Chapter 3: Signaling in the Nervous System

Along with muscle cells, neurons are unique in that they are excitable; that is, they respond to stimuli by generating electrical impulses. Electrical responses of neurons (modifications of the electrical potential across their membranes) may be local (restricted to the place that received the stimulus) or propagated (may travel through the neuron and its axon). Propagated electrical impulses are termed action potentials. Neurons communicate with each other at synapses by a process called synaptic transmission.

The membranes of cells, including nerve cells, are structured so that a difference in electrical potential exists between the inside (negative) and the outside (positive). This results in a resting potential across the cell membrane, which is normally about -70 mV.

The electrical potential across the neuronal cell membrane is the result of its selective permeability to charged ions. Cell membranes are highly permeable to most inorganic ions, but they are almost impermeable to proteins and many other organic ions. The difference (gradient) in ion composition inside and outside the cell membrane is maintained by ion pumps in the membrane, which maintain a nearly constant concentration of inorganic ions within the cell (Fig 3–1 and Table 3–1). The pump that maintains Na⁺ and K⁺ gradients across the membrane is Na, K-ATPase; this specialized protein molecule extrudes Na⁺ from the intracellular compartment, moving it to the extracellular space, and imports K⁺ from the extracellular space, carrying it across the membrane into the cell. In carrying out this essential activity, the pump consumes adenosine triphosphate (ATP).

Two types of passive forces maintain an equilibrium of Na⁺ and K⁺ across the membrane: A chemical force tends to move Na⁺ inward and K⁺ outward, from the compartment containing high concentration to the compartment containing low concentration, and an electrical force (the membrane potential) tends to move Na⁺ and K⁺ inward. When the chemical and electrical forces are equally strong, an equilibrium potential exists.

For an idealized membrane that is permeable to only K⁺, the Nernst equation, which describes the relationship between these forces, is used to calculate the equilibrium potential (ie, the membrane potential at which equilibrium exists). Normally, there is a much higher concentration of K⁺ inside the cell ([K⁺]_i) than outside the cell ([K⁺]_o) (see Table 3–1). The Nernst equation, which is used to determine membrane potential across a membrane permeable only to K⁺ ions, is as follows:

where

• E = equilibrium potential (no net flow across the membrane)

• K = potassium

• T = temperature

• R = gas constant

• F = Faraday constant (relates charge in coulombs to concentration in moles)

• N = valence (for potassium, valence = 1)

• [K⁺]_i = concentration of potassium inside cell

• [K⁺]_o = concentration of potassium outside cell

At physiologic temperatures

The equilibrium potential (E_Na) for sodium can be found by substituting [Na⁺]_i and [Na⁺]_o in the Nernst equation; this potential would be found across a membrane that was permeable only to sodium. In reality, most cell membranes are permeable to several ionic species. For these membranes, potential is the weighted average of the equilibrium potentials for each permeable ion, with the contribution for each ion weighted to reflect its contribution to total membrane permeability. This is described mathematically, for a membrane that is permeable to Na⁺ and K⁺, by the Goldman–Hodgkin–Katz equation (also known as the constant field equation):

where

• [Na]_i = concentration of sodium inside cell

• [Na]_o = concentration of sodium outside cell

• P_Na = membrane permeability to sodium

• P_K = membrane permeability to potassium

As seen in this equation, membrane potential is affected by the relative permeability to each ion. If permeability to a certain ion increases (eg, by the opening of pores or channels specifically permeable to that ion), membrane potential moves closer to the equilibrium potential for that ion. Conversely, if permeability to that ion decreases (eg, by closing of pores or channels permeable to that ion), membrane potential moves away from the equilibrium potential for that ion.

In the membrane of resting neurons, K⁺ permeability is much higher (~20-fold) than Na⁺ permeability; that is, the P_K–P_Na ratio is approximately 20:1. Thus, when a neuron is inactive (resting), the Goldman–Hodgkin–Katz equation is dominated by K⁺ permeability so that membrane potential is close to the equilibrium potential for K (E_K). This accounts for the resting potential of approximately -70 mV.

