Bruce R. Ransom
Extracellular fluid in the brain provides a highly regulated environment for central nervous system neurons
Everything that surrounds individual neurons can be considered part of the neuronal microenvironment. Technically, therefore, the neuronal microenvironment includes the extracellular fluid (ECF), capillaries, glial cells, and adjacent neurons. Although the term often is restricted to just the immediate ECF, the ECF cannot be meaningfully discussed in isolation because of its extensive interaction with brain capillaries, glial cells, and cerebrospinal fluid (CSF). How the microenvironment interacts with neurons and how the brain (used here synonymously with central nervous system, or CNS) stabilizes it to provide constancy for neuronal function are the subjects of this discussion.
The concentrations of solutes in brain extracellular fluid (BECF) fluctuate with neural activity, and conversely, changes in ECF composition can influence nerve cell behavior. Not surprisingly, therefore, the brain carefully controls the composition of this important compartment. It does so in three major ways. First, the brain uses the blood-brain barrier to protect the BECF from fluctuations in blood composition. Second, the CSF, which is synthesized by choroid plexus epithelial cells, strongly influences the composition of the BECF. Third, the surrounding glial cells “condition” the BECF.
The brain is physically and metabolically fragile
The ratio of brain weight to body weight in humans is the highest in the animal kingdom. The average adult brain weight is ~1400 g in men and ~1300 g in women, approximately the same weight as the liver (see Chapter 46). This large and vital structure, which has the consistency of thick pudding, is protected from mechanical injury by a surrounding layer of bone and by the CSF in which it floats.
The brain is also metabolically fragile. This fragility arises from its high rate of energy consumption, absence of significant stored fuel in the form of glycogen (~5% of the amount in the liver), and rapid development of cellular damage when ATP is depleted. However, the brain is not the greediest of the body’s organs; both the heart and kidney cortex have higher metabolic rates. Nevertheless, although it constitutes only 2% of the body by weight, the brain receives ~15% of resting blood flow and accounts for ~20% and 50%, respectively, of total resting oxygen and glucose utilization. The brain’s high metabolic demands arise from the need of its neurons to maintain the steep ion gradients on which neuronal excitability depends. In addition, neurons rapidly turn over their actin cytoskeleton. Neuroglial cells, the other major cells in the brain, also maintain steep transmembrane ion gradients. More than half of the energy consumed by the brain is directed to maintain ion gradients, primarily through operation of the Na-K pump (see Chapter 5). An interruption of the continuous supply of oxygen or glucose to the brain results in rapid depletion of energy stores and disruption of ion gradients. Because of falling ATP levels in the brain, consciousness is lost within 10 seconds of a blockade in cerebral blood flow. Irreversible nerve cell injury can occur after only 5 minutes of interrupted blood flow.
CEREBROSPINAL FLUID
CSF is a colorless, watery liquid. It fills the ventricles of the brain and forms a thin layer around the outside of the brain and spinal cord in the subarachnoid space. CSF is secreted within the brain by a highly vascularized epithelial structure called the choroid plexus and circulates to sites in the subarachnoid space where it enters the venous blood system. The composition of CSF is highly regulated, and because CSF is in slow diffusional equilibrium with BECF, it helps regulate the composition of BECF. The choroid plexus can be thought of as the brain’s “kidney” in that it stabilizes the composition of CSF, just as the kidney stabilizes the composition of blood plasma.
CSF fills the ventricles and subarachnoid space
The ventricles of the brain are four small compartments located within the brain (Fig. 11-1A). Each ventricle contains a choroid plexus and is filled with CSF. The ventricles are linked together by channels, or foramina, that allow CSF to move easily between them. The two lateral ventricles are the largest and are symmetrically located within the cerebral hemispheres. The choroid plexus of each lateral ventricle is located along the inner radius of this horseshoe-shaped structure (Fig. 11-1B). The two lateral ventricles each communicate with the third ventricle, which is located in the midline between the thalami, through the two interventricular foramina of Monro. The choroid plexus of the third ventricle lies along the ventricle roof. The third ventricle communicates with the fourth ventricle by the cerebral aqueduct of Sylvius. The fourth ventricle is the most caudal ventricle and is located in the brainstem. It is bounded by the cerebellum superiorly and by the pons and medulla inferiorly. The choroid plexus of the fourth ventricle lies along only a portion of this ventricle’s tent-shaped roof. The fourth ventricle is continuous with the central canal of the spinal cord. CSF escapes from the fourth ventricle and flows into the subarachnoid space through three foramina: the two laterally placed foramina of Luschka and the midline opening in the roof of the fourth ventricle, called the foramen of Magendie. We shall see later how CSF circulates throughout the subarachnoid space of the brain and spinal cord.
Figure 11-1 The brain ventricles and the cerebrospinal fluid. A, This is a transparent view, looking from the left side of the brain. The two lateral ventricles communicate with the third ventricle, which in turn communicates with the fourth ventricle. B, Each ventricle contains a choroid plexus, which secretes CSF. The CSF escapes from the fourth ventricle and into the subarachnoid space through the two lateral foramina of Luschka and the single foramen of Magendie.
The brain and spinal cord are covered by two membranous tissue layers called the leptomeninges, which are in turn surrounded by a third, tougher layer. The innermost of these three layers is the pia mater; the middle is the arachnoid mater (or arachnoid membrane); and the outermost layer is the dura mater (Fig. 11-2). Between the arachnoid mater and pia mater (i.e., the leptomeninges) is the subarachnoid space, which is filled with CSF that escaped from the fourth ventricle. The CSF in the subarachnoid space completely surrounds the brain and spinal cord. In adults, the subarachnoid space and the ventricles with which they are continuous contain ~150 mL of CSF, 30 mL in the ventricles and 120 mL in the subarachnoid spaces of the brain and spinal cord.
Figure 11-2 The meninges and ependymal cells. The figure represents a coronal section through the anterior portion of the brain. The upper inset shows the three layers of meninges: the dura mater, which here is split into two layers to accommodate the superior sagittal sinus (filled with venous blood); the arachnoid mater, which is formed by cells that are interconnected by tight junctions; and the pia mater, which closely adheres to a layer composed of astrocyte endfeet that are covered by a basement membrane (glia limitans). The lower inset shows ependymal cells lining the interior of the frontal horn of the left ventricle. Both the subarachnoid space and the cavities of the ventricles are filled with CSF.
The pia mater (Latin for “tender mother”) is a thin layer of connective tissue cells that is very closely applied to the surface of the brain and covers blood vessels as they plunge through the arachnoid into the brain. A nearly complete layer of astrocytic endfeet—the glia limitans—abuts the pia from the brain side and is separated from the pia by a basement membrane. The pia adheres so tightly to the associated glia limitans in some areas that they seem to be continuous with each other; this combined structure is sometimes called the pial-glial membrane or layer. This layer does not restrict diffusion of substances between the BECF and the CSF.
The arachnoid membrane (Greek for “cobweb-like”) is composed of layers of cells, resembling those that make up the pia, linked together by tight junctions. The arachnoid isolates the CSF in the subarachnoid space from blood in the overlying vessels of the dura mater. The cells that constitute the arachnoid and the pia are continuous in the trabeculae that span the subarachnoid space. These arachnoid and pial layers are relatively avascular; thus, the leptomeningeal cells that form them probably derive nutrition from the CSF that they enclose as well as from the ECF that surrounds them. The leptomeningeal cells can phagocytose foreign material in the subarachnoid space.
The dura mater is a thick, inelastic membrane that forms an outer protective envelope around the brain. The dura has two layers that split to form the intracranial venous sinuses. Blood vessels in the dura mater are outside the blood-brain barrier (see later), and substances could easily diffuse from dural capillaries into the nearby CSF if it were not for the blood-CSF barrier created by the arachnoid.
The brain floats in CSF, which acts as a shock absorber
An important function of CSF is to buffer the brain from mechanical injury. The CSF that surrounds the brain reduces the effective weight of the brain from ~1400 g to less than 50 g. This buoyancy is a consequence of the difference in the specific gravities of brain tissue (1.040) and CSF (1.007). The mechanical buffering that the CSF provides greatly diminishes the risk of acceleration-deceleration injuries in the same way that wearing a bicycle helmet reduces the risk of head injury. As you strike a tree, the foam insulation of the helmet gradually compresses and reduces the velocity of your head. Thus, the deceleration of your head is not nearly as severe as the deceleration of the outer shell of your helmet. The importance of this fluid suspension system is underscored by the consequences of reduced CSF pressure, which sometimes happens transiently after the diagnostic procedure of removal of CSF from the spinal subarachnoid space (see the box titled Lumbar Puncture). Patients with reduced CSF pressure experience severe pain when they try to sit up or to stand because the brain is no longer cushioned by shock-absorbing fluid and small gravity-induced movements put strain on pain-sensitive structures. Fortunately, the CSF leak that can result from lumbar puncture is only temporary; the puncture hole easily heals itself, with prompt resolution of all symptoms.
