CHAPTER 2: Cellular Basis of Communication

• Neurons are the principal cells in the brain that process information. There is a great diversity of neuronal cell types based on morphology, chemistry, location, and connections.

• The nucleus and major cytoplasmic organelles in the cell body of neurons synthesize and process proteins, which are subsequently transported to their appropriate locations within the neuron.

• The axon transports molecules and conducts action potentials to presynaptic terminals to initiate communication with other neurons, which occurs at synapses.

• Dendrites, multiple fine processes that extend from the neuronal cell body, together with the cell body, serve as the primary structure for the reception of synaptic contacts from other neurons.

• The cytoskeleton—the inner scaffold of a neuron formed by a system of interconnected protein filaments called microtubules, intermediate filaments, and actin filaments—plays a key role in the structure of neurons and in the transport of various proteins and organelles from the cell body to axonal and dendritic processes.

• Three major classes of glia—astrocytes, oligodendrocytes, and microglia—play important roles in brain function.

• The blood–brain barrier, formed by tight junctions between endothelial cells of capillaries in cerebral vascular beds, allows only small lipophilic substances to enter the brain from the general circulation.

• In their resting state, neurons maintain a negative electrical potential in relation to the extracellular environment. This results from differences between the intracellular and extracellular concentrations of K⁺, Na⁺, and Cl⁻ and the relative permeability of the cell membrane to these and other ions. The energy-consuming Na⁺/K⁺ pump helps to maintain appropriate ionic gradients across the membrane.

• The generation of all-or-none action potentials relies on the activities of voltage-dependent ion channels, highly specialized proteins that allow the flow of a specific ion (K⁺, Na⁺, or Ca²⁺) across neuronal membranes in response to changes in neuronal membrane potential.

• Sodium channels are the targets of many important drugs including local anesthetics and some antiseizure medications.

• The three general classes of potassium channels include voltage-gated potassium channels, calcium-activated potassium channels, and inward rectifiers.

• Entry of calcium into neurons through voltage-dependent calcium channels, of which there are five major classes—L-type, N-type, T-type, P/Q-type, and R-type—is important for neurotransmitter release and activation of intracellular signaling cascades. L-type calcium channel blockers are used to treat ischemic heart disease and hypertension.

• Mutations in ion channels are the cause of several neurologic disorders, including certain inherited neuromuscular disorders, epilepsy, and migraine syndromes.

It has been estimated that there are at least 100 billion neurons in the human brain, although this number in itself reveals little of the brain’s complexity. Unlike other organs, the brain contains an enormous diversity of cell types. Depending on the definition of a neuronal cell type, there may be thousands in the brain—each with their own structural, biochemical, and functional properties. Yet, the complexity of the brain as an information-processing organ extends beyond its cellular diversity. Neurons in the brain form diverse circuits that can range in scale from small local neuronal groups to long-distance projections. Many neurons function in more than one circuit and can communicate with thousands of other neurons.

Neuronal communication depends on both electrical and chemical carriers of information. The electrical mechanisms rely on the ability of each neuron to control the flow of ions across its membrane and thus to process and store information. Chemical signals are the means by which the vast majority of neurons communicate with each other. Neurons are not physically connected but are separated by a minute space, as small as 20 to 40 nm, called a synapse (2–1). The simplest synaptic arrangement involves an electrical impulse in one neuron, the presynaptic neuron, that triggers the release of a chemical substance, or neurotransmitter, which diffuses across the synaptic cleft and binds to specific receptors on another, the postsynaptic neuron. The binding of a neurotransmitter to its appropriate receptors precipitates changes in the electrical activity of the postsynaptic neuron, which in turn leads to the release of a neurotransmitter and further interneuronal communication. In a small fraction of cases, neurons physically connect with one another so that direct electrical transmission between the neurons can occur.

Neurons in the brain can form thousands of synapses with other neurons; extreme examples are a Purkinje cell in the cerebellum or a monoamine-containing cell in the brainstem that may form more than 100,000 synapses. Overall, in a single human brain, there are likely to be more than 100 trillion synapses. Complex processes of brain development that are not fully understood result in connections among neurons—some local, some over long distances—that are both highly specific and highly plastic.

The overall patterns of neural connectivity in the mammalian central nervous system (CNS) are dictated by a complicated set of genetically programmed interactions. Nevertheless, both spontaneous neural activity and neural activity that occurs in response to stimulation during gestation and throughout life have profound influences on the fine-tuning of an individual’s pattern of synaptic connections. Although there are finite critical periods of early postnatal development during which the pattern of neural connectivity in some circuits is markedly influenced by experience, the adult brain is far more plastic than previously thought and synaptic connectivity is modified throughout life. Thus, unlike computers, which are sometimes represented as artificial brains, the neural circuitry of the mammalian brain is not hardwired but instead constantly reacts and adapts to an ever-changing environment. The advent of optogenetics and related tools has vastly expanded the ability to study the function of precise neural circuits in the brain 2–1.

Although the neuron is the critical cell type for communication in neural networks, essential roles are played by glial cells. The brain contains three major classes of glia: astrocytes, oligodendrocytes, and microglia. Astrocytes have the most diverse functions, which include maintenance of the extracellular milieu (the composition, including ion concentrations, of extracellular fluid in the brain) for healthy neuronal function, metabolism of certain neurotransmitters, and formation of the blood–brain barrier. Astrocytes also are critical in the CNS response to injury. Oligodendrocytes produce myelin sheaths that encase axons and facilitate the conduction of action potentials. Microglia, together with lymphocytes and macrophages that migrate to the CNS from the periphery, are the cellular components of the brain’s immune system. All types of glia elaborate soluble factors, including neurotrophic factors and cytokines, which are involved in the maintenance of the nervous system and in its adaptation to changes in the environment (Chapter 8). Indeed, astrocytes are important in modulating the process of synaptic transmission. Glia also are key components in guiding the migration of growing neurons during development.

This chapter focuses on the basic features of neurons and glia and on the electrical excitability of neurons. The molecular and cellular basis of synaptic transmission and signal transduction are covered in subsequent chapters.

Neurons are highly asymmetric (polarized) cells that have three major components: a cell body (also known as a soma or perikaryon), a single long process called an axon, and a varying number of branching processes known as dendrites 2–2. Aggregations of neuronal cell bodies form the gray matter of the brain (named for its appearance on freshly cut sections). Axons make up the white matter, whose appearance results from the myelin sheaths that insulate many axons. Although neurons share a common set of features, they have a variety of sizes and shapes 2–3 and serve very different functions within the networks in which they operate.

Overview of the Neuron

The cell body contains the nucleus and major cytoplasmic organelles such as the rough and smooth endoplasmic reticulum (ER) and Golgi apparatus 2–4. It is primarily responsible for synthesizing and processing proteins, which are subsequently transported to their appropriate locations within the neuron. The nucleus contains genomic DNA that is transcribed into mRNA; mRNA is exported from the nucleus to the cytoplasm where it is translated into protein on ribosomes (Chapter 4). Although most mRNA remains in the cell body, some of it is transported to dendrites and axon terminals. Polyribosomes, which are multiple ribosomes arrayed on mRNA, and ER are found in dendrites, often right beneath synapses, where they presumably permit localized protein synthesis. Local control of protein synthesis may permit alterations to be made to specific synapses within a neuron, a mechanism that may underlie very precise changes in neural circuit function.

The cell body is the smallest part of the neuron; the bulk of cytoplasmic volume is distributed throughout the axon and the dendritic arbor. Yet, because it must produce components that sustain the rest of the neuron, the metabolic and synthetic demands on the neuronal cell body are immense. Thus, the cell body contains large numbers of mitochondria, which are the sites of oxidative phosphorylation and provide the main form of energy (adenosine triphosphatase [ATP]) used by all eukaryotic cells.

The axon is a fine tubular process that extends from the neuronal cell body; it conducts electrical impulses from the cell body to the axon terminals (presynaptic boutons) that form the presynaptic component of a synapse. Projection neurons are those that send axons to another region of the CNS or to the periphery, as opposed to interneurons whose axons remain within the CNS region of origin. Neurons generally have a single axon, the length of which varies from less than 1 mm for interneurons to more than 1 m for motor neurons that innervate the extremities. The axons of long projection neurons typically are myelinated; those of local circuit neurons usually do not have myelin sheaths. The axon normally emerges from a region of the cell body called the axon hillock 2–2, the region of the neuron from which an action potential is most often generated. As it approaches its terminal field of innervation, an axon may branch to varying degrees, depending on the number of neurons with which it makes synaptic contact. Axons also may give rise to recurrent collaterals that innervate other neurons in their vicinity and thereby serve feedback regulatory functions.

Dendrites are multiple fine processes that extend from the neuronal cell body and, together with the cell body, serve as the primary structure for the reception of synaptic contacts from other neurons. The geometry of dendritic arbors can be very complex and indeed beautiful (see 2–3). The precise location and extent of a dendritic arbor determine the role of a cell in a network. Many types of neurons have discrete spines protruding from their dendrites 2–5. Such spines typically receive the major excitatory inputs directed to their respective cells. Moreover, the spines are thought to structurally and biochemically isolate synapses so that each synapse serves as a small, individual unit of information processing (Chapter 3). Dendritic spines are not fixed structures; neurons regulate the number and morphology of spines, in some cases over minutes and hours, in response to neural activity and environmental signals.

