Zebrafish.

Wildtype AB zebrafish embryos were obtained from the zebrafish core facility at Karolinska Institutet. Breeding groups of adult fish (3 males and 2 females) were housed together overnight in 10 L spawning tanks containing environmental enrichment (commercially available aquaria decoration made of non-toxic plastic). Thirty minutes after turning the light on (9:00 a.m.), the fertilized eggs were collected and stored at 28.5°C until further processing. The eggs were washed twice with fresh E3 water (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.33 mM MgSO₄; 0.05% methylene blue; pH 7.4) in order to remove all debris. This procedure was performed on a clean open lab bench, at room temperature, under constant illumination (approximately 500 lx). Stock PFOS was further diluted in DMSO to yield a final concentration of 0.1% DMSO in the rearing water. The zebrafish embryos were exposed to PFOS concentrations of either 0.1, or 1 mg/L (0.186 or 1.858 µM, respectively). The doses were selected based on our experience and earlier reports [30]–[32] to avoid a significant increase in the rate of embryonal malformations and lethality. Control embryos were exposed to 0.1% DMSO in E3 water. The exposure to PFOS was initiated about 2 hours post fertilization (hpf). The zebrafish larvae were then plated and maintained individually in 48-well plates (cylindrical wells, 10 mm inner diameter; VWR, Leuven, Belgium) in 750 µl E3 water at 28°C in a 10:14 h dark-light cycle (light on at 9 a.m.; 300 lx intensity, daylight-matching spectrum white light) until behavioral testing at 6 days post fertilization (dpf). The exposure followed a static, non-replacement regime (i.e. the rearing water was not refreshed or replaced throughout the experiment, according to OECD guidelines [33], [34] and earlier reports [30], [32], [35]). All treatments were present in each plate, and the larvae were distributed such that an equal proportion from each group would be placed in wells at the periphery of the plate. The mortality, successful hatching, and the occurrence of embryonal malformations as assessed at the 24 (developmental failure/coagulation, light-induced coiling movements), 48 (developmental failure/coagulation, blood circulation, pericardial oedema, pigmentation), 72, and 120 hpf (hatching, eye development, swimming bladder inflation, morphological abnormalities such as scoliosis or bent spine), as well as after completing the behavioural experiments (6 dpf) using an inverted microscope (Nikon Ti-S) in brightfield configuration with a 4× objective. The larvae displaying morphological abnormalities at 6 dpf were excluded from analyses.

Determination of PFOS in zebrafish larvae by UPLC-MS/MS. The zebrafish larvae were killed by rapid cooling in ice-cold water immediately after the behavioural recording at 6 dpf. The larvae were then collected and stored in Eppendorf tubes with a minimal amount of rearing water to preserve the integrity of the organism until further processing. All larvae from an experiment were pooled per treatment to yield a sufficient amount of tissue (22±14 mg per treatment group from 6 independent experiments). The internal dosimetry was performed by liquid chromatography coupled with mass spectrometry using an Acquity UPLC system (Waters, Milford, MA, USA) coupled with a triple quadrupole Waters TQD mass spectrometer. The samples were homogenized with acetonitrile (100 µL x 10 mg) using an ultrasonic homogenizer (Omni-Ruptor 250, Omni International). An internal standard (Perfluorononanoic acid, PFNA) was added to 100 µL of supernatant and then the samples (2 µL) were diluted 1:40 with 2 mM ammonium acetate - methanol solution (50∶50, v/v) before injection in UPLC. A calibration curve was prepared by spiking PFOS and PFNA standard to a pool of homogenized control larvae in the range from 100 to 1600 absolute ng of PFOS. Chromatographic separation was performed on an Acquity UPLC BEH C18 column (2.1×100 mm, 1.7 µm) maintained at 30°C, by gradient elution with a mixture containing variable proportions of 2 mM ammonium acetate buffer and methanol delivered at a flow rate of 0.5 mL/min. Under these conditions isomers of PFOS eluted at 1.6, 1.9, 2.1 and 2.5 minutes whereas PFNA eluted at 2.4 minutes. The total run time was 9 min. For the MS/MS detection, quantitative analysis was obtained in multiple reaction monitoring (MRM) in negative ion mode (ESI-). In particular for all the isomers of PFOS 499.0→79.8 m/z was used for quantification and 499.0 → 98.8 for confirmation; for PFNA 463.0→418.9 was selected. The limit of detection for PFOS (at a signal-to-noise ratio of 3) was 0.2 ng.

