Developmental exposure to ethylbenzene: Elucidating the effects of volatile organic compounds on phenotypic and transcriptomic endpoints using zebrafish (Danio rerio)

Abstract

Volatile organic compounds (VOCs) are ambient anthropogenic pollutants prevalent in urban environments; sources include vehicle exhaust and industrial emissions. Ethylbenzene, a ‘possible human carcinogen’ is of public health concern due to high levels in residential drinking water wells near waste sites and leaking underground sources entering indoor air through vapor intrusion. Exposure has also been linked to non-cancer adverse health outcomes, yet few studies focus on the immediate effects of ethylbenzene exposure. In this study, we utilized the zebrafish model to carry out controlled VOC exposures during the critical window of early development. Our novel, validated method using sealed glass vials was employed for optimized chemical delivery. Phenotypic data collection was conducted following 0.001-10 ppm ethylbenzene exposure after 5 days post fertilization (dpf) and tissue samples were collected for transcriptomic endpoints. Behavioral assays were carried out at 1 and 5 dpf to characterize very early signs of neurotoxicity. At 5 dpf, exposure to ethylbenzene was significantly associated with a decrease in larval behavior at almost every concentration (p<0.001) during the dark periods and significantly associated with a decrease in behavior at 10 ppm (p<0.001) in the light periods. Mortality and swim bladder inflation were significantly affected at 1 and 10 ppm ethylbenzene. Dysregulated genes (including crtc2 and pdca) and pathways of interest are associated with carcinogenesis, the endocrine systems, neurological system, and reproductive system. The high conservation of molecular pathways between zebrafish and mammals provides further insight into potential mechanisms and adverse health impact associated with volatile organic compound exposure.

Graphical Abstract

1. Introduction

Volatile Organic Compounds (VOCs) are widely distributed environmental pollutants with industrial, commercial and biogenic origins. VOCs are commonly released as byproducts of fuel production, building furnishing coatings and industrial run-off [1]. They are also emitted anthropologically from vehicles and coal or wood combustion [1]. Exposure to VOCs amongst general populations can occur in the home or workplace and vary in duration, concentration and route of exposure. The use of common household products like paint, perfume or cleaners can lead to the release of these harmful chemicals [1]. VOCs are a broad category of chemicals, described by the EPA as an organic compound which “participates in atmospheric photochemical reactions” [2]. A subset of classifications exists within VOC compounds, defined by some as being dependent on their boiling point and by others the ease in which they’re emitted [2]. Of these carbon-containing compounds, the second most persistent group in the environment consists of benzene, toluene, ethylbenzene, and xylenes, commonly referred to as BTEX [3]. Amongst the various existing VOC subgroups, BTEX compounds are found and characterized relatively commonly in environmental studies of VOCs, likely due to their ubiquitous nature, while the developmental effects of these individual components remain largely understudied [3]. BTEX exposures are known to lead to carcinogenic outcomes, neurological impairment, and other adverse health effects [4]. Detection and treatment methods of BTEX-contaminated waters have improved, and novel methods have been developed in recent years to combat the ever-present nature of these chemicals, and yet, they persist [1].

Ethylbenzene, a singular component of the BTEX tetrad, is an aromatic hydrocarbon that presents as a colorless liquid [5]. It is described as possibly carcinogenic by the National Institutes of Health (NIH) and the Environmental Protection Agency (EPA) acknowledges the limited information available on the carcinogenic effects of ethylbenzene [4,6]. Ethylbenzene naturally occurs in petroleum products and is mainly used to produce styrene [6,7]. Commercial products such as carpet glue, pesticides, tobacco and gasoline all contain ethylbenzene [6]. Ethylbenzene persists across all four spheres of the environment- in the air, water, soil and living organisms. Recent studies conducted in an urban environment (Detroit) have detected ethylbenzene levels in soil (0.009 ppm) and sewer vapor (5.1 ppm) that were higher than the EPA’s risk-based screening level [8]. Once in the environment, ethylbenzene is more likely to be absorbed through inhalation, ingestion, or percutaneously. However, the main exposure route of ethylbenzene exposure for humans is inhalation of ambient air [4]. In fact, early toxicokinetic studies have found that when inhaled at high concentrations (23–85 ppm) a small sample was shown to retain 60% of the inspired vapor [9]. Percutaneous absorption of ethylbenzene in its liquid form can also occur and is more likely in occupational settings than ambient or household settings [6]. According to the EPA, a lifetime exposure to 0.7 ppm ethylbenzene (Maximum Containment Level) in drinking water is not expected to cause any harmful effects [5]. However, the known effects of ethylbenzene are dependent on length of exposure and the concentration.

