Zebrafish embryonically exposed to valproic acid present impaired retinal development and sleep behavior

Prenatal exposure to valproic acid (VPA), a drug widely used to treat epilepsy and bipolar disorder, is an environmental risk factor for autism spectrum disorder (ASD). VPA has been used to reproduce the core symptoms of ASD in animal model organisms, including zebrafish. Visual system functioning is essential in the interpretation of social conditions and plays an important role of several behavioral responses. We hypothesized that behavioral deficits displayed by ASD patients may involve impaired visual processing. We used zebrafish as model organism to investigate the visual system after embryonic exposure to VPA using histological, behavioral and gene expression analysis. We analyzed the pineal gland of zebrafish and sleep‐like behavior to study how VPA exposure alters photo‐sensibility of zebrafish. VPA‐exposed zebrafish showed a delay in the development of the retina and optic nerve, which normalized at five days post fertilization. At larval stage, VPA‐exposed zebrafish showed sleep disturbances associated with a reduced number of serotonin‐producing cells of the pineal gland. In addition, the number of hypocretin/orexin (hcrt) expressing neurons in the rostral hypothalamus at 6 and 14 days post fertilization was reduced. In conclusion, we demonstrated that although VPA exposure leads to a delay in visual system development, it does not affect larval visual function. The novel finding that VPA alters significantly cells involved in sleep regulation and the sleep‐like state itself may be relevant for understanding sleep disturbances in ASD patients.


INTRODUCTION
Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by deficits in social behavior and communication and stereotypies (Willsey et al., 2022).The etiology of ASD is not fully understood, but environmental and genetic factors are known to have a role in the development of this disorder (Lord et al., 2020).Exposure to valproic acid (VPA), a drug used to treat epilepsy and psychiatric disorders, has been associated with a 5-fold increase in ASD risk in general population (Błaszczyk et al., 2022;Wood et al., 2015).A wide range of co-occurring symptoms is observed in ASD patients, including altered sensory integration/processing such as visual defects, and sleep disturbances (Lane et al., 2022).
The possible causes of insomnia in ASD include impairments in systems that promote sleep and establish a regular sleep-wake cycle, such as melatonin (Martinez-Cayuelas et al., 2022).Alterations in the serotonergic system and hypofunction in melatonin secretion have been reported in patients with ASD (Lalanne et al., 2021;Pourhamzeh et al., 2022).The central pacemaker in nonmammalian vertebrates resides at the pineal gland, which regulates the circadian rhythms of animal behavior and physiology (Li et al., 2012).In the zebrafish, the pineal gland is located in the dorsal diencephalon and possesses pineal photoreceptors which are key players in controlling this activity, as they secrete the sleep-promoting hormone melatonin (Falc on et al., 2010;Sapède et al., 2020).More complete understanding of the developmental and exogenous influences of altered serotonin (5-HT) function on ASD core and associated symptoms may yield new opportunities, such as the development of new therapeutic approaches for this disorder.For that, model organisms are important tools.
A combination of well-characterized CNS development and availability of diverse tools for genetic manipulation make zebrafish an useful model to study the molecular underpinnings of ASD (Lillesaar & Gaspar, 2019;Stewart et al., 2015).Embryonic exposure to VPA has been used to generate a ASD-like phenotype, which is characterized by molecular and neurochemical abnormalities (Chen et al., 2018), which persist into adulthood and accompany impaired sociability (Baronio et al., 2018).
Even though VPA exposed zebrafish present impaired social behavior (Baronio et al., 2018;Chen et al., 2018), the visual system of these animals has not been thoroughly studied.This assessment is important because social preference in zebrafish is highly visuallydriven (Bollmann, 2019).Additionally, it has been hypothesized that the behavioral deficits displayed by ASD patients might be related with impaired visual processing (Cheng et al., 2020;Guimarães-Souza et al., 2019).Visual system functioning is essential in the interpretation of social conditions and plays an important role of various behavioral responses and circadian cycle (Lockley et al., 2007).Thus, we investigated the development of the visual system of zebrafish embryonically exposed to VPA and assessed behavioral responses do different visual stimuli.Furthermore, considering the role of 5-HT in important physiological processes, such as sleep regulation, and its proposed role in the pathophysiology of ASD, we also analyzed its distribution in the retina and pineal gland of these animals.Finally, we studied the locomotor activity of VPA-exposed zebrafish larvae during 24 h to reveal possible disturbances in sleep wake cycle.

