Abstract

   Arsenic trioxide (As[2]O[3]) is a ubiquitous heavy metal in the
   environment. Exposure to this toxin at low concentrations is
   unremarkable in developing organisms. Nevertheless, understanding the
   underlying mechanism of its long-term adverse effects remains a
   challenge. In this study, embryos were initially exposed to As[2]O[3]
   from gastrulation to hatching under semi-static conditions. Results
   showed dose-dependent increased mortality, with exposure to 30–40 µM
   As[2]O[3] significantly reducing tail-coiling and heart rate at early
   larval stages. Surviving larvae after 30 µM As[2]O[3] exposure showed
   deficits in motor behavior without impairment of anxiety-like responses
   at 6 dpf and a slight impairment in color preference behavior at 11
   dpf, which was later evident in adulthood. As[2]O[3] also altered
   locomotor function, with a loss of directional and color preference in
   adult zebrafish, which correlated with changes in transcriptional
   regulation of adsl, shank3a, and tsc1b genes. During these processes,
   As[2]O[3] mainly induced metabolic changes in lipids, particularly
   arachidonic acid, docosahexaenoic acid, prostaglandin, and
   sphinganine-1-phosphate in the post-hatching period of zebrafish.
   Overall, this study provides new insight into the potential mechanism
   of arsenic toxicity leading to long-term learning impairment in
   zebrafish and may benefit future risk assessments of other
   environmental toxins of concern.

   Keywords: arsenic toxicity, zebrafish, long-term learning impairment,
   behaviors, metabolomics

1. Introduction

   Among the various organs affected by bioaccumulation of heavy metals,
   brain damage is of particular concern due to its high susceptibility to
   environmental chemicals [[42]1]. Exposure to heavy metals during
   neurodevelopment is believed to cause more types of neurodevelopmental
   disorders (NDDs) than in adulthood. However, the complexity of
   etiological pathways makes elucidation of this mechanism challenging
   [[43]2]. The effects of environmentally relevant concentrations of
   heavy metals on the nervous system are usually slow-onset,
   irreversible, and often overlooked due to non-teratogenic effects.
   Children with NDDs have difficulties with sensory and motor function,
   communication, learning, and memory. NDDs include intellectual
   disabilities (ID), autism spectrum disorders (ASDs),
   attention-deficit/hyperactivity disorder (ADHD), and motor development
   disorders [[44]3]. Impaired cognitive function interferes with normal
   human functioning and is often associated with neuropsychiatric
   disorders to varying degrees [[45]4,[46]5]. The global increase in NDDs
   [[47]6,[48]7] highlights the need for multidisciplinary efforts to
   understand the effects of low-concentration heavy metal exposure to
   fill data gaps in developmental neurotoxicity (DNT).

   Arsenic (As) is a toxic, ubiquitous metalloid found in food, water, the
   environment, and various tissues of the human body [[49]8] and has
   become a global health concern. The double-edged potential of As became
   a topic of discussion when the US Food and Drug Administration (FDA)
   approved arsenic trioxide (As[2]O[3]) for the treatment of acute
   promyelocytic leukemia (APL) in 1996 [[50]9]. This biologically active
   form of As further pollutes the environment via patient excretions, as
   well as improper waste manufacturing and disposal [[51]10]. Impacts
   during early life development are determined by the dynamics of As
   transit through the placental barrier, which regulates important
   aspects of embryonic development [[52]11]. Exposure during pregnancy
   has been associated with neurological impairments in the prenatal,
   perinatal, and childhood periods [[53]12]. Although it is not yet clear
   whether ingestion of drinking water contaminated with As at low
   concentrations affects children’s brains, epidemiological studies have
   indicated that cognitive deficits associated with As exposure
   [[54]13,[55]14] may have irreversible cumulative adverse effects years
   later [[56]15]. A recent global As risk assessment predicted that
   85–90% of people particularly who live in south Asia, are potentially
   exposed to high levels of As in groundwater from their domestic water
   supply [[57]16], even at low concentrations [[58]17]; thus, it is
   critical to understand the mechanism of action of As toxicity. The
   changing landscape, ponds, and extensive use of pesticides on palm oil
   plantations in the Langat Basin, Malaysia, are considered to be the
   major sources of increased arsenic concentration in the Langat River
   (0.98–21.94 µg/L), which exceed the Malaysian Ministry of Health (MOH)
   and World Health Organization (WHO) permissible limit for arsenic of 10
   µg/L [[59]18].

   To reduce costs and shorten the duration of basic toxicity DNT studies,
   zebrafish (Danio rerio) have attracted considerable interest due to
   their practical benefits, including high fecundity, small size, short
   embryonic period, high permeability to small molecules, and
   transparency, which facilitate their anatomical characterization
   [[60]19]. The zebrafish is universally used by biologists to study
   brain development [[61]20], as it has key pathways relevant to human
   disease [[62]21] with conserved gene expression for neurodevelopment
   [[63]22] and brain homology [[64]23]. Zebrafish also exhibit behavioral
   responses similar to those of rodents when exposed to toxicants
   [[65]24], along with a behavioral repertoire that can be automatically
   quantified as a valuable indicator of altered brain function.

   In this study, zebrafish were used as a model organism to understand
   the developmental effects of embryonic exposure to As[2]O[3]. To assess
   the long-term toxic effects of As[2]O[3] at the functional level, motor
   activity, anxiety, and learning behavior were measured at different
   life stages. With the aim of deciphering the neurochemical changes upon
   exposure to As[2]O[3] at environmentally relevant concentrations,
   zebrafish behavioral analysis, untargeted liquid chromatography–mass
   spectrometry (LCMS)-based metabolomics, and gene expression analyses
   were performed.

