Fish husbandry, spawning, and all experiments were performed in accordance with the University of California, Davis Institutional Animal Care and Use Committee (IACUC) protocols 17645 and 19391. Wildtype zebrafish (Tropical 5D strain) were originally obtained from the Sinnhuber Aquatic Research Laboratory at Oregon State University with subsequent generations raised at UC Davis. Zebrafish were maintained under standard laboratory conditions with 14 h light (∼850 lux)/10 h dark cycles (Harper and Lawrence 2011). Water was maintained at 28.5 ± 0.5°C, pH 7.2 ± 0.4 and conductivity at 700 ± 100 µS using bicarbonate and Instant Ocean Sea Salt (Blacksburg, VA, United States), respectively. Adult fish were fed twice daily with a combination of live Artemia nauplii (INVE Aquaculture, Inc., Salt Lake City, UT, United States) and commercial flake food comprised of Zeigler Zebrafish Granule (Zeigler Bros, Inc., Gardners, PA, United States), Spirulina Flake (Zeigler Bros, Inc.), Cyclopeeze (Argent Aquaculture, Redmond, WA, United States), and Golden Pearl (Brine Shrimp Direct, Ogden, UT, United States). Embryos were obtained by natural group spawning in system fish water (FW) and kept in an incubator at 28.5°C.

PCB congeners 28, 66, 95, and 153 were purchased from AccuStandard (Accustandard Inc., New Haven, CT, United States), while PCBs 84 and 138 were synthesized at the University of Iowa as described previously (Sethi et al., 2019). The purity of PCBs was determined on an Agilent 6890 chromatograph coupled with an Agilent 5975 Inert Mass Selective Detector (Agilent Technologies, Santa Clara, CA, United States) operated in electron ionization mode using a capillary Supelco SLB-5MS GC column (30 m × 250 μm × 0.25 µm, Sigma-Aldrich). Purity was determined in the total ion chromatogram (TIC) mode using the following temperature program (Telu et al., 2010; Holland et al., 2017a; Li et al., 2018): starting temperature 50°C, 10°C/min to 150°C, then 5°C/min to 280°C, hold for 6 min, and 10°C/min to 300°C, hold for 5 min. Helium flow rate was 2 ml/min. The inlet and auxiliary temperatures were both 280°C. The purity was calculated based on the relative peak area. A congener specific analysis method in the select ion monitoring (SIM) mode was employed to identify PCB impurities present in the PCB samples (Hu et al., 2015; Holland et al., 2017a). This method allowed the analysis of 209 PCB congeners as 162 peaks of individual or co-eluting peaks. The temperature program was modified based on the published method: hold at 80°C for 1 min, 2°C/min to 160°C, 1°C/min to 170°C and hold for 15 min, 1°C/min to 180°C and hold for 15 min, 1°C/min to 245°C, then 10°C/min to 300°C and hold for 15 min. The helium flow rate was 1.5 ml/min. The injector temperature was 280°C. The MS temperatures were 280°C, 230°C, and 150°C for transfer line, source and quadrupole, respectively. The GC analysis showed that all six PCB congeners had a purity >99.9%, and that no other PCB congeners were detected. The gas chromatograms and the mass spectra for all PCB congeners are shown in Supplementary Figures S1–S12. PCB congeners were dissolved in 100% DMSO (Sigma-Aldrich, St. Louis, MO, United States) and stored in amber glass vials as 1000X stocks. On exposure days, stocks were freshly diluted at 2X in embryo media (EM) (Westerfield and ZFIN, 2000).