Animals and VPA exposure
Zebrafish of the Turku strain at embryonic, larval and juvenile stages of development were used in this study.In total, 12 embryos (2 days post fertilization; dpf), 87 larvae (5-8-dpf) and 74 juveniles (1 month post fertilization; mpf) were used in the behavior or histological experiments (each experiment was performed in duplicate from different breedings).In addition, we collected 95 larvae from each experimental group (5-dpf) to perform qPCR analysis.After fertilization, embryos were distributed randomly in a 6-well plate (30 embryos per well).VPA (Sigma P4543, Darmstadt, Germany) diluted in E3 medium (5.00 mM NaCl, 0.44 mM CaCl 2 , 0.33 mM MgSO4 and 0.17 mM KCl) at a concentration of 25 μM was added into wells at 10 hours postfertilization (hpf).At 24-hpf, embryos were rinsed with E3 medium, placed into petri dishes and transferred to an incubator at 28.5 C (Baronio et al., 2018).Different animals were used for behavioral tests and staining experiments.

Immunohistochemistry
We performed an immunohistochemical analysis to evaluate the histological changes in the retina and brain of VPA-exposed fish and control fish.Animals were fixed in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS; pH 7.4) overnight at 4 C.After washes in PBS, the samples of larvae at 8-dpf and juveniles at 1-mpf (n = 5 per group) were cryoprotected in 30% sucrose, embedded in Tissue-Tek OCT Compound (Sakura, USA) and cryosectioned at 14 μm thickness.Sections containing the brain and the retina of animals were analyzed.Embryos (2-dpf) and larvae (5-dpf) were directly incubated in blocking solution for staining of whole animals (n = 6 per group).To analyze serotonergic cells in the pineal gland we dissected the whole brain from 6-dpf fish (n = 6 per group).Immunohistochemistry was performed as previously described (Baronio et al., 2018;DeOliveira-Mello et al., 2019) by using primary and secondary antibodies shown in Table 1, diluted in 0.02% (sections) or 0.3% (whole tissues) Triton X-100 in PBS.Blocking steps were done by using 5% goat serum.After several washes in PBS, sections were coverslipped in Fluorescence Mounting Medium (DAKO, Denmark) and whole embryos/larvae were mounted in glycerol.

Length measurement of optic nerve zones
The structure of the optic nerve of teleost fish has been divided in different regions for its study (Lillo et al., 1998;Maggs & Scholes, 1990), which include prechiasmatic segments (the optic nerve head and the intraorbital segment), the chiasm, and the postchiasmatic segment.The optic nerve head is the region where the axons of ganglion cells pass through the retinal layers to form the optic nerve.The intraorbital segment is the region where the axons get out of the retina until the chiasm.We measured the thickness of the intraorbital segment of embryonic optic nerve at 2-dpf (both eyes of 6 embryos per group), and the optic nerve head and the intraorbital segment lengths of zebrafish larvae (5-dpf; from 6 larvae of each group).We used the quantify tool of Leica Application Suite X (LAS X) software.We drew a line (Region of Interest; ROI) to define the distance between two points and extracted the length information in μm to an excel file.We then calculated the proportion of the optic nerve thickness in embryos, and analyzed the differences in length of each zone in the larvae.

In situ hybridization
For in situ hybridization we used the already published protocol (Sundvik et al., 2011) to detect hypocretin/ orexin (hcrt) expression in whole brains of zebrafish at 6-dpf (n = 6 per group) and 14-dpf (n = 9 per group).Briefly, larvae were killed in ice-cold water and fixed in PFA 4% overnight at 4 C.After that, the brains were dissected and stored in 100% methanol at À20 C until used.Samples were then rehydrated, and digested with proteinase K for 20 min and pre-hybridized for 1 h at 60 C in 50% formamide, 5Â saline-sodium citrate solution (SSC), 0.1% Tween20, 50 μg/mL of heparin and 500 μg/mL of tRNA (pH 6).Between the steps, samples were washed in PBS with 0.1% Tween20 (PBST) during 5 min each.The samples were hybridized overnight at 60 C in similar pre-incubation solution, with added 250 μL of digoxigenin riboprobes against hcrt (Sundvik et al., 2011).During the following day, the samples were washed in pre-incubation solution for 5 min.Preincubation solution was removed in four washes of decreasing concentrations with 2xSSC for 10 min each.Afterwards, 2xSSC was replaced for 0.2xSSC in two washes during 10 min each.An antibody to detect digoxigenin (in a concentration of 1:10,000) was then used, followed by incubation with NBT/BCIP according to the manufactured instructions as described earlier (Sundvik et al., 2011).Samples were then mounted in 80% glycerol diluted in PBS for imaging.

