Classification of advanced methods for evaluating neurotoxicity

As fields such as neurotoxicity evaluation and neuro-related drug research are increasing in popularity, there is a demand for the expansion of neurotoxicity research. Currently, neurotoxicity is assessed by measuring changes in weight and behavior. However, measurement of such changes does not allow the detection of subtle and inconspicuous neurotoxicity. In this review, methods for advancing neurotoxicity research are divided into molecule-, cell-, circuit-, and animal model-based methods, and the results of previous studies assessing neurotoxicity are provided and discussed. In coming decades, cooperation between universities, national research institutes, industrial research institutes, governments, and the private sector will become necessary when identifying alternative methods for neurotoxicity evaluation, which is a current goal related to improving neurotoxicity assessment and an appropriate approach to neurotoxicity prediction. Many methods for measuring neurotoxicity in the field of neuroscience have recently been reported. This paper classifies the supplementary and complementary experimental measures for evaluating neurotoxicity.


Introduction
According to a global burden disease study, from 1990 to 2016, there was a 29.9% increase in years lived with disability (YLDs) for subjects with bipolar disorder (from 376.4 to 489.1 YLDs in thousands), a 65.6% increase for dementia (from 360.8 to 597.6 YLDs in thousands), a 13.6% increase for alcohol-related disorders (from 558.2 to 633.9 YLDs in thousands), a 36% increase for schizophrenia (from 503.3 to 685.2 YLDs in thousands), a 30.5% increase for opioid use disorders (from 1256.2 to 1638.9 YLDs in thousands), a 30.8% increase for anxiety disorders (from 1341.7 to 1755.0 YLDs in thousands), a 27.2% increase for migraine (from 1580.3 to 2010.1 YLDs in thousands), and a 27.0% increase for major depressive disorder (from 1726.2 to 2193.0 YLDs in thousands) (Collaborators, U.S.B.o.D. et al. 2018).
Neuromodulation technologies are required for the assessment of treatments and drugs administered for various diseases. Such technologies are used to improve cognition in stroke or dementia patients, aid recovery from mental illness, and rapidly assess the effects of drugs.
Current methods used to measure neurotoxicity involve assessing the effects of drugs on the central nervous system (CNS) by evaluating behaviors. However, behaviors represent many physiological functions and therefore cannot be used to assess specific neurotoxic effects. As many studies in neuroscience are being performed, various methods for measuring neurotoxicity are being developed. Systems biology, bioinformatics, and rapid assay technologies may be alternatives to the existing behavioral tests (Gibb 2008). Methods for assessing the risk of neurotoxicity are shifting from traditional animal toxicity tests to various mechanistic studies designed to clarify the toxicity mechanism and identify techniques for coping with the adverse effects of toxic substances (Krewski et al. 2009). In this paper, we present a classification system for various methods that may be used to assess neurotoxicity.

Synapses as targets for assessing neurotoxicity
The evaluation of drug safety requires measuring CNS, cardiovascular system, and respiratory system functions. In particular, current methods used for evaluating the safety of drugs in the CNS primarily involve behavior-based assessments. The behavioral tests commonly used for rodents are the open field test, which measures exploratory behavior and general activity, and the Morris water maze test, which assesses spatial learning and memory. Unfortunately, behavioral evaluations in the laboratory can be subjective and nonreproducible due to the characteristics of the person performing the experiment and various environmental factors, including housing conditions. Furthermore, behavior involves the integration of complex processes, including the actions of various molecules, synaptic activity, and a summation of various neural circuits; thus, behavioral results are nonspecific. Therefore, there is a need to specifically study each step in the behavior being assessed by separating the complex stages of the behavioral output. The first specific target could be the synapses that allow neurons to communicate with other neurons through the transmission of a chemical or electrical signal. In the field of neuroscience, identifying the cellular mechanism underlying synaptic modification helps to elucidate behavioral changes. Synapses contain various ion channels that regulate neuronal activity and synaptic plasticity, and they are directly linked to behavioral changes, so they could be considered principal targets for neurotoxicity assessments.

