Virtually nothing can enter or leave the central nervous system parenchyma without passing through an astrocytic interphase

The complex packing achieved by the processes and cell bodies of astrocytes underscores their involvement in brain metabolism. Although astrocytes traditionally have been subdivided into protoplasmic and fibrous astrocytes [4], these two forms probably represent the opposite ends of a spectrum of the same cell type. However, Raff et al. [22] have suggested that the two groups might derive from different progenitors and that the progenitor of the fibrous astrocyte is the same as that of the oligodendrocyte. The structural components of fibrous and protoplasmic astrocytes are identical; the differences are quantitative. In the early days of EM, differences between the two variants were more apparent owing to imprecise techniques, but with the development of better procedures, the differences became less apparent.

Protoplasmic astrocytes range in size from 10 to 40 μm, frequently are located in gray matter in relation to capillaries and have a clearer cytoplasm than do fibrous astrocytes (Fig. 1-12). Within the perikaryon of both types of astrocyte are scattered 9-nm filaments and 24-nm microtubules (Fig. 1-13); glycogen granules; lysosomes and lipofuscin-like bodies; isolated cisternae of the rough ER; a small Golgi apparatus opposite one pole of the nucleus; and small, elongated mitochondria, often extending together with loose bundles of filaments along cell processes. A centriole is not uncommon. Characteristically, the nucleus is ovoid and the nucleochromatin homogeneous, except for a narrow, continuous rim of dense chromatin and one or two poorly defined nucleoli. The fibrous astrocyte occurs in white matter (Fig. 1-13). Its processes are twig-like, being composed of large numbers of 9-nm glial filaments arranged in tight bundles. The filaments within these cell processes can be distinguished from neurofilaments by their close packing and the absence of side-arms (Figs. 1-13 and 1-14). Desmosomes and gap junctions occur between adjacent astrocytic processes.

Figure 1-12

A protoplasmic astrocyte abuts a blood vessel (lumen at L) in rat cerebral cortex. The nucleus shows a rim of denser chromatin, and the cytoplasm contains many organelles, including Golgi and rough endoplasmic reticulum. ×10,000. Inset: Detail (more...)

Figure 1-13

A section of myelinating white matter from a kitten contains a fibrous astrocyte (A) and an oligodendrocyte (O). The nucleus of the astrocyte (A) has homogeneous chromatin with a denser rim and a central nucleolus. That of the oligodendrocyte (O) is denser (more...)

In addition to protoplasmic and fibrous forms, regional specialization occurs among astrocytes. The outer membranes of astrocytes located in subpial zones and adjacent to blood vessels possess a specialized thickening. Desmosomes and gap junctions are very common in these regions between astrocytic processes. In the cerebellar cortex, protoplasmic astrocytes can be segregated into three classes, each ultrastructurally distinct: the Golgi epithelial cell, the lamellar or velate astrocyte and the smooth astrocyte [1].

Astrocyte functions have long been debated. Their major role is related to a connective tissue or skeletal function since they invest, possibly sustain and provide a packing for other CNS components. In the case of astrocytic ensheathment around synaptic complexes and the bodies of some neurons, such as Purkinje cells, it has been speculated that the astrocyte isolates these structures.

One well-known function of the astrocyte is concerned with repair. Subsequent to trauma, astrocytes invariably proliferate, swell, accumulate glycogen and undergo fibrosis by the accumulation of filaments, expressed neurochemically as an increase in glial fibrillary acidic protein (GFAP). This state of gliosis may be total, in which case all other elements are lost, leaving a glial scar, or it may be a generalized response occurring against a background of regenerated or normal CNS parenchyma. Fibrous astrocytosis can occur in both the gray and white matter, thereby indicating common links between protoplasmic and fibrous astrocytes. With age, both fibrous and protoplasmic astrocytes accumulate filaments. In some diseases, astrocytes become macrophages. It is interesting to note that the astrocyte is probably the most disease-resistant component in the CNS because very few diseases, other than alcoholism, cause depletion of astrocytes.