Optomotor behavior
To evaluate the visual perception of VPA-exposed fish and control animals we performed an optomotor assay adapted from the protocol of Orger et al. (2000).Dr. Stephan Neuhauss from University of Zurich kindly provided us the strip movies, which were displayed in an Asus MB169 screen connected to a conventional laptop computer (Dell).Single fish were placed in an acrylic chamber (length 33 cm Â width 1 cm Â depth 1 cm) (Supplementary Figure 1), under which a visual stimulus lasting for 30 s was displayed.All individuals (n = 19 per group) were tested twice, using both directions of the moving strips, with a 15 min interval between the trials.In total, we recorded 38 videos from each group (VPAexposed and controls) using an Iphone 7 camera (Supplementary Figure 1).
All videos were analyzed frame by frame to determine the exact location of the fish at the beginning and at the end of the trial.Distances between these locations were measured by using Fiji software (v.1.53c) (protocol available in https://imagej.nih.gov/ij/docs/pdfs/ImageJ.pdf).Color preference Zebrafish show innate preference and avoidance for certain colors (de Abreu et al., 2021).Thus, we evaluated if control and VPA-exposed fish (1-mpf) displayed any differences in color perception.The assay was conducted in a glass tank (length 9.5 cm Â width 7.5 cm Â depth 6 cm), filled with 180 mL of system water.One-half of the tank area with a size of approximately 4.75 Â 3.75 cm was covered with yellow tape and the other half was covered with red tape.This tank was placed inside the DanioVision observational chamber (Noldus Information Technology, Wageningen, Netherlands).Fish were individually placed into the tank (n = 13 per group) and had 5 min of habituation time before the 10 min trial started.The time spent on each side of the tank was measured with EthoVision 13 software (Noldus Information Technology, Wageningen, Netherlands).

h locomotor activity
Evaluation of the sleep-like behavior was performed at larval stage (6-7-dpf) according to a previous report (Puttonen et al., 2018).Briefly, larvae (n = 20 to controls and n = 21 to VPA-exposed animals, run in duplicate) were automatically tracked on a 48-well plate with E3 medium during 24 h using the DanioVision system (Noldus Information Technology, Wageningen, Netherlands), with the light conditions (250 steps $330 lux) following the regular light/dark cycle of the larvae.Both groups were acclimatized in the experimental chamber for 10 min before tracking starts.Using the EthoVision software, we analyzed the total distance moved during the 24 h in 30 min bins.We also calculated the periods of sleep-like state during the night time (period of darkness).To determine the sleep-like state we extracted the distance moved during darkness in 1 s bins, and counted the number of immobility bouts lasting 60 s in one hour based in previously described protocol (Puttonen et al., 2018).

Quantitative PCR (qPCR)
Animals were killed on ice-cold water for RNA extraction.Total RNA was extracted using the RNeasy mini kit (Qiagen, Germany) following the manufacturer's instructions.Each sample consisted of a pool of 30-33 dissected eyes at 5-dpf, and 1.0 μg of total RNA was reverse-transcribed using Super Script TM III reverse transcriptase (Invitrogen, USA).qPCR analysis was performed with the LightCycler 480 real-time PCR system and the LightCycler 480 SYBR green I master kit (Roche Applied Science, Germany).Sequences of primer-BLAST (NCBI) designed are shown in Table 2.All samples were tested in duplicate to ensure the reliability of single values.Fluorescence changes were monitored with SYBR Green after every cycle.Dissociation curve analysis was performed (0.2 C per s increase from 60 to 95 C with continuous fluorescence readings) at the end of cycles to ensure that only a single amplicon was obtained.Cycling parameters were as follows: 95 C for 30 s and 45 cycles of the following, 95 C for 10 s and 62 C for 45 s. Results were analyzed with the LightCycler 480 software (v.1.5),and quantification was performed by Ct value comparison by using the Ct value of β-actin as the reference control.

