Silencing of Activity During Hypoxia Improves Functional Outcomes in Motor Neuron Networks in vitro

Abstract

The effects of hypoxia, or reduced oxygen supply, to brain tissue can be disastrous, leading to extensive loss of function. Deoxygenated tissue becomes unable to maintain healthy metabolism, which leads to increased production of reactive oxygen species (ROS) and loss of calcium homoeostasis, with damaging downstream effects. Neurons are a highly energy demanding cell type, and as such they are highly sensitive to reductions in oxygenation and some types of neurons such as motor neurons are even more susceptible to hypoxic damage. In addition to the immediate deleterious effects hypoxia can have on neurons, there can be delayed effects which lead to increased risk of developing neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS), even if no immediate consequences are apparent. Furthermore, impairment of the function of various hypoxia-responsive factors has been shown to increase the risk of developing several neurodegenerative disorders. Longitudinal assessment of electrophysiological network activity is underutilised in assessing the effects of hypoxia on neurons and how their activity and communication change over time following a hypoxic challenge. This study utilised multielectrode arrays and motor neuron networks to study the response to hypoxia and the subsequent development of the neuronal activity over time, as well as the effect of silencing network activity during the hypoxic challenge. We found that motor neuron networks exposed to hypoxic challenge exhibited a delayed fluctuation in multiple network activity parameters compared to normoxic networks. Silencing of activity during the hypoxic challenge leads to maintained bursting activity, suggesting that functional outcomes are better maintained in these networks and that there are activity-dependent mechanisms involved in the network damage following hypoxia.

Keywords: activity-dependent mechanisms; hypoxia; longitudinal; motor neuron disease; multielectrode array (MEA) recording; network activity.

FIGURE 1

Experiment timeline. (A) The experimental timeline. Cells were seeded on day 0 after coating with polyethyleminine (PEI) and laminin. RI: Rock inhibitor Y-27623 2HCL, PS, Penicillin-Streptomycin antibiotics; ICC, Immunocytochemistry. (B) Detailed timeline of hypoxic challenge. Following day 56 activity was recorded every day, and media was half replaced every other day. (C) Network bursts were identified by bins of high firing rate, with the threshold indicated by the red horizontal line.

FIGURE 2

Immunocytochemistry. After maturation, neurons expressed motor neuron specific markers as well as relevant neurotransmitter receptors and markers of mature cytoskeleton. (A) Overlap of expression of motor neuron specific marker Islet1 and neuronal marker NeuN, alongside heavy neurofilament, indicates the presence of mature motor neurons. Scale bar 50 μm. (B) The co-expression of motor neuron markers HB9 and ChAT further confirms motor neuron identity. Scale bar 50 μm. (C) Expression of receptors for GABA and glutamate confirms the capacity for excitatory signalling within the motor networks, as well as the susceptibility of the networks to the GABA-inhibition. Scale bar 100 μm.

FIGURE 3

Activity of neural networks before and after the hypoxic challenge. The activity of the motor neuron networks is described in terms of firing rate (A), fraction of spikes in network bursts, or bursting propensity (B), and coherence index, a measure of network synchrony (C). The stippled line at 54 DIV indicates when the hypoxic challenge was carried out, and the data points presented for this day are baseline recordings pre-hypoxia. Red labels along the Culture Age axis indicates that cell media was replaced these days. Lines and shaded regions indicate median ± median absolute deviation.

FIGURE 4

Dispersion in network activity pre and post the hypoxic challenge. The dispersion of the electrophysiology parameters was assessed by the RCV_Q of the activity of each network and compared by Wilcoxon rank sum test, prior to and following the hypoxic challenge. The longitudinal dispersion pre-hypoxia showed no significant differences in terms of firing rate (A), bursting propensity (B), or synchrony (C). Post-hypoxia the hypoxic networks showed similar significant increases in longitudinal dispersion in terms of firing rate (D) and synchrony (F). However, only the uninhibited hypoxic networks showed a significant increase in the dispersion of bursting propensity, while the inhibited hypoxic networks showed significantly less longitudinal dispersion, in line with control networks (E). *p < 0.0083, **p < 0.0017, ***p < 0.00017. Bars and error bars indicate median ± median absolute deviation.

FIGURE 5

Hypoxic challenge and GABA inhibition. The hypoxia exposed, uninhibited networks showed similar activity to uninhibited networks normoxic networks, while inhibited networks were completely silenced. (A) Data during the hypoxia session was divided into 1-min bins and is shown here throughout the 1-h recording. Uninhibited networks showed stable activity throughout the episode, while inhibited networks were completely silenced. Lines and shaded regions indicate median ± median absolute deviation. (B–G) There were no significant differences between the baseline electrophysiological activity and the activity during the hypoxic exposure in terms of firing rate, burst propensity or synchrony, for normoxic controls or hypoxic networks, assessed by Wilcoxon rank sum test, all p > 0.05. Bars and error bars indicate median ± median absolute deviation.