Developmental Toxicity in Zebrafish (Danio rerio) Exposed to Uranium: A Comparison with Lead, Cadmium, and Iron

Abstract

Populations of plants and animals, including humans, living in close proximity to abandoned uranium mine sites are vulnerable to uranium exposure through drainage into nearby waterways, soil accumulation, and blowing dust from surface soils. Little is known about how the environmental impact of uranium exposure alters the health of human populations in proximity to mine sites, so we used developmental zebrafish (Danio rerio) to investigate uranium toxicity. Fish are a sensitive target for modeling uranium toxicity, and previous studies report altered reproductive capacity, enhanced DNA damage, and gene expression changes in fish exposed to uranium. In our study, dechorionated zebrafish embryos were exposed to a concentration range of uranyl acetate (UA) from 0 to 3,000 μg/L for body burden measurements and developmental toxicity assessments. Uranium was taken up in a concentration-dependent manner by 48 and 120 hour post fertilization (hpf)-zebrafish without evidence of bioaccumulation. Exposure to UA was not associated with teratogenic outcomes or 24 hpf behavioral effects, but larvae at 120 hpf exhibited a significant hypoactive photomotor response associated with exposure to 3 μg/L UA which suggested potential neurotoxicity. To our knowledge, this is the first time that uranium has been associated with behavioral effects in an aquatic organism. These results were compared to potential metal co-contaminants using the same exposure paradigm. Similar to uranium exposure, lead, cadmium, and iron significantly altered neurobehavioral outcomes in 120-hpf zebrafish without inducing significant teratogenicity. Our study informs concerns about the potential impacts of developmental exposure to uranium on childhood neurobehavioral outcomes. This work also sets the stage for future, environmentally relevant metal mixture studies.

Graphical Abstract

Summary

Uranium exposure to developing zebrafish causes hypoactive larval swimming behavior similar to the effect of other commonly occurring metals in uranium mine sites. This is the first time that uranium exposure has been associated with altered neurobehavioral effects in any aquatic organism.

Introduction

Abandoned uranium mines are scattered throughout the world, located on every continent excluding Antarctica, including thousands of sites in the southwestern region of the United States, mainly on the Colorado Plateau (Moore-Nall 2015). Drainage from these mine sites can lead to uranium entering nearby waterways (Blake et al. 2017), contaminating groundwater and local soils (Abdelouas 2006), and being found in blowing dust from the mine surface soils (Zychowski et al. 2018). Despite governments setting strict limits on uranium in drinking water and food sources, many individuals living in rural areas, such as the southwestern region of the United States, rely on private, unregulated wells for drinking water, livestock water, and water to irrigate crops. In total, 12.5 % of the unregulated water sources sampled in the eastern region of the Navajo Nation were found to have uranium levels exceeding the United States Environmental Protection Agency (U.S. EPA) maximum contaminant level of 30 μg/L (Hoover et al. 2018). Groundwater sampled near abandoned uranium mines in the western region of the Navajo Nation range from 0.04 to 490 μg/L uranium (Credo et al. 2019), while surface waters from the Rio Paguate adjacent to the Jackpile uranium mine in Laguna Pueblo measured uranium levels up to 772 μg/L (Blake et al. 2017). The potential routes of exposure leave plants, animals, and humans in proximity at risk of exposure to elevated levels of the toxic metal (Waseem et al. 2015). ²³⁸U has been shown to accumulate in vegetables grown near abandoned mines sites, coinciding with elevated uranium in soils (Carvalho et al. 2014). Bioaccumulation of uranium in the food chain may leave higher consumers, particularly humans, more susceptible to the negative impacts of uranium exposure.

Little is known about the developmental health effects of uranium exposure. A 1981 study showed an association between maternal proximity to uranium mines, tailings or waste sites, and adverse birth outcomes in offspring (Shields et al. 1992). Another study determined that compared to adults, children were more sensitive to uranium exposure from contaminated well water (Magdo et al. 2007). Uranium exposure is associated with several neurobehavioral disorders, but majority of the studies conducted so far have been in adult rodent models (Dinocourt et al. 2015). Some studies have demonstrated that different species of fish exposed to contaminated waters from nearby mine sites exhibit altered hatching times (Pyle et al. 2002), increased oxidative stress (Bessa et al. 2016), altered metabolic activity (Bessa et al. 2016), increased micronucleus frequencies in erythrocytes (Annamalai et al. 2017), and decreased acetylcholinesterase, an important enzyme that regulates the neurotransmitter acetylcholine (Gagnaire et al. 2015). However, uranium-associated neurobehavioral data, specifically during development, is lacking.