Toxicant metabolism data is vital to understand likely health effects of ethylbenzene exposure but thorough epidemiological and toxicological studies are also necessary to categorize public and human health risk. In the early 2000s, it was established that the major enzyme responsible for hydroxylation of ethylbenzene in the human liver is CYP2E1 from the cytochrome P450 (CYPP450) superfamily and subsequently, mandelic acid and phenylglyoxylic acid are the excreted urinary metabolites [9-11]. Though human data are limited, one toxicology study conducted inhalation exposures at high exposure levels in mice, rats, and rabbits and established no observed adverse effect level (NOAEL) of 382 ppm (rats and mice) with a lowest observed adverse effect level (LOAEL) of 782 ppm [12]. This study also saw significant increases in liver weights in rats exposed to high levels of ethylbenzene [12]. Additionally, a study in rats found that gestational exposure to high levels of ethylbenzene (100-2000 ppm) resulted in no evidence of teratogenic effects but did observe fetal toxicity evidenced by significant decreases in fetal body weights at 1000 ppm exposure level [13].

Early studies have been important for establishing the knowledge base on ethylbenzene, however, a focus on environmentally relevant exposure levels should be prioritized moving forward given the possibility that VOCs elicit effects at sub-cancerous (low) levels in other animal models [14]. There is currently no literature on the effects of early ethylbenzene exposure during key developmental windows, or that considers the subsequent reproductive health of developmentally exposed organisms at sexual maturity and fate of their offspring. While there are data showing inhaled ethylbenzene in immature animals is linked to small decreases in weight gain [5], there is a lack of recent data on low-level effects of developmental exposure to various levels of ethylbenzene, necessitating the use of a high-throughput animal model to address immediate, adult-onset and transgenerational adverse health effects.

The purpose of this study is to analyze phenotypic and transcriptomic changes and their associated diseases and pathways due to a controlled, wide range of environmentally relevant ethylbenzene exposures (.001, .01, 0.1, 1 and 10 ppm) during early embryonic development. This study uses zebrafish as a high-throughput animal model, allowing for immediate studies in early development, lifespan, and transgenerational studies in reproductive health. Since zebrafish shared 70% homology with humans, these results will be highly translational to human health effects due to environmental ethylbenzene exposure. The sealed-exposure experimental design as previously described will improve and stabilize dissolved ethylbenzene content, generating more consistent and exact exposure doses. The results of this study will have implications for human health and aid in the development of more rigorous guidelines for regulating ethylbenzene levels in our environment.

2. Materials and Methods

Animal Husbandry

Adult AB/TU strain zebrafish were housed in buffered reverse osmosis (RO) water (60 mg/L Instant Ocean Salts; Aquarium Systems, Mentor, OH, USA) at 28 °C on a recirculating system at a maximum density of 5 fish per liter with a standard light/dark cycle (14/10 h). The zebrafish were fed a mixture of Zeigler Adult Zebrafish Diet (Zeigler Bros. Inc., Gardners, PA, USA), Spirulina Flake Fish Food (Ocean Star International, Snowville, UT, USA), 300–500-micron Golden Pearls (Aquatic Foods Inc., Fresno, CA, USA) twice daily, and supplemental brine shrimp (Artemia International, Fairview, TX, USA) once daily. The animal use protocols were approved by the Institutional Animal Care and Use Committee at the University of Florida according to the National Institutes of Health Guide to the Care and Use of Laboratory Animals (Protocol No. #IACUC202200000046).

2.1. Spawning

Zebrafish embryos were obtained by spawning adult AB/TU zebrafish. These spawns were initiated the evening prior to ethylbenzene exposure at a ratio of either 2 females:1 male, 2 females:3 males, or 1 female:1 male in 1.0 L plastic spawning tanks (Aquaneering Inc., San Diego, CA, USA) with plastic dividers separating the sexes. The dividers were removed the following morning, and embryos were collected 2-2.5 hours post-fertilization (hpf).

2.2. Egg Collection

The embryos were rinsed thoroughly with RO water to remove fish waste and soaked in a dilute bleach solution (RO water with 0.6 g/L Instant Ocean Salts and 600 uL/L of 6% sodium hypochlorite) for 10 min. The embryos were then rinsed with RO water to remove any remaining bleach solution. Once cleaned, all embryos were moved to a beaker with egg water (RO water with 0.6 g/L Instant Ocean Salts) and mixed to encourage randomization. Then, the randomized embryos were put into Petri dishes containing egg water and sorted based on viability. Viable embryos were transferred into 30 mL clear glass vials containing 20 mL of egg water at a density of 30 embryos per vial. The exposure included a total of 30 vials per concentration (n = 900 embryos per concentration).

