23:  23.5 Overview of Neuronal Outgrowth

Individual Neurons Can Be Identified Reproducibly and Studied

A prerequisite for understanding how specific neural contacts form is identifying specific neurons in the developing embryo and observing them as their axons elongate and form contacts. Embryos of grasshoppers, Drosophila, and other insects in which the central nervous system is formed by a string of ganglia, one per body segment, provide excellent experimental systems. A single ganglion may contain a thousand or so nerve-cell bodies; some of these have a characteristic size and position, permitting precise identification of the same cell in different embryos. When a nerve cell is sufficiently large, it can be microinjected with a fluorescent dye (e.g., lucifer yellow) that spreads throughout its cytosol. The cell’s projections then can be visualized in the fluorescence microscope (Figure 23-30a). Analogous techniques can be used to identify vertebrate motor neurons. For example, the enzyme peroxidase injected into a nerve near its terminus is transported back to the cell body by retrograde axonal transport (Chapter 19). The enzyme can then be detected in fixed tissue by histochemical staining. In this way, it is possible to localize in the spinal cord the cell body of a particular neuron that innervates a specific muscle. Lipophilic dyes (e.g., DiO and DiI) also have been particularly useful in studying neuronal development in vertebrates. They diffuse within the membrane highlighting the entire extent of the axon and neuritic processes. In contrast to peroxidase labeling, lipophilic dyes can be used to stain fixed specimens or to follow events in living tissue using confocal microscopy.

Important insights into neuronal wiring have come from study of the simple soil worm, C. elegans, which exists in two forms—a self-fertilizing hermaphrodite and a male that facilitates cross-fertilization. The adult hermaphrodite contains 959 somatic cells; 301 of these are neurons (see Figure 5-13c). Because of the small number of neurons in C. elegans, researchers have been able to reconstruct the complete neuronal wiring pattern from thin sections visualized by electron microscopy. Moreover, the short generation time of C. elegans (52 hours at 25 °C) and the ease of genetic analysis in this system have permitted identification of many different mutations affecting neuronal wiring. Labeling methods, such as that illustrated in Figure 23-30b, allow analysis of the effects of such mutations on outgrowth of individual neurons in living worms.

Growth Cones Guide the Migration and Elongation of Developing Axons

During differentiation of the nervous system, precursors of neurons, called neuroblasts, arise at particular locations at specific times. Such cells lack axonal or dendritic projections. Certain neuroblasts migrate to specific destinations where they form clusters called ganglia. Vertebrate cells migrate extensively in the central nervous system. As a newly born neuron begins to differentiate, it grows one or more axonal projections. At the leading edge of the elongating axon is the highly motile growth cone, which possesses cell-surface receptors for extracellular signals (see Figure 23-29).

As the growth cone moves outward, the cell body stays put and the axon elongates due, in part, to the polymerization of tubulin into microtubules, which give the axon its rigidity. Ligand binding to receptors on growth cones triggers intracellular signaling that regulates their motility, in part, by controlling actin polymerization in the filopodia and lamellipodia at the distal end. Recognition of signaling molecules by receptors located at different positions along the periphery of advancing growth cones is thought to steer them toward the appropriate target cells.

Different Neurons Navigate along Different Outgrowth Pathways

Although the nervous systems of vertebrates and invertebrates are different in their structure and complexity, similar principles of axon guidance are used: growth cones of elongating axons use a changing set of cell-surface receptors to move along specific extracellular-matrix fibers and also along specific cells. The notion that specific molecules guide growth cones to their targets was first proposed nearly a century ago. Here we describe a few examples that argue persuasively for the importance of specific cues in directing growth cones to their targets.

In vertebrates, the cell bodies of motor neurons are located in the spinal cord, and axons extend out of the central nervous system by ventral roots. Axons that innervate a specific muscle are bound together to form a peripheral nerve. Obviously it is essential for each motor neuron to innervate only the appropriate muscle, and indeed this high degree of specificity is found throughout the nervous system.

In certain favorable experimental systems, scientists can observe individual motor neurons as their axons exit the spinal cord and elongate. During embryogenesis in the zebrafish, for example, the axons of three pioneer neurons are the first to emerge from the spinal cord in each segment of the embryo (Figure 23-31). Investigators identified the cell bodies of these three motor neurons and followed the trajectories of their axons by microinjection with a fluorescent dye. These studies showed that the axons of all three pioneer neurons initially grow out of the spinal cord along a common pathway. They then diverge: one axon grows upward to innervate dorsal muscles; one downward to innervate ventral muscles; and one laterally to innervate both muscles. These pioneer axons are never seen to send branches or growth cones off in an inappropriate direction.