Zebrafish (Danio rerio) is a well-established vertebrate model and a valuable tool for understanding developmental toxicity from xenobiotic exposures (Garcia et al. 2016; Nishimura et al. 2015; Teraoka et al. 2003). Zebrafish embryos are optically transparent; they develop rapidly and externally and thus support the rapid assessment of toxic effects on organ development and neurobehavior (Bugel et al. 2016; Geier et al. 2018a). The zebrafish genome is highly annotated and shares high genetic homology with humans; 76% of human protein-coding genes have a zebrafish counterpart, and 82% of human genes that cause disease are present in zebrafish, increasing the translational value of the zebrafish model (Howe et al. 2013). In addition, zebrafish brain anatomy is similar to mammals making zebrafish a popular model for central nervous system research as well as to understand various brain disorders, and the neurobehavioral impact of chemical exposure (Bailey et al. 2013; Stewart et al. 2014).

Several studies have investigated the effects of chronic uranium exposure in adult zebrafish. Uranium caused genotoxicity (Lerebours et al. 2013; Simon et al. 2018), modified reactive oxygen species production and phenoloxidase-like activity (Gagnaire et al. 2013), altered zebrafish metabolism (Augustine et al. 2015), negatively impacted the olfactory and lateral line systems (Faucher et al. 2012), and the reproductive system (Armant et al. 2017; Bourrachot et al. 2014; Simon et al. 2018; Simon et al. 2011). Other studies identified significant gene expression alterations and skeletal muscle malformations in the progeny of adult zebrafish exposed to 20 μg/L depleted uranium, suggesting that developmental zebrafish are extremely sensitive to uranium exposure (Armant et al. 2017; Gombeau et al. 2017). Waterborne exposure to 250 μg/L depleted uranium delayed hatching of zebrafish embryos, reduced prolarval body length, and increased mortality (Bourrachot et al. 2008). It was noted that uranium localized primarily in the chorion, which hindered uranium uptake at lower exposure concentrations and may have blocked more severe developmental effects (Bourrachot et al. 2008).

We compared the uranium toxicity data collected in this study with the zebrafish developmental toxicity of other metals: cadmium, iron, and lead using the same exposure paradigm. Similar to uranium, these metals are measured at detectable levels in unregulated water sources near abandoned uranium mine and mill sites with varying percentages above U.S. national drinking water standards which are currently set at 5 μg/L for cadmium, 300 μg/L for iron, and 15 μg/L for lead (U.S. EPA 2020a , U.S. EPA 2020b; Credo et al. 2019; Hoover et al. 2018). The metals enter the water primarily via the oxidation of metal sulfide minerals that are present in both uranium ore and in waste materials, and thus are likely to occur as co-contaminants with uranium (National Research Council 2011). Early-life exposures to lead and cadmium have previously been associated with neurological effects in children (Hou et al. 2013; Kippler et al. 2012). Additionally, animal models have been utilized to identify the sub-lethal effects of iron. Neonatal oral iron exposure causes long-term neurobehavioral effects in mice (Fredriksson et al. 1999). These studies demonstrate that exposure to environmentally relevant concentrations of lead, cadmium, and iron is capable of altering neurocognitive functioning. Despite these metals being detected together in the environment, to the best of our knowledge, their developmental toxicity has not been compared in the same experimental platform.

The present study investigated the effects of exposure of dechorionated zebrafish embryos to uranyl acetate starting at 6 hours post fertilization (hpf). Inductively-coupled plasma mass spectrometry (ICP-MS) revealed the embryo uranium uptake was concentration-dependent at 48 hpf and 120 hpf. We did not detect significant teratogenicity associated with exposure to uranium, but we detected significant behavioral effects in larval zebrafish. Our results will inform research, mitigation, and regulatory concerns about the potential impacts of prenatal, perinatal, and postnatal exposure to uranium on childhood developmental and neurobehavioral outcomes. In addition, the individual developmental toxicity effects of lead, cadmium, and iron were investigated in this study. Embryonic exposures to all three metals altered normal neurodevelopmental outcomes despite having no effect on teratogenicity. By comparing the toxicity of the four metals in the same model organism with the same exposure paradigm, we are providing a foundation for future research to investigate potential, environmentally relevant metal mixture effects.

Materials and Methods

Zebrafish husbandry

Tropical wild type 5D zebrafish (Danio rerio) were housed at the Sinnhuber Aquatic Research Laboratory (SARL) at Oregon State University (Corvallis, OR, USA). Larval, juvenile, and adult fish were fed GEMMA Micro 75, 150, and 500 (Skretting, Inc., Fontaine Les Vervins, France) respectively. Larval and juvenile fish were fed thrice daily while adult fish were fed twice daily (Barton et al. 2016). Adult fish were maintained on a 28°C recirculating water system with Instant Ocean salts in 50-gallon tanks at densities of 10 fish/gallon of water under a 14-hour:10-hour light:dark cycle. Fish were set up for group spawns the night before with spawning funnels placed in the tanks. After collection and cleaning in the morning, fertilized embryos were placed in a 28 °C incubator in Petri dishes with embryo media (EM) (Kimmel et al. 1995). EM comprised 15 mM NaCl, 0.5 mM KCl, 1 mM MgSO₄, 0.15 mM KH₂PO₄, 0.05 mM Na₂HPO₄, and 0.7 mM NaHCO₃.(Westerfield 2000). All zebrafish handling and use were conducted according to Institutional Animal Care and Use Committee protocols (ACUP 5113, date: 11th October, 2018).