Image acquisition and analysis
To analyze immunostained samples, we used a Leica SP2 AOBS confocal microscope.We scanned and acquired images of whole-mounts and sections using a HC PL APO 20x/0.70 and HCX PL APO 40x/0.75dry objectives.Laser excitations at 488 nm and 568 nm were used in sequential scans, and emission was collected at 500-550 nm and 601-647 nm, respectively.Images were

Statistical analysis
Data from EthoVision automatic tracking analysis were exported to an excel file.We counted 5-HTimmunoreactive cells (previously identified as pinealocytes (Lee et al., 2021;Lillesaar, 2011;Sun et al., 2002) in the pineal gland at 6-dpf and from the histological sections of the retina at 8-dpf (n = 3-5 sections from 5 animals per group).In addition, we quantified the photoreceptors in the pineal gland (positive for Zpr-1) and the cells expressing hcrt in the rostral hypothalamus at 6-dpf.The statistical analysis was performed using GraphPad Prism software (v.9.

RESULTS
Larvae, but not juvenile zebrafish, present retinal abnormalities after embryonic exposure to VPA We used neurolin (Zn5), Sox2 and Zpr-1 antibodies to analyze retinal development (Figure 1).In zebrafish, Zn5 has been described as an adhesion protein in differentiating neurons and their growing axons (Laessing & Stuermer, 1996).We found that the structure of developing neurons of VPA-exposed animals at 2-dpf was highly altered (Figure 1a, b).The retina of VPA-exposed embryos was flattened in comparison to controls (Figure 1a, b).The ventral view of the optic nerve at 2-dpf showed differences in the thickness and localization between the groups (Figure 1a, b).The Zn-5 positive fibers of embryos exposed to VPA were more than 40% thinner compared with controls (mean of VPA = 0.92 ± 0.36 μm and mean of controls = 2.27 ± 0.38 μm).At larval stages (5-dpf; Figures 1c, d and 2), the intraorbital optic nerves (Lillo et al., 1998) were very similar in length in the two groups (Figure 2).The optic nerve head was smaller in VPA-exposed zebrafish (Figures 1c, d and 2).
In addition, the retina of VPA-treated larvae showed Zn5 positive cells in the central retina (Figures 1c, d and 2), in contrast to controls in which Zn5 positive cells were restricted to peripheral germinal zone (Figure 2) (Raymond et al., 2006).No significant differences were observed in juvenile retinas (Figure 1e, f) at 1-mpf.Sox2 is a transcription factor essential in the eye formation.
Although it is known that the main function of Sox2 is maintenance of proliferative cells, its function in differentiated cells is still a matter of debate (DeOliveira-Mello et al., 2019;Whitney et al., 2014).Sox2 positive cells did not form the typical layered retina in embryos exposed to VPA at 2-dpf (Figure 1g.h).However, in the retina of larvae (Figure 1i, j) and juveniles (Figure 1k, l) this structure was apparently identical and layered in VPAexposed animals and controls.Zpr-1 antibody allows us to analyze photoreceptor formation, more specifically cone cells (Santos-Ledo et al., 2011).At 2-dpf, the retina of control embryos showed a ventral population of cells expressing Zpr-1 (Figure 1g) while VPA treated animals had only few cells at this stage (Figure 1h).In 5-dpf larvae, cone cells were well-formed in the outer nuclear layer of the retina of both VPA-exposed and control zebrafish (Figure 1i, j).5-HT functions as an important neuromodulator and neuroprotective factor in the retina and is synthesized and released by amacrine cells in the inner nuclear layer of the retina (Masson, 2019;Yazulla & Studholme, 2001).We sectioned larvae at 8-dpf in two histological planes to obtain reliable 3D results.Control larvae (Figure 3a, b) and VPA treated (Figure 3c, d) animals showed cells positive for 5-HT in the correct layer of the retina in a normal distribution.We found no statistically significant difference in the number of cells (Figure 3e) between the groups.