The main functions of the dendritic tree include the reception, processing, and integration of incoming synaptic communications. Dendrites are both electrically and biochemically quite complex. Dendrites, like axons, contain voltage-dependent ion channels and thus can fire action potentials and actively propagate information to the soma. They also contain a wide variety of intracellular signaling molecules (Chapter 4) that are activated during synaptic communication and alter neuronal function.

The Cytoskeleton and the Transport of Proteins

The cytoskeleton, which represents the inner structure, or scaffold, of a neuron, is formed by a system of interconnected molecular filaments termed microtubules, intermediate filaments, and actin filaments. Microtubules are made of polymers of tubulin, a globular protein that forms a heterodimer between α and β tubulin. Microtubules copurify with several microtubule-associated proteins (MAPs) such as tau, which have significant roles in the assembly of microtubules, in cross-linking them to other filaments, and in transport functions. Intermediate filaments of neurons, called neurofilaments, are formed by three polypeptide subunits of low, middle, and high molecular masses. Actin filaments (also called microfilaments) are made of actin, a globular protein that self-assembles into a linear polymer. Microtubules and intermediate filaments are cross-linked to form a longitudinal scaffold for axons and dendrites. Actin microfilaments form a network underneath the entire surface membrane of the neuron; in dendrites and axons they are connected to microtubules and intermediate filaments. Actin microfilaments are heavily concentrated in dendritic spines and growth cones, both of which are dynamic structures that can respond to extracellular signals by changing shape. Many and perhaps all neurons in the CNS possess cilia. They are important for the migration of neurons during development but may also continue to serve important functions in the adult brain.

The cytoskeleton not only has important structural functions but also controls the transport of proteins between the cell body and its axonal and dendritic processes. Both fast anterograde and retrograde transport of proteins (100–400 mm per day) and slow anterograde transport (0.1–3.0 mm per day) occur in axons. Fast anterograde axonal transport involves the movement of transport vesicles, derived from the Golgi, along axonal microtubules. These vesicles contain many of the proteins necessary for the functioning of the presynaptic terminal.

The power to move vesicles along microtubules by fast axonal transport is derived from two force-generating proteins, kinesin and dynein. These proteins are ATPases that, by binding to microtubules, are stimulated to transport vesicles. Because microtubules have an intrinsic polarity, with plus ends pointing toward presynaptic terminals and minus ends pointing toward the cell body, they can serve as a compass to direct vesicle traffic. Kinesin and dynein also have an intrinsic polarity; kinesin moves vesicles only toward the plus end of microtubules and therefore is the motor protein for anterograde fast transport, and dynein moves them only toward the minus end and thus is the motor protein for retrograde fast transport. Retrograde axonal transport functions to return various membrane molecules to the soma for elimination and is also important for communicating information from nerve terminals to the soma.

The molecular mechanisms responsible for slow axonal transport and for the transport of proteins to dendrites are not well understood. Slow axonal transport appears to utilize the proteins that comprise the cytoskeleton itself—a dynamic rather than a static structure that is continually renewed. Proteins that are specifically directed toward dendrites rather than axons must be guided by molecular recognition signals that dictate the direction of transport. The dendritic cytoskeleton is different from that of axons. Unlike axons, in which the polarity of microtubules is fixed, dendrites have microtubules whose polarity varies because equal numbers of microtubules are oriented in each direction. The proteins associated with dendritic microtubules also differ from those of axonal microtubules; MAP2 is found only in dendrites and in the cell body, whereas the protein tau is found almost exclusively in axons. Perhaps these differences help guide the transport of specific proteins to specified sites within neurons.

The Synapse

A synapse is a specialized structure involved in the transmission of information from one neuron to another. Synapses typically are composed of a single presynaptic element—the presynaptic terminal or bouton of an axon—and a postsynaptic element, which for excitatory synapses is often a dendritic spine. Although the synaptic cleft lies between these two elements, it is incorrect to think of the cleft as empty space that separates two independent structures. Instead, the presynaptic and postsynaptic elements of a synapse are tightly bound to one another and to the extracellular matrix (composed of proteoglycans, other polysaccharides, and proteins) by means of numerous scaffolding proteins, such as cell adhesion molecules (CAMs), and by astrocytes (see 2–2). The synapse should therefore be considered an individual unit whose sole purpose is to transmit, process, and store information.

Most synapses in the brain utilize chemical transmission. Neurotransmitters 2–2 are typically released from presynaptic terminals, and diffuse across a synaptic cleft and bind to specialized receptor proteins on postsynaptic cells (Chapters 3 and 4). The presynaptic terminal contains specialized cellular structures that allow it to remain, to a certain extent, metabolically and functionally independent from the neuronal cell body. It contains large numbers of mitochondria to provide energy, enzymes to synthesize and degrade neurotransmitters, and synaptic vesicles to store substantial concentrations of neurotransmitters while waiting for a signal to be released. The postsynaptic dendritic membrane is markedly enriched with appropriate neurotransmitter receptors and elaborate intracellular signaling machinery.

Synapses that involve the innervation of a postsynaptic dendrite by a presynaptic nerve terminal represent just one anatomic arrangement of chemical synapses. Other arrangements, in particular, where substances are released by dendrites that act on nerve terminals, are described in 2–3. Overall, chemical synapses are predominant in the CNS but do not represent the only form of synaptic transmission. In a small minority of cases, neurons are connected by means of a gap junction rather than separated by an intervening space. Gap junctions, which are formed by a large number of tightly packed proteins (connexons), produce so-called electrical synapses that permit electrical currents to flow directly between cells.

Every heartbeat, every nerve impulse, every movement, and every thought is critically dependent on the tightly controlled and precisely timed flow of ions across cell membranes. A disruption of this flow, for example, by the puffer fish poison tetrodotoxin, can be fatal. Ion channel abnormalities are responsible for many human diseases (now called channelopathies). Mutations of ion channels can result in severe deficits in neuromuscular functioning, including episodic ataxias and paralyses, myotonia, and long QT syndrome of the heart as well as seizure disorders 2–1. Moreover, ion channels are also the targets of widely used and efficacious pharmacologic agents. For example, phenytoin and carbamazepine, which are used to treat epilepsy, act by altering Na⁺ channel kinetics. Lidocaine and procaine, common local anesthetics, block voltage-gated Na⁺ channels and prevent the conduction of nerve impulses that signal the occurrence of tissue damage and therefore pain. An awareness of ion channel structure and function is crucial for understanding neuropharmacology and neuronal disease processes.

Electrical Potential in Cells

An animal’s nervous system receives information from the environment, integrates this information with past experience, and creates a behavioral response that promotes the survival of the organism. Sensory neurons receive information from both the external environment—for example, sounds and sights—and the internal environment, such as the sensation of hunger or the stretch of a muscle, and relay messages to other neurons. Neurons that receive these messages are responsible for directing appropriate signals to various locations that may include muscles, internal organs, glands, or other neurons. The swift communication of these signals typically benefits an organism. It can, for example, enable a prey’s rapid response to the appearance of a predator, or produce a reflexive postural correction that is necessary to prevent a fall.

Neurons convey signals rapidly by alternately maintaining and varying their electrical potentials in relation to the extracellular environment. All neurons maintain a negative electrical potential (also known as a “membrane potential”) relative to their extracellular environment. This negative potential provides a driving force for charged particles; in the absence of other forces, positive charges tend to be drawn into the cell, and negative charges tend to be repelled from the cell. When a neuron is activated by an external stimulus such as a chemical signal from another neuron, or by an event in the environment, it may depolarize; that is, its electrical potential may become less negative relative to the extracellular milieu. A neuron may, for example, depolarize from −70 to −50 mV. If a neuron undergoes significant depolarization, it may generate an action potential—a brief, all-or-none depolarization and repolarization of the membrane potential 2–6. Such substantial, rapid depolarization will often stimulate a neuron to release neurotransmitters and thus convey information to other cells through chemical signals; for example, a signal conveyed to muscle cells may cause the contraction of a muscle, and a signal conveyed to an internal organ may stimulate or attenuate its activity.

How Neurons Maintain Electrical Potential

Two characteristics of nerve cells contribute to their ability to maintain an electrical potential. First, different types of ions are unequally distributed across the neuronal cell membrane. Generally, the neuron’s interior contains a higher concentration of K⁺ and lower concentrations of Na⁺, Ca²⁺, and Cl⁻ than does the extracellular space. These ionic gradients, which occur not only in neurons but also in most cell types,¹ are maintained through ion-specific pumps that require energy. Moreover, the neuron’s interior contains many negatively charged, membrane-impermeable proteins that do not exist in the extracellular milieu.

Second, the neuronal cell membrane is differentially permeable to ions. In the resting state, most neurons are highly permeable to K⁺, somewhat permeable to Cl⁻, and very slightly permeable to Na⁺ and Ca²⁺. Because the cell membrane is a lipid bilayer, it would be completely impermeable to ions if it did not contain specialized proteins for ion transit. Ions strongly prefer interaction with the polar water molecules in intracellular and extracellular spaces to interaction with the hydrophobic lipid groups that comprise the bulk of the cell membrane. The selective permeability of the cell membrane depends on the numbers and states of its various ion channels. These neuronal properties—unequal permeability of ions across the cell membrane and unequal distribution of ions—are theoretically sufficient to maintain an electrical potential in a cell.