The choroid plexuses secrete CSF into the ventricles, and the arachnoid granulations absorb it
Most of the CSF is produced by the choroid plexuses, which are present in four locations (Fig. 11-1): the two lateral ventricles, the third ventricle, and the fourth ventricle. The capillaries within the brain appear to form a small amount of CSF. Total CSF production is ~500 mL/day. Therefore, the entire volume of CSF, ~150 mL, is replaced or “turns over” about three times each day.
Secretion of new CSF creates a slight pressure gradient, which drives the circulation of CSF from its ventricular sites of origin into the subarachnoid space through three openings in the fourth ventricle, as discussed earlier. CSF percolates throughout the subarachnoid space and is finally absorbed into venous blood in the superior sagittal sinus, which lies between the two cerebral hemispheres (Fig. 11-2). The sites of absorption are specialized evaginations of the arachnoid membrane into the venous sinus (Fig. 11-3A). These absorptive sites are called pacchionian granulations or simply arachnoid granulations when they are large (up to 1 cm in diameter) and arachnoid villi if their size is microscopic. These structures act as pressure-sensitive, one-way valves for bulk CSF clearance; CSF can cross into venous blood, but venous blood cannot enter CSF. The actual mechanism of CSF absorption may involve transcytosis (see Chapter 20), the formation of giant fluid-containing vacuoles that cross from the CSF side of the arachnoid epithelial cells to the blood side (Fig. 11-3A). CSF may also be absorbed into spinal veins from herniations of arachnoid cells into these venous structures. Net CSF movement into venous blood is promoted by the pressure of the CSF, which is higher than that of the venous blood. When intracranial pressure (equivalent to CSF pressure) exceeds ~70 mm H2O, absorption commences and increases with intracranial pressure (Fig. 11-3B). In contrast to CSF absorption, CSF formation is not sensitive to intracranial pressure. This arrangement helps stabilize intracranial pressure.
Figure 11-3 Absorption of CSF. A, Arachnoid villi—or the larger arachnoid granulations (not shown)—are specialized evaginations of the arachnoid membrane through the dura mater and into the lumen of the venous sinus. The absorption of CSF may involve transcytosis. Note that arachnoid villi and granulations serve as one-way valves; fluid cannot move from the vein to the subarachnoid space. B, The rate of CSF formation is virtually insensitive to changes in the pressure of the CSF. On the other hand, the absorption of CSF increases steeply at CSF pressures above ~70 mm H2O.
If intracranial pressure increases, CSF absorption selectively increases as well so that absorption exceeds formation (Fig. 11-3B). This response results in a lower CSF volume and a tendency to counteract the increased intracranial pressure. However, if absorption of CSF is impaired even at an initially normal intracranial pressure, CSF volume increases and causes an increase in intracranial pressure. Such an increase in intracranial pressure can lead to a disturbance in brain function.
Lumbar Puncture
One of the most important diagnostic tests in neurology is the sampling of CSF by lumbar puncture. Critical information about the composition of CSF and about intracranial pressure can be obtained from this procedure. The anatomist Vesalius noted in 1543 that the ventricles are filled with a clear fluid, but the diagnostic technique of placing a needle into the lumbar subarachnoid space to obtain CSF was not introduced until 1891 by the neurologist Heinrich Quincke. The method of lumbar puncture is dictated by spine anatomy. In adults, the spinal cord ends at the interspace between L1 and L2 (see Chapter 10). A hollow needle for sampling of CSF can be safely inserted into the subarachnoid space at the level of the L3-L4 interspace, well below the end of the spinal cord. (See Note: Heinrich Quincke)
Once the needle is in the subarachnoid space, the physician attaches it to a manometer to measure pressure. With the patient lying on the side, normal pressure varies from 100 to 180 mm H2O, or 7 to 13 mm Hg. With the subject in this position and in the absence of a block to the free circulation of CSF, lumbar CSF pressure roughly corresponds to intracranial pressure. The physician can demonstrate direct communication of the pressure in the intracranial compartment to the lumbar subarachnoid space by gently compressing the external jugular veins in the neck for 10 seconds. This maneuver, called the Queckenstedt test, rapidly increases intracranial pressure because it increases the volume of intracranial venous blood. It quickly leads to an increase in lumbar pressure, which just as rapidly dissipates when the jugular pressure is removed.
CSF pressure can become elevated because of a pathological mass within the cranium, such as a tumor or collection of blood, or because the brain is swollen as a result of injury or infection (see the later box titled Cerebral Edema). If a “mass lesion” (i.e., any pathological process that occupies intracranial space) is large or critically placed, it can displace the brain and cause interference with the free circulation of CSF. For example, an expanding mass in the cerebellum can force the inferior part of the cerebellum into the foramen magnum and block flow of CSF into the spinal subarachnoid space. Under these conditions, performance of lumbar puncture can precipitate a neurological catastrophe. If a needle is placed in the lumbar subarachnoid space and fluid is removed for diagnostic examination or leaks out after the needle is removed, the ensuing decrease in pressure in the lumbar space creates a pressure gradient across the foramen magnum and potentially forces the brain down into the spinal canal. This disaster is called herniation. For this reason, a computed tomographic scan or magnetic resonance image of the head is usually obtained before a lumbar puncture is attempted; the imaging study can rule out the possibility of a large intracranial lesion that might raise intracranial pressure and increase the risk of herniation when the subarachnoid space is punctured and CSF withdrawn. The Queckenstedt test must also be avoided when an intracranial mass is suspected because it could enhance the pressure gradient and hasten herniation.
The epithelial cells of the choroid plexus secrete the CSF
Each of the four choroid plexuses is formed during embryological development by invagination of the tela choroidea into the ventricular cavity (Fig. 11-4). The tela choroidea consists of a layer of ependymal cells covered by the pia mater and its associated blood vessels. The choroid epithelial cells (Fig. 11-4, first inset) are specialized ependymal cells and therefore contiguous with the ependymal lining of the ventricles at the margins of the choroid plexus. Choroid epithelial cells are cuboidal and have an apical border with microvilli and cilia that project into the ventricle (i.e., into the CSF). The plexus receives its blood supply from the anterior and posterior choroidal arteries; blood flow to the plexuses—per unit mass of tissue—is ~10-fold greater than the average cerebral blood flow. Sympathetic and parasympathetic nerves innervate each plexus, and sympathetic input appears to inhibit CSF formation. A high density of relatively leaky capillaries is present within each plexus; as discussed later, these capillaries are outside the blood-brain barrier. The choroid epithelial cells are bound to one another by tight junctions that completely encircle each cell, an arrangement that makes the epithelium an effective barrier to free diffusion. Thus, although the choroid capillaries are outside the blood-brain barrier, the choroid epithelium insulates the ECF around these capillaries (which has a composition more similar to that of arterial blood) from the CSF. Moreover, the thin neck that connects the choroid plexus to the rest of the brain isolates the ECF near the leaky choroidal capillaries from the highly protected BECF in the rest of the brain.
Figure 11-4 Secretion of CSF by the choroid plexus. The top panel shows the location of the choroid plexuses in the two lateral ventricles and the third ventricle. The middle panel shows the organization of a single fold of choroidal epithelial cells, with the basolateral membranes of the epithelial cells overlying capillaries and the apical membranes facing the CSF. The bottom panel shows a single choroid epithelial cell and several of the transporters and channels that are believed to play a role in the isosmotic secretion of CSF. CA, carbonic anhydrase.
Normal-Pressure Hydrocephalus
Impaired CSF absorption is one mechanism proposed to explain a clinical form of ventricular enlargement called normal-pressure hydrocephalus. This condition is somewhat misnamed because the intracranial pressure is often intermittently elevated. Damage to the arachnoid villi can occur most commonly from infection or inflammation of the meninges or from the presence of an irritating substance, such as blood in the CSF after a subarachnoid hemorrhage. A spinal tap reveals normal pressure readings, but computed tomography or magnetic resonance imaging of the head shows enlargement of all four ventricles. Patients with normal-pressure hydrocephalus typically have progressive dementia, urinary incontinence, and gait disturbance, probably caused by stretching of axon pathways that course around the enlarged ventricles. A flexible plastic tube can be placed in one of the lateral ventricles to shunt CSF to venous blood or to the peritoneal cavity, thereby reducing CSF pressure. This procedure may reduce ventricular size and decrease neurological symptoms. The “shunting” procedure is also used for patients with obstructive hydrocephalus. In this condition, CSF outflow from the ventricles is blocked, typically at the aqueduct of Sylvius.