2.3. Ethylbenzene Exposures

Five exposure concentrations (0.001 ppm, 0.01 ppm, 0.1 ppm, 1 ppm and 10 ppm) were made from a stock solution of 99.8% pure ethylbenzene (Ethylbenzene, 99.8%, pure, Thermo Scientific Chemicals, Waltham, MA, USA). The stock and dilution concentrations were analyzed by GC-MS/MS (TSQ 9610 GCMSMS AEI NV) at the Lumigen Instrument Center, Wayne State University, Detroit MI (Supplemental Table 1). The 99.8% ethylbenzene stock was diluted in a septum-sealed amber vial with 40 mL egg water to create a standard solution of 170 ppm ethylbenzene. This standard was then used to make the 10 ppm and 1 ppm ethylbenzene stock dilutions, which were then spiked into vials in varying volumes to create the exposure concentrations (10 ppm, 1 ppm, 0.1 ppm, 0.01 ppm and 0.001 ppm). In preparation for ethylbenzene exposure, egg water was removed from the vial of embryos to balance the volume of stock dilutions that were injected into vials with a gastight syringe for final concentrations. Control vials were re-injected with clean egg water to control any potential effects of exposure administration. To minimize ethylbenzene loss through volatilization, the ethylbenzene stock dilutions were mixed fresh daily, and the caps were never removed from the septum-sealed vials during or after ethylbenzene injection into vials housing embryos (i.e., exposure vials). The control vials contained 20 mL egg water and the exposure vials totaled 20 mL egg water and ethylbenzene solution. The control and exposure vials were then gently rocked by hand and placed in a 28 °C incubator. Over the next 4 days, this procedure was repeated daily for each vial to refresh the exposure dilution or egg water. In addition to daily water changes, embryos/larvae were screened daily, and any that were dead (as indicated by coagulation of the embryos or a lack of heartbeat in larvae) were removed. At 5 dpf, embryonic development was completed, and the larval fish were rinsed with egg water 3 times to end the exposure. The larval fish were then moved to a recirculating system to be raised past sexual maturity (>4 months post-fertilization) under standard husbandry conditions. Replicates were carried out in the larval exposures consisting of at least 3 separate biological replicates with exposures that were conducted identically for each replicate. Exposure paradigm used has been previously described [14-16]

2.4. Spontaneous Tail Coiling

Spontaneous tail coiling (STC) was measured at 23 ± 2 hpf to evaluate early motor activity and developmental neural activity. 60-second videos were recorded and segmented into image stacks at a rate of 2 frames/second using FFmpeg software. STC frequency for embryos with a visually developed tail was analyzed using a workflow developed in KNIME Analytics Platform [17]. A subset of videos was manually counted. Results are reported as average STC count per minute, and statistical significance was evaluated using a Kruskal-Wallis test. Samples used for the STC endpoint were excluded from further assays.

2.5. Abnormality, Mortality and Behavioral Screening

Zebrafish embryos were screened daily for mortality throughout the 5-day exposure period. Survival data is reported as percent mortality each day, and data was analyzed using a Chi-square test with Bonferroni correction. On day 5, the following morphological abnormalities were assessed for live larvae: numbers of unhatched embryos, skeletal deformities, lack of inflated swim bladder in hatched larvae, yolk sac edema, and cardiac edema. Incidence of abnormalities was analyzed using Fisher’s exact test with a Bonferroni correction in R. Then, fish with inflated swim bladders and no morphological abnormalities were selected for behavioral testing. For each exposure concentration and control group, larval fish were divided evenly into 24-well plates, with each well housing a single fish in 2 mL of egg water. The plated fish were allowed to acclimate at 28.0° C for at least one hour, then placed into a ZebraBox Revolution (Viewpoint Life Sciences, Montreal, QC, Canada) for a 36-minute behavioral assay. After a 12-minute acclimation period inside of the ZebraBox, an alternating 3-minute light and 3-minute dark cycle commenced and repeated 4 times. Distance traveled was tracked inside the behavioral chamber using the ZebraLab automated fish behavior software (Viewpoint Life Sciences, Montreal, QC, Canada). Movement data was collected between 14:00 and 19:00, and average distance moved (cm) was analyzed using one-way ANOVA followed by Tukey HSD in R.

2.6. Tissue Collection

Tissue was collected from exposed larval zebrafish at 5 dpf. First, larvae were euthanized in a solution of 0.4 g/L tricaine methane sulfonate (MS-222) buffered with 0.66 g/L sodium bicarbonate. Then, 5 whole larvae were pooled per sample for each exposure concentration and control group. The larval tissues from these 6 conditions (5 exposure groups and 1 control group) were immediately deposited into 300 μL RNAlater (Invitrogen, Carlsbad, CA, USA) for tissue preservation and stored at 4°C for up to one week. Then, the RNAlater was removed, and the samples were stored at −80 °C for downstream analysis.

