What happens when the action potential reaches the end of the axon at the axon terminal How does one neuron communicate with another neuron and complete the circut?

Axons

The axon originates either from the soma or from the proximal segment of a dendrite at a specialized region free of Nissl granules, the axon hillock. Action potentials are initiated here, at the junction with the proximal axon (axon initial segment) (Lörincz and Nusser 2010). The axonal plasma membrane (axolemma) is undercoated at the hillock by a concentration of cytoskeletal molecules, including spectrin and actin fibrils, which are important in anchoring numerous voltage-sensitive channels to the membrane (seeBender and Trussell (2012)). The axon hillock is unmyelinated and often participates in inhibitory axo-axonal synapses. It is unique because it contains ribosomal aggregates immediately below the postsynaptic membrane (Kole and Stuart 2012).

In the CNS, small, unmyelinated axons lie free in the neuropil, whereas in the PNS they are embedded in Schwann cell cytoplasm. Myelin, which is formed around almost all axons of >2 μm diameter by oligodendrocytes in the CNS and by Schwann cells in the PNS, begins at the distal end of the axon hillock. Nodes of Ranvier are specialized regions of myelin-free axon where action potentials are generated and where an axon may branch. In both CNS and PNS, the territory of a myelinated axon between adjacent nodes is called an internode; the region close to a node, where the myelin sheath terminates, is called the paranode; and the region just beyond that is the juxtaparanode. Myelin thickness and internodal lengths are, in general, positively correlated with axon diameter. The density of sodium channels in the axolemma is highest at nodes of Ranvier, and very low along internodal membranes; sodium channels are spread more evenly within the axolemma of unmelinated axons. Fast potassium chanels are present in the paranodal regions of myelinated axons. Fine processes of glial cytoplasm (astrocytic in the CNS, Schwann cell in the PNS) surround the nodal axolemma.

The terminals of an axon are unmyelinated. Most expand into presynaptic boutons, which may form connections with axons, dendrites, neuronal somata or, in the periphery, muscle fibres, glands and lymphoid tissue. Exceptions include the free afferent sensory endings in, for example, the epidermis, which are unspecialized structurally, and the peripheral terminals of afferent sensory fibres with encapsulated endings (seeFig. 3.29). Axon terminals contain abundant small clear synaptic vesicles and large dense-core vesicles. The former contain a neurotransmitter (e.g. glutamate, γ-aminobutyric acid (GABA), acetylcholine) that is released into the synaptic cleft on the arrival of an action potential at the terminal and which then binds to cognate receptors on the postsynaptic membrane. Depending on the nature of the transmitter and its receptors, the postsynaptic neurone will become excited or inhibited. The dense-core vesicles contain neuropeptides, including brain-derived neurotrophic factor (BDNF;Dieni et al 2012). Axon terminals may themselves be contacted by other axons, forming axo-axonal presynaptic inhibitory circuits (Veres et al 2014,Burke et al 2017).

The axon compartment comprises the axon hillock, initial segment, shaft and terminal arbor

These regions differ ultrastructurally in membrane morphology and cytoskeletal organization. The axon hillock may contain fragments of Nissl substance, including abundant ribosomes, which diminish as the hillock continues into the initial segment. Here, the various axoplasmic components begin to align longitudinally. A few ribosomes and the smooth ER persist, and some axoaxonic synapses occur. The axolemma of the initial segment where the action potential originates exhibits a dense granular layer similar to that seen at the nodes of Ranvier, consistent with a specialized membrane cytoskeleton. Also present in this region are microtubules, neurofilaments and mitochondria. The arrangement of the microtubules in the initial segment is distinctive in forming fascicles interconnected by side arms. Beyond the initial segment, the axon maintains a relatively uniform caliber even after branching with little or no diminution until the very terminal arbors (Fig. 1-7). One exception is a reduction of caliber for myelinated axons at the peripheral node of Ranvier (Hsieh et al., 1994) (see Fig. 1-2 and below). Myelinated axons show granular densities on the axolemma at nodes of Ranvier (Raine, 1982) that correspond to adhesion molecules and high densities of sodium channels. In myelinated fibers, there is a concentration of sodium channels at the nodal axon, a feature underlying the rapid, saltatory conduction of such fibers (Ch. 4).