Delayed retinal development does not affect visual functions in juvenile fish
The optomotor assay demonstrated that 1-mpf fish showed a normal response to visual stimuli (Figure 4a).Zebrafish optomotor response consists of animal moving in the direction of moving strips.Normally, zebrafish swim in the direction of moving dark strips, referred to as optomotor response (Orger et al., 2000).Almost all fish (more than 85%) swam following moving strips as the applied protocol described (Orger et al., 2000).A small number of fish from both control (3 of 38 videos) and VPA (5 of 38 videos) groups displayed freezing behavior and were excluded from the analysis.Fish embryonically exposed to VPA swam in the direction of moving strips similarly with control fish.Distances covered by treated and control animals demonstrated no differences in visual acuity (Figure 5a).
In order to analyze color perception we performed a color preference assay, by measuring the time that the animals stayed on each side of the colored chamber (Figure 5b).Automatic tracking detector showed that VPA-exposed fish stayed on the yellow or red side for an equal amount of time in comparison with controls (Figure 5b, c).Both VPA-treated and control individuals showed a clear preference to the red side (Figure 5c).24 h locomotor activity and neurotransmitters are affected by embryonic VPA exposure 24 h locomotor activity assay revealed a hypoactive phenotype of VPA-exposed animals during the first light h of the test (Figure 6a).After 2 h in which no significant differences were found, treated animals became hyperactive in comparison with the controls (Figure 6a).When we analyzed the locomotor activity of both groups in 30 min bins during the dark period (subjective night), no significant differences were detected.However, sleep-like state of embryonically VPA exposure animals was significantly shorter than controls (Figure 5b).After nighttime, the lights were automatically turned on until the trial was finished.In this last period of light, fish of both groups increased their locomotor activity, with VPA exposed animals showing signs of hyperactivity (Figure 6a).
In order to examine potential causes of an altered sleep-like state in VPA-exposed larvae, we analyzed the F I G U R E 4 qPCR analysis of mRNA expression in the eyes of zebrafish larvae at 5-dpf.The expression of sox2 and th1 is upregulated in VPAexposed zebrafish larvae.No changes are detected in developmental (neurod1), serotonergic (sertb), glial (pax2a and apoeb) or apoptotic related (casp3a) genes.Data are expressed as mean ± SEM; the level of significance is set at *p < 0.05 and **p < 0.01.t-Student test was applied for the analysis.
pineal gland of control and VPA-exposed groups by examining cells immunoreactive to 5-HT (Figure 7a) and to the pineal cone-cell marker Zpr1 (Figure 7b) (Appelbaum et al., 2009;Ben-Moshe et al., 2014).These cells are photosensory pinealocytes that are involved in the conversion of 5-HT into melatonin, which means that photoreceptor-like cells (Zpr1 positive cells) are positive for 5-HT.Melatonin has major roles in different physiological processes, including circadian rhythmicity.It is important to mention that, besides the cells closely related to retinal cones (Zpr1-positive), the zebrafish pineal gland also possess rod-like cells and that they both contribute to circadian regulation (Ben-Moshe et al., 2014;Li et al., 2012).VPA-exposed fish showed a reduced number of 5-HT positive cells when compared with controls (Figure 7d).Curiously, the number of Zpr1-positive cell was not significantly different when VPA-exposed and control zebrafish were compared (Figure 7d), which could indicate that not all cone-like cells express 5-HT and/or that failure in 5-HT production is related to rod-like cells.We verified this by double staining of 5-HT and Zpr-1 expressing cells, which revealed that not all 5-HT immunoreactive cells express detectable Zpr-1 (Supplementary Figure S2).We also quantified the hcrt-positive cells in the rostral hypothalamus (Figure 7c), which are also closely related with sleep regulation (Appelbaum et al., 2009;Sundvik et al., 2011).The VPA-treated embryos also showed a reduced number F I G U R E 5 Visual behavior assays of zebrafish at 1-mpf.(a) In the optomotor assay, distances moved during the 30 seconds of strips video revealed that VPA-exposed animals and control fish follow strips without significant differences.(b, c) Both VPAtreated and control animals show a very similar preference between yellow and red zones.Data are expressed as mean ± SEM, n = 19 for optomotor test and n = 13 for color preference teste (per group), *p-value <0.05.
F I G U R E 6 Locomotor activity test during 24 h at 6-7 dpf.(a) Locomotion was evaluated in light during 10 hours, followed by a dark period of the same length (gray), followed by another 4 h light period.Analysis of distance moved in light showed significant differences between VPA-exposed and control larvae (asterisks).(b) Although tracking showed no differences in distance moved during the night, we found significant differences in the sleep-like state.We considered sleep-state as at least 60 s of no movement in a row and calculated how many minutes per hour VPA-treated animals and controls were immobile.VPA-exposed animals showed significantly less movement than controls in the night (b).Data are expressed as mean ± SEM, n = 20 in controls and n = 21 in treated animals, *p-value <0.05. of cells expressing hcrt at 6-dpf (Figure 7c, d).Interestingly, this significant reduction in the number of hcrtpositive neurons in the brains of VPA-treated embryos persisted until 2 weeks post fertilization (Figure 7e).