2.7. RNA Isolation

Prior to isolation, the samples were removed from the −80 °C freezer and allowed to thaw briefly on ice. A lysis buffer, Buffer RLT (Qiagen, Hilden, Germany) and DL-1,4-Dithiothreitol (Thermo Scientific Chemicals, Waltham, MA, USA) was added to the samples before promptly proceeding to isolation steps. For RNA isolation, the RNeasy^® Mini Kit (Qiagen, Hilden, Germany) was used according to the manufacturer’s instructions. The Qubit^® 3.0 Fluorometer and Qubit^® RNA Broad Range Assay Kit (Invitrogen, Carlsbad, CA, USA) were used to determine the RNA concentrations. RNA concentrations and RINS are reported in Supplemental Table 2. For further quality assurance, following TapeStation the RNA samples were then kept at −80 °C until QuantSeq library preparation.

2.8. Transcriptomic Analysis

Next, 3′ mRNA-seq libraries were prepared from isolated RNA using the QuantSeq FWD 3′ mRNA-Seq V2 w/ UDI 12 nt Set B1 Library Prep Kit and the PCR Add-on Reamplification Module (Lexogen, Vienna, Austria), according to the manufacturer’s instructions. The samples were amplified at 18 cycles. The libraries were quantified using a Qubit^® 3.0 Fluorometer and Qubit^® 1X dsDNA HS Assay Kit (Invitrogen, Carlsbad, CA, USA), and run on an Agilent TapeStation 4150 (Agilent Technologies, Santa Clara, CA, USA) for quality control. The samples were sequenced on a NovaSeq 3298 (Illumina, San Diego, CA, USA) in rapid mode (single-end 75 bp reads). The sample size was n = 5. Reads were aligned to D. rerio (Build danRer11) using STAR via the Kangooroo Platform (Lexogen, Vienna, Austria). The differential gene expression between the control and exposed zebrafish was evaluated using DEseq2 (available through GenePattern; Broad Institute, Cambridge, MA, USA). Genes with significant changes in expression, as defined by an absolute log2 fold change value ≥ 0.5 and p – value < 0.05, were uploaded into Ingenuity Pathway Analysis software version 22.0.2 (IPA; QIAGEN Bioinformatics, Redwood City, CA, USA) for analysis using RefSeq IDs as identifiers. The complete list of significant gene data is included in Supplemental Table 3. The significant canonical pathways and associated diseases and biofunctions were defined as having p– values < 0.05, (or −log10(p– value) > 1.3) based on a right-tailed Fisher’s exact test. A right-tailed Fisher’s Exact test is justified because of unequal proportions between groups, finite populations, and because we are sampling without replacement. The z-score was calculated through IPA functionality to predict the activation (≥2) or inhibition (≤−2) of a pathway.

3. Results

Significant increases in mortality were observed over the course of the 5-day exposure. At 2-dpf, exposure to ethylbenzene at 1 ppm was associated with increased mortality (p < 0.01). At 5 dpf, exposure to 0.01 (p < 0.01), 1 (p < 0.001), and 10 (p < 0.01) ppm was associated with increased mortality (Figure 1).

Figure 1.

Percent of zebrafish larvae surviving (total dead/total alive) at 1 – 5 days post fertilization (dpf) following ethylbenzene exposure (0.001 - 10 ppm). * Indicates p < 0.05 based on Chi-Square test with Bonferroni correction.

For phenotypic responses, cardiac edema, yolk sac edema, skeletal deformities, and swim bladder inflation were evaluated with no significant response observed (p > 0.05). Exposure to ethylbenzene was associated with a significant increase in the number of fish with uninflated swim bladders at concentrations of 1 (p < 0.01) and 10 (p < 0.001) ppm (Figure 2a). No statistically significant changes in spontaneous tail coiling at 1 dpf were observed at any concentration of ethylbenzene (Figure 2b).

Figure 2.

A) Percentage of zebrafish larvae with morphological abnormalities per concentration at 5-days post fertilization (dpf) following exposure to ethylbenzene. * indicates p < 0.05 based on Fisher's Exact Test with Bonferroni correction. B) Average spontaneous tail coiling per minute per ethylbenzene concentration in zebrafish at 1 day post fertilization (dpf). Error bars represent standard error.

Exposure to ethylbenzene was associated with a significant decrease in larval locomotion (distance traveled) at 0.001, 0.01, 0.1 and 10 ppm (p < 0.001) during the dark cycles and with a decrease in locomotion at 10 ppm (p < 0.001) in the light cycles (Figure 3).