At a later stage of development, axons of other neurons grow out of the central nervous system to innervate the same muscles. Growth cones of these axons migrate along the surfaces of the three pioneer axons, which guide the secondary neurons to the correct muscles. However, experimental destruction of one of the pioneer neurons by an intense laser beam does not affect the ability of the secondary motor neurons to select the pathway that leads to the appropriate target muscle. This finding suggests that by the time a growth cone exits the spinal cord, the neuron is already programmed to follow a specific pathway. Experiments with the chick embryo—in which nerve transplantation experiments are easier—support this contention. Here, motor neurons from four adjacent segments of the spinal cord innervate various muscles of the hindlimb. Multiple axons destined for different muscles grow out of the spinal cord together and follow the same path to plexuses; from there the individual neurons follow different paths to the appropriate muscle. If, before nerve outgrowth, a piece of spinal cord containing motoneuron cell bodies is transplanted from one segment to an adjacent one, these motoneuron axons will exit the spinal cord at the “wrong” site. However, they will grow into the proper plexus and still innervate the correct muscle. A similar result is obtained if a segment is inverted, so that anterior neurons are moved to the posterior.

Clearly, the specificity in guidance of any axon does not depend on where it leaves the spinal cord, nor on its transplanted position within a segment. It seems to depend on its intrinsic properties, perhaps an inherent set of receptors on each growth cone that allow it to recognize the surrounding environment. For instance, in Figure 23-31 the neuron that grows along the ventral pathway does so because it specifically recognizes components of the extracellular matrix or glial or other cells located in this region. A growth cone most probably moves from one short-range target to another as it guides an axon to its destination.

Although we have chosen these examples from the vertebrate, similar principles probably apply to the outgrowth of the first axons in insect nervous systems. For instance, the growth cones of certain pioneer axons in the grasshopper central nervous system always make a turn at a specific glial cell. If the glial cell is destroyed by a laser beam, the growth cones do not turn but continue in the original direction. These growth cones use glial cells as landmarks or guides, as do those of other pioneer neurons in both vertebrate and invertebrate central nervous systems. Evidence suggests very intimate contact between pioneer neurons and guidepost glial cells; gap junctions, for instance, form transiently between the two contacting cells.

Various Extracellular-Matrix Components Support Neuronal Outgrowth

The extracellular matrix and some of its individual constituents (e.g., laminin) have been shown to regulate neuronal outgrowth. The experiment described in Figure 23-32 indicates that growth cones select specific matrix substrates over others for outgrowth. Some researchers concluded from this finding that the pathway taken by a growth cone to its target is determined by a series of choice points between substrates of different adhesiveness. Subsequent studies, however, showed that no simple relationship exists between the degree of adhesiveness and growth-cone motility on a variety of substrates. Indeed, under certain conditions, laminin has anti-adhesive properties. Hence, it is probably more accurate to view outgrowth as reflecting the ability of various extracellular-matrix components (or components expressed on the surfaces of cells along the pathway) to promote motility through intracellular signaling pathways regulating cytoskeletal dynamics in the growth cone (e.g., integrin-laminin interactions).

Although many common extracellular-matrix components support growth-cone movement, they do not appear capable of determining the directionality of outgrowth. For instance, neurons in culture do not show directed outgrowth even on a steep gradient of laminin. Before discussing how directionality is determined, we consider the outgrowth of axons along other axon tracts.

Growth Cones Navigate along Specific Axon Tracts

As pioneer axons extend using extracellular-matrix and cell-surface cues, more and more of the space in the central nervous system becomes occupied by axonal processes. As additional neurons differentiate, their growth cones navigate on the surfaces of other axons and their axons eventually bundle together forming fascicles. Different growth cones select different axonal surfaces in different fascicles on which to migrate. This can be seen vividly in the development of the grasshopper central nervous system.

Grasshopper and Drosophila embryos are divided into segments, most of which have a similar pattern of neurons. Each segment has two ganglia, one on each side of the embryo, which contain the neuron cell bodies. Bundles of axons (longitudinal fascicles) run along either side of the embryo and bundles of axons (transverse fascicles) cross the embryo, connecting the segmental ganglia in both directions. Of the hundreds of neurons in each segmental ganglion, many sprout axons whose growth cones migrate to the opposite side of the embryo by growing along a transverse fascicle. Six such identifiable neurons are depicted in Figure 23-33a. When these six growth cones reach the opposite side, each chooses a different longitudinal fascicle to follow or a different direction in which to migrate.