Metal Exposures

The chorions of 4-hours post fertilization (hpf) zebrafish embryos were enzymatically removed with pronase (Sigma-Aldrich, catalog # 81750) using a custom automated dechorionator to eliminate a potential barrier for exposure to the metal (Mandrell et al. 2012). Additionally, metals such as iron have been shown to cause hardening of the zebrafish chorion which prevented normal hatching (Hassan et al. 2020); this could lead to potential effects on the developmental endpoints tested in this study. For the uranyl acetate (UA) exposures, dechorionated zebrafish (n = 32-68 per exposure concentration) at 6 hpf were placed one per well in round-bottom 96-well polystyrene plates (BD Falcon) prefilled with 100 μL exposure solution. The large sample size range is a result of experiments conducted on two different days; n = 68 is for the concentrations that overlapped between the two days, and for the control. UA was purchased from Electron Microscopy Sciences (USA) with a 99.9% U²³⁸ / 0.1% U²³⁵ composition and handled according to the regulations set forth by the Radiation Safety office at Oregon State University. In an aerobic environment, such as some soils and water, uranium compounds exist with a +VI valence (Garnier–Laplace et al. 2001; Waseem et al. 2015). UA was chosen because the uranyl ion has a +VI valence in aqueous medium, mimicking an environmental exposure. For exposures to lead (II) acetate trihydrate (Pb(CH₃CO₂)2 · 3H₂O, Sigma), cadmium chloride (CdCl₂, Sigma), and iron (III) chloride (FeCl₃, Sigma), we used an automated embryo placement system (Mandrell et al. 2012). Embryos were placed one per well (n = 36 per exposure concentration) in round bottom 96-well plates prefilled with 100 μL EM. A Hewlett Packard D300e chemical dispenser was used to dispense 100% dimethyl sulfoxide (DMSO) chemical stock solutions and final DMSO concentrations were normalized to 1% (V/V). The exposure concentrations of the metals tested in this study are listed in Table 1. After exposure, the plates were sealed with Parafilm to minimize evaporation of the exposure solutions, and shaken overnight at 28 °C in the dark on an orbital shaker (235 rpm) to enhance solution uniformity (Truong et al. 2016). Embryos were statically exposed in a 28 °C incubator for the remaining duration of the exposure (until 48 hpf for body burden assessments for UA, or until 120 hpf for body burden assessments for UA, and teratogenicity and behavior assessments for all metals). The pH of the UA exposure solutions were measured using the API pH Test Kit over the course of five days, and we confirmed that the UA did not alter pH of the exposure solution. pH was measured to be ~7.6 in both control and UA solutions (data not shown).

Table 1.

Chemicals tested with sampling time point and purpose (Abbreviations: EM – Embryo media, hpf – hours post fertilization, DMSO – dimethyl sulfoxide).

Metal	CAS number	Form of metal	Exposure concentrations (μg/L)	Control	Time point	Purpose
Uranyl Acetate	541-09-3	238U	0, 30, 300, 3,000	EM	48 and 120 hpf	Body burden
			0, 0.1, 3, 10, 30, 100, 300, 1,000, 3,000	EM	24 and 120 hpf	Teratogenicity and behavior assessments
Lead (II) acetate trihydrate	6080-56-4	Pb2+	0, 379, 759, 1,707, 3,414, 6,828, 12,897, 25,415	1% DMSO
Cadmium chloride	10108-64-2	Cd2+	0, 0.183, 1, 1.83, 5, 18.3, 183, 1830
Iron (III) chloride	7705-08-0	Fe3+	0, 162, 412, 1,046, 2,660, 5,977, 12,133, 16,220

Body burden of UA with ICP-MS

After UA exposure (0, 30, 300, 3,000 μg/L), embryos were collected for body burden at 48 and 120 hpf. The zebrafish were euthanized on ice for 30 minutes, which was confirmed by observing the lack of heartbeat under the microscope. At each time point, three replicates of 30 zebrafish each per concentration were collected in 15 mL plastic tubes. Embryos were washed five times with 1 mL ice-cold Chelex-treated water (Millipore Sigma, MA, USA). Each wash consisted of removal of liquid from the tube, addition of 1 mL ice-cold Chelex-treated water, and swirling of fish. After the last wash, all liquid was removed from the tubes before storage at −80 °C until further processing. Pooled embryos were digested in 0.5 mL full strength OmniTrace Nitric Acid (Millipore Sigma, MA, USA), overnight at room temperature and diluted with 4.5 mL Chelex-treated MilliQ water. Samples were diluted further with Chelex-treated MilliQ water for an overall dilution of 1:40. Samples were stored at −20 °C. ICP-MS analysis for uranium was performed by the Arizona Laboratory for Emerging Contaminants (ALEC) at the University of Arizona, Tucson, AZ. Samples were run on an Agilent 8900 ICP-QQQ following an analytical QA/QC protocol that was adapted from U.S. EPA Method 200.8 for ICP-MS analysis. Calibration standards for uranium were prepared from multi-element stock solutions resulting in a curve including at least seven points with a correlation coefficient of 0.9998. The detection limit for uranium was 0.7445 ng/L.