Another putative role of the astrocyte is its involvement in transport mechanisms (see Chap. 5) and in the BBB system (see Chap. 32). Astrocytes interact with neurons in various metabolic and transport processes. It was believed for some time that transport of water and electrolytes was effected by the astrocyte, a fact never definitively demonstrated and largely inferred from pathological or experimental evidence. It is known, for example, that damage to the brain vasculature, local injury due to heat or cold and inflammatory changes produce focal swelling of astrocytes, presumably owing to disturbances in fluid transport. The astrocytic investment of blood vessels suggests a role in the BBB system, but the studies of Reese and Karnovsky [23] and Brightman [24] indicate that the astrocytic end-feet provide little resistance to the movement of molecules and that blockage of the passage of material into the brain occurs at the endothelial cell-lining blood vessels (see Chap. 32). CNS endothelial cells display selective transport by transcytosis. During inflammation, these mechanisms are disrupted and there are alterations in permeability of endothelial tight junctions and formation of edema. Astrocytes also are involved in reuptake of the neurotransmitter glutamate (see Chaps. 5 and 15). Finally, it is believed that astrocytes are responsible for the regulation of local pH levels and local ionic balances.

Molecular markers of astrocytes. Although antigenically distinct from other cell types by virtue of its expressing GFAP [17], there is no documented evidence of astrocytic disease related to an immunological response to GFAP on any astroglial molecule. GFAP remains singularly the most used cytoplasmic marker of astrocytes. A reliable marker for astrocytic membranes remains to be described. Interestingly, there is increasing evidence demonstrating the ability of astrocytes to serve as accessory cells of the immune system in a number of immune-mediated conditions [27,28]. In this regard, astrocytes are known for their ability to express class II MHC antigens in vitro, which are molecules essential for the presentation of antigen to helper/inducer CD4⁺ T cells, as well as their ability to synthesize a number of cytokines, such as interleukin-1, tumor necrosis factor and interferon γ (see Chaps. 35 and 39). It appears, therefore, that in circumstances in which the BBB is interrupted, the astrocyte is a facultative phagocyte with the potential to interact with lymphocytes.

Oligodendrocytes are myelin-producing cells in the central nervous system

The ultrastructural studies of Schultz and co-workers (1957) and Farquhar and Hartman (1957) (discussed in [4]) were among the first to contrast the EM features of oligodendrocytes with astrocytes (Fig. 1-12). The study of Mugnaini and Walberg [4] more explicitly laid down the morphological criteria for identifying these cells, and apart from subsequent technical improvements, our EM understanding of these cells has changed little since that time [5,29].

As with astrocytes, oligodendrocytes are highly variable, differing in location, morphology and function, but definable by some morphological criteria. The cell soma ranges from 10 to 20 μm and is roughly globular and more dense than that of an astrocyte. The margin of the cell is irregular and compressed against the adjacent neuropil. Few cell processes are seen, in contrast to the astrocyte. Within the cytoplasm, many organelles are found. Parallel cisternae of the rough ER and a widely dispersed Golgi apparatus are common. Free ribosomes occur, scattered amid occasional multivesicular bodies, mitochondria and coated vesicles. Distinguishing the oligodendrocyte from the astrocyte are the apparent absence of glial filaments and the constant presence of 24-nm microtubules (Fig. 1-13). Microtubules are most common at the margins of the cell, in the occasional cell process and in the cytoplasmic loops around myelin sheaths. Lamellar dense bodies, typical of oligodendrocytes, are also present [5]. The nucleus is usually ovoid, but slight lobation is not uncommon. The nucleochromatin stains heavily and contains clumps of denser heterochromatin; the whole structure is sometimes difficult to discern from the background cytoplasm. Desmosomes and gap junctions occur between interfascicular oligodendrocytes [5].

Ultrastructural and labeling studies on the developing nervous system (see Chap. 27) have demonstrated variability in oligodendrocyte morphology and activity. Mori and Leblond (see [5]) separated oligodendrocytes into three groups based on location, stainability and DNA turnover. Their three classes correspond to satellite, intermediate and interfascicular, or myelinating, oligodendrocytes. Satellite oligodendrocytes are small (~10 μm), restricted to gray matter and closely applied to the surface of neurons. They are assumed to play a role in the maintenance of the neuron and are potential myelinating cells. Interfascicular oligodendrocytes are large (~20 μm) during myelination but, in the adult, range from 10 to 15 μm, with the nucleus occupying a large percentage of the cell volume. Intermediate oligodendrocytes are regarded as satellite or potential myelinating forms. The nucleus of these cells is small, the cytoplasm occupying the greater area of the soma.

