Zinc oxide nanoparticles disrupt development and function of the olfactory sensory system impairing olfaction-mediated behaviour in zebrafish

: 20 Zinc (Zn) is an essential metal present in numerous enzymes throughout the body, playing a vital 21 role in animal and human health. However, the increasing use of zinc oxide nanomaterials 22 (ZnONPs) in a diverse range of products has raised concerns regarding their potential impacts on 23 health and the environment. Despite these concerns, the toxicity of ZnONP exposure on animal 24 health remain poorly understood. To help address this knowledge gap, we have developed a 25 highly sensitive oxidative stress (OS) biosensor zebrafish capable of detecting cell/tissue-specific 26 OS responses to low doses of various oxidative stressors, including Zn, in a live fish embryo. 27 Using live-imaging analysis with this biosensor zebrafish embryo, we discovered that the 28 olfactory sensory neurons in the brain are especially sensitive to ZnOP exposure. Furthermore, 29 through studies monitoring neutrophil migration and neuronal activation in the embryonic brain 30 and via behaviour analysis, we have found that sub-lethal doses of ZnONPs (ranging from 0.033 31 to 1 mg/L nominal concentrations), which had no visible effect on embryo growth or morphology, 32 cause significant localised inflammation, disrupting the neurophysiology of olfactory brain tissues 33 and ultimately impaired olfaction-mediated behaviour. Collectively, these findings establish a 34 potent and important effect mechanism for ZnONP toxicity, indicating the olfactory sensory 35 system as the primary target for ZnONPs as environmental toxicant in aquatic environments. Our 36 result also highlights that even low doses of ZnONPs can have detrimental effects on the olfactory 37 sensory system, surpassing previous expectations. The importance of olfaction in environment 38 sensing, sex behaviours and overall fitness across species raises concerns about the potential 39 impact of ZnONPs on olfaction-mediated brain function and behaviour in animals and humans. 40


ABSTRACT:
Zinc (Zn) is an essential metal present in numerous enzymes throughout the body, playing a vital role in animal and human health.However, the increasing use of zinc oxide nanomaterials (ZnONPs) in a diverse range of products has raised concerns regarding their potential impacts on health and the environment.Despite these concerns, the toxicity of ZnONP exposure on animal health remain poorly understood.To help address this knowledge gap, we have developed a highly sensitive oxidative stress (OS) biosensor zebrafish capable of detecting cell/tissue-specific OS responses to low doses of various oxidative stressors, including Zn, in a live fish embryo.
Using live-imaging analysis with this biosensor zebrafish embryo, we discovered that the olfactory sensory neurons in the brain are especially sensitive to ZnOP exposure.Furthermore, through studies monitoring neutrophil migration and neuronal activation in the embryonic brain and via behaviour analysis, we have found that sub-lethal doses of ZnONPs (ranging from 0.033 to 1 mg/L nominal concentrations), which had no visible effect on embryo growth or morphology, cause significant localised inflammation, disrupting the neurophysiology of olfactory brain tissues and ultimately impaired olfaction-mediated behaviour.Collectively, these findings establish a potent and important effect mechanism for ZnONP toxicity, indicating the olfactory sensory system as the primary target for ZnONPs as environmental toxicant in aquatic environments.Our result also highlights that even low doses of ZnONPs can have detrimental effects on the olfactory sensory system, surpassing previous expectations.The importance of olfaction in environment sensing, sex behaviours and overall fitness across species raises concerns about the potential impact of ZnONPs on olfaction-mediated brain function and behaviour in animals and humans.

