Preclinical zebrafish model for organophosphorus 1 intoxication : neuronal hyperexcitation , behavioral 2 abnormalities and subsequent brain damages 3

As key compounds for modern cultivation practices, organophosphorus (OP)-containing pesticides have become an important public health and environmental issues, worldwide, causing millions human intoxications each year. OP poisoning induces cholinergic syndrome, associating irreversible brain damages with epileptic seizures, possibly ending in life-threatening status epilepticus. Existing countermeasures are life-saving, but insufficiently effective to prevent long lasting neuronal consequences, emphasizing the dire need for animal models mimicking OP poisoning as tools to identify novel anti-OP countermeasures. Here, we used diisopropylfluorophosphate (DFP), a prototypic and moderately toxic OP compound, to generate a zebrafish OP intoxication model and study the consequences of DFP exposure on neuronal activity, larvae behaviour and neuron network organization. DFP poisoning caused marked acetylcholinesterase (AChE) inhibition, resulting in paralysis, decreased oxygen consumption, overexpression of c-Fos neuron activity marker, increased neuron apoptosis and epileptiform seizure-like activity, which was partially alleviated by diazepam treatment. DFP-exposed larvae also showed altered neuron networks with increased accumulation of NR2B-NMDA receptor combined with decreased GAD65/67 and gephyrin protein accumulation. Thus, we described a zebrafish model of DFP poisoning, which should (i) provide important insights into the pathophysiological mechanisms underlying OP intoxication and ensuing brain damage, and (ii) help identify novel therapeutic agents to restore CNS functions following acute OP poisoning.


INTRODUCTION 46
Organophosphorus (OP) compounds are highly toxic molecules used as lethal weapons in both war situations and terrorist attacks, but also as key chemical pesticides to combat pests and 48 parasites. As the result of their massive use for agricultural purposes worldwide, OP poisoning 49 represents today a major public health issue with 3 million severe intoxications reported annually

Acetylcholinesterase activity 121
Five dpf zebrafish larvae (20 larvae per sample) were collected at 2 h, 4 h, 6 h post-exposure 122 to DFP and stored at −80 °C for further analysis. Samples were homogenized in 50 mM phosphate 123 buffer (pH 7.4)/0.5% Tween using Precellys® homogenizer with 1.4 mm ceramic beads and 124 centrifuged at 10,000 × g (4 °C) for 10 min. The resulting supernatants were collected and stored 125 at −80 °C. Total protein concentrations were determined using the DC Protein Assay (Bio-Rad) 126 and all the samples were diluted to 1.2 mg/mL. AChE activity was determined by adding 1 mM 127 acetylthiocholine (Sigma) and 0.22 mM 5,5′-dithiobis-2-nitrobenzoic acid (DTNB, Sigma) to the 128 sample (0.03 mg/mL of total proteins), the formation of the product resulting from the reaction 129 between thiocholine and DTNB at 25 °C was monitored for 30 min at 412 nm with a microplate 130 reader. All the samples were assayed in duplicate. The final results were expressed as percentages 131 of average control activity. 132

Measurement of oxygen consumption 133
Five dpf larvae were exposed to DFP or vehicle (1% DMSO) for 4.5 h and then transferred 134 to 96-well microplates (Greiner Bio-One International) (7 individuals per well) containing 90 µL 135 of E3 medium and 10 µL of 35 µg/mL MitoXpress Xtra (MitoXpress Xtra Reagent Pack, Agilent 136 Technologies) that enables real-time measurement of extracellular oxygen consumption in living 137 larvae. A volume of 100 µL of mineral oil (MitoXpress Xtra Reagent Pack, Agilent Technologies) 138 was added to seal the wells and isolate the reaction medium from ambient air oxygen. Oxygen 139 consumption was then measured in real time for 90 min at 28 °C in a 96-well plate using a 140 spectrofluorimeter (Tecan Spark: λ excitation 380 nm, λ emission 650 nm). The areas under the linear 141 portion of the curve were used to determine O 2 consumption rates. 142

Hematoxylin/eosin staining 143
Five dpf larvae were exposed to DFP or vehicle (1% DMSO) and then anesthetized using 144 0.01% tricaine, fixed with 10% formaldehyde, paraffin-embedded and sectioned. Sections were 145 deparaffinized and rehydrated before hematoxylin and eosin staining. Freshly stained sections were 146 treated with ethanol and xylene, and then mounted in Pertex medium. Sections were imaged using 147 a Nikon Eclipse microscope (E-200) equipped with a digital sight (Nikon). 148