Microtubules are a prominent feature of all axons. Axonal microtubules are aligned with the long axis of the axon and have a uniform polarity with plus ends distal to the soma (Ch. 6). Microtubules are present in loose groupings rather than bundles and vary in their spacing (Fig. 1-7A). Vesicles and mitochondria are typically seen in association with these microtubule domains, consistent with their movement in fast axonal transport (Ch. 8). In axons less than a micron in diameter, which are usually unmyelinated, microtubules are the primary structural cytoskeletal elements, with sparse neurofilaments and gaps in the neurofilament cytoskeleton. As axons get larger, the number of neurofilaments increases dramatically, becoming the primary determinant of axonal caliber. For large, myelinated axons, neurofilaments occupy the bulk of an axon cross-section (Ch. 6) with microtubules found in small groups along with membrane profiles.

Although neuroscientists typically draw neurons with a single unbranched axon and one presynaptic terminal, most axons are extensively branched into terminal arbors, often producing hundreds or thousands of presynaptic terminals (Fig. 1-3). In addition, many axons in the CNS have en passant presynaptic specializations (Peters et al., 1991) that allow a single axon to have many presynaptic specializations in series. Parallel fibers in the cerebellar cortex may have thousands of these specializations. When en passant synapses occur on myelinated fibers, these synaptic specializations are seen at the nodes of Ranvier. The terminal portion of the axon arborizes and enlarges to form presynaptic specializations at sites of synaptic contact (Chs. 7 and 12).

Neuronal Types in the Cerebral Cortex

The cerebral cortex is an intricate complex of neuronal somata and fibres, neuroglia and blood vessels. The neocortex contains two major neuronal cell types: excitatory pyramidal cells with long axons (projection neurones) and stellate cells (Fig. 32.4). Pyramidal cells are the most abundant type and represent about 75% of the cortical neurones. Stellate cells have been subdivided into spiny stellate cells that express glutamate as neurotransmitter, and GABA-containing inhibitory aspinous interneurones with relatively short, locally terminating axons (Fig. 32.5). All types have been further divided on the basis of their size and shape, connectivity, neurotransmitter and gene expression, and electrophysiological characteristics.

Pyramidal cells have a flask-shaped or triangular soma that ranges from 10 to 80 μm in diameter and gives rise to a single thick apical dendrite and multiple basal dendrites. The apical dendrite ascends towards the cortical surface, tapering and branching (seeFig. 32.4). Some dendrites end in the most superficial lamina, the molecular layer (layer I), others end at the level of the outer granular (layer II) or outer pyramidal layer (layer III). From the basal surface of the soma, dendrites spread horizontally or obliquely for distances up to 1 mm, depending on the size of the soma. Like apical dendrites, the basal dendrites branch profusely along their length. All pyramidal cell dendrites are studded with dendritic spines that become more numerous as the distance from the cell soma increases. A single slender axon arises from the axon hillock, which is usually situated centrally on the basal surface of a pyramidal neurone. For the vast majority, if not in all, pyramidal neurones, the axon leaves the cortical grey matter to enter the white matter. Before an axon reaches the white matter, it gives rise to a recurrent branch that terminates in the cortical region of origin of the parent axon. Pyramidal cells use an excitatory amino acid, either glutamate or aspartate, as neurotransmitter. Various types of pyramidal cells can be identified by their connectivity (e.g. axonal projections to the ipsi- or contralateral hemisphere, cortico-striate, cortico-thalamic, cortico-pontine and cortico-spinal projections), electrophysiological (e.g. regular fast or slow spiking, or intrinsically bursting types) and genetic characteristics. The giant Betz cells and the solitary Meynert cells are very large pyramidal cells with special morphological features.