Behavior and Teratogenicity Assessments

Behavior Assessments

Embryonic Photomotor Response (EPR)

A custom-built photomotor response assay tool (PRAT) was utilized to measure the Embryonic Photomotor Response (EPR) (Noyes et al. 2015). After exposure to each metal, the zebrafish were kept in the dark until the evaluation of the EPR test at approximately 24 hpf. EPR consisted of 30 s of darkness (IR light, Background Interval); first 1 s pulse of intense white light; 9 s of darkness (Excitation Interval); second 1 s pulse of intense light; 10 s of darkness (Refractory Interval). To quantify total movement for each embryo across the period of the assay, pixel changes between video frames were recorded. Embryos were examined immediately after the EPR assay, and embryos with mortality were excluded from the EPR analysis (Truong et al. 2016). Statistical significance was assessed separately for each interval using the non-parametric Kolmogorov-Smirnov (KS) test (Bonferroni-corrected p value threshold = 0.01) against the vehicle control animals with the additional criterion that the mean area under the curve (AUC, Excitatory interval) estimate for each treatment group had to be at least 50% different in either direction from the control group AUC (Reif et al. 2016). Raw data collected from the EPR assay can be found in Table S1.

Larval Photomotor Response (LPR)

To determine if developmental exposure to UA, lead, cadmium, and iron altered larvae photomotor swimming behavior, the Larval Photomotor Response (LPR) was evaluated at 120 hpf using the ZebraBox system (ViewPoint Life Sciences, Lyon, France) (Knecht et al. 2017; Truong et al. 2012). The assay consisted of a total of three light-dark cycles, each cycle consisting of three minutes in visible light, and three minutes in the dark (IR light). Only the first light/dark cycle was included in the statistical analysis as it typically presents the greatest dynamic range. Larvae that exhibited any teratogenic effects at 120 hpf (described below in “Teratogenicity Assessments”) were excluded from the LPR analysis. Statistical significance was quantified using a KS test (p<0.05) on pairwise comparisons of mean treatment area under the curve (AUC) relative to the control group AUC, with the additional criterion that the relative AUC ratio calculated as (Treatment-Control/Control) (Relative Ratio) (Zhang et al. 2017) had to be at least 10% (−10% for hypoactivity and 10% for hyperactivity). To allow for acclimatization, we analyzed only the third light-dark cycle to determine significant effects. Raw data collected from the LPR assay for uranium, lead, cadmium, and iron can be found in Tables S2, S3, S4 and S5, respectively.

Teratogenicity Assessments

At 24 hpf, zebrafish embryos were systematically screened for mortality, developmental progression, notochord formation, and spontaneous motion. Upon completion of the LPR assay at 120 hpf, larvae were euthanized with tricaine methanesulfonate (MS-222) and visually assessed for 18 developmental endpoints: mortality, yolk sac edema, pericardial edema, body axis, trunk length, caudal fin, pectoral fin, pigmentation, somite, eye, snout, jaw, otolith, brain, notochord and circulatory malformations, swim bladder presence and inflation, and touch response. A binary presence or absence response was recorded for each zebrafish for each teratogenicity endpoint, with “presence” indicating any deviation from normal development for each endpoint. The data were entered into a laboratory information management system (LIMS), and statistically analyzed as described previously (Truong et al. 2014). Briefly, a lowest effect level (LEL) was computed for each endpoint measured as the concentration at which the incidence of the endpoint exceeded a significance threshold over the background rate, estimated using a Fisher’s exact (binomial) test.

Results and Discussion

Uranium uptake

To quantify uranium uptake, we performed ICP-MS analysis at 48 and 120 hpf on dechorionated embryos exposed to 0, 30, 300, and 3,000 μg/L UA from 6 hpf. Fig. 1 reveals a concentration-dependent accumulation of uranium in the exposed zebrafish. We note that the uptake concentration from exposure to 300 μg/L UA was <0.05 μg uranium/g wet weight at both 48 and 120 hpf in our chorion-absent study. This was lower than the uptake concentration in a chorion-intact study (Bourrachot et al. 2008) where >2 μg uranium/g dry weight in embryos, and >10 mg uranium/g dry weight in the chorion, was detected at 48 hpf after exposure to 250 μg/L uranyl nitrate (UN) (Bourrachot et al. 2008). The study showed that uranium primarily adsorbed to the chorion independent of uranium’s isotopic nature (Bourrachot et al. 2008). The chorion, with an approximate pore size of 0.6 – 0.7 μm, is an acellular envelope surrounding embryos until they hatch and acts as a chemical barrier irrespective of chemical size or hydrophilicity (Bonsignorio et al. 1996; Braunbeck et al. 2005; Lee et al. 2007; Ozoh 1980). The presence of the chorion could generate false-negative results in such studies (Henn et al. 2011) and thus we sought to quantify zebrafish uranium uptake in the absence of the chorion.