Zebrafish larval locomotor activity 149
Locomotor activity of 5 dpf zebrafish larvae was performed as previously described in 150 Brenet et al, 2019 11 . 151

Neuronal calcium uptake imaging 161
Calcium imaging of 5 dpf zebrafish larvae was performed as previously described in Brenet 162 et al, 2019 11 . 163

Diazepam treatment 164
Five dpf Tg[Huc:GCaMP5G] larvae were exposed to 15 µM DFP for 5 h and then 165 pancuronium-paralyzed and embedded in 1.1% low-melting agarose in the center of a 35 mm glass-166 bottomed dish covered with E3 solution containing 300 µM pancuronium bromide. Calcium 167 uptakes were recorded for 30 min prior to diazepam (40 µM DZP, Sigma) addition and calcium 168 activity was then monitored for an additional hour. Calcium activity was measured as described in 169 Brenet et al, 2019 11 . 170

Apoptosis labeling 171
Neuronal cell death was visualized and quantified as previously described in Brenet et al, 172 2019 11 . 173

Immunohistochemistry 174
For synapse protein immunostaining, zebrafish larvae were fixed using 4% formaldehyde, 175 then directly immersed in 10% sucrose at 4 °C and incubated overnight, embedded in 7.5% 176 gelatin/10% sucrose solution, flash frozen in isopentane at -45 °C and stored at -80 °C until use. 177 When needed, frozen embedded zebrafish larvae were cut into 20 µm cryostat sections, which were 178 corresponding standard errors of the mean (SEM), and to assess statistical significance of the 210 differences observed between DFP-treated and control larvae. 211

RESULTS 212
3.1 Larvae exposed to DFP showed paralysis and acetylcholinesterase (AChE) inhibition 213 Prior to the development of a zebrafish model of DFP poisoning, we first measured the 214 stability of this compound after dilution in fish water. Ranging amounts of DFP were diluted in 215 fish water and DFP concentrations were determined until 6 hours. Results showed that diluted DFP 216 was stable in fish water, with an average 2% loss per hour, approximately ( Figure S1). Next, to 217 determine in vivo DFP toxicity in zebrafish, 5 days post-fertilization (dpf) larvae were exposed to 218 15, 20, 30, and 50 µM DFP and studied over a 24 h period. Results showed that all larvae incubated 219 in 20 µM DFP or in higher concentration, either died prior to 6 h exposure or displayed gross 220 phenotypic defects, including a curly tail and marked reductions of the head's and eyes' volumes 221 ( Figure S2). As we sought to investigate DFP neurotoxicity and subsequent brain damages, we 222 selected 15 μM DFP and an exposure time of 6 h ( Figure 1B), an experimental setup that did not 223 induce any visible phenotype (n = 20) when compared to control larvae exposed to 1% DMSO 224 (n = 20) (Figure 1C), nor any significant increase in larvae lethality. Histopathological analysis 225 confirmed that larvae exposed to 15 µM DFP showed no visible neurological abnormalities ( Figure  226 S3). 227 Figure 1. DFP-exposed zebrafish larvae displayed reduced motility, AChE inhibition and 229 respiratory failure. A: In the experimental set-up, 5 dpf larvae were exposed to 15, 20, 30 and 50 230 µM DFP, and larvae lethality, phenotypic defects, locomotor activity and AChE activity were 231 studied for 6 h. B: Lethality rates of 5 dpf larvae exposed for 6 h to 15, 20, 30 and 50 µM DFP led 232 us to select 15 µM DFP as optimal concentration (LC20). C: 5 dpf larvae exposed for 6 h to either 233 15 µM DFP or vehicle (DMSO), are phenotypically indistinguishable. D: Quantification of AChE 234 activity in larvae exposed to either 15 µM DFP (n = 5) or vehicle (DMSO) (n = 5), for 2, 4, and 6 235 h (Student unpaired t-test: **, p < 0.01; ***, p < 0.001). E: Locomotor activity of 5 dpf larvae 236 exposed to either 15 µM DFP (n = 48) or vehicle (DMSO) (n = 48) (Mann-Whitney test: *, p < 237 0.05). F: Real-time measurement of oxygen consumption by 5 dpf larvae exposed to either 15 µM 238 DFP or vehicle (DMSO). G: Quantification of oxygen consumption rate (OCR) of larvae exposed 239 to either 15 µM DFP (n = 301) or vehicle (DMSO) (n = 189) (Student unpaired t-test: ***, p < 240 0.001). Abbreviations: Ey, eye; Br, brain; Ea, ear. 241