Introduction
Zinc (Zn) is an essential metal required for the health of animals and humans.It exists in numerous enzymes and proteins present across various tissues in the body, serving integral catalytic and structural functions.The cellular Zn levels are tightly controlled by the activity of different Zn transporters, ensuring proper Zn homeostasis (Haraet al. 2017;Plumet al. 2010).
Zinc oxide nanoparticles (ZnONPs) are among the most highly produced engineered metal oxide nanomaterials with an estimated annual global manufacture of 8,000 tons (Schulteet al. 2019).
With their many beneficial physicochemical properties, including strong absorption of UV light, anti-bacterial/anti-fungal/anti-cancer actions and catalytic function, they are used in a very wide range of products, including in pharmaceuticals, biomedical products, cosmetics, sunscreen/UV protection, food additives, and industrial products (Sirelkhatimet al. 2015;Vimercatiet al. 2020).
Many in vitro and animal model studies have shown that oxidative stress (OS) responses are one of the most likely biological mechanisms underlying the toxicity of metal-based nanomaterials, including for ZnONPs.In cell-based in vitro systems, exposure to ZnONPs has been shown to cause acute/chronic inflammation, damage proteins, lipids and/or nucleic acids, and induce apoptosis due to the production of excess reactive oxygen species (ROS) (Adamcakova-Doddet al. 2014;Huanget al. 2010;Jacobsenet al. 2015;Saptarshiet al. 2015;Sharmaet al. 2012;Xiaet al. 2008).
Such information is needed to understand more fully the potential hazards of ZnONPs in animals, including humans, and to help establish biomarkers and endpoints suitable for a more accurate evaluation of ZnONP (and other nanomaterials)-induced toxicities.
The zebrafish embryo has become an important vertebrate experimental model for (eco)toxicological assessments (Takesonoet al. 2022a).It is particularly useful for monitoring the effect(s) of environmental pollutants on the development and physiology in the embryo-larvae, due to their transparency and rapid development.Well-established gene knock-out and transgenic technology in the zebrafish has also greatly facilitated the examination of how pollutant exposure affects specific gene function and signaling.In addition to these technical advantages, the fundamental stages of organogenesis in the zebrafish are completed within five days post fertilisation (dpf) before it is categorised as "regulated animal" under UK home office guidance (Office) as well as EU Directive 2010/63/EU (Strähleet al. 2012).Thus, the use of the zebrafish embryo for animal testing and scientific research complies favourably with the requirement for Replacement, Reduction, and Refinement of the use of animals in research (the 3Rs).
We have previously established an OS biosensor transgenic zebrafish model, tg(electrophile response element (EpRE):mCherry), that allows visualisation of cell-tissue-specific gene transcription in response to oxidative stress across the whole body of developing zebrafish embryos (Mourabitet al. 2019).This biosensor line utilises the Nrf2/Keap1-mediated RedOx system, where ROS triggers Nrf2 activation through Keap1 release, which in turn translocates to the nucleus and activates the EpRE in the genome to induce the expression of cytoprotective genes (Kobayashi and Yamamoto 2006).Combined with non-invasive imaging methods, our EpRE:mCherry model shows a high signal-to-noise sensitivity for a wide range of OS-inducing materials, spanning metals to pharmaceuticals, and is particularly useful for quantifying a stressor-dependent and tissue-specific OS response (Mourabitet al. 2019).
In this paper, we set out to identify the target cells/tissues and olfactory mediated consequences of early-life exposure to ZnONP, employing the tg(EpRE:mCherry) zebrafish model in combination with imaging methods.For this work we employed ZnONP concentrations that occur in polluted natural environments (i.e.33 µg/L) and a higher (but non toxic) level to help illustrate the effects mechanisms on olfactory development and function via the imaging work.

Results
Characterisation of OS-responding target cells/tissues to acute exposure of ZnONPs in OS biosensor zebrafish embryos.
To identify spatiotemporal OS responses for exposure to ZnONPs, the embryos of an OS biosensor transgenic zebrafish model, Tg(EpRE:mCherry), were exposed to ZnONPs (NM110, JRC) under a static exposure condition (detailed in Materials and Methods).
To verify the particle exposure conditions, the dosimetry of ZnONP and ZnSO4•7H2O (dissolved metal control) in the zebrafish exposure media (the egg water) were characterized.We verified that the exposure conditions used were consistent throughout the experimental periods (Suppl. Info. and Fig. S1).We also confirmed that the uptake of Zn in the exposed embryo-larvae was dose-dependent (see Suppl.Info.and Fig. S1D).
In the control group, EpRE:mCherry embryo-larvae generally showed minimal basal mCherry expression except for the somewhat higher basal mCherry expression in the eye lens Fig.S2A and S2B).With ZnONP exposure, however, we observed an intense induction of OS responses (mCherry expression) specifically in the OE and OB, two primary brain tissues involved in olfaction (Fig. 1Aiii, indicated with a white dotted square).These unique OS responses occur in a subset of OSNs, that originate from the OE and project to the OB.We also observed a relatively weaker mCherry expression in ganglions of the peripheral sensory system in ZnONP exposed embryo-larvae (Fig. 1Aiii, white arrows and Fig. S2A and S2B), suggesting ZnONPs also induce OS responses in mechanosensory cells.In addition, ZnONP exposure caused OS responses in a specific type of skin epithelial cells called ionocytes (Fig. S2A and S2B and Takesono et al. in preparation).The OE and OB, however, were the two distinctive tissues in the brain showing ZnONP-induced OS responses.
We found that the OS responses in OSNs were detectable from exposure concentrations of 33 µg/L (≈ 405.3 nM, nominal) (Fig. 1C, indicated with white arrows) being increased in a dosedependent manner (Fig. 1D-F, white arrows; Fig. 1J).Dissolution analysis revealed that ZnONPs dissolved in the egg water over time, with 78.2 ± 3.7% dissolution at 72h after the initiation of the incubation (Fig. S1E).These data strongly suggest that most (almost 80%) of ZnONPs were dissolved releasing Zn ions into the media during the exposure period.Indeed, we found that dissolved Zn (as ZnSO4•7H2O) equally induced the OS response in OSNs (Fig. 1G-I).The doseresponse curves of ZnONP and ZnSO4•7H2O were similar when the molar concentrations were compared (Fig 1J; this is a non-linear curve fit using GraphPad Prizm version 9.3.1, and involving five independent experiments for ZnONPs and two independent experiments for ZnSO4•7H2O, with 8 replicates for each condition in each experiment), supporting the hypothesis that Zn ion is the major cause of the OSN specific OS response.Of note, no significant effects were seen on hatching rate at 72 hpf (Fig. S2C), lethality (Fig. S2D) or on embryo growth at 96 hpf (Fig S2E ), regardless of the dose of ZnONP or ZnSO4•7H2O used in this study.The mean effective concentration (EC50) for the ZnONP-induced OS response in EpRE:mCherry embryo-larvae was calculated as 179 µg/L (= 2.2 µM), which is approximately 45 times lower than the LC50 for ZnONP cytotoxicity in PMA-differentiated THP-1 macrophage cells (LC50 = 8.1 µg/ml) (Safaret al. 2019) and 200-335 times lower than the 96h LC50 for zebrafish embryos (LC50 = 35.88mg/L -60 mg/L) (Azevedoet al. 2016;Duet al. 2017).Using the EpRE:mCherry model our data indicate that live fish embryo-larvae show remarkable sensitivity to ZnONP-induced OS responses, even at doses relevant to those measured in some polluted surface waters, without visible effects on embryo development or morphology (discussed in "Wider implications of ZnONP exposure on olfaction" section).

