Cell adhesion molecules influence axonal outgrowth

During development, axons extend from the neuronal cell body and grow, frequently over long distances, in order to make very precise connections with a target. Despite the complexity of the nervous system, where over 10¹² axons must find their targets, the pattern of axonal outgrowth for any one axon is highly reproducible from one individual to another. The flattened tip of a growing axon is called the growth cone, which resembles the palm of the hand, with processes extending from it like fingers (see Chaps. 8 and 27). It is the growth cone that responds to the environment and determines what direction the growing axon will take. Environmental cues that the growth cone encounters can be either fixed or diffusible, and it is the integration of these signals that determines the final direction. The majority of fixed signals are CAMs, expressed by glia, other cells or older axons that have already traversed that particular pathway. In addition, growth cones use integrins to select a pathway of ECM molecules on which to extend processes.

This has been demonstrated from both tissue culture studies and studies in vivo [23–25]. In culture, members of all three families of adhesion molecules found in the nervous system have been shown to promote axonal growth. This was carried out by transfecting the cDNA for the molecule to be tested in a cell line that does not normally express it, usually a fibroblast cell line. Isolated neurons are then grown on a monolayer of the CAM-expressing cells, and neurite length is compared to that obtained by contact with cells not expressing the adhesion molecule. Experiments like these have shown that CAMs, including the cadherins, the integrins and members of the Ig family, in particular NCAM and L1, are very potent promoters of axonal growth from a variety of primary neurons. The same neuron has been shown to respond to different CAMs, indicating that there is not any one unique “CAM cue” for a particular neuron. Instead, it is likely that a variety of CAMs contribute to the effect on axonal growth. As well as being able to promote axonal growth when fixed in the ECM, some CAMs, such as L1, have been shown to promote axonal growth when added in a soluble form to neurons in culture. This strengthens the suggestion that CAMs are not just “sticky” molecules but can exert their effects by activation of a signal-transduction pathway. Indeed, a number of soluble forms of various CAMs, including L1, NCAM, MAG and cadherins, have been found in the extracellular milieu of living cells. It remains to be determined whether these soluble CAMs in fact influence axonal growth in vivo.

In addition to CAM-expressing fibroblasts, another more physiological cell substrate has been used to demonstrate the effects of CAMs on axonal growth in culture. When dorsal root ganglion neurons are grown on a monolayer of nonmyelinating Schwann cells, which are very permissive for growth in vivo when not synthesizing myelin proteins, there is robust axonal outgrowth. Only when a combination of antibodies against NCAM, L1, cadherins and integrins was added to the cultures was outgrowth significantly reduced. Together, these experiments indicate that axonal outgrowth at any one time can be influenced by several CAMs.

This concept is supported by observations in vivo. Knockout mice lacking NCAM expression are viable and display subtle abnormalities in the nervous system. These abnormalities include a reduction in the size of the olfactory bulb, which is attributed to a decrease in migration of the neuronal cells that form this structure. Possibly, the inability to migrate normally reflects an increase in cellular adhesivity, due to the absence of the highly sialylated form of NCAM, which in the wild-type or normal animal tends to repel apposed membranes on which it is expressed. Also, a reduction in the density of mossy fibers in the hippocampus has been noted in the NCAM knockout mouse, revealing a defect in axonal growth. The subtle phenotype in these mice, relative to the dramatic effect of NCAM on axonal growth in culture, suggests that there may be a certain amount of redundancy in nervous tissue, whereby another CAM or combination of CAMs can in part compensate for the absence of NCAM.

