Although recent studies have shown the existence of scattered ribosomes in axons which are likely to have an important role in normal axonal function (Koenig et al. 2000) and in pathological processes (Court et al. 2008), the most essential macromolecules must be synthesized in the cell body and transported centrifugally. An elaborate architecture and physiology have evolved in order to carry out this process (recently reviewed in Brown 2013). The normal axon contains microtubules, neurofilaments, mitochondria, membrane-bound tubulovesicles, and various dense bodies and granules (Fig. 4.1a, b). Many of these organelles are involved in the axonal transport process.

Although the axon is a single, sometimes branched, continuous tube extending from cell body to nerve terminal, myelinated axons do have a periodic architecture as a result of their intimate interaction with segmentally organized Schwann cells. Axon internodal diameter (excluding the myelin sheath) ranges from 1 to 7 μm for myelinated axons and 0.1–3 μm for unmyelinated axons in the human sural nerve (Behse 1990). The composition and organization of axonal contents vary depending on whether nodal or intermodal regions are examined.

The internodal organization of axonal contents is relatively simple: Neurofilaments and microtubules align longitudinally and relatively homogenously and intermingle with tubulovesicular structures and the occasional mitochondrion (Fig. 4.1a, b). A normal internode is 200–1,800 μm in length, roughly proportional to axon diameter.

The node of Ranvier (Fig. 4.1c–e) is the unmyelinated axon segment between two Schwann cells and their myelin sheaths, which is typically 1 μm in length. On either side of the node are the paranodal segments, 10–75 μm in length. Approaching the node, an axon loses its rounded contour and develops grooves that appear imposed by the invagination of the overlying myelin (Fig. 5.​5b). These grooves taper towards the nodal segment, where the axon diameter decreases to 50 % or less of its value at the internode (Fig. 4.2). The degree of nodal constriction is proportionally greater for large fibers than for small fibers (Reles and Friede 1991). The constricted axon segment is 5–10 μm in length, with the central 1 μm being the unmyelinated true node of Ranvier (Rydmark and Berthold 1983) and the peripheral 3–5 μm on either side being the site where myelin loops attach to the axon (Figs. 4.1e and 4.3a).

In the nodal–paranodal region microtubules and membrane-bound organelles become concentrated centrally, the former at a density several times greater than that seen in internodal cross sections. Neurofilament density does not change significantly along the axon (Berthold 1978; Reles and Friede 1991). Tubulovesicular profiles assume a longitudinal rather than reticular pattern (Fig. 4.3a, b). Retrogradely transported organelles (dense lamellated bodies and multivesicular bodies) tend to accumulate at the distal half of the nodal and paranodal axon, while anterogradely transported organelles (especially tubulovesicular elements) accumulate at the proximal end (Berthold et al. 1993) (Fig. 4.3b). Both types of organelles occur at far greater concentrations at the node of Ranvier than in the internode. Light microscopy reveals a “veil” of toluidine blue-staining substance in approximately 50 % of nodal–paranodal segments (Figs. 4.1e and 4.2). The ultrastructural correlate of this is a diffuse electron-dense granular material composed of particles 5 nm in diameter (Berthold et al. 1993).

In contrast to the segmental organization of myelinated fibers, axons of unmyelinated fibers have same appearance throughout their length, containing neurofilaments, microtubules, approximately one mitochondrion per cross-sectional view, and relatively infrequent tubulovesicular elements and dense bodies.

The bulk of neuronal protein synthesis occurs in the perikaryon rather than the axon. Recent studies have, however, shown that ribosomes, initiation and elongation factors, transfer and messenger RNAs, as well as proteins and microRNAs involved in the regulation of mRNA stability and translation are located in the axon. Locally synthesized axonal proteins include cytoskeletal proteins, protein chaperones, metabolic enzymes, some membrane proteins, and secreted proteins. It is likely that local axonal synthesis has an important and dynamic role in local structure and maintenance, synaptic function, and reactions to injury; however, by itself, local synthesis is insufficient for axonal maintenance. Although mRNA was once thought to be excluded from the axonal compartment, the existence of protein synthesis in growing or regenerating axons in culture is now generally accepted. However, its extent and functional importance remains a subject of intense investigation. Furthermore, evidence of mRNA axonal transport and local gene translation in vivo has been shown in axons of the developing zebra fish (Danio rerio) embryo, with frequent accumulation at the growth cones. Additionally, while supporting cells might provide some of the axon’s metabolic needs (Gainer et al. 1977), this contribution is insufficient by itself, requiring that most axonal material must be transported centrifugally from the cell body over long distances. For several decades investigators have appreciated the existence of a special mechanism for this purpose (Griffin and Watson 1988; Ochs and Brimijoin 1993), recently reviewed in detail (Brown 2013).

The purpose of axonal transport is not always to move a cargo from the cell body directly to a destination at the most distal end of the axon, rather some cargos (e.g., mitochondria) are functionally active during their transit providing long stretches of axons with an ongoing energy supply. Axonal transport functions dynamically, recruiting and redistributing these cargos in response to the changing physiologic and metabolic needs of the axon.