ZnONP exposure affects olfactory development.
To confirm the cell types responding to ZnONP exposure, ZnONP-exposed EpRE:mCherry embryolarvae were co-stained with anti-keyhole limpet hemocyanin (KLH) antibody, a marker for OSNs, and with anti-mCherry antibody for OS responding cells.In control, the termini of KLH positive OSNs form spherical clusters in the OB, which are synaptic neuropils called olfactory glomeruli (Fig. 2Ai, mediodorsal glomeruli, mdGs; Fig. 2Aii, dorsolateral glomerulous, dlG and Fig. 2E).In vertebrates, OSNs that express one type of olfactory receptor innervate a single glomerulus.The excitability patterns in these olfactory glomeruli play a crucial role in odour sensing, forming a topologically distinct representation known as the odour sensory map (Miyasakaet al. 2013).We found that ZnONP exposure induced mCherry expression in the somata of OSNs in the OE (Fig. 2B and 2C) as well as in their projections extending to olfactory glomeruli (Fig. 2Bi-ii and 2Ci-ii).
Dissolved Zn also triggered OS responses in OSNs (Fig. 2D).In the embryo-larvae exposed to ZnONPs and dissolved Zn, the axonal termini of OSNs in olfactory glomeruli exhibited a loss of KLH expression and instead showed mCherry expression.This indicates the activation of OSmediated transcription in those regions, accompanied by morphological distortions (Fig. 2Ai-ii vs 2Bi-ii, 2Ci-ii and 2Di-ii; Fig. 2E vs. 2F).These data suggest that ZnONP exposure specifically induced OS-mediated apoptosis in OSNs, resulting in the defasciculation of the axonal bundle of the OSNs.
To further characterise the cell types of the OSNs, ZnONP-exposed EpRE:mCherry embryo-larvae were stained with an anti-calretinin antibody, which is a marker for microvillous OSNs in embryonic zebrafish (Koideet al. 2009).In ZnONP-exposed samples, calretinin-positive microvillous OSNs maintained a typical spherical glomerular structure with no mCherry expression.Notably, other OSNs expressing mCherry were present in the same olfactory brain regions of the same embryo (Fig. 3A vs. 3B).The optical section images confirmed that the majority of microvillous OSNs did not induce OS response even at a high concentration of ZnONP exposure (1 mg/L).The expression of mCherry and calretinin thus showed a mutually exclusive pattern in both cell bodies and glomerular innervations (Fig. 3Ci-iii).These data suggest that the sensitivity to ZnONP differs among different types of OSNs, with KLH-positive/calretinin negative OSNs (likely ciliated OSNs (Koideet al. 2009)) being more severely affected compared to calretinin-positive microvillous OSNs.In addition, ZnONP exposure in the embro-larvae led to a reduction in the size of the OE rosette area (Fig. 3D, * p<0.05, by t-test, GraphPad Prizm version 9.3.1,n=7), without affecting other brain regions (Fig. 3E and 3F, ns (not significant) by ANOVA with Sidak's post hoc test, GraphPad Prizm version 9.3.1,n=7).These findings are consistent with the observation that OS responses were exclusive to OSNs in the brain of ZnONP-exposed animals (Fig. S2A).