2. Materials and Methods

2.1. Zebrafish Husbandry

   All experiments were performed in accordance with the Institutional
   Animal Care and Use Committee of Universiti Putra Malaysia (UPM)
   (UPM/IACUC/AUP-R049/2019), with an approval date of 23 July 2019. Adult
   wild-type zebrafish were maintained in freshwater at the Natural
   Medicines and Product Research Laboratory (NaturMeds), Institute of
   Bioscience, UPM. Fish were maintained at 25–27 °C with a light cycle of
   14 h light: 10 h dark [[66]25] to induce the reproductive cycle of the
   fish. Fertilized eggs were collected 30 min after lights were turned
   on. The collected eggs were incubated at 28 °C in an egg buffer
   solution [[67]26]. Only embryos with intact chorionic membranes that
   had reached the gastrulation stage (50% epiboly) were chosen [[68]27].
   All dead or unfertilized eggs were removed. After behavioral assessment
   at age 6 dpf, surviving larvae were fed formulated diets and brine
   shrimp (Artemia salinaa) twice until the adult stage [[69]28]. A total
   of 90 of the 6 dpf larvae from each group were evenly distributed to a
   different 3 L freshwater aquarium equipped with dripped water. At 14
   dpf, larvae were provided with a small stream of circulating water
   [[70]29] and reared in this system until the adult stage (3 months).

2.2. Chemical Exposure

   The selected As concentrations correspond to the As concentrations
   reported in the domestic water supply [[71]16]. A stock solution of 100
   mM As[2]O[3] (≥99.99%, Sigma-Aldrich, St. Louis, MO, USA) was diluted
   in 1 M NaOH (0.01%), which then further diluted to final concentrations
   of 20, 30, 40, and 50 μM in egg buffer solution. Zebrafish embryos were
   exposed to a varying range of As[2]O[3] from 5 hpf to 72 hpf under
   semi-static conditions. For behavioral, biochemical, and metabolomics
   analysis, we chose a 30 μM As[2]O[3] concentration, which does result
   in any morphological abnormalities at 5 hpf under semi-static
   conditions until hatching (72 hpf). All experiments were performed in
   triplicate and repeated at least three times (n = 90 embryos per
   exposure group). For larval anxiety-like response assay, an anxiogenic
   (100 mg/L Caffeine, Sigma-Aldrich C53) or anxiolytic drug (5 mg/L
   Buspirone, Sigma-Aldrich B7418) was used in 6 dpf larvae
   [[72]30,[73]31]. Both caffeine and buspirone were dissolved in egg
   buffer solutions at the selected concentration. Larvae were treated
   with caffeine or buspirone 2 h prior to behavioral recording and
   maintained in the same solutions during the behavioral recording
   [[74]32]. To investigate learning behavior at 11 dpf, zebrafish larvae
   were exposed to cognition-impairing MK-801 as a positive control (M107,
   Sigma-Aldrich, St Louis, MO, USA). MK-801 was dissolved in sterilized
   water to prepare a 10 mM stock solution. The MK-801 working solution
   was freshly diluted from the concentrated stock solution with egg
   buffer solution to a final concentration of 200 μM 4 h before the
   experiments [[75]33].

2.3. Embryonic Toxicity Test

   At 4 hpf, normal fertilized embryos with intact chorion membranes were
   selected prior to As[2]O[3] exposure. A comprehensive toxicity
   assessment was performed for each As[2]O[3]-exposed zebrafish group,
   including mortality rate, morphological deformities, survival to adult
   stage, heartbeat (count/minute), and percentage of hatching rate. The
   morphological deformities after As[2]O[3] exposure in zebrafish
   included impairment of fin folds and tail primordium; body axis
   curvature (kink in tail, lordosis, or scoliosis); and abnormal shape of
   yolk, heart, and eyes. Body length, swim bladder diameter, and swim
   bladder volume of surviving larvae were measured at 6 dpf. Swim
   bladders were observed with an SMZ-745T stereomicroscope (Nikon, Nikon
   Instruments Inc., New York, NY, USA). Swim bladder volume was measured
   as follows: 4/3 πab^2, (a) major horizontal axis and (b) minor vertical
   axis [[76]34]. Image analysis was performed using the freely available
   ImageJ software (version 1.48, Wayne Rasband, National Institutes of
   Health, Bethesda, MD, USA from [77]http://rsb.info.nih.gov/ij/web page
   (accessed on 15 July 2021). The swim bladder elongated
   anterior–posterior was flattened, resulting in decreased volume.
   As[2]O[3]-exposed embryos were raised to adult stage, and their
   survival was recorded throughout the growth phase.

2.4. Assessment of Locomotor, Anxiety, and Color Preference in Zebrafish

   After As[2]O[3] exposure, we examined locomotor activity and
   anxiety-like responses at the larval stage 6 dpf. Vertical swimming
   behavior was defined as the ability of larvae to reach the water
   surface [[78]34]. Down preference in this assay denotes the percentage
   of larvae positioned at the bottom of the glass cylindrical column. To
   further examine the effect of swim bladder changes on vertical swimming
   behavior, we measured swim bladder diameter and swim bladder volume of
   larvae 6 dpf after the vertical swimming behavior was recorded. The
   exploratory activity of larvae was examined using an open field test
   that measured the distance traveled [[79]35]. For anxiety-like response
   assay, we measured the percentage of down and edge preference, swimming
   speed, and percentage of rest. All larval behavioral procedures were
   performed in triplicate and repeated at least three times (n = 90
   embryos per exposure group). An aversive stimulus ([80]Figure S3)
   represented by a red moving ball was introduced to 6 dpf larvae for 5
   min using Microsoft PowerPoint (version 2010, Robert Gaskins and Dennis
   Austin, Santa Rosa, CA, USA) [[81]26]. ImageJ and Microsoft Excel were
   used to auto-generate the percentage of down and edge preference,
   swimming speed, and percentage of rest [[82]32]. Assessment of
   avoidance response (down preference) in the anxiety-like response assay
   represented the percentage of larvae positioned at the bottom part of
   the well, as the aversive stimulation from left to right. Larval color
   preference was assessed in zebrafish larvae at 11 dpf by percentage of
   exploration maze and color preference (yellow, green, blue, and red
   represented by total distance traveled) [[83]36,[84]37,[85]38], whereas
   exploratory test and color preference (green or red) were assessed in
   adult zebrafish at 3 months of age [[86]39]. A summary of the
   behavioral assessments is provided in [87]Supplementary Materials
   (Figures S1–S3A).