Fig. 1. Uranium uptake by developing zebrafish.

Concentrations of uranium detected by ICP-MS analyses in 48 hpf and 120 hpf zebrafish exposed from 6 hpf to 30, 300, or 3,000 μg/L UA. As exposure concentration increased, the concentration detected in zebrafish also increased. The UA did not bioaccumulate between 48 hpf and 120 hpf. A two-way ANOVA followed by a Sidak’s post-hoc multiple comparison test was used to compare samples. Graph represents mean ± SEM of at least 3 independent replicates (ns = not significant, **p ≤ 0.01).

Our results demonstrate the total body burden of UA was lower relative to previous zebrafish studies with UN (Bourrachot et al. 2014; Bourrachot et al. 2008). While chemical differences between UA and UN could be contributing to uptake discrepancy, the difference in the pH of the exposure solutions between our study and previous studies could also be playing a role. While the exposure solution in our study was maintained at pH = 7.6 throughout the duration of the experiment, the pH of the exposure solutions in Bourracho et al. was lowered to 6.5 ± 0.2 to maximize the bioavailability of uranium (Bourrachot et al. 2014; Bourrachot et al. 2008). It has been shown that the aqueous speciation of uranium drastically changes within the pH range of 4 to 8, which can significantly alter uranium uptake into organisms (Fournier et al. 2004; Simon et al. 2014). In the current study, we maintained the pH of the exposure solution at 7.6 for two reasons: 1). To avoid the potential confounding effects of lowered pH to developing zebrafish. While zebrafish are well-known for being an acid-tolerant species (Kwong et al. 2014), one study demonstrated that development is altered, and hatching and survival rates of embryonic zebrafish are significantly decreased at pH ≤ 5 and ≥ 10 (Andrade et al. 2017). Additionally, the impact of pH on the sensitive neurobehavior endpoints evaluated in this study is unknown, and thus, we maintained pH of the exposure solution at control levels. 2). To maintain the environmental relevance of our study. The pH of groundwater from a uranium-contaminated well on the Navajo Nation was found to be 7.36 (Austin et al. 2012). Additionally, contaminated surface waters from the Rio Paguate adjacent to the Jackpile Uranium Mine measured at circumneutral pH levels ranging from 6.8 and 8.6 (Blake et al. 2017). Thus, the pH utilized in our study is within the range of the pH of the water detected near uranium mine sites. While pH may be playing a significant role in the uptake discrepancy between our study and previous studies, further research investigating uranium uptake specifically into developing zebrafish over a pH range is necessary to explain the observed differences. Here, we utilize the same exposure conditions on dechorionated zebrafish for our uptake experiment as well as for the neurobehavioral assays in order to directly correlate nominal water exposure concentration, uptake concentration, and developmental neurobehavioral toxicity. In our uptake study, the uranium concentration per embryo remained constant between the 48 and 120 hpf time points demonstrating that the uranium did not bioaccumulate. Our result is in contrast to previous studies where adult zebrafish accumulated uranium or a mixture of ²³³U-depleted uranium over several days of waterborne exposure (Barillet et al. 2007; Simon et al. 2018). This is not surprising, as the primary uptake route for metals in fish is via the gills (Bucher et al. 2016; Bywater et al. 1991), and zebrafish do not have fully functional gills until 14 days post fertilization (Rombough 2002). The difference in life stages between studies, and their concomitant physiological and behavioral traits, may explain the lack of bioaccumulation in our developing zebrafish. Overall, our results suggested that 1) dechorionated zebrafish embryos were able to take up uranium from a waterborne exposure, 2) uptake was rapid and detectable by 48 hpf at exposure concentrations ≥ 300 μg/L UA, and 3) uranium did not bioaccumulate in zebrafish between 48 and 120 hpf.

Behavior Assessments

Embryonic Photomotor Response (EPR)

Following uptake estimation, dechorionated embryos were exposed to 0, 0.1, 3, 10, 30, 100, 300, 1,000, and 3,000 μg/L UA from 6 to 120 hpf to assess developmental toxicity. To assess if subteratogenic UA might have caused behavioral changes, we examined the embryonic photomotor response (EPR) at 24 hpf and the larval photomotor response (LPR) at 120 hpf. None of the UA exposure concentrations used in our study were associated with an EPR profile that was statistically different from the control embryos in any of the three intervals of the assay (Fig. 2). Similar to the control embryos, the UA-exposed zebrafish responded to the first light flash (Excitation Interval) and failed to respond to the second flash (Refractory Interval). It is possible that either insufficient uranium was taken up into the 24-hpf zebrafish embryos to cause behavioral effects, or that uranium exposure does not alter 24-hpf behavior of zebrafish. Since we measured uranium uptake only at 48 hpf and 120 hpf, we cannot confirm which of these hypotheses are true. Future work quantifying uranium uptake at earlier developmental time points will help explain this result.