ZnONP exposure induces a local inflammation around the OS responding olfactory tissues.
OS responses induce ROS, which subsequently triggers the progression of inflammation (Mittalet al. 2014).During the initial phase of acute inflammation in zebrafish (within 30 min after an injury), neutrophils are recruited to the site of the injury by sensing the ROS gradient (i.e., H2O2) (Niethammeret al. 2009;Yooet al. 2011).To examine whether the ZnONP-induced OS response in OSNs caused acute inflammation, we developed and applied a double transgenic zebrafish embryo carrying both EpRE:mCherry and mpx:GFP (neutrophil-specific) reporter genes, which allows for the simultaneous monitoring of OS-responding OSNs (mCherry) and neutrophils (GFP).
With this line, we investigated whether ZnONP exposure affects the migratory behaviour of neutrophils, that reflects their infiltration and activation status upon a local inflammation.We also examined the spatial relationship between OS-responding OSNs and neutrophils in the olfactory tissues.We found that the number of neutrophils in the head/brain regions was significantly increased by ZnONP exposure (Fig. 4A, left-end column).This increase was dosedependent and particularly evident in the anterior part of the brain, in the area in close proximity to the OS-responding OSNs (Fig. 4Bi and Fig. 4Ci; details in Materials and Methods).Applying trajectory analyses for migrating neutrophils, we found that many neutrophils were retained around OS-responding OSNs in ZnONP-exposed embryo-larvae: in the anterior head/brain region, total distance moved (Fig. 4Bii), displacement (start-end distance: Fig. 4Biii) and migration speed (Fig. 4Biv) of neutrophils were consistently reduced by ZnONP exposure in a dosedependent manner.In contrast, in the mid-region of the head, which is distant to the OSresponding OSNs, no change in the migratory behaviour of neutrophils was observed except for a reduction in the migration speed in the animals exposed to I mg/L ZnONPs (Fig 4Cii -iv).These data indicate that neutrophils can detect the location of the OS-responding OSNs in ZnONPexposed animals and likely contribute to the acute inflammation in the local microenvironment.
ZnONP exposure alters the intrinsic/spontaneous neuronal activity in the olfactory brain regions in the embryonic brain.
Next we explored the neurophysiological consequence of the ZnONP toxicity in relation to OSNspecific OS responses.For this, we employed the live calcium imaging analysis using elavl3:GCaMP6s transgenic zebrafish model that has been established previously in our laboratory (details in (Takesonoet al. 2022b;Winteret al. 2017)).Using this system, a full brain volume of live imaging of the calcium indicator, GCaMP6s, was obtained and the region of interest (ROI)-specific neuronal activity data was subsequently extracted.With this, we found that ZnONP-exposure led to a significant increase in intrinsic GCaMP signals primarily in the anterior forebrain regions (group I in Fig. 5B; details in Materials and Methods).This increase was particularly evident in the OE and OB, the specific olfactory brain tissues exhibited pronounced OS responses to ZnONPs.Similar effects were observed in some surrounding ROIs in the diencephalon (group II) and in a specific brain region in the midbrain (mesencephalon)(group III).
ZnONP effects on neurophysiology of wider ROIs may be due to the functional interconnectivity between these olfactory tissues and the other brain regions.Collectively, our data indicate that ZnONP exposure alters intrinsic neuronal activity in a region-specific manner, correlating with its effects on region-specific OS responses, alterations in OSN development, and the resulting local inflammation.

ZnONP exposure impairs olfaction-mediated avoidance behaviour.
Finally, we addressed whether this OSN-specific toxicity of ZnONP could affect olfactionmediated behaviour.To do so, we examined the odour-evoked avoidance behaviour using a death-associated odour, cadavarine, as an aversive olfactory cue (Takesonoet al. 2022b).A group of 10 embryo/larvae were placed in the middle area of a test chamber which was partitioned into three areas by removable dividers (Fig 6A).The movement of the test individuals was videorecorded and automatically tracked using the multiple object tracking system in idTracker software (Pérez-Escuderoet al. 2014).During the acclimation period (before applying cadavarine), both control and ZnONP-exposed embryo-larvae showed no significant difference in swimming capability (total distance travelled during the last 5 min of acclimation: 98.03 ± 5.40 mm for control vs. 85.33 ± 7.64 mm for ZnONP-exposed group)(Fig.6C, acclimation, left).With cadavarine administration following divider removal, the test animals were subsequently exposed to the cadavarine gradient.We found that control fish responded to cadavarine by avoiding the side of the tank with a higher cadavarine gradient (Fig. 6B, control, left), travelling greater distances (Fig. 6C, cadavarine, right), and spending more time at the opposite end of the chamber (Fig. 6D).In contrast, ZnONP-exposed fish showed reduced responsiveness to cadavarine, remaining around the centre of the chamber where the fish were originally placed (Fig. 6E).Thus, the total distance travelled during cadavarine exposed period (Fig. 6C, cadavarine, right: 866 ± 98.06 mm for control vs. 390 ± 42.38 mm for ZnONP groups) and the centroid distance from the cadavarine administrated site over the acquisition time (avoidance index) were significantly reduced in ZnONP-exposed groups (Fig. 6F).These data indicate that ZnONP exposure impairs olfaction-mediated avoidance behaviour in zebrafish embryos-larvae, reflecting specific OS responses in OSNs and resultant defects in olfactory development and function.