2.5. Fourier Transform Infrared Spectroscopy (FTIR)

   The 6 dpf zebrafish larvae from both control and As[2]O[3]-exposed
   groups were fixed with 4% paraformaldehyde (PFA) and washed three times
   with phosphate-buffered saline (PBS) for 5 min each. All 180 larvae (90
   = control, 90 = As[2]O[3]-exposed) were dried in a lyophilizer
   (Benchtop Freeze Dryer Labconco, Kansas City, MO, USA) at 50 °C for 12
   h to remove the water content in the samples before grinding in an
   agate mortar and pestle to obtain zebrafish larvae powder. Larvae
   powder was completely mixed with dried potassium bromide (100 mg) and
   subjected to a pressure of 5 t in an evacuated disc for 5 min to
   produce a clear, transparent KBr disc with a diameter of 13 mm and a
   thickness of 1 mm for use in FT-IR analyses [[88]40]. FT-IR analyses of
   the freeze-dried samples were performed using a Thermo Nicolet Nexus
   Smart Orbit spectrometer (Ramsey, NJ, USA). The spectra were recorded
   in the middle infrared (IR) region (500–4000 cm^−1, in triplicate for
   each sample).

2.6. LC-MS Analysis and Metabolomics

   At the end of the behavioral assessment, pools of 90 zebrafish larvae
   at 6 dpf were extracted; the freeze-dried tissues were homogenized in
   300 µL of extraction solvent (80:20 v/v, cold methanol/water) in a 2 mL
   Eppendorf tube. All samples were analyzed as a single batch in random
   order to minimize analytical error and subjective interference and to
   minimize column retention shift. UHPLC analysis was performed using a
   Bruker impact II quadrupole time-of-flight (QTOF)–mass spectrometry
   system (Bruker Daltonics, Bremen, Germany) equipped with an
   electrospray ionization source (ESI). Chromatographic separations were
   performed in an Inertsil phenyl-3 column (150 × 4.6 mm with a particle
   size of 5 µm) (GL Sciences Inc., Rolling Hills Estates, CA, USA) for
   positive- and negative-ion analyses [[89]41]. The injection volume was
   10 µL, with filtration using a 0.22 µm hydrophobic PTFE membrane at a
   flow rate of 0.4 mL/min. The mobile phases consisted of water with 0.1%
   formic acid (solution A) and methanol with 0.1% formic acid (solution
   B). The elution gradient at 50 °C was as follows to ensure improved
   repeatability between runs: (1) 5% solution B for 1 min, (2) 5–50%
   solution B for 11 min, (3) 100% solution B for 23 min, (4) new 100%
   solution B for 35 min, and (5) 5% solution B for 37–50 min. The
   acquisition time for time-of-flight (TOF) mass spectrometry (MS) was
   0.25 s, with a scan range of 70–1250 Daltons (Da). The collision energy
   was set to 35 V, with a collision energy spread of 15 V. A summary of
   sample preparation and LCMS data analysis [[90]42,[91]43,[92]44,[93]45]
   is provided in [94]Figures S3B and S4 in the Supplementary Materials.

2.7. Quantitative Expression Analysis (qPCR)

   To investigate the effects of As[2]O[3] exposure on the expression of
   ASD-associated genes (adenylosuccinate lyase (adsl), SH3 and multiple
   ankyrin repeat domains 3A (shank3a), and tuberous sclerosis complex 1
   (tsc1b) [[95]46,[96]47]), qPCR was performed in triplicate on 6 dpf
   zebrafish larvae (n = 30 larvae per sample). After chemical exposure
   and behavioral recording at 6 dpf, the larvae were transferred into a
   1.5 mL centrifuge tube and flash-frozen for euthanization purposes.
   Total RNA was extracted using an RNeasy UCP micro kit (QIAGEN, Hilden,
   Germany, 2019), with concentration and quality checked with an ND-1000
   spectrophotometer (NanoDrop Technologies, Wilmington, ED, USA). During
   the RNA extraction, genomic DNA was selectively removed with the
   clearing agent that was included in the purification kit. cDNAs were
   synthesized by reverse transcription using a ReverTra AceTM qPCR RT
   master mix with gDNA Remover (Toyobo, Japan). The cDNA concentrations
   were also measured using an ND-1000 spectrophotometer. The samples were
   then diluted with purified water, followed by the addition of 2 µL of
   4× DN master mix incubated at 35 °C for 5 min. A control experiment
   without RNA was used to validate whether amplicons originate from cDNA
   and/or genomic DNA. We used β-actin as a reference housekeeping gene.
   The sequence of primers for the target genes and reference gene
   (β-actin) for zebrafish are shown in [97]Table S1 [[98]48]/Target genes
   were amplified using a CFX96 real-time PCR detection machine (Bio-Rad
   Laboratories, Hercules, CA, USA). The PCR reaction mixture (total 20
   μ/L) contained 10 μL of SensiFAST™ SYBR No-ROX kit master mix (Meridian
   Bioscience, Cincinnati, OH, USA), 0.8 μL of each forward and reverse
   primer (10 μM), 8.4 μL of purified PCR-grade water, and 0.8 μL of cDNA
   sample. The thermal cycle profile was as follows: preincubation at 95
   °C for 2 min; 40 cycles of amplification at 95 °C for 5 s and 60 °C for
   20 s; and annealing at 65 °C for 10 s and 72 °C for 10 s. Variations in
   target gene expression were normalized by using β-actin expression as a
   reference. Delta delta Ct values (ΔΔCt) were used to calculate the
   relative level of gene transcription. The Ct value was determined to
   calculate ΔCt by subtracting the Ct value of the treated and control
   samples. The ΔΔCt value was obtained by subtracting the ΔCt value of
   the target gene of the treated sample from the ΔCt value of the
   housekeeping gene. The expression value of each gene was represented by
   the fold changed, which was calculated as follows: (x = 2^−ΔΔCt).