Figure 3.

Distance traveled (cm) during 3-minute light/dark intervals for ethylbenzene-exposed (0.001 - 10 ppm) zebrafish larvae at 5 –days post fertilization (dpf). *** indicates p<0.001 based on one-way ANOVA with Tukey’s HSD tests. Error bars represent standard errors.

Larval fish tissue RNA concentrations averaged 168 ng/uL (standardized input 150 ng) with RNA integrity numbers (RINs) above 7 (average 9.1) (Supplemental Table 2). Significant perturbation in gene expression was evident in all ethylbenzene exposure groups as compared to the control group (Supplemental Figure 1; Supplemental Table 3). The highest number of differentially expressed genes (DEGs) were present in the highest exposure group (10 ppm) and the lowest number of DEGs present in the 0.01 ppm exposure group (Table 1). For every group, differential gene expression was skewed towards upregulation.

TABLE 1.

Total differentially expressed genes (DEGs) for each ethylbenzene exposure group as compared to the control group with p value <0.05 and log2 fold change greater than ∣0.5∣. DEGs stratified by direction of change for total over- or under-expressed.

Differential Gene Expression
Concentration (ppm)	Total # DEGs	# Upregulated	# Downregulated
0.001	74	48	26
0.01	57	30	27
0.1	69	39	30
1	79	41	38
10	138	84	54

While there were no genes consistently altered in every exposure group, there were two genes (pdca, crtc2) that were changed in every group except the 0.01 ppm ethylbenzene exposure group. This was the highest degree of overlapping effects across groups that we observed in the larvae as shown in Figure 5.

FIGURE 5.

Upset Plot showing the frequency and overlap of differentially expressed (DE) genes and gene sets after exposure to ethylbenzene with p – value <0.05 and log2 fold change greater than ∣0.5∣ [18]. The inset table identifies the two genes in the set highlighted.

Based on pathway analysis conducted on all significant DEGs, cancer was a commonality among the top disease and disorder pathways across all 5 exposure groups. Included in the top 5 of all groups was organismal injury and abnormality, and in all groups except the 0.01 ppm exposure group, we observed disruption to pathways linked to endocrine system disorders.

Pathway analysis was then broken down from disease and biological function categories to annotated functions based on the DEGs involved in associated pathways. These pathways were largely associated with carcinoma, tumor development, and cancer cell line proliferation. The most consistent disease pathway across exposure groups was extracranial solid tumor which was increased, except in the 0.01 ppm group (Figure 6). The highest exposure condition had the largest number of enriched pathways, many with high enough z-scores to predict activation of the pathways. Pathways involved in endocrine system associated functions were also changed, especially at the highest and lowest exposure levels. Resulting from analysis of canonical pathways, the actin cytoskeleton signaling, AMPK signaling, tight junction signaling, and Vitamin D (calciferol) metabolism were significantly altered ( −log (p–value) > 1.3). Actin cytoskeleton signaling was altered in the three highest doses while AMPK signaling, tight junction signaling, and vitamin D metabolism pathways were changed in the two highest concentration groups (1 ppm, 10 ppm). LXR/RXR activation, acute phase response signaling, and xenobiotic metabolism signaling were all significantly changed in multiple exposure groups. All three of the aforementioned pathways were changed in the 10 ppm group, with LXR/RXR activation also in 0.001 group, xenobiotic metabolism signaling changed in the 0.1 ppm group, and acute phase response signaling was changed in all three (0.001, 0.1, and 10 ppm) ( −log (p–value) > 1.3).

FIGURE 6.

Heatmap shows values for ethylbenzene-exposed larval zebrafish for each exposure concentration (ppm), displaying the statistically significant pathways linked to diseases and biological functions for cancer, neurological system, organismal injury and abnormality (OI&A), endocrine system, and cellular development and functioning. Colored boxes indicate z – score while grey boxes indicate significance without predictive action (z-score).

Some key genes were differentially expressed across multiple exposure groups or were altered to a high degree which are of interest for this analysis. Examples of genes that were changed in three exposure groups were si:ch211-241e1.3 and otx2a. Other genes were shown to have relatively high fold changes compared to most other genes in this dataset (cln3, actn2b, ppp1R27, trib1) or were active in multiple pathways (Table 2). The AMPK signaling pathway results include changes to both crtc2 and ppp1r27a while map3k4 is active in the xenobiotic metabolism signaling pathway. Additionally, the gene cubn is involved in the vitamin D metabolism pathway that was also altered due to ethylbenzene exposure.

TABLE 2.