Discussion
ZnONP exposure induces the OSN-specific OS response, resulting in adverse effects on olfactory tissue development, neurophysiology and behaviour.
We have shown previously that our tg(EpRE:mCherry) model can effectively identify distinct tissuespecific OS responses to a diverse range of chemical and other stressors.Illustrating this, acetaminophen (paracetamol) triggers OS responses in the liver, cisplatin in hair cells, phenylhydrazine in red blood cells and Cu 2+ ions in skin ionocytes (Mourabit M. 2019).Thus, our OS biosensor zebrafish model discerns chemical-specific target cells/tissues, reflecting its specificity in responding cell types and its mode of action (Mourabit M. 2019).Using the OS biosensor zebrafish model, here we show that ZnONP-exposure causes olfactory tissue-specific developmental neurotoxicity via the induction of OS response in the OSNs.This neurotoxicity was induced by sub-lethal exposure concentrations of ZnONPs, with which no effect on overall morphological development, hatching rate, embryo size (growth) or survival rate was observed in zebrafish embryo-larvae.Using whole-mount immunohistochemistry and other transgenic zebrafish models, we showed that ZnONP-exposure leads to defects in OSN development and olfactory tissue-specific inflammation, and subsequently alters the neurophysiology in the forebrain regions and disrupts olfaction-mediated behaviour in the zebrafish embryo-larvae.In light of this, our data unveil OSNs as a primary target tissue for ZnONP neurotoxicity, and illustrate that sublethal exposure concentrations of ZnONP impact neurodevelopment, physiology, and behaviour in live fish embryos.Whether this effect may occur for other metal based nanomaterials has yet to be explored fully, but for studies on Cu 2+ exposure (0.6-6.5 µg/L) in tg(EpRE:mCherry) embryos we have not found evoked OS responses in OSNs but rather induce marked responses for this metal in skin ionocytes (Mourabit M. 2019).
The olfactory epithelium: the primary uptake site of ZnONPs in the brain.
In our OS biosensor model, ZnONP exposure induced OS response exclusively in a few specific cell types, including OSNs in the brain (this study) and skin and gill epithelial cells (Fig. S2A-B, Takesono et al. in preparation).These cells are commonly in direct contact with the aquatic environment and therefore likely to be especially vulnerable to chemical exposures.From this viewpoint, the OE is likely to be the first entry site for ZnONPs into the brain of fish embryo.This is supported by our data that intense OS responses (indicated by mCherry expression) were seen only in the cell bodies of OSNs in the OE and their axons in the OB, while no mCherry expression was observed in other cells in the brain (Fig. 1Aiii, Fig. 2A and 2C).These findings suggest that ZnONP-induced OS responses occur along peripheral (the OE) -central (the OB) axis of OSNs in the zebrafish embryo-larvae.
Under our exposure conditions, it appears that the dissolved Zn ion, rather than the particulate ZnONP, is responsible for the OS responses in OSNs.This is supported by the fact that the by 96 hours of exposure the majority (over 80%) of ZnONPs had dissolved in the exposure media (Fig. 1SE) and given the approximate EC50 values for the OS response for both ZnONP and Zn ion exposures (Fig. 1J).The dissolved Zn ion from ZnONPs is likely taken up into the cytoplasm of OSNs by specific isotypes of Zn transporters expressed in their transmembrane.In zebrafish, there are total 18 Zn transporters which are involved in either Zn influx or efflux (Feeneyet al.

2005).
Zn plays a crucial role in the brain, regulating transcription during neurogenesis and differentiation, neurotransmission and neuronal apoptosis following injury or ischemia (Fredericksonet al. 2005;Plumet al. 2010).Previous studies have shown that intracellular concentrations of Zn are particularly high in specific brain regions in rodents, including the olfactory bulb.Zinc autometallography (AMG) and immunostaining of Zn transporters, such as ZnT1 and ZnT3 involved in Zn efflux, have revealed the complementary enrichment of Zn ion and Zn transporters in the outer part of each olfactory glomerulous which consists of OSN termini and dendrites from mitral cells and periglomerular cells (Joet al. 2000;Sekleret al. 2002).These findings suggest specific roles of Zn in the development and function of olfactory glomeruli.OSNs may possess an inherent ability to efficiently incorporate Zn ions into their cytosol, which could lead to disruptions in Zn homeostasis upon ZnONP exposure and heightened susceptibility to ZnONP toxicity.The distinct OS responses in OSNs may be due to the specific expression of Zn transporters within these cells.Further study into the specific expression patterns and roles of these transporters in OSNs in zebrafish embryo-larvae would enhance our understanding of the potential mechanisms involved in Zn uptake and efflux related to ZnONP toxicity.

