Expression of L1 and N-CAM cell adhesion molecules during development of the mouse olfactory system.

The expression of the neural adhesion molecules L1 and N-CAM has been studied in the embryonic and early postnatal olfactory system of the mouse in order to gain insight into the function of these molecules during development of a neural structure which retains neuronal turnover capacities throughout adulthood. N-CAM was slightly expressed and L1 was not significantly expressed in the olfactory placode on Embryonic Day 9, the earliest stage tested. Rather, N-CAM was strongly expressed in the mesenchyme underlying the olfactory placode. In the developing nasal pit, L1 and N-CAM were detectable in the developing olfactory epithelium, but not in regions developing into the respiratory epithelium. At early developmental stages, expression of the so-called embryonic form of N-CAM (E-N-CAM) coincides with the expression of N-CAM, whereas at later developmental stages and in the adult it is restricted to a smaller number of sensory cell bodies and axons, suggesting that the less adhesive embryonic form is characteristic of morphogenetically dynamic neuronal structures. Moreover, E-N-CAM is highly expressed at contact sites between olfactory axons and their target cells in the glomeruli of the olfactory bulb. L1 and N-CAM 180, the component of N-CAM that accumulates at cell contacts by interaction with the cytoskeleton are detectable as early as the first axons extend toward the primordial olfactory bulb. L1 remains prominent throughout development on axonal processes, both at contacts with other axons and with ensheathing cells. Contrary to N-CAM 180 which remains detectable on differentiating sensory neuronal cell bodies, L1 is only transiently expressed on these and is no longer detectable on primary olfactory neuronal cell bodies in the adult. Furthermore, whereas throughout development L1 has a molecular form similar to that seen in other parts of the developing and adult central nervous systems, N-CAM and, in particular, N-CAM 180 retain their highly sialylated form at least partially throughout all ages studied. These observations suggest that E-N-CAM and N-CAM 180 are characteristic of developmentally active structures and L1 may not only be involved in neurite outgrowth, but also in stabilization of contacts among fasciculating axons and between axons and ensheathing cells, as it has previously been found in the developing peripheral nervous system.

Development and plasticity of adrenal chromaffin cells: cues based on in vitro studies.

Neural crest derived precursors of the sympathoadrenal cell lineage give rise to two major cell types that differ in a number of morphological, ultrastructural, and biochemical characteristics: principal sympathetic neurons and chromaffin cells of the adrenal medulla. The present article reviews experimental studies performed on cultured adrenal medullary cells and designed to unravel the nature of epigenetic signals governing the developmental choice between the endocrine chromaffin and the neuronal sympathetic phenotype. Emphasis is placed on the role of glucocorticoids in initiation, development, and maintenance of the endocrine chromaffin phenotype and apparently antagonistic influences exerted by nerve growth factor (NGF) in vitro, resulting in the acquisition of neuronal properties by differentiated chromaffin cells. Experimental data from in vitro studies are compatible with the following conclusions. Glucocorticoids represent the decisive signal for the initial induction of endocrine differentiation. Moreover, high steroid hormone concentrations, as present in the adrenal medulla, are a prerequisite for the maturation of chromaffin cells. Even in a differentiated state, the endocrine phenotype is unstable in the absence of glucocorticoids, and the cells seem to reenter the neuronal developmental pathway. Under these conditions, cellular survival and differentiation into sympathetic neurons become NGF-dependent, as in normal sympathetic development. Thus, the effects of NGF survival, neurite outgrowth, and transmitter synthesis of cultured chromaffin cells probably do not reflect the induction of a specific phenotype, but they may be interpreted as a general neurotrophic support observable with other responsive cell types.

Mosaic Drosophila wings reveal regional heterogeneity in the guidance of ectopic axons.