Differentially expressed genes (DEGs) of interest with p – value <0.05. Log2 fold change indicated by level of ethylbenzene zebrafish larvae were exposed to.

Dysregulated genes
Gene symbol	Gene name	0.001 ppm	0.01 ppm	0.1 ppm	1 ppm	10 ppm
pdca	phosducin a CREB regulated transcription	−0.52636		−0.57617	−0.59888	−0.6514
crtc2	coactivator 2	0.5502		0.519583	0.52787	0.55378
si:ch211-241e1.3	si:ch211-241e1.3	0.53646		−0.55061	−0.54041
otx2a	orthodenticle homeobox 2a		0.5513	−0.52857	−0.53579
actn2b	actinin, alpha 2b	−0.71166		0.518015
rnaseh1	ribonuclease H1 heat shock			−0.50566		−0.5215
hsf2	transcription factor 2			−0.50249
cubn	cubilin				0.55894	0.57584
tube1	tubulin, epsilon 1 natriuretic peptide	−0.60119
npr1a	receptor 1a CLN3 lysosomal/endosomal transmembrane	0.56221
cln3	protein, battenin pbx/knotted 1	0.80465
pknox2	homeobox 2		−0.5525
rfx4	regulatory factor X, 4 solute carrier family		−0.6159
slc13a5a	13 member 5a GRB10 interacting		0.61472
gigyf1a	GYF protein 1a mitogen-activated protein kinase kinase		0.58591
map3k4	kinase 4 vitamin K epoxide reductase complex,			0.609103
vkorc1l1	subunit 1-like 1			−0.50403
ano6	anoctamin 6 hydroxyacid oxidase				0.51364
hao1	(glycolate oxidase) 1 protein phosphatase 1, regulatory subunit				−0.54031
ppp1R27a	27a apolipoprotein A-IV					0.8795
apoa4b.3	b, tandem duplicate 3 tribbles					0.61343
trib1	pseudokinase 1					0.84617

4. Discussion & Conclusion

This is the first study to our knowledge to use zebrafish as a toxicological model for ethylbenzene exposure. The study of commonplace VOCs continues to be necessary to contextualize ambient and occupational exposures, especially where it involves critical windows of development (i.e. during pregnancy) which program health and disease outcomes for the offspring. Our results show that developmental exposure to ethylbenzene levels as low as 0.001 ppm elicited transcriptomic changes, and environmentally relevant levels induced phenotypic abnormalities as well as mortality in concentrations as low as 0.01 ppm.

Among phenotypic responses, mortality and swim bladder abnormality were statistically significant at and above 1 ppm. For behavioral phenotypes, lower exposure level (0.001 ppm, 0.01 ppm, and 0.1 ppm) effects were observed in addition to high (10 ppm) level effects. It is accepted that the molecular drivers of swim bladder inflation are not well defined, and any neurological or locomotor impairment can prevent the organism from reaching the water surface to gulp air and ultimately lead to this abnormality [19]. Swim bladder inflation is a landmark in proper zebrafish development and is vital to organism survival [20]. In turn, when these developmental landmarks are not hit, whether due to environmental stress or neuromotor deficits, other systems can show subsequent and notable effects like altered response to stimuli or elevated mortality. Consistent with this theory, the two most affected exposure groups also had the highest number of dysregulated genes overall. Zebrafish survival was also significantly decreased in the two groups exhibiting swim bladder effects (1 and 10 ppm) with the addition of day 5 decreased survival in the 0.01 ppm exposure group. Swim bladder abnormalities have important implications for proper development and specifically thyroid system development and function [21]. Surprisingly, thyroid receptor genes (thraa and thrab) were not significantly changed. While swim bladder is likely not the sole reason survival was decreased, it is an indicator that normal development, and likely, neurodevelopment was disrupted due to ethylbenzene exposure.

The neurological and cognitive effects often seen associated with VOC exposure necessitated multiple behavior and motor function assays [22-24]. Zebrafish behavioral response assays are a robust method to characterize toxicogenetic results and provide a scaffold for categorizing brain function. To capture immediate-onset effects, we conducted a spontaneous tail coiling (STC) assay within 1 day after ethylbenzene exposure. STC showed no significant change, though responses trended non-monotonically, likely reflecting the relatively low-level and brief exposure compared with the robust effects observed at 120 h. Future work could employ a different early photomotor response assay to probe stimulus-evoked behavior and clarify ethylbenzene’s early-life neurodevelopmental impact. The second behavioral assay was completed at a later stage of development and tested zebrafish locomotor activity following a light or dark stimuli. Zebrafish are more active in the dark period but here we noticed significant hypoactivity in both the dark and light periods. This was evident for every group except the 1 ppm group in the dark and for the 10 ppm group only in the light. This is an interesting finding and suggests ethylbenzene may impair normal behavior patterns which could be indicative of suppressed long-term survival. Behavioral assays such as the locomotor assay demonstrated here have been used to model complex brain disorders and developmental delays in learning and memory making them pivotal in elucidating long term-effects of changing environmental conditions [25]. Especially important for translational research, zebrafish neurobiology is largely conserved with mammals [26-28]. Long term follow up with the addition of behavioral endpoints later in life would further improve our understanding of ethylbenzene-induced phenotypes and future studies will focus on rearing these cohorts to adulthood to evaluate lasting or latent effects across the lifespan.

