Spike Generation and Excitable Domains in Myelinated and Demyelinated Axons

The electrically active nerve cells, also called ‘neurons’, consist of functionally and morphologically distinct branched structures. Information is received and integrated in dendrites and rapidly transmitted to thousands of other neurons via the small-diameter output branches called ‘axons’. The axonal membrane is highly enriched in anchoring proteins, which cluster voltage-gated ion channels to specialized membrane regions involved in the generation of action potentials (APs), in the axon initial segment (AIS), and the conduction of these APs via the nodes of Ranvier. The axolemma between nodes is ensheathed by myelin membranes. Myelination accelerates impulse conduction velocity by a process called saltatory conduction. Here, we asked the question how sensitive AP generation is to changes in the cellular architecture and axonal domains? Recent studies have shown that the AIS and nodal domains are affected by disease and are activity-dependent plasticity. But is APs generation dependent on the AIS location and myelination? In order to disentangle the role of excitable domains in relation to dendritic morphology and myelination we used correlated high-resolution recording and imaging of large pyramidal neurons in the rodent and human neocortex.
Experimental and theoretical analysis of the structure-function relationship of AIS location in the neocortical layer 5 pyramidal neurons, revealed a scaling principle between the size of the dendritic tree and AIS location. Our data revealed that the AIS location is not random by co-varies with proximal dendritic morphology for optimal AP waveform. Unique access to human acute slices obtained from temporal love epilepsy surgeries, allowed us to investigate the spike generation in human layer 5 pyramidal neurons. Our data revealed also in human pyramidal neurons AIS location is correlated with the dendritic tree, and the increased compartmentalization of the human AIS produces slower spike rise times which were offset by increased dendritic electrogenesis. Subsequently, we examined in mouse layer 5 axons the consequences of myelin loss on the AIS and cellular excitability. Using a mouse model for demyelination we found changes in the location and stability of ion channel clustering both in the AIS as well as the nodes of Ranvier. Single pyramidal neurons and the neocortical circuit became hyperexcitable and APs were in some cases initiated from ectopic locations downstream from the AIS. These findings are very relevant to our understanding of neurological symptoms in multiple sclerosis patients. Building further on this demyelination model, we extended the analysis beyond the main axon and measured how APs invade and distribute into the unmyelinated axon collaterals, using electrical and optical voltage recordings from demyelinated main axons and en passant pre-synaptic boutons. Loss of myelin abolished saltatory conduction independent of the voltage-gated ion channels. Also, AP propagation in higher-order branches of non-myelinated axonal collaterals was significantly and frequency-dependent hampered by myelin loss.
This thesis sheds light on the fundamental role of the cellular architecture of axons showing that AIS location is species-specific and with micrometer precision tuned to the dendritic morphology. Furthermore, axonal demyelination has not only a widespread impact on spike reliability but also makes the neocortical circuit hyperexcitable.