A Simple Cell Model

Consider a very simple model of a cell. The interior of the cell (I) contains 100 mM K⁺A⁻, where A⁻ is an impermeant anion such as a negatively charged protein. Exterior to the cell (O), the concentration of K⁺A⁻ is 5 mM 2–7. To simplify this model, the effects of osmosis are ignored.

If the cell’s lipid bilayer is impermeable to both K⁺ and A⁻, no movement of ions occurs across the bilayer, and the concentration of ions on each side of the bilayer remains constant. However, if the lipid bilayer of this cell became permeable to K⁺, and only to K⁺—for example, from the opening of K⁺-specific ion channels in the lipid bilayer—K⁺ ions but not A⁻ ions would be free to move across the lipid bilayer in both directions. Because I contains 20 times more K⁺ than O, many more K⁺ ions would be likely to travel from I to O than from O to I. Consequently, a net efflux of K⁺ from I to O would occur; because A⁻ would remain impermeable, it would not cross the membrane.

As soon as K⁺ ions begin to move to O, however, a net positive charge develops in O because K⁺ has left the cell without accompanying A⁻. This net positive charge repels K⁺ from O. Thus, two tendencies act in opposition to one another: (1) the tendency of K⁺ to move out of I because more K⁺ is present in I than in O and (2) the tendency of K⁺ to be drawn into I because of its relative negative potential.

Eventually, these two tendencies reach an equilibrium whereby the net efflux of K⁺ from I (favored by the concentration gradient) is equal to the net efflux of K⁺ from O (promoted by the electrical potential). Only a minuscule fraction of K⁺ must venture from I to O to create an electrical potential strong enough to balance the tendency of K⁺ to leave I along its concentration gradient. As a consequence, a minuscule fraction of the A⁻ ions in I exists without accompanying K⁺ ions, and this tiny separation of charge creates a negative potential of −75 mV in I relative to O. This negative charge within the cell exerts just enough force on the K⁺ ions that the net flow of K⁺ ions across the membrane is zero, despite the existing concentration gradient of K⁺. The transmembrane electrical potential at which this equilibrium occurs (−75 mV) can be determined by the Nernst equation² and is termed the equilibrium potential (or the Nernst potential) for K⁺.

A More Complicated Cell Model

A basic understanding of membrane potential drawn from the previous cell model can assist in understanding a model that more closely resembles a mammalian nerve cell. It involves: (1) a more complete complement of extracellular and intracellular ions and (2) more realistic ionic permeabilities. 2–2 describes several features of this cell model. First, the net total charge on each side of the cell is zero, a condition that is necessary for electrical neutrality.³ Second, the cell membrane is far more permeable to K⁺ than to any other ion, which is the case for most neurons at rest. All other permeabilities are described relative to the permeability of K⁺.

What is the electrical potential at which the net flux of charge across the cell membrane equals zero? If the cell were only permeable to K⁺, the membrane potential would rest at the equilibrium potential of K⁺, because all other ions would be trapped on one side of the cell and could not migrate to contribute to the generation of an electrical potential. However, the cell is permeable to other ions, although to a much lesser degree than it is to K⁺. If the membrane were permeable only to Na⁺, the resting potential would be at the equilibrium potential of Na⁺. In this case, the cell interior would develop a positive potential relative to the exterior of the cell. The positive potential in the cell interior would repel Na⁺ ions and balance the statistical tendency of these ions to flow down their concentration gradient and into the cell.

The true equilibrium potential for this cell model would be expected to lie somewhere between the equilibrium potentials for the various ions. The membrane permeability for each type of ion determines the relative contribution of ion on the inside and outside of the membrane to the equilibrium potential. The equilibrium potential for this cell, −68 mV, lies between the equilibrium potentials estimated for K⁺, Na⁺, and Cl⁻.⁴ Because the membrane’s permeability to K⁺ is so much greater than its permeability to the other ions, the neuron’s resulting membrane potential is much closer to the equilibrium potential of K⁺ (−75 mV) than to that of Na⁺ (+58 mV) or Cl⁻ (−59 mV).

Many electrophysiologic concepts and the roles of ion channels and ion channel–targeting drugs in physiologic processes can be understood intuitively by understanding the change in membrane potential caused by changing the permeability of the membrane to different types of ions via the opening and closing of ion channels. When only one type of ion channel opens, it drives the membrane potential of the cell toward the equilibrium potential of that ion. For example, if many of the Na⁺ channels in this cell model were suddenly opened, causing Na⁺ to be three times more permeable than K⁺, the membrane potential would move toward the equilibrium potential for Na⁺ (in this case +19.6 mV). If the Na⁺ channels were to close suddenly, bringing the permeability ratio of Na⁺ to K⁺ back to its original value of 0.01, the membrane potential would return to −68 mV in response to the efflux of K⁺ through the many open K⁺ channels.

It also is important to emphasize that the equilibrium potential for a particular ion depends on the relative concentrations of that ion inside and outside the cell, which can vary considerably among different cell types and tissues. For example, the equilibrium potential for Cl⁻ can range from −60 to −90 mV.

Maintenance of Membrane Potential by ATP-Dependent Pumps

The equilibrium potential provides for an equal exchange of cations back and forth across the cell membrane: for every excess Na⁺ ion that sneaks across the membrane, a K⁺ ion moves out, holding the membrane stable at the predicted potential. This exchange occurs slowly enough that the ionic concentrations may be considered constant for short period of time. However, if a slow exchange of Na⁺ for K⁺ were allowed to continue for hours or days, the concentration gradients would eventually degenerate and the membrane potential would slowly begin to dissipate. Na⁺–K⁺ pumps maintain the ionic gradients across a cell membrane by extruding Na⁺ from and pumping K⁺ into the cell against their respective concentration gradients at the cost of energy 2–8. Each pump is a multimeric integral membrane protein consisting of transmembrane α subunits, which possess the catalytic and iontophoretic (ion pore–containing) domains of the pump, accessory transmembrane β subunits, which mediate the trafficking of the α subunit to the cell membrane, and tissue-specific regulatory subunits from the FXYD protein family.

The catalytic subunit of each pump has extracellular binding sites for K⁺ and intracellular binding sites for Na⁺ and ATP. ATP transfers its terminal phosphate to the catalytic subunit in a Na⁺-dependent manner; because the pump is a protein that cleaves ATP by means of ion-dependent enzymatic activity, it often is referred to as an Na⁺/K⁺-ATPase. At the expense of the energy of hydrolysis of ATP, typically three Na⁺ ions are transported out of the cell and two extracellular K⁺ ions are transported in. Thus, the pump is electrogenic: it exports more positive charge than it imports. The phosphorylated catalytic subunit is subsequently hydrolyzed in the presence of K⁺ ions, returning the catalytic subunit to its resting state. The resting potential of a cell with active Na⁺–K⁺ pumps is usually a few millivolts more negative than would be predicted on the basis of ion distribution and relative permeabilities alone.

The importance of Na⁺–K⁺ pumps in the maintenance of cellular membrane potential becomes quite evident when they are inhibited by pharmacologic agents. Cardiac glycosides such as ouabain and digoxin, which increase the contractile force of cardiac muscle and are used in the treatment of congestive heart failure and some cardiac dysrhythmias, are the best known inhibitors of the Na⁺–K⁺ pump. Normally, cardiac myocytes maintain low levels of intracellular Ca²⁺, partly through a Na⁺–Ca²⁺ pump that uses energy from the movement of Na⁺ down its concentration gradient to transport Ca²⁺ out of the cell. Because cardiac glycosides inhibit Na⁺–K⁺ pumps and increase intracellular Na⁺ concentrations, they make Na⁺–Ca²⁺ pumps less effective. Consequently, intracellular Ca²⁺ concentrations increase, which increases the contractile force of cardiac muscle. These drugs also can slowly reduce the resting potential of neurons eventually to zero; in large neurons, this decline in potential occurs after several hours. This action of the drugs in the CNS accounts for common side effects of cardiac glycosides, including disturbed vision, confusion, and delirium.

In the cell model described previously, it was possible to create a potential across the membrane by making it selectively permeable to the ions distributed unequally across it. It also was possible to alter the potential of the membrane by changing the relative permeabilities of various ions, for example, by increasing the permeability of Na⁺ relative to K⁺. Likewise, the membrane’s original potential could be restored by reinstating the original permeabilities. In addition to these features, real neurons have two biophysical properties that affect the movement of charge and the development of potential across a neuronal membrane.

First, unlike the movement of charge in the cell model, charge is not transferred instantaneously from one neuronal compartment to another. Electrically, the membrane can be thought of as a resistor and a capacitor 2–9. Resistance describes a membrane’s ability to pass ions, and it is determined by the number and properties of the ion channels, and the thickness of the membrane. Thick membranes, such as myelinated axon membranes, are high resistors that are difficult to let ions through. Capacitance describes a membrane’s ability to store charge; the larger the membrane’s capacitance, the more charge required to raise a membrane’s potential. Several properties such as size and, most significantly, thickness can affect a membrane’s capacitance. A very large membrane area requires more stored charge to bring it to a given potential, or, in other words, it has a greater capacitance, than does a small membrane area. Very thick membranes are poor capacitors: because ions separated across a greater distance possess greater potential energy, fewer ions are required to reach a certain membrane potential. Because a myelin sheath increases a membrane’s thickness, a myelinated axonal membrane is a high resistor and poor capacitor compared with a nonmyelinated membrane.