In most studies of axon guidance in the peripheral tissues of insects, the ability of experimentally perturbed axons to pathfind was examined only along their normal pathways. This means that regions normally devoid of axons have not been sampled for their ability to influence axonal trajectories. To examine this question, we have induced the formation of single sensory neurons in a variety of abnormal locations in the developing wing of Drosophila and have examined the course taken by their axons. The axons of such ectopic neurons have a regionally varying tendency to grow in the normal, proximal direction. This proximal bias approaches 100% for neurons located in the distal part of vein L2 and 70% in distal vein L4 but falls to chance (50%) along vein L5. Thus, neurons forming in ectopic regions of the wing, especially those found near the normal axon pathways (veins L1 and L3), have a high probability of growing axons in the correct direction. We conclude that information relevant to axon outgrowth is not restricted to the normal pathways. Whether this information is intrinsic or extrinsic to the neurons, and why its strength shows such conspicuous regional variation, awaits further study.

Formation of the transverse nerve in moth embryos. II. Stereotyped growth by the axons of identified neuroendocrine neurons.

We are interested in the cellular mechanisms that guide neuroendocrine axons to their neurohaemal target regions and that regulate the extent and positioning of their terminal arbor. The neurohaemal organ we have studied is the segmentally repeated transverse nerve of the moth Manduca. In the mature animal, two motor neurons and a heterogeneous set of identified neuroendocrine neurons project to this nerve; the latter release hormonal peptides from along its length. In the preceding report, we demonstrated that during embryogenesis, the position, trajectory and extent of the transverse nerve are anticipated by two sets of nonneuronal cells, the strap and the bridge. In this paper we show that four identified neuroendocrine neurons (L1 and B1-3), like the identified motor neurons before them, elaborate growth cones that use this preexisting scaffolding as a substrate for axonal elongation. Moreover, growth cone navigation by these neuroendocrine neurons is as precise and invariant as that displayed by the motor neurons. One feature that differentiates the behavior of the developing neuroendocrine cells from that of the motor neurons is a stereotyped interaction that the L1 and B1-3 axons undergo with an identified syncytial cell that lies in close proximity to the strap. Each neuroendocrine neuron specifically adheres to the syncytium by extending numerous filopodia, and an occasional large lamellopodium, over its surface. These contacts are maintained by the neuroendocrine axons after their growth cones have left the vicinity of the syncytium and proceeded into the strap/bridge complex. Adhesion to the syncytium is transient and specific to the neuroendocrine neurons: although motor neuron axons are present at this same time and place, they display no affinity for the syncytium. This distinction correlates with the fact that the neuroendocrine neurons go on to elaborate arbor within the confines of the transverse nerve, while the motor neurons do not. We suggest that the syncytium may act as a "fictive target" for these neurons to aid in the differentiation of features that are specific to their cellular phenotype.

Axonal regeneration and growth direction in square-shaped mesothelial chambers.

A preformed rectangular mesothelial chamber was used as an experimental model to study possible neurotrophic and chemotactic effects on outgrowing axons. The experiments were performed on rats. The proximal end of a severed sciatic nerve was introduced into one corner of the chamber. In the opposite or the diagonally opposite corner the distal end of the nerve or a nerve transplant was introduced leaving a gap of about 10 mm between the proximal and distal nerve segments. In other series a silicon tube was introduced into a distal corner and some chambers were kept closed distally. After three and six months respectively the contents of the chambers were analysed, using histological and neurophysiological techniques. In all cases there was a preferential growth of regenerating nerve fibres towards and into the distal nerve tissue regardless of the position in the chamber. With a silicon tube or no tissue at all introduced distally the regenerating axons grew only 3-5 mm. These experiments indicate that nerve tissue has a neurotrophic and chemotactic influence on outgrowing axons.

Neurofascin: a novel chick cell-surface glycoprotein involved in neurite-neurite interactions.

We have identified neurofascin, a novel chick cell-surface glycoprotein involved in neurite-neurite interactions. Neurofascin is defined by its reactivity with monoclonal antibody (MAb) F6, which detects two polypeptides (160 and 185 kd) in immunotransfers of brain plasma membrane proteins. Immunoaffinity chromatography using immobilized MAb F6 yields major molecular mass bands at 185, 160, 135-110, and 92 kd. Fingerprint analyses show that these polypeptides are related. Neurofascin is expressed primarily in fiber-rich areas of embryonic cerebellum, spinal cord, and retina. Fab fragments of polyclonal antibodies to neurofascin interfere with the outgrowth of retinal and sympathetic axons in two different in vitro bioassays. Neurofascin is immunologically distinct from other known neurite-associated surface glycoproteins.