Behavioral tests and analyses in zebrafish larvae. We recorded the locomotor activity of 6 dpf zebrafish larvae at 50 frames per second (fps) under constant infrared illumination using integrated high throughput systems for videotracking and environmental control (ZebraBox™, ViewPoint, Lyon, France, and DanioVision, Noldus, Wageningen, The Netherlands). The larvae were detected by automated online image analysis using a dynamic subtraction algorithm with individual frame weight of 4%. All videotracking data was visually inspected for detection accuracy before inclusion in further analyses. The raw tracking data were exported and analyzed using routines custom developed in Matlab™ R2012b environment (MathWorks Inc., Natick, MA, USA). To minimize the effect of missing samples (e.g. when the fish moved too fast over few frames, or too little to meet the automated detection criteria), the position of center of gravity was approximated using linear interpolation of x and y coordinates between the samples at the beginning and at the end of the missing samples sequence. For the analysis of bouts, we used a threshold of 0.2 mm/frame to filter out detection noise. A bout of activity was defined as the activity detected in a continuous sequence of frames during which the measured displacement of the center of gravity is above the threshold (see insert in Fig 1 A). Bouts separated by less than 40 ms (2 frames) were considered as a single bout. The following parameters were calculated for measuring the spontaneous activity bouts: frequency of occurrence, expressed in number of bouts per time unit (bouts/min); and average distance moved within the bout (mm).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. The structure of spontaneous locomotion in zebrafish larvae and mice during the active phase of the circadian cycle (light for zebrafish larvae, dark for mice).

(A) Representative sequence of spontaneous activity (20 s) recorded at 6 dpf in one control and one larva exposed to 1 mg/L PFOS. In the control larva, the spontaneous activity consists of short, evenly spread bouts of swimming. In the larva exposed to 1 mg/L PFOS, the spontaneous activity consists of clusters of intense activity separated by extended periods of inactivity. Background noise (displacement below 0.2 mm/frame) shaded in gray; inset in top panel – magnification of a representative bout of spontaneous activity. (B) Quantification of frequency of spontaneous bouts of activity. Note the dramatic decrease in bout frequency induced by exposure to 1 mg/L PFOS. (C) Average distance moved during one spontaneous bout of activity. Zebrafish larvae exposed to 1 mg/L PFOS have a hyperactive phenotype characterized by a 2.5 fold higher distance swum during spontaneous bouts of activity. (D) Illustrative sequence of spontaneous locomotion during the active phase of the circadian cycle in one control and one mouse exposed to 3 mg/kg/dy PFOS during gestation. The spontaneous locomotor activity (visits) in the homecage is integrated over consecutive, non-overlapping 10 min bins and spline-interpolated for clarity. (E, F) Similar to the pattern found in zebrafish larvae, the mice exposed to 3 mg/kg PFOS display less frequent (E), but more intense (F) bouts of activity. B, C, E, F – one-way ANOVA followed by Dunnett's post-hoc test; * p<0.05 PFOS exposed vs. control. The number of independent observations is indicated at the bottom of each column in B, C, E and F.

https://doi.org/10.1371/journal.pone.0094227.g001

Spontaneous activity. Before initiating the behavioural experiments, the temperature control system of the recording chamber (DanioVision™, Noldus) was allowed to run continuously until equilibrium was reached, and the temperature did not vary by more than 0.2°C over 5 min. The experimental workflow ensured that (1) the handling of the multi-well plate was limited to less than 30 s in total; (2) utmost care was taken to limit the vibrations transferred to the plate during handling; (3) the larvae did not experience major fluctuations in light intensity, particularly not towards sudden decrease (the light sources inside the recording chamber were on when the multi-well plate was placed inside, and the intensity of the lighting did not change except when testing the VMR). The larvae were allowed to acclimatize allowed to acclimatize for at least 15 min after transfer to the recording chamber. This procedure was found sufficient (pilot tests; not shown) for the spontaneous activity to be reliably recorded and analysed over reasonably long periods free from trends or trailing effects of changes in the environment. We estimated the average frequency of occurrence, and distance moved within spontaneous bouts over at least 3 min after the acclimatization period and prior to any further experimental challenge.