Sensory, Synaptic, and Action Potentials

Neurons exploit their ability to rapidly change their transmembrane potentials in order to receive information from the environment and to relay messages. Input from a variety of sources, including other neurons, can cause a neuron’s membrane potential to fluctuate. If a neuron depolarizes enough or reaches threshold, it produces an action potential: a rapid, all-or-none depolarization that propagates down its axon. The firing of an axon generally leads to the release of neurotransmitter from the axon’s terminals, which in turn conveys the neuron’s signal to muscle cells, to effector organs such as glands, or to other neurons.

Receiving information

Some neurons are equipped with specialized systems that enable them to receive information from the environment. Hair cells in the cochlea, for example, are sensitive to vibrational energy: vibrations cause the movement of tiny cilia on the surface of these cells. Such movement activates mechanosensitive ion channels that increase the membrane’s permeability to Na⁺ that in turn leads to depolarization of the hair cell. Photoreceptor cells in the retina can respond to light because photons activate a series of chemical reactions that cause changes in the ionic permeability of the cell membrane, which in turn leads to changes in the membrane potential of the cell.

Neurons generally receive signals from other neurons through chemical neurotransmitters. They typically release neurotransmitter at synapses where it binds to receptors on an adjacent neuron’s cell membrane. The binding of neurotransmitter to a receptor routinely leads to changes in the receiving neuron’s ion permeability. This change in permeability may occur directly; many neurotransmitter receptors are ligand-gated ion channels. However, changes in permeability also take place indirectly; many types of neurotransmitter receptors activate second messenger systems within a cell that in turn modify ion channels, causing changes in membrane potential (Chapters 3 and 4).

Integrating information

The opening of ion channels in a localized area, such as a synapse, produces a transmembrane current that changes the membrane potential of a cell. However, this change in potential does not occur instantaneously or remain localized to the narrow region of membrane in which the current was generated. The integration of local changes in the transmembrane voltage of a neuron is affected by two basic types of summation: spatial summation and temporal summation 2–10.

Neurons are “decision makers”—they must continually decide whether to respond to particular sets of stimuli by firing action potentials and in turn communicate with fellow neurons or with effector organs. A motor neuron in the spinal cord, for example, receives thousands of excitatory and inhibitory inputs from pathways that descend from the brain. These descending pathways deliver enormous amounts of information integrated from many areas of the brain, including motor planning, vestibular, and visual centers. The currents produced by excitatory and inhibitory inputs continually undergo summation to produce membrane potentials that fluctuate in time and space. A neuron reads fluctuating potentials at the base of its axon, the axon hillock, where it has a high concentration of voltage-dependent Na⁺ channels. If the sum of these potentials produces a sufficient level of depolarization, an action potential is triggered.

Action potential and neurotransmitter release

An action potential is a rapidly propagating depolarization of the axonal membrane that can lead to the release of neurotransmitter from axon terminals. This phenomenon is best understood in terms of the responses of Na⁺ channels to changes in voltage. In this section the Na⁺ channel is discussed as an abstract entity; its molecular structure is described in detail in the next section.

The opening of Na⁺ channels (voltage gated or voltage dependent) on membrane depolarization leads to an inward current that in turn causes further depolarization because the equilibrium potential of Na⁺ is positive compared with the resting potential of the cell membrane. Voltage-gated Na⁺ channels continue to open and to permit further depolarization in a repetitive, self-perpetuating manner. In response to this positive feedback loop, the membrane of the cell’s axon is quickly driven toward the equilibrium potential for Na⁺.

However, two phenomena quickly return the membrane potential to its resting, hyperpolarized state 2–11. First, Na⁺ channels spontaneously close, or inactivate, after a brief period of time (1–2 ms), thus reducing the membrane’s permeability to Na⁺ and minimizing the contribution of the Na⁺ gradient to the membrane’s potential. Second, voltage-activated K⁺ channels open, although their response to membrane depolarization is slower than that of Na⁺ channels. Subsequently, the transient influx of Na⁺ is rapidly overwhelmed by an outflow of K⁺ along its concentration gradient, and the membrane responds by resuming its resting potential near the K⁺ equilibrium potential. Although the automatic closing of the Na⁺ channels and the membrane’s high resting conductance to K⁺ are sufficient to return the membrane to its resting potential, voltage-gated K⁺ channels accelerate the process. In fact, the amount and properties of the various K⁺ channels in a neuron play a critical role in determining the duration of a neuron’s action potentials and the neuron’s rate of firing.

Because there are Na⁺ channels along the length of the axon, the action potential propagates down the axon and invades the presynaptic nerve terminals, where it triggers the influx of Ca²⁺ by activating voltage-dependent Ca²⁺ channels and subsequently leads to the Ca²⁺-dependent release of neurotransmitter (Chapter 3). In nonmyelinated, small-diameter axons, action potential propagation is fairly slow because propagation is continuous and each segment of membrane must be depolarized to threshold. In myelinated axons, action potential propagation is saltatory: action potentials appear to jump from one node of Ranvier, a small gap in the myelin sheath, to the next at a much faster rate 2–12. Because myelin increases the resistance and decreases the capacitance of the axonal membrane, current entering the axon at one node does not leak out of the internodal membrane but instead travels rapidly down the axon to depolarize the adjacent node until it reaches threshold. Disruption of myelination of axons under pathological conditions, such as multiple sclerosis, impairs the conductance of action potentials, and causes neurologic symptoms such as vision loss and muscle weakness (Chapter 12).

Dendrites, the recipients of most incoming synaptic activity, are not electrically passive during this process; they contain a constellation of voltage-dependent Na⁺, Ca²⁺, and K⁺ channels and are capable of generating action potentials. Dendritic action potentials presumably amplify incoming synaptic signals so that they can be detected at the soma.

Thus far, ion channels have been discussed only in the most abstract sense. However, modern advances in molecular biology, such as the isolation, cloning, and mutagenesis of ion channels, and in cell physiology, such as patch clamping techniques that allow the observation of single ion channels, have provided scientists with remarkably detailed knowledge of the operation of ion channels. The crystallization of simple ion channels also has yielded important functional and structural information. Consequently, we know the exact amino acid composition of hundreds of ion channels, have detailed models about how these channels assemble in cellular membranes, and understand which amino acids are responsible for features of ion channels such as voltage gating, selective permeability, and the binding of pharmaceutical agents.

This section focuses on the voltage-gated–like (VGL) ion channel superfamily that includes but is not limited to voltage-gated Na⁺, Ca²⁺, and K⁺ channels (Na_V, Ca_V, and K_V). The extended family members also include Ca²⁺-activated K⁺ channels (K_Ca), cyclic nucleotide–modulated ion channels (cyclic nucleotide gated [CNG] and hyperpolarization-activated cyclic nucleotide gated [HCN]), transient receptor potential channels (TRP), inward rectifier K⁺ channels (K_ir), and two-pore K⁺ channels (K_2P). (Note that the term rectifier describes a channel that passes current much more efficiently in one direction; for example, an inward rectifier preferentially passes current into a cell.) Members of this gene family are principal players in regulating the electrical signals in neurons. We review these channels in considerable detail with the expectation that these proteins represent valuable targets for the development of future treatments for psychiatric and neurologic disorders. Another major family of ion channels, ligand-gated ion channels, which is responsible for translating chemical signals into electrical signals via generation of synaptic potentials, is discussed in subsequent chapters.

Structure of VGL Ion Channels

The fundamental features of an ion channel are its aqueous pore that controls ion permeation, the channel’s gating mechanism, and its modulation. The relationship among the >140 protein members of this ion channel family is built on the amino acid sequences of the minimum pore region 2–13. The founding members of this family are voltage-gated Na⁺ channels. Their principal α subunit consists of four internally homologous domains that surround a central pore. Each of the domains contains six membrane-spanning α helices (S1–S6), and a membrane reentrant loop between S5 and S6, also called SS1–SS2, that is believed to form a hairpin loop that dips down into the channel and forms the lining of the ion pore 2–14.

Other voltage-gated channels display structures remarkably similar to those of the Na⁺ channel. Most similar are the principal α subunits of Ca²⁺ channels, which also contain four internally homologous domains, each of which possesses six putative membrane-spanning regions 2–14. Voltage-gated K⁺ channels appear to be composed of subunits that correspond to only one of the four internally homologous domains of the Na⁺ and Ca²⁺ channels. It is believed that four of these smaller subunits multimerize in the plasma membrane to form a channel that is similar in structure to the Na⁺ and Ca²⁺ channels. Several other families of ion channels (K_Ca, CNG, HCN, and TRP) also have this tetrameric structure. However, although the inwardly rectifying K⁺ channels are also tetramers, each of the subunits only has two transmembrane segments, named M1 and M2, which are analogous to the S5 and S6 segments of voltage-gated Na⁺, Ca²⁺, and K⁺ channels. On the other hand, two of these pore motifs are linked together to generate the two-pored K⁺ channels. In mammalian cells, each of these channels is composed of several subunits in addition to their primary α subunits; the role of these accessory proteins is discussed in 2–4.

Selectivity of Ion Channels

K⁺ channels are 100 to 1000 times more permeable to K⁺ than Na⁺. Na⁺ channels are 12 times more selective for Na⁺ than for any other ion, and some Ca²⁺ channels are 1000 times more selective for Ca²⁺ than for other cations. A channel’s ability to select one cation over another suggests remarkably sophisticated protein design, and how this is accomplished has been the focus of intense investigation.

