Axonal cytoskeletal changes after non-disruptive axonal injury

In animal models of human diffuse axonal injury, axonal swellings leading to secondary axotomy occur between 2 and 6 h after injury. But, analysis of cytoskeletal changes associated with secondary axotomy has not been undertaken. We have carried out a quantitative analysis of cytoskeletal changes in a model of diffuse axonal injury 4 h after stretch-injury to adult guinea-pig optic nerves. The major site of axonal damage was the middle portion of the nerve. There was a statistically significant increase in the proportion of small axons with a diameter of 0.5 μm and smaller in which there was compaction of neurofilaments. Axons with a diameter greater than 2.0 μm demonstrated an increased spacing between cytoskeletal elements throughout the length of the nerve. However, in the middle segment of the nerve these larger axons demonstrated two different types of response. Either, where periaxonal spaces occurred, there was a reduction in axonal calibre, compaction of neurofilaments but no change in their number, and a loss of microtubules. Or, where intramyelinic spaces occurred there was an increased spacing between neurofilaments and microtubules with a significant loss in the number of both. Longitudinal sections showed foci of compaction of neurofilaments interspersed between regions where axonal structure was apparently normal. Neurofilament compaction was correlated with disruption of the axolemma at these foci present some hours after injury. We suggest that the time course of these axonal cytoskeletal changes after stretch-injury to central axons is shorter than those changes documented to occur during Wallerian degeneration.     

Ontogeny of the retina and optic nerve of Xenopus laevis: IV. Ultrastructural evidence of early ganglion cell differentiation

Ultrastructural evidence indicates that Xenopus retinal ganglion cell axons differentiate early, between stages 28 and 32. Light microscope studies indicated the presence of argryophilic material in the ventral retina and optic stalk of early embryos. Ultrastructural analysis of this region confirmed the presence of axons in the stalk and interstices of ventral retinal cells. Axons containing aligned microtubules and neurofilaments and elongated mitochondria with a paucity of other cell inclusions are found with increasing frequency in the ventral retina from stages 28 through $\text{33}\text{34}$. Central and dorsal regions of the retinas examined show little or no evidence of axons. A discrete, small bundle of axons is found in the optic stalk of stage 28 embryos and by stage $\text{30}\text{31}$ the number of axons in bundles has increased, suggesting early fasciculation. Between stages 28 and $\text{33}\text{34}$ (± 12 hr) extracellular space surrounding early axons diminishes and processes from neuroretinal cells in contact with axons surround developing axon bundles. The evidence presented suggests that axon initiation occurs in stages much earlier than previously reported. Other investigators have failed to detect ganglion cell differentiation prior to stage 32 possibly because they examined regions of the retina with few axons. Thus, experiments which rotate the retina in the orbit may have to be reevaluated since regenerating axons may use previously established pathways to organize and “home in” on tectal target cells.     

Microtubules and filaments in the axons and astrocytes of early postnatal rat optic nerves

Changes in the population of microtubules and filaments within the cytoplasm of maturing axons and astrocytes have been studied during the early postnatal development of rat optic nerves. At birth, all of the axons are unmyelinated; most have a diameter of 0.2–0.3 µ and contain 4–10 microtubules. Neurofilaments do not occur with any frequency until about 5 days postnatal when they appear as individual groups, each containing 4–12. Subsequently, the neurofilaments of each group disperse so that they become more evenly distributed in mature axons. Developing astrocytes show similar but rather more dramatic changes. Most astrocytic processes contain only microtubules at birth, but during maturation filaments begin to appear in increasing numbers while microtubules become less common. This process continues until, in the mature fibrous astrocytes, filaments pack the cytoplasm and microtubules are rare. These observations suggest that the filaments within axons and astrocytes may be formed by the breakdown of microtubules.     