The composition of CSF differs considerably from that of plasma; thus, CSF is not just an ultrafiltrate of plasma (Table 11-1). For example, CSF has lower concentrations of K+ and amino acids than plasma does, and it contains almost no protein. Moreover, the choroid plexuses rigidly maintain the concentration of ions in CSF in the face of large swings in ion concentration in plasma. This ion homeostasis includes K+, H+/HCO−3, Mg2+, Ca2+, and, to a lesser extent, Na+ and Cl−. All these ions can affect neural function, hence the need for tight homeostatic control. The neuronal microenvironment is so well protected from the blood by the choroid plexuses and the rest of the blood-brain barrier that essential micronutrients, such as vitamins and trace elements that are needed in very small amounts, must be selectively transported into the brain. Some of these micronutrients are transported into the brain primarily by the choroid plexus and others primarily by the endothelial cells of the blood vessels. In comparison, the brain continuously metabolizes relatively large amounts of “macronutrients,” such as glucose and some amino acids.
Table 11-1 Composition of Cerebrospinal Fluid
CSF forms in two sequential stages. First, ultrafiltration of plasma occurs across the fenestrated capillary wall (see Chapter 20) into the ECF beneath the basolateral membrane of the choroid epithelial cell. Second, choroid epithelial cells secrete fluid into the ventricle. CSF production occurs with a net transfer of NaCl and NaHCO3 that drives water movement isosmotically (Fig. 11-4, large transepithelial arrow in the right inset). The renal proximal tubule (see Chapter 35) and small intestine (see Chapter 5) also perform near-isosmotic transport, but in the direction of absorption rather than secretion. In addition, the choroid plexus conditions CSF by absorbing K+ (Fig. 11-4, small transepithelial arrow in the right inset) and certain other substances (e.g., a metabolite of serotonin, 5-hydroxyindoleacetic acid).
The upper portion of the right inset of Figure 11-4 summarizes the ion transport processes that mediate CSF secretion. The net secretion of Na+ from plasma to CSF is a two-step process. The Na-K pump in the choroid plexus, unlike in other epithelia (see Chapter 5), is unusual in being located on the apical membrane, where it moves Na+ out of the cell into the CSF—the first step. This active movement of Na+out of the cell generates an inward Na+gradient across the basolateral membrane, energizing basolateral Na+ entry—the second step—through Na-H exchange and Na+-coupled HCO−3 transport. In the case of Na-H exchange, the limiting factor is the availability of intracellular H+, which carbonic anhydrase generates, along with HCO−3, from CO2 and H2O. Thus, blocking of the Na-K pump with ouabain halts CSF formation, whereas blocking of carbonic anhydrase with acetazolamide slows CSF formation.
The net secretion of Cl−, like that of Na+, is a two-step process. The first step is the intracellular accumulation of Cl− by the basolateral Cl-HCO3 exchanger. Note that the net effect of parallel Cl-HCO3exchange and Na-H exchange is NaCl uptake. The second step is efflux of Cl− across the apical border into the CSF through either a Cl− channel or a K/Cl cotransporter.
HCO−3 secretion into CSF is important for neutralizing acid produced by CNS cells. At the basolateral membrane, the epithelial cell probably takes up HCO−3 directly from the plasma filtrate through electroneutral Na/HCO3cotransporters (see Fig. 5-11F) and the Na+-driven Cl-HCO3 exchanger (see Fig. 5-13C). As noted before, HCO−3 can also accumulate inside the cell after CO2 entry. The apical step, movement of intracellular HCO−3 into the CSF, probably occurs by an electrogenic Na/HCO3 cotransporter (see Fig. 5-11D) and Cl− channels (which are generally permeable to HCO−3).
The lower portion of the right inset of Figure 11-4 summarizes K+ absorption from the CSF. The epithelial cell takes up K+ by the Na-K pump and the Na/K/Cl cotransporter at the apical membrane (see Fig. 5-11G). Most of the K+recycles back to the CSF, but a small amount exits across the basolateral membrane and enters the blood. The concentration of K+ in freshly secreted CSF is ~3.3 mM. Even with very large changes in plasma [K+], the [K+] in CSF changes very little. The value of [K+] in CSF is significantly lower in the subarachnoid space than in choroid secretions, which suggests that brain capillary endothelial cells remove extracellular K+ from the brain.
Water transport across the choroid epithelium is driven by a small osmotic gradient favoring CSF formation. This water movement is facilitated by expression of the water channel aquaporin 1 on both the apical and basal membranes as in renal proximal tubule (see Chapter 35).
BRAIN EXTRACELLULAR SPACE
Neurons, glia, and capillaries are packed tightly together in the CNS
The average width of the space between brain cells is ~20 nm, which is about three orders of magnitude smaller than the diameter of either a neuron or glial cell body (Fig. 11-5). However, because the surface membranes of neurons and glial cells are highly folded (i.e., have a large surface-to-volume ratio), the BECF in toto has a sizable volume fraction, ~20%, of the total brain volume. The fraction of the brain that is occupied by BECF varies somewhat in different areas of the CNS. Moreover, because brain cells can increase volume rapidly during intense neural activity, the BECF fraction can reversibly decrease within seconds from ~20% to ~17% of brain volume.
Figure 11-5 Tight packing of neurons and astrocytes. This is an electron micrograph of a section of the spinal cord from an adult rat showing the intermingling and close apposition of neurons and glial cells, mainly astrocytes. Neurons and glial cells are separated by narrow clefts that are ~20 nm wide and not visible at this magnification. The BECF in this space creates a tortuous path for the extracellular diffusion of solutes. Astrocyte processes are colored. As, astrocytes; At, en passant synapses; Ax, unmyelinated axons, Ax1 and Ax2, myelinated axons; Den, dendrites; f, astrocytic fibrils; nf, neurofilaments; S, synapses; SR, smooth endoplasmic reticulum. (Modified from Peters A, Palay SL, Webster H: The Fine Structure of the Nervous System. Philadelphia: WB Saunders, 1976.)
Even though the space between brain cells is extremely small, diffusion of ions and other solutes within this thin BECF space is reasonably high. However, a particle that diffuses through the BECF from one side of a neuron to the other must take a circuitous route that is described by a parameter called tortuosity. For a normal width of the cell-to-cell spacing, this tortuosity reduces the rate of diffusion by ~60% compared with movement in free solution. Decreases in cell-to-cell spacing can further slow diffusion. For example, brain cells, especially glial cells, swell under certain pathological conditions and sometimes with intense neural activity. Cell swelling is associated with a reduction in BECF because water moves from the BECF into cells. The intense cell swelling associated with acute anoxia, for example, can reduce BECF volume from ~20% to ~5% of total brain volume. By definition, this reduced extracellular volume translates to reduced cell-to-cell spacing, further slowing the extracellular movement of solutes between the blood and brain cells (see the box titled Cerebral Edema).
The BECF is the route by which important molecules such as oxygen, glucose, and amino acids reach brain cells and by which the products of metabolism, including CO2 and catabolized neurotransmitters, leave the brain. The BECF also permits molecules that are released by brain cells to diffuse to adjacent cells. Neurotransmitter molecules released at synaptic sites, for example, can spill over from the synaptic cleft and contact nearby glial cells and neurons, in addition to their target postsynaptic cell. Glial cells express neurotransmitter receptors, and neurons have extrajunctional receptors; therefore, these cells are capable of receiving “messages” sent through the BECF. Numerous trophic molecules secreted by brain cells diffuse in the BECF to their targets. Intercellular communication by way of the BECF is especially well suited for the transmission of tonic signals that are ideal for longer term modulation of the behavior of aggregates of neurons and glial cells. The chronic presence of variable amounts of neurotransmitters in the BECF supports this idea.
The CSF communicates freely with the BECF, thereby stabilizing the composition of the neuronal microenvironment
CSF in the ventricles and the subarachnoid space can exchange freely with BECF across two borders, the pia mater and ependymal cells. The pial-glial membrane (Fig. 11-2, upper inset) has paracellular gaps (see Chapter 2) through which substances can equilibrate between the subarachnoid space and BECF. Ependymal cells (Fig. 11-2, lower inset) are special glial cells that line the walls of the ventricles and form the cellular boundary between the CSF and the BECF. These cells form gap junctions between themselves that mediate intercellular communication, but they do not create a tight epithelium (see Chapter 5). Thus, macromolecules and ions can also easily pass through this cellular layer through paracellular openings (some notable exceptions to this rule are considered later) and equilibrate between the CSF in the ventricle and the BECF.
