Axonal propagation: does the spike stop here?

Nerve cells in the brain are characterized by a complex axon which may project locally and distally to several thousands of postsynaptic targets. The main function of the axon is to conduct the presynaptic action potential from the cell body to the nerve terminals, thus allowing a spread of activity over long distances in the brain (Fig. 1). Axonal propagation is driven by sodium channels and prevented by activation of potassium channels. Thus, the axon is generally considered to be a very reliable transmission cable in which stable propagation occurs once an action potential is generated. However, the high divergence of the axonal tree (i.e. the presence of branch points) and the presence of small varicosities along axons are parameters that decrease the safety factor for axonal conduction. Several theoretical studies have highlighted the role of these local perturbations in axonal geometry on propagation (Goldstein & Rall, 1974). In addition, the biophysical properties of ion channels may confer some filtering properties to the axon (Debanne et al. 1997; Kopysova & Debanne, 1998). However, direct experimental assay of axonal propagation is difficult because the small diameter of the axon prevents electrophysiological recording beyond the axonal initial segment. More recently, studies using optical imaging of calcium transients associated with spike propagation have indicated a high fidelity of axonal transmission (Mackenzie et al. 1996; Cox et al. 2000; Koester & Sakmann, 2000). These sophisticated methods may still, however, overestimate propagation, and the consequences of calcium buffering by the fluorescent probe have not yet been precisely evaluated. Figure 1 The action potential (AP), which is initiated near the initial segment of the axon, follows a branched path. Reliability of propagation could be altered at branch points or axonal swellings. Propagation can be assessed by extracellular recording from ... In this issue of The Journal of Physiology, Morten Raastad & Gordon Shepherd (2003) used an elegant method to directly assess the conduction properties of Schaffer collateral axons in the CA1 area of the adult rat hippocampus. Single axons could be isolated and recorded extracellularly with a large patch pipette that allowed confinement of a small portion of the axon following gentle suction (Fig. 1, inset). In some cases, the recorded axon was visualized by pre-labelling with a lipophilic fluorescent dye (DiI). This technique is not invasive and permits a superior temporal resolution to that possible with calcium imaging methods. In addition to the technical originality, several important findings are reported in this paper. Raastad & Shepherd (2003) show that at low frequency, axonal conduction is stable and reliable in more than 95 % of stimulations. Interestingly, the population of axons was not totally homogeneous and some units (∼10 %) exhibited higher levels of propagation failure. It will be important to identify in the future the determinants of unreliability in these particular axons. The authors were able to assess the refractory period of CA3 pyramidal cell axons (i.e. the minimum time required for regenerating a second action potential). When a pair of axonal stimuli was delivered with a small delay (5–10 ms), a large reduction in the amplitude of the second action potential was observed, suggesting that during bursting activity at a frequency above 100 Hz, the last action potentials will fail to propagate. Thus, the hippocampal axon can be considered to be a low-pass filter with a cut-off frequency near 400 Hz at 37 °C. The paper by Raastad & Shepherd (2003) strongly supports the reliability of axonal propagation but it also opens numerous new areas of investigation to assess in detail the physiology of axonal propagation in mammalian neurons. For instance, the combination of Raastad & Shepherd's technique with intracellular recording from the cell body will allow direct evaluation of the transfer function of central axons. Several important questions are still pending such as the developmental profile of axonal conduction. The role of sodium channel inactivation during repetitive axonal stimulation will also need further exploration (Brody & Yue, 2000; He et al. 2002). It is undeniable that the use of techniques such as those described by Raastad & Shepherd (2003) will soon help to answer the question whether and why spike propagation stops along thin axons.