Towards Zebrafish Models of CNS Channelopathies

Channelopathies are a large group of systemic disorders whose pathogenesis is associated with dysfunctional ion channels. Aberrant transmembrane transport of K+, Na+, Ca2+ and Cl− by these channels in the brain induces central nervous system (CNS) channelopathies, most commonly including epilepsy, but also migraine, as well as various movement and psychiatric disorders. Animal models are a useful tool for studying pathogenesis of a wide range of brain disorders, including channelopathies. Complementing multiple well-established rodent models, the zebrafish (Danio rerio) has become a popular translational model organism for neurobiology, psychopharmacology and toxicology research, and for probing mechanisms underlying CNS pathogenesis. Here, we discuss current prospects and challenges of developing genetic, pharmacological and other experimental models of major CNS channelopathies based on zebrafish.


Introduction
Membrane-bound ion channels play an important role in the development and functioning of multiple physiological systems, including the central nervous system (CNS) [1][2][3]. While transmembrane transport of K + , Na + , Ca 2+ and Cl − by these channels in the brain is key for the transmission of nerve impulses [1,4], experimentally induced (e.g., genetic or pharmacological) deficits of ion transport may lead to a wide range of systemic disorders termed 'CNS channelopathies' [2,5] (Table 1). The global prevalence of this disorder group is estimated as~35 per 100,000 [3], with most common pathologies including epilepsy, followed by migraine, as well as various movement-and psychiatric disorders [5].
Animal models are a widely used tool for studying neurological diseases, including channelopathies, as well as their mechanisms and strategies for therapy [6,7]. Complementing rodent evidence, the zebrafish (Danio rerio) is rapidly gaining popularity as a translational model organism in biomedicine, including neurobiology and preclinical genetic and CNS drug screening [8,9]. Due to substantial genetic and physiological homology to humans, rapid growth and development, high fertility and medium-to-high throughput, zebrafish have become a powerful system for studying CNS pathologies [10,11]. The presence of all major mediator systems and neurotransmitters, as well as their receptors and transporters, in zebrafish makes them suitable for modeling of a wide range of CNS channelopathies. Zebrafish also demonstrate the ease of behavioral, genetic and pharmacological manipulations, hence further fostering the development of sensitive models of SCN1B, SCN1A [2,24] Generalized epilepsy with febrile seizures plus type 2 (GEFS+2) Generalized tonic/clonic convulsions, hyperexcitability SCN1A [25,26] Benign familial infantile epilepsy Partial seizures in occipito-parietal areas, slow deviation of head and eyes to one side, cyanosis, unilateral limb jerks SCN2A [27] Cognitive impairment with or without cerebellar ataxia Dementia, poor verbal memory and executive dysfunction, visuospatial dysfunction, intellectual impairment, limb ataxia, dysarthria, loss of proprioception, ataxia SCN8A [28] Congenital indifference to pain, autosomal-recessive Insensitivity to pain, absence of physiological responses to noxious stimuli SCN9A [29]
Voltage-gated Ca 2+ channels play an important role in the brain, regulating neuronal excitability and the release of neurotransmitters [65]. Ca 2+ channels include L (long lasting), N (neuronal), P (Purkinje cell), Q (granule cell), R (toxin-resistant), and T (transient) types [1]. The L-, N-, P-, Q-and R-types of Ca 2+ channels regulate neurotransmitter release [1], whereas mutations in the P-and Q-type channel genes cause ataxia and migraine in both human and animals, due to their high density in cerebellar Purkinje and granule cells that control locomotion [1]. Mutations in these channels commonly lead to overt epilepsy and ataxia phenotypes in several animal models. For example, mouse leaner mutants show seizures similar to human absence seizures, and progressive cerebellar degeneration caused by splicing mutation impact on the α subunit of the P/Q-type Ca 2+ channel [1]. The tottering mutant mice also demonstrate absence-like seizures, whereas rolling mutants display no seizure activity, but poor limb coordination, falling and rolling over [1,66]. Mouse lethargic mutants exhibit absence-like seizures and ataxia without neurodegeneration, stargazer mice show spike-like wave seizures (common in absence seizures) in the cerebellum [1], and Roker mice present absence-like seizure and ataxia resembling the tottering mouse phenotype discussed above. Interestingly, ablating the P/Q-type Ca 2+ channel in the serotonergic neurons increases aggression in male mice without affecting their anxiety, which not only implicates these channels in the regulation of some evolutionarily conservative behaviors (e.g., aggression), but suggests novel ways of alleviating aggressive behaviors in people with CNS channelopathies [67].