Ions do not flow through ion channels like water through a pipe; such a model cannot explain how some channel pores are selective for Na⁺ and others are selective for K⁺. Instead, an ion binds transiently to one or more sites in a channel, with its permeability presumably determined by the amount of energy released during its binding to amino acid residues and by the energy required for its dissociation from some or all surrounding water molecules. The speed with which an ion can dissociate from a pore and escape from a channel is also likely to influence selectivity.

Crystallization of the bacterial KcsA K⁺ channel has provided great insight into how ionic selectivity may be accomplished in K⁺ channels. The selectivity filter is the narrowest part of the pore, formed by the SS1–SS2–like region of the KcsA channel, with a very conserved GYG sequence lying at the heart of the selectivity filter. Voltage-gated Na⁺ and Ca²⁺ channels probably use a different strategy to achieve selectivity. Both types of channels use negatively charged residues, and perhaps some backbone carbonyls as well, to coordinate the cation.

Opening of Ion Channels

Na⁺ channels responsible for propagating action potentials, K⁺ channels responsible for terminating action potentials, and Ca²⁺ channels that admit Ca²⁺ to nerve terminals in response to action potentials are all voltage activated: they open when a membrane is depolarized. It is now known that voltage-activated ion channels display a gating charge, or a shift in the distribution of charge across the membrane, that occurs concomitantly with channel activation and is believed to result from the movement of a putative voltage sensor within the ion channel itself. The S4 transmembrane domain, marked by a unique pattern of positively charged amino acids, may behave as a gating particle 2–14. However, the exact conformational change that results from movement of the S4 domain and causes the opening of a channel remains controversial.

The addition of regulatory domains to the carboxyl terminus of K_ir, K_Ca, CNG, and HCN channels yields gating by binding of small intracellular ligands such as Ca²⁺, ATP, and cyclic nucleotides or by interactions with protein ligands. Ligand binding to these domains is thought to rotate the S6 segment to open the pore. For K_Ca and HCN channels, ligand binding and membrane depolarization act in concert to open the pore.

Closing of Ion Channels

In some cases, the closing of a voltage-gated ion channel is simply the opposite of its opening: the channel undergoes conformational changes in response to a particular membrane voltage and subsequently returns to its resting conformation when the membrane voltage change subsides. This process is called deactivation. For example, the closure of voltage-gated ion channels such as the delayed rectifier K⁺ channel, which restores an axon’s resting potential after an action potential, does not appear to require additional regulatory mechanisms.

In response to depolarization, however, a voltage-gated Na⁺ channel closes immediately after it is activated, even if the depolarization is maintained, such as during an action potential. This process of inactivation is distinctly different from that of deactivation, the return of the Na⁺ channel to its resting state. The “ball and chain” model of inactivation postulates the movement of a positively charged segment of a channel protein into an open ion channel, which prevents further conductance 2–14. Site-directed mutagenesis indicates that the ball and chain portion of the Shaker K⁺ channel (a prototype of K_V channels discussed below) is most likely its N terminus, and that of the Na⁺ channel the cytoplasmic region between domains III and IV. In some Ca²⁺ channels, residues in the S6 segment of domain I, and the loop between domain I and II, appear to be functionally important in voltage-dependent inactivation. Some subtypes of Ca²⁺ channels also exhibit Ca²⁺-dependent inactivation, which involves the C-terminal region 2–14.

Several toxins and pharmaceutical agents act by modulating the process of inactivation. Some toxins selectively slow Na⁺ channel inactivation, causing these channels to remain open for a longer period of time; a class of scorpion toxins that act in this manner can lead to spastic paralysis, seizures, and death. Other substances, including some drugs used in the treatment of epilepsy (eg, phenytoin and carbamazepine), selectively stabilize the inactivated state of the Na⁺ channel, thereby decreasing neuronal excitability.

Types of Ion Channels

The voltage-dependent cation channels are similar enough in general architecture that it is possible to discuss their basic properties as a group. However, rapid advances in molecular biology, pharmacology, and electrophysiology have made it possible to distinguish hundreds of variations of ion channels. Distinct channels arise from different gene products, splice variants of the same gene transcript, posttranslational modifications, and combinations of individual channel subunits. The following sections of this chapter serve as an abbreviated catalogue of the various types and subtypes of voltage-gated cation channels and other types of cation channels from the VGL family 2–13. The properties, molecular composition, and medical relevance of each are summarized in the most general terms; more information about each channel type can be found in the references listed at the end of the chapter and from online resources.

Sodium Channels

Na⁺ channels are a relatively homogeneous group of ion channels. Variations do exist: in mammals, nine principal α subunit genes of Na⁺ channels have been cloned, and they show different tissue distribution patterns, and different sensitivities to the typical Na⁺ channel blocker tetrodotoxin. However, across species and tissue types, Na⁺ channel properties are remarkably similar in terms of voltage dependence and activation–inactivation kinetics.

Pharmacology and toxicology of sodium channels

Toxins that bind to Na⁺ channels are uncommon, but their effects can be dramatic. The isolation of the Na⁺ channel was initially possible because it binds tightly and specifically to toxins such as tetrodotoxin, a heterocyclic compound found in many marine species, including the Japanese puffer fish, the globe fish, and the blue-ringed octopus. Approximately 200 humans experience tetrodotoxin poisoning each year, in most cases after consuming improperly prepared puffer fish. Symptoms of tetrodotoxin poisoning include facial numbness, headache, and increasing paralysis and respiratory distress. Although victims are completely paralyzed, they may be conscious and lucid until shortly before death, which generally occurs within 4 to 6 hours. Among those poisoned, the mortality rate is approximately 50%; survival is usually attributed to rapid treatment (gastric lavage) or the victim’s ingestion of only a small amount of toxin. Because tetrodotoxin binds to the outside of the Na⁺ channel at the S5–S6 loop of the α subunit, site-directed mutagenesis of amino acids in this region greatly reduces its ability to bind to the channel (see 2–14). Changes in amino acid sequence in this region in certain Na⁺ channel subunits cause these channels to be resistant to tetrodotoxin to varying degrees.

Other Na⁺ channel toxins act not by blocking the channel but by activating it, usually by slowing or eliminating its voltage-dependent inactivation, which leads to the hyperexcitability of neurons and muscle tissue. Victims of these toxins, which have been isolated from scorpion and sea anemone species, may undergo convulsions and spastic paralysis.

Substances used to treat human disease generally have less dramatic effects on the Na⁺ channel. Phenytoin and carbamazepine, which are used to treat epilepsy, act on Na⁺ channels by slowing their recovery from an inactivated state. Because prolonged inactivation of these channels limits the firing rates of affected neurons, these drugs can prevent seizures. Local anesthetics, such as cocaine, and synthetic derivatives such as lidocaine and procaine, which are membrane permeable in their nonionized form, bind to the cytoplasmic side of the Na⁺ channel and also cause Na⁺ to bind to the inactivated state of the Na⁺ channel, creating a use-dependent blockade of the channel.

The sodium channel in disease

Some inherited neurologic disorders, including certain myotonias and periodic paralyses, are linked to Na_V1.4 (also known as SkM1), the Na⁺ channel in adult skeletal muscle. Although myotonia, which is characterized by muscle excitability, and periodic paralysis, which is characterized by muscle hypoexcitability, seem to be diametrically opposed, the underlying Na⁺ channel defects are similar. Electrophysiologic studies indicate that both disorders involve slowed or impaired inactivation of the Na⁺ channel. In individuals who experience periodic paralysis, the defect in Na⁺ channel inactivation is severe enough that muscle remains at a depolarized potential and becomes refractory to further action potentials; in those who experience myotonia, the defect is less severe and results in repetitive firing. Mutations in neuronal and cardiac Na⁺ channel subunits that slow or impair the inactivation of Na⁺ channels are also linked to familial disorders such as familial neonatal seizures, generalized epilepsy, and long QT syndrome 2–1. Of great interest is the recent finding that certain sodium channel subunits, such as Na_V1.7, Na_V1.8, and Na_V1.9, are preferentially expressed in peripheral sensory nerves and are critically important in the transmission of pain signals, raising the possibility that blockers of these channels might be useful in the treatment of pain syndromes (Chapter 11).

Potassium Channels

K⁺ channels generally act as a stabilizing force. Their diverse functions include setting a cell’s resting potential, repolarizing the cell after an action potential, and controlling an action potential’s rate of firing and shape. Proper K⁺ channel functioning is a critical determinant of neuronal activity, and altered expression levels of specific K⁺ channel subunits in response to prolonged periods of neuronal activation or inhibition play an important role in maintaining normal cell excitability, a process termed whole cell, nonsynaptic, or homeostatic plasticity. Mutations in K⁺ channels lead to a variety of disorders, ranging from ataxia to cardiac arrhythmias 2–1. A malfunctioning K⁺ channel generally causes some form of hyperexcitability in affected tissue; for example, the long QT syndrome, a delay in ventricular repolarization that can cause heart arrhythmias, results from a malfunction of the K_V11.1 (HERG) K⁺ channel, which is responsible for repolarizing the ventricle after contraction. Indeed, binding to HERG channels is a common cause of cardiac side effects of many classes of medications and is now used as a routine screen during the drug discovery process. Mutations of the related K_V7.2/7.3 (KCNQ2/3) K⁺ channel cause benign familial neonatal convulsions, which resolve as the affected child grows. Mutation of another K⁺ channel K_V1.1, which is expressed at high levels in the cerebellum, is linked to a type of episodic ataxia believed to be caused by an abnormal increase in the firing of cerebellar cells. Tetraethylammonium (TEA), which selectively blocks most types of voltage-gated K⁺ channels, has been a useful research tool.