Cell interactions and histogenesis in embryonic neural aggregates

Mixtures of cell suspensions from the optic lobe of 3-day-old and 6-day-old chick embryos form aggregates which show many ‘rosettes’ resulting from the invagination of peripheral ‘placodes’. However, in aggregates formed exclusively by optic lobe cells of either 3-day-old or 6-day-old embryos no rosettes are observed. By means of radioautographic studies of the combined aggregates, it was shown that rosettes are formed almost exclusively by cells from 3-day-old optic lobe, while those of 6-day-old embryos are predominantly found at the periphery of the aggregates. Cells from 3-day old optic lobe also form rosettes when cultured in combination with wing cells. Similar observations were done in aggregates formed by mixtures of 3-day-old optic lobe and 6-day-old liver cells. The cells from the optic lobe of 3-day-old embryos forming the concave surface of invaginating placodes and rosettes in the ‘combined’ aggregates appear polarized and wedge- or ‘bottle’-shaped. They are aligned in a highly ordered way, and abundant zonula adhaerentes, usually parallel to the cell major axis, are found at the cell apex where bundles of longitudinally oriented microtubules are present. The 3-day-old optic lobe cell which form the convex external surface of the ‘pure’ aggregates have a similar shape and organization, but in this situation apical cell bulges with randomly oriented microtubules interdigitate with those from neighbouring cells, and are joined by zonula adhaerentes usually perpendicular to the longitudinal axis of the cell.     

Gene replacement in mice reveals that the heavily phosphorylated tail of neurofilament heavy subunit does not affect axonal caliber or the transit of cargoes in slow axonal transport

The COOH-terminal tail of mammalian neurofilament heavy subunit (NF-H), the largest neurofilament subunit, contains 44-51 lysine-serine-proline repeats that are nearly stoichiometrically phosphorylated after assembly into neurofilaments in axons. Phosphorylation of these repeats has been implicated in promotion of radial growth of axons, control of nearest neighbor distances between neurofilaments or from neurofilaments to other structural components in axons, and as a determinant of slow axonal transport. These roles have now been tested through analysis of mice in which the NF-H gene was replaced by one deleted in the NF-H tail. Loss of the NF-H tail and all of its phosphorylation sites does not affect the number of neurofilaments, alter the ratios of the three neurofilament subunits, or affect the number of microtubules in axons. Additionally, it does not reduce interfilament spacing of most neurofilaments, the speed of action potential propagation, or mature cross-sectional areas of large motor or sensory axons, although its absence slows the speed of acquisition of normal diameters. Most surprisingly, at least in optic nerve axons, loss of the NF-H tail does not affect the rate of transport of neurofilament subunits.     

Neurochemical Characteristics of Myelin-like Structure in the Chick Retina

Abstract: Certain characteristics of myelin-like structures in the chick retina were examined morphologically and biochemically. Developmental changes of 2', 3'-cyclic nucleotide 3'-phosphohydrolase (CNPase) in the chick retina and optic nerve were examined. The measurable activity in the retina was first detected at 16 days of incubation and thereafter, it increased rapidly until 4 weeks post-hatching. By contrast, CNPase activity in the optic nerve reached the maximum level at 4 days post-hatching and maintained a constant level thereafter. The purifed myelin fraction from the chick retina showed higher activity of CNPase, whereas its activity in the retinal homogenate was very low. Hence, it was considered that the myelin fraction from the chick retina is similar to that of CNS myelin with respect to CNPase. Protein profiles of the purified myelin fractions isolated from the chick optic tectum, optic nerve, retina and sciatic nerve were analysed by SDS-polyacrylamide gel elec-trophoresis. Myelin fractions from the chick optic tectum and optic nerve contained basic protein (BP) and Folch-Lees proteolipid protein (PLP). Myelin fraction from the chick sciatic nerve contained BP, P2 and two glycoproteins (PO and 23K). In contrast, retinal myelin fraction contained only BP. PLP, PO, 23K and P2 proteins were definitely undetectable. Electron micrographs revealed that some axons in the optic nerve fiber layer of the chick retina were wrapped by a spiral-structured myelin-like sheath, which showed some differences from those of CNS and PNS myelin sheaths. It was suggested that the origin of the myelin-like structure in the chick retina is other than from oligodendroglia or Schwann cells.     

Contribution of cytoskeletal elements to the axonal mechanical properties

Background

Microtubules, microfilaments, and neurofilaments are cytoskeletal elements that affect cell morphology, cellular processes, and mechanical structures in neural cells. The objective of the current study was to investigate the contribution of each type of cytoskeletal element to the mechanical properties of axons of dorsal root and sympathetic ganglia cells in chick embryos.

Results

Microtubules, microfilaments, and neurofilaments in axons were disrupted by nocodazole, cytochalasin D, and acrylamide, respectively, or a combination of the three. An atomic force microscope (AFM) was then used to compress the treated axons, and the resulting corresponding force-deformation information was analyzed to estimate the mechanical properties of axons that were partially or fully disrupted.