The volume of the somata of Betz cells, located in lamina V of the primary motor cortex, is ten times larger than that of common layer V pyramidal cells. A Betz cell gives rise to a 10 μm thick myelinated axon that terminates on motor neurones in the brainstem and anterior horn of the spinal cord via the cortico-nuclear and cortico-spinal tracts, respectively. In contrast to common pyramidal cells, which have a single apical dendrite and dendritic arbors that exit the soma from basal angles, Betz cells have many dendrites that arise directly from the entire circumference of the soma. Meynert cells are located at the border between layers V and VI of the primary visual cortex; their somata are three times larger than those of common layer V pyramidal cells. In contrast to other pyramidal cells, most of their spines are found on the extensively arborized basal dendrites in layers V and VI, which extend below several blobs and interblobs (see below). Meynert cells are involved in motion detection, mediated via their connections with area MT/V5, and with the superior colliculus via axon collaterals. The von Economo neurone (spindle neurone) is a subtype of projection neurone that occurs preferentially in the deeper layer V of fronto-insular and anterior cingulate regions and has also been described in the entorhinal cortex, the hippocampal formation, and the dorsomedial Brodmann’s area 9 (BA9). These neurones are characterized by a single apical dendrite, a single basal dendrite and a long projection axon and are probably glutamatergic: their selective degeneration is found in early stages of fronto-temporal dementia.

F Response Procedure

To obtain an F response, the setup is essentially the same as that for a routine motor conduction study using distal stimulation. Several adjustments must be made to the EMG machine to record F responses, however. The gain should be increased to 200 µV (because the amplitude of the F response is quite low), and the sweep speed should be increased to 5 or 10 ms, depending on the length of the nerve being studied. Supramaximal stimulation must always be used, and the stimulator should be turned around so that the cathode is more proximal (Figure 4–5). Although F responses typically can be obtained with the stimulator in the standard position (cathode distal), there is the theoretical possibility of anodal block (wherein the nerve hyperpolarizes under the anode, blocking antidromic travel of the action potential from the depolarization site under the cathode). One should stimulate at a rate no faster than once every two seconds (0.5 Hz). This is done in order to avoid the effects of the previous stimulus on a subsequent response. In addition, stimulating at this rate is much more comfortable for the patient and avoids the “temporal summation of pain” that occurs when the stimulation frequency is too fast (i.e., the patient is stimulated again before recovering from the discomfort of the previous one).

Because each F response varies in latency and amplitude, it is important to obtain at least ten F responses, preferably on a rastered trace. Indeed, the normal values of F waves are based on doing at least ten stimulations. If one is unable to obtain an F response, first ensure that the nerve has been stimulated supramaximally. Second, the Jendrassik (reinforcement) maneuver can be of help in “priming” the anterior horn cells. The patient can be asked to make a fist with the contralateral hand or clench the teeth prior to each stimulation. This maneuver often will elicit an F response where one was not present at rest. It should be noted that one should not do the Jendrassik maneuver unless the F responses are difficult to elicit. Paradoxically, performing a Jendrassik maneuver when not necessary can actually decrease the likelihood of obtaining F responses.*

Of the various F response measurements (minimal latency, chronodispersion [maximal minus minimal F response latency], and F wave persistence [Figure 4–6]), the minimal F wave latency is the most reliable and useful measurement, although occasionally side-to-side differences in F wave persistence and chronodispersion help in identifying an abnormality. Unfortunately, because F responses are quite small, there is often some inherent error in placing the latency markers. It is best to place the latency marker on the F response at the point where it departs from the baseline, with either a positive or negative deflection. In addition, superimposing the rastered traces once all the responses are obtained is often helpful in determining the minimal latency.

It is important to emphasize that although F responses usually are thought of as assessing the proximal nerve segments, they actually check the entire nerve. For example, any nerve with a prolonged distal motor latency on routine nerve conduction studies will also have prolonged F responses, because the F response must travel through the distal segment of nerve as well as the proximal segment. This situation is commonly seen in patients with median neuropathy at the wrist, wherein the median minimal F latency is often prolonged; in this situation, the depolarization travels antidromically from the stimulation site at the wrist up the nerve to the anterior horn cell, and then back down the nerve to the point of stimulation. However, once the depolarization proceeds past the point of stimulation, through the area of slowing at the wrist, this results in prolongation of the F response. Likewise, if there is generalized conduction velocity slowing from a polyneuropathy, the F response will also be slowed, reflecting the slowed conduction velocity of the entire nerve. The F response latency is shorter in the arms than in the legs, reflecting the shorter length of nerve traveled. Therefore, it should be no surprise that taller patients have longer F responses than do shorter patients. Thus, the distal motor latency, the conduction velocity, and the height of the patient must all be taken into account before a prolonged F response is interpreted as indicating a proximal nerve lesion.