Ion channels
AMPARs and NMDARs are ionotropic glutamate receptors that regulate neuronal activity by mediating excitatory synaptic transmission in the brain. AMPAR-and NMDAR-mediated synaptic transmission triggers long-term changes in synaptic plasticity, leading to changes in behavior (Bredt and Nicoll 2003;Malinow and Malenka 2002;Newpher and Ehlers 2008). GABA, the major inhibitory neurotransmitter, reduces neuronal excitability throughout the nervous system, and GABAA receptor-mediated neurotransmission is critical for various behavioral actions related to ethanol and is associated with anesthesia, sedation, hypnosis, and anxiolysis (Lobo and Harris 2008). The GABAA receptor has also been reported to be involved in drug addiction behavior, including nicotine dependence (Agrawal et al. 2008a), cannabis abuse (Agrawal et al. 2008b), cocaine addiction (Dixon et al. 2010), and heroin dependence (Enoch et al. 2010), in both human and animal studies. In addition, through the use of genetic, behavioral, and pharmacological techniques, it has been shown that a GABAA receptor subunit modulates anxiety-and depression-related behaviors (Smith and Rudolph 2012). The transient receptor potential (TRP) channel is involved in the nonspecific conductance of cations and is highly permeable to Ca 2+ . The activation of TRP channels is directly related to nociceptive transmission, mechanosensation, axonal guidance, and gastrointestinal sensation (Sadler and Stucky 2019;Ramsey et al. 2006;Yu et al. 2016). Measurement of currents flowing through channels in the synapse using electrophysiological methods can directly and specifically assess neurotoxicity under different conditions.

Synaptic plasticity
Various changes in behavior, including long-term potentiation (LTP) and long-term depression (LTD), most likely depend on synaptic plasticity, which can be mediated by trafficking or removal of AMPARs. Activity-dependent synaptic plasticity is induced at appropriate synapses during memory formation and is critical for information storage (Kandel and Schwartz 1982;Lynch and Baudry 1984;Korol et al. 1993;Morris and Frey 1997). NMDAR-dependent synaptic plasticity in hippocampal synapses is critically involved in hippocampus-dependent spatial learning and memory. The first studies of the role of LTP in memory formation demonstrated that the persistence of LTP is significantly correlated with the rate of learning and spatial memory ability over time (Barnes 1979;Barnes and McNaughton 1985). Synaptic plasticity is also involved in amygdala-dependent fear learning and memory. The fear-conditioning paradigm, which involves memory formation, induces a form of LTP in the amygdala (Rogan et al. 1997;Rogan and LeDoux 1995;McKernan and Shinnick-Gallagher 1997). Furthermore, several studies have suggested that changes in behavior caused by addictive drugs can be mediated by synaptic plasticity in the mesolimbic reward system (Kauer and Malenka 2007).

Action potentials
Action potentials (APs), which are generated by transient changes in the movement of Na + and K + ions across the neuronal membrane, are required for neural communication. APs 1 3 contain information based on a diverse coding scheme. In motor neurons, the frequency or rate of APs determines the force exerted by a muscle during a voluntary contraction; this coding scheme is called the 'rate code'. The rate code model of neuronal communication through APs suggests that the rate of APs increases as the intensity of a stimulus increases (Stein et al. 2005;Gerstner et al. 1997;Adrian and Zotterman 1926). On the other hand, according to the 'temporal code' scheme, information is transmitted in the auditory, visual, and somatosensory systems via the precise timing of APs (Gerstner and Kistler 2002;Dayan and Abbott 2001). Several studies have shown that the temporal resolution of APs on the millisecond scale is critical for transmitting neural information (Eckmiller et al. 1990;Butts et al. 2007;Gollisch and Meister 2008). Another neural coding scheme for neuronal communication through APs is the 'synchrony code', in which APs are synchronously generated within less than a millisecond. Purkinje cells in the cerebellum synchronize their APs for information coding during behavior. A working hypothesis is that the synchronous APs of Purkinje cells regulate the activity of postsynaptic neurons, thereby transmitting information from the cerebellum (Gauck and Jaeger 2000;Bell and Grimm 1969;Person and Raman 2011;Han, et al. 2018). Measurement of APs using electrophysiological methods can be used to directly assess neurotoxicity related to cellular mechanisms and behaviors that are encoded by frequency, timing, or synchronicity of AP activity.