DISCUSSION
Exposure to VPA has been used to model ASD-like phenotypes in zebrafish.This approach generates several neurobehavioral impairments in these animals, including abnormal aminergic neurotransmitter systems and impaired social behavior (Baronio et al., 2018).Zebrafish are highly dependent on vision (Bollmann, 2019), and abnormalities in the eyes, impairments in perception of visual stimuli or visual pathways may produce abnormal responses that can affect social behaviors.Previous studies have demonstrated that animal models of ASD and patients with this disorder show visual impairments, but the roles that they play in their social conduct have been underestimated (Cheng et al., 2020;Guimarães-Souza et al., 2019).Social impairments related to the absence of the typical predisposed preferences for motion stimuli have been detected in VPA-injected chicks (Lorenzi et al., 2019;Sgadò et al., 2018).Our aim was to study the extent to which behavioral alterations of zebrafish embryonically exposure to VPA, a well-established model for autism research, are due to visual impairments.
In this study, we analyzed the developing retina of zebrafish exposed to VPA, which presented delayed retinal development at larval stage and altered gene expression in the eye.We found that visual behavior is normal fish embryonically exposed to VPA at larval stage.However, zebrafish embryonically exposed to VPA showed an abnormal pattern of activity during darkness and histological abnormalities in the pineal gland.
The visual system of zebrafish has been postulated as an ideal model to study vertebrate retina due to the high similarity in development and structure among vertebrate species (Bilotta & Saszik, 2001;Gramage et al., 2014).The stratified organization of the retina consists of nuclear and plexiform layers.The first cells to be specified in the retina are the ganglion cells at 24-hpf (Zn5 positive cells) (Bejarano-Escobar et al., 2013).The different cell types of the retina are sequentially specified while the optic nerve is colonized by glial cells expressing pax2 (Parrilla et al., 2009;Tao & Zhang, 2014).The growing axons of ganglion cells form the chiasm and finally reach the optic tectum at 48-hpf (Baier & Wullimann, 2021;Burrill & Easter, 1994).At 3-dpf, the retina is completely layered and starts growing from the peripheral zone at 5-dpf.The newborn neurons of the peripheral zone express Zn5 until their complete differentiation during the entire zebrafish lifespan (DeOliveira-Mello et al., 2019).Thus, differences in Zn5 positive cells could indicate alterations in retinal neurogenesis.We observed a delay in development of the optic nerve during chiasm formation stage (48 hpf) and in the neurons of the retina at 5-dpf.However, juvenile zebrafish exposed to VPA showed typical normal structure, in which Zn5 expression is restricted to new neurons in the periphery and their axons in the neural fiber layer.Further, expression of neuronal developmental genes (notch1a and neurod1) showed no significant differences between the groups.Overall, despite the delayed formation of neural cells in the retina, zebrafish exposed to VPA managed to reestablish the expected histological structure.More broadly, developmental delays associated with ASD-like phenotypes may affect physiological functions later in life as a consequence of e.g.abnormal neuronal connectivity in the CNS.
Sox2 is a transcription factor widely expressed throughout the presumptive neural retina (Yang et al., 2009) and it plays an important role in the visual system development.Absence of Sox2 in embryonic development is an important cause of microphthalmia or anophthalmia (Hever et al., 2006;Kautzman et al., 2018;Reis et al., 2010).Sox2 remains in a pool of stem cells in the peripheral zone after retinal formation, which is responsible for continuous life-long growth of the retina (DeOliveira-Mello et al., 2019;Easter & Nicola, 1996).In the differentiated retina, sox2 is expressed by various cell types and its functions are still a matter of debate (DeOliveira-Mello et al., 2019;Whitney et al., 2014).Our histological analysis did not show any clear difference in Sox2 positive cells between VPA-exposed and control zebrafish.However, sox2 mRNA was upregulated in the eyes of the VPA-treated animals, which could indicate differences in protein translation and/or that the qPCR technique is more sensitive than immunohistochemical detection, which is not quantitative.
Our results showed no differences in the expression of glial genes (apoeb, pax2a, and sox10) in the retina of VPA-treated animals in comparison with controls.The retina has a wide range of amacrine cell subtypes (Weruaga et al., 2000;Yan et al., 2020), including dopaminergic cells (Li & Dowling, 2000;Witkovsky, 2004).Curiously, expression of tyrosine hydroxylase (th1) was upregulated in VPA-exposed zebrafish eyes, suggesting that dopaminergic amacrine cells of the retina are overexpressing tyrosine hydroxylase enzyme as a consequence of VPA exposure.Previous studies have shown a downregulation of th1 in the brains of adult fish and whole larvae embryonically exposed to VPA (Baronio et al., 2018), which includes different populations of dopaminergic and noradrenergic cells.Further studies are needed to determine if dopamine from the retina and/or brain have a compensatory role to modulate the effects of VPA.
Several reports indicate that a subset of ASD patients have an impaired perception of environmental stimuli, including visual, tactile and auditory ones (Hsieh et al., 2022).This same phenotype is displayed by animal models of the disorder (Martin et al., 2022).Zebrafish models of ASD often display an exacerbated response to changes in illumination and acoustic/vibrational stimulus (Baronio et al., 2018;Santistevan et al., 2022).In mice prenatally exposed to VPA submitted to the tail-flick test, the latency to avoid a noxious stimulus was significantly longer than in control mice, indicating an elevated nociceptive threshold (Baronio et al., 2015).Innate color preference or aversion enables the recognition of the environment and can provide information on how to appropriately respond to environmental stimuli.We found that photoreceptor cells responsible for color perception (cones; Zpr-1 positive cells) have a developmental delay at 48-hpf as a consequence of VPA exposure.Nevertheless, in zebrafish larvae, histological analysis showed that cones are in the correct layer and structurally similar in VPA-exposed and control animals.We also evaluated the response of VPA-exposed zebrafish and controls to different colors that have been reported to be attractive and aversive to this model organism, namely red and yellow, respectively.The purpose of this test was also to verify if the transient developmental delay following VPA exposure on the zebrafish retina affected the color vision of these animals.However, both groups displayed a similar behavioral phenotype during the test, showing a significant preference for the color red.
Sleep disorders, such as insomnia, are common in patients with ASD (Lin et al., 2022).Zebrafish embryonically exposed to VPA showed a hypoactive phenotype at the beginning of 24 h locomotor tracking assay.A recent study showed abnormalities in sleep fragmentation of VPA treated zebrafish at 6-dpf (Robea et al., 2021).They analyzed the 24 h locomotor activity in 1 h bins, and used a different timing of exposure to VPA.Interestingly, our analysis showed that differences in locomotor activity between VPA treated and control larvae were suppressed after 3 h, most likely due to a reduction in activity by the controls and an increase in the activity of VPA-treated larvae as the nighttime was approaching.The 30 min bin analysis showed no overall difference between VPAexposed larvae and controls in total distance moved during the nighttime.However, when analyzed in periods of 1 min of immobility as a sleep-like state VPA-exposed larvae were more active than control larvae.
Hypocretin innervation modulates melatonin production in zebrafish (Appelbaum et al., 2009).Interestingly, we detected a reduced number of hcrt expressing cells in the brains of VPA-exposed larvae that persisted until 2 weeks post fertilization, as well as a reduced number of 5-HT-immunoreactive cells in the pineal gland, which could indicate a deficit in melatonin production (Lee et al., 2021).This hypothesis corroborates the behavioral pattern of VPA-exposed larvae during 24 h tracking, which displayed a reduced sleep-like state in comparison with controls.Consistent with our results, there are publications that used genetic zebrafish ASD models (Rea & Van Raay, 2020), including models that showed sleep disturbances (Ruzzo et al., 2019).However, further studies are necessary to analyze the hcrt expressing cells and the pineal gland of these gene-modified animals.Our results highlighted that VPA-exposed zebrafish can serve as a relevant model for further studies on sleep disturbances of the ASDs.In the retina, 5-HT acts as a neuromodulator and is synthesized and released by a subtype of amacrine cells, which are responsive to light stimulus and play a role in circadian cycle regulation as they express clock genes (Masson, 2019).We did not find differences in the number of these cells or their localization in the retina of the VPA treated and control animals, or in the expression of the paralogue b of serotonin transporter (sertb).
In summary, the present study shows that embryonic exposure to VPA in zebrafish affects retinal development during the larval stage.However, this seems to be a transient effect as no significant differences were detected between the VPA group and control fish in immunohistochemical analysis of the retina and behavioral assays that evaluated visual perception at later stages of development.These findings may indicate that the behavioral impairments detected in adult animals, such as impaired social interaction, are independent of an abnormal visual perception.Interestingly, VPA exposure affected the pineal gland, hcrt system and sleep-like behavior in zebrafish.The results support the concept that several central nervous systems and pineal mechanisms contribute to the behavioral abnormalities associated with VPA-induced ASD-like condition in zebrafish.