Sensitization of the zebrafish RYR by ortho-PCBs was measured using [³H]ryanodine ([³H]Ry) ligand binding assays as previously described (Fritsch and Pessah, 2013). Briefly, crude microsomal protein preparations were created from the tail (post dorsal fin) muscle (primarily white skeletal muscle) of six adult zebrafish (∼1 year old) that was homogenized in a buffer containing 320 mM sucrose, 5 mM HEPES, 1 mM PMSF, 10 mM NaF, 2 mM β-Glycerol, 5 mM Na₄P₂O₇, 0.5 mM Na₃VO₄ and 2 μg/L leupeptin (pH 7.2). The homogenate was centrifuged at 5,000 × g (4°C) for 15 min. The pellet was re-homogenized in 5 ml homogenization buffer and centrifuged at 5,000 × g. Supernatants were combined for ultracentrifugation at 100,000 g (4°C for 1 h). The resultant pellet was re-suspended in a buffer of 320 mM sucrose and 5 mM HEPES (Sigma-Aldrich, St. Louis, MO, United States) at pH 7.2. Protein concentrations were determined in triplicate using a BCA Assay (Fisher Scientific, Waltham, MA, United States). [³H]Ry ligand binding assays were conducted with 40 μg/ml of the crude protein homogenate in the presence of DMSO (0.5%) or 0.1–50 μM of individual PCB congeners in 0.5% DMSO. Tests were run using 5 nM [³H]Ry (Perkin Elmer Waltham, MA, United States), 250 mM KCl, 20 mM HEPES, 15 mM NaCl, and 50 μM CaCl₂ (pH 7.1) and incubated for 16 h at 25°C. Non-specific binding was determined with the addition of 10 µM ryanodine (Tocris Bioscience, Bristol, United Kingdom) and 200 μM EGTA (Sigma-Aldrich, St. Louis, MO, United States). Assays were terminated by rapid filtration through Whatman GF/B filter paper (Brandel, Gaithersburg, MD, United States) and washed with ice-cold buffer containing 140 mM KCl, 10 mM HEPES, and 0.1 mM CaCl₂, pH 7.4. Bound radioligand was measured by liquid scintillation counting (Beckman Coulter, Indianapolis, IN, United States). PCB-induced [³H]Ry binding was assessed in two different protein preparations, and each assay was run in triplicate. Individual PCB concentration-response curves were determined using non-linear regression (GraphPad Prism 6.0, GraphPad Software Inc., San Diego, CA, United States).

At 4 hpf, embryos were treated with 50 µl of 41 mg/ml protease from Streptomyces griseus (P5147, Sigma-Aldrich, St. Louis, MO, United States) in 25 ml of FW for a maximum of 6 min to remove the chorion (Truong et al., 2011). At 5 hpf, embryos were placed in individual wells of 96-well plates filled with 100 µl of EM. At 6 hpf, embryos were exposed to a 2× concentration of the selected PCB congeners or controls that were diluted in 100 μl EM to produce a 1× concentration in a total of 200 µl solution. Exposure concentrations were 0 (DMSO control), 0.1, 0.3, 1, 3, and 10 μM. After exposure, plates were covered with Parafilm M (Bemis NA, Neenah, WI, United States) to reduce evaporation and placed in an incubator at 28.5°C with 14 h light (∼300 lux)/10 h dark cycles. Final DMSO concentrations were 0.2% (vol/vol) for all exposures. Fish were statically exposed until 5 dpf.

Exposed fish were examined daily for any gross morphological malformations and mortality using an Olympus Stereo Microscope Model SZ61 (Olympus, Japan) up to ×4.5 magnification. On days of behavioral testing (3, 4, and 5 dpf), examination was conducted immediately after the conclusion of the behavioral assay. Teratological endpoints that were evaluated include yolk sac edema, pericardial edema, somite formation, deformities of the body axis, notochord, craniofacial structures, or caudal fin. Malformations were scored in a binary manner of either present or not present. Fish that were scored as neither malformed nor dead were deemed viable. For each exposure, the number of fish that were viable, malformed, or dead was divided by the total number of biological replicates within each group and multiplied by 100 to obtain percent incidence.

Larval locomotor tests were performed at 3, 4, and 5 dpf. Vehicle and PCB-exposed larvae in 96-well plates were placed in a DanioVision behavior system (Noldus Information Technology, Leesburg, VA, United States) and subjected to a 35 min light/dark locomotor test. System and plate temperature were maintained at 28.5 ± 0.5°C using a temperature control unit from Noldus. The testing paradigm consisted of a 10 min light period (∼1,900 lux) to allow for acclimation (5 min) and to record baseline swimming (5 min). This was followed by a 5 min dark period (∼0 lux) to stimulate increased swimming behavior and a 5 min light period (∼1,900 lux) to stimulate freezing behavior. The last 15 min consisted of a dark period (∼0 lux) to observe increased swimming behavior and acclimation to the dark conditions.

The distance traveled binned per minute by each larva during the 35-min observation test period was recorded via continuous tracing using a DanioVision system camera with an infrared filter to allow recording in both light and dark conditions. Movement was tracked using EthoVisionXT software (Noldus Information Technology) and exported to Excel and Prism 6.0 (GraphPad Software) for analysis and visualization. Dead and/or malformed larvae (as noted by observation under dissecting scope) were excluded from the behavior analysis.