AChE inhibition, leading to paralysis and respiratory failure in DFP exposed larvae 242
As AChE inhibition and muscle paralysis are hallmarks of OP poisoning, we measured the 243 motor activity of DFP-exposed and control larvae, by measuring the distance swum by these 244 individuals over a 30-min period. As expected, larvae exposed to DFP (n = 48) showed 245 significantly decreased motor activity compared to their control siblings (n = 48) ( Figure 1D). We 246 next estimated AChE activity in larvae exposed to either 15 µM DFP (n = 5) or vehicle (DMSO 247 1%) (n = 5). We observed a 50% inhibition of AChE activity as early as 2 h after DFP exposure 248 ( Figure 1E). As respiratory failure is an early consequence of OP poisoning 28,29 , we next quantified 249 the respiration of larvae exposed to 15 µM DFP by calculating the extracellular oxygen 250 consumption rate (ORC) of living larvae using the MitoXpress Xtra oxygen consumption assay, a 251 simple kinetic measurement of oxygen consumption ( Figure 1F). Results showed that the OCR of 252 larvae exposed to 15 µM DFP were 88.78 ± 1.00% of that observed in controls ( Figure 1G, p < 253 0.001). Thus, after DFP exposure, zebrafish larvae displayed strong inhibition of AChE activity, 254 decreased oxygen consumption, and reduced motor activity. 255