Cerebral Edema
Almost any type of insult to the brain causes cell swelling. This swelling is frequently accompanied by a net accumulation of water within the brain that is referred to as cerebral edema. Cell swelling in the absence of net water accumulation in the brain does not constitute cerebral edema. For example, intense neural activity causes a rapid shift of fluid from the BECF to the intracellular space, with no net change in brain water content. In cerebral edema, the extra water comes from the blood, as shown in Figure 11-6.
Figure 11-6 Cerebral edema. A, In cerebral edema, the brain fluid that accumulates comes from the vascular compartment. Cell swelling due to the mere shift of fluid from the extracellular to the intracellular fluid is not cerebral edema. B, Although small increases in intracranial volume have little effect on pressure, additional increases in volume cause potentially life-threatening increases in pressure. Note that compliance (i.e., ΔV/ΔP) falls at increasing volumes.
The mechanisms by which glial cells and neurons swell are not completely understood. Neuron cell bodies and dendrites, but not axons, swell when they are exposed to high concentrations of the neurotransmitter glutamate. This transmitter, along with others, is released to the BECF in an uncontrolled fashion with brain injury. Activation of ionotropic glutamate receptors (see Chapter 13) allows Na+ to enter neurons, and water and Cl− follow passively. Glial cells, both astrocytes and oligodendrocytes, swell vigorously under pathological conditions. One mechanism of glial swelling is an increase in [K+]o, which is a common ionic disturbance in a variety of brain pathological processes. This elevated [K+]o causes a net uptake of K+, accompanied by the passive influx of Cl− and water.
Cerebral edema can be life-threatening when it is severe. The problem is a mechanical one. The skull is an inelastic container housing three relatively noncompressible substances: brain, CSF, and blood. A significant increase in the volume of CSF, blood, or brain rapidly causes increased pressure within the skull (Fig. 11-3). If the cerebral edema is generalized, it can be tolerated until intracerebral pressure exceeds arterial blood pressure, at which point blood flow to the brain stops, with disastrous consequences. Fortunately, sensors in the medulla detect the increased intracerebral pressure and can partially compensate (Cushing reflex), to a point, by increasing arterial pressure (see Chapter 24). Focal cerebral edema (i.e., edema involving an isolated portion of the brain) causes problems by displacing nearby brain tissue. This abnormality may result in distortion of normal anatomical relationships, with selective pressure on critical structures such as the brainstem.
Clinical evidence of cerebral edema results directly from the increased intracranial pressure and includes headache, vomiting, altered consciousness, and focal neurological problems such as stretching and dysfunction of the sixth cranial nerve.
Hyperventilation is the most effective means of combating the acute increase in intracranial pressure associated with severe cerebral edema. Hyperventilation causes a prompt respiratory alkalosis (see Chapter 28) that is rapidly translated to an increase in the pH surrounding vascular smooth muscle, thereby triggering vasoconstriction and reduced cerebral blood flow (see Chapter 24). Thus, total intracranial blood content falls, with a rapid subsequent drop in intracranial pressure. Alternatively, the brain can be partially dehydrated by adding osmoles to the blood in the form of intravenously administered mannitol (see Chapter 5).
Because CSF and BECF can readily exchange with one another, it is not surprising that they have a similar chemical composition. For example, [K+] is ~3.3 mM in freshly secreted CSF and ~3 mM in both the CSF of the subarachnoid space (Table 11-1) and BECF. The [K+] of blood is ~4.5 mM. However, because of the extent and vast complexity of the extracellular space, changes in the composition of CSF are reflected slowly in the BECF and probably incompletely.
CSF is an efficient waste management system because of its high rate of production, its circulation over the surface of the brain, and the free exchange between CSF and BECF. Products of metabolism and other substances released by cells, perhaps for signaling purposes, can diffuse into the chemically stable CSF and ultimately be removed on a continuous basis either by bulk resorption into the venous sinuses or by active transport across the choroid plexus into the blood. For example, choroid plexus actively absorbs the breakdown products of the neurotransmitters serotonin (i.e., 5-hydroxyindoleacetic acid) and dopamine (i.e., homovanillic acid).
The ion fluxes that accompany neural activity cause large changes in extracellular ion concentration
As discussed in Chapter 7, ionic currents through cell membranes underlie the synaptic and action potentials by which neurons communicate. These currents lead to changes in the ion concentrations of the BECF. It is estimated that even a single action potential can transiently lower [Na+]o by ~0.75 mM and increase [K+]o by a similar amount. Repetitive neuronal activity causes larger perturbations in these extracellular ion concentrations. Because ambient [K+]ois much lower than [Na+]o, activity-induced changes in [K+]o are proportionately larger and are of special interest because of the important effect that [K+]o has on membrane potential (Vm). For example, K+ accumulation in the vicinity of active neurons depolarizes nearby glial cells. In this way, neurons signal to glial cells the pattern and extent of their activity. Even small changes in [K+]o can alter metabolism and ionic transport in glial cells and may be used for signaling. Changes in the extracellular concentrations of certain common amino acids, such as glutamate and glycine, can also affect neuronal Vm and synaptic function by acting at specific receptor sites. If the nervous system is to function reliably, its signaling elements must have a regulated environment. Glial cells and neurons both function to prevent excessive extracellular accumulation of K+ and neurotransmitters.
THE BLOOD-BRAIN BARRIER
The blood-brain barrier prevents some blood constituents from entering the brain extracellular space
The unique protective mechanism called the blood-brain barrier was first demonstrated by Ehrlich in 1885. He injected aniline dyes intravenously and discovered that the soft tissues of the body, except for the brain, were uniformly stained. Even though aniline dyes, such as trypan blue, extensively bind to serum albumin, the dye-albumin complex passes across capillaries in most areas of the body, but not the brain. This ability to exclude certain substances from crossing CNS blood vessels into the brain tissue is due to the blood-brain barrier. We now recognize that a blood-brain barrier is present in all vertebrates and many invertebrates as well.
The need for a blood-brain barrier can be understood by considering that blood is not a suitable environment for neurons. Blood is a complex medium that contains a large variety of solutes, some of which can vary greatly in concentration, depending on factors such as diet, metabolism, illness, and age. For example, the concentration of many amino acids increases significantly after a protein-rich meal. Some of these amino acids act as neurotransmitters within the brain, and if these molecules could move freely from the blood into the neuronal microenvironment, they would nonselectively activate receptors and disturb normal neurotransmission. Similarly, strenuous exercise can increase plasma concentrations of K+ and H+ substantially. If these ionic changes were communicated directly to the microenvironment of neurons, they could disrupt ongoing neural activity. Running a foot race might temporarily lower your IQ. Increases in [K+]o would depolarize neurons and thus increase their likelihood of firing and releasing transmitter. H+ can nonspecifically modulate neuronal excitability and influence the action of certain neurotransmitters. A broad range of blood constituents—including hormones, other ions, and inflammatory mediators such as cytokines—can influence the behavior of neurons or glial cells, which can express receptors for these molecules. For the brain to function efficiently, it must be spared such influences.
The choroid plexus and several restricted areas of the brain lack a blood-brain barrier; that is, they are supplied by leaky capillaries. Intra-arterially injected dyes can pass into the brain extracellular space at these sites through gaps between endothelial cells. The BECF in the vicinity of these leaky capillaries is similar to blood plasma more than to normal BECF. The small brain areas that lack a blood-brain barrier are called the circumventricular organsbecause they surround the ventricular system; these areas include the area postrema, posterior pituitary, median eminence, organum vasculosum laminae terminalis, subfornical organ, subcommissural organ, and pineal gland (Fig. 11-7). The ependymal cells that overlie the leaky capillaries in some of these regions (e.g., the choroid plexus) are linked together by tight junctions that form a barrier between the local BECF and the CSF, which must be insulated from the variability of blood composition. Whereas dyes with molecular weights up to 5000 can normally pass from CSF across the ependymal cell layer into the BECF, they do not pass across the specialized ependymal layer at the median eminence, area postrema, and infundibular recess. At these points, the localized BECF-CSF barrier is similar to the one in the choroid plexus. These specialized ependymal cells often have long processes that extend to capillaries within the portal circulation of the pituitary. Although the function of these cells is not known, it has been suggested that they may form a special route for neurohumoral signaling; molecules secreted by hypothalamic cells into the third ventricle could be taken up by these cells and transmitted to the general circulation or to cells in the pituitary.