Idiopathic or primary epilepsy is typically caused by mutations in genes encoding ion channels [1], including neuronal nicotinic acetylcholine receptor (nAChR) subunits, as well as voltage-gated K + (Kv), Na + , and Ca 2+ channels. A popular genetic rodent model of idiopathic epilepsy involves WAG/Rij rats (an outbred subline of Wistar rats) developed in the 1920s [77]. WAG/Rij rats can also be a valid genetic animal model of absence seizures [78], given their similar EEG and behavioral patterns with those in human absence seizures. Autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE), BFNC and severe myoclonic epilepsy in infancy (SMEI) are common types of idiopathic epilepsy [79], whose candidate genes include CACNA1A, CACNA1G, CACNA1H, CACNA1I, CACNAB4, CACNAG2 and CACNG3 that encode subunits of Ca 2+ channels, and GABRG2, GABRAB3, GABRA5, GABA(B1) and GABA(B2) that encode subunits of the Cl − channels on the gamma aminobutyric acid (GABA)-A receptors [80], respectively.
Zebrafish have become popular model organisms in translational neuroscience, physiology and preclinical screening mostly because of their low cost, easy maintenance, as well as high genetic and physiological homology with mammals, including humans [81]. Fast development and transparent embryos make zebrafish a useful tool for genetic manipulations, with multiple established genetic models of CNS disorders, including autism [82], schizophrenia [83], attention deficit hyperactivity disorder (ADHD) [84] and epilepsy [81]. Zebrafish are also susceptible to a wide range of pharmacological agents, including a conventional GABA-lytic drug pentylenetetrazole (PTZ) that blocks Cl − ionophore on GABA-A receptors and induces seizures in zebrafish that resemble those observed in mammals [85].
One of the best-characterized genetic epilepsy models in zebrafish is presently an Nethyl-N-nitrosourea (ENU)-generated model of Dravet's syndrome, involving mutation in the SCN1A gene that affects voltage-gated Na + channels [86] (Table 2). For example, the scn1Lab zebrafish mutants have spontaneous seizures that resemble those in patients with Dravet's syndrome, both behaviorally and electrophysiologically [85]. Another interesting zebrafish model of epilepsy is based on the kcnj10 morphants, recapitulating the EAST (Epilepsy, Ataxia, Sensorineural deafness, Tubulopathy) syndrome [87]. KCNL10 encodes a K + channel protein that regulates K + levels in glial cells and hence modulates their excitability [88]. Other promising candidates for genetic models of epilepsy in zebrafish involve temperature-gated 'transient receptor potential vanilloid' (TRPV) TRPV1 and TRPV4 ion channels. TRPVs are a family of ion channels with high Ca 2+ permeability, thus regulating neuronal excitability in both mammals [89,90] and zebrafish [91]. Activation of thermosensitive TRPV1 and TRPV4 ion channels generally increases pathological excitability of neurons and causes febrile seizures in zebrafish [92]. Ataxia is another complex neurological symptom that involves poor coordination of movements or gait, and some forms of which are commonly caused by channelopathies. For example, ataxia is observed in both clinical EAST syndrome and in zebrafish knockdown of kcnj10, which presents as abnormal swimming, frequent spontaneous tail flicks, circling, bursts of speed and other aberrant movements [102]. Spinocerebellar ataxia type 6 (SCA6) is a neurological disease resulting from a mutation in CACNA1A [103]. Due to teleostspecific duplication of the zebrafish genome, there are two gene copies encoding the CaV2.1 Ca 2+ channels, cacna1aa and cacna1ab. Two mutations of the cacna1ab gene include tb204a [104] and fakir [105], both disrupting Ca 2+ channel function in zebrafish [106]. The tb204a mutation (which replaces tyrosine with asparagine (Y1662N) at the carboxyl end of the CaV2.1a channel) reduces the mobility of zebrafish larvae and lowers intracellular Ca 2+ in presynaptic neuromuscular junctions [104]. The fakir cacna1ab mutation (referred to as L356V, substituting leucine 356 with a valine), impairs larval locomotion and causes depolarizing shifts during the Ca 2+ channel activation [105]. Although both cacna1a and -b mutations in zebrafish have similar effects on locomotion, there are some differences in the mechanism of occurrence of these deficits: while fakir mutants have decreased touch sensitivity and altered transmission between motor neurons and slow-twitch fibers in muscles [107], tb204a mutant fish demonstrate aberrant transmission in neuromuscular junction [104]. Spinocerebellar ataxia type 13 (SCA13) is neurodegenerative disease caused by mutated KCNC3 gene encoding the Kv3.3 K + ion channel, presenting as the atrophy of cerebellar neurons, motor dysfunction and intellectual disability followed by progressive ataxia in adulthood [108][109][110]. In zebrafish, kncn3 mutant larvae also display reduced startle response to touch [108], supporting the validity of this model of ataxia. As already noted, the Dravet's syndrome is a severe childhood epileptic encephalopathy due to a mutation in the SCN1A gene encoding