Potassium channel classifications

K⁺ channels can be placed into two broad classes, six TM (transmembrane domain) and two TM channels 2–3. Six TM channels include voltage-gated K⁺ channels (K_V1–12 families) and Ca²⁺-activated K⁺ channels (K_Ca1–5 families). Two TM channels are inwardly rectifying K⁺ channels (K_ir1–7 families). The differences in these channels reflect the variety of ways that they can affect the firing properties of a neuron.

Voltage-gated K⁺ channels subsequently have two different functional classes that are not categorized by subfamilies, delayed rectifiers and A-type channels. Delayed rectifiers undergo delayed activation after depolarization and inactivate slowly; they thereby facilitate repolarization by remaining open for a prolonged period of time. These channels also shape action potentials; blocking their activity increases the duration of action potentials that in turn depend solely on the inactivation of Na⁺ channels for repolarization. A-type channels, also referred to as Shaker-related channels (after the first K⁺ channel cloned from Drosophila), are transiently activated when a cell is depolarized after a period of hyperpolarization. They can function to decrease the frequency of action potential firing by prolonging the interspike interval.

Ca²⁺-gated K⁺ channels open in response to the binding of Ca²⁺, which usually occurs after a depolarization-induced influx of Ca²⁺. Generally, these channels can remain open for prolonged periods, such as a few seconds, and are responsible for several different types of long after-hyperpolarization (AHP), which is a hyperpolarization of the membrane that occurs after an action potential or series of action potentials. A long AHP can profoundly affect the firing pattern of a neuron; for instance, when a steady stimulus produces a train of action potentials, an AHP often is generated in response and can gradually slow the rate of firing. This phenomenon, known as spike-frequency adaptation, occurs throughout the nervous system and has many functional consequences.

SK channels (named for their small conductance) are a subtype of Ca²⁺-gated K⁺ channels that are important for AHPs; they have become attractive drug targets because of their roles in modulating neuronal firing. SK channels are blocked by apamin, a honeybee neurotoxin.

Inward rectifiers structurally resemble truncated Shaker-type channels that are missing S1–S4 segments; that is, they comprise only the putative pore-lining segments S5, S6, and SS1–SS2. As with the Shaker-type channels, four of these subunits are believed to multimerize to form a functional channel. Such rectifiers are described as anomalous because they open when the membrane is hyperpolarized and close as the membrane depolarizes. They most likely stabilize a membrane’s potential when the potential is near rest, without inhibiting depolarization. By conducting outward current in the voltage range just slightly positive to E_K, they maintain a resting potential near E_K; however, if the cell is sufficiently depolarized, they shut off, freeing the membrane to undergo further depolarization. This type of K⁺ channel is found in the heart, in striated muscle, and in neurons.

Seven subfamilies of inward rectifiers—K_ir1–7—have been identified. Channels formed by members of the K_ir3 subfamily are especially noteworthy because of the role that G proteins play in their regulation (Chapter 4). G protein–coupled K_ir3 channels, termed GIRKs, are critical to many physiologic functions; for example, they mediate the slowing of heart rate by acetylcholine. The binding of acetylcholine to muscarinic acetylcholine receptors in the atrium activates a G protein and liberates its βγ complex, which in turn leads to the activation of a GIRK. The K⁺ current produced by this activation slows the depolarization of sinoatrial cells to their firing threshold. This process explains the positive ionotropic and chronotropic effects of muscarinic cholinergic antagonists such as atropine. GIRKs are also expressed in the brain where, among other functions, they mediate a neuron’s electrical responses to the activation of many types of G protein–coupled neurotransmitter receptors, especially those coupled with G proteins from the G_i family (Chapter 4).

K(ATP) channels (K_ir6.2), another atypical member of the K_ir family, play an important role in cellular and systemic metabolic regulation. Functionally, these channels are voltage insensitive; they close in the presence of high (100 μM to 1 mM) intracellular ATP concentrations, and open when these concentrations are low, resulting in a more hyperpolarized membrane potential. They may serve a protective function in neurons by reducing neural firing when neural energy reserves are depleted. Structurally, each of these ATP-sensitive K⁺ channels is a multimeric complex comprising four K_ir6.2 subunits and one sulfonylurea receptor (SUR) protein. Sulfonylureas such as tolbutamide block K(ATP) channels and are used to treat adult-onset diabetes. The drugs act by increasing insulin secretion from pancreatic β cells. Mutations in K_ir6.2 channels cause congenital hyperinsulinemia 2–1. Iptakalim is a structurally novel opener of K(ATP) channels, which shows promise as a neuroprotective agent in animal models. Minoxidil, used to restore hair growth, is also an activator of K(ATP) channels, although whether this action is related to its effects on hair growth remains unknown.

Calcium Channels

Ca²⁺ is an important signaling molecule that is present in low concentrations in extracellular fluid (1–5 mM) and in minute concentrations in most cell interiors (approximately 0.1–0.2 μM). The opening of Ca²⁺ channels is the critical link between cell depolarization and Ca²⁺ entry, which can result in local intracellular Ca²⁺ concentrations as great as 100 μM. The subsequent binding of Ca²⁺ to intracellular molecules leads to many biologic responses, including muscle contraction; the triggering of neurotransmitter release from nerve terminals (Chapter 3); the activation of second messenger systems that cause many changes, including alterations in gene expression (Chapter 4); and, in extreme cases, neuronal self-destruction (Chapters 18 and 20). Some Ca²⁺ channels also impart electrical properties to the cells in which they are expressed; for example, such cells may show Ca²⁺ spikes—action potentials in which the depolarizing current is carried predominantly by Ca²⁺.

Ca²⁺ channels are categorized in terms of their voltage dependence, kinetics (speed of activation and inactivation), and pharmacology. Ten Ca²⁺ channel genes have been isolated, many of which correspond to Ca²⁺ channel types that were originally classified according to their electrophysiologic properties 2–4.

L channels

L-type (large current or “long open time”) Ca²⁺ channels are activated by large depolarizations (approximately −20 mV) and can remain open for a long time before inactivating (500 ms or more). Four major L-channel subunits have been cloned: Ca_V1.1–4. The Ca_V1.1 (α1S) subunit resides exclusively in skeletal muscle, where it is abundant in transverse tubules; Ca_V1.2 (α1C) subunits are found in cardiac muscle, smooth muscle, and brain; and Ca_V1.3 (α1D) subunits predominate in cochlear hair cells, and endocrine and kidney cells, but also reside in the brain. Ca_V1.4 (α1F) subunits are found only in the retina, with mutations in this gene linked to congenital stationary night blindness.

Clinically, these channels are targets for antianginal and antihypertensive drugs. L-type Ca²⁺ channel blockers, phenylalkylamines (eg, verapamil), dihydropyridines (eg, nifedipine), and benzothiazepines (eg, diltiazem), decrease myocardial contractile force and thereby reduce myocardial oxygen requirements, or reduce smooth muscle contractility and thereby decrease arterial and intraventricular pressure. L-type Ca²⁺ channels are located primarily on the cell bodies and proximal dendrites of neurons. They admit Ca²⁺ to the cell body during periods of strong depolarization, and this influx causes second messenger activation and changes in gene transcription. Yet substances that block these channels rarely have noticeable neurologic side effects, despite the fact that some—for example, nimodipine—can cross the blood–brain barrier. One possibility is that one of the neuronal L-type Ca²⁺ channels, Ca_V1.3 channels, are less sensitive to these typical L channel blockers.

T channels

T-type (tiny current or transient) Ca²⁺ channels are activated by depolarizations near resting potential; half-maximal activation of these channels occurs at approximately −40 mV. They exhibit voltage-dependent inactivation such that depolarization produced by their activation ultimately triggers their inactivation. Because of this self-limiting property, these channels are excellent oscillators; in fact, they are believed to provide a pacemaker current in thalamic neurons that generate the rhythmic cortical discharge associated with absence seizures (petit mal). Ethosuximide, which blocks T-type current, is an effective therapy for these seizures (Chapter 19). The cloning of T-type channel subunits Ca_V3.1–3 (α1G, H, I) should facilitate the development of drugs that interact more specifically with this important Ca²⁺ channel subtype.

P/Q, N, and R channels

The α subunits of these channels belong to one subfamily Ca_V2.1, 3 (α1A, B, and E). The Ca_V2 channels are high-voltage activated, not sensitive to L channel blockers, yet can be blocked with high affinity by peptide toxins derived from spiders and sea snails 2–4. This subfamily of Ca²⁺ channels is best known for their regulation of neurotransmitter and hormone release: Ca²⁺ flows into nerve terminals through these channels in a voltage-dependent manner that triggers this release (Chapter 3). P/Q (Ca_V2.1) and N (Ca_V2.2) channels contribute most of the Ca²⁺ influx that triggers neurotransmitter and hormone release. P/Q-type channels also control neurotransmitter release at neuromuscular junctions, whereas N-type channels control neurotransmitter release in sympathetic neurons. A synthetic peptide blocker of N-type Ca²⁺ channels (ziconotide), derived from a marine snail conotoxin, is under development for treating patients unresponsive to intrathecal opiates for relief of chronic pain (Chapter 11). Ca_V2.3 channels underlie a subpopulation of R (resistant)-type Ca²⁺ channels. The functions of Ca_V2.3 channels are less well defined, but may include neurotransmitter and hormone release and the generation of dendritic calcium transients.