The generator (receptor) potential is a local, nonpropagated response that occurs in some sensory receptors (eg, muscle stretch receptors and pacinian corpuscles, which are touch-pressure receptors) where mechanical energy is converted into electric signals. The generator potential is produced in a small area of the sensory cell: the nonmyelinated nerve terminal. Most generator potentials are depolarizations, in which membrane potential becomes less negative. In contrast to action potentials (see the next section), which are all-or-none responses, generator potentials are graded (the larger the stimulus [stretch or pressure], the larger the depolarization) and additive (two small stimuli, close together in time, produce a generator potential larger than that made by a single small stimulus). Further increase in stimulation results in larger generator potentials (Fig 3–2). When the magnitude of the generator potential increases to about 10 mV, a propagated action potential (impulse) is generated in the sensory nerve.

Neurons communicate by producing all-or-none electrical impulses called action potentials. Action potentials are self-regenerative electrical signals that tend to propagate throughout a neuron and along its axon. The action potential is a depolarization of about 100 mV (a large signal for a neuron). The action potential is all or none. Its size is constant for each neuron.

Neurons can generate action potentials because they contain specialized molecules, called sodium channels, that respond to depolarization by opening (activating). When this occurs, the relative permeability of the membrane to Na⁺ increases, and the membrane moves closer to the equilibrium potential for Na⁺, as predicted by the Goldman–Hodgkin Katz equation, thus causing further depolarization. When a depolarization (from a generator potential, synaptic potential, or oncoming action potential) impinges on a neuronal membrane, sodium channels activate and, as a result, the membrane begins to further depolarize. This action tends to activate still other sodium channels, which also open and cause depolarization. If a sufficient number of sodium channels are activated, there is a depolarization of about 15 mV, and threshold is reached so that the rate of depolarization increases sharply to produce an action potential (Fig 3–3). Thus, the membrane generates an explosive, all-or-none action potential. As the impulse passes, repolarization occurs rapidly at first and then more slowly. Membrane potential thus returns to resting potential. The action potential tends to last for a few milliseconds.

Voltage-sensitive ion channels are specialized protein molecules that span the cell membrane. These doughnut-shaped molecules contain a pore that acts as a tunnel, permitting specific ions (eg, Na⁺ or K⁺), but not other ions, to permeate. The channel also possesses a voltage sensor, which, in response to changes in potential across the membrane, either opens (activates) or closes (inactivates) the channel.

The neuronal membrane has the ability to generate impulses because it contains voltage-sensitive Na⁺ channels, which are selectively permeable to Na⁺ and tend to open when the membrane is depolarized. Because these channels open in response to depolarization, and because by opening they drive the membrane closer to Na⁺ equilibrium potential (E_Na), they tend to further depolarize the membrane (Fig 3–4). If a sufficient number of these channels are opened, there is an explosive, all-or-none response, termed the action potential (see Fig 3–3). The degree of depolarization necessary to elicit the action potential is called the threshold.

Other voltage-sensitive ion channels (voltage-sensitive K⁺ channels) open (usually more slowly than Na⁺ channels) in response to depolarization and are selectively permeable to K⁺. When these channels open, the membrane potential is driven toward the K⁺ equilibrium potential (E_K), leading to hyperpolarization.

Myelin is present around some axons within the peripheral nervous system (PNS) (where it is produced by Schwann cells) and within the central nervous system (CNS) (where it is produced by oligodendrocytes). Myelination has profound effects on the conduction of action potentials along the axon.

Nonmyelinated axons, in the mammalian PNS and CNS, generally have a small diameter (less than 1 μm in the PNS and less than 0.2 μm in the CNS). The action potential travels in a continuous manner along these axons because of a relatively uniform distribution of voltage-sensitive Na⁺ and K⁺ channels. As the action potential invades a given region of the axon, it depolarizes the region in front of it, so that the impulse crawls slowly and continuously along the entire length of the axon (Fig 3–5). In nonmyelinated axons, activation of Na⁺ channels accounts for the depolarization phase of the action potential, and activation of K⁺ channels produces repolarization.