The trapezoidal rule was used to compute the area under the curve (AUC) for each fish per day per lighting condition. The first 5 min of the 35-min observation period was removed to account for initial acclimation. Therefore, data were analyzed for the following time epochs: Light 1 (L1) = 6–10 min, Dark 1 (D1) = 11–15 min, Light 2 (L2) = 16–20 min, and Dark 2 (D2) = 21–35 min. Mixed effects regression models, including zebrafish-specific random effects, were used to assess differences between groups. DMSO groups from each spawn date (≥3 independent spawns) and plate were combined in the analysis. Separate models were fit for each congener. New variables were defined to capture differences in the AUC between Light 1 and Dark 1, Dark 1 and Light 2, and Light 2 and Dark 2, to measure change in distance traveled when transitioning from light to dark or dark to light. Contrasts were specified to compare these transitions and the AUC during each of the lighting conditions between the exposed groups and the DMSO control group. Exploratory analysis indicated that a natural logarithmic transformation was needed for the AUC to stabilize the variance and meet the underlying assumptions of the mixed effects models. Due to some zeroes in the AUC, all values were shifted by 0.1 prior to taking the natural logarithm. Concentration (0.1, 0.3, 1, 3, 10 µM and DMSO), day (3, 4, 5 dpf), and the transition variables were all of interest in the models, including their interactions. Akaike information criterion was used for model selection and Wald tests for comparing groups. All analyses were conducted using SAS version 9.4.

PCB body burden was assessed at 5 dpf by pressurized liquid extraction Dionex ASE200 system (Thermo Scientific, Sunnyvale, CA, United States) (Wu et al., 2015). Briefly, zebrafish larvae samples (n = 36–51 per sample) were mixed with diatomaceous earth (2 g) and placed in an extraction cell (33 ml) containing Florisil (60–100 mesh, 12 g, Fisher Scientific, Fair Lawn, NJ, United States). PCB 117 (100 ng) was added to each sample as surrogate recovery standard, and the cells were extracted with hexane-acetone (1:1, v/v) at 100°C and 1,500 psi (10 MPa) with preheat equilibration for 6 min, 35% of cell flush volume and 1 static cycle of 5 min. Method blanks containing only Florisil and diatomaceous earth and sample blanks with naïve (unexposed) and/or DMSO-exposed zebrafish larvae were extracted in parallel with each sample set. The extracts were concentrated to approximately 0.75 ml using a Turbo Vap II (Biotage, Charlotte, NC, United States) and transferred to glass tubes. Samples were further cleaned up, as described previously, to remove sulfur and other impurities (Kania-Korwel et al., 2007). All samples were spiked with PCB 204 (100 ng) as internal standard (volume corrector) prior to gas chromatographic analysis.

Analysis of PCBs was performed in the splitless mode on an Agilent 7890 gas chromatograph equipped with a ⁶³Ni micro electron capture detector (µECD) and a SPB-1 capillary column (length, 60 m; inner diameter, 250 μm; and 0.25 µm film thickness; Supelco, St. Louis, MO, United States) (Wu et al., 2013). The following temperature program was used based on a published method (Kania-Korwel et al., 2011): hold at 80°C for 1 min, 20°C/min to 215°C, then 0.1°C/min to 219.5°C, and 20°C/min to 280°C, then hold at 280°C for 3 min. A constant helium flow rate of 1 ml/min was used for all analyses. The injector and detector temperatures were 250°C and 300°C, respectively. The ⁶³Ni-μECD used for the PCB analysis was linear up to concentrations of 1,000 ng/ml for all analytes investigated (R ² > 0.999). The recovery of PCB 117 was 83 ± 7% (range: 71–97%). A detailed summary of the limits of detection, limits of quantification, and background levels of the analyzed PCBs is presented in Supplementary Table S1.