Figure 11-7 Leaky regions of the blood-brain barrier: the circumventricular organs. The capillaries of the brain are leaky in several areas: the area postrema, the posterior pituitary, the subfornical organ, the median eminence, the pineal gland, and the organum vasculosum laminae terminalis (OVLT). In these regions, the neurons are directly exposed to the solutes of the blood plasma. A midline sagittal section is shown.
Neurons within the circumventricular organs are directly exposed to blood solutes and macromolecules; this arrangement is believed to be part of a neuroendocrine control system for maintaining such parameters as osmolality (see Chapter 40) and appropriate hormone levels, among other things. Humoral signals are integrated by connections of circumventricular organ neurons to endocrine, autonomic, and behavioral centers within the CNS. In the median eminence, neurons discharge “releasing hormones,” which diffuse into leaky capillaries for carriage through the pituitary portal system to the anterior pituitary. The lack of a blood-brain barrier in the posterior pituitary is necessary to allow hormones that are released there to enter the general circulation (see Chapter 47). In the organum vasculosum laminae terminalis, leakiness is important in the action of cytokines from the periphery, which act as signals to temperature control centers that are involved in fever (see Chapter 59).
Continuous tight junctions link brain capillary endothelial cells
The blood-brain barrier should be thought of as a physical barrier to diffusion from blood to brain ECF and as a selective set of regulatory transport mechanisms that determine how certain organic solutes move between the blood and brain. Thus, the blood-brain barrier contributes to stabilization and protection of the neuronal microenvironment by facilitating the entry of needed substances, removing waste metabolites, and excluding toxic or disruptive substances.
The structure of brain capillaries differs from that of capillaries in other organs. Capillaries from other organs generally have small, simple openings—or clefts—between their endothelial cells (Fig. 11-8A). In some of these other organs, windows, or fenestrae, provide a pathway that bypasses the cytoplasm of capillary endothelial cells. Thus, in most capillaries outside the CNS, solutes can easily diffuse through the clefts and fenestrae. The physical barrier to solute diffusion in brain capillaries (Fig. 11-8B) is provided by the capillary endothelial cells, which are fused to each other by continuous tight junctions (or zonula occludens; see Chapter 2). The tight junctions prevent water-soluble ions and molecules from passing from the blood into the brain through the paracellular route. Not surprisingly, the electrical resistance of the cerebral capillaries is 100 to 200 times higher than that of most other systemic capillaries.
Figure 11-8 The blood-brain barrier function of brain capillaries. A, Capillaries from most other organs often have interendothelial clefts or fenestrae, which makes them relatively leaky. B, Brain capillaries are not leaky and have reduced transcytosis. C, Continuous tight junctions connect the endothelial cells in the brain, making the capillaries relatively tight.
Elsewhere in the systemic circulation, molecules may traverse the endothelial cell by the process of transcytosis (see Chapter 20). In cerebral capillaries, transcytosis is uncommon, and brain endothelial cells have fewer endocytic vesicles than do systemic capillaries. However, brain endothelial cells have many more mitochondria than systemic endothelial cells do, which may reflect the high metabolic demands imposed on brain endothelial cells by active transport.
Other interesting features of brain capillaries are the thick basement membrane that underlies the endothelial cells, the presence of occasional pericytes within the basement membrane sheath, and the astrocytic endfeet (or processes) that provide a nearly continuous covering of the capillaries and other blood vessels. Astrocytes may play a crucial role in forming tight junctions between endothelial cells; experiments have shown that these glial cells can induce the formation of tight junctions between endothelial cells derived from capillaries outside the CNS. The close apposition of the astrocyte endfoot to the capillary also could facilitate transport of substances between these cells and blood.
Uncharged and lipid-soluble molecules more readily pass the blood-brain barrier
The capacity of the brain capillaries to exclude large molecules is strongly related to the molecular mass of the molecule and its hydrated diameter (Table 11-2). With a mass of 61 kDa, prealbumin is 14 times as concentrated in blood as in CSF (essentially equivalent to BECF for purposes of this comparison), whereas fibrinogen, which has a molecular mass of 340 kDa, is ~5000 times more concentrated in blood than in CSF. Diffusion of a solute is also generally limited by ionization at physiological pH, by low lipid solubility, and by binding to plasma proteins. For example, gases such as CO2 and O2 and drugs such as ethanol, caffeine, nicotine, heroin, and methadone readily cross the blood-brain barrier. However, ions such as K+ or Mg2+ and protein-bound metabolites such as bilirubin have restricted access to the brain. Finally, the blood-brain barrier is permeable to water because of the presence of water channels in the endothelial cells. Thus, water moves across the blood-brain barrier in response to changes in plasma osmolarity. When dehydration raises the osmolality of blood plasma (see the box titled Disorders of Extracellular Osmolality in Chapter 5), the increased osmolality of the CSF and BECF can affect the behavior of brain cells.
Table 11-2 Comparison of Proteins in Blood Plasma versus Cerebrospinal Fluid
Cerebral capillaries also express enzymes that can affect the movement of substances from blood to brain and vice versa. Peptidases, acid hydrolases, monoamine oxidase, and other enzymes are present in CNS endothelial cells and can degrade a range of biologically active molecules, including enkephalins, substance P, proteins, and norepinephrine. Orally administered dopamine is not an effective treatment of Parkinson disease (see Chapter 13), a condition in which CNS dopamine is depleted, because dopamine is rapidly broken down by monoamine oxidase in the capillaries. Fortunately, the dopamine precursor compound L-dopa is effective for this condition. Neutral amino acid transporters in capillary endothelial cells move L-dopa to the BECF, where presynaptic terminals take up the L-dopa and convert it to dopamine in a reaction that is catalyzed by dopa decarboxylase.
Transport by capillary endothelial cells contributes to the blood-brain barrier
Two classes of substances can pass readily between blood and brain. The first consists of the small, highly lipid soluble molecules discussed in the preceding section. The second group consists of water-soluble compounds—either critical nutrients entering or metabolites exiting the brain—that traverse the blood-brain barrier by specific transporters. Examples include glucose, several amino acids and neurotransmitters, nucleic acid precursors, and several organic acids. Two major transporter groups provide these functions: the SLC superfamily and ABC transporters (see Chapter 5). As is the case for other epithelial cells, capillary endothelial cells selectively express these and other membrane proteins on either the luminal or basal surface. (See Note: System L Amino Acid Transporters in Brain Capillary Endothelial Cells)
Although the choroid plexuses secrete most of the CSF, brain endothelial cells produce some interstitial fluid with a composition similar to that of CSF. Transporters such as those shown in Figure 11-8C are responsible for this CSF-like secretion as well as for the local control of [K+] and pH in the BECF.
GLIAL CELLS
Glial cells constitute half the volume of the brain and outnumber neurons
The three major types of glial cells in the CNS are astrocytes, oligodendrocytes, and microglial cells (Table 11-3). As discussed in Chapter 10, the peripheral nervous system (PNS) contains other, distinctive types of glial cells, including satellite cells, Schwann cells, and enteric glia. Glial cells represent about half the volume of the brain and are more numerous than neurons. Unlike neurons, which have little capacity to replace themselves when lost, neuroglial (or simply glial) cells can proliferate throughout life. An injury to the nervous system is the usual stimulus for proliferation.
Table 11-3 Glial Types
Historically, glial cells were viewed as a type of CNS connective tissue whose main function was to provide support for the true functional cells of the brain, the neurons. This firmly entrenched concept remained virtually unquestioned for the better part of a century after the early description of these cells by Virchow in 1858. Knowledge about glial cells has accumulated slowly because these cells have proved far more difficult to study than neurons. Because glial cells do not exhibit easily recorded action potentials or synaptic potentials, these cells were sometimes referred to as silent cells. However, glial cells are now recognized as intimate partners with neurons in virtually every function of the brain.
Astrocytes supply fuel to neurons in the form of lactic acid
Astrocytes have great numbers of extremely elaborate processes that closely approach both blood vessels and neurons. This arrangement led to the idea that astrocytes transport substances between the blood and neurons. This notion may be true, but it has not been proved. Throughout the brain, astrocytes envelop neurons, and both cells bathe in a common BECF. Therefore, astrocytes are ideally positioned to modify and to control the immediate environment of neurons. Most astrocytes in the brain are traditionally subdivided into fibrous and protoplasmic types. Fibrous astrocytes (found mainly in white matter) have long, thin, and well-defined processes; protoplasmic astrocytes (found mainly in gray matter) have shorter, frilly processes (Fig. 11-9). Astrocytes are evenly spaced. In cortical regions, the dense processes of an individual astrocyte define its spatial domain, into which adjacent astrocytes do not encroach. The cytoskeleton of these and other types of astrocytes contains an identifying intermediate filament (see Chapter 2) that is composed of a unique protein called glial fibrillar acidic protein (GFAP). The basic physiological properties of both types of astrocyte are similar, but specialized features, such as the expression of neurotransmitter receptors, vary among astrocytes from different brain regions.