Cell adhesion molecules are responsible for axonal fasciculation

Usually, when growing toward their target, axons fasciculate (Fig. 7-6) and grow in bundles. However, there are points along the pathway where axons must defasciculate and different axons must take different paths. Hence, in addition to promoting axonal outgrowth, CAMs can affect the direction an axon takes at key decision points along the way. One model that has been proposed for explaining how an axon is selected to remain fasciculated with respect to other axons or to separate from the bundle and grow alone is based on the relative adhesiveness of the substrate vs. the adhesiveness of other axons. This model was proposed based on observations of adhesion and axonal growth with two forms of NCAM: one form that carries the highly charged sialic acid and polysialic acid (PSA) and another form with very little or no PSA. It has been suggested that the high negative charge on PSA—NCAM renders it less adhesive than its unsialylated counterpart [26]. Because the negatively charged sialic acid moieties tend to sequester water, occupy a large volume in the extracellular space and strongly repel one another across the bilayer, they probably keep plasma membranes sufficiently far apart so as to prevent other adhesion molecules, such as L1 and the cadherins, from getting close enough to interact. Following from this, fasciculated axons, which adhere tightly to each other, express the adhesive form of NCAM that does not carry PSA. In contrast, the PSA form of NCAM is more effective than the unsialylated form at promoting axonal growth.

Figure 7-6

Neural cell adhesion molecule (NCAM) carries different levels of polysialic acid. A: NCAM is depicted on a growing axon, without polysialic acid. Under these conditions, NCAM and L1 molecules interact homophilically and axons fasciculate. B: NCAM carries (more...)

As stated above, the complexity and precision of connections within the nervous system would appear to require a very rigid set of rules or cues. The ubiquitous expression of individual CAMs throughout the nervous system and during development implies that they could not provide the specificity required for such a precise signal for individual neurons. However, the full repertoire of CAMs used in vivo for any single axon to reach its target has yet to be characterized. It is likely, as suggested from both in vivo and in vitro results, that many CAMs affect the growth of any one axon. In addition, as demonstrated by studies with different isoforms of NCAM, that is, with and without the VASE sequence or with and without PSA, it is highly likely that subtle differences in isoform expression of any single CAM can have dramatic effects on axonal response. What has been determined regarding the role of adhesion molecules in axonal guidance during development to date is most likely an underestimate of their full involvement.

Cell adhesion molecules may also function in regeneration

As well as growing during development, axons grow and regenerate under certain conditions after injury [27]. Axonal regeneration is often successful in the PNS, where it may be accompanied by full restoration of function. In contrast, in the mammalian CNS, there is usually little or no regrowth and function is often lost. The main difference between axonal growth during development and during regeneration is in whether the damaged axon is allowed by the existing conditions to regrow, not if it is able (Chap. 29). In the PNS, after injury, damaged tissue, consisting mostly of myelin debris, is cleared away by macrophages. Coincident with this event, Schwann cells downregulate the expression of the myelin-specific proteins and upregulate a number of growth-promoting adhesion molecules, including L1 and NCAM. These Schwann cells now resemble Schwann cells in the developing nervous system and are very permissive for axonal growth (Fig. 7-7). In contrast, myelin is not cleared from the CNS after injury and the myelin-forming cells in the CNS, oligodendrocytes, continue to express myelin proteins and do not upregulate any growth-promoting adhesion molecules. There is now substantial evidence to support the idea that the presence of myelin and the absence of growth-promoting molecules are two factors largely responsible for the lack of regeneration in the mammalian CNS [28]. Myelin membranes have been shown to be inhibitory for regeneration both in vitro and in vivo, and a number of myelin-specific molecules have been shown to be potent inhibitors of axonal growth. One of these is MAG, which is a member of the siglec subfamily. MAG can promote as well as inhibit axonal growth depending on the age and type of neuron. The ability of MAG to inhibit or promote regeneration when presented to the neuron either as a substrate or in a soluble form strongly suggests that MAG binds to a neuronal receptor and activates a signal-transduction mechanism in the neuron that effects changes in axonal growth [29,30].

Figure 7-7

A simplified model for regeneration. A: Transection of a peripheral nerve causes degeneration. B: Schwann cells divide and dedifferentiate. Myelin debris is cleared by the Schwann cells and macrophages. C: The axon regenerates by growing over the growth-permissive (more...)