voltage-gated Na + channels (Nav1.1) [111]. In addition, the SCN2A, SCN8A, GABRA1 and STXBP1 genes (that encode Na + , Cl − channels and syntaxin-associated proteins, respectively) also play a role in Dravet's pathogenesis [76,112]. Mutations within the SCN1A gene (encoding the Nav1.1 channel) disrupt Na + flow into neurons [113], hence affecting the action potential of the GABA-ergic neurons in brain structures that are critical for modulating seizure activity (e.g., in cortex and the hippocampus) [114]. The scn1lab zebrafish mutants show good survival rate, increased synaptic activity and reduced inhibitory tone in the brain [115]. In addition, by 7th day postfertilization (dpf), the scn1lab mut/mut fish show fewer GABA-ergic neurons and increased glutamatergic signaling [115]. Interestingly, the number of glial cells in these mutants is also markedly higher, which may be useful for studying (and, eventually, parsing) the role of glia in the pathogenesis of Na + channelopathies [115].

Towards Valid Zebrafish Models Relevant to CNS Channelopathies
Overall, ion channels play a critical role in functioning and control of various physiological systems, including CNS [1] (Table 1). Genetic mutations of K + , Na + , Ca 2+ and Cl − channels in neurons lead to severe cardiovascular, respiratory and CNS diseases [1,4] (Tables 1-4). The search for novel treatments for these prevalent and debilitating maladies represents an urgent unmet biological problem, necessitating reliable, throughput, sensitive and efficient animal models. As already discussed, the use of zebrafish to model CNS channelopathies is beginning to develop rapidly in translational research, also offering an indispensable and effective tool for rapid CNS drug screening. Indeed, zebrafish are highly sensitive to a wide range of major CNS drugs, including antidepressant [116], antiepileptic [117], antipsychotic [118], anxiolytic [119], nootropic [120] and other psychoactive substances [121,122], including addictive drugs, such as alcohol [123,124], nicotine [123,125], caffeine [126] and synthetic salts [127][128][129]. The low cost of maintenance, easy breeding, fast development cycle, the availability of "two models in one" (i.e., both larvae and adult fish), as well as the simplicity of treatment with acute and chronic substances, all make this model organism an indispensable tool for drug discovery and high-throughput preclinical screening [130].  Tottering mice Ca 2+ Episodic attacks of ataxia due to stress [141] Leaner mice Ca 2+ Degeneration of differentiated granule, Golgi, and Purkinje cells in cerebellum [141] Rolling Nagoya mice Ca 2+ Severe ataxia without motor seizures [141] V408A/+ mice K + Loss of motor coordination, increased frequency and amplitude of spontaneous GABAergic inhibitory postsynaptic currents (IPSCs), motor impairment, [142] Ducky mice Ca 2+ Ataxia, and paroxysmal dyskinesia, "ducky" gait [143] Migraine S218L mutant mice Ca 2+ Severe migraine with seizures, coma, severe cerebral edema following mild head trauma, propensity to multiple cortical spreading depression (CSD) events [68] R192Q knock-in mice Ca 2+ Increased glutamate release in the cerebellar cortex, decreased threshold for CSD, decreased neuronal response to nociceptive activation, photophobia; unilateral head grooming; lateralized winking/blinking, migraine-like headache [144][145][146] S218L knock-in mice Ca 2+ Exquisite sensitivity to CSD (reduced triggering threshold, increased propagation velocity and frequency of CSC events), ataxia, seizures, migraine-like headache, photophobia; unilateral head grooming; lateralized winking/blinking [145,146] Atp1a2 +/R887 mice Na,K-ATPase Decreased induction threshold, increased velocity of propagation CSD, higher susceptibility to fear and anxiety [147] A unique practical advantage of using zebrafish for pharmacological modeling of epilepsy-related phenotypes of channelopathies is the possibility of treating multiple fish with a water-diluted convulsant drug at once, evoking simultaneous and rapid development of seizures in these zebrafish and allowing a rapid screening of a wide range of anticonvulsants at the same time [148]. For example, as we calculated based on our own 20-year experience with experimental epilepsy models in both mammals and fish, a typical time-consuming full-day rodent experiment with systemic (e.g., intraperitoneal, i.p.) PTZ injection to induce seizures would require~10 times more time than a zebrafish study of a similar experimental design and sample size, where the drug is administered via water immersion to multiple fish at once, and several fish are recorded per trial, to assess their seizure-specific locomotor endpoints. The simplicity of pharmacological model of epilepsy also highlights the similarity of the observed phenotypes of zebrafish seizure behavior (i.e., similar clonic and tonic-clonic seizures, hyperlocomotion, corkscrew-like circular swimming and ataxia) to those in rodent models [149]-as discussed above, another highly relevant phenotype to modeling CNS channelopathies. Table 4. Selected zebrafish models of CNS channelopathies.