F
I G U R E 1 Visual system analysis of VPA-exposed and control zebrafish.(a, b) Neurolin (Zn5) staining shows a delay in development of the retina and the optic nerve (white arrows) of embryos (2-dpf).(c, d) The retina of VPA-exposed larvae (5-dpf) was underdeveloped in comparison with control larvae, showing a flatten structure (white arrows).(e, f) A complete recovery of forming neurons can be seen in juvenile (1-mpf) samples, where the Zn5-positive cells were restricted to peripheral retina and its axons pass thought the neural fiber layer until form the optic nerve.(g-j) Photoreceptor development by using Zpr1 antibody (green) in embryos and larvae.(g-l) Staining for Sox2 transcription factor showed positive cells in the layered retina in embryos (g, h) larvae (i, j) and juveniles (k, l).Sox2 positive cells in the PGZ and in different layers of differentiated retina has been previous described(DeOliveira-Mello et al., 2019;Santos-Ledo et al., 2022) and was similar in VPA-exposed zebrafish and controls.(g, h) Cones showed a delay in embryos (2-dpf) exposed to VPA (white arrow; magnification in h').(i, j) Recovered structure of cone cells in larvae (5-dpf) exposed to VPA. (k, l).In juveniles (1-mpf), the structure of VPA-treated animals and controls are very similar.PGZ, peripheral germinal zone.Scale bar 50 μm.

