Cerebral hypoxia-ischemia models have been widely used to evaluate mechanisms of neuroprotection [4]. The pathophysiology underlying hypoxic brain damage includes complex mechanisms including loss of ATP, excitotoxicity, production of free radicals, inflammation, overactivation of the immune system, and cell death [27]. Although numerous mammalian studies have extensively investigated ischemic stroke, most have failed to develop therapeutic treatments for ischemia. Therefore, some progress may be made by studying hypoxic-tolerant organisms such as a fish [27].
The zebrafish is a relatively small and simple vertebrate organism, who’s genetic composition is similar to that of mammals, including humans. Thus, similar genes are likely to be associated with similar functions in humans [17]. Braga et al. reported spontaneous behavioral recovery in zebrafish after hypoxia [24]. In addition, several zebrafish models have been developed to demonstrate the usefulness of assessing cognitive function, learning, and memory [26, 28]. The zebrafish T-maze is based on visual discrimination learning [29]. Sison and Gerlai evaluated associative memory using visual perception in zebrafish [30]. Notably, zebrafish have specific color preferences [31]. We previously employed a similar zebrafish behavior model using color preferences in the absence of food rewards [21, 25]. Similar to our previous findings, a four-day training period was sufficient to allow zebrafish to develop a preference for a particular target compartment, even in the absence of colored cellophane and food during the testing period. Approximately ten minutes of hypoxia reversed the effects of training and induced cerebral injury, consistent with our previous results [21, 25].
Magnesium is important for proper functioning in many tissues and organs including those of the cardiovascular, neuromuscular, and nervous systems. It plays an important role in synaptic plasticity [32] by reducing the calcium dependent post-burst after hyperpolarization of membrane potential, and regulating voltage-dependent blockade of N-methyl-D-aspartate (NMDA) glutamate receptors [23, 33]. Previous studies have shown that an increase in magnesium ion concentration in the extracellular fluid causes long-term enhancement of synaptic plasticity in hippocampal neurons [34]. Thus, increased magnesium in the brain may improve cognitive function.
Nonetheless, McKee et al. [11] reported limited neuroprotective effects of MgSO4 in patients with acute brain injury. Intravenous MgSO4 administration appeared to be hindered by the blood-brain barrier, leading to low levels of magnesium in the CSF, reflective of brain bioavailability [11, 35]. Compared with MgSO4, MgT has a different molecular structure that consists of a magnesium ion and threonate. Interestingly, Slutsky et al. [13] showed that sodium-L-threonate with/without magnesium chloride did not affect memory, while MgT enhanced memory recall. Sun et al. [12] reported several threonate effects, including increased mitochondrial function, glutamatergic synapse density, and neuronal intracellular magnesium ions in hippocampal neuronal cultures. They suggested that threonate may induce magnesium ion transport into hippocampal neurons.
Although reliable increases in magnesium levels are mostly safe [13], magnesium overdose may result in adverse events including lowered blood pressure, slowed heart rate, and cardiac arrhythmia or arrest. Because there was no prior study regarding MgT concentration, we first evaluated MgT toxicity in zebrafish embryos. After confirming that an MgT concentration below 25 mM showed no adverse effect on development or survival, we then performed a cellular experiment. In aerobic conditions, MgT and MgSO4 did not alter viability of human neuronal cell cultures. However, in hypoxic conditions, MgT treatment showed significantly improved cell viability, while MgSO4 treatment was not significant. In addition, the concentration (1 mM or 10 mM) of either MgT or MgSO4, did not affect outcome. This finding suggests that the regulation of the magnesium occurs at a cellular level at a relatively wide concentration range. We then performed an in vivo experiment with MgT. As expected, zebrafish pretreated with MgT maintained a preference in time, distance, and frequency of entries to the target compartment after hypoxic insult. The absorbance of zebrafish brain after TTC staining in the MgT + HYP groups was significantly higher than those in the PBS + HYP group. This indicates that MgT preconditioning reduced brain infarction and protected against hypoxic insult, in agreement with the behavioral results.