2.8. Statistical Analysis

   All experiments were repeated three times and performed in triplicate.
   Data were analyzed with SPSS statistical analysis software (version
   22.0, IBM Corp., Armonk, NY, USA) using the probit analysis statistical
   method. The LC[50] values (with 95% confidence limits) were calculated.
   Differences among the results were considered to be statistically
   significant when the p value was <0.05. MS Excel 2007 was used to
   determine the regression equation (Y = mortality; X = concentrations),
   and the LC[50] was derived from the obtained best-fit line. One-way
   ANOVA followed by post hoc Tukey test and two-way ANOVA followed by
   Duncan’s test and t test were applied to determine significant
   differences in teratogenicity, behavior, learning impairment
   assessment, and gene expression between exposed and control groups.
   Data are presented as mean values ± standard error of the mean (SEM),
   with significant differences relative to the control (p-values ≤ 0.05).
   GraphPad Prism statistical software (GraphPad Software, San Diego, CA,
   USA) was used for all graphs.

3. Results

3.1. Developmental Toxicity Effects of Embryonic Exposure to As[2]O[3]

   The total percentage of mortality was represented by dead embryos that
   exhibited coagulation, lack of somite formation, non-detachment of the
   tail, and no heartbeat. [99]Figure 1A shows the percentage of mortality
   in zebrafish larvae until hatching (24–72 hpf). The mortality in
   as-exposed embryos was increased in a dose-dependent manner. Exposure
   to concentrations equal to or greater than 30 µM resulted in a
   significant increase in mortality; in contrast, exposure to 20 µM
   As[2]O[3] showed no significant difference compared to the control
   group. The lethal concentration (LC[50]) of As[2]O[3] killing 50% of
   zebrafish embryos at 96 hpf was 27.10 µM, as shown in [100]Figure 1B.
   Furthermore, no severe morphological malformations (scoliosis, yolk sac
   edema, or tail kinks) were observed in the exposed larval groups
   throughout the exposure period. Exposure to 30 and 40 µM As[2]O[3]
   resulted in a decrease in the incidence of tail coiling compared to the
   control group ([101]Figure 1C), whereas no tail coiling occurred in
   embryos exposed to 50 µM As[2]O[3], as all embryos were dead after 24
   hpf.

Figure 1.

   [102]Figure 1
   [103]Open in a new tab

   The toxicity effects of As[2]O[3] on zebrafish (Danio rerio) embryos.
   (A) Exposure to As[2]O[3] increased embryo mortality in a
   dose-dependent manner. (B) LC[50] for As[2]O[3]. (C) Exposure to 30 µM
   and 40 µM As[2]O[3] significantly decreased the incidence of tail
   coiling in 24 hpf embryos. (D) Exposure to 40 µM As[2]O[3]
   significantly decreased the heartbeat of embryos examined at 48 hpf.
   (E) Exposure to increased As[2]O[3] concentrations significantly
   delayed hatching between 48 hpf and 72 hpf. Data are presented as mean
   ± SEM of triplicate wells (n = 90 embryos per exposure group), with
   significant differences relative to the control group. ∗ (p ≤ 0.05); ^a
   no tail coiling, and heartbeats were recorded for 50 µM As[2]O[3]
   exposed embryos, as all embryos were dead at 24 hpf.

   At 48 hpf, heartbeat was significantly decreased in larvae exposed to
   40 µM As[2]O[3], whereas no significant alterations in heartbeat were
   observed at lower concentrations when compared with the control group
   ([104]Figure 1D). However, no heartbeat was recorded in larvae exposed
   to 50 µM As[2]O[3], as all larvae were dead at 24 hpf. As shown in
   [105]Figure 1E, the percentage of hatched fish decreased over time in
   As[2]O[3]-exposed embryos. Furthermore, 48 h of exposure to 20 µM (8.9
   ± 2.2%) to 40 µM (2.8 ± 2.2%) As[2]O[3] inhibited embryo hatching by up
   to 75% compared to the control group (35.1 ± 2.2%). A similar
   significant trend was also observed after 72 h with 40 µM (68.3 ± 2.2%)
   and 50 µM (0.1667 ± 2.188%) As[2]O[3] exposure compared to the control
   group (98.5 ± 2.2%). In contrast, at 72 h, no significant differences
   in hatching were observed in 20 µM (97.3 ± 2.2%) and 30 µM As[2]O[3]
   (95.4 ± 2.2%)-exposed larvae, suggesting that delaying the hatching of
   zebrafish embryos in the presence of As[2]O[3] may result in abnormal
   organ function in subsequent developmental stages, reducing their
   ability to survive to adulthood. To better understand the toxic effects
   of As, 30 µM As[2]O[3]-exposed larvae were selected due to the
   significant effects of this concentration on total toxicity levels,
   survival to adulthood, and similarity to the arsenic concentration
   found in Langat Basin, Malaysia (0.98–21.94 µg/L) [[106]18].