Myelinated axons, in contrast, are covered by myelin sheaths. The myelin has a high electrical resistance and low capacitance, permitting it to act as an insulator. The myelin sheath is not continuous along the entire length of the axon. On the contrary, it is periodically interrupted by small gaps (approximately 1 μm long), called the nodes of Ranvier, where the axon is exposed. In mammalian myelinated fibers, the voltage-sensitive Na⁺ and K⁺ channels are not distributed uniformly. Na⁺ channels are clustered in high density (about 1000/μm²) in the axon membrane at the node of Ranvier, but are sparse in the internodal axon membrane, under the myelin. K⁺ channels, on the other hand, tend to be localized in the "internodal" and "paranodal" axon membrane, that is, the axon membrane covered by the myelin (Fig 3–6).

Because the current flow through the insulating myelin is very small and physiologically negligible, the action potential in myelinated axons jumps from one node to the next in a mode of conduction that has been termed saltatory (Fig 3–7).

There are several important consequences to this saltatory mode of conduction in myelinated fibers. First, the energy requirement for impulse conduction is lower in myelinated fibers; therefore, the metabolic cost of conduction is lower. Second, myelination results in an increased conduction velocity. Figure 3–8 shows conduction velocity as a function of diameter for nonmyelinated and myelinated axons. For nonmyelinated axons, conduction velocity is proportional to (diameter)^1/2. In contrast, conduction velocity in myelinated axons increases linearly with diameter. A myelinated axon can conduct impulses at a much higher conduction velocity than a nonmyelinated axon of the same size. To conduct as rapidly as a 10-μm myelinated fiber, a nonmyelinated axon would need a diameter of more than 100 μm. By increasing the conduction velocity, myelination reduces the time it takes for impulses to travel from one region to another, thus reducing the time needed for reflex activities and permitting the brain to operate as a high-speed computer.

Types of Fibers

Nerve fibers within peripheral nerves have been divided into three types according to their diameters, conduction velocities, and physiologic characteristics (Table 3–2). A fibers are large and myelinated, conduct rapidly, and carry various motor or sensory impulses. They are most susceptible to injury by mechanical pressure or lack of oxygen. B fibers are smaller myelinated axons that conduct less rapidly than A fibers. These fibers serve autonomic functions. C fibers are the smallest and are nonmyelinated; they conduct impulses the slowest and serve pain conduction and autonomic functions. An alternative classification, used to describe sensory axons in peripheral nerves, is shown in Table 3–3.

Synapses are the junctions between neurons that permit them to communicate with each other. Some synapses are excitatory (increasing the probability that the postsynaptic neuron will fire), whereas others are inhibitory (decreasing the probability that the postsynaptic neuron will fire).

In the most general sense, there are two broad anatomic classes of synapses (Table 3–4). Electrical (or electrotonic) synapses are characterized by gap junctions, which are specialized structures in which the presynaptic and postsynaptic membranes come into close apposition. Gap junctions act as conductive pathways, so electrical current can flow directly from the presynaptic axon into the postsynaptic neuron. Transmission at electrical synapses does not involve neurotransmitters. Synaptic delay is shorter at electrical synapses than at chemical synapses. Whereas electrical synapses occur commonly in the CNS of inframammalian species, they occur only rarely in the mammalian CNS.

The second broad class of synapse, which accounts for the overwhelming majority of synapses in the mammalian brain and spinal cord, is the chemical synapse. At a chemical synapse a distinct cleft (about 30 nm wide) represents an extension of the extracellular space, separating the pre- and postsynaptic membranes. The pre- and postsynaptic components at chemical synapses communicate via diffusion of neurotransmitter molecules; some common transmitters that consist of relatively small molecules are listed with their main areas of concentration in the nervous system in Table 3–5. As a result of depolarization of the presynaptic ending by action potentials, neurotransmitter molecules are released from the presynaptic ending, diffuse across the synaptic cleft, and bind to postsynaptic receptors. These receptors trigger the opening of (or, in some cases, closing of) ligand-gated ion channels. The opening (or closing) of these channels produces postsynaptic potentials. These depolarizations and hyperpolarizations are integrated by the neuron and determine whether it will fire or not (see Excitatory and Inhibitory Synaptic Actions section).