Wider implications of ZnONP exposure on olfaction
The concentration range of ZnONP used in this study, with nominal concentrations between 33 µg/L and 1 mg/L and measured concentrations between 54.18±16.25 and 762±18.18µg/L (Fig. S1A), is considerably lower than those reported in previous studies on ZnONP toxicity in zebrafish embryo-larvae (Azevedoet al. 2016;Duet al. 2017) and with relevance to environmental concentrations of Zn.Our dissolution analysis showed that most of ZnONP (about 80%) was dissolved in the water by 72 hours of the exposure (Fig. S1E), and the toxicities of ZnONP and Zn ion for OSNs did not appear to differ (Fig. 1 and Fig. 2).Thus, the toxic effects on OSN are most likely to be due to Zn ions rather than ZnONP particles themselves.
Zn is a relatively abundant metal, and ZnONP may contribute considerably to the bioavailable source in the environment (Larneret al. 2012).The environmental quality standard (EQS) values for Zn vary globally (Vorkamp 2016), with EQS values ranging from 10.9 µg /L In UK freshwaters for Zn in Japan between 1991 to 2002 across 3347 sites found that approximately 20% of the test sites exceeded the national EQS (30 µg/L), including near the municipal wastewater treatment plants for household wastes (concentrations of Zn ranged between non detectable to 2.9 mg/L) (Naitoet al. 2010).Thus, the ZnONP exposure concentrations in our study can be considered as within the environmental concentration range, making our findings on ZnONP-induced olfactorydisruption relevant to freshwater bodies worldwide.
Olfaction is essential for various animal behaviours such as foraging, kin-recognition, predator avoidance, mating, and eco-location (Wyatt 2003).Olfaction dysfunction may also affect the social and sex behaviours and fitness of the animals.Importantly, the principal mechanisms underlying the organisation of olfactory system and odour processing are highly conserved across diverse phyla and indeed most animal species, including worms, insects, fish, rodents and humans (Ache and Young 2005).Thus, neurotoxicity of ZnONP reported in this study suggests a potential global risk to animal and even human health.To what extent the ZnONP-mediated developmental neurotoxicity in OSNs affects odour sensing in later-life and/or the adulthood remain unknown and warrants further investigation.

Conclusions
Using an OS biosensor zebrafish model, we have discovered a unique neurotoxic effect of ZnONPs on OSNs during developmental stage of zebrafish embryo-larvae.This specific OS response in OSNs leads to defects in OSN development and localised inflammation in olfactory tissues, and disruptions in neurophysiology and olfaction-mediated behaviour.Remarkably, these OSmediated defects in olfactory sensory system occur even at low exposure concentrations of ZnONP, without any visible effects on embryo growth or morphology, highlighting their direct relevance as a primary mechanism of ZnONP toxicity in the environment.Our study emphasises the importance of research into olfactory sensing systems of fish (and other organisms) when assessing the potential hazards associated with exposures to metal based nanomaterials and potentially other types of toxicants.

Experimental zebrafish lines and fish husbandry
The oxidative stress biosensor transgenic zebrafish line, Tg(EpRE:mCherry), used in this study has been described previously in (Mourabitet al. 2019).The double transgenic of EpRE:mCherry and mpx:GFP (neutrophils line; generously provided by Stephen A. Renshaw, University of Sheffield), Tg(EpRE:mCherry; mpx:GFP), were produced from a pair-cross of homozygous parents of each TG fish line (detailed in SI, Suppl.Materials and Methods).The elavl3(huC):GCaMP6s transgenic zebrafish line, Tg(elavl3:GCaMP6s), has also been described previously in (Takesonoet al. 2022b;Winteret al. 2021;Winteret al. 2017) and was originally supplied by Misha B. Ahrens (Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, Virginia, USA).Adult zebrafish were maintained in flow through aquaria in the aquatic resource centre at the University of Exeter on a g 14/10 light/dark cycle at 28 ± 1 °C.All procedures for fish husbandry and the experiments conducted with zebrafish were in accordance with U.K. Home Office regulations for the use of Animals in Scientific Procedures Act (1986) and followed local ethical review guidelines for ensuring their humane treatment.