3.2. Effects of Embryonic As[2]O[3] Exposure on Anxiety-Related Responses in
6 dpf Larvae

   To further understand the toxic effects of As[2]O[3] exposure on
   zebrafish larval development, we examined their anxiety-related
   behavior. Assessment of anxiety-related responses in this study
   included percent of edge preference (thigmotaxis), down preference
   (avoidance response), rest, and speed upon aversive stimulation.
   Aversive stimulus was represented by a red moving ball from left to
   right displayed in Microsoft PowerPoint, whereas no stimulus was
   represented by a blank background. As[2]O[3] exposure resulted in no
   changes in anxiety-related responses in 6 dpf larvae under either
   condition (without/with aversive stimulus) ([107]Table 1 and
   [108]Figure S3A in Supplementary Materials, respectively) when compared
   to controlled larvae.

Table 1.

   Changes in anxiety-related responses in As[2]O[3]-, caffeine-, or
   buspirone-exposed larvae.
   Anxiety-Related Response Controlled Larvae 30 µM As[2]O[3]-Exposed
   Larvae 100 mg/L
   Caffein-Exposed Larvae 5 mg/L
   Buspirone-Exposed Larvae
   Aversive Stimulus Aversive Stimulus Aversive Stimulus Aversive Stimulus
   Without With Without With Without With Without With
   Edge preference 84.0% ± 3.4 87.0% ± 3.4 83.0% ± 3.4 84.6% ± 3.4 89.9% ±
   3.3 92.1% ± 3.7 73.9% ± 5.6 79.3% ± 4.8
   Down preference 48.0% ± 3.9 54.0% ± 3.9 43.1% ± 3.9 49.6% ± 3.9 47% ±
   2.8 58.9% ± 2.9 57.9% ± 4.1 74.3% ± 3.9
   Speed 42 ± 3.1 mm/min 39 ± 3.0 mm/min 42.8 ± 3.0 mm/min 47.3 ± 3.0
   mm/min 9 ± 2.1 mm/min 11.0% ± 2.1 mm/min 47 ± 2.1 mm/min 43 ± 2.1
   mm/min
   Rest 20% ± 4.6 24% ± 4.6 17.4% ± 3.0 18.8% ± 3.0 72% ± 2.1 74% ± 2.1 0%
   0%
   [109]Open in a new tab

   Exposure to anxiogenic caffeine increased edge preference under both
   conditions, reduced downward preference, reduced swimming speed, and
   increased rest. These results suggest that caffeine exaggerated edge
   preference and reduced larval avoidance behavior compared to control
   larvae. In contrast, exposure to anxiolytic buspirone decreased edge
   preference and increased down preference and swimming speed, and no
   resting larvae were detected, indicating that all larvae moved under
   both conditions compared to the control group. These results suggest
   that buspirone had minimal effects on larval edge preference and
   significantly enhanced larval avoidance behavior. Overall, the data
   show that As[2]O[3], buspirone, and caffeine each have different
   effects on larvae.

3.3. As[2]O[3] Affects Survivability and Induces Behavioral Defects during
Juvenile to Adult Stages

   After the exposure period, zebrafish larvae were rinsed and reared to
   maturity under normal laboratory conditions. The long-term deleterious
   effects of As[2]O[3] were evidenced by reduced survivability at 72 hpf,
   increase in swim bladder volume with impaired vertical swimming
   behavior at 6 dpf, and a persistent decrease in exploratory behavior
   until adulthood. However, increased swim bladder volume did not affect
   survivability itself at 72 hpf after As[2]O[3] exposure. For both
   control and As[2]O[3]-treated larvae, survival was 100% by day 12 and
   decreased to 86.7% by day 13 ([110]Figure 2D). The survivability was
   maintained after day 14 in control and after 23 dpf in
   As[2]O[3]-exposed fish during juvenile and adult growth. Throughout the
   rearing process, not a single death was recorded after 72 hpf in
   As[2]O[3]-exposed fish, comparable to the control group. Exposure to
   As[2]O[3] (1.4 ± 0.05 mm^3) resulted in a smaller diameter of the
   posterior lobe of the swim bladder versus the control group (1.7 ± 0.05
   mm^3, [111]Figure 2A). This was supported by the fact that swim bladder
   volume was significantly enlarged in As[2]O[3]-exposed larvae (0.4 ±
   0.07 mm^3) compared to that of the control group (0.3 ± 0.07 mm3,
   [112]Figure 2B). Increased swim bladder volume affected the neutral
   buoyancy of larvae, ultimately affecting their survival in later
   stages, as well as subsequent behavior. Larvae exposed to As[2]O[3]
   (44.6 ± 1.1%) showed a significantly decreased down preference when
   compared with the control group (61.8 ± 1.1%) ([113]Figure 2C). These
   excessive floating attempts were consistent with the higher swim
   bladder volume observed in As[2]O[3]-exposed larvae, indicating an
   impaired control of neutral buoyancy affecting swim behavior [[114]34].

Figure 2.

   [115]Figure 2
   [116]Open in a new tab

   Effects of embryonic As[2]O[3] exposure in 6 dpf larvae and adult
   zebrafish. At 6 dpf, As[2]O[3] exposure increased swim bladder
   inflation/volume (B), although its anterior–posterior length was
   shortened (A). (C) With change in swim bladder volume, As[2]O[3]
   exposure also affected swimming behavior, with reduced down preference.
   (D) In adults, embryonic As[2]O[3] exposure resulted in a reduced
   percentage of survivability in As[2]O[3] exposed larvae (x), which was
   mainly detected at 13–23 dpf, compared to the control group (□). Data
   are presented as mean ± SEM, (∗ p ≤ 0.05) n = 30–90 per group), with
   significant differences relative to the control group ∗ (p ≤ 0.05).