Conclusion

We have found that the mechanical stiffness was most reduced in microtubules-disrupted-axons, followed by neurofilaments-disrupted- and microfilaments-disrupted-axons. This suggests that microtubules contribute the most of the mechanical stiffness to axons.

     

Microtubule-associated protein 2 within axons of spinal motor neurons: associations with microtubules and neurofilaments in normal and beta,beta'-iminodipropionitrile-treated axons

We have examined the distribution of microtubule-associated protein 2 (MAP2) in the lumbar segment of spinal cord, ventral and dorsal roots, and dorsal root ganglia of control and beta,beta'-iminodipropionitrile- treated rats. The peroxidase-antiperoxidase technique was used for light and electron microscopic immunohistochemical studies with two monoclonal antibodies directed against different epitopes of Chinese hamster brain MAP2, designated AP9 and AP13. MAP2 immunoreactivity was present in axons of spinal motor neurons, but was not detected in axons of white matter tracts of spinal cord and in the majority of axons of the dorsal root. A gradient of staining intensity among dendrites, cell bodies, and axons of spinal motor neurons was present, with dendrites staining most intensely and axons the least. While dendrites and cell bodies of all neurons in the spinal cord were intensely positive, neurons of the dorsal root ganglia were variably stained. The axons of labeled dorsal root ganglion cells were intensely labeled up to their bifurcation; beyond this point, while only occasional central processes in dorsal roots were weakly stained, the majority of peripheral processes in spinal nerves were positive. beta,beta'- Iminodipropionitrile produced segregation of microtubules and membranous organelles from neurofilaments in the peripheral nervous system portion and accumulation of neurofilaments in the central nervous system portion of spinal motor axons. While both anti-MAP2 hybridoma antibodies co-localized with microtubules in the central nervous system portion, only one co-localized with microtubules in the peripheral nervous system portion of spinal motor axons, while the other antibody co-localized with neurofilaments and did not stain the central region of the axon which contained microtubules. These findings suggest that (a) MAP2 is present in axons of spinal motor neurons, albeit in a lower concentration or in a different form than is present in dendrites, and (b) the MAP2 in axons interacts with both microtubules and neurofilaments.     

Metabolism of 4-hydroxyaminoquinoline-1-oxide and its binding to DNA in chick embryos

The metabolic pathway of 4-hydroxyaminoquinoline-1-oxide (4HAQO) and its binding to DNA was studied in 2-day chick embryos administered [G-3H]4HAQO in a shell-less culture. The 4HAQO rapidly metabolized into non-carcinogenic compounds and 1 h after administration only very small amounts of free 4HAQO could be detected in the embryo cells. The amount of DNA-bound 4HAQO in the embryo cells reached a maximum 2 h after administration, then began to decrease. The maximum extent (mu mol/mol P of nucleotide) was 18.2, equivalent to 1 molecule of 4HAQO-purine adducts per 2.8 X 10(4) base pairs of DNA. It was possible to detect removal of 4HAQO-purine adducts from DNA in chick embryo cells in a shell-less culture. A dose-response relationship for the killing effect of 4HAQO on 2-day embryos was observed in the range of 0.24-24 nmol 4HAQO per embryo. The practicality of the present method of administration of 4HAQO for 'flash administration' of compounds to chick embryo and the advantages of the shell-less culture method which provides access for biochemical and developmental studies of chick embryos were also discussed.     

[32P]orthophosphate and [35S]methionine label separate pools of neurofilaments with markedly different axonal transport kinetics in mouse retinal ganglion cells in vivo

Newly synthesized neurofilament proteins become highly phosphorylated within axons. Within 2 days after intravitreously injecting normal adult mice with [³²P]orthophosphate, we observed that neurofilaments along the entire length of optic axons were radiolabeled by a soluble³²P-carrier that was axonally transported faster than neurofilaments.³²P-incorporation into neurofilament proteins synthesized at the time of injection was comparatively low and minimally influenced the labeling pattern along axons.³²P-incorporation into axonal neurofilaments was considerably higher in the middle region of the optic axons. This characteristic non-uniform distribution of radiolabel remained nearly unchanged for at least 22 days. During this interval, less than 10% of the total³²P-labeled neurofilaments redistributed from the optic nerve to the optic tract. By contrast, newly synthesized neurofilaments were selectively pulse-labeled in ganglion cell bodies by intravitreous injection of [³⁵S]methionine and about 60% of this pool translocated by slow axoplasmic transport to the optic tract during the same time interval. These findings indicate that the steady-state or resident pool of neurofilaments in axons is not identical to the newly synthesized neurofilament pool, the major portion of which moves at the slowest rate of axoplasmic transport. Taken together with earlier studies, these results support the idea that, depending in part on their phosphorylation state, transported neurofilaments can interact for short or very long periods with a stationary but dynamic neurofilament lattice in axons.Special issue dedicated to Dr. Sidney Ochs.     

