Axon-glial relationships in early CNS–PNS transitional zone development: an ultrastructural study

The CNS–PNS transitional zone of rat cervical ventral rootlets develops in two stages: first, axon segregation, then transitional node formation. This ultrastructural study examines the former. Material was prepared by standard methods. Shortly after they grow out from the neural tube, ventral motoneuron axon bundles are extensively segregated by a matrix of fine processes forming a barrier across the rootlet, just distal to the cord surface. These processes arise from cell clusters on the rootlet surface. This barrier is prominent until the period around birth, when it is replaced by a second in which the axons are completely segregated from one another. The perikarya and processes forming this barrier resemble those of the first, but lie at or just below the cord surface. Thus, beginning at the earliest stage, a barrier crosses the axon bundle and segregates its axons before axon segregation is advanced either in the PNS or (especially) in the CNS. This may prevent central Schwann cell migration. Evidence is presented suggesting that the second barrier may arise through a relative proximal relocation of the first, as the cord grows radially. Near the cord surface, a complete, funnel-shaped sleeve of glial processes surrounds the axon bundle. This is continuous at the cord surface with the glia limitans. It constitutes an integral part of the transitional zone apparatus. It is also continuous centrally with the sheath which enfolds the bundle of ventral motoneuron axons as they run between the ventral horn and the transitional zone. Axon segregation at the cord surface, and therefore the formation of the definitive astrocytic CNS–PNS barrier occur relatively (and perhaps surprisingly) late at the cord surface. The definitive sharp discontinuity of central and peripheral tissue types characteristic of the transitional zone is established only after birth.     

The ventral motoneurone axon bundle in the CNS — a cordone system?

In the developing CNS neighbouring structures are commonly separated by transient barriers termed cordones, some of which coincide with glial elements. Where ventral motoneuron axons cross the spinal white matter as intramedullary bundles to reach the CNS-PNS transitional zone they are surrounded from early development by a glial sleeve resembling a cordone. This becomes better developed with age and, like some cordones, persists into adult life. This could provide a radial conduit which might underlie the capacity of central segments of mature ventral motoneurone axons to regenerate. It may also provide a pathway for glial migration from the central cord to more superficial levels, including the transitional zone, where they help form the CNS-PNS barrier. Axons in the intramedullary bundle and in the surrounding ventral white column mature at different rates. Glial sleeve cells of the intramedullary bundles are apposed to both. Morphometric analysis of the axon-glial relationships of the two populations indicates that glial development proceeds at a different rate in relation to each axon class and that this is influenced by the degree of axonal maturation, which may in turn be related to target contact. Furthermore, early axon glial relationships differ between the two populations. For ventral motoneurone axons these take place in two stages: firstly, glial segregation of axons (resembling that in the PNS) and secondly, oligodendrocytic contact and ensheathment, which leads on to myelination. Axon-glial relationships in the ventral white column begin with the second of these events, as is more typical of early CNS myelination in general.     

Three markers of adult non-myelin-forming Schwann cells, 217c(Ran-1), A5E3 and GFAP: development and regulation by neuron-Schwann cell interactions

Immunohistochemical methods are used to investigate in detail the development and regulation of three proteins (217c(Ran-1), A5E3 and GFAP) specifically associated with adult non-myelin-forming Schwann cells in the rat sciatic nerve, from embryo day 15 to maturity. 217c(Ran-1), which is probably the NGF-receptor, and A5E3 are expressed by the majority of cells in the nerve at embryo day 15 and by essentially all cells at embryo day 18. GFAP first appears at embryo day 18; this is an intrinsically programmed developmental event which occurs in cultured Schwann cells even in the absence of serum. Postnatally, the expression of 217c(Ran-1), A5E3 and GFAP is suppressed in cells that form myelin but retained in non-myelin-forming Schwann cells. Mature myelin-forming cells nevertheless maintain the potential to express all three proteins but will only do so if removed from contact with myelinated axons. In neuron-free cultures Schwann cells express all three proteins. This work, together with our previous observations on N-CAM, shows that removal of a diverse set of surface proteins and a change in intermediate filament expression is one of the major consequences of axon to Schwann cell signalling during myelination in the rat sciatic nerve. Unlike myelin-forming cells, adult non-myelin-forming Schwann cells remain very similar to embryonic and newborn cells with respect to expression of surface proteins, in contrast to the previously established developmental changes that occur in their surface lipids.     

