was developed to analyze both compounds in gastropods. The procedure involves an extraction of about 10 milligrams of matrix by 500 µL of a mixture of acetonitrile: water:hexa- ne (50/20/30) and 100 mg of citrate buffer. Recoveries were 87% for fluoxetine and 86% for norfluoxetine. Nano-LC-nano-ESI MS/MS analysis was performed with a nano Ultimate3000 (Dionex®) coupled with a Qtrap3200 detector (AB Sciex®). Data were processed with Analyst 1.5. The chromato- graphic separation was performed in two steps: first a preconcentration step on a precol- umn C18 (5 µm, 300 Å, LC Packings®) followed by a separation on a Pepmap C18 column (C18 3 µm x 75 µm x 100 Å, Dionex®). The column oven temperature was set to 60°C; injection volume was 1 µLin µLpickup mode; the flow rate was 300 nL/min. Samples were analyzed in positive mode with the mobile phase (A) CH3CN/FA/H2O (2/0.1/97.9) and (B) CH3CN/FA/H2O (80/0.08/19.92), with the fol- lowing elution program: start at 60% (A), from 60 to 0% in 5 min and from 5 to 20 min, 100% (B). The column was re-equilibrated 20 min between runs. MS/MS detection was per- formed in the multiple reaction monitoring (MRM) mode using a NanoSpray® II source (AB Sciex®). Fluoxetine and norfluoxetine were identified by their retention times, two characteristic ion transitions (fluoxetine: 310>148 and 310>91, norfluoxetine: 296>134 and 296>30) and specific ratios of the respons- es of the transitions (deviation <20% with respect to analytical standards ratios). The quantification was carried out by double injec- tion. The use of matrix-matched standards was selected. The matrix blank (snail) was pre- pared as for analysis of samples. Six-point cal- ibration curves were built from the injection of these matrix prepared standards. This calibra- tion intended to compensate for matrix effects. In these conditions, the limit of quantification for fluoxetine and norfluoxetine in snail sam- ples was respectively 7.7 and 3.8 ng/g.

Fluoxetine accumulation on days 7 and 14, was assessed by calculating bioaccumulation factor (BCF) estimates according to Nakamura et al. [9] The estimate (BCF_totBCF_tot = C_snail/ C_water) was calculated, using the fluoxetine concen- trations measured in snails after 7 and 14 days (C_snail) for P. antipodarum and 7 days for V. piscinalis.

Results and Discussion

During the exposure, mean water tempera- ture was 21.0±0.3°C, mean conductivity 332±25 µS/cm and mean pH 7.9±0.3. Chemical measurements showed a fluoxetine recovery in the exposure system depending on the test- ed concentration (from 25 to 98%), with mean measured exposure concentrations of 2.9, 24 and 79 µg fluoxetine/L (

Table 1. Measured concentrations of fluoxetine in water (µg/L), snails (µg/g), and of norfluoxetine in snails (µg/g).

D0, D7 and D14: days 0, 7 and 14; LOD, limit of detection; LOQ, limit of quantification.

). Even if the gradient of exposure still was present, fluoxe- tine measured concentrations were lower than expected ones, with the lowest recovery rates for the lowest tested concentration. This is consistent with what was previously observed in similar exposure systems.2 These lower exposure concentrations can certainly be explained by adsorption on the devices used for exposure system, as fluoxetine in a rela- tively stable compound [13].

Fluoxetine was measured in both P. antipo- darum and V. piscinalis (Table 1). No signifi- cant differences (P>0.05) were noted between the fluoxetine concentrations in P. antipo- darum after 7 or 14 days, suggesting that the bioaccumulation reached steady state after 7 days of exposure. In the Japanese medaka, bioaccumulation steady state was achieved after only 3 days of exposure [8]. On the basis of these results which showed quick accumula- tion for fish and snail, we assumed that steady state could also be reached in V. piscinalis after 7 days of exposure and we compared fluoxetine body burden in the two species on day 7. Concentrations of fluoxetine and estimate BCF_tot in P. antipodarum were significantly higher (P<0.05) than in V. piscinalis (Table 1 and

Table 2. Bioaccumulation factor (BCF) and pseudo-BCF in aquatic organisms.

BCF, bioaccumulation factor; B, body; L, liver, B+L, body + liver; M, muscle; AT, adipose tissue.

) [3,8,9,14], implying a higher bioaccumula- tion of the parent compound in P. antipodaru- mafter 7 days of exposure. Exposure systems were rigorously similar for both species. As flu- oxetine was waterborne, the contamination probably preferentially occurred across the gills. However, fluoxetine can adsorb on food as it does on sediment13 and thus oral route is not excluded and total exposure could be mod- ified by various food consumption. Indeed, both snails have different feeding modes, P. antipodarum being a deposit feeder and V. piscinalis being a scraper [15].

Many studies showed that fluoxetine con- centrations were higher in lipid rich tissues, than in muscles (Table 2). Thus, it can be assumed that a difference in lipid content in both snails could also explain the differences in BCF_tot. TG content in P. antipodarum was around seven times higher than in V. piscinalis (Figure 1), which probably explains the higher bioaccumulation of fluoxetine in the mudsnail compared to the valve snail. Therefore, the lesser sensitivity of V. piscinalis to waterborne fluoxetine compared to P. antipodarum [2] could be at least partly explained by its lesser bioac- cumulation in the whole body.

BCF estimates in both snails were far lower than in mussel, but higher than those calculat- ed in fishes (Table 2). Fluoxetine bioaccumu- lation is strongly dependant of Ph [9], and no data is available concerning the exposure pH of E. complanata,3 thus limiting the comparison between species. Dietary route of exposure and lower exposure concentrations might also be in cause.

As for fluoxetine, norfluoxetine was meas- ured in both P. antipodarum and V. piscinalis (Table 1). To our knowledge, this is the first study to investigate in vivo metabolism of a pharmaceutical in mollusk. Contrary to parental compound, norfluoxetine concentra- tions measured in both species were not sig- nificantly different (P>0.05), and norfluoxe- tine levels in snails represented less than 1% of total fluoxetine body burden. This suggests that N-demethylation is not the primary metabolite as in fish [16].

Since norfluoxetine was measured in both snails, pseudo-BCFs estimates (pseudo-BCF_tot) were calculated according to Nakamura et al. [9], the denominator being the concentration of fluoxetine in water and not the concentration of norfluoxetine. Pseudo-BCF estimates (Table 2) were much lower than those determined for fluoxetine. In fishes, higher norfluoxetine con- centrations in comparison to fluoxetine are measured and pseudo-BCF estimates were far higher than in gastropods (Table 2). Thus N- demethylation capacity is higher in fishes than in gastropods. In snails, total accumulation of fluoxetine can be reduced to fluoxetine accu- mulation.

The biological effects in gastropods are sup- ported by the higher bioaccumulation of fluox- etine in the mudsnail. In humans, the norflu- oxetine is more active than the parent com- pound [17], as in a protozoan and a crustacean 24 h lethality-test [18]. However, the presumed active metabolite is measured in similar very low quantities in both species, implying that metabolite activation into norfluoxetine does not explain the interspecific differences previ- ously observed. However, other metabolites cannot be excluded. Thus further studies are needed in order to better assess these issues.