Neurotransmitter in presynaptic terminals is contained in membrane-bound presynaptic vesicles. Release of neurotransmitter occurs when the presynaptic vesicles fuse with the presynaptic membrane, permitting release of their contents by exocytosis. Vesicular transmitter release is triggered by an influx of Ca²⁺ into the presynaptic terminal, an event mediated by the activation of presynaptic Ca²⁺ channels by the invading action potential. As a result of this activity-induced increase in Ca²⁺ in the presynaptic terminal, SNARE proteins facilitate fusion of synaptic vesicles with the presynaptic membrane. The release process and diffusion across the synaptic cleft account for the synaptic delay of 0.5 to 1.0 ms at chemical synapses.

Directly Linked (Fast)

Transmitter molecules carry information from the presynaptic neuron to the postsynaptic neuron by binding at the postsynaptic membrane with either of two types of postsynaptic receptor. The first type is found exclusively in the nervous system and is directly linked to an ion channel (a ligand-gated ion channel). By binding to the postsynaptic receptor, the transmitter molecule acts directly on the postsynaptic ion channel. Moreover, the transmitter molecule is rapidly removed. This mode of synaptic transmission takes only a few milliseconds and is rapidly terminated; therefore, it is termed "fast." Depending on the type of ion channel that is open or closed, fast synaptic transmission can be either excitatory or inhibitory (see Table 3–4).

Second-Messenger Mediated (Slow)

A second mode of chemical synaptic transmission, which is closely related to endocrine communication in nonneural cells, uses receptors that are not directly linked to ion channels; these receptors open or close ion channels or change the levels of intracellular second messengers via activation of G-proteins and production of second messengers. When the transmitter is bound to the receptor, the receptor interacts with the G-protein molecule, which binds guanosine triphosphate (GTP) and is activated. Activation of the G-protein leads to production of cyclic adenosine monophosphate (cAMP), diacylglycerol (DAG), or inositol triphosphate (IP₃). Cyclic AMP, DAG, and IP₃ participate in the phosphorylation of ion channels, thus opening channels that are closed at the resting potential or closing channels that are open at the resting potential. The "slow" cascade of molecular events, leading from binding of transmitter at these receptors to opening or closing of channels, takes hundreds of milliseconds to seconds, and the effects on channels are relatively long-lasting (seconds to minutes). G-protein coupled receptors have been identified for a broad range of neurotransmitters, including dopamine, acetylcholine (muscarinic ACh receptor), and neuropeptides (Tables 3–6 and 3–7).

In contrast to fast synaptic transmission, which is highly targeted and acts on only a single postsynaptic element, second-messenger-linked transmission is slower and may affect a wider range of postsynaptic neurons. Thus, this mode of synaptic transmission serves an important modulatory function.

Excitatory postsynaptic potentials (EPSPs) are produced by the binding of neurotransmitter molecules to receptors that result in the opening of channels (eg, Na⁺ or Ca²⁺ channels) or the closing of channels (eg, K⁺ channels), thus producing depolarization. In general, excitatory synapses tend to be axodendritic. In contrast, inhibitory postsynaptic potentials (IPSPs) in many cases are caused by a localized increase in membrane permeability to Cl^– or to K⁺. This tends to cause hyperpolarization and most commonly occurs at axosomatic synapses, where it is called postsynaptic inhibition (Fig 3–10).