Mutations of the human Ca_V2.1 (α1A) gene have been linked to familial hemiplegic migraine, spinocerebellar ataxia, and episodic ataxia. Although this gene is richly expressed in Purkinje cells, how the mutant P/Q channels cause the abnormalities seen in these disorders is still under study.

Cyclic Nucleotide–Regulated Channels

The family of cyclic nucleotide–regulated channels includes the CNG channels and the HCN channels. These channels are six TM segment channels, structurally similar to voltage-gated K⁺ channels. They are nonselective cation channels, in that monovalent ions (K⁺, Na⁺) can pass through with no discrimination, and divalent ions (Ca²⁺) can pass as well.

The activation of CNG channels is mediated by the direct binding of cGMP or cAMP to the C-terminal region of the channel protein. These channels are expressed in the cilia of olfactory neurons and in outer segments of rod and cone photoreceptors, where they transduce odorant signals and photons into electrical signals 2–6. Photoreceptor channels strongly discriminate between cGMP and cAMP, whereas the olfactory channels are almost equally sensitive to both ligands.

CNG channels are heterotetramers composed of homologous A subunits (CNGA1–4) and B subunits (CNGB1–3). Each subunit binds a single molecule of cAMP or cGMP. Although the binding of only one subunit can cause a channel to open, maximum activation of the channel requires the binding of all four subunits to cyclic nucleotide. Mutations in CNG genes are linked to retinitis pigmentosa and achromatopsia.

In mammals, the HCN family comprises four members (HCN1–4). They are expressed in neurons, cardiac pacemaking cells, and photoreceptors. In contrast to most other VGL channels, HCN channels open on hyperpolarization and close at positive potentials. The currents carried by these channels are also known as I_h, I_f, or I_q currents. cAMP and cGMP can bind to the C-terminal region of HCN channels and enhance channel activity by shifting the activation curve to more positive voltages. Because their activation causes the cell to depolarize, often beyond spike threshold, these channels help drive a repetitive cycle of rhythmic firing and, hence, are called “pacemaker channels.” Mutations in HCN4 are associated with sick sinus node disease 2–1.

Given the key role of HCN channels in cardiac pacemaking, these channels are promising pharmacologic targets for the development of drugs used in the treatment of cardiac arrhythmias and ischemic heart disease. For example, ivabradine, a heart rate–lowering agent, is a use-dependent blocker of HCN channels. Similar compounds include zetabradine and cilobradine. The effect of cAMP on these channels is exploited by several common drugs used to modify heart rate; for example, propranolol, a β-adrenergic antagonist, reduces cellular cAMP levels, which decreases heart rate in part by inhibiting I_h channels. Because the rhythmic activity of neurons is integral to functions such as arousal, maintenance of the sleep–wake cycle, and respiratory rate, these pacemaker channels are important targets for future drug development. Indeed, antagonists of these channels are currently undergoing evaluation for use in the treatment of a range of neuropsychiatric disorders.

TRP Channels

TRP channels are named after their role in Drosophila phototransduction. Mutations in the channel caused the response to light to be transient, and resulted in a decrease in the level of light-induced Ca²⁺ influx. To date, approximately 28 genes have been identified in mammals encoding TRP channels. They are divided into six subfamilies (TRPC, TRPV, TRPM, TRPP, TRPA, TRPML), all of which are six TM domain channels that lack the complete set of positively charged residues in S4 that act as voltage sensors. TRP channels are permeable to Ca²⁺ and Na⁺, and, in some cases, Mg²⁺. No single activation mechanism is shared by all TRP channels. In general, TRP channels can be described as Ca²⁺-permeable cation channels with polymodal activation mechanisms, including receptor activation (ie, certain G protein–coupled receptors and receptor tyrosine kinases activate phospholipases that can subsequently modulate TRP channel activity), ligand activation (ie, exogenous small organic molecules, endogenous lipids or products of lipid metabolism, purine nucleotides, inorganic ions), and direct activation (ie, change of temperature, mechanical stimuli, conformational coupling to inositol triphosphate [IP₃] receptors, channel phosphorylation). Because most TRP channels are responsive to multiple stimuli, they can integrate these stimuli, and couple them to downstream signal amplification cascades through elevation of intracellular Ca²⁺ and membrane depolarization.

TRP channels are of particular importance in sensory perception, including vision, taste, smell, hearing, mechanosensation, and thermosensation. As just one example, TRPV1 is activated by capsaicin, the active ingredient in hot pepper. TRP channels, particularly their involvement in pain, are discussed in greater detail in Chapter 11 (see 11–1). TRP channels also allow individual cells to sense changes in the local environment, such as alterations in fluid flow and mechanical stress. As cellular sensors, TRP channels are proposed to participate in various physiologic processes, such as regulating vessel tone, fertilization, Mg²⁺ absorption, and neurite outgrowth. Mutations in TRP channels are associated with diseases such as hypomagnesemia, polycystic kidney disease, and a neurodegenerative disease, mucolipidosis 2–1.

Chloride Channels

While this chapter has focused on cation channels, neurons and most other cells are also permeable to Cl⁻. Cl⁻ channels are critical for many physiologic processes, and mutations of these channels have been implicated in a variety of diseases 2–1.

Cl⁻ channels serve two primary functions. First, they dampen electrical excitability. In most cells, including neurons and myocytes, the intracellular concentration of Cl⁻ is close to, but below, its electrochemical equilibrium. In a manner similar to that of K⁺ channels, Cl⁻ channels provide a force—the concentration gradient of Cl⁻—that pulls the cell toward the equilibrium potential of Cl⁻, which is generally a hyperpolarized value. As a result, the inactivation of Cl⁻ channels can lead to hyperexcitability; for example, mutations in the Cl⁻ channel protein CLC-1 in skeletal muscle lead to muscle hyperexcitability, which results in myotonia. Similarly, the binding of strychnine to Cl⁻-conducting glycine receptors on, among other cells, motor neurons of the spinal cord can lead to violent convulsions and death (Chapter 5).

Cl⁻ channels also control the osmotic flow of water across the cell membrane. Some Cl⁻ channels that regulate osmotic balance are in fact activated by the swelling of a cell; activation of these channels allows Cl⁻ to exit the swollen cell, accompanied by cations and water. These channels play an important role in secretory cells, such as those of the mucosal epithelium, and in the kidney. Indeed, the so-called loop diuretics, such as furosemide, block Cl⁻ channels in the ascending loop of Henle and are widely used in the treatment of congestive heart failure.

As might be expected, the channels that permit Cl⁻ to cross plasma membranes differ in molecular composition from the voltage-gated cation channels. Cl⁻ channels comprise three molecular varieties 2–15. Ligand-gated Cl⁻ channels are among the most important in the brain; they act as signaling proteins for the inhibitory neurotransmitters γ-aminobutyric acid (GABA) and glycine and are discussed further in Chapter 5. CLC channels, which are believed to function as multimers of CLC proteins, are a family of homologous proteins characterized by 12 presumptive transmembrane domains (see 2–15). Nine CLCs have been cloned (CLC-1–7, CLC-Ka, and CLC-Kb); among them CLC-1, -2, -Ka, and -Kb reside on plasma membranes to control Cl⁻ flux and membrane potential, and CLC-3–5, possibly 6 and 7 as well, are located on the membrane of intracellular vesicles and are thought to be important in maintaining the pH of these vesicles. Human mutations in CLC channels are linked to several diseases, including several renal disorders, neurodegeneration, and possibly epilepsy.

Cystic fibrosis transmembrane conductance regulator (CFTR) channels are activated by the binding of ATP to two nucleotide-binding domains and by the phosphorylation of key serine residues on the regulatory domain (see 2–15; also see 2–5 for a discussion of ion channel phosphorylation). Like CLC channels, CFTRs are believed to possess 12 transmembrane domains; however, CFTR channels are distinguished from CLC channels by features such as their nucleotide-binding domains. The CFTR is among the most intensively studied ion channels because mutations of this channel cause cystic fibrosis, a relatively common hereditary disease that affects 1 in 3000 Caucasian newborns. In individuals with cystic fibrosis, the loss of CFTR Cl⁻ channels in several types of epithelial cells limits the egress of Cl⁻ ions into the lumen. Through mechanisms that are not completely understood, the disease also affects epithelial Na⁺ channels, whose malfunctioning causes the production of a thick, desiccated mucus. This abnormal secretion leads to obstruction of the biliary and pancreatic tracts and to a greatly increased incidence of pulmonary disease.

Astroctyes

Astrocytes constitute between 25% and 50% of the cellular volume of most brain regions. Although a subset of these cells is morphologically reminiscent of stars as their name implies, their form varies widely. In gray matter, the predominant form is the protoplasmic astrocyte, whereas in white matter the predominant form is the fibrous astrocyte. In addition to their shape, astrocytes can be identified by the unique intermediate filaments composed of glial fibrillary acidic protein (GFAP) that fill their processes. Accordingly, antibodies to GFAP robustly stain astrocytes but not neurons.