Human amnion membrane serves as a substratum for growing axons in vitro and in vivo.

The epithelial cell layer of human amnion membrane can be removed while the basement membrane and stromal surfaces remain morphologically intact. Such a preparation has been used as a substratum for the in vitro culture of dissociated neurons. Embryonic motor neurons from chick ciliary ganglion attached to both surfaces but grew extensive neurites only on the basement membrane. On cross sections of rolled amnion membranes, regenerating axons of cultured neurons were guided along pathways of basement membrane that were immunoreactive with an antibody to laminin. In addition, when rolled amnion membranes were implanted into a lesion cavity between the rat septum and hippocampus, cholinergic neurons extended axons through the longitudinally oriented implant into the hippocampus. Thus, this amnion preparation can serve as a bridge to promote axonal regeneration in vivo in damaged adult brain.

Regulation of motor axon sprouting.

Our studies on the amphibian and mammalian motor systems suggest that sprouting of intact motoneurons and synapse formation can be regulated by three mechanisms: peripheral, central, and transneuronal. Peripheral mechanisms provide the means of a direct mode of interaction between the periphery of the nerve cell and the target, to determine the extent of target innervation. The central mechanism enables target muscles to signal the cell bodies of their innervating motoneurons to regulate axonal growth and synapse formation, and thus again determine the extent of their innervation. The transneuronal mechanism provides a vehicle by which the pattern of innervation of a muscle can be altered by nerve cells that do not themselves innervate the muscle, but are an integral part of the entire system.

Morphology of retinogeniculate X and Y axon arbors in monocularly enucleated cats.

We examined the terminal arbors of single, physiologically identified retinogeniculate X and Y axons from the remaining retinas of adult cats raised from birth with monocular enucleation. These were compared with arbors of X and Y axons in normally reared cats. We used intra-axonal injections of horseradish peroxidase to label each axon after recording its response properties. While the axons in monocularly enucleated cats exhibited normal response properties, both X and Y axons in these cats had abnormally large terminal arbors. Each of the hypertrophied X arbors appeared to be completely confined to the single geniculate lamina A or A1 appropriate to its eye of origin (i.e., lamina A for the contralateral retina and lamina A1 for the ipsilateral retina). In contrast, in addition to their normal terminations, most of the Y arbors seemed to extend well into laminae normally innervated only by the retina that was removed. Thus most or all of the translaminar sprouting previously reported for monocularly enucleated cats appears to reflect extensions of Y axon arbors. These data, in addition to earlier, analogous data from young kittens and cats reared with monocular lid suture, suggest the following sequelae during postnatal development: the retinogeniculate X arbors mature first and develop exuberant arbors that are later competitively pruned as the Y axons expand their innervation of the lateral geniculate nucleus; monocular lid suture prevents the Y axons from succeeding in this competition, so they fail to establish normal arbors and cannot reduce the exuberant X arbors; monocular enucleation offers a less resistant path in the denervated laminae for the rapidly growing Y arbors from the remaining eye, and the expansion of these arbors there reduces the competitive pressure on the exuberant X arbors. Thus, in monocularly enucleated cats, sprouting is limited to Y axons, either because only they possess the capacity to sprout or because they are in the midst of a period of relatively rapid growth at the time of the neonatal enucleation. The X axon arbors are also abnormally large within their appropriate laminae. This occurs presumably because they are able to maintain their immature exuberance, although we cannot rule out the possibility that they are pruned and later regrow to the final size seen in our experiments.

Role of competitive interactions in the postnatal development of X and Y retinogeniculate axons.