ZnONP and ZnSO4•7H2O exposure
A stock ZnONP (NM110, JRC) suspension was made up at a concentration of 2.56 mg/ml in MilliQ water was prepared following the NANoReg_Ecotox dispersion protocol version 6 (Booth 2015) using a probe-sonicator (Cole-parmer, CPX 130 ultrasonic processor).For this, the amplitude and duration of the sonication collectively delivered a total acoustic power (Pac) of 7.35±0.05Watts.
The stock ZnONP suspension was then left for overnight at room temperature to achieve suspension equilibrium.Prior to the initiation of the embryo exposures, the ZnONP stock suspension was re-suspended via a brief (30 sec) water-bath sonication and the solution was then used immediately to make up a series of working exposure dilutions in the zebrafish embryos exposure media (egg water; 60 mg/L artificial sea salt, The Tropical Marine Centre).
EpRE:mCherry embryos were collected by natural spawning, cleaned and cultured in the egg water at 28°C until 24 hpf.Healthy 24 hpf EpRE:mCherry embryos were then exposed under static systems to freshly prepared ZnONP suspensions at concentrations of 33, 100, 330 µg/L and 1 mg/L , for 72 hours (until 96 hpf) at 28°C.This was carried out in 24 well plates at a density of 5 embryos/2ml media/well.Further sets of EpRE:mCherry embryos were exposed to ZnSO4•7H2O (dissolved Zn) at water concentrations ranging between 100 µg/L and 10 mg/L under the same exposure conditions as described for ZnONP above, that lead to comparative OS responses in olfactory tissues as those induced by used ZnONP concentrations.Control EpRE:mCherry embryos were prepared by exposing to the egg water only in the same incubation conditions.All experiments were repeated at least twice and each experiment was conducted with a different batch of EpRE:mCherry embryos, freshly prepared ZnONP/Zn ion suspension, and with 8 replicates per treatment condition.

Imaging and quantification of tissue specific OS responses
For imaging of tissue specific OS responses, 96 hpf control, ZnONP or dissolved Zn exposed EpRE:mCherry embryos were anaesthetised with 0.03% MS222 in egg water and quickly mounted in 0.7% low melting point agarose in a 35 mm diameter glass-bottom dish (MatTek) for microscopic observation.For dose-response analyses, epifluorescent images of oxidative stress responses (read-out as tissue specific mCherry expressions) in control and ZnONP and dissolved Zn exposed embryos were acquired using fluorescence light microscopy (Zeiss observer Z1; Zeiss, Cambridge, UK) with a metal halide fluorescent light source (HXP 120C, Zeiss, Cambridge, UK) and a 20x objective lens.The imaging parameters (e.g. the exposure time and intensity of the fluorescent light, z-step size) were optimised to avoid the saturation of mCherry signal in the embryo-larvae samples at the highest dose of ZnONP or dissolved Zn exposure and to ensure inclusion of all signals in the OE and the OB, and the imaging parameters were kept consistent for each experiment.The acquired images were processed using Fiji with set parameters for brightness/contrast adjustment and background subtraction.To quantify OS responses in the OE, the average intensity projection image of each sample was obtained and the mean grey value of mCherry signal within the OE was measured.The OE region was identified by tracing the outline of the olfactory rosette in the DIC image.The background grey value in the equivalent area size of non-fluorescent brain region was subtracted from the raw grey value of the mCherry signals in the OE for each sample.Data (Fig. 1J) are shown as Log (OS response = grey vale) plotted against log (M) of ZnONP or ZnSO4•7H2O concentration, and curve fits calculated and applied using GraphPad Prism 9.0.For representative images (Fig 1A -I), Zeiss LSM880 airyscan operated by Zen Black software was used keeping at a set (and/or optimal) z step size.mCherry Images were airy processed in Zen Black software and DIC images were processed using extended depth focus (EDF) function in Zen Blue software.Processed images were merged and further processed using Fiji with set parameters for the adjustment.

Assessment of local inflammation in live zebrafish embryos
A double transgenic of EpRE:mCherry and mpx:GFP (neutrophils), Tg(EpRE:mCherry; mpx:GFP), was used to observe the ZnONP-induced OS response in the OSNs and simultaneously monitor the migratory behaviour of neutrophils in the same embryo-larvae.Exposure conditions and sample preparation for live imaging are as described in "Imaging and quantification of tissuespecific OS responses" section.Live imaging of mCherry-expressing OSNs and GFP-expressing neutrophils was carried out using Zeiss 880 in fast acquisition mode with Airyscan, which achieves nine consecutive optical z-section images extending through the entire forebrain-midbrain regions (scan depth 120 µm, 12 µm step each) in 2.2 s, allowing tracking of migratory trajectories of neutrophils in the whole forebrain-midbrain regions.The time-lapse images were obtained every 2 minutes for 20 minutes.The acquired images were airy processed in Zen Black software and were further processed using Fiji with set parameters for the adjustment.Migratory parameters of neutrophils in the anterior brain and the mid-brain regions were analysed using Manual tracking plugging in Fiji.Data were obtained from two independent experiments, with four individual fish embryos as experimental replicates, and analysing migratory behaviour of 13-15 neutrophils per embryo.Data analyses were carried out using ANOVA with Tukey's post hoc test using GraphPad Prizm version 9.3.1.