During hypoxia, many changes, including glutamate alteration, NMDA receptor stimulation, and neuronal degeneration occurred [36]. There are several explanations for neuroprotective effects of magnesium. Stevenson et al. [10] suggested that magnesium protects the central nervous system from hypoxic injuries through the prolyl hydroxylase or factor inhibiting hypoxia-inducible factor pathways. They focused on a specific genetic pathway (ephrinB2a with a hypoxia-inducible transcription factor 1 pathway) in neurodevelopment of zebrafish embryos. Others focused on tumor necrosis factor alpha to explain the neuroprotective effect of magnesium [14, 15]. Other plausible explanations include stabilization of the cell membrane, maintenance of ionic homeostasis by attenuating reductions in Na+K-ATPase activity, neuronal effects via reduction in NMDA receptor-mediated calcium entry into the cell, and vascular effects by improving cerebral blood flow [11, 36].
In this study, we investigated the role of EAAT4 as a neuroprotective mechanism of magnesium. Once released into the synapse, glutamate is rapidly cleared by transporters (high-affinity EAATs) to limit excitotoxicity [37]. Keeping a low concentration of extracellular glutamate is also required for high signal-to-noise ratios during synaptic transmission [6]. Among the glutamate transporters, EAAT4 is expressed in several sites, including the cerebellum, hippocampus, and spinal cord [38–40]. Compared to EAAT subtypes 1–3, EAAT4 is associated with higher chloride conductance, which is not coupled to glutamate uptake, and therefore acts as an inhibitory glutamate receptor [38, 41]. By switching glutamate transport and chloride channel activity, EAAT4 may dampen cellular excitability during glutamate uptake and prevent a reduction in transport rate [41]. A previous study found that Purkinje cells die more easily in the event of EAAT4 deficiency after global brain ischemia [42]. Yi et al. [39] reported upregulated EAAT4 in rat hippocampal astrocytes 3–7 days after traumatic brain injury. In addition, Sachs et al. [43] suggested a role of EAAT4 in neuroprotection using a mutant neurodegenerative rat model. These findings suggest that EAAT4 may play a role, although the neuroprotective mechanisms are complex [44]. In this study, MgT groups showed upregulation of EAAT4 protein, which is approximately 61 kDa in size. Based on our results, MgT upregulates EAAT4 by 65–110%, whereas hypoxia downregulates EAAT4 by 18–36%. Taken together with previous experimental results, hypoxia may initially deplete EAAT4 and then induce EAAT4 reactivity in a later recovery phase [39]. Upregulation of EAAT4 by MgT appeared to protect the brain from hypoxic injury. A schematic diagram of the proposed mechanism of hypoxia and magnesium interaction is described in Fig. 8.
Western blotting demonstrated two unidentified protein bands expressed consistently in all groups. A consignment test also showed the two additional unknown bands mentioned above (data not shown). We referred to the antibody manufacturer’s guidelines to verify whether these bands may represent non-specific binding, or be due to incomplete antibody validation in the zebrafish. The antibody used was rabbit-derived, and validated in mice. One of the limitations of our study was that there was no suitable commercial antibody for zebrafish, making further evaluation of the mechanism limited. Moreover, it would be interesting to confirm the expression of EAAT4 several days following hypoxia. Another limitation of this study was that we performed tricaine anesthesia for oral administration of MgT or PBS prior to hypoxia and behavior testing. Although time under anesthesia was minimal, it may have influenced the outcomes via anesthetic preconditioning or toxic effects.
Despite these limitations, this study has several advantages. To the best of our knowledge, this study is the first application of drug-induced memory preservation in the hypoxic zebrafish behavior model. Because zebrafish are inexpensive and easy to manage, using our protocol may be useful to apply similar experiments designed to confirm the protective or toxic effects of other drugs. This finding suggests that hypoxic-tolerant organisms appear to have adaptive mechanisms to overcome hypoxic damage. Further experiments and observations are required to evaluate the specific mechanisms in order to contribute to the clinical application of potential treatments.