The enantioselective analysis of PCB 84 was performed on a 6890 Agilent gas chromatograph equipped with a ⁶³Ni-µECD detector and CP-Chiralsil-Dex CB capillary column with cyclodextrin directly bonded to dimethylpolysiloxane (25 m length, 0.25 mm inner diameter, 0.25 µm film thickness; Varian, Palo Alto, CA, United States). The temperature program was as follows (Kania-Korwel et al., 2006): starting temperature 50°C, hold for 1 min, 10°C/min to 140°C for 100 min, then 10°C/min to 200°C for 5 min. The enantioselective analysis of PCB 95 was performed on a 7890 Agilent gas chromatograph equipped with a ⁶³Ni-µECD detector and Chiraldex B-DM (BDM) capillary column (length, 30 m × inner diameter, 250 μm × 0.12 µm film thickness, Supelco; St. Louis, MO, United States). The following temperature program was used (Uwimana et al., 2016): starting temperature 50°C for 1 min, 10°C/min to 160°C for 90 min, then 10°C/min to 200°C for 10 min. For both instruments, the injector was operated in the splitless mode at 250°C, and the detector temperature was 250°C. A constant helium flow rate of 3 ml/min was used for all enantioselective analyses. Enantiomer fractions (EF) were calculated with the drop valley method (Asher et al., 2009; Uwimana et al., 2016) as EF = Area E₁/(Area E₁ + Area E₂), were Area E₁ and Area E₂ denote the peak area of the first and second eluting atropisomer, respectively. The resolution of two atropisomers of PCB 84 and PCB 95 were 0.79 and 1.2, respectively. The EF values were 0.5 for the racemic PCB 84 and PCB 95 standards.

The neurotoxic potential of the congeners tested in this study were ranked based on the three endpoints measured in this study: RYR binding, behavior, and body burden. Congeners were ranked relative to each other for each individual endpoint from 1 (highest) to 6 (lowest), and then the three scores for each congener averaged to obtain an overall relative ranking for each congener. Ranking for RYR activation was based on EC_2X values and their relative comparisons. Behavior ranking represents the total change from control for each congener across all epochs (L1, D1, L2, and D2) and all days (3, 4, and 5 dpf) tested. AUC values were normalized to the appropriate DMSO control (AUC value/DMSO). Each value was then subtracted from 1 to determine the deviation from the control values [1 − (AUC value/DMSO)]. Absolute values were combined for all concentrations of each congener across all time points (epochs and testing days). PCB body burden was ranked as the percentage uptake using the following calculation: (uptake/total PCB) × 100, where, the uptake values were the measured PCB levels in the samples, and the total PCB values were calculated using the exposure concentration (0.1, 1, or 10 μM) and volume (200 μl). The resulting values were averaged across the three assayed concentrations and rounded to the closest integer to obtain the final ranking order.

Adult zebrafish tail muscle homogenates were used to test RYR binding of PCBs 28, 66, 84, 95, 138, and 153 (Figure 1A). [³H]Ry binds with high affinity to the open RYR channel, where increased binding signifies increased channel activity in the presence of PCBs. PCB 28 and 66 displayed limited RYR activity, where even concentrations above 10 µM failed to cause a 2-fold (EC_2X) increase in channel activity. The environmentally prevalent PCB 138 and 153 presented moderate RYR activity (Figure 1 and Table 1). Interestingly, PCB 84 exhibited a biphasic effect (Supplementary Figure S13) in which concentrations below 3 µM caused a significant reduction in RYR activity. Conversely, concentrations above 3 µM caused modest activation with a maximal effect of 400% compared to the DMSO control. Similar to that seen in mammalian species, PCB 95 was the most efficacious at sensitizing RYR, causing an approximate 800% increase in [³H]Ry-binding, a measure of channel activation. Relative RYR activity is summarized in Figure 1B. As predicted, PCBs with less than two chlorines in the ortho-position displayed weak or negligible RYR activity. In contrast, congeners with meta substitutions adjacent to ortho-chlorines (PCB 84, 95, and 138) displayed higher activity at the channel. PCB 95 and 84, which lack para substitutions, were the most potent RYR sensitizers.

Zebrafish were statically exposed to PCBs starting at 6 hpf through 5 dpf and observed daily. Each of the six PCB congeners was tested at the same five concentrations ranging from 0.1 to 10 µM. Within this concentration range, there were no significant changes in zebrafish viability or teratogenic effects (yolk sac or pericardial edema, malformation of caudal fin, pectoral fins, eyes, craniofacial structures, axis, or notochord) of PCB-exposed larvae when compared to the corresponding vehicle control at 1, 3, 4, and 5 dpf (Figure 2). However, exposure to 30 µM PCB 95 resulted in nearly 100% loss of viability by 3 dpf (data not shown). The other PCB congeners were not tested at 30 µM due to solubility constraints.