Figure 11-9 Astrocytes. The endfeet of both fibrous and protoplasmic astrocytes abut the pia mater and the capillaries.
During development, another type of astrocyte called the radial glial cell (see Chapter 10) is also present. As discussed in Chapter 10, these cells create an organized “scaffolding” by spanning the developing forebrain from the ventricle to the pial surface. Astrocytes in the retina and cerebellum are similar in appearance to radial glial cells. Like astrocytes elsewhere, these cells contain the intermediate filament GFAP. Retinal astrocytes, called Müller cells, are oriented so that they span the entire width of the retina. Bergmann glial cells in the cerebellum have processes that run parallel to the processes of Purkinje cells.
Astrocytes store virtually all the glycogen present in the adult brain. They also contain all the enzymes needed for metabolizing glycogen. The brain’s high metabolic needs are primarily met by glucose transferred from blood because the brain’s glucose supply in the form of glycogen is very limited. In the absence of glucose from blood, astrocytic glycogen could sustain the brain for only 5 to 10 minutes. As implied, astrocytes can share with neurons the energy stored in glycogen, but not by the direct release of glucose into the BECF. Instead, astrocytes break glycogen down to glucose and even further to lactate, which is transferred to nearby neurons, where it can be aerobically metabolized (Fig. 11-10). The extent to which this metabolic interaction takes place under normal conditions is not known, but it may be important during periods of intense neuronal activity, when the demand for glucose exceeds the supply from blood.
Figure 11-10 Role of astrocytes in providing lactate as fuel for neurons. Neurons have two fuel sources. They can obtain glucose directly from the blood plasma, or they can obtain lactate from astrocytes. In the direct path, the oxidation of one glucose molecule provides 30 ATP molecules to the neuron. In the transastrocyte path, conversion of two lactates to two pyruvates, and then the subsequent oxidation of the pyruvate, provides 28 molecules of ATP to the neuron. GLUT1 and GLUT3, glucose transporters; MCT1 and MCT3, monocarboxylate cotransporters.
Astrocytes can also provide fuel to neurons in the form of lactate derived directly from glucose, independent of glycogen. Glucose entering the brain from blood first encounters the astrocytic endfoot. Although it can diffuse past this point to neurons, glucose may be preferentially taken up by astrocytes and shuttled through astrocytic glycolysis to lactic acid, a significant portion of which is excreted into the BECF surrounding neurons. Several observations support the notion that astrocytes provide lactate to neurons. First, astrocytes have higher anaerobic metabolic rates and export much more lactate than do neurons. Second, neurons and their axons function normally when glucose is replaced by lactate, and some neurons seem to prefer lactate to glucose as fuel. Note that when they are aerobically metabolized, the two molecules of lactate derived from the breakdown of one molecule of glucose provide nearly as much ATP as the complete oxidation of glucose itself (28 versus 30 molecules of ATP; see Table 58-4 on p. 1231). The advantage of this scheme for neuronal function is that it provides a form of substrate buffering, a second energy reservoir that is available to neurons. The availability of glucose in the neuronal microenvironment depends on moment-to-moment supply from the blood and varies as a result of changes in neural activity. The concentration of extracellular lactate, however, is buffered against such variability by the surrounding astrocytes, which continuously shuttle lactate to the BECF through the metabolism of glucose or by breaking down glycogen.
Astrocytes are predominantly permeable to K+ and also help regulate [K+]o
The membrane potential of glial cells is more negative than that of neurons. For example, astrocytes have a Vm of about −85 mV, whereas the resting neuronal Vm is about −65 mV. Because the equilibrium potential for K+ is about −90 mV in both neurons and glia, the more negative Vm in astrocytes indicates that glial membranes have higher K+ selectivity than neuronal membranes do (see Chapter 6). Although glial cells express a variety of K+ channels, inwardly rectifying K+ channels seem to be important in setting the resting potential. These channels are voltage gated and are open at membrane potentials that are more negative than about −80 mV, close to the observed resting potential of astrocytes. Astrocytes express many other voltage-gated ion channels that were once thought to be restricted to neurons. The significance of voltage-gated Na+ and Ca2+ channels in glial cells is unknown. Because the ratio of Na+ to K+ channels is low in adult astrocytes, these cells are not capable of regenerative electrical responses such as the action potential.
One consequence of the higher K+ selectivity of astrocytes is that the Vm of astrocytes is far more sensitive than that of neurons to changes in [K+]o. For example, when [K+]o is raised from 4 to 20 mM, astrocytes depolarize by ~25 mV versus only ~5 mV for neurons. This relative insensitivity of neuronal resting potential to changes in [K+]o in the “physiological” range may have emerged as an adaptive feature that stabilizes the resting potential of neurons in the face of the transient increases in [K+]o that accompany neuronal activity. In contrast, natural stimulation, such as viewing visual targets of different shapes or orientations, can cause depolarizations of up to 10 mV in astrocytes of the visual cortex. The accumulation of extracellular K+ that is secondary to neural activity may serve as a signal—to glial cells—that is proportional to the extent of the activity. For example, small increases in [K+]o cause astrocytes to increase their glucose metabolism and to provide more lactate for active neurons. In addition, the depolarization that is triggered by the increased [K+]o leads to the influx of HCO−3 into astrocytes by the electrogenic Na/HCO3cotransporter (see Chapter 5); this influx of bicarbonate in turn causes a fall in extracellular pH that may diminish neuronal excitability. (See Note: Glial Modulation of Neuronal Excitability via Extracellular K+ and pH)
Not only do astrocytes respond to changes in [K+]o, they also help regulate it (Fig. 11-11A). The need for homeostatic control of [K+]o is clear because changes in brain [K+]o can influence transmitter release, cerebral blood flow, cell volume, glucose metabolism, and neuronal activity. Active neurons lose K+ into the BECF, and the resulting increased [K+]o tends to act as a positive feedback signal that increases excitability by further depolarizing neurons. This potentially unstable situation is combated by efficient mechanisms that expedite K+ removal and limit its accumulation to a maximum level of 10 to 12 mM, the so-called ceiling level. [K+]o would rise far above this ceiling with intense neural activity if K+ clearance depended solely on passive redistribution of K+ in the BECF. Neurons and blood vessels can contribute to K+ homeostasis, but glial mechanisms are probably most important. Astrocytes can take up K+ in response to elevated [K+]o by three major mechanisms: the Na-K pump, the Na/K/Cl cotransporter, and the uptake of K+ and Cl− through channels. Conversely, when neural activity decreases, K+ and Cl− leave the astrocytes through ion channels.
Figure 11-11 K+ handling by astrocytes. ECS, extracellular space.
Gap junctions couple astrocytes to one another, allowing diffusion of small solutes
The anatomical substrate for cell-cell coupling among astrocytes is the gap junction, which is composed of membrane proteins called connexins that form large aqueous pores connecting the cytoplasm of two adjacent cells (see Chapter 6). Coupling between astrocytes is strong because hundreds of gap junction channels may be present between two astrocytes. Astrocytes may also be weakly coupled to oligodendrocytes. Ions and organic molecules that are up to 1 kDa in size, regardless of charge, can diffuse from one cell into another through these large channels. Thus, a broad range of biologically important molecules, including nucleotides, sugars, amino acids, small peptides, cAMP, Ca2+, and inositol 1, 4, 5-trisphosphate (IP3), have access to this pathway.