Fig. 2. UA exposure did not affect embryonic photomotor response (EPR).

Dechorionated zebrafish embryos (n=32-68) were exposed to 0, 0.1, 3, 10, 30, 100, 300, 1,000, and 3,000 μg/L UA from 6 hpf, and the EPR assay was conducted at approximately 24 hpf. Normal zebrafish embryos exhibit baseline activity in IR light prior to the first light flash (Background), an immediate burst of tail bending in response to the first light flash (Excitation) that rapidly decays to immobility, and little or no movement in response to the second light flash (Refractory). The K-S test (p < 0.05, delta > 50%) was used to determine that there were no alterations from normal behavior before and after the two light pulses in UA-exposed zebrafish. Any dead or malformed animals were excluded from the analysis.

The zebrafish EPR assay has been previously utilized to investigate the toxicity of a wide variety of compounds including neuroactive drugs and chemicals (Carbaugh et al. 2020; Geier et al. 2018b; Kokel et al. 2011). Additionally, the developmental neurotoxicity of environmentally relevant concentrations of metals has also been evaluated (Olivares et al. 2016). While gallium exposure does not alter 24-hpf photomotor activity, indium and arsenic exposures are associated with hypoactivity relative to controls in the Excitation Interval of the assay. We highlight that the EPR assay is highly predictive of teratogenic outcomes later in life (Reif et al. 2016) indicating that the absence of an altered EPR response from uranium exposure in our study suggests the lack of long-term detrimental effects. Of note, however, is that this assay has a low sensitivity for detecting chemicals known to be neuroactive in mammals (Carbaugh et al. 2020). Other studies have found conflicting results between the EPR assay and other neurobehavioral assays for a specific chemical, demonstrating the utility of multi-endpoint developmental assessments in identifying potential neurotoxicity (Geier et al. 2018a). Thus, by presenting the negative results for this assay in our study, we emphasize the importance of evaluating more than one behavior endpoint to truly capture the potential of a chemical to alter neurobehavioral functioning (Audira et al. 2018).

Larval Photomotor Response (LPR)

In addition to EPR, we conducted a second photomotor behavior assay at 120 hpf to evaluate UA-induced alterations in larval swimming activity in response to rapid light-dark transitions. A normal LPR consists of low total movement in the light phase and 5 – 10 fold more movement in the dark phase. Any alterations to the normal LPR upon UA exposure could indicate impacts to sensory or neuromuscular development. Fig. 3A depicts results from the three alternating light-dark phases of the assay, and reveals that there is a clear sustained hypoactivity from UA exposure across all cycles despite the decreasing maximum activity of the control zebrafish. Embryos exposed to 10, 30, 100, and 3,000 μg/L UA exhibited statistically significant decrease (hypoactive) in swimming (p < 0.01; Relative Ratio ≤ −10%) in the light phase relative to control fish, while embryos exposed to 1,000 μg/L UA were hyperactive in the light phase (p < 0.01; Relative Ratio ≥ 10%). The 0.1, 3, and 300 μg/L UA concentrations displayed a trend toward hypoactivity, but they were not statistically significant (Fig. 3B). Zebrafish exposed to 3, 30, and 100 μg/L UA were significantly hypoactive in the dark phase of the assay (p < 0.01; Relative Ratio ≤ −10%). Zebrafish exposed to all other concentrations displayed a trend toward hypoactivity. Despite low uptake of uranium at the 30, 300, and 3,000 μg/L UA exposure concentrations (0.003, 0.020, and 0.460 μg uranium/g wet weight, respectively at 120 hpf) (Fig. 1), we still detected a behavioral impact at 120 hpf at all three concentrations, which underscores the value of including subteratogenic endpoints in toxicity evaluations.

Fig. 3. UA exposure altered larval photomotor response (LPR).

Developmental exposure of zebrafish (n=32-68) to UA was associated with hypoactivity in the light (quiescent) and dark (active) periods. (A) LPR assay consisted of three 3-minute periods of alternating light and dark periods. Dechorionated zebrafish embryos exposed to 10, 30, 100, 300, and 3,000 μg/L UA had a significant decrease (p < 0.05, Relative Ratio ≤ −10%) in photo-motor activity in the light phase, and embryos exposed to 3, 30, and 100 μg/L UA had a significant decrease in the dark phase (p < 0.05, Relative Ratio ≥ 10%). (B) Table shows area under the curve (AUC), AUC Relative Ratio (%), p value (calculated by the K-S test), and the detected significant effect, if any, calculated during the third light-dark cycle of the assay.