   As[2]O[3] caused a persistent reduction in larval exploratory behavior
   until adulthood. The exploratory behavior of zebrafish at 6 dpf,
   represented by the average distance traveled, was significantly
   decreased in As[2]O[3]-exposed larvae (219.6 ± 5.7 mm) compared to the
   control group (271.7 ± 5.7 mm, [117]Figure 3A). These findings are
   consistent with the positive control, caffeine-treated larvae (58.1 ±
   5.7 mm), which showed a significant reduction in average distance
   traveled. However, buspirone treatment (negative control) resulted in a
   longer average distance traveled (242.2 ± 5.7 mm), which was
   significantly greater than that of the other exposed groups. The
   exploratory activity of zebrafish was further evaluated at 11 dpf and 3
   months of age. The results showed that exploratory activity of
   As[2]O[3]-exposed larvae (83.6 ± 12.8 mm) was significantly reduced
   compared to control larvae (131.4 ± 12.8 mm [118]Figure 3B).
   Coincidentally, zebrafish larvae that were treated with cognitive
   impairer MK-801, which mechanistically blocks NMDA/glutamatergic
   signaling, showed a severe reduction in exploratory activity (11.4 ±
   12.8 mm) compared to control larvae ([119]Figure 3B). In addition, no
   morphological deformities were observed in MK-801-exposed larvae
   throughout the exposure period. Locomotor behavioral assessment,
   represented as exploratory activity, performed prior to any aversive or
   color stimulation showed that these detrimental effects persisted into
   the adult stage, as As[2]O[3] (417.7 ± 91.8 mm) significantly decreased
   exploratory activity compared to control larvae (640.5 ± 91.8 mm)
   ([120]Figure 3C).

Figure 3.

   [121]Figure 3
   [122]Open in a new tab

   Effects of long-term impairment of 30 µM As[2]O[3] exposure on
   exploratory activity. Exploratory activity was persistently decreased
   in larvae from 6 dpf (A) and 11 dpf (B) to adult stage (C). Data are
   presented as mean ± SEM, (∗ p ≤ 0.005), n = 30 larvae per group, n = 22
   adults per group), with significant differences relative to the control
   group, ∗ (p ≤ 0.05).

3.4. As[2]O[3] Exposure Affects Color Preference and Learning Impairment

   An innate color preference test was performed in a plus maze with four
   different-colored sleeves to assess the effects of embryonic As[2]O[3]
   exposure in 11 dpf zebrafish larvae. This color test showed that
   control zebrafish larvae exhibited distinct color discrimination and
   color preference. We observed a reduction in color preference for blue
   in As[2]O[3]-exposed larvae (749.3 ± 132.4 s) compared to the control
   group (1052 ± 133.5 s) ([123]Figure 4A). No significant differences
   were observed in color preference for red and green in control (336.1 ±
   132.4 s vs. 271.6 ± 132.4 s) or As[2]O[3]-exposed larvae (262.7± 132.4
   s vs. 250.2 ± 132.4 s). However, MK-801-exposed larvae showed a change
   in color preference (blue: 303.3 ± 132.4 s, red: 305.3 ± 132.4 s,
   green: 579.8 ± 132.4 s, and yellow: 361.6 ± 132.4 s). MK-801-treated
   larvae swam longer in the center of the maze without showing a clear
   color preference.

Figure 4.

   [124]Figure 4
   [125]Open in a new tab

   Effects of embryonic As[2]O[3] exposure on the color preference of
   larval (11 dpf, A) and adult zebrafish (B). (A) In the 11 dpf larval
   test, a cross maze with four different-colored cambers was used.
   Reduction in color preference for blue in As[2]O[3]-exposed larvae
   compared to the control group. No significant differences were observed
   in color preference for red and green in control or As[2]O[3]-exposed
   larvae. However, MK-801-exposed larvae showed no clear color preference
   (B) Five-minute video tracking of color preference in adult fish after
   associative learning. Adult fish were acclimated to the three-chamber
   maze for 6 days with red food association before the test on the 7th
   day. As[2]O[3]-exposed zebrafish showed no significant preference for
   either green or red color Data are presented as mean ± SEM, (∗ p ≤
   0.005). n = 30 larvae per group, n = 22 adults per group. ∗
   Significance at p ≤ 0.05 between left/center/right and between two
   color arms for each test, n.s: not significant.

   Because the color preference test was previously used to evaluate adult
   zebrafish behavior for learning and memory [[126]49,[127]50], we
   examined color preference in 3-month-old adult zebrafish using a
   three-chamber apparatus with red and green sleeves to assess long-term
   effects of embryonic As[2]O[3] exposure on color preference in adult
   fish. It was previously established that red is the most preferred
   color in zebrafish and that associations with red are easily learned
   from food color (for example, brine shrimp) during rearing [[128]51].
   Thus, we used red color preference for food-associative learning in
   adult zebrafish. To increase food color-associated learning in the
   color preference test, adult fish were acclimated in the maze for 6
   days with red color brine shrimp feeding before the test on the 7th
   day. In the three-chamber color preference test, control fish showed a
   tendency to prefer the red-colored zone over the green zone (red, 139.1
   ± 22.0 s; green, 78.5 ± 22.0 s; center, 81.9 ± 22.0 s). However, no
   such response was observed in the As[2]O[3]-exposed adult group,
   showing no significant preference for either green or red color (red,
   74.6 ± 22.0 s; green, 62.2 ± 22.0 s; center: 162.6 ± 22.0 s)
   ([129]Figure 4B). This lack of color preference and lost directional
   preference ([130]Figure S12) in the adult stage may indicate an
   association with color impairment in the As[2]O[3]-exposed group.