Information processing by neurons involves the integration of synaptic inputs from many other neurons. If they occur close enough in time, EPSPs (depolarizations) and IPSPs (hyperpolarizations) tend to sum with each other. As a neuron integrates the incoming synaptic information, it weighs the excitatory and inhibitory signals. Depending on whether or not threshold is reached at the impulse initiation zone (usually the axon initial segment), an action potential is either generated or not. If an action potential is initiated, it propagates along the axon to impinge, via its synapses, on still other neurons. The rate and pattern of action potentials carry information.

The nervous system can learn and store information in the form of memories. It has long been suspected that memory has its basis in the strengthening of particular synaptic connections. Long-term potentiation, characterized by the enhanced transmission at synapses that follow high-frequency stimulation, was first observed at synapses in the hippocampus (a part of the brain that plays an important role in memory) and may play a role in associative learning. Long-term potentiation depends on the presence of N-methyl-D-aspartate (NMDA) receptors in the postsynaptic membrane. These specialized glutamate receptors open postsynaptic Ca²⁺ channels in response to binding of the transmitter glutamate but only if the postsynaptic membrane is depolarized. Depolarization of the postsynaptic element requires the activation of other synapses, and the NMDA receptor-linked Ca²⁺ channels open only when both sets of synapses are activated. Thus, these synapses sense the "pairing" of two synaptic inputs in a manner analogous to conditioning to behavioral stimuli. Recent work suggests that, as a result of increased Ca²⁺ admitted into postsynaptic cells by this mechanism, protein kinases are activated and, via actions that are not yet fully understood, alter the synapse so as to strengthen it. These structural changes, triggered by specific patterns of synaptic activity, may provide a basis for memory.

The production of second messengers by synaptic activity may also play a role in regulation of gene expression in the postsynaptic cell. Thus, second messengers can activate enzymes that modify preexisting proteins or induce the expression of new proteins. This activation provides a mechanism whereby the synaptic activation of the cell can induce long-term changes in that cell. This is an example of plasticity within the nervous system.

Presynaptic inhibition provides a mechanism for controlling the efficacy of transmission at individual synapses. It is mediated by axoaxonal synapses (see Fig 3–10). Binding of neurotransmitters to the receptors mediating presynaptic inhibition leads to a reduction in the amount of neurotransmitter secreted by the postsynaptic axon. This reduction is caused either by a decrease in the size of the action potential in the presynaptic terminal as a result of activation of K⁺ or Cl^– channels or by reduced opening of Ca²⁺ channels in the presynaptic terminal, thereby decreasing the amount of transmitter release. Presynaptic inhibition thus provides a mechanism whereby the "gain" at a particular synaptic input to a neuron can be reduced without reducing the efficacy of other synapses that impinge on that neuron.

The axons of lower motor neurons project through peripheral nerves to muscle cells. These motor axons terminate at a specialized portion of the muscle membrane called the motor end-plate, which represents localized specialization of the sarcolemma, the membrane surrounding a striated muscle fiber (Fig 3–11). The nerve impulse is transmitted to the muscle across the neuromuscular synapse (also called the neuromuscular junction). The end-plate potential is the prolonged depolarizing potential that occurs at the end-plate in response to action potential activity in the motor axon. The transmitter at the neuromuscular synapse is ACh. Small amounts of ACh are released randomly from the nerve cell membrane at rest; each release produces a minute depolarization, a miniature end-plate potential, about 0.5 mV in amplitude. These miniature end-plate potentials, called quanta, reflect the random discharge of ACh from single synaptic vesicles. When a nerve impulse reaches the myoneural junction, substantially more transmitter is released as a result of the synchronous discharge of ACh from many synaptic vesicles. This causes a full end-plate potential that exceeds the firing level of the muscle fiber.

A large number of molecules act as neurotransmitters at chemical synapses. These neurotransmitters are present in the synaptic terminal, and their action may be blocked by pharmacologic agents. Some presynaptic nerves can release more than one transmitter; differences in the frequency of nerve stimulation probably control which transmitter is released. Some common transmitters are listed in Table 3–5.

Some neurons in the CNS also accumulate peptides. Some of these peptides act much like conventional transmitters; others appear to be hormones. Some relatively well-understood neurotransmitters are discussed next.