The UA-induced altered LPR profile not only shows that some aspect of zebrafish sensory or neurodevelopment has been misregulated at 120 hpf, but is indicative of potential long-term cognitive effects as seen in recent studies conducted by our group (Garcia et al. 2018; Knecht et al. 2017). Previous uranium work in other vertebrate models are also consistent with this hypothesis. While several studies have found alterations to both locomotion and neurocognitive behaviors in adult rodents from uranium exposure (Dinocourt et al. 2015), only a handful of studies have focused on the effects of developmental uranium exposure. Rodents exposed to uranium during development displayed altered locomotor activity (Lestaevel et al. 2016), a deficit in spatial working memory (Dinocourt et al. 2017; Houpert et al. 2007), and delayed hyperactivity that lasted up to 9 months of age (Houpert et al. 2007). One study also found that prenatal uranium exposure caused a depressive-like behavior phenotype in postnatal rats (Legrand et al. 2016). The results of these studies not only demonstrate that uranium can target the central nervous system, but also emphasize the long-term detrimental impacts of developmental uranium exposure.

The altered LPR response diagnosed in our study is the first time a behavioral outcome has been identified in developing zebrafish exposed to uranium. While the larval swimming behavior disruption may be due to altered sensory system development, the potential for the uranium exposure to cause misregulation of neurogenesis is not unlikely. Uranium-induced impairment of nervous system development has been hypothesized in previous rodent studies (Dinocourt et al. 2017), however more research is needed to gain an in-depth mechanistic understanding of uranium neurotoxicity. As a first step, we have identified an aberrant behavioral response in the high-throughput zebrafish model which also has a high physiological and genetic similarity to humans. Future work conducting gene expression analyses in developing zebrafish exposed to uranium will give us insight into the mechanisms of uranium toxicity. Further, assessment of multiple behavioral endpoints at later life stages of zebrafish will determine whether long-term neurocognitive health effects exist from exposure to uranium during development.

Comparison of UA with other metals

Similar to UA, we also exposed developing zebrafish to lead, cadmium, and iron, and assessed their impact on development and swimming behavior at 120 hpf. All three of these metals have been detected at elevated levels in water sources near uranium mine sites. Exposure to a broad concentration range of lead or iron resulted in a hypoactive larval photomotor response at 120 hpf in both the light and dark phases of the assay (Fig. 4A). Exposure to cadmium resulted in a hypoactive LPR in the dark phase of the assay and did not induce an altered LPR in the light. The lowest effect level (LEL) to cause the adverse effect was 759 μg/L for lead (dark phase), 0.183 μg/L cadmium (dark phase), and 1,046 μg/L for iron (light phase). We note that the LPR results for all three metal exposures are similar to the LPR hypoactivity associated with UA exposure (LEL = 3 μg/L). At all the exposure concentrations for UA, lead, cadmium, and iron, we did not detect significant teratogenicity (Fig. 4B). Fewer than five animals per concentration (<16%) of UA, lead, and iron had any teratogenic effects, while cadmium exposure was associated with a higher rate of mortality (up to ~23% per concentration; not statistically significant) relative to other metals. These results not only highlight zebrafish as a sensitive model for exposure assessment, but also emphasize the importance of conducting both teratogenic and subteratogenic studies when determining the toxicological impact of chemical exposure.

Fig. 4. Comparison of developmental toxicity of UA with lead, cadmium, and iron.

(A) Developmental exposure of zebrafish to lead and iron was associated with hypoactivity in the light (quiescent) and dark (active) phases, and exposure to cadmium resulted in hypoactivity in the light phase. The lowest effect levels (LEL) in the dark and light phases, and the lowest sample sizes of all concentrations for each chemical are indicated in the table. (B) Developmental exposures to uranyl acetate, lead, cadmium, and iron did not cause teratogenic effects till 120 hpf. “Any effect” includes all 18 developmental endpoints described in the Teratogenicity Assessments section of the Methods (Abbreviations: MO24 = mortality at 24 hpf, MORT = mortality at 120 hpf).