GCaMP6s imaging
The detailed procedures for sample preparation for GCaMP6s imaging and for imaging acquisition using a custom-built LSM are described in (Takesonoet al. 2022b;Winteret al. 2017).
Exposure of Tg(elavl3:GCaMP6s) embryo-larvae to 125, 250 µg/L ZnONP was conducted as described in the section" ZnONP and ZnSO4•7H2O exposure".At 96 hpf (72 h exposure), exposed fish were pre-screened for a similar basal GCaMP expression level in the brain before LSM imaging.All experimental treatments were conducted using the same batch of embryos, and each experiment was repeated on two separate occasions, to account for possible batch-to-batch variation.Data analyses were carried out using mixed effect generalised linear model in R v. 3.2 (Team 2013).N=7-9.

Whole-mount immunohistochemistry
All whole-mount immunohistochemistry was conducted with 4% paraformaldehyde (PFA)-fixed 4 dpf EpRE:mCherry embryos that had been exposed to ZnONP or dissolved Zn or non-exposed controls, as described above.See further details in (Takesonoet al. 2022b).

Olfaction-mediated avoidance behaviour assays
EpRE:mCherry embryos were cultured with or without 1 mg/L ZnONP as described above at a density of 50-70 embryos in 75 ml of the egg water in a glass dish.Olfaction-mediated avoidance behaviour was assessed in infrared light in an enclosed box using automated video-tracking system (PGRFlyCap software) with a video camera at 20 frames per second.At 5 dpf, 10 healthy fish were placed in the middle zone in a test chamber (10.5 cm x 3.5 cm x 1.7 cm)(Fig.5A) containing 15 ml of the egg water held in an enclosed light box with a camera attached to the top of the box.The fish in the test chamber were acclimated for 15 minutes and their movement during the last 5 minutes of the acclimation period was video-recorded to assess their migratory activity (acclimation phase).A fear-associated aversive odorant, cadavarine (20 µl of 10 mM), was subsequently administrated at the edge of either side of the arena (the side of administration was alternated between trials in a semi-random manner).After a 5-minute equilibration period, the dividers were removed and the movement of fish video-recorded for 10 min (test phase).
Acclimation and test phase videos were analysed using the multiple subject tracking system in idTracker software v2.1.(Pérez-Escuderoet al. 2014).In each trial the movement of individuals was analysed and extracted in the form of x-y coordinates.Using the extracted coordinates, the overall distance travelled for each individual was calculated, both for the acquisition and for the test phase.Individual fish that remained static for >50% of the frames analysed per trial were excluded from further analysis; these were evenly distributed across experimental conditions.
The mean centre of mass (centroid) was calculated for each experimental group and the shift of the centroid from the centre of the arena was determined as a measure of the response to the cadaverine administration during the test phase.
The distance travelled during the acquisition phase as well as the test phase was analysed using generalised linear mixed effect models in the 'lme4' v1.29 R package (Bates 2010); the full models included condition (Control/ZnONP) as a fixed factor and an individual trial-level as random effects.The shift of the centroid from the centre of the test arena during the test phase was also analysed using a generalised linear mixed effect model using the 'lme4' R package.The full model included the experimental condition (Control/ZnONP), the side of administration (Left/Right to control for any bias), the standardised average of the distance moved by the group of individuals (calculated separately for each treatment as a z-score); day was included as a random factor, while the sequence in which the animals were tested was found not to explain any of the variance and was therefore omitted.In all cases the family (the error structure, i.e. the probability distribution of the errors) and link function (i.e. the function relating the mean value of y to the fixed factor) were chosen for the best fitting-model.All models met their assumptions and were validated through diagnostic plots.All statistical analyses were carried out in R v. 3.2 (Team 2013).

(
Affairs. 2014) to 30 µg /L in Japan(Naitoet al. 2010; Shikazonoet al. 2008).In Europe the EQS for ZnONPs has been estimated at 10 µg/L, calculated based on a probabilistic material flow modelling analysis (Gottschalket al. 2009).Importantly, surface water monitoring data have shown Zn concentrations exceeding EQS levels in various locations, most notably around industrial areas and mining sites.In the case of the latter Zn has been measured in excess of 445 µg/L (e.g. in the River Teign, UK (Jordanet al. 2020) and Shiine river, Japan (Shikazonoet al. 2008)) and up to 935 µg/L in the River Hayle, UK (Minghettiet al. 2014).Surface water monitoring data

Figure 4 .
Figure 4. ZnONP exposure induces a local inflammation in close proximity to OS responding Figure 5.