The pathway analysis revealed that pathways involved in cancer were the most affected at every exposure level. This was not unexpected as there is enough animal toxicity data to classify ethylbenzene as a carcinogen and potentially carcinogenic to humans (group 2B) [29]. This validates our findings, especially when considering mechanisms of action. Organismal injury and abnormality pathways are biological functions with effects that largely impact the whole organism and are generally observable effects that often, but not always, correlate to phenotypes that we have observed- such as skeletal deformities in benzene exposed larvae [15]. In addition, endocrine system disorders were on the list of top altered pathways. This adds to the increasing but incomplete evidence that VOCs (separately or in tandem with personal care products) can have endocrine disrupting properties [30]. This has also been considered for BTEX constituents specifically linked to disruption of hormone signaling along the CNS-pituitary-ovarian axis [31].

Based on analysis of canonical pathways, actin cytoskeleton signaling, AMPK signaling, tight junction signaling, and Vitamin D (calciferol) were consistently perturbed. The actin cytoskeleton signaling pathway plays a vital role in cellular motility, shape, cytokinesis, and phagocytosis [32]. AMPK signaling is a key regulator of cell energy and metabolism and is involved in stress response [33]. Further, upregulation of AMPK pathway inhibits cellular death and stress in β cells which is important for preventing the development of diabetes [34]. It is evident that AMPK and insulin secretion are intertwined and highly important for the endocrine and reproductive systems. Insulin receptors are widely distributed in ovarian cells and there is strong evidence for the direct effect of insulin on steroid production and the importance of insulin-signaling pathways in ovulation control [35,36]. These findings in conjunction with a growing body of evidence suggesting a link between BTEX exposure and adverse reproductive outcomes necessitate the in-depth evaluation of fertility and reproductive success in the wake of ethylbenzene exposure. The xenobiotic metabolism signaling pathway is also highly relevant for stress response. Both the AMPK and xenobiotic metabolism signaling pathways were altered in a previous study after 0.1 ppm or 1 ppm exposure to another BTEX constituent, benzene [15]. This study also found that after 1 ppm exposure to benzene, the LXR/RXR activation pathway and acute phase response signaling were statistically significant which is consistent with some of our findings here [15]. The LXR/RXR activation pathway regulates biosynthesis of fatty acid during early embryogenesis and is associated with inflammatory responses [37]. The acute phase response signaling pathway is also responsible for an inflammatory response wherein the organism defends against injury or infection by synthesizing acute phase proteins in the liver in response to increased levels of certain cytokines [38,39]. Adverse birth outcomes are associated with BTEX exposure likely through inflammatory and immune pathways [40]. Our previous studies with benzene identified inflammatory and immune response resultant of the VOC exposures and this response has been a theorized mediator for reproductive effects in humans [15]. While many other genes and pathways are involved in inflammatory processes such as synthesis of inflammatory cytokines, this process is intertwined in a complex balance with AMPK activation and oxidative stress- another key finding from previous low -dose VOC toxicity studies [15]. Lastly, the vitamin D metabolism pathway was significantly affected which has implications for the endocrine system and even some cancers because evidence from animal models show vitamin D has beneficial effects on various cancers, blood pressure, heart disease, immunologic disorders meaning disrupted metabolism could have widespread effects [41].

These genes and pathways play a pivotal role in insulin signaling which has been the most consistent result of VOC zebrafish toxicity studies. Insulin signaling plays a major role in regulating the endocrine system which is responsible for determining reproductive function as dictated through hormone production and signals. The current understanding of insulin has expanded from simply being a regulator of energetic homeostasis via glucose maintenance to include the regulation of reproduction as a “major signal linking metabolism and reproductive status” [42]. Further, insulin influences the hypothalamic-pituitary system and HPG axis (hypothalamic-pituitary-gonad) [43]. While many other genes and pathways are involved in inflammatory processes such as synthesis of inflammatory cytokines, this process is intertwined in a complex balance with AMPK activation and oxidative stress- another key finding later in life for benzene [44].