To the best of our knowledge, mixed metal exposures using the metals tested in this study have not been investigated in zebrafish (or other aquatic) behavioral models. To prepare for environmentally relevant mixed metal exposures, we first needed to compare the toxicity effects of each metal alone in the same biological model system. Thus, we evaluated the morphological and behavioral impact of the exposure to uranium, lead, cadmium, and iron in this study. While an increasing number of studies are evaluating developmental toxicity of individual metals in zebrafish (Roy et al. 2015; Zhu et al. 2012), there is only limited research making direct comparisons between them. Unlike the lack of teratogenicity that we observed from exposure to cadmium and lead in our study, a previous study reported that exposures to 25 μmol/L cadmium (4,583 μg/L) and 2.5 μmol/L lead (948.32 μg/L) were associated with both morphological malformations and behavioral effects in developing zebrafish (Tu et al. 2017). The exposure concentration difference for cadmium between our study and Tu et al. likely accounts for toxicity differences. Additionally, while we conducted a static exposure over the entire course of the exposure period, Tu et al. renewed the exposure solutions every 24 hours. This difference in experimental design likely accounts for the lack of significant teratogenicity from our lead exposures. The disparity reiterates the importance of toxicity comparisons using the same experimental platform. Indeed, we expect to observe varying results in behavior depending on which metal mixtures are investigated and at what exposure concentrations. While metal mixture toxicity studies are scarce in the literature, some individual metals have been shown to interfere with the cellular uptake and bioaccumulation of other metals when fish are exposed to metal mixtures. When juvenile Prochilodus lineatus (streaked prochilod) were exposed to a mixture consisting of zinc, manganese, and iron, the tissue distribution of the metals was altered compared to each metal alone (de Oliveira et al. 2018). Futhermore, binary combinations of copper with zinc and copper with cadmium have been shown to cause synergistic effects on Gobiocypris rarus (Chinese rare minnow) larvae mortality (Zhu et al. 2011). Any altered tissue distribution or mortality caused by individual metals when presented as a mixture, could also have synergistic behavioral consequences. We plan to follow up our study by exposing zebrafish embryos to mixtures of the individual metals presented in this manuscript.

The metals tested in this study are found in unregulated water sources in the southwestern U.S. in the same region as abandoned uranium mine sites, possibly acting as co-contaminants. Several metals such as calcium, manganese, magnesium, nickel, uranium, zinc, and iron have been detected at concerning levels in both groundwater, and flora and fauna near uranium mine sites in Portugal (Pereira et al. 2014). Additionally, Wang et al. reported elevated concentrations of uranium in combination with thorium, iron, zinc, lithium, and cobalt in surface waters near uranium mine and milling sites in Northern Guangdong Province, China (Wang et al. 2012), demonstrating that the co-occurrence of metals is not unique to just the U.S. Along with uranium, these trace metals have the ability to be taken up by crops grown around uranium mining areas potentially leading to exposure of humans and subsequent health effects from the elevated levels of metals (Neves et al. 2012). In unregulated water sources in the Navajo Nation of northeastern Arizona and southeastern Utah, an area with a long history of uranium mining, cadmium was detected at levels of up to 11 μg/L whereas the U.S. EPA guideline is 5 μg/L (Credo et al. 2019). The 75^th percentile of iron concentrations was 2,100 μg/L, with 38.6% of the samples testing over the National Drinking Water Standard of 300 μg/L, and the 75^th percentile of lead concentrations was 6.3 μg/L, with 2.6% testing over the Standard of 15 μg/L (Hoover et al. 2018). The concentration range utilized to test lead, cadmium, and iron in developing zebrafish in our study included environmentally relevant concentrations, and we detected significant behavioral toxicity at these concentrations suggesting the potential for adverse health impacts from human exposure.

Conclusions

Uranium toxicity is an important issue for animals and humans living in close proximity to abandoned uranium mine sites. The current study demonstrates both the sensitivity and utility of developmental zebrafish to model human toxicity by our ability to detect subteratogenic impacts of depleted uranium on neurobehavioral outcomes. The photomotor behavior effect that we identified would be missed with a more traditional toxicology approach, which would have read 3,000 μg/L UA, and its corresponding low level of uptake, as putatively benign. Traditionally, the nervous system has not been considered a primary target of uranium toxicity, but our results should inform further research on the potential neurobehavioral effects of exposure to uranium. Logical next steps will be to 1) assess these same animals for gene expression changes that might be anchored to the LPR phenotype, forming a basis for mechanistic understanding of how uranium exerts developmental toxicity, and 2) determine whether any cognitive and social behavioral deficits from developmental exposure are manifest in adulthood.

Rarely do single metals exist as sole contaminants near mine land use sites, but they are present as mixtures of metals. This study reveals the adverse neurobehavioral impacts of uranium and its potential co-contaminants lead, cadmium, and iron, on developing zebrafish without causing significant teratogenicity. Organisms including humans that live around uranium mine sites are at a high risk of environmental exposure to both uranium and trace metals, and very little is known about the adverse health effects associated with exposure. Our results provide a benchmark for conducting developmental zebrafish toxicity studies with metal mixtures, and reveal a benefit of using a sensitive animal to model the potential human health effects of exposure to environmentally relevant metals.

Highlights.

1. Uranyl acetate (UA) does not cause morphology malformations or mortality in developing zebrafish

2. 3 μg/L UA exposure alters 120-hour post fertilization (hpf) zebrafish swimming behavior

3. First study to detect behavioral malformations in developing zebrafish exposed to UA

4. Lead, cadmium, and iron impacted larval zebrafish behavior with same exposure paradigm as UA

5. Our study sets the stage for future metal mixture studies using zebrafish