Chemical and structural changes of neurofilaments in transected rat sciatic nerve

The sequence of changes occurring in transected rat sciatic nerve was examined by electron microscopy and by sodium dodecyl sulfate (SDS) polyacrylamide disc gel electrophoresis. Representative segments of transected nerves were processed for ultrastructural examinations between 0 and 34 days after the transection of sciatic nerves immediately below the sacro-sciatic notch. The remainder of the transected nerves and the intact portions of sciatic nerves were desheathed and immediately homogenized in 1 percent SDS containing 8 M urea and 50 mM dithioerythritol. Solubilized proteins were analyzed on 12 percent gels at pH 8.3 in a discontinuous electrophoretic system. Initial changes were limited to the axons of transected nerve fibers and were characterized by the loss of microtubules and neurofilaments and their replacement by an amorphous floccular material. These changes became widespread between 24 and 48 h after transection. The disruption of neurofilaments during this interval occurred in parallel with a selective loss of 69,000, 150,000 and 200,000 mol wt proteins from nerve homogenates, thus corroborating the view that these proteins represent component subunits of mammalian neurofilaments. Furthermore, the selective changes of neurofilament proteins in transected nerves indicate their inherent lability and suggest their susceptibility to calcium-mediated alterations. Electrophoretic profiles of nerve proteins during the 4-34-day interval after nerve transection reflected the breakdown and removal of myelin, the proliferation of Schwann cells and the deposition of endoneurial collagen. A marked increase of intermediate-sized filaments within proliferating Schwann cell processes was not accompanied by the appearance of neurofilamentlike proteins in gels of nerve homogenates.     

Retrograde axonal transport in lateral motor neurons of the chick embryo prior to limb bud innervation

Horseradish peroxidase was injected into the distal limb buds of 3-, 3.5-, and 4-day chick embryos and was allowed to diffuse into the path of outgrowing axons. In the majority of the embryos injected, reaction product was found in cells of the ventral spinal cord ipsilateral to the injection site. The reaction product in the most clearly stained cells appeared smooth and diffuse rather than the typical granular appearance found in later embryos and adults. A granular background was seen in 4-day embryos, however, although distinct granular cell outlines were infrequent. This study indicates that retrograde axonal transport is a very early cellular feature during neurogenesis, demonstrable almost from the inception of neurite outgrowth. It is speculated that this transport capacity might function in relaying positional information from growing fiber tips back to cell bodies to aid in the formation of specific synaptic connections or in guiding directed axonal growth, or it may provide a means for trophic interactions vital to the survival of the young neuroblast.     

The morphology of the Limulus visual system

Summary The lateral optic nerve of Limulus polyphemus, the horseshoe crab, contains 4 types of axons, which originate from eccentric cells, retinula cells, rudimentary eye cells, and from unidentified cells in the brain that give rise to the efferent fibers. Though small in diameter in a young animal, the eccentric cell axons in the adult grow to the same size as the rudimentary eye axons, which are originally the largest fibers in the nerve of the small Limulus. Cytoplasmic content, particularly the orderly distribution of microtubules, is identical in the three types of visual fibers. The segregation of rudimentary eye axons into a separate grouping within the optic nerve in small animals gives way to a homogeneous distribution in the adult. Interrupting the optic nerve leads to a proximal pile-up of secretory granules in a few fibers. The identity of these granules with those in the synaptoid terminations of photoreceptors establishes these fibers as efferent. The same operation leads to a conspicuous hypertrophy of subsurface cisternae within retinula cell axons.This study constitutes Publication No. 483 from the Oregon Regional Primate Research Center, supported by Grants FR00163 and EY 00392 from the National Institutes of Health and by a Bob Hope Grant-in-Aid by Fight-for-Sight, Inc., New York City.The author wishes to thank Mrs. Audrey Griffin for patient and excellent technical assistance.     