Myelinating oligodendrocytes have been studied extensively [5,30] (see Chap. 4). Examination of the CNS during myelinogenesis (Fig. 1-15) reveals connections between the cell body and the myelin sheath [31]; however, connections between these elements have never been demonstrated in a normal adult animal, unlike the PNS counterpart, the Schwann cell. In contrast to the Schwann cell (see below), the oligodendrocyte is capable of producing many internodes of myelin simultaneously. It is estimated that oligodendrocytes in the optic nerve produce between 30 and 50 internodes of myelin [5]. In addition to this heavy structural commitment, the oligodendrocyte possesses a slow mitotic rate and a poor regenerative capacity. Damage to only a few oligodendrocytes, therefore, can be expected to produce an appreciable area of primary demyelination. In most CNS diseases in which myelin is a target, oligodendrocytes are among the most vulnerable elements and the first to degenerate (see Chap. 39).

Figure 1-15

A myelinating oligodendrocyte, nucleus (N), from the spinal cord of a 2-day-old kitten extends cytoplasmic connections to at least two myelin sheaths (arrows). Other myelinated and unmyelinated fibers at various stages of development, as well as glial (more...)

Somewhat analogous to the neuron, the relatively small oligodendrocyte soma produces and supports many more times its own volume of membrane and cytoplasm. For example, consider an average 12-μm oligodendrocyte producing 20 internodes of myelin [5]. Each axon has a diameter of 3 μm and is covered by at least six lamellae of myelin, each lamella representing two fused layers of unit membrane. By statistical analysis, taking into account the length of the myelin internode, which is possibly 500 μm, and the length of the membranes of the cell processes connecting the sheaths to the cell body (~12 μm), the ratio between the surface area of the cell soma and the myelin it sustains is approximately 1:620. In most cases, however, this ratio is probably in the region of 1:3,000. In rare instances, oligodendrocytes elaborate myelin around structures other than axons in that myelin has been documented around neuronal somata and nonaxonal profiles.

Molecular markers of oligodendrocytes. The oligodendrocyte is potentially highly vulnerable to immune-mediated damage since it shares with the myelin sheath many molecules with known affinities to elicit specific T- and B-cell responses, which lead to its destruction. Chapter 39 describes the immune process in demyelination. Many of these molecules, such as myelin basic protein, proteolipid protein, myelin-associated glycoprotein, myelin/oligodendrocyte protein, galactocerebroside, myelin oligodendrocyte glycoprotein (MOG) and others, have been used to generate specific antibodies, which are routinely applied to anatomical analyses of oligodendrocytes in vivo and in vitro. However, unlike the astrocyte, the oligodendrocyte expresses no class I or II MHC molecules suggestive of interactions with the immune system [32].

The microglial cell plays a role in phagocytosis and inflammatory responses

Of the few remaining types of CNS cells, the most interesting, and probably the most enigmatic, is the microglial cell, a cell of mesodermal origin, located in the normal brain in a resting state and purported to become a very mobile, active macrophage during disease (see Chap. 35). Microglia can be stained selectively and demonstrated by light microscopy using Hortega's silver carbonate method, but no comparable technique exists for their ultrastructural demonstration. The cells have spindle-shaped bodies and a thin rim of densely staining cytoplasm difficult to distinguish from the nucleus. The nucleochromatin is homogeneously dense, and the cytoplasm does not contain an abundance of organelles, although representatives of the usual components can be found. During normal wear and tear, some CNS elements degenerate and microglia phagocytose the debris (Fig. 1-16). Their identification and numbers, as determined by light microscopy, differ from species to species. The CNS of rabbit is richly endowed. In a number of disease instances, such as trauma, microglia are stimulated and migrate to the area of injury, where they phagocytose debris. The relatively brief mention of this cell type in the major EM textbooks [3] and the conflicting EM descriptions [33] are indicative of the uncertainty attached to their identification. Pericytes are believed by some to be a resting form of microglial cell. Perivascular macrophages, which are of bone marrow origin and are distinct from parenchymal microglia, also have been described.

Figure 1-16

A microglial cell (M) has elaborated two cytoplasmic arms to encompass a degenerating apoptotic oligodendrocyte (O) in the spinal cord of a 3-day-old kitten. The microglial cell nucleus is difficult to distinguish from the narrow rim of densely staining (more...)