Vehicle [0.2% (v/v) DMSO] control larvae exhibited increased locomotion in the dark epochs (D1 and D2) and reduced locomotion in the light epochs (L1 and L2), consistent with normal larval photomotor behavior (Basnet et al., 2019). An overall developmental effect was observed in which locomotion in both light and dark epochs increased at 4 dpf compared to 3 dpf and increased at 5 dpf compared to either 4 dpf or 3 dpf (3 dpf < 4 dpf < 5 dpf) (Table 2). To quantify this behavior further, we analyzed light epoch transitions, which are defined as the difference in distance moved from L1-D1, D1-L2, and L2-D2. As expected, in vehicle control larvae there was a significant increase in the distance moved with the L1-D1 (p < 0.001) and L2-D2 (p < 0.001) transitions and a significant decrease in distance moved with the D1-L2 transition (p < 0.001). Furthermore, the transitions between light epochs (L1-D1, D1-L2 and L2-D2) decreased with increased larval age (3 dpf > 4 dpf > 5 dpf) (Figure 3).

Exposure to PCB 66 had the least impact on behavior in the photomotor assay. Exposure to 0.11 µM (p = 0.03) or 1 µM (p = 0.001) PCB 66 significantly decreased distance traveled in L1 relative to age-matched vehicle control larvae, on average, across larval ages (p < 0.03; Figure 3 and Table 2), and the difference did no significantly vary across age. The other four PCB congeners had a greater impact on photomotor behavior with significant effects observed at multiple concentrations with respect to both distance moved and transitions between light epochs. The results for distance traveled are summarized in Table 2 and the transition findings are summarized in Supplementary Figures S14–S19.

At 3 dpf, larvae exposed to 3 and 10 µM PCB 28 traveled significantly less distance in D2 than vehicle control larvae (p = 0.049 and p < 0.001, respectively). At 4 dpf, all PCB 28-exposed larvae, except for those exposed to 10 μM, translocated significantly more than the vehicle controls in D1 (0.1 µM: p = 0.02; 0.3 µM: p = 0.005; 1 µM: p = 0.03; 3 µM: p = 0.03), with those exposed to 0.3 µM traveling less in L1 than vehicle controls (p = 0.01). The L1-D1 transition also significantly increased relative to vehicle controls [0.1 µM moving 0.9 mm more (p = 0.009), 0.3 µM moving 1.3 mm more (p < 0.001), 1 µM moving 0.87 mm more (p = 0.009), 3 µM moving 0.9 (p = 0.009) and 10 µM moving 0.8 more mm (p = 0.02)] (Supplementary Figure S15). At 5 dpf, larvae exposed to 10 µM moved less than vehicle controls in L2 (p = 0.048).

Similarly, at 3 dpf, larvae exposed to 1 or 10 µM PCB 153 moved more in D1, D2, and L2 than vehicle controls (p < 0.03); the L1-D1 transition was larger in the larvae exposed to 1 µM or 10 µM PCB 153 than vehicle controls (p < 0.025). However, at 4 dpf, larvae exposed to 0.3 µM PCB 53 moved more than vehicle controls in L1 (p = 0.02), attenuating the L1-D1 transition (p = 0.046), while larvae exposed to 10 µM PCB 153 moved more than vehicle controls in L2 (p < 0.01). At 5 dpf, larvae exposed to 10 µM PCB 153 moved more in L1 than vehicle controls (p = 0.02), which resulted in larvae in an attenuated L1-D1 transition compared to controls (p = 0.06, Supplementary Figure S16); this group also exhibited more movement in L2 than controls (p = 0.002).

Increased distanced moved in L1 and L2 compared to controls was observed in 3 dpf larvae exposed to PCB 84 at 10 µM (p < 0.02), while 3 dpf larvae exposed to 3 µM PCB 84 moved more in L2 (p = 0.04) than controls. PCB 138 and PCB 84-exposed larvae also exhibited significantly increased distance moved in D1 and D2 at 3 dpf. Specifically, for PCB 138, the 1 µM (p = 0.03), 3 µM (p = 0.002) and 10 µM (p < 0.001) exposure groups moved more than the vehicle controls in D1 while the 3 µM (p < 0.001) and 10 µM (p < 0.001) exposure groups moved more than vehicle controls in D2. For PCB 84, the 10 µM exposure group moved more (p < 0.001) in D1 and D2 than the age-matched controls. However, there were also differing effects of these two congeners on photomotor behavior. For PCB 84, the 0.1 µM exposure group moved less in D1, L2 and D2 at 3 dpf (D1: p = 0.01; L2: p = 0.003; D2: p < 0.001). PCB 84 exposed larvae also had altered movement at 4 dpf, with the 0.3 and 1 µM exposure groups exhibiting significantly less movement in L1 (p = 0.004 and p = 0.002, respectively). At 5 dpf, the 0.3, 3, and 10 µM exposure groups moved more in L1 (p = 0.047, p < 0.001 and p = 0.02, respectively). In contrast, 5 dpf larvae exposed to PCB 138 at 10 µM moved less in D2, than the vehicle control fish (p = 0.004).