The visual motor response (VMR) is one of the earliest behavioral responses that can be recorded in freely moving zebrafish larvae (already at 4 dpf), and consists of a robust increase in swimming activity in response to a sudden decrease in the environmental light intensity [36]–[38]. Yet, at 4 dpf, the energy required for growth and swimming is provided by the protein-rich yolk, and although the larvae display a robust locomotor response to changes in light intensity, the spontaneous activity varies widely between individuals and does not support reliable estimations of baseline activity [39]. Therefore, the analysis of activity induced by dark pulses was performed at 6 dpf as follows: the larvae were placed in the DanioVision™ chamber and allowed to acclimatize for at least 15 min before starting the experiment (environmental temperature maintained at 28°C by means of a custom-built system for circulating temperature-controlled water; light intensity 3200–3500 lx). The baseline activity under constant illumination was recorded during the last 10 min of the acclimatization period. The larvae were then exposed to a dark pulse (10 min; light not detectable in the visible spectrum). The transition between light and dark was virtually instantaneous.

To characterize the hyperactivity induced by the dark pulse, we used two orthogonal parameters derived from the cumulative activity curve: the total locomotor activity and the index of curvature (IOC). IOC was calculated according to the formula below, as described by Fry et al. [40]: , where R_i denotes the value of the i-th bin in the cumulative histogram. The IOC can assume values between -(n-1)/n and (n-1)/n (approaching -1 and 1, respectively, for very large n), where n is the number of observation bins (time intervals). If the activity is recorded predominantly in the beginning of the observation interval (e.g. if the animals habituate to the environment), IOC is negative, and the absolute value increases with the bias of distribution in activity measures. In contrast, an IOC value close to 0 indicates a uniform distribution of activity within the observation period. The rationale for calculating a parameter to describe the shape of the curve is that (1) the novelty-induced hyperactivity is a consistent finding in all animal models; (2) the locomotor activity necessarily decays with time as the animals habituate to the environment [41], [42]; and (3) the shape of the habituation curve conveys ethologically relevant information, such as the rate of habituation or acquisition of reinforcement [40]. As compared to the measures of habituation described earlier [42], [43], the IOC is a longitudinal measure of the activity and has the advantage of orthogonality in relation to global measures of activity.

The startle response consists of a brief increase in swimming activity (typically a single bout that is more intense than spontaneous bouts) shortly after the application of a stimulus (vibrational, acoustic, electric, or tactile), followed by a period of inactivity before resuming spontaneous activity [44] (Movie S1). The acoustic/vibrational startle response in particular develops in parallel with the auditory system and consistent behavioral responses can be recorded no sooner than 5 dpf in normal zebrafish larvae [45], [46]. We have used a custom-built system consisting of a solenoid hitting the chassis of the recording chamber. The stimulus was triggered from the videotracking software (EthoVision XT 9, Noldus, Wageningen, The Netherlands) and was synchronized with the behavioral observation. The fish were allowed to habituate to the experimental environment for 4 minutes before delivering the first stimulus. The baseline activity was recorded during the last minute of the habituation period. A series of 10 equally spaced taps (1 min intervals) were delivered during each session. A startle response was validated only if we recorded a bout of activity within 0.5 s (25 frames) after triggering the stimulus. After validation, we analyzed the following parameters: average distance moved per time unit (1 s); latency to startle (measured as the delay between the application of the stimulus and the beginning of the first subsequent bout); distance moved within the bout; inactive period following the startle response (measured as the delay between the end of the startle response and the beginning of the first next bout) (see also Figure S1). We found no evidence of habituation (decreasing number of responsive larvae, or decreasing intensity of the startle response) in our system (data not shown), and estimated group means after averaging the values for each parameter across the validated startle responses for each larva.