Acetylcholine

ACh is synthesized by choline acetyltransferase and is broken down after release into the synaptic cleft by acetylcholinesterase (AChase). These enzymes are synthesized in the neuronal cell body and are carried by axonal transport to the presynaptic terminal; synthesis of ACh occurs in the presynaptic terminal.

ACh acts as a transmitter at a variety of sites in the PNS and CNS. ACh is responsible for excitatory transmission at the neuromuscular junction (N-type, nicotinic ACh receptors). It is the transmitter in autonomic ganglia and is released by preganglionic sympathetic and parasympathetic neurons. Postganglionic parasympathetic neurons, as well as one particular type of postganglionic sympathetic axon (ie, the fibers innervating sweat glands), use ACh as their transmitter (M-type, muscarinic receptors).

Within the CNS, several well-defined groups of neurons use ACh as a transmitter. These include neurons that project widely from the basal forebrain nucleus of Meynert to the cerebral cortex and from the septal nucleus to the hippocampus. Cholinergic neurons, located in the brain stem tegmentum, project to the hypothalamus and thalamus, where they use ACh as a transmitter.

Glutamate

The amino acid glutamate has been identified as a major excitatory transmitter in the mammalian brain and spinal cord. Four types of postsynaptic glutamate receptors have been identified. Three of these are ionotropic and are linked to ion channels. These receptors are named for drugs that bind specifically to them. The kainate and AMPA types of glutamate receptor are linked to Na⁺ channels, and when glutamate binds to these receptors they produce EPSPs. The NMDA receptor is linked to a channel that is permeable to both Ca²⁺ and Na⁺. The NMDA-activated channel, however, is blocked (so that influx of these ions cannot occur) unless the postsynaptic membrane is depolarized. Thus, NMDA-type synapses mediate Ca²⁺ influx, but only when activity at these synapses is paired with excitation via other synaptic inputs that depolarize the postsynaptic neuron. The Ca²⁺ influx mediated by these synapses may lead to structural changes that strengthen the synapse. It has been hypothesized that this alteration may provide a basis for memory.

A metabotropic type of glutamate receptor has also been identified. When the transmitter glutamate binds to this receptor, the second messengers, IP₃ and DAG, are liberated. This liberation can lead to increased levels of intracellular Ca²⁺, which may activate a spectrum of enzymes that alter neuronal function and structure.

It has been suggested that excessive activation of glutamatergic synapses can lead to very large influxes of Ca²⁺ into neurons, which can cause neuronal cell death. Because glutamate is an excitatory transmitter, excessive glutamate release might lead to further excitation of neuronal circuits by positive feedback, resulting in a damaging avalanche of depolarization and calcium influx into neurons. This excitotoxic mechanism of neuronal injury may be important in acute neurologic disorders, such as stroke and CNS trauma, and possibly in some chronic neurodegenerative diseases, such as Alzheimer's.

Catecholamines

The catecholamines norepinephrine (noradrenaline), epinephrine (adrenaline), and dopamine are formed by hydroxylation and decarboxylation of the essential amino acid phenylalanine. Phenyl-ethanolamine-N-methyltransferase, the enzyme responsible for converting norepinephrine to epinephrine, is found in high concentration primarily in the adrenal medulla. Epinephrine is found at only a few sites in the CNS.

Dopamine is synthesized, via the intermediate molecule dihydroxyphenylalanine (DOPA), from the amino acid tyrosine by tyrosine hydroxylase and DOPA decarboxylase. Norepinephrine, in turn, is produced via hydroxylation of dopamine. Dopamine, like norepinephrine, is inactivated by monoamine oxidase (MAO) and catechol-O-methyltransferase (COMT).

