Exposure Effects of Environmentally Relevant Concentrations of the Tricyclic Antidepressant Amitriptyline in Early Life Stage Zebrafish

Antidepressants are one of the most globally prescribed classes of pharmaceuticals, and drug target conservation across phyla means that nontarget organisms may be at risk from the effects of exposure. Here, we address the knowledge gap for the effects of chronic exposure (28 days) to the tricyclic antidepressant amitriptyline (AMI) on fish, including for concentrations with environmental relevance, using zebrafish (Danio rerio) as our experimental model. AMI was found to bioconcentrate in zebrafish, was readily transformed to its major active metabolite nortriptyline, and induced a pharmacological effect (downregulation of the gene encoding the serotonin transporter; slc6a4a) at environmentally relevant concentrations (0.03 μg/L and above). Exposures to AMI at higher concentrations accelerated the hatch rate and reduced locomotor activity, the latter of which was abolished after a 14 day period of depuration. The lack of any response on the features of physiology and behavior we measured at concentrations found in the environment would indicate that AMI poses a relatively low level of risk to fish populations. The pseudopersistence and likely presence of multiple drugs acting via the same mechanism of action, however, together with a global trend for increased prescription rates, mean that this risk may be underestimated using current ecotoxicological assessment paradigms.


Flow rates
Supp Table 1.Average flow rate of the amitriptyline stock solution and RO water over the 28-day exposure period (mean ± SEM, n = 5).

Water quality parameters
Water quality parameters, including; pH, temperature, dissolved oxygen (FireStingO2 fiber-optic oxygen meter, PryoScience Sensor Technology) and ammonia (Tropic Marin Ammonia/Ammonium Test Kit) were measured throughout the experiment on a weekly basis (pH was measured twice a week).
Supp Table 2. Tank water quality parameters, including pH, temperature and dissolved oxygen.Outlined as the mean (SEM) over the entire 42-day duration of the exposure where n = 5.

Measurement of AMI in water samples
For analysis, 800 µl of tank solution, plus 200 µl of AMI-d3 hydrochloride solution (CAS: 342611-00-1) that had been diluted to 96 nM in methanol, was transferred to a small glass vial prior to vortexing.Once well mixed, 100 µl was transferred to a deep 96 well plate for analysis by LC-MS/MS.Analyses of water samples were performed using a TSQ Vantage triple quadrupole mass spectrometer.The mass spectrometer was equipped with a heated electrospray (HESI II) source (ThermoFisher Scientific, Hemel Hempstead, UK).The HESI probe was operating in positive mode; an ion-spray voltage of 4.0 kV, heated capillary temperature was set at 270 °C and the vaporizer temperature was 350 °C.Nitrogen was employed as a sheath and auxiliary gas at a pressure of 60 and 2 arbitrary units, respectively.The argon CID gas was used at a pressure of 1.5 mTorr and the optimum CE for each transition was selected.
Quantification was performed by monitoring two characteristic MRM transitions for AMI and one for the d3-AMI internal standard.Chromatographic separation was achieved using a reversed-phase, 3 µm particle size, C18 Hypersil GOLD column (50 mm × 2.1 mm id., Thermo Scientific, San Jose CA, USA).All analytes were separated using a linear gradient with a flow rate of 500 µL min -1 .The autosampler temperature was maintained at 8˚C, while the column was kept at room temperature.

Rationale and aim
To avoid dilution of the exposure solution that would come with administering Artemia nauplii and minimise any contribution of trophic transfer from live prey and biomagnification of amitriptyline in tissues, a dry food ONLY diet was used.
Furthermore, this allowed control of the mass of feed in each tank relative to the number of surviving individuals i.e. this allowed adjustment of the mass of dry food to account for mortalities.First, the growth of healthy zebrafish larvae was established from 5 to 42 dpf (duration of the experiment where fish were to be fed exogenously) when fed an optimum standard aquarium diet.Using this information, 3 feeding regimes were tested to determine the optimum feeding protocol with dry feed only.

Zebrafish weight measurements from 5 to 42 dpf
Zebrafish were housed in accordance to the standard aquarium guidelines as follows: from 5-7 dpf, larvae were housed in 1 L static tank and fed 1 dry feed (1/2 scoop <100 ZEBRAFEED by Sparos) and 3 rotifer feeds (3ml) daily.From 7 to 20 dpf, larvae were housed in 3.5 L tanks with a drip rate of 1-2 drips per second and fed 1 dry feed (1 scoop <100 ZEBRAFEED by Sparos) and 3 rotifer feeds (5 ml) daily.From 21-30 dpf, larvae were housed in 3.5 L tanks with a drip rate of 1-2 drips per second and fed 2 dry feeds (1/2 scoop 100-200 ZEBRAFEED by Sparos), 1 rotifer feed (5 ml) and 1 artemia feed (1 ml) daily.From 31-40 dpf, larvae were housed in 3.5 L tanks with a drip rate of 2-3 drips per second and fed 2 dry feeds (1 scoop 100-200 ZEBRAFEED by Sparos) and 1 artemia feed (2 ml) daily.From 41 dpf onwards, larvae were housed in 8 L tanks with a drip rate of 2-3 drips per second and fed 2 dry feeds (2 scoops 200-400 ZEBRAFEED by Sparos) and 1 artemia feed (5 ml) daily.At 5-day intervals commencing from 5 to 45 dpf, 10 fish were removed from stock tanks (2/3 from each tank) and placed in a plastic Petri dish containing 50ml reverse osmosis (RO) water.Following schedule 1 termination, fish length (measured using Leica M205C with a Leica DMC4500 camera using 1x panio objective and LAS X programme) and weight were measured (Figure S2).At 5 and 10 dpf, to increase measurement accuracy due to their small size, body weight measurements were made by pooling two fish and then dividing this measurement by 2 to provide an average individual weight for those two individual fish.From 15dpf onwards, individual fish weights were taken.