High-throughput assays
Methods for recording spikes (Kodandaramaiah 2018) and evaluating the activity of sodium channels (Zhang et al. 2020), potassium channels (Kutchinsky et al. 2003), chloride channels (Billet et al. 2017), and cation channels (McPate et al. 2014) through an automated patch-clamp technique are well established. However, such methods require professional-level quality control and can cost a considerable amount of money; thus, they are challenging to perform in small laboratories.
Microelectrode arrays (MEAs) are in vitro measurement devices used to measure the rhythmicity of heart cells. Such devices are relatively easy to use; thus, there is a demand for their use when assessing neuronal excitability (Malerba et al. 2018;Kizner et al. 2019;Taga et al. 2019). Human-derived stem cells are required for MEA-based research technology.

In vivo assays
Electroencephalography (EEG) is an electrophysiological method used to record fluctuations in ionic current voltages and, thus, to assess neuronal activity in vivo (Niedermeyer and Silva 2005). Even though the electrical activity of a single neuron cannot be detected, the summed activity of numerous neuronal populations can be measured by EEG. Generally, EEG has been used to assess neuronal activity in the context of assessing epilepsy, seizure disorders, and other conditions, including brain tumors, brain damage, brain inflammation, stroke, and sleep disorders. The advantages of EEG are its lower start-up cost compared to other techniques (Vespa et al. 1999), its high temporal resolution (Montoya-Martinez et al. 2021), and it does not require the use of radioligands (Yasuno et al. 2008). Furthermore, EEG can be performed simultaneously with neuroimaging techniques such as functional magnetic resonance imaging (fMRI) and positron emission tomography (PET). Therefore, the recording of electrical signals via EEG in vivo might be useful when assessing the association of neurotoxicity with various diseases.

Imaging-based assays
Imaging-based assays indirectly measure ionic movement based on membrane potential or ion concentration changes and provide more spatial information than that obtained from the patch-clamp technique. Voltage changes are measured by voltage-sensitive dyes that exhibit potential-dependent changes in their properties. Bis-trimethine oxonol (DiBAC 4 (3)) has been applied in high-throughput screening and optimized for detecting steady-state rather than kinetic changes in potential (Yamada et al. 2001). FMP, another membrane potential dye, shows a high level of temporal resolution when measuring kinetic changes (Whiteaker et al. 2001). Fluorescence resonance energy transfer (FRET)-based assays are also widely used to measure membrane hyperpolarization or depolarization. Specific ion concentrations are measured with ion-specific probes, including Fura-2, Fluo-3, and Fluo-4 as Ca 2+ indicators , FluxOR and PBFI for K + ions (Beacham et al. 2010;Meuwis et al. 1995), and SBFI for Na + ions .
Various neurotransmitters, such as glutamate, dopamine, GABA, 5-HT, acetylcholine, and norepinephrine, (Sun et al. 2018(Sun et al. , 2020 can be monitored with sensors that provide high sensitivity and spatiotemporal resolution (Banerjee et al. 2020;Sun et al. 2018Sun et al. , 2020. The incorporation of nanomaterials, biomolecules, and polymers is helpful in the development of sensors for neurotransmitters. In addition, a voltammetry-based approach for electrochemical sensing has been recently developed for in vivo measurement.