CNS Disorder Modulated Ion Channel Zebrafish Model Phenotypes References
Dravet's syndrome Na + scn1Lab mutants Seizures [85] scn1lab mutants Increased synaptic activity and reduced inhibitory tone in the brain [115] Epilepsy Ca 2+ kcnj10 morphants Epilepsy, ataxia and sensorineural deafness [87] Ca 2+ kcnj10 knockdown Abnormal swimming, frequent spontaneous tail flicks, circling, bursts of speed and other aberrant movements [102] Spinocerebellar ataxia type 6 Ca 2+ tb204a mutants Reduced the locomotor activity, accompanied by reduced intracellular Ca 2+ in presynaptic neuromuscular junctions [104] Spinocerebellar ataxia type 13 K + kcnc3 mutants Reduced startle response to touch [108] Furthermore, with the additional genome duplication event in teleost fishes, the zebrafish becomes an excellent model for various genetic manipulations and modifications, including the creation of morphants and knockout mutants of known human CNS diseases associated with specific mutations [81,150,151], including those affecting ion channels in the brain [152,153]. An important conceptual advantage of zebrafish models in this regard is that with partially duplicated genome, it becomes possible to create mutants whose genetic analogues in rodents are impossible due to the lethality of these mutations [130] (a situation not uncommon for mutations in genes encoding critical ion channels with so many key physiological functions). This aspect paves the way for the creation of more accurate model systems for studying the pathogenesis of the development of mutations in ion channels in aquatic vertebrates with 70+% genetic similarity to humans [154].
Furthermore, behavioral phenotypes of zebrafish are quite simple, clear-cut and generally comparable to rodents in all major domains [155]. There are also multiple behavioral tests and paradigms to assess zebrafish motor phenotypes, similarly to rodents, hence making behavioral phenotyping in general (including targeting epilepsy and, likely, other related CNS channelopathies) a simple, fast and efficient procedure [149,155]. In addition, zebrafish models seem to be more resilient (than rodent models) to variation in external testing procedures, which may contribute to increased reproducibility and reliability of laboratory data using zebrafish [156].
Interestingly, the use of a 3D system software for studying zebrafish behavior [157] enables the creation of libraries of locomotor responses to various pharmacological agents [155]. Despite the simplicity of demonstrated phenotypes, zebrafish have complex behavioral responses to various stimuli, including social behavior in a group, sufficient to study and model complex behavioral paradigms [158]. Developing high-efficiency software, along with 3D behavioral tracking, facilitates automated behavioral screening of chemical agents using artificial intelligence (AI) [159]. It is therefore likely that AI-based drug screening in zebrafish, including tracking of their tonic and clonic seizures in epilepsy simulation, will greatly advance preclinical studies of anticonvulsants and other drugs to treat channelopathies. Zebrafish also present several protocols to assess CNS channel activity or action potentials using various electrophysiological methods [160][161][162]. For instance, electrophysiological characterization of heterologously expressed zebrafish Kcnh1a and Kcnh1b channels reveals functional patterns similar to those of human K + channels [163]. Such methods, particularly useful and feasible in zebrafish with their smaller transparent brains, enable the evaluation of the action potential and direct linking aberrant (e.g., mutant) ion channel activity to CNS disorders.