Molecular markers of microglial cells. There has been a veritable explosion of activity in the field of microglial cell biology with the realization that this cell type is capable of functioning as a highly efficient accessory cell of the immune system. While no particularly microglia-specific molecule has been identified, a number of antibodies raised against monocytic markers and complement receptor molecules stain microglial cells in situ and in vitro. There is strong evidence that microglia express class II MHC upon activation [34–37], frequently in the absence of a T-cell response. This suggests that class II MHC expression may represent a marker of activation or in some way elevate the cells to a state of immunological awareness. Microglia are also producers of a number of proinflammatory cytokines with known effects upon T cells. Taken in concert, the increasing evidence of an immunological role for microglia in a wide spectrum of conditions probably supports the putative monocytic origin of this cell type.

Ependymal cells line the brain ventricles and the spinal cord central canal

Ependymal cells are arranged in single-palisade arrays and line the ventricles of the brain and central canal of the spinal cord. They are usually ciliated, their cilia extending into the ventricular cavity. Their fine structure has been elucidated by Brightman and Palay [38]. They possess several features that clearly differentiate them from any other CNS cell. The cilia emerge from the apical pole of the cell, where they are attached to a blepharoplast, the basal body (Fig. 1-17), which is anchored in the cytoplasm by means of ciliary rootlets and a basal foot. The basal foot is the contractile component that determines the direction of the ciliary beat. Like all flagellar structures, the cilium contains the common microtubule arrangement of nine peripheral pairs around a central doublet (Fig. 1-17). In the vicinity of the basal body, the arrangement is one of nine triplets; at the tip of each cilium, the pattern is one of haphazardly organized single tubules. Also, extending from the free surface of the cell are numerous microvilli containing actin microfilaments (Fig. 1-17). The cytoplasm stains intensely, having an electron density about equal to that of the oligodendrocyte, whereas the nucleus has a similar density to that of the astrocyte. Microtubules; large whorls of filaments; coated vesicles; rough ER; Golgi apparatus; lysosomes; and abundant small, dense mitochondria are also present. The base of the cell is composed of involuted processes that interdigitate with the underlying neuropil. The lateral margins of each cell characteristically display long, compound, junctional complexes (Fig. 1-18) made up of desmosomes, termed zonula adherentes, and gap junctions [3]. Overlying specialized secretory zones around the ventricles, the so-called subventricle organs and choroid plexus, the ependymal lining is different and the cells are connected at their apical poles by tight junctions called zonula occludentes. Desmosomes and gap junctions are also present at the lateral aspects of the cells [39].

Figure 1-17

The surface of an ependymal cell contains basal bodies (arrows) connected to the microtubules of cilia, seen here in longitudinal section. Several microvilli are also present. ×37,000. Inset: Ependymal cilia in transverse section possess a central (more...)

Figure 1-18

A typical desmosome (d) and gap junction (g) between two ependymal cells. Microvilli and coated pits (arrows) are seen along the cell surface. ×35,000.

The biochemical properties of these structures are known. Desmosomes display protease sensitivity, divalent cation dependency and osmotic insensitivity; and their membranes are mainly of the smooth type. In direct contrast to desmosomes, the tight junctions as well as gap junctions and synapses display no protease sensitivity, divalent cation dependency or osmotic sensitivity, while their membranes are complex. These facts have been used in the development of techniques to isolate purified preparations of junctional complexes.

The Schwann cell is the myelin-producing cell of the peripheral nervous system

When axons leave the CNS, they lose their neuroglial interrelationships and traverse a short transitional zone, where they are invested by an astroglial sheath enclosed in the basal lamina of the glia limitans. The basal lamina then becomes continuous with that of axon-investing Schwann cells, at which point the astroglial covering terminates. Schwann cells, therefore, are the axon-ensheathing cells of the PNS, equivalent functionally to the oligodendrocyte of the CNS (see Chap. 4). Along the myelinated fibers of the PNS, each internode of myelin is elaborated by one Schwann cell and each Schwann cell elaborates one internode [30]. This ratio of one internode of myelin to one Schwann cell is a fundamental distinction between this cell type and its CNS analog, the oligodendrocyte, which is able to proliferate internodes in the ratio of 1:30 or greater. Another distinction is that the Schwann cell body always remains in intimate contact with its myelin internode (Fig. 1-19), whereas the oligodendrocyte extends processes toward its internodes. Periodically, myelin lamellae open up into ridges of Schwann cell cytoplasm, producing bands of cytoplasm around the fiber, Schmidt-Lanterman incisures, reputed to be the stretch points along PNS fibers. These incisures usually are not present in the CNS. The PNS myelin period is 11.9 nm in preserved specimens, which is some 30% less than in the fresh state, in contrast to the 10.6 nm of central myelin. In addition to these structural differences, PNS myelin differs biochemically and antigenically from that of the CNS (see Chap. 4). Not all PNS fibers are myelinated, but in contrast to nonmyelinated fibers in the CNS, nonmyelinated fibers in the PNS are suspended in groups within the Schwann cell cytoplasm, each axon connected to the extracellular space by a short channel, the mesaxon, formed by the invaginated Schwann cell plasmalemma.