Axonal transport mechanisms have traditionally been organized on the basis of their direction (anterograde and retrograde) and speed (fast or slow) (Brown 2013; Lasek et al. 1984; Guzik and Goldstein 2004). The prototype examples include:

Investigators have clarified much of the machinery of fast axonal transport in recent years (Sheetz et al. 1989; Brown 2013). Vesicles can be directly imaged moving along paths demarcated by microtubules at fast anterograde and retrograde rates (Kachar et al. 1987). Deprivation of energy or calcium prevents fast transport, as do agents that depolymerize microtubules. Axoplasmic motor proteins important in axonal transport have been identified: kinesin for fast anterograde and dynein for fast retrograde transport. These proteins have a hinge region and bind to both cargo organelles (tail regions) and microtubules (head region) and, using their ATPase activity, move along microtubules in the presence of ATP (Sheetz et al. 1989; Vallee et al. 1989). Experimental data thus provides strong evidence that fast axonal transport is performed using a microtubule railway upon which a cytoplasmic peptide motor carries vesicles and their contents to the appropriate destination via a calcium- and energy-dependent mechanism. In support of this, mutations in a critical region of β-tubulin (TUBB3) or a motor adaptor protein KIF1Bβ both result in neuropathy (Niwa et al. 2013; Brown 2013). Although the retrograde and anterograde fast transport mechanisms have much in common, specific inhibitors can affect them differently (Edstrom et al. 1988; Guzik and Goldstein 2004).

Slow axonal transport carries cytoskeletal components (neurofilaments predominantly anterogradely) and probably governs the rate of axonal growth and regeneration (Black and Lasek 1979). The classic “structural hypothesis” of slow transport paints a picture of the axonal cytoskeleton composed of tubules and filaments, assembled in the cell body and moving ponderously forward as a whole at a constant rate (Lasek et al. 1984). A competing “unitary hypothesis” suggested that the cytoskeleton is stationary and that its components are transported distally in an unassembled state and assembled locally (Ochs et al. 1989; Bamburg 1988). Recent work (detailed in Brown 2013) suggests that fast and slow transport both take place in association with microtubules; however, they differ in the time the cargo stays in contact with the highway, which has been called the “duty ratio” (the proportion of the time that the cargo spends moving). Neurofilament movement is fast but intermittent, with each filament spending most of its time pausing between short bursts of rapid movement (Tang et al. 2013). Rather than cross-linked neurofilaments and microtubules, the axonal environment may more accurately behave as a polymer solution. Microtubules (all orientated with their minus ends pointing proximally) and microfilaments (relatively flexible, two-stranded, 5–7 nm filamentous polymers of actin capable of multidirectional orientation) serve as the tracks for long-range axial movements in axons and short-range motion, respectively. Neurofilament polymers themselves move along microtubule tracks. However, microfilament myosin motors may influence the long-range transport behavior of mitochondria by delivering these organelles to and from their microtubule tracks. Most motors interact with their cargos via adapter proteins which bind to receptors on the cargo. A single motor may often be called on to interact with multiple different cargoes.

Neurofilaments play a central role in determining axonal diameter (Gold et al. 1985; Hoffman et al. 1984, 1987). This conclusion correlates with the observation that localized accumulation of neurofilaments constitutes the substrate of focal axonal swelling and that impairment of neurofilament translocation to the distal axons underlies at least some types of axonal atrophy (vide infra). It is now thought that sidearms produced by NFM and NFH proteins present on neurofilaments preserve space between adjacent filaments and help maintain axonal size, which is critical for conduction velocity (Brown 2013). Additionally, a rigid cross-linked network of neurofilaments would likely retard the movement of other axonally transported cargos. In mutant animals lacking neurofilaments, axons fail to attain their normal caliber and have delayed conduction velocities.

Studies in animal models and human pathology have provided important insights into the spatiotemporal sequence of distal axonopathy (Spencer and Schaumburg 1977; Prineas 1969a, b; Sahenk and Mendell 1980). Most changes of distal axonal degeneration are nonspecific, but some etiologies have specific patterns. For example, acrylamide and the hexacarbon neuropathies tend to produce neurofilamentous accumulations and axonal swelling; TOCP and zinc pyridinethione produce tubulovesicular aggregates; and thallium intoxication results in mitochondrial abnormalities. However, we cannot catalog all the various combinations of ultrastructural changes that investigators have observed in the numerous animal toxic models and human ultrastructural examinations conducted to date. Such an undertaking would not prove diagnostically useful in the practice of peripheral nerve pathology. On the contrary, the discussion below will emphasize the nonspecificity of ultrastructural changes seen in axonal degeneration. Similar changes occur in axonal degeneration in the CNS (Lampert 1967). No single change is diagnostic, or even necessarily indicative of pathology, as the wide variation of normal appearance, age-related changes, and artifacts induced by the trauma of nerve biopsy and fixation can be misleading. Many of the same alterations are seen in Wallerian degeneration following mechanical or ischemic nerve transection.