The cat's retinogeniculate pathway is largely composed of X and Y axons, which represent two distinct neuronal streams organized in parallel. Our earlier data, summarized in the previous paper, suggest that the postnatal development of retinogeniculate axon arbors is characterized by competitive interactions between the X and Y axons. Thus, during development, X arbors in lamina A or A1 are initially broad or exuberant before the Y arbors begin to develop adultlike arbors; the X arbors then shrink to their adult form as the Y arbors grow and establish their mature complement of connections; monocular lid suture prevents the rapid growth of Y arbors, which in turn prevents the pruning of X arbors; and monocular enucleation at birth allows X arbors from the remaining eye to retain their exuberance although completely confined to their appropriate lamina A or A1, whereas the Y arbors develop aberrant extensions into adjacent, previously denervated laminae. We now provide additional evidence for the role of competition between retinogeniculate X and Y axons during development. The addition of visual deprivation by lid suture of the remaining eye to monocular enucleation at birth causes no apparent change in the morphology of X arbors in laminae A and A1. In contrast, the Y arbors of such cats continue to form extensive translaminar sprouts in the previously denervated laminae despite severely reduced terminations in the lamina A or A1 normally innervated by the remaining eye. We interpret these new data, in conjunction with our earlier data, as follows. If retinogeniculate X and Y arbors complete for synaptic space during postnatal development, terminations of Y axons can be affected by lid suture only in geniculate laminae where terminations of X axons are also present. Thus, Y axon arbors are severely reduced in deprived lamina A or A1 following lid suture whether or not the other eye is removed. Where X arbors are not present, such as in lamina C or the laminae inappropriate for the remaining eye after removal of the other, the lid suture has no obvious effect on development of the Y arbors.

Elimination of action potentials blocks the structural development of retinogeniculate synapses.

Although the influence of electrical activity on neural development has been studied extensively, experiments have only recently focused on the role of activity in the development of the mammalian central nervous system (CNS). Using tetrodotoxin (TTX) to abolish sodium-mediated action potentials, studies on the visual system show that impulse activity is essential both for the normal development of neuronal size and responsivity in the lateral geniculate nucleus (LGN), and for the eye-specific segregation of geniculo-cortical axons. There have been no anatomical studies to investigate the influence of action potentials on CNS synaptic development. We report here the first direct evidence that elimination of action potentials in the mammalian CNS blocks the growth of developing axon terminals and the formation of normal adult synaptic patterns. Our results show that when TTX is used to eliminate retinal ganglion-cell action potentials in the cat from birth to 8 weeks, the connections made by ganglion cell axons with LGN neurones, retinogeniculate synapses, remain almost identical morphologically to those in the newborn kitten.

Pioneer growth cone steering along a series of neuronal and non-neuronal cues of different affinities.

We have analyzed the morphology of over 5000 Ti1 pioneer growth cones labeled with anti-HRP, which reveals the disposition of axons, growth cone branches, and filopodia. Ti1 axon pathways typically consist of a sequence of 7 characteristically oriented segments, with a single, distinct reorientation point between each segment. Growth cones exhibit the same orientations and reorientations in a given region as do axon segments at later stages. The single, distinct reorientations suggest that growth cones make discrete switches between guidance cues as they grow. Ti1 growth cones are guided by various types of cues. A set of 3 immature identified neurons serves as nonadjacent guidepost cells and lies at the proximal end of 3 of the axon segments. To form another segment, growth cones reorient along a limb segment boundary within the epithelium. Growth cones also respond consistently to, and orient toward, a specific mesodermal cell, which may be a muscle pioneer. Thus, growth cones respond to at least 3 different types of cells in the leg. Ti1 growth cones exhibit a hierarchy of affinity for these cues. Guidepost neurons are the dominant cues in that contact with them reorients growth cones from guidance by the other types of cues. Growth cone branches are exclusively oriented to specific cues. Growth cones reorient by extending a branch directly to the cue of highest affinity and by withdrawing any branches that are extended to a cue of lesser affinity. A single filopodium in direct contact with a guidepost neuron can reorient a growth cone that still has multiple filopodia or even prominent branches specifically oriented to a previous cue of lesser affinity. These observations suggest that growth cone steering may not result simply from passive adhesion and filopodial traction, but may involve more active processes.

Axonal pathfinding in organ-cultured embryonic avian retinae.