Axons Convey Electrical Signals Over Long Distances

The single axon of each neuron looks different from the dendrites. Rather than being a tapered extension of the neuronal cell body, the axon is a cylindrical process that arises abruptly from anaxon hillock on one side of the neuronal cell body or one of its proximal dendrites. Bundles of microtubules, accompanied by neurofilaments and mitochondria, funnel through the axon hillock into theinitial segment of the axon (Fig. 1.17). Some RNA makes it into the axon, but Nissl substance stays behind (seeFig. 1.12B); despite some local axonal protein synthesis, the relatively vast volume of axonal cytoplasm depends on the soma for most of the macromolecules needed by it and its synaptic terminals. The initial segment is typically the most electrically excitable part of the neuron; as described inChapters 7 and8, all the synaptic inputs to the dendrites, cell body, and initial segment are summed up here to determine the electrical response that will be propagated along the axon.

Beyond the initial segment, many axons are encased in a spiral wrapping of glial membranes calledmyelin (Fig. 1.18). As discussed inChapter 7, myelin is a mammalian invention that greatly increases the speed of propagation of electrical signals along axons.

Axons and Axon Terminals

The axon arises from the cell body at a small elevation called the axon hillock. The proximal part of the axon, adjacent to the axon hillock, is the initial segment. The cytoplasm of the axon (axoplasm) contains dense bundles of microtubules and neurofilaments (Figs. 2.1 and 2.5A, B). These function as structural elements, and the microtubules also play key roles in the transport of metabolites and organelles along the axon. Axons are typically devoid of ribosomes, a feature that distinguishes them from dendrites at the ultrastructural level.

In contrast to dendrites, axons may extend for long distances before branching and terminating. An example is the axon of a corticospinal tract neuron with a cell body in the motor cortex and an axon that reaches the caudal portion of the spinal cord. The axon of such a neuron accounts for approximately 99.8% of the total volume of the neuron. The surface area of an axon can be several thousand times the surface area of the parent cell body. Axons are sometimes referred to as nerve fibers, although strictly speaking, a nerve fiber includes both the axon and a sheath that is provided by support cells (described in a subsequent section).

Axons in the CNS often end in fine branches known as terminal arbors (Fig. 2.5C). In most neurons, each axon terminal is capped with small terminal boutons (boutons terminaux, terminal buttons) (Figs. 2.1 and 2.3C, E). These correspond to functional points of contact (synapses) between nerve cells. In some cells, boutons are found along the length of the axon, where they are called boutons en passant. Other axons contain swellings, or varicosities, that are not button-like but still can represent points of cell-to-cell information transfer.

The site at which an axon terminal communicates with a second neuron, or with an effector tissue, is called a synapse (from the Greek word meaning “to clasp”). In general, the synapse can be defined as a contact between part of one neuron (usually its axon) and the dendrites, cell body, or axon of a second neuron. The contact can also be made with an effector cell such as a skeletal muscle fiber. Synapses are considered later in this chapter in the section Neurons as Information Transmitters.

Myelinated axons

The site of action potential origination in myelinated axons was also shown to be the axon hillock about 50 years ago. More recently, with improvements in the ability to measure voltage changes optically and with patch clamp electrodes, the precise location of the origin of the action potential has been identified. Action potentials can originate not only at the axon hillock, but also in the axon initial segment, 30–40 μm from the soma and close to the first myelinated segment. In some neurons the action potential even originates at the first node of Ranvier, where sodium channels are highly concentrated (Figure 1). For both myelinated and unmyelinated axons, once the action potential begins in the axon, it not only propagates orthodromically toward the nerve terminals but also propagates antidromically, back into the soma and dendrites.

Action Potentials

The action potential is an all-or-nothing electrical wave that is initiated at the axon hillock and propagates toward the axon terminal via highly coordinated sequential activation of various ion channels that have differential selectivities with respect to ion permeation (Figure 1). Action potentials are not graded signals like EPSPs or IPSPs, and are either initiated fully or not at all. Although the shape of an action potential is relatively consistent across the many types of neurons within the nervous system, variations in ion channel properties do introduce some differences (Kress and Mennerick, 2009).