The transition results followed a pattern similar that observed for distance moved amongst larvae exposed to PCB 138 or 84. For PCB 138, at 4 dpf, the L1-D1 transition was increased in the 0.3 µM group (0.821 more; p = 0.01) and the 1 µM group (0.68 more; p = 0.04) and decreased in the 10 µM group though not quite statistically significant (0.6 less; p = 0.06) compared to the control group (Supplementary Figure S17). At 5 dpf, the 3 µM group moved less during the L1-D1 transition compared to controls (0.7 less; p = 0.03). For PCB 84 at 4 dpf, the 0.3 and 1 µM groups moved more in the L1-D1 transition (p = 0.007 and p = 0.005, respectively), while the 10 µM group moved less during the transition (p = 0.04) (Supplementary Figure S18). At 5 dpf, the PCB 84 3 and 10 µM groups increased less than the DMSO group in the L1-D1 transition (p < 0.001 and p = 0.004, respectively).

Larvae exposed to 10 µM PCB 95 moved significantly more in D1 and D2 at 3 dpf (p < 0.001 for both) and significantly less at 4 dpf (in D1; p = 0.004) and 5 dpf (in both D1 and D2; p < 0.001) than vehicle controls (Figure 3). During light conditions, larvae exposed to 10 µM PCB 95 moved significantly more in L1 and L2 at 3 dpf (p < 0.001) and 4 dpf (p < 0.001) and more in L1 at 5 pdf (p < 0.001) than age-matched controls. Larvae exposed to 3 µM moved more in L2 and D2 at 3 dpf (p < 0.02) and more in L1 at 5 dpf (p < 0.001) than controls. As shown in Supplementary Figure S19, the 3 µM PCB 95 exposure group moved significantly less than DMSO in the L1-D1 transition at 5 dpf (p < 0.001), while in the 10 µM PCB 95 exposure group, the L1-D1 transition was decreased on 4 and 5 dpf (p < 0.001). In contrast, the larvae exposed to 0.1 µM PCB 95 moved more during the L1-D1 transition at 4 dpf (p = 0.04) than vehicle controls.

PCB body burden was measured in duplicate for three exposure concentrations (0.1, 1, and 10 µM) of each congener. Exposure began at 6 hpf and tissues were harvested at 5 dpf. For all five PCB congeners, larval body burden was found to increase with increasing PCB concentration (Figure 4). Uptake of PCB 153 was the lowest followed by PCB 28 and PCB 66. PCBs 84 and 95 had comparable body burden, while PCB 138 exposed larvae possessed the highest body burden of the tested PCBs.

Enantioselective GC analysis of PCB 84 and PCB 95 was used to assess possible enantioselective uptake. No significant atropisomeric enrichment of either PCB congener at the concentrations investigated was found, with average EF values of 0.50 ± 0.01 (n = 6) and 0.50 ± 0.01 (n = 14) for PCB 84 and PCB 95, respectively.

Rankings were assigned to each congener based on the average results of ryanodine binding, photomotor behavior, and PCB body burden. The relative ranking for RYR sensitization (Figure 1) was 66 < 28 < 153 < 138 < 84 < 95. Behavior scores differed noticeably between PCB 95 and the five other congeners: PCB 28 = 10.83 < PCB 66 = 14.39 < PCB 153 = 23.05 < PCB 84 = 28.93 < PCB 138 = 29.32 < PCB 95 = 81.46. To determine the relative ranking for PCB body burden, the % uptake values (measured uptake/theoretical uptake × 100) were averaged across the three concentrations to which zebrafish were exposed: PCB 153 = 1.4% < PCB 28 = 4.5% < PCB 66 = 5.6% < PCB 95 = 7.4% < PCB 84 = 9.2% < PCB 138 = 13.8%. The absolute relative rankings for each endpoint (from 1 to 6) were averaged to determine the relative ranking of the overall congener neurotoxic potency in this zebrafish model: PCB 28 < PCB 66 < PCB 153 < PCB 84 < PCB 138 < PCB 95 (Figure 5). Interestingly, the overall ranking of PCB congeners for potential neurotoxic outcomes follows a similar pattern of their potency in sensitizing the RYR.