DFP exposure promoted neuronal hyperexcitation and apoptosis 256
Increase in c-Fos expression, a molecular marker of activity, is observed during seizures 30 , 257 and has been shown to promote epileptogenesis 31-34 . As a first attempt to evaluate the consequences 258 of DFP poisoning on neuronal activity, we studied c-Fos expression in DFP-treated and control 259 sibling larvae. Interestingly, qRT-PCR revealed a significant increase in c-Fos mRNA 260 accumulation following DFP poisoning ( Figure 2B, p < 0.001), suggesting increased neuronal 261 excitation. We then sought to visualize neuronal activity in live brains of larvae exposed to DFP 262 using calcium imaging. Indeed, transient calcium uptakes in neurons, as revealed by GCaMP5G 263 fluorescent protein, fully correlates neuronal excitation in zebrafish epilepsy models, allowing to 264 visualize seizures in vivo at the level of a whole brain 11, 35 . Five dpf larvae from the transgenic line 265 Tg[Huc:GCaMP5G] were treated with DFP and neuronal activity was recorded during the 6 h of 266 incubation time using time-lapse confocal microscopy ( Figure 2E, and supplementary video 2). As 267 early as 20 min following DFP addition, some intense transient calcium uptake events were 268 detected; their number and intensity progressively increased over the next 2 hours ( Figure 2F, G). 269 Then, 2 -3 hours following DFP addition, all DFP-treated larvae (n = 5) displayed massive, brief 270 and synchronous calcium uptake events in both neuropils of the optic tectum neurons, strongly 271 reminiscent of those seen during generalized seizures in zebrafish epilepsy models ( Figure 2C2). 272 To confirm that the increase in neuronal calcium uptakes observed in DFP-exposed larvae 273 did correspond to actual neuronal hyperexcitation, we ascertained whether administration of 274 diazepam, the main drug administered to epileptic patients to relieve seizures 36 , alleviated calcium 275 uptake activity. Larvae were first exposed to DFP (15 μM) for 5 h, and their neuronal calcium 276 activity was recorded for 30 minutes; diazepam (40 μM) was then added and their neuronal calcium 277 activity was recorded for an additional 60 min. Interestingly, exposure to diazepam significantly 278 decreased neuronal excitation induced by DFP ( Figure 2H, I). All these data confirm that DFP 279 exposure caused an intense neuronal hyperexcitation. Tg[Huc:GCaMP5G] larvae were exposed to either 15 µM DFP or vehicle (DMSO), and transient 283 calcium uptakes were then recorded in brain neurons over 6 h using calcium imaging. B: qRT-PCR 284 demonstrated markedly increased expression of the C-Fos gene in larvae exposed to 15 µM DFP 285 (n = 19) when compared to that in control larvae (DMSO) (n = 19) (Student unpaired t-test: ***, p 286 < 0.001). C, Snapshot views of calcium imaging of 5 dpf Tg[Huc:GCaMP5G] larvae brain showing 287 baseline calcium activity (C1 in Figure 2D) and seizure-like hyperactivity (C2 in Figure 2E  Exposure to highly toxic OP compounds (soman, sarin, VX) has been shown to cause 301 elevated neuronal loss in both humans and animal models 24,37-40 . We thus examined whether the 302 larvae exposed to DFP showed an increase in neuronal death. Using acridine orange (AO), a vital 303 marker that labels dying cells, we first observed in living larvae a marked increase in the number 304 of cells showing AO staining in DFP-exposed larvae compared to controls ( Figure 3C, D, E). Anti-305 activated-caspase-3 immunolabeling confirmed that DFP exposure promotes neuronal apoptosis 306 ( Figure 3F, G, H). sub-unit, a major excitatory glutamate receptor, was affected following DFP exposure, brain 326 sections of larvae exposed to DFP and control siblings were analyzed by immunocytochemistry 327 using an anti-NR2B-NMDA receptor antibody ( Figure 4A, excitatory synapse). Interestingly, 328 results showed a clear increase in NR2B-NMDA accumulation in the brains of DFP-exposed 329 larvae, compared to untreated controls ( Figure 4C, Supplementary video 3 and 4). The increased 330 NR2B-NMDA accumulation induced by DFP poisoning was confirmed by quantification of 331 NR2B-NMDA staining using Imaris software (Bitplane Inc., Version 9.1.2) ( Figure 4G). To further 332 characterize neuronal networks in DFP-treated larvae, we analyzed the accumulation of glutamate 333 decarboxylase (GAD65/67), an enzyme involved in GABA synthesis in presynaptic inhibitory 334 synapses, and gephyrin, a protein that anchors postsynaptic GABA receptors to the cytoskeleton, 335 using anti-GAD65/67 and anti-gephyrin antibodies, respectively ( Figure 4A, inhibitory synapse). 336 Interestingly, following DFP exposure, we observed a significant decrease in the accumulation of 337 both GAD65/67 ( Figure 4D) and gephyrin ( Figure 4F). Labeling quantification using Imaris 338 software (Bitplane Inc., Version 9.1.2) confirmed the decreased accumulation of both GAD65/67 339 ( Figure 4H) and gephyrin in DFP-treated larvae ( Figure 4I). 340 341 Figure 4. DFP exposure provoked increased NR2B-NMDA receptor accumulation. A: As 342 experimental set-up, 5 dpf larvae were exposed to either 15 µM DFP or vehicle (DMSO) for 6 343 hours, prior to NR2B-NMDA immunolabelling of glutamatergic/excitatory synapses. B: Scheme 344 of 5 dpf larvae head with the red box showing the region of interest in the brain uncovering the 345 optic tectum (OT). C: Anti-NR2B-NMDA receptor immunolabelling of glutamatergic synapses 346 (C1 to C2') and DAPI staining (C1', C2') in 5 dpf larvae exposed for 6 hours to either vehicle 347 (DMSO) (C1, C1') or 15 µM DFP (C2, C2'). Scale bar: 10 µm. 3D image reconstruction of NR2B-348 NMDA labeled neuron branch details in 5 dpf larvae exposed for 6 hours to either vehicle (DMSO) 349

