Introduction

The structural plasticity of neurons is a key mechanism that brain circuits can respond to, and correct, wholesale changes in activity. Activity-dependent structural changes have been described across all neuronal compartments, including the axon initial segment (AIS); with timescales ranging from as short as a few hours to as long as weeks [1-2]. Due to its unique cytoskeleton, the AIS accumulates voltage-gated ion-channels and is the primary location of action potential initiation [3]. Changes to AIS length have been proposed as a mechanism of homeostatic plasticity of neuronal intrinsic physiology, in vitro and in vivo [4], and has been linked to typical and neuropathological development (e.g. Fragile X Syndrome [5]). Until now, mammalian studies of AIS plasticity have largely relied on data from inbred mouse strains [1, 2, 4], which we have previously failed to replicate in acute mouse brain slices – despite the presence of robust homeostatic physiological plasticity [5]. Whether AIS structure and neuronal physiology display activity-dependent plasticity common to many species, including humans, remains unexplored.

Methods and results:

This research examines the anatomical (i.e. AIS length) and physiological (i.e. cellular excitability) properties of neurons following modulation of neuronal activity. These experiments are performed under the same ex vivo conditions in mice, rats, and resected human brain tissue; and in vivo in mice. All statistics were performed using linear mixed-effects models to avoid potential pseudoreplication. All ethical approvals for animal and human tissue collection were present prior to experiments.

 

Under control conditions we find that, irrespective of species, AIS length correlates well with many action potential properties in Layer 2/3 neocortical neurons. However, such correlation displayed region and species differences, with the AIS length of human neurons generally correlating less with excitability. In contrast to previous studies [2, 4], where we manipulated circuit activity ex vivo over a short period of time (1-3 hours), we found no evidence for altered length of the AIS in mice (N=5), out-bred rats (N=7), or adult human brain slices (N=6) following either direct depolarisation (15 mM K⁺), and increased (20 μM bicuculline) or decreased (300 nM TTX + 50 μM AP-5) circuit activity. Surprisingly, cortical neurons from living human brain slices displayed no evidence for altered intrinsic or synaptic function even after 3 hours of depolarisation – changes that have previously been observed in mice. Finally, we modulated neuronal activity in vivo through deprivation of visual inputs by dark rearing mice for 2 weeks (N=10) or 4 weeks (N=6). In these dark-reared mice we observed no change in AIS length or neuronal excitability in layer 2/3 of the primary visual cortex, consistent with an absences of activity-dependent AIS plasticity in vivo.

Conclusion and implications:

Taken together, our data suggest that whilst AIS length correlates with action potential function across species, we observe no evidence of rapid remodelling of neuronal excitability mediated by changes to AIS length following wholesale changes in activity. These data question the role that AIS plasticity could play in homeostatic plasticity of human brain circuits.