Gap junctions may coordinate the metabolic and electrical activities of cell populations, amplify the consequences of signal transduction, and control intrinsic proliferative capacity. The strong coupling among astrocytes ensures that all cells in the aggregate have similar intracellular concentrations of ions and small molecules and similar membrane potentials. Thus, the network of astrocytes functionally behaves like a syncytium, much like the myocytes in the heart (see Chapter 21). In ways that are not yet clear, gap junctional communication can be important for the control of cellular proliferation. The most common brain cell–derived tumors in the CNS arise from astrocytes. Malignant astrocyte tumors, like malignant neoplasms derived from other cells that are normally coupled (e.g., liver cells), lack gap junctions. (See Note: Astrocytomas)
The coupling among astrocytes may also play an important role in controlling [K+]o by a mechanism known as spatial buffering. The selective K+ permeability of glia, together with their low-resistance cell-cell connections, permits them to transport K+ from focal areas of high [K+]o, where a portion of the glial syncytium would be depolarized, to areas of normal [K+]o, where the glial syncytium would be more normally polarized (Fig. 11-11B). Redistribution of K+ proceeds by way of a current loop in which K+ enters glial cells at the point of high [K+]o and leaves them at sites of normal [K+]o, with the extracellular flow of Na+ completing this circuit. At a site of high neuronal activity, [K+]o might rise to 12 mM, which would produce a very large depolarization of an isolated, uncoupled astrocyte. However, because of the electrical coupling among astrocytes, the Vm of the affected astrocyte remains more negative than the EK predicted for a [K+]o of 12 mM. Thus, K+ would tend to passively enter coupled astrocytes through channels at sites of high [K+]o. As discussed in the preceding section, K+ may also enter the astrocyte by transporters. (See Note: K+ Siphoning by Muller Cells)
Astrocytes synthesize neurotransmitters, take them up from the extracellular space, and have neurotransmitter receptors
Astrocytes synthesize at least 20 neuroactive compounds, including both glutamate and γ-aminobutyric acid (GABA). Neurons can manufacture glutamate from glucose or from the immediate precursor molecule glutamine (Fig. 11-12). The glutamine pathway appears to be the primary one in the synthesis of synaptically released glutamate. Glutamine, however, is manufactured only in astrocytes by use of the astrocyte-specific enzyme glutamine synthetase to convert glutamate to glutamine. Astrocytes release this glutamine into the BECF through the SNAT3 and 5 transporters (SLC38 family; see Table 5-4) for uptake by neurons through SNAT1 and 2. Consistent with its role in the synthesis of glutamate for neurotransmission, glutamine synthetase is localized to astrocytic processes surrounding glutamatergic synapses. In the presynaptic terminals of neurons, glutaminase converts the glutamine to glutamate, for release into the synaptic cleft by the presynaptic terminal. Finally, astrocytes take up much of the synaptically released glutamate to complete this glutamate-glutamine cycle. Disruption of this metabolic interaction between astrocytes and neurons can depress glutamate-dependent synaptic transmission.
Figure 11-12 Role of astrocytes in the glutamate-glutamine cycle. Most of the glutamate of glutamatergic neurons is generated from glutamine, which the neurons themselves cannot make. However, astrocytes take up some of the glutamate that is released at synapses (or produced by metabolism) and convert it into glutamine. The glutamine then enters the neuron, where it is converted back to glutamate. This glutamate also serves as the source for γ-aminobutyric acid in inhibitory neurons.
Glutamine derived from astrocytes is also important for synthesis of the brain’s most prevalent inhibitory neurotransmitter, GABA. In the neuron, the enzyme glutamic acid decarboxylase converts glutamate (generated from glutamine) to GABA (see Fig. 13-8A). Because astrocytes play such an important role in the synthesis of synaptic transmitters, these glial cells are in a position to modulate synaptic efficacy.
Astrocytes have high-affinity uptake systems for the excitatory transmitter glutamate and the inhibitory transmitter GABA. In the case of glutamate uptake, mediated by EAAT1 and EAAT2 (SLC1 family; see Table 5-4), astrocytes appear to play the dominant role compared with neurons or other glial cells. Glutamate moves into cells accompanied by two Na+ ions and an H+ ion, with one K+ ion moving in the opposite direction (Fig. 11-12). Because a net positive charge moves into the cell, glutamate uptake causes membrane depolarization. The presynaptic cytoplasm may contain glutamate at a concentration as high as 10 mM, and vesicles may contain as much as 100 mM glutamate. Nevertheless, the glutamate uptake systems can maintain extracellular glutamate at concentrations as low as ~1 μM, which is crucial for normal brain function.
Neurotransmitter uptake systems are important because they help terminate the action of synaptically released neurotransmitters. Astrocyte processes frequently surround synaptic junctions and are therefore ideally placed for this function. Under pathological conditions in which transmembrane ion gradients break down, high-affinity uptake systems may work in reverse and release transmitters, such as glutamate, into the BECF.
Astrocytes express a wide variety of ionotropic and metabotropic neurotransmitter receptors that are similar or identical to those present on neuronal membranes. As in neurons, activation of these receptors can open ion channels or generate second messengers. In most astrocytes, glutamate produces depolarization by increasing Na+ permeability, whereas GABA hyperpolarizes cells by opening Cl− channels, similar to the situation in neurons (see Chapter 13). Transmitter substances released by neurons at synapses can diffuse in the BECF to activate nearby receptors on astrocytes, thus providing, at least theoretically, a form of neuronal-glial signaling.
Astrocytes apparently can actively enhance or depress neuronal discharge and synaptic transmission by releasing neurotransmitters that they have taken up or synthesized. The release mechanisms are diverse and include stimulation by certain neurotransmitters, a fall in [Ca2+]o, or depolarization by elevated [K+]o. Applying glutamate to cultured astrocytes increases [Ca2+]i, which may oscillate. Moreover, these increases in [Ca2+]i can travel in waves from astrocyte to astrocyte through gap junctions or through a propagated front of extracellular ATP release that activates astrocytic purinergic receptors, thereby increasing [Ca2+]i and releasing more ATP. These [Ca2+]i waves—perhaps by triggering the release of a neurotransmitter from the astrocyte—can lead to changes in the activity of nearby neurons. This interaction represents another form of glial-neuronal communication.
Astrocytes secrete trophic factors that promote neuronal survival and synaptogenesis
Astrocytes, and other glial cell types, are a source of important trophic factors and cytokines, including brain-derived neurotrophic factor, glial-derived neurotrophic factor, basic fibroblast growth factor, and ciliary neurotrophic factor. Moreover, both neurons and glial cells express receptors for these molecules, which are crucial for neuronal survival, function, and repair. The expression of these substances and their cognate receptors can vary during development and with injury to the nervous system.
The development of fully functional excitatory synapses in the brain requires the presence of astrocytes, which act at least in part by secreting proteins called thrombospondins. Indeed, synapses in the developing CNS do not form in substantial numbers before the appearance of astrocytes. In the absence of astrocytes, only ~20% of the normal number of synapses form.
Astrocytic endfeet modulate cerebral blood flow
Astrocytic endfeet surround not only capillaries but also small arteries. Neuronal activity can lead to astrocytic [Ca2+]i waves, as previously described, that spread to the astrocytic endfeet or to isolated increases in endfoot [Ca2+]i. In either case, the result is a rapid increase in blood vessel diameter and thus in local blood flow. A major mechanism of this vasodilation is the stimulation of phospholipase A2 in the astrocyte, the formation of arachidonic acid, and the liberation through cyclooxygenase 1 (see Fig. 3-11) of a potent vasodilator that acts on vascular smooth muscle. This is one mechanism of neuron-vascular coupling—a local increase in neuronal activity that leads to a local increase in blood flow. Radiologists exploit this physiological principle in a form of functional magnetic resonance imaging (fMRI) called blood oxygen level–dependent (BOLD) MRI, which uses blood flow as an index of neuronal activity.
Excitatory Amino Acids and Neurotoxicity
The dicarboxylic amino acid glutamate is the most prevalent excitatory neurotransmitter in the brain (see Chapter 13). Although glutamate is present at millimolar levels inside neurons, the BECF has only micromolar levels of glutamate, except at sites of synaptic release (Fig. 11-12). Excessive accumulation of glutamate in the BECF—induced by ischemia, anoxia, hypoglycemia, or trauma—can lead to neuronal injury. Astrocytes are intimately involved in the metabolism of glutamate and its safe disposition after synaptic release.
In anoxia and ischemia, the sharp drop in cellular levels of ATP inhibits the Na-K pump, thereby rapidly leading to large increases in [K+]o and [Na+]i. These changes result in membrane depolarization, with an initial burst of glutamate release from vesicles in presynaptic terminals. Vesicular release, however, requires cytoplasmic ATP and probably halts rapidly. The ability of astrocytes to remove glutamate from the BECF is impeded by the elevated [K+]o, elevated [Na+]i, and membrane depolarization. In fact, the unfavorable ion gradients can cause the transporter to run in reverse and dump glutamate into the BECF. The action of rising levels of extracellular glutamate on postsynaptic and astrocytic receptors reinforces the developing ionic derangements by opening channels permeable to Na+ and K+. This vicious cycle at the level of the astrocyte can rapidly cause extracellular glutamate to reach levels that are toxic to neurons—excitotoxicity.