Transcriptomic changes of ethylbenzene exposure were also related to the endocrine and neurological systems. Among the genes associated with endocrine pathways and functions are pdca, crtc2, cubn, and map3k4. Phosducin (pdca/PDC) is known to be involved in transmission of sensory signals, regulating blood pressure, and thyroid gland endocrine functions [45]. This gene is also predicted to act upstream of G-protein-coupled receptor signaling pathway[46]. The current literature for CRTC2 is focused on glucose and lipid metabolism contributing to metabolic homeostasis [47]. In zebrafish, crtc2 is predicted to act upstream of the CREB transcription factor and within the AMPK signaling pathway [48]. Another toxicology study with known endocrine disruptors bisphenol A (BpA) and phthalates found Crtc2 was a mediator in toxicity after gestational BPA exposure, which can be associated with lipid metabolic disorders in later life [49]. Also cubn, a vitamin D-related gene, is theorized to play a role in the pathogenesis of thyroid carcinoma- one of the most prevalent endocrine malignancies [50]. For map3k4, the evidence for endocrine activity is even stronger. This gene is known to play a role in not only fetal/placental growth but also in controlling the insulin-like growth factor 1 receptor (IGF1R), insulin receptor (IR), and Akt signaling pathway [51]. Interestingly, akt2, active in the Akt signaling pathway as well as multiple insulin-like growth factors, was consistently dysregulated in our previous studies of 1,4-dioxane [52]. These genes and pathways play a pivotal role in insulin signaling which has been the most consistent result of VOC zebrafish toxicity studies.

Involved in neurological pathways were some highly dysregulated genes otx2a and actn2b. Important for brain development, Otx2 which in mice is coexpressed with phsoducin (Pdc) at the embryonic stage, is relevant as a marker of photoreceptors [53]. In humans, the early expression of OTX2 and OTX1 in the fetal brain supports the concept that OTX2 is important in neuronal cell development and differentiation [54]. Disturbed neuronal cell development can be a player in abnormal locomotion, especially due to light/dark stimuli which was observed here. Also interesting was that a zebrafish study with actn2 morphant embryos exhibited slow swimming in response to touch in addition to other maladaptive behavior phenotypes [55]. This supports the idea that otx2a and actn2b dysregulation could be involved in altered zebrafish behavior, though, neither of these were significant in the highest exposure group, also indicating a need to uncover the mechanism of dose-dependent ethylbenzene induced hypoactivity. A dose-dependent pattern was not clearly established as the ultra-low (0.001 ppm) exposure group was more dysregulated than the median doses (0.01, 0.1 ppm). Non-monotonic effects can be an indicator of endocrine disruption [56]. While ethylbenzene is not a well-established Endocrine disrupter, there is some aforementioned evidence for its EDC potential.

This study shows that ethylbenzene exposure of even 0.001 ppm can induce transcriptomic alterations that are associated with cancer, endocrine system, immune system and inflammatory responses. The endocrine associated molecular effects add to growing evidence that BTEX constituent exposure may play a role in environmental endocrine disruption. In addition, the number of dysregulated genes implicated for the top canonical pathways, on average, increased with concentration of ethylbenzene exposure. The behavioral and anatomical phenotypes at and above 0.001 ppm indicate low-level effects of ethylbenzene that haven’t been previously described. Further, neurological outcomes have yet to be linked to low-level ethylbenzene exposure and should be characterized further with learning and memory assays. These low-level exposure effects have policy implications and should be leveraged in future risk assessments. Future mechanistic studies should be conducted to investigate these transcriptomic changes, as well test if ethylbenzene-induced transcriptomic changes during early development have adult-onset consequences at environmentally relevant levels.

Environmental implications statement:

Volatile Organic Compounds (VOCs) are a major contributor to indoor and outdoor air pollution. Ethylbenzene, a singular component of the BTEX tetrad, is an aromatic hydrocarbon that poses several environmental risks, even at environmentally relevant concentrations, due to its volatility, toxicity to aquatic organisms, and potential to contaminate air, soil, and water. Ethylbenzene is possibly carcinogenic and naturally occurs in petroleum products and is mainly used to produce styrene. This study introduces a novel exposure model during early development to provide a foundational perspective of environmentally-relevant ethylbenzene toxicity that is translational to human and environmental health outcomes and regulation.

Highlights.

• Key genes were highly differentially expressed across multiple exposure concentrations

• Behavior findings suggest low-level ethylbenzene may impair normal behavior patterns

• Observed abnormal swim bladders have implications for thyroid system adverse outcome pathways

• Pathways involved in cancer were the most affected across exposure levels

• Endocrine system pathways were significantly changed post exposure