Fine structure of the retino-optic nerve junction in the chicken

The aim of this study is to investigate a fine structure of the retino-optic nerve junction in the chicken. We especially focused on the myelin sheaths and astrocytes in the intraocular optic nerve (ION) and its adjoining parts. A part of the axons of retinal nerve fiber layer (NFL) were myelinated. Ganglion cell axons were ensheathed by loose myelin in the NFL and by a compact one in the ION and optic nerve (ON). Myelin structure changed from loose type to a compact one within the very narrow NFL-ION junction. Loose myelin forming cells are dark type of oligodendrocytes in the retina. From the most peripheral ON to the choroidal part of ION, astrocytes contained abundant microtubules. The optic nerve around the lamina cribrosa is exposed to mechanical force during eye movement. It is suggested that these microtubules may perform the cytoskeletal function. Astrocytes in the retinal part of ION had longer processes filled with abundant gliofilaments. They may provide the mechanical support for the ganglion cell axons, which are exposed directly to intraocular pressure. Although astrocytes in the retinal level of ION extended their processes into the retina, their soma was never found in the retina.     

Developmental changes in phosphorylation state of neurofilament proteins in the chick embryonic optic nerve

Developmental changes in the phosphorylation state of neurofilament proteins (NFPs) in the chick embryonic optic nerve were histochemically and biochemically studied using monoclonal antibody (MAb) 82E10 specific to the highly phosphorylated components of high (180K)- and middle (160K)-molecular-weight subunits of neurofilament (NF) in the chicken. Cross sections of developing embryonic optic nerve were studied by enzyme immunohistochemistry using this MAb. The staining pattern showed marked changes with the developmental stage. In 6-day embryos (E6) the entire cross section was stained, whereas in E10 only about a ventroposterior half of the cross section was stained. In E14 nearly the entire area of the cross section became unstained. Thereafter, the immunoreactivity reappeared and gradually increased, such that in E20 the entire cross section became immunopositive again. Electrophoretic and immunoblot analyses were made on optic nerves dissected out of embryos of various stages. The 82E10 immunoreactivity at the position of NF-M underwent a transient loss in E14 in parallel with the time course of histochemical change. Two-dimensional gels stained for protein further showed that the highly phosphorylated form of NF-M is transiently lost from embryonic optic nerve in E14, while the less phosphorylated form persists throughout the embryonic developmental stages. In order to understand the orderly loss of the 82E10 immunoreactivity in relation to retinotopic and chronotopic organizations of the fibers in the embryonic optic nerve, retinal injection of a fluorescent dye DiI as an anterograde tracing marker for selected fibers was utilized. An ordered arrangement of the fibers was present within the embryonic optic pathway, suggesting that the orderly loss of the 82E10 immunoreactivity in the embryonic optic nerve reflects the chronological order of the optic axons. These changes in the phosphorylation state of NFPs in the embryonic optic nerve presumably reflect dynamic changes of the neuronal cytoskeleton at certain stages during development.     

The fine structure of the vertical lobe of octopus brain

Although much is known about the structural organization and connexions of the various lobes of the octopus brain from light microscopy, this is the first attempt at a detailed analysis of one of the lobes- the vertical lobe, with the electron microscope. The vertical lobe consists of five lobules. The median superior frontal (MSF) axons enter each lobule from the MSF lobe. The MSF axons contain both microtubules and neurofilaments. The varicosities of the MSF axons contain both agranular and dense-cored vesicles and synapse with trunks of the amacrine cells. These trunks run together in bundles termed amacrine tracts into the centres of the lobules. The amacrine trunks contain microtubules but no neurofilaments. The trunks contain large and small agranular synaptic vesicles and synapse with what are in all probability branches of the trunks of the large cells. These trunks contain microtubules but no neurofilaments. They run out through the bases of the lobules probably without forming synaptic contacts within the lobule. Fibres signalling 'pain' (nocifensor) enter the lobules from below. They can be recognized by their content of neurofilaments. Their terminals contain numerous very small synaptic vesicles and a few larger and dense-cored ones. These 'pain' fibres appear to synapse mostly with processes of the large cells. J. Z. Young has shown that the vertical lobe is especially concerned with the integrative action of the visual system, linked with the chemo-tactile system. Electron microscopy supports Young's suggestion that the superior frontal and interconnected vertical lobe systems constitute a loop which could sustain a positive feed-back mechanism (MSF -- amacrine -- large cell -- lateral superior frontal -- MSF) while the 'pain' (nocifensor) input could exert a suppressor (inhibitory) effect on the loop by its action on the large cells.     