The axon hillock contains a very high concentration of voltage-gated Na+ channels that become activated once a critical membrane potential is reached, the threshold potential. The threshold potential is a membrane depolarization of approximately 10 mV from rest. This change in voltage results in a conformational change of the voltage-gated Na+ channel that opens up a central pore that is highly selective for Na+, allowing its influx as directed by its electrochemical gradient (Bezanilla, 2006). This influx further depolarizes nearby areas of the cell. Voltage-gated Na+ channels are a single polypeptide chain of amino acids that contain four repeating homologous domains each containing six transmembrane segments (S1–S6) and a re-entrant pore-lining region (P region) that lies between the S5 and S6 segments (Payandeh et al., 2011). The four P regions of the protein contain chemical groups that effectively replace the hydration shell of Na+ and stabilize the charge on the dehydrated Na+ as it passes through the selectivity filter of the pore (which is specific for Na+, and impermeant to K+). The S4 region, although mostly hydrophobic, contains positively charged amino acids that sense the voltage difference across the cell membrane. It is hypothesized that this portion of the protein moves in response to membrane depolarization, thereby causing a conformation change in the channel allowing for the opening of an activation gate and flux of Na+ through the channel. Voltage-gated Na+ channels inactivate in a manner that is distinct from just the simple closing of the activation gate. The inactivation gate of voltage-gated Na+ channels also closes as a result of depolarization, however, more slowly than the opening of the activation gate. This results in a brief window (milliseconds) during which both gates are open and (Na+) can pass and allow for the neuronal membrane potential to approach ENa (typically ~+55 mV). This sequence of events results in the upswing or rising phase of the action potential. It is not until the membrane potential returns to near rest that both gates reset and voltage-gated Na+ channels resume their resting conformation of a closed activation gate and open inactivation gate. Until voltage-gated Na+ channels return to their resting conformation, another action potential cannot be initiated since the channels are inactivated and unable to allow Na+ influx. The time it takes for voltage-gated Na+ channels to return to their resting confirmations and therefore allow for the initiation of another action potential is called the absolute refractory period.

The downswing, or falling phase, of action potentials results from the subsequent opening of voltage-gated K+ channels, as well as voltage-gated Na+ channel inactivation. Voltage-gated K+ channels also contain six transmembrane regions (S1–S6) and a re-entrant P region between S5 and S6 (Jiang et al., 2003). However, each gene product of voltage-gated K+ channels contains only one copy of this sequence. Four subunits assemble to produce the functional channel that is highly selective for K+. Similar to voltage-gated Na+ channels, it has been shown that the P region lines the central pore and the S4 region is involved in channel activation of voltage-gated K+ channels. Voltage-gated K+ channels are activated by membrane depolarization as well, however their activation occurs more slowly than voltage-gated Na+ channels. When activated, K+ flows out of the cell along its electrochemical gradient, thereby bringing the membrane potential back near EK (typically ~−80 mV), which is more negative than the resting membrane potential (after hyperpolarization). The return of the membrane potential toward rest results in the closing of voltage-gated K+ channels (unlike voltage-gated Na channels, they do not contain an inactivation gate). The slow closing of voltage-gated K+ channels also makes it more difficult to initiate another action potential since the latent permeability to K+ will keep the neuronal membrane further from the threshold potential necessary to activate voltage-gated Na+ channels. This relative refractory period lasts until voltage-gated K+ channels return to their closed resting confirmations. During the relative refractory period, it is possible to generate another action potential, however, a stronger stimulus is required.

A. Amyloid Precursor Protein Immunocytochemistry

The traditional stains for visualizing axons—the single, slender processes that emerge from neurons at the axon hillocks and extend with a relatively uniform caliber to the terminals up to a meter or more away—are silver stains that bind to the neurofilaments. With such stains damaged axons appear swollen because of the interruption in their fast transport system and the proximal accumulation of organelles and fluid.