For instance, the G neuron growth cone initially makes contact with about 100 axons in 25 longitudinal fascicles, but it only follows the A-P fascicle, which is composed of the axons of four neurons called A1, A2, P1, and P2. More detailed studies have shown that the G growth cone actually moves along only two of the four axons in the fascicle, P1 and P2 (Figure 23-33b). If the A1 and A2 neurons are experimentally destroyed by a laser beam, the differentiation of the G neuron as well as most of the other neurons is unaffected; the G growth cone moves normally along the P1 and P2 axons. However, if the P1 and P2 neurons are destroyed, the G neuron grows abnormally, and its growth cone behaves as if it were undirected and does not bind to any other axon. Thus, migration of the G growth cone relies absolutely on binding to the P axons.

These observations suggest that the P1 and P2 axons bear a unique surface marker that is recognized by a receptor present on the G growth cone but not on growth cones of other neurons that do not follow the P axons. To identify these guide molecules, investigators prepared monoclonal antibodies to preparations of grasshopper neuronal membranes and screened them for their ability to stain specific subsets of fascicles. With these anti-fascicle antibodies, several cell-surface fascicle proteins were identified and purified; the genes encoding these proteins then were identified and cloned using techniques described in Chapter 7. Remarkably, when four of these proteins—fasciclin I, II, III and neuroglian—were individually expressed in transfected nonadherent tissue-culture cells, they each promoted aggregation of cells expressing the protein, suggesting that they function as cell-adhesion molecules, or CAMs (Section 22.3).

To further characterize these proteins and study their role in neuronal outgrowth, their Drosophila homologs were identified. All are transmembrane proteins with a large extracellular domain; fly neuroglian and fasciclin II contain immunoglobulin-type repeats and fibronectin type III repeats, whereas fasciclin III contains only immunoglobulin repeats. Researchers then introduced various mutations that disrupted the structure of these proteins. Surprisingly, these mutations had little or no effect on the formation of specific axon fascicles in the fly CNS. Similar results have been obtained on N-CAM, the vertebrate homolog of fasciclin II, which has long been viewed as a critical determinant of neuronal development, including formation of fascicles. For instance, knockout mice that produce no functional N-CAM are viable and fertile, and their neuronal organization exhibits few abnormalities. However, recent studies have demonstrated that L1 (a vertebrate neuroglian-like cell-adhesion molecule) is required for the formation of nerve tracts in mice and human. Mice lacking L1 show defects in some but not all axon tracts. In humans, mutations in L1 lead to neurological deficits and defects in some nerve tracts.

In view of the multitude of CAMs identified on fascicles and the often mild effects of mutations in any one protein, it is likely that considerable redundancy exists in the mechanisms guiding movement of growth cones along other axons. Indeed, such redundancy has been directly demonstrated in antibody-disruption experiments with cultured chick embryonic cells. In this system, the interactions between three different CAMs and their receptors on growth cones must be neutralized to prevent outgrowth of ciliary ganglion cells over Schwann-cell membranes. The emerging view is that selection of specific fascicles by a particular growth cone involves recognition of a unique combination of widely expressed cell-surface molecules rather than unique fascicle-specific molecules.

Soluble Graded Signals Can Attract and Repel Growth Cones

So far we have seen that components of the extracellular matrix and certain cell-adhesion molecules participate in neuronal outgrowth. Growth cones can also respond to soluble factors. Small molecules such as neurotransmitters and various soluble peptides have been shown to modulate growth cone motility in vitro. These studies, as well as the observation that axons often travel long distances toward their targets, support the long-held notion that specific secreted molecules can act as long-range attractants. In the next section, we discuss the recent identification of specific proteins that act as graded signals to regulate the directionality of axon outgrowth.

SUMMARY

• During development, neurons elaborate remarkably specific and reproducible patterns of connections with other neurons (and with other target cells, e.g., muscle cells). Techniques for labeling individual neurons and their processes have been critical in investigating neuronal wiring.

• The leading edge of an elongating axon is a sensorimotor structure called a growth cone (see Figure 23-29).

• Elongation and migration of growth cones is promoted and guided by components of the extracellular matrix, cell-adhesion molecules (CAMs) on other neurons, and secreted, diffusible proteins.

• Each growth cone expresses a combination of cell-surface receptors on its surface that determines its pathway of choice. Ligand binding to these receptors is thought to trigger a signaling pathway that controls growth-cone motility and in some cases steers the growth cone toward its target.

• Graded diffusible signals that attract or repel growth cones are critical in determining the direction of neuronal outgrowth.