Pharmacological challenges in zebrafish larvae. Apomorphine (Apoteket, Sweden), dexamfetamine (S(+)-amfetamine sulfate, Cat. No. A-127; RBI, Natick, MA, USA), quinpirole (Tocris Bioscience, Bristol, UK), and SKF-81297 (Tocris Bioscience, Bristol, UK) were dissolved in 0.1% DMSO E3water immediately before adding to the rearing water. The pharmacological compounds were delivered in 20 µL E3 medium/well 30 min prior to recording the swimming activity. The final concentrations in the rearing water were as follows: dexamfetamine – 1 and 10 µM; apomorphine – 2.5 and 5 µM; quinpirole – 10 and 20 µM; SKF-81297 – 25 and 50 µM. The doses and the timing of behavioural recording relative to the administration of the selected compounds were set based on previous reports (for dexamfetamine, apomorphine, and quinpirole see [47], [48]) and dose-response curves tested in our lab (for SKF-81297). In particular the doses of dexamfetamine were selected so that the low dose would induce hyperactivity, and the high dose would suppress spontaneous activity in control larvae [48]. All plates were treated similarly in each experiment, i.e. all larvae in one plate received the same concentration of the compound, and the different concentrations were divided between plates originating from the same stock and spawning session. All experiments were repeated at least 3 times, and used fertilized eggs produced from different stocks. This design was carefully observed in order to minimize the bias of genetic background differences across stocks.

Historical behavioral data in mice.

In light of the phenotypical alterations found in zebrafish, as well as to strengthen the relevance of zebrafish larvae as experimental model for developmental neurotoxicity, we further analyzed historical behavioral data from mice. The developmental exposure to PFOS in mice has been described earlier [19]. Briefly, pregnant C57Bl/6/Bkl dams (Scanbur BK, Sweden) received PFOS at a dose of 0.3 or 3 mg/kg/day throughout gestation (N = 6 per treatment). The locomotor activity was recorded in 8 male pups (5–6 weeks of age), randomly selected from all litters (maximum 2 pups from any litter were included). At weaning, the pups were implanted with sterile subcutaneous radio frequency identification (RFID) tags (ID-100A, Trovan Ltd., UK) under 4% isoflurane anesthesia. The transponders allowed unambiguous identification of animals throughout the experiment.

Monitoring and analysis of spontaneous locomotion in freely moving mice. The locomotor activity of freely moving mice was monitored using the TraffiCage™ system (NewBehavior, Zürich, Switzerland), as described elsewhere [19]. Briefly, the system consists of an array of radio frequency antennas to be placed under the cage with group-housed, freely moving mice. The antennas would read the RFID tags and provide an approximate location of each animal with a time resolution of 50 ms. The series of RFID detection were exported and analyzed using custom routines developed in Matlab™ R2012b. The time interval during which an animal is detected constantly by the same antenna is defined as “visit”, and is described by time of first detection of a specific RFID and by its duration. When an animal is changing its location to a large enough extent to be detected by another antenna, a new visit is recorded. Visits that are longer than 10 min are considered inactive periods, consistent with resting/sleeping time (see also [19]). The number of visits was used as estimate of locomotor activity. The spontaneous locomotion is typically characterized by clusters of relatively short visits bounded by inactive periods, which we defined as “activity bouts”. A bout of activity is therefore described by the moment of appearance (for classification of occurrence in relation to the phases of the circadian cycle), the intensity of activity it comprises (number of visits); and the duration (defined as the time interval between the end of an inactive period, and the beginning of the next inactive period). Given that the spontaneous activity analyzed during the dark phase is part of a virtually infinite sequence, the number of bouts is equal to the number of inactive periods. In order to match the zebrafish data on spontaneous activity, we analyzed only the dark phase of the circadian cycle (when mice are normally active). In addition, we analyzed the novelty-induced hyperactivity in mice. Thus, we monitored the locomotor activity for 2 h after transferring the animals to a new cage with fresh bedding (otherwise identical to the home cage). The handling and the sudden change in the environment typically induces an abrupt increase in exploration/locomotor activity, which subsides within 2–3 h in group-housed animals (see also [19]). We assessed the habituation to a new environment by estimating the total locomotor activity and the IOC.