Axon growth in embryonic chick and quail retinal whole mounts in vitro

Whole retinae of 4- to 10-day-old chick and quail embryos were spread on membrane filters and kept in culture for up to 4 days. Axon growth during culture was demonstrated by silver staining, anterograde labeling of fibers with RITC, time-lapse recording, and SEM. Fiber growth was observed in specimens from chick embryos up to 7 days old, with a growth maximum at E6 and from quail embryos up to E6 with the maximum at E5. Newly growing axons followed the optic fiber pattern already existing and, like axons in vivo, grew predominantly toward the optic fissure. Directional and orientational adaptation of newly growing axons to the preexisting fibers increased with the donor age. Retinae from donors up to E5 in chick and up to E4 in quail showed a high proportion of axons which crossed the optic fissure during the culture period and invaded the opposite retinal fiber layer. These fibers showed a correct radial orientation while growing in the opposite direction to normal. Likewise, in cultures from these young donors some fibers grew out initially in the diametrically opposite direction to normal toward the tissue periphery. Since all of the wrongly directed axons grew at the same rate as normal and adapted correctly to the already formed axon pattern, this suggests independent signals for the direction and orientation of growing fibers. Treatment of mounted retinae with collagenase or trypsin removed the vitreal retinal surface, leaving the existing axon pattern intact. Subsequently, new axons grew profusely in culture, but lost both their orientational and directional characteristics.     

The Prosencephalon Has the Capacity to Differentiate into the Optic Tectum: Analysis by Chick-Specific Monoclonal Antibodies in Quail-Chick-Chimeric Brains

The alar plate of the prosencephalon of the quail embryo was heterotopically transplanted into the alar plate of the mesencephalon of the chick embryo at the 7–10 somite stage. Chick and quail cells in chimeric brains were distinguished after Feulgen-Rossenbeck staining and/or immunohistochemical staining with a species specific monoclonal antibody MAb-37F5 which recognized cytoplasmic components of chick brain cells. Neural connections between the transplant and the host were studied by monoclonal antibodies, MAb39-B11, which recognizes a species specific antigen on chick nerve fibers, and MAb-29B8, which reacts to 160 kD neurofilaments of both chick and quail.
When the transplant was completely integrated into the host mesencephalon, the transplant developed a laminar morphology closely resembling that of the optic tectum. Immunohistochemical staining with MAb-39B11 showed that the host optic nerve fibers innervated both the host tectum and the tectum-like transplant. However, optic nerve fibers did not invade transplants that failed to develope a laminar structure characteristic of the tectum. These findings suggest that the prosencephalon has a capacity to differentiate into the optic tectum at the 7–10 somite stage.     

ULTRASTRUCTURE AND FUNCTION OF GROWTH CONES AND AXONS OF CULTURED NERVE CELLS

Dorsal root ganglion nerve cells undergoing axon elongation in vitro have been analyzed ultrastructurally. The growth cone at the axonal tip contains smooth endoplasmic reticulum, vesicles, neurofilaments, occasional microtubules, and a network of 50-A in diameter microfilaments. The filamentous network fills the periphery of the growth cone and is the only structure found in microspikes. Elements of the network are oriented parallel to the axis of microspikes, but exhibit little orientation in the growth cone. Cytochalasin B causes rounding up of growth cones, retraction of microspikes, and cessation of axon elongation. The latter biological effect correlates with an ultrastructural alteration in the filamentous network of growth cones and microspikes. No other organelle appears to be affected by the drug. Removal of cytochalasin allows reinitiation of growth cone-microspike activity, and elongation begins anew. Such recovery will occur in the presence of the protein synthesis inhibitor cycloheximide, and in the absence of exogenous nerve growth factor. The neurofilaments and microtubules of axons are regularly spaced. Fine filaments indistinguishable from those in the growth cone interconnect neurofilaments, vesicles, microtubules, and plasma membrane. This filamentous network could provide the structural basis for the initiation of lateral microspikes and perhaps of collateral axons, besides playing a role in axonal transport.