Partly because of their high concentration of filaments, astrocytes give structure to the brain. Their processes extend to the pia mater, one of the membranes that cover the brain, to the ependyma, which lines the brain’s ventricular system, and to the serosal surface of capillaries that penetrate the brain parenchyma. Moreover, astrocytes extend processes to the cell bodies of neurons and form sheaths around breaks in myelin (nodes of Ranvier) and around many synapses as well 2–16. Astrocytes also help to keep the brain isolated from the general circulation. They play a central role in building and maintaining the blood–brain barrier by inducing endothelial cells that line capillaries in the brain to form tight junctions between cells.

Astrocytes serve many functions in the brain. Radial glia, which are the embryonic precursors of astrocytes, provide a scaffolding, or path, for the migration of neurons—a critical event in the development of the mature architecture of the brain. Astrocytes also synthesize adhesion molecules and extracellular matrix proteins, including CAMs, fibronectin, and laminin. These proteins are involved in neuronal migration and may regulate axonal pathfinding and dendritic remodeling. As previously mentioned, astrocytes also elaborate neurotrophic factors and cytokines throughout life—molecules that are involved in the development and maintenance of both neurons and glia (Chapter 8).

Because astrocytes take up excess neurotransmitters and ions, including K⁺, that are a product of neurotransmitter release and receptor activation, they help maintain the extracellular milieu that surrounds neurons during synaptic transmission. Among transporters that promote the reuptake of excitatory amino acid neurotransmitters, the highest density is found on astrocytes. This is consistent with the anatomic organization of astrocytes in surrounding most excitatory synapses (see 2–1). Astrocytes also may play a heretofore unappreciated role in neuronal communication because they can be coupled to each other by means of gap junctions and can generate intracellular Ca²⁺ waves in response to activity in neurons. Moreover, some astrocytes express proteins normally considered neuronal, such as ion channels and neurotransmitter receptors. Although the physiologic significance of these properties remains unclear, they suggest that astrocytes can alter their function in response to neuronal activity in their vicinity. In fact, evidence is accumulating that astrocytes can influence synaptic function via the uptake of neurotransmitters such as glutamate and the activity-dependent release of a number of different factors.

Astrocytes also are involved in disease. After brain injury, inflammation, and some types of infection, and in response to neurodegenerative processes such as Alzheimer disease, astrocytes hypertrophy and may proliferate. Such reactive gliosis is accompanied by alterations in gene expression in the reactive glia and the elaboration of cytokines and other soluble factors. Some of these factors recruit additional glia to respond to injury, which under extreme circumstances can result in damage to large numbers of nearby neurons.

Oligodendrocytes and Schwann Cells

Oligodendrocytes in the CNS and Schwann cells in the peripheral nervous system produce and ensheathe axons with myelin. By forming a relatively thick layer of insulation, myelin enables rapid electrical conduction down axons by markedly diminishing the capacitance of the neuronal membrane and increasing its resistance. Whereas Schwann cells can produce only one myelin sheath, oligodendrocytes can produce several of them. This is accomplished by the oligodendrocyte membrane wrapping itself in multiple layers around the axon 2–16. As mentioned previously, the axons express a high level of voltage-gated Na⁺ channels at their nodes of Ranvier, small gaps in myelin between each myelinated segment, which permit the regeneration of an action potential.

Although the major chemical components of myelin in the CNS differ from those of myelin in the peripheral nervous system, both types contain hydrophobic proteolipids that function as insulators. Thus, the destruction of myelin sheaths can lead to marked deficits in nerve conduction velocity and a variety of sensory and motor deficits. Such deficits are characteristic of multiple sclerosis, a disease that involves the destruction of oligodendrocytes and the subsequent loss of myelin, presumably in response to an autoimmune mechanism (Chapter 12).

Microglia

Microglia are an important component of the defense mechanisms against infectious diseases of the CNS. Although the source of microglial cells remains controversial, it is believed that they are derived primarily from monocytic progenitor cells of bone marrow that enter the brain and undergo differentiation; thus, many microglia appear to be related to peripheral macrophages. Evidence suggests that some microglia also may be derived from glial lineages, although this pathway remains incompletely characterized. Like astrocytes, microglia respond to insult and injury with alterations in their morphology and number and thereby are important for tissue defense and repair. However, as with immune cells in other tissues, excessive activation of microglia can be deleterious and may contribute to inflammatory, and perhaps neurodegenerative, disorders of the CNS. Microglia are, for example, major mediators of the CNS effects of human immunodeficiency virus (HIV) infection.

Compared with other organs in the human body, the brain consumes a disproportionate share of energy and oxygen and consequently is richly vascularized. Its vascular bed is supplied by perforating arteries that branch into the subarachnoid space and subsequently enter the brain parenchyma. These arterioles in the brain’s gray matter, which is more densely vascularized than the white matter, give rise to a large number of capillary beds. These in turn drain into a venous network that terminates in large venous sinuses before returning blood to the general circulation.

The strict local control of cerebral circulation correlates well with neural activity. This correlation is the basis of current brain imaging techniques that utilize cerebral blood flow, such as positron emission tomography (PET) with ¹⁵O water, or that rely on blood oxygenation, such as magnetic resonance imaging (MRI) with endogenous blood oxygen level–dependent contrast (BOLD), as a surrogate to gauge neural activity. The chemical signals that couple neural activity to regulation of the vasculature remain an important matter of investigation.

Because presynaptic nerve terminals exhibit tremendous metabolic activity and contain a large fraction of the mitochondria in the brain, imaging signals related to blood flow or to the use of oxygen or glucose may reflect the functioning of nerve terminals rather than the activity of cell bodies or dendrites. This tendency greatly complicates the interpretation of brain imaging studies; for example, strong imaging signals that result from increased activity of inhibitory nerve terminals may reflect a decrease in the activity of postsynaptic cell bodies. This type of problem challenges scientists to better understand the biologic processes that underlie images of the living human brain and to develop markers that more closely approximate the firing of neurons.

Blood–Brain Barrier

Signaling in the nervous system depends on precise concentrations of extracellular ions and on the carefully controlled release, reuptake, and metabolism of neurotransmitters. If the brain were not isolated from the general circulation, changes in ion concentration related to diet or hydration status would routinely disrupt neural function in profound ways. Ingestion of common nutrients, such as amino acids, that also serve as neurotransmitters or their precursors, or the release of peripheral hormones, would wreak havoc on brain function.

These catastrophic occurrences are prevented by the existence of a blood–brain barrier that creates an isolated homeostatic milieu for the brain and spinal cord. This barrier is formed by tight junctions between endothelial cells of capillaries in cerebral vascular beds, which restrict the passage of soluble molecules 2–17. The endothelial cells line the lumen of the vessels and are surrounded by a continuous basal lamina. Potential space around the capillaries is taken up by pericytes that surround the capillary walls; these cells are involved in secretion of proteins that contribute to the basement membrane. Because the end feet (terminal processes) of perivascular astrocytes intercalate between the pericytes and contact the basal lamina, they are believed to be involved in the formation and maintenance of the tight junctions between the cerebral endothelial cells 2–18. In contrast, small spaces between endothelial cells in the general circulation permit blood solutes to enter the surrounding tissues.

Only highly lipophilic molecules can diffuse passively through the membranes of vascular cells to enter the brain; thus, any peripherally administered drug that targets the CNS must be relatively lipophilic. In contrast, hydrophilic molecules are generally excluded from the brain; only water-soluble molecules that are recognized by ATP-dependent transport processes, such as glucose, amino acids, and nucleosides, can enter the CNS. Some other substances gain access to the brain by means of receptor-mediated processes (eg, iron–transferrin, insulin, and lipoproteins; 2–7, 2–19). Many additional substances (eg, leptin) appear to penetrate the brain, although the underlying mechanisms remain poorly understood.

Some areas of the brain that surround the cerebral ventricles, including the area postrema of the medulla, the circumventricular organs, and regions of hypothalamus, lack a blood–brain barrier. These regions have fenestrated capillaries that allow peptides, hydrophilic molecules, and other solutes to enter the brain. In these periventricular regions, the brain is able to sample the general circulation, detect circulating hormones and cytokines, and subsequently direct adaptive responses to these signals. The area postrema, for example, contains chemoreceptors that trigger vomiting in response to circulating toxins.

In addition to the physical impediment provided by the blood–brain barrier, transmembrane pumps are found in the CNS that can rapidly transport drugs out of the CNS, even if the drugs have penetrated the blood–brain barrier. One such family of pumps, the P-glycoproteins, was first investigated as a cause of resistance to cancer chemotherapy agents; P-glycoproteins are expressed in cells at the blood–brain barrier.

Because the blood–brain barrier excludes all but small, highly lipophilic substances from the CNS, many potential pharmacotherapeutic agents cannot be used successfully. Accordingly, neuropharmacologic research continues to focus on developing strategies for penetrating the blood–brain barrier to improve drug delivery (see 2–7). Increasingly drugs can be screened to see whether they are substrates for P-glycoprotein pumps and thus unlikely to achieve therapeutic concentrations at neurons. There are also attempts to generate inhibitors of P-glycoproteins as a means of enhancing the actions of certain drugs in the CNS. Certain L-type Ca²⁺ channel blockers, for instance, inhibit P-glycoproteins. Use of such drugs could be beneficial if planned as a strategy to increase brain concentrations of a drug, but can cause toxicity if a patient inadvertently takes a penetrant CNS drug at regular doses that is normally removed from the CNS by P-glycoproteins.