Despite multiple obvious advantages, zebrafish models also have significant limitations as well. For example, no animal model is an exact reproduction of complex human CNS diseases [164]. While there is a large number of genetic models of CNS diseases in zebrafish, these animals are often unable to reproduce due to the loss of fertility. Another limitation of zebrafish models is the methods of administering the test substances, mainly by water immersion or i.p. injection, which complicates studying water-insoluble compounds and requires the use of special solvents [165]. Another important technical limitation of zebrafish as a model is the likely difference in dosing between humans and fish, which requires a specific recalculation of the used dose of the drug for preclinical studies when translating the results to humans [166,167].
The reproduction of core clinical symptoms in rodents and zebrafish represents yet another key translational issue with modeling CNS channelopathies in fish. For example, one of broad clusters of clinical signs of CNS channelopathies (beyond epilepsy) is migraine and headaches of various types and etiologies (Table 1) [4]. Indeed, K + channels are found mainly in the heart and brain, where they modulate and stabilize action potentials of cells by regulating K + current [168]. The loss of ion channel function due to mutation of the KCNK18 gene is a common cause of migraine [169]. In addition to K + channels, mutation in CACNA1A Ca 2+ channel gene contribute to the pathogenesis of type 2 episodic ataxia and hemiplegic migraine [170]. First, assessing the level of headache and the ability to accurately diagnose this condition in animal models are a challenging task [171]. For instance, in genetic models of migraine-like conditions in rodents, it is impossible to accurately assess the presence and the severity of headache. Thus, altered sensitivity of animals to various stimuli is rather tested instead, similar to clinical signs of migraine, such as photophobia, increased tactile and sound sensitivity, and preference for darkness.
It is therefore impossible to assume that migraine-like pain models in rodents cause migraines per se, and the same problem applies a priori to any such zebrafish models as well. In addition, when modeling migraine, only that accompanied by the aura lends itself to quantitative and qualitative assessment. In contrast, it seems that migraine without the aura cannot be reliably diagnosed in animal models. Accordingly, it remains unclear whether it is possible to simulate migraine and migraine-like pain in fish, and whether fish are even capable of experiencing headaches similar to those observed clinically. Likewise, we do not know what pharmacological and genetic tools can help model these conditions best in both rodents and fish.
In summary, the high global prevalence of diseases directly related to abnormal ion channels necessitates our improved understanding of the pathological mechanisms underlying CNS channelopathies, as well as the development of novel pharmacological agents for their treatment. For this, zebrafish models can be a useful tool for identifying evolutionary-conservative mechanisms underlying CNS channelopathies, and a powerful efficient drug-screening platform, to induce or prevent these conditions in vivo. With multiple conceptual and methodological questions remaining open in the field (Table 5), this collectively calls for a wider use of zebrafish models in studying CNS channelopathies. Table 5. Selected open questions related to zebrafish-based modeling of CNS channelopathies.

•
Given generally high overall genetic homology (70%) of zebrafish to humans, do CNS channelopathies share similarly overlapping genetic substrate between the two model organisms? • Zebrafish have a generally high (85+%) percentage of gene orthologs for known human-disorder-associated genes. Is there a similar (i.e., comparably high) percentage of disorder-associated genes for CNS channelopathies? • What is the percentage of unique, human-specific channelopathies that are not present in zebrafish, and vice versa? • What is the exact impact of teleost fish-specific genome duplication on the expression of disordered phenotypes in CNS channelopathies? • What are reliable behavioral or neurological tests that can distinguish between mental and neurological manifestations of CNS channelopathies in experimental animal models in general, and in zebrafish in particular? • Are there overt sex and strain differences in the pathogenesis of CNS channelopathies in zebrafish models? • Is there a direct correlation between physiological aspects of CNS channelopathies (e.g., Ca 2+ channel modulation) and behavioral phenotypes (e.g., seizure)? If so, is this correlation translational (e.g., from zebrafish to humans)? • How can specific diet (e.g., high-Na + or low-Ca 2+ diet) influence various phenotypes of CNS channelopathies? Can other dietary factors (e.g., high-fat/high-sugar diet) play a role in CNS channelopathies? • What non-pharmacological therapies can be effective for CNS channelopathy treatment? Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.