F
I G U R E 2 Developing optic nerve and retina in larval zebrafish at 5-dpf.(a, b) Developing neurons expressing Zn-5 at 5-dpf were restricted to peripheral retina in controls (white arrows), but were extended in entire retina of VPA-exposed zebrafish.(d) Although the intra-orbital optic nerve length showed no differences in VPA-treated larvae in comparison with controls, the length of the optic nerve head of larvae exposed to VPA was smaller than in control larvae.Data are expressed as mean ± SEM, n = 5 per group, *p-value <0.05.F I G U R E 3 Serotonergic neurons in the retina of larvae zebrafish at 8-dpf.(a, d) Immunostaining for 5-HT (green) in sagittal and horizontal sections of larval (8-dpf) retinae.5-HT positive cells (green) are seen in the inner nuclear layer of the retina in both VPA-treated and control larvae in a similar distribution.(e) Quantification of the number of cells per section showed no statistical differences between treated animals and controls.Neurons were counterstained with HuD/HuC label (red).Data are expressed as mean ± SEM, n = 5 per group, *pvalue <0.05.5-HT, serotonin.Scale bar 50 μm.

F
I G U R E 7 Histological analysis of pineal gland and hypocretin/orexin (hcrt) neurons.(a, b) Serotonin-immunoreactive (5-HT) cells (a) and photoreceptor-like (Zpr1) cells (b) in the pineal gland were studied at 5-dpf of VPA-exposed and control larvae.(c) Hcrt positive neurons of VPAexposed and control larvae in rostral hypothalamus were analyzed at 6 dpf.(d) Quantification of the number of cells in a, b,and cindicates that VPA-exposed animals have significantly fewer serotonin-immunoreactive cells in the pineal gland and cells expressing hcrt in the brain than control larvae, while photoreceptors in the pineal are apparently normal in VPA treated larvae.(e) Hcrt positive neurons in rostral hypothalamus were also counted at 14-dpf, and VPA-exposed larvae had still fewer cells expressing hcrt in comparison with control larvae.Data are expressed as mean ± SEM, n = 6 in larvae at 5 and 6-dpf and n = 9 in 14-dpf zebrafish, *p-value <0.05.5-HT: serotonin; Zpr1: photoreceptors; Hcrt, hypocretin/orexin.Scale bar 50 μm.
List of antibodies used.
List of primes used for qPCR.
T A B L E 2 Behavioral data was analyzed by 2-way measures ANOVA followed by Sidak's post hoc test.Results are shown as means ± SEM and p-value <0.05 was considered statistically significant.
3.1).The normality of data was tested using Shapiro-Wilk test.For data with normal distribution, we applied Student's t test, and for non-parametric analyses we used Mann-Whitney-U test (applying Bonferroni correction to multiple comparisons).