3.5. Disturbance of Lipid and Fatty Acid Metabolites

   To assess changes in biochemical information and understand both
   molecular structure and molecular composition, freeze-dried samples of
   6 dpf control (whole body) and 30 µM As[2]O[3]-treated zebrafish larvae
   were subjected to FTIR analysis. As[2]O[3] exposure causes biochemical
   alterations in proteins, lipids, carbohydrates, and nucleic acids of
   larvae. The regions of transmittance in the FTIR spectra are directly
   proportional to concentration of the molecules. The FTIR spectrum of 6
   dpf zebrafish larvae showed a complex of several bands originating from
   functional groups belonging to lipids, proteins, nucleic acids, and
   carbohydrates ([131]Table S2 and Figure S6 in Supplementary Materials).
   However, a more detailed picture of altered metabolites could be
   identified using a more sophisticated LCMS-based metabolomics tool to
   detect a broad spectrum of affected metabolites with high sensitivity
   and resolution.

   One approach to detect significant similarities and differences among
   affected metabolites in large metabolomics datasets is multivariate
   data analysis (MVDA) methods, including unsupervised principal
   component analysis (PCA) and supervised partial least
   squares-discriminant analysis (PLS-DA). In this MVDA method, the tested
   samples are clustered based on their variance by exposing them to
   different principal components (PCs). The metabolites are believed to
   be responsible for group separation by PCA ([132]Figure S7) and PLS-DA
   loadings ([133]Figure S8). To further identify the significant
   metabolites contributing to the discrimination, a second precise and
   straightforward comparison was performed between the As[2]O[3]-exposed
   zebrafish larval group and the control group, as shown in the
   supervised orthogonal projections to latent structure discriminant
   analysis (OPLS-DA) ([134]Figure 5). OPLS-DA using the S-plot was
   helpful in clarifying and identifying biomarkers in both groups.

Figure 5.

   [135]Figure 5
   [136]Open in a new tab

   OPLS−DA score plot and loading scatter plot (A,B) of negative and
   positive (C,D) modes based on the zebrafish larvae normalized data
   exposed to 30 µM As[2]O[3] in comparison to control larvae. (E)
   Differential expression of metabolites in As[2]O[3]-exposed larvae
   produced by hierarchical clustering of the most significantly
   upregulated (red) and downregulated (blue) metabolites obtained from in
   negative- and positive-ion modes compared to the control group based on
   the log2 fold change value. (F) Metabolic set enrichment analysis of
   lipid metabolites in 6 dpf zebrafish showed the biosynthesis of
   unsaturated fatty acids, arachidonic metabolism, and sphingolipid were
   dysregulated after embryonic exposure to As[2]O[3]. Color intensity
   (yellow–to–orange/red) represents increasing statistical significance,
   whereas circular diameter is related to pathway impact. The graph was
   obtained by plotting−log of p-values from pathway enrichment analysis
   on the y-axis and the pathway impact values derived from pathway
   topology analysis on the x-axis.

   The potential biochemical biomarkers for the long-term effects of
   As[2]O[3] compared to control larvae were further computed using a
   supervised OPLS-DA analysis (score and loading plots of negative
   ([137]Figure 5A,B) and positive ([138]Figure 5C,D) mode) to correlate
   the changes in metabolites in As[2]O[3]-exposed larvae with
   neurotoxicity and behavioral abnormalities. The generated OPLS-DA score
   plot shows significant differences between the As[2]O[3]-exposed
   zebrafish larvae and the control groups ([139]Figure 5A,C). In negative
   mode, four components were produced ([140]Figure S9A,B) with acceptable
   goodness of fit and predictability of the model, as evidenced by R2Y
   and Q2Y values of 0.997 and 0.912, respectively. In contrast, the
   positive mode generated five components ([141]Figure S9C,D), with an
   R2Y value of 1 and a Q2Y value of 0.657. One hundred random
   permutations of the y variables confirmed the minimal validity of the
   current model of y-axis intercepts below zero, as indicated in SIMCA
   ([142]Figure S11). The validity of metabolite changes was supported by
   observed vs. predicted plots, with excellent correlation between all
   features (y data) and metabolites (x data), as represented by
   regression lines with a value of 1 ([143]Figures S10 and S11).

   Of the total 150 identified features, 13 metabolites were found to be
   significantly altered between the As[2]O[3]-exposed and control groups.
   [144]Table S3 lists the identified metabolites and their changes with
   significant p and FDR values. The generated p and FDR values are <0.01,
   which is below the threshold of 0.05, indicating that the observed
   changes are statistically robust. The log2 fold change in metabolite
   levels, presented as red pixels (for upregulation) and blue pixels (for
   downregulation) in [145]Table S3 (in Supplementary Materials), are
   shown in the heatmap ([146]Figure 5E). The results show that nine
   metabolites were significantly downregulated in As[2]O[3]-exposed
   zebrafish larvae, including arachidonic acid, docosahexaenoic acid
   (DHA), palmitic acid, 9,10-epoxyoctadecenoic acid,
   sphinganine-1-phosphate, L-palmitoylcarnitine, prostaglandin,
   cholesterol, and sulfate tetrahydrocorticosterone. In contrast, four
   metabolites, namely 5,6-epoxy-8,11,14-eicosatrienoic acid,
   7α-hydroxy-3-oxo-4-cholestenoate, stearic acid, and homogentisic acid,
   were upregulated. Following FTIR and LCMS-based metabolomics analyses,
   which showed that As[2]O[3] primarily affected lipids, further targeted
   data analysis of lipid metabolites was performed, which included
   pathway enrichment analysis combined with topology analysis to identify
   the major metabolic pathways affected by As[2]O[3] ([147]Figure 5F).
   Based on KEGG metabolic pathways, As[2]O[3] was found to be responsible
   for disrupting three major metabolic pathways, namely unsaturated fatty
   acid biosynthesis, arachidonic acid, and sphingolipid metabolism in
   zebrafish larvae.