Eye cups from stage 14-28 (E2 to E5) chick and quail embryos consisting of neural retina, lens, and vitreous body were cultured for 1 or 2 days. These eyes expanded by proliferation of the retinal cells and the surface areas of the retinae increased several-fold. The area covered by ganglion cells and axons also expanded in vitro. [3H]Thymidine labeling showed extensive proliferation of the neuroepithelial cells including the formation of new ganglion cells. Culturing eyes from embryos before stage 17 results, as in vivo, in the generation of the first ganglion cells of the retina, but unlike in the in vivo situation, the outgrowing axons always formed a random fiber net in the central portion of the retina. A defined axonal pattern identical to the in vivo developed only in specimens from embryos of stage 17 and older. Some aberrant axons, however, were also observed at the retinal periphery in specimens from embryos of more advanced stages (20-24), but only during the second day of culturing. Axons in retinae from embryos of stages 23 to 26 heading toward the optic fissure often crossed the fissure and, in contrast to the situation in vivo, invaded the opposite retinal side. These axons of wrong polarity followed the pathways of axons growing centripetally but in reverse direction. This suggests that the polarity of growing nerve fibers and their course are determined by different factors. Culturing the eyes of embryos from stages 20 to 25 in the presence of antibodies showed that the antibodies penetrated the entire retina with 6 hr. Neither anti-N-CAM nor the T-61 antibody--both recognizing membrane proteins of retinal cells--affected the growth of the eyes in vitro. The development of the axonal pattern in vitro was not affected by incubation with N-CAM-antibodies at concentrations up to 500 micron/ml, whereas the T-61 antibody which is known to block neurite extention in vitro (S. Henke-Fahle, W. Reckhaus, and R. Babiel (l984). "Developmental Neuroscience: Physiological, Pharmacological, and Clinical Aspects," pp. 393-398. Elsevier, Amsterdam/New York) showed inhibition of axonal growth in retina cultures at 50 micron/ml. These results indicate that the eye cultures can be used as a test system for antibodies against antigens which could be involved in axon extension and neurite pathfinding in situ.

Peripheral nerves in shiverer----dystrophic mouse chimeras: evidence that a non-Schwann cell component is required for axon ensheathment in vivo.

Peripheral nerves in dystrophic mice express multiple axon ensheathment abnormalities. If an intrinsic deficiency expressed within the Schwann cells themselves were to account for this neuropathy, then such cells, existing in a chimera preparation, would be expected to express the same ensheathment abnormalities, whereas a coexisting population of non-dystrophic Schwann cells should not be similarly affected. The genotype of myelinated Schwann cells in shiverer----dystrophic chimera was established with immunocytochemical techniques. Shiverer myelin lacks the P1 component of myelin basic protein (MBP), whereas dystrophic myelin appears to contain normal levels of MBP. No correlation between the ensheathment characteristics of the chimera spinal roots and the genotype of the local Schwann cell population was found; both dystrophic Schwann cell populations expressing normalized ensheathment characteristics and shiverer Schwann cells failing to respond to the local presence of naked axons were observed. These results require that a defective extra Schwann cell component is involved in the pathogenesis of the dystrophic neuropathy. Moreover, the normal realization of that component appears to be a necessary prerequisite for shiverer Schwann cells to achieve full ensheathment competence. Although a definitive identification of the cell type(s) that expresses the dy gene locus has not been achieved in this chimera preparation, the observations are consistent with defective endoneurial fibroblast function.

Injured neurons in the olfactory bulb of the adult rat grow axons along grafts of peripheral nerve.