Supp Figure 2. Zebrafish larva growth over time as depicted by weight, when fed standard aquarium diet of Artemia nauplii, rotifers and dry food combined (mean ± SEM, N = 5 at 5 and 10dpf and 10 from 15dpf onwards).
The y-axis is presented on a logarithmic scale.

Feeding trial
Adult fish were spawned and embryos collected as outlined in section 2.3.
Embryos-larvae were kept in Petri Dishes containing 50 ml aquarium water (75% daily water change) at 28  1C until 5 dpf.From the onset of independent feeding ( 5dpf), larvae were fed Zebrafeed, Sparos dry food at one of three quantities, 50, 175 or 300% of total body weight (as calculated above), or the standard aquarium feed (outlined above).The daily body weights of larvae were selected from Figure S2.
Each treatment had 2 replicate tanks, each housing 55 larvae at 5 dpf and then until 20 dpf, at 5-day intervals, fish were removed from each tank for length and weight measurements.At 5 dpf, to increase measurement accuracy due to their small size, body weight measurements were made by pooling two fish and then dividing this measurement by 2 to provide an average individual weight for those two individual fish (2 fish sampled per tank).From 10dpf onwards, individual fish weights were taken (10 fish sampled per tank).
As seen from Figure S3, food build up in the 175 and 300% tanks was excessive, particularly in the 300% tanks which is likely why at 20 dpf, larvae in one of the 300% tanks died (ammonia/bacteria accumulation) due to excess debris.
Furthermore, it was observed that larvae in the 50% tank spent the majority of their time swimming at the water's surface, indicative of searching/waiting for food.It should be noted that tanks were not cleaned over the 15-day period.Larvae fed on the control diet (standard aquarium diet) had the highest growth rate which was expected as this Supp Figure 4. Zebrafish larvae growth over time when fed different percentages of their body weight in dry food compared to the standard aquarium control diet (mean ± SEM, N = 2 at 5 dpf and 10 from 10 dpf onwards).
feeding regime has been established to be optimum for growth in this aquarium environment (Figure S4).All three dry feed only regimes resulted in larvae having reduced growth compared to the control (Figure S4).The cleanliness of the tanks diminished (build-up of debris) as the amount of food added increased, and is ultimately what led to unfavourable living conditions, particularly for those on the 300% diet.However, the 300% treatment yielded body weights most similar to larvae fed on the standard aquarium diet so to enable sufficient growth, for the main exposure study, larvae were fed 300% of their body weight in dry feed from 5 dpf and tanks were cleaned daily.When it was clear this ration was too great (i.e.food continued to accumulate in the tanks as fish were not eating it all), the % body weight of food given was adjusted accordingly.

Section 3: Measurement of AMI in tissue samples
Once defrosted, individual samples (WB, WH and HNE) were transferred to a new 2 ml Eppendorf containing 600 µl of acetonitrile:water 1:1 mix with 25 µg L -1 internal standards of AMI and NOR spiking solution, alongside a stainless steel ball bearing.Samples were then homogenised in a Tissuelyser (Qiagen) at 300 Hz for 3 minutes before being centrifuged at 160,000g for 1 hour at room temperature.250 µl of the supernatant was then transferred to HPLC glass vials lined with inserts for LCMS/MS analysis.At the highest treatment level, AMI concentration was measured in head samples (with and without the eyes), alongside the remaining trunk for the same animal.Heads were analysed singularly for the higher concentrations but were pooled in groups of 3 for the lower concentrations.
The analysis of AMI and NOR was again undertaken on a TSQ Vantage triple quadrupole mass spectrometer as detailed above.All conditions were the same as for the water samples except that the analytes were separated using a linear gradient of water and methanol, both containing 0.1% formic acid.The initial conditions for the gradient consisted of 20% methanol with formic acid, which was increased to 100% in 1.5 min and maintained for another 1.5 min before returning to the initial 5% (flow rate was again 500 µL min -1 ).In addition, the autosampler temperature was set at 6˚C in this case.To assess analyte recoveries, two WB tissues from the 28 dpf-sampled control group were processed as above, but after homogenization spiked with a known amount of AMI and NOR and left for 30 minutes to allow for matrix binding.The samples were then processed as before.Analyte recoveries were found to be 90 ± 1% for AMI and 103 ± 7% for NOR.It should be noted that for some of the lower concentrations, tissues pooled as levels were too low for analysis on individual animals.

