Surprising diversity in axonal properties between the different functional classes of neurone in peripheral nerves

Many different types of neurone send axons in peripheral nerves including somatic motoneurones, sympathetic post-ganglionic neurones, and numerous classes of afferent neurone. These neurones have characteristic morphology and functions at their terminals, but the axons are generally thought of as rather uniform links between the two ends of the neurone. They may be myelinated (A fibres) or unmyelinated (C fibres), and there are some broad correlations between fibre size (and hence conduction velocity) and function. For example most of the specialized mechanoreceptor afferents have large A fibres whilst most nociceptors have unmyelinated or small myelinated axons. In this issue of The Journal of PhysiologySerra et al. (1999) show, in an elegant microneurography study in man, that different classes of C fibres with the same conduction velocity differ markedly in their recovery of excitability after firing and also in spike duration and amplitude.

Serra et al. (1999) find that nociceptive C fibres in a cutaneous nerve slow markedly in conduction velocity during repetitive stimulation whilst cold-sensitive thermoreceptors slow rather little and a third group of C fibres that could not be excited from the skin do not slow at all at the frequencies used (up to 2 Hz). This observation fits broadly with earlier findings in rat cutaneous nerves (Thalhammer et al. 1994; Gee et al. 1996). In the rat it was possible to look additionally at C mechanoreceptor afferents and at post-ganglionic sympathetic neurones. These two groups both showed conduction velocity (CV) slowing intermediate between the nociceptor and the thermoreceptor afferents (Gee et al. 1996). The differences in CV slowing are quite reliable and allow, in both man and rat, identification of C fibres as to functional class without needing to test the properties of terminals.

As well as differences in post-spike recovery, there are also differences in spike shape. Serra et al. (1999) found that their inexcitable C fibres with no conduction slowing had unusually small spikes of relatively short duration. Again differences in axonal spike duration for C fibres with different function have also been found in other species (Gee et al. 1999) with nociceptors having longer duration spikes than non-nociceptors and similar differences have recently been described for the cell body spikes of these neurones (Djouhri et al. 1998). The differences in spike duration, in contrast to those in conduction slowing, are not sufficiently clear-cut to provide unambiguous functional identification.

The different electrical properties of axons will reflect the particular mix of pumps and channels in the axonal membrane. Some of the likely candidates are shown in Fig 1. Spike duration differences will reflect differences in fast voltage-gated channels for sodium, calcium and potassium. There is good evidence from molecular studies of dorsal root ganglion cells that several different sodium channel isoforms are expressed in the small ganglion cells with C fibre axons. In particular it has been found that the relatively slowly activating TTX-resistant channel (SNS, Akopian et al. 1996) is found mainly in capsaicin and bradykinin-sensitive neurones, i.e. probably nociceptors. The differences in recovery after firing may reflect several processes including the amount of electrogenic sodium pumping, the presence or absence of inward rectification (I_h) and the opening of calcium-activated potassium channels (K_(Ca)), assuming there is significant calcium entry during the spike. In general, differences in membrane mechanisms between cells are thought to have evolved to minimize the metabolic cost of maintaining ion gradients. In cells that are continuously active, such as thermoreceptor afferents or post-ganglionic sympathetic neurones, the best mix of channels will differ from the situation in nociceptor afferents, where activation is intermittent (hopefully very intermittent!). It is also clear that the expression of channels can change when conditions change. Thus it is known that with peripheral inflammation there is an upregulation in the expression of the SNS sodium channel (Tanaka et al. 1998), possibly related to the need to maintain the continuous firing that occurs in nociceptor afferents under these conditions.

Figure 1.

Ionic mechanisms in C fibres that may vary with modality. Note that I_h is a non-selective cation channel, but after the action potential will largely carry a sodium current.

The presence of particular channels preferentially in nociceptor afferents opens up the possibility of developing highly selective analgesics. The fast voltage-gated channels are one obvious target and reports of agents that block some of the relevant sodium channels and show selective analgesic action are already appearing (Trezise et al. 1998). The channels that control recovery of excitability may also be a possible target. Holding K_(Ca) channels open or holding I_h channels closed might not lead to complete block, but might limit firing frequency to a therapeutically useful extent.

It certainly appears that axons are a good deal more interesting than we thought. Examining axonal properties has already given us some useful experimental tools. Finding out the molecular basis for these differences will give new insights into axonal function and may lead to the development of a whole new generation of selective local analgesic drugs.