Immunocytochemical staining for β-amyloid precursor protein (APP) more sensitively detects axons that have impaired fast axonal transport (Fig. 1). In normally functioning axons, the protein is transported in this way and never builds up to a concentration that allows its detection in tissue sections. However, axons that have this transport system disrupted rapidly accumulate APP proximal to the disrupted segment. This occurs before the development of conventional morphological evidence of axonal damage (e.g., in the form of axonal end-bulbs), so the immunocytochemical method for detecting APP is more sensitive than routine histological methods for detecting axon damage. Other proteins transported by fast axonal transport also accumulate, but antibodies to APP have been shown to be the most sensitive for detecting this type of damage (Grady et al., 1993; Gultekin and Smith, 1994; Li et al., 1995; Ng et al., 1994; Pesini et al., 1999; Sherriff et al., 1994a). Experimental animal studies of brain trauma have shown that some axonal damage is reversible, but it is not known if axonal damage severe enough to be detected with APP immunoreactivity in humans is ever reversible; in general this has been thought unlikely. With regard to timing of damage detectable in this way, the immunoreaction in damaged axons for APP becomes positive in head trauma 1 to 3 hours after the insult (McKenzie et al., 1996; Oehmichen et al., 1998; Sherriff et al., 1994b) and remains positive for up to 1 month (Geddes et al., 2000). The distal parts of irreversibly damaged axons will undergo wallerian degeneration, which can be detected by such traditional methods as the Marchi technique on tissue sections and in living subjects by neuroimaging with (Banati et al., 2000) or without (Simon et al., 2000) novel ligands. An estimate of the scale of long-previous irreversible axonal damage can be obtained by performing estimates of axon numbers, a relatively straightforward task using modern computerized image analysis facilities.

10.04.3.1 Axonal Action Potential

A neuron receives synaptic inputs from numerous neighboring neurons via its dendrites. These inputs are combined at soma in the axon hillock, which controls the firing of an action potential down the axon.

Hodgkin and Huxley first studied the ionic basis of action potentials in squid giant axons (Hodgkin and Huxley, 1952a,b; Hodgkin et al., 1952). Using the voltage clamp technique, they could hold the membrane potential constant and measure the resultant current flowing through the membrane. At hyperpolarizing voltages, there is little measurable current because most channels are closed. Upon depolarization, there is first a brief but large inward current followed by a more sustained outward current. By clamping the voltage at the equilibrium potential of different ions and by manipulating extracellular ion concentrations, they were able to identify two distinct currents responsible for generating the action potential: Na+ and K+ currents.

Now it is known that when the membrane potential exceeds a certain threshold, NaV channels rapidly open, causing a sudden rush of Na+ into the cell. The influx of Na+ brings the membrane potential toward the Na+ equilibrium potential and depolarizes the membrane, generating the upstroke of the action potential. Quickly following activation, NaV channels inactivate and shut down Na+ conductance. The depolarization of the membrane activates KV channels. The efflux of K+ out of the cell hyperpolarizes the cell and returns the membrane potential back to resting levels. This depolarizing waveform is propagated down the axon and feeds onto synapses with neighboring neurons.

In addition to uncovering the currents behind the action potential, Hodgkin and Huxley developed an empirical kinetic model to describe the voltage‐dependent Na+ and K+ conductances and action potentials in terms of the ion conductances (Hodgkin and Huxley, 1952c). A simple schematic of the electrical circuit of the membrane is shown in Figure 2(b). From this schematic, the equation describing the circuit is

[3]dVmdt =−1CmIion

where Vm is membrane voltage, Cm is membrane capacitance, and Iion is the ionic current across the membrane.

The individual ionic currents are

[4] Iion=INa+IK+IL=gNaVm−ENa+gKVm−EK +gLVm−EL

where INa, IK, and IL are Na+, K+, and leak currents, respectively; gNa, gK, and gL are Na+, K+, and leak conductances, respectively; and ENa, EK, and EL are Na+, K+, and leak reversal potentials, respectively.

The Na+ and K+ conductances are described by

[5]gNa=g¯Nam3h

[6]gK=g¯Kn4

where m represents probability of one Na+ activation gate to be in position to open the channel, h represents probability of the Na+ inactivation gate to be in position to inactivate the channel, and n represents the probability of one K+ activation gate to be in position to open the channel. m, n, and h are unitless parameters. g¯Na and g¯K are the maximum conductances for Na+ and K+, respectively. From the model, Na+ channels require three activated gates while K+ channels require four activated gates for channels to open.

Substituting eqns [5] and [6] into eqn [4] yields

[7]Iion=g¯Nam3hVm−ENa+g¯Kn4Vm−EK+g¯LVm−EL

The transitions for each gate can be written as a set of differential equations with corresponding rate constants.

[8]didt =αiV1−i−βiVi

where i = m, h, n; α and β are the forward and reverse rate constants, respectively.

These rate constants can be determined by fitting the equations to experimental recordings of Na+ and K+ currents.