Figure 1-19

A myelinated PNS axon (A) is surrounded by a Schwann cell, nucleus (N). Note the fuzzy basal lamina around the cell, the rich cytoplasm, the inner and outer mesaxons (arrows), the close proximity of the cell to its myelin sheath and the 1:1 (cell:myelin (more...)

Ultrastructurally, the Schwann cell is unique and distinct from the oligodendrocyte. Each Schwann cell is surrounded by a basal lamina made up of a mucopolysaccharide approximately 20 to 30 nm thick that does not extend into the mesaxon (Fig. 1-19). The basal laminae of adjacent myelinating Schwann cells at the nodes of Ranvier are continuous, and Schwann cell processes interdigitate so that the PNS myelinated axon is never in direct contact with the extracellular space. These nodal Schwann cell fingers display intimate relationships with the axolemma (Figs. 1-20 and 1-21), suggesting that the entire nodal complex might serve as an electrogenic pump for the recycling of ions [10]. A similar arrangement between the nodal axon and the fingers of astroglial cells is seen in the CNS. The Schwann cells of nonmyelinated PNS fibers overlap, and there are no nodes of Ranvier.

Figure 1-20

Low-power electron micrograph of a node of Ranvier in longitudinal section. Note the abrupt decrease in axon diameter and the attendant condensation of axoplasmic constituents in the paranodal and nodal regions of the axon. Paranodal myelin is distorted (more...)

Figure 1-21

A transverse section of the node of Ranvier (7 to 8 μm across) of a large fiber shows a prominent complex of Schwann cell fingers around an axon highlighted by its subaxolemmal densification and closely packed organelles. The Schwann cell fingers (more...)

The cytoplasm of the Schwann cell is rich in organelles. A Golgi apparatus is located near the nucleus, and cisternae of the rough ER occur throughout the cell. Lysosomes, multivesicular bodies, glycogen granules and lipid granules, also termed pi granules, also can be seen. The cell is rich in microtubules and filaments, in contrast to the oligodendrocyte. The plasmalemma frequently shows pinocytic vesicles. Small, round mitochondria are scattered throughout the soma. The nucleus, which stains intensely, is flattened and oriented longitudinally along the nerve fiber. Aggregates of dense heterochromatin are arranged peripherally [3].

In sharp contrast to the oligodendrocyte, the Schwann cell responds vigorously to most forms of injury (see Chap. 39). An active phase of mitosis occurs following traumatic insult, and the cells are capable of local migration. Studies on their behavior after primary demyelination have shown that they are able to phagocytose damaged myelin. They possess remarkable reparatory properties and begin to lay down new myelin approximately 1 week after a fiber loses its myelin sheath. Studies on PNS and CNS remyelination [40] have shown that by 3 months after primary demyelination, PNS fibers are well remyelinated, whereas similarly affected areas in the CNS show relatively little proliferation of new myelin (see Chap. 29). Under circumstances of severe injury, such as transection, axons degenerate and the Schwann cells form tubes, termed Büngner bands, containing cell bodies and processes surrounded by a single basal lamina. These structures provide channels along which regenerating axons might later grow. The presence and integrity of the Schwann cell basal lamina is essential for reinnervation.

The extracellular space between peripheral nerve fibers is occupied by bundles of collagen fibrils, blood vessels and endoneurial cells

Endoneurial cells are elongated, spindle-shaped cells with tenuous processes relatively poor in organelles except for large cisternae of the rough ER. There is some evidence that these cells proliferate collagen fibrils. Sometimes mast cells, the histamine producers of connective tissue, can be seen. Bundles of nerve fibers are arranged in fascicles emarginated by flattened connective tissue cells forming the perineurium, an essential component in the blood—nerve barrier system. Fascicles of nerve fibers are aggregated into nerves and invested by a tough elastic sheath of cells known as the epineurium [41].