3.6. Alterations in ASD-Associated Genes

   As[2]O[3] exposure resulted in significant overexpression of adsl (3.57
   ± 0.19) and shank3a (1.8 ± 0.31) genes and downregulation of tcs1b
   (0.49 ± 0.15) compared to control larvae (adsl: 0.8585 ± 0.1949,
   shank3a: 0.87± ± 0.30, tsc1b: 1.106 ± 0.1466) ([148]Figure 6). Overall,
   the results of this study highlight that embryonic exposure to
   environmentally relevant As concentrations in zebrafish embryos could
   be associated with NDD, particularly ASD. ASD-associated genes were
   selected due to the widespread prevalence of ASD-like symptoms, such as
   cognitive deficit, and the anticipated burden of ASD in children in
   Malaysia. A proportion of 3.3% of infants in Malaysia were reported to
   exhibit developmental delays, with 4.7% of children born with
   disabilities [[149]52]. Besides genetics, environmental exposure to
   arsenic is a potential risk factor for ASD [[150]53], which is usually
   accompanied subtle effects in childhood with potentially late onset
   [[151]54].

Figure 6.

   [152]Figure 6
   [153]Open in a new tab

   Embryonic exposure to As[2]O[3] induced alterations in transcriptional
   regulation of ASD-associated genes. As[2]O[3] exposure resulted in
   significant overexpression of adsl and shank3a, whereas tsc1b was
   downregulated. Data are presented as the mean ± SEM, (* p < 0.05) (n =
   90 per group).

4. Discussion

   This study demonstrated that embryonic exposure to As[2]O[3] in
   zebrafish increased mortality, decreased heart rate, and reduced the
   incidence of tail coiling in a dose-dependent manner. Additionally,
   As[2]O[3]-exposed larvae showed motor behavior deficits, followed by an
   impairment in color preference at 11 dpf and later in adulthood.
   Preliminary biochemical evaluation by FTIR showed that 30 µM As[2]O[3]
   induced changes in lipid, protein, carbohydrate, and nucleic acid
   profiles. Metabolomics analysis further revealed disruption of lipid
   metabolites involving arachidonic, sphingolipid, and biosynthesis of
   unsaturated fatty acid metabolism in As[2]O[3]-treated zebrafish
   larvae. By integrating metabolite dysregulation, behavioral alteration,
   and altered regulation of ASD-associated genes, these results support
   the idea that embryonic As[2]O[3] exposure could be involved in NDD
   pathogenesis. Lipid alteration associated with cognitive deficit, which
   have been reported in both ASD and ADHD [[154]55], highlights the
   possibility that a similar metabolism could be affected likewise in
   NDD.

   Exposure to concentrations equal to or greater than 30 µM As[2]O[3]
   caused a significant increase in mortality and induced weak heartbeats.
   This result is consistent with previous studies, which reported that
   zebrafish exposed to a range of As in later stages (15–96 hpf)
   developed edema and scoliosis, which could lead to cardiac
   malformations and mortality [[155]56]. In contrast, 20–50 µM
   As[2]O[3]-exposed embryos (5–72 hpf) showed no morphological
   abnormalities. However, a significant increase in embryo mortality was
   observed after exposure to 40–50 µM As[2]O[3]. The increase in
   mortality and occurrence of abnormalities after exposure to
   thousand-fold As (2 mM) began prior to gastrulation (4 hpf) and post
   gastrulation (6 hpf), signifying the vulnerability of the exposure
   window [[156]57,[157]58]. This suggests that zebrafish larvae are more
   sensitive to toxic effects of As[2]O[3] at younger ages, even at
   micromolar concentrations, resulting in reduced survival into
   adulthood, also affecting defective swimming activity [[158]34].
   As[2]O[3] exposure also affected swim bladder inflation and vertical
   swimming behavior. However, further exploration is required to identify
   the molecular mechanism underlying swim bladder development or
   inflation [[159]59,[160]60]. Larvae stayed less on the edge after
   exposure to As[2]O[3] and buspirone, suggesting that anxiety was
   reduced, although non-significantly. In contrast, larvae tended to stay
   on the edge after exposure to caffeine, suggesting that anxiety was
   increased. This result is consistent with previous reports that high
   levels of caffeine can increase anxiety [[161]32,[162]61] and increase
   edge preference in response to visual stimuli [[163]32].

   Although no noticeable malformations were observed, micromolar
   As[2]O[3]revealed its toxic effect by reducing the incidence of tail
   coiling and causing a persistent reduction in larval locomotor activity
   until the adult stage. This cumulative effect suggests that embryos
   treated with As[2]O[3] from the gastrulation stage could be impaired in
   essential neurodevelopmental processes [[164]60,[165]61]. However,
   further investigation is required to identify the exact molecular
   mechanism. The obtained results are comparable to those of a previous
   embryonic exposure study investigating thousand-fold NaAsO[2] exposure
   (1 mM) (0–120 hpf), which reduced locomotor activity in zebrafish
   [[166]62]. A similar trend was observed in rats, with locomotor
   decreased activity by 0.10 mg/L AsNaO[2], which was associated with an
   increase in oxidative stress and inhibition of AChE in the striatum
   [[167]63]. Another study [[168]64] reported decreased myoblast
   proliferation and a reduced number of muscle fibers, resulting in a
   long-lasting impairment of locomotor activity in fish that persisted
   into adulthood [[169]65] after exposure to 1 µM arsenite for 72 h,
   which is consistent with the results of the present study. However,
   As[2]O[3] exposure had no effect on anxiety responses in the present
   study.

   Innate color preferences are vital abilities for fish to learn and make