Supp Table 5. Results of the chemical analysis tissue samples from the 28-day exposure and 14 day depuration of zebrafish to AMI (2 s.f.).
Average tissue concentrations of amitriptyline and nortriptyline in the whole-body tissues, whole head tissues and head tissues with the eyes removed, of zebrafish exposed to amitriptyline, at a range of concentrations, for 28 days and then following a 14-day period of depuration in clean water up to 42 dpf (mean ± SEM, n = see table ).

Section 4: Assessment of apical endpoints
Supp Table 6.Effects of amitriptyline exposure on zebrafish apical endpoints.

Section 5: Brain Monoamine Analysis via HPLC
At 28 and 42dpf, a subset of fish from each of the control, 0.096nM and 960nM treatments (representing pharmacologically and environmentally relevant concentrations respectively) were terminated and their brains removed in E3 medium (see recipe below) under a dissection microscope.Samples were then snap frozen in Eppendorfs in liquid nitrogen and stored at -80°C until further processing.For analysis of monoamine levels, brains were weighed, 100 µl of 0.1 N perchloric acid was added, and then they were homogenised using a micro-pestle.Samples were then centrifuged at 4C, 12000 x g for 15 minutes, before the supernatant was collected.HPLC analysis of the selected neurotransmitter supernatants was carried out using the method of Carreno Gutierrez et al. ( 2018), on an ANTEC-HPLC system with electrochemical detection 1 .The mobile phase consisted of 75 mM sodium dihydrogen phosphate, 1 mM EDTA and 0.6 mM octane sulphonic acid (OSA) in deionised water containing 10% methanol.Samples were quantified based upon a one-point calibration using a 50 nM standard of each of the neurotransmitters being analysed.Neurotransmitters measured were: 5-HT, 5-hydroxyindoleacetic acid (5-HIAA, the main metabolite of 5-HT), norepinephrine (NE), dopamine (DA), 3,4-Dihydroxyphenylacetic acid (DOPAC, the major metabolite of DA) and homovanillic acid (HVA, the product of degraded DA).
The ratios of 5-HIAA/5-HT, DOPAC/DA and HVA/DA were calculated to assess monoamine turnover in the brain.exposure.For example, all three genes have been found to be differentially expressed in zebrafish embryo-larvae following an 80-hour exposure to the SSRI FLX 8 .

Primer selection and optimisation
Primer sequences specific for the target mRNA were taken from published studies 9,10 and purchased from Eurofins Genomics (Ebersberg, Germany).The optimal primer pair annealing temperature (Ta) for hrt2c was optimised by running a temperature gradient as Cunha et al. (2018), the source from which the target gene primer sequences was taken, did not explicitly state the optimum Ta 8 .The optimisation followed methods described previously 11 .PCR conditions included a denaturation step (heated to 95°C for 3 minutes), followed by 40 cycles of: denaturation at 95°C for 10 seconds, followed by annealing at each specific temperature in the gradient for 30 seconds (CFX96 Touch Real-Time PCR Detection System, Bio-rad).The highest temperature with the lowest CT value, and where replicates were most consistent, was selected as the optimal primer annealing temperature as this suggested the most efficient amplification.The Ta for the remaining three genes, slc6a4a, htr1aa and rpl8 had already been established in our lab 9 .

c). d).
(a) mortality, represented as a cumulative measure at 28 dpf, (b) hatching rate, which was recorded at 24-hour intervals until 120 hpf, (c) body weight and (d) body length (mean ± SEM, n = 5 for (a) and (b) and N = 38-60 for (c) and (d)).Differences in endpoints between drug-exposed fish and the controls were assessed using Kruskal-Wallis and Dunn's Multiple Comparison tests, where significant differences are represented by ** (P<0.01).

Section 7 :Supp Figure 5 .
Behaviour analysis Pattern of fish zone preference displayed as the % total distance travelled in the outer versus the inner zone at 28 dpf during the (a) light and (b) dark phase and at 42 dpf during the (c) light and (d) dark phase (mean ± SEM, N = 16).Here, the dotted line represents equal travel across both areas.Differences in zone preference were assessed using Kruskal-Wallis and Dunn's Multiple Comparison tests, where significant differences are represented by * (P<0.05),**(P<0.01),*** (P<0.001) and **** (P<0.0001)versus the control treatment.

Table 3 .
Measured concentrations of amitriptyline in exposure water during the saturation phase (mean ± SEM, n = 5).Note only the highest four concentrations of AMI in water were above the LOQ of the LC-MS/MS method used. S6Exposure

and depuration period concentrations Supp Table 4. Results of the chemical analysis of water samples from the 28-day exposure and 14 day depuration of zebrafish to AMI (2 s.f.). Measured
concentrations of amitriptyline in exposure water across the duration of the 42-day experiment expressed as the % of the nominal (mean ± SEM, n = 5).

Table 7 .
Neurochemical analysis of zebrafish whole brain tissues following a 28-day amitriptyline exposure and 14-day depuration period in clean water.Here, (a) the concentrations of single monoamines including 5-HT, 5-HIAA, DA, DOPAC, HVA and S16Supp