DISCUSSION 363
Because of their use for agricultural purposes worldwide, acute poisoning by OP 364 compounds is a major public health problem, with several millions of intoxications reported each 365 year 44,45 . In this work, we took advantage of the possibilities offered by zebrafish larvae to develop 366 an animal model of OP poisoning and study the consequences of OP exposure on neuronal network 367 activity. Interestingly, as described in mammalian models of OP poisoning 46,47 , zebrafish larvae 368 exposed to DFP displayed marked AChE inhibition, the hallmark of OP intoxication, validating 369 this small fish as a good model for investigating the consequences of OP poisoning. One of the 370 most devastating features of OP intoxication in both humans and rodents is full, sometimes fatal, 371 respiratory failure 28 . It is important to note that in epileptic OP intoxication models that use 372 mammals, the induced respiratory failure must be prevented by the simultaneous addition of 373 cholinergic inhibitors (atropine) and AChE reactivators (oximes) to avoid premature death 29 . By 374 contrast, although we observed that DFP-treated larvae showed significantly decreased oxygen 375 consumption, there was no need to protect larvae with cholinergic inhibitors. Thus, DFP-exposed 376 zebrafish larvae appeared as a powerful and simple model to test the effects of anti-convulsive 377 agents in absence of either muscarinic antagonists or cholinesterase reactivators. 378 It has been demonstrated that acute OP exposure causes epileptic-like seizures, which, if 379 not treated, can eventually lead to life-threatening status epilepticus 47 . We therefore investigated 380 neuronal excitation in larvae exposed to DFP. We first found that larvae exposed to DFP showed 381 increased expression of c-Fos, a marker of neuronal activity, which is overexpressed after 382 seizures 48-50 . We next recorded neuronal calcium uptakes in living larvae exposed to DFP, a 383 technology that enables visualization of epileptic seizures 11, 35 . In DFP-exposed larvae, as early as 384 20 minutes following OP addition, we observed neurons showing massive calcium uptake events 385 that were never seen in control siblings, and which number increased over the next 2 h. Moreover, 386 the OP-induced neuronal activity was potently alleviated by diazepam treatment, confirming that 387 larvae exposed to DFP show neuronal hyperexcitation reflecting epileptiform seizures. In humans, 388 if victims are not treated within the first 30 minutes, seizures caused by OP intoxication can end in 389 status epilepticus, a major life-threatening neurologic disorder 6,51 , also leading to long term brain 390 damages 52,53 . This 30-minute long status epilepticus window frame also appears to be a critical 391 period during which long-term brain lesions are generated 51 . In the Tokyo subway attack, 392 approximately 3% of OP-poisoned victims suffered convulsions 54 . Interestingly, 2 -3 h after DFP 393 exposure, we observed that all the DFP-exposed larvae showed massive synchronous calcium 394 uptake events, strongly reminiscent of generalized seizures seen in zebrafish epilepsy models 11, 35 , 395 suggesting that these larvae displayed a status epilepticus-like phenotype. 396 Together with AChE inhibition and neuronal seizures, massive neuronal death is another 397 hallmark of OP poisoning 20 , also observed in the DFP-exposed zebrafish larvae. At the cellular 398 level, it has long been known that hyperactivity of cholinergic receptors induces a massive release 399 of glutamate, leading to over-activation of glutamatergic receptors and neuronal 400 hyperexcitability 41-43 . Specifically, it has been shown that acute OP intoxication induces the 401 activation of NMDA receptors 55 . Moreover, in a mammalian model of OP poisoning, activation of 402 NMDA receptors plays essential roles in seizure activity and apoptosis 52,56 . In the brain of zebrafish 403 larvae exposed to DFP, we observed a decreased accumulation of both gephyrin and GAD65/67, 404 two proteins specifically accumulated in inhibitory synapses, while NR2B-NMDA receptor was 405 significantly overexpressed. This suggests that following acute DFP poisoning, neuronal 406 hyperexcitation results from a shift in the synaptic balance of brain neurons toward excitatory 407

states. 408
We report here a vertebrate model of OP poisoning that displays phenotypes and symptoms 409 of acute toxicity, faithfully recapitulating those described in humans, i.e. AChE inhibition, 410 respiratory deficit, neuronal apoptosis and epileptiform seizures. The zebrafish is thus a model of 411 choice for large-scale screening of entities that could restore CNS functions after OP poisoning and 412 mitigate the long-term neurological sequelae of acute OP poisoning in humans. 413

DISCLOSURE 416
The authors declare that the research was conducted in the absence of any commercial or financial 417 relationships that could be construed as a potential conflict of interest.  Figure 4C1', showing the 440 optic tectum from a 5 dpf larva exposed for 6 hto vehicle (DMSO) and labeled with an anti-NR2B-441 NMDA antibody (green) and counterstained with DAPI (blue). 3D images were generated using 442 Imaris software (Bitplane Inc., Version 9.1.2). 443