Dopamine

Dopaminergic neurons generally have an inhibitory effect. Dopamine-producing neurons project from the substantia nigra to the caudate nucleus and putamen (via the nigrostriatal system) and from the ventral tegmental area to the limbic system and cortex (via the mesolimbic and mesocortical projections). In Parkinson's disease, there is degeneration of the dopaminergic neurons and the substantia nigra. Thus, dopaminergic projections from the substantia nigra to the caudate nucleus and putamen are damaged, and the inhibition of neurons in the caudate nucleus and putamen is impaired. The dopaminergic projection from the ventral tegmental area to the limbic system and cortex may be involved in schizophrenia.

Dopamine-containing neurons have also been found in the retina and the olfactory system. In these areas they appear to mediate inhibition that filters sensory input.

Norepinephrine

Norepinephrine-containing neurons in the PNS are located in the sympathetic ganglia and project to all of the postganglionic sympathetic neurons except those innervating sweat glands, which are innervated by axons that use ACh as a transmitter. Norepinephrine-containing cell bodies in the CNS are located in two areas: the locus ceruleus and the lateral tegmental nuclei. Although the locus ceruleus is a relatively small nucleus containing only several hundred neurons, it projects widely into the cortex, hippocampus, thalamus, midbrain, cerebellum, pons, medulla, and spinal cord. The noradrenergic projections from these cells branch extensively and are distributed widely. Some of the axons branch and project to both the cerebral cortex and the cerebellum. Noradrenergic neurons in the lateral tegmental areas of the brain stem appear to have a complementary projection, projecting axons to regions of the CNS that are not innervated by the locus ceruleus.

The noradrenergic projections from the locus ceruleus and the lateral tegmental area appear to play a modulatory role in the sleep–wake cycle and in cortical activation and may also regulate sensitivity of sensory neurons. Some evidence suggests that abnormal paroxysmal activity in the locus ceruleus can result in panic attacks.

Serotonin

Serotonin (5-hydroxytryptamine) is an important regulatory amine in the CNS. Serotonin-containing neurons are present in the raphe nuclei in the pons and medulla. These cells are part of the reticular formation, and they project widely to the cortex and hippocampus, basal ganglia, thalamus, cerebellum, and spinal cord. Serotonin-containing neurons can also be found in the mammalian gastrointestinal tract.

Serotonin-containing neurons, along with norepinephrine-containing neurons, appear to play an important role in determining the level of arousal. Firing levels of neurons in the raphe nuclei, for example, are correlated with sleep level and show a striking cessation of activity during rapid eye movement sleep. Serotonin-containing neurons may also participate in the modulation of sensory input, particularly for pain. Selective serotonin reuptake inhibitors, which increase the amount of serotonin available at the postsynaptic membrane, are used clinically as antidepressants.

Gamma-Aminobutyric Acid

Gamma-aminobutyric acid (GABA) is present in relatively large amounts in the gray matter of the brain and spinal cord. It is an inhibitory substance and probably the mediator responsible for presynaptic inhibition. GABA and glutamic acid decarboxylase (GAD), the enzyme that forms GABA from L-glutamic acid, occur in the CNS and the retina. Two forms of GABA receptor, GABA_A and GABA_B, have been identified. Both mediate inhibition but by different ionic pathways (see Table 3–6). Inhibitory interneurons containing GABA are present in the cerebral cortex and cerebellum and in many nuclei throughout the brain and spinal cord. The drug baclofen acts as an agonist at GABA_B receptors; its inhibitory actions may contribute to its efficacy as an antispasticity agent.

Endorphins

The general term endorphins refers to some endogenous morphine-like substances which bind to opiate receptors in the brain. Endorphins (brain polypeptides with actions like opiates) may function as synaptic transmitters or modulators. Endorphins appear to modulate the transmission of pain signals within sensory pathways.

Enkephalins

Two closely related polypeptides (pentapeptides) found in the brain that also bind to opiate receptors are methionine enkephalin (met-enkephalin) and leucine enkephalin (leu-enkephalin). The amino acid sequence of met-enkephalin has been found in alpha-endorphin and beta-endorphin, and that of beta-endorphin has been found in beta-lipotropin, a polypeptide secreted by the anterior pituitary gland.