Astrocytic modulation of blood flow is complex, and increases in [Ca2+]i in endfeet can sometimes lead to vasoconstriction.
Oligodendrocytes and Schwann cells make and sustain myelin
The primary function of oligodendrocytes as well as of their PNS equivalent, the Schwann cell, is to provide and to maintain myelin sheaths on axons of the central and peripheral nervous systems, respectively. As discussed in Chapter 7, myelin is the insulating “electrical tape” of the nervous system (see Fig. 7-21B). Oligodendrocytes are present in all areas of the CNS, although their morphological appearance is highly variable and depends on their location within the brain. In regions of the brain that are dominated by myelinated nerve tracts, called white matter, the oligodendrocytes responsible for myelination have a distinctive appearance (Fig. 11-13A). Such an oligodendrocyte has 15 to 30 processes, each of which connects a myelin sheath to the oligodendrocyte’s cell body. Each myelin sheath, which is up to 250 μm wide, wraps many times around the long axis of one axon. The small exposed area of axon between adjacent myelin sheaths is called the node of Ranvier (see Chapter 10). In gray matter, oligodendrocytes do not produce myelin and exist as perineuronal satellite cells.
Figure 11-13 Myelination of axons by oligodendrocytes and Schwann cells.
During the myelination process, the leading edge of one of the processes of the oligodendrocyte cytoplasm wraps around the axon many times (Fig. 11-13A, upper axon). The cytoplasm is then squeezed out of the many cell layers surrounding the axon in a process called compaction. This process creates layer on layer of tightly compressed membranes that is called myelin. The myelin sheaths remain continuous with the parent glial cells, which nourish them.
In the PNS, a single Schwann cell provides a single myelin segment to a single axon of a myelinated nerve (Fig. 11-13B). This situation stands in contrast to that in the CNS, where one oligodendrocyte myelinates many axons. The process of myelination that occurs in the PNS is analogous to that outlined for oligodendrocytes. Axons of unmyelinated nerves are also associated with Schwann cells. In this case, the axons indent the surface of the Schwann cell and are completely surrounded by Schwann cell cytoplasm (Fig. 11-14).
Figure 11-14 Ensheathed versus myelinated axons. A, Ensheathed axons. This transmission electron micrograph shows a Schwann cell surrounding several unmyelinated peripheral axons, some of which are marked with an asterisk. The arrows point to the basal lamina. The arrowhead points to collagen fibrils. B, Myelinated axons. This transmission electron micrograph shows a Schwann cell (nucleus on right side of picture) surrounding a peripheral axon with several layers of myelin. The lower star shows the beginning of the spiraling myelin sheath; the upper star indicates the termination of the spiral and a small region of noncompacted cytosol. The final magnification is ~14,000 in both panels. (Reproduced from Bunge RP, Fernandez-Valle C: In Kettemann H, Ransom RR [eds]: Neuroglia, pp. 44-57. New York: Oxford University Press, 1995; courtesy of Mary Bartlett Bunge.)
Myelin has a biochemical composition different from that of the oligodendrocyte or Schwann cell plasma membrane from which it arose. Although PNS myelin and CNS myelin look similar, some of the constituent proteins are different (Table 11-4). For example, proteolipid protein is the most common protein in CNS myelin (~50% of total protein) but is absent in PNS myelin. Conversely, P0 is found almost exclusively in PNS myelin.
Table 11-4 Proteins in Myelin
|
Protein |
CNS (% of total myelin proteins) |
PNS (% of total myelin proteins) |
|
MBP |
30 |
<18 |
|
PLP |
50 |
<0.01 |
|
MAG |
<1 |
<50.1 |
|
CNP |
<4 |
<50.4 |
|
P0 |
<0.01 |
>50 |
|
P2 |
<1 |
1-15 |
|
PMP22 |
<0.01 |
5-10 |
|
MOG |
<0.05 |
<0.01 |
CNP, cyclic nucleotide phosphodiesterase; MAG, myelin-associated glycoprotein; MBP, myelin basic protein; MOG, myelin oligodendrocyte glycoprotein; PLP, proteolipid protein; PMP22, peripheral myelin protein 22.
Myelination greatly enhances conduction of the action potential down the axon because it allows the regenerative electrical event to skip from one node to the next rather than gradually spreading down the whole extent of the axon. This process is called saltatory conduction (see Chapter 7). Besides being responsible for CNS myelin, oligodendrocytes play another key role in saltatory conduction: they induce the clustering of Na+ channels at the nodes (see Fig. 12-5C on p. 318), which is essential for saltatory conduction.
It is well known that severed axons in the PNS can regenerate with restoration of lost function. Regrowth of these damaged axons is coordinated by the Schwann cells in the distal portion of the cut nerve. Severed axons in the CNS do not show functional regrowth, in part because of the growth-retarding nature of CNS myelin (see Chapter 10).
Oligodendrocytes are involved in pH regulation and iron metabolism in the brain
Oligodendrocytes and myelin contain most of the enzyme carbonic anhydrase within the brain. The appearance of this enzyme during development closely parallels the maturation of these cells and the formation of myelin. Carbonic anhydrase rapidly catalyzes the reversible hydration of CO2 and may thus allow the CO2/HCO−3 buffer system to be maximally effective in dissipating pH gradients in the brain. The pH regulation in the brain is important because it influences neuronal excitability. The classic example of the brain’s sensitivity to pH is the reduced seizure threshold caused by the respiratory alkalosis secondary to hyperventilation (see Chapter 28).
Oligodendrocytes are the cells in the brain most involved with iron metabolism. They contain the iron storage protein ferritin and the iron transport protein transferrin. Iron is necessary as a cofactor for certain enzymes and may catalyze the formation of free radicals under pathological circumstances, such as disruption of blood flow to the brain.
Oligodendrocytes, like astrocytes, have a wide variety of neurotransmitter receptors. Unmyelinated axons can release glutamate when they conduct action potentials, and in principle, this glutamate could signal nearby oligodendrocytes. Ischemia readily injures oligodendrocytes, in part by releasing toxic levels of glutamate. Even white matter, therefore, can suffer excitotoxicity.
Microglial cells are the macrophages of the CNS
Microglial cells are of mesodermal origin and derive from cells related to the monocyte-macrophage lineage. Microglia represent ~20% of the total glial cells within the mature CNS. These cells are rapidly activated by injury to the brain, which causes them to proliferate, to change shape, and to become phagocytic (Fig. 11-15). When activated, they are capable of releasing substances that are toxic to neurons, including free radicals and nitric oxide. It is believed that microglia are involved in most brain diseases, not as initiators but as highly reactive cells that shape the brain’s response to any insult.
Figure 11-15 Microglial cells. Resting microglial cells become activated by injury to the brain, which causes them to proliferate and to become phagocytic.
Microglia are also the most effective antigen-presenting cells within the brain. Activated T lymphocytes are able to breech the blood-brain barrier and enter the brain. To become mediators of tissue-specific disease or to destroy an invading infectious agent, T lymphocytes must recognize specific antigenic targets. Such recognition is accomplished through the process of antigen presentation, which is a function of the microglia.
REFERENCES
Books and Reviews
Brown PD, Davis SL, Speake T, Millar ID: Molecular mechanisms of cerebrospinal fluid production. Neuroscience 2004; 129:957-970.
Kettenmann H, Ransom BR eds: Neuroglia. 2nd ed, New York: Oxford University Press; 2005.
Nedergaard M, Ransom BR, Goldman SA: New roles for astrocytes: Redefining the functional architecture of the brain. TINS 2003;26:523-530.
Ohtsuki S: New aspects of the blood-brain-barrier transporters; its physiological roles in the central nervous system. Biol Pharm Bull 2004;27:1489-1496.
Ullian EM, Christopherson KS, Barres BA: Role for glia in synaptogenesis. Glia 2004;47:209-216.
Journal Articles
Christopherson KS, Ullian EM, Stokes CC, et al: Thrombospondins are astrocyte-secreted proteins that promote CNS synaptogenesis. Cell 2005;120:421-433.
Kaplan MR, Meyer-Franke A, Lambert S, et al: Induction of sodium channel clustering by oligodendrocytes. Nature 1997;386:724-728.
Newman EA, Zahs KR: Modulation of neuronal activity by glial cells in the retina. J Neurosci 1998;18:4022-4028.
Takano T, Tian GF, Peng W, et al: Astrocyte-mediated control of cerebral blood flow. Nat Neursci 2006;9:260-267.
Wender R, Brown AM, Fern R, et al: Astrocytic glycogen influences axon function and survival during glucose deprivation in central white matter. J Neurosci 2000;20:6804-6810.