Cell-based toxicity assays
In some cases, the cause of the neurological disease may be the death of nerve cells within the brain. Therefore, measurement of neuronal viability can be used to assess 1 3 neurotoxicity. Prenatal exposure to lipopolysaccharide induces depression-like behavior by reducing the number of BrdU-positive cells, which are newly generated proliferating neuronal cells (Lin and Wang 2014) (Table 1).
It has been reported that neuronal differentiation is associated with neurodevelopmental toxicity. Class 3 betatubulin, a marker of neuronal differentiation, has been used to detect neuronal differentiation to determine whether 9-cis retinoic acid, methylmercury, and 22(R)-hydroxycholesterol affect the differentiation of NT/D1 cells with neuronal stem cells characteristics (Taylor et al. 2019).
The balance between neuronal proliferation and apoptosis, a form of programmed cell death, has an important effect on normal brain development. Therefore, toxic substances, which can be produced due to apoptosis, can affect normal brain morphology and function. The enzymatic activity of caspase-3/7 has been used to assess apoptosis (Druwe et al. 2015), whereas β-tubulin type III has been used as a molecular marker of the migration of human Ntera2 (NT2) cells (Stern et al. 2014). Disorganizations within the ventricular zone, the neuroependyma, radial glial scaffolds, or tangential fibers have been used  Borchelt et al. 1997;Holcomb et al. 1998;Dawson et al. 2010 as markers of neuronal migration defects (Moroni et al. 2011).

Neural circuit-based toxicity assays
Neural circuits can be assessed by measuring neurite growth, synaptogenesis, and myelination (Brat and Brimijoin 1992). Glycidol exposure increases the gene expression of Kl, Igf2, and Igfbp2 in the corpus callosum to promote myelination, while Efnb3, Tnc, and Cd44 gene expression increases in the cingulate cortex promote neurite growth (Akane et al. 2014). The safety of galactose (Powell et al. 1986) and methylmercury (Parran et al. 2003) has been assessed by measuring neurite growth. The thickness of myelin around axons and the nodal gap length can be determined by measuring the expression of myelin-mediated molecules such as pan-sodium and Caspr1 (Suminaite et al. 2019;Dutta et al. 2018).

Animal model-based toxicity assays
Cuprizone injection has been used in the establishment of an animal model of multiple sclerosis through the induction of demyelination (Torkildsen et al. 2008). Injections of 6-hydroxydopamine, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine, paraquat, and rotenone have been used to generate animal models of Parkinson's disease through induction of dopaminergic neuron death in the substantia nigra compacta (Zeng et al. 2018). In addition, valproic acid injection has been used to establish a model of autism spectrum disorder, a psychiatric disorder (Schneider and Przewlocki 2005). Apolipoprotein E4 (apoE4) is very strongly associated with Alzheimer's disease. Expression of both mutant presenilin 1 (PS1) and mutant amyloid precursor protein (APP) is associated with the deposition of amyloid aggregates and Alzheimer's disease-like behaviors (Borchelt et al. 1997;Holcomb et al. 1998). Mutations in LRRK2, alpha-synuclein, parkin, DJ-1, and PINK1 are strongly related to Parkinson's disease (Dawson et al. 2010).

Conclusion
In this review, we suggest that neurotoxicity should be assessed by analyzing a variety of measures, including measurement of parameters related to synapses, ion channels, synaptic plasticity, and APs, as well as the application of electrophysiological and imaging techniques in cell, circuit, and animal models. To date, standardized criteria for measuring neurotoxicity have not been fully established. In the future, there is a need to develop a classification system and standardized methods for use in neurotoxicity research. It is also necessary for government-funded research institutes to propose an active research methodology to obtain research results that follow the regulatory direction of the government. Cooperation between governments, industry, institutes, and universities is necessary for the further development of technology for assessing neurotoxicity.
Human and animal rights This article does not contain any studies with human participants or animals performed by any of the authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.