Certain neurons in the central nervous system (CNS) of adult mammals extend axons for several cm along peripheral nerve grafts inserted into the brain or spinal cord. It is not clear, however, if these nerve cells constitute a special population or are examples of a general capacity of the injured mammalian CNS to regrow processes under these experimental conditions. Furthermore, because the new axons could originate by collateral sprouting from uninjured neurons, it is important to prove that the interruption of a central axonal projection can be followed by extensive fiber regrowth from the damaged neurons. In this anatomical study, we examined whether: (1) nerve cell type; and (2) axotomy, influence CNS axon regrowth along peripheral nerve grafts. For this purpose, we grafted segments of sciatic nerve into the olfactory bulb (OB) of adult rats and used combinations of neuroanatomical tracers (horseradish peroxidase and the fluorescent dyes True Blue and Nuclear Yellow) to investigate axonal regrowth from the different neurons that normally populate the OB. We demonstrate that OB axons extending along peripheral nerve grafts originate from mitral and tufted cells near the graft tip, rather than from the smaller OB neurons (periglomerular, short axon, and granule cells). Most of the mitral and tufted cells that extend new axons in grafted peripheral nerve segments lose their normal projections through the lateral olfactory tract because of axotomy at the time of grafting. Neuronal type, damage, and proximity to the graft appear to be prerequisites of this regenerative response from the OB.

Essentiality of a specific cellular terrain for growth of axons into a spinal cord lesion.

To date, there are no reports of growth of significant numbers of axons into or across a lesion of the mammalian spinal cord. However, recent studies showing that CNS axons will grow into PNS environments indicate that comparable growth into spinal cord lesions could be achieved if ischemic necrosis could be prevented and the lesion site repopulated by astrocytes and ependymal cells rather than by the macrophages, lymphocytes, and fibroblasts that generally accumulate at sites of CNS injury. To examine this possibility, we made a laminectomy at T5 in rats and crushed the spinal cord for 2 s with a smooth forceps (leaving the dura mater intact to prevent ingrowth of connective tissue). At 1 week, the lesion was filled with mononuclear cells, degenerating nerve fibers, and capillaries that were oriented parallel to the long axis of the spinal cord. By 2 weeks, longitudinally oriented cords of ependymal cells and astrocytes had migrated into the lesion from the adjacent spinal cord, and similarly oriented nerve fibers had begun to grow into the lesion along these capillaries and cellular cordons. The mononuclear cells had now assumed phagocytic activity and were engorged with myelin and other cellular debris. After 3 weeks, the astrocytes had elaborated thick cell processes. The nerve fibers in the lesion were still oriented longitudinally but had increased in number and were often arranged in small fascicles. These observations provide the first histological evidence of growth of nerve fibers into a lesion of the rat spinal cord. We conclude that the intrinsic regenerative capacity of the spinal cord can be expressed if ischemic necrosis and collagenous scarring are prevented and the spinal cord parenchyma is first reconstructed by its nonneuronal constituents.

Enhancement of axonal growth into a spinal lesion by topical application of triethanolamine and cytosine arabinoside.

We have demonstrated that a brief compression lesion of the rat spinal cord produces axotomy with minimal necrosis or scarring and that axons grow into such a lesion along longitudinally oriented capillaries and similarly oriented cordons of ependymal cells and astrocytes. Inasmuch as extensive, oriented growth of axons into a spinal lesion is never seen after transection, concussion, or other models of spinal cord injury, this new surgical procedure appeared to be applicable to the in vivo testing of pharmacological agents designed to promote neuritic outgrowth. The spinal cord of anesthetized rats was crushed extradurally for 1 s with a smooth jeweler's forceps. After 2 days when edema had subsided, the animals were reoperated. The dura mater was opened, and a polyethylene tube was implanted so that one end was fastened over the injury site and the other end was exteriorized at the back of the neck. The lesion site was superfused with 0.1 ml of control or test solutions four times daily for 2 weeks and then the animals were anesthetized and killed by vascular perfusion with fixative. After decalcification, the vertebral column and spinal cord were embedded in paraffin and stained by several histologic procedures including the protargol silver impregnation method for nerve fibers. Treatment with triethanolamine and cytosine arabinoside, substances which promote neuritogenesis in cultured spinal ganglia of chick embryos, markedly stimulated the growth of axons into the lesion of the rat spinal cord. We conclude (i) that it is possible to pharmacologically enhance the intrinsic growth capacity of CNS neurons and (ii) that brief compression provides a type of injury that is well suited to the evaluation of treatments aimed at promoting axonal regeneration.