Distribution of glial-associated proteins in the developing chick auditory brainstem

In the avian brainstem, nucleus magnocellularis (NM) projects bilaterally to nucleus laminaris (NL) in a pathway that facilitates sound localization. The distribution of glia during the development of this pathway has not previously been characterized. Radial glia, astrocytes, and oligodendrocytes facilitate many processes including axon pathfinding, synaptic development, and maturation. Here we determined the spatiotemporal expression patterns of glial cell types in embryonic development of the chick auditory brainstem using glial-specific antibodies and histological markers. We found that vimentin-positive processes are intercalated throughout the NL cell layer. Astrocytes are found in two domains: one in the ventral neuropil region and the other dorsolateral to NM. GFAP-positive processes are primarily distributed along the ventral margin of NL. Astrocytic processes penetrate the NL cell layer following the onset of synaptogenesis, but before pruning and maturation. The dynamic, nonoverlapping expression patterns of GFAP and vimentin suggest that distinct glial populations are found in dorsal versus ventral regions of NL. Myelination occurs after axons have reached their targets. FluoroMyelin and myelin basic protein (MBP) gradually increase along the mediolateral axis of NL starting at E10. Multiple GFAP-positive processes are directly apposed to NM-NL axons and MBP, which suggests a role in early myelinogenesis. Our results show considerable changes in glial development after initial NM-NL connections are made, suggesting that glia may facilitate maturation of the auditory circuit.     

Hyaluronate receptor CD44 is expressed by astrocytes in the adult chicken and in astrocyte cell precursors in early development of the chick spinal cord.

The role of the hyaluronate receptor, CD44, is well known in adult mammal astrocytes where it modulates neuron-glia interactions. However, no data exist regarding its expression in other vertebrates during their development. In order to detect the expression of CD44 in the chicken and its possible involvement in glial precursor migratory patterns during spinal cord development, a monoclonal antibody (MoAb) against the mammalian standard isoform, CD44-H, was used in immunohistochemical and immunoblot assays. With these methods, CD44 hyaluronate receptors were found on mature astrocyte membranes of adult chicken spinal cord. Astrocytes were identified using a MoAb against GFAP. During development, small clusters of CD44 labelled cells were seen lining the central canal starting from embryonic stage E10. These labelled cells were dispersed in the dorsal, lateral and ventral funiculi of the spinal cord in the subsequent stages. After stage E15, the CD44 labelled cells were identified as astrocytes because of their GFAP immunoreactivity. We conclude that CD44 receptors on immature astrocyte precursors should be considered as early astrocyte markers which have a possible role during cell migratory dispersal.     

Immunoblot identification of glial fibrillary acidic protein in rat sciatic nerve,brain, and spinal cord during development

The appearance of the glial fibrillary acidic protein (GFAP) during embryonic and postnatal development of the rat brain and spinal cord and in rat sciatic nerve during postnatal development was examined by the immunoblot technique. Cytoskeletal proteins were isolated from the central and peripheral nervous system and separated by SDS slab gel electrophoresis or two-dimensional gel electrophoresis. Proteins from the acrylamide gels were transferred to nitrocellulose sheets which were treated with anti-bovine GFAP serum and GFAP was identified by the immunoblot technique. GFAP was present in the embryonic rat brain and spinal cord at 14 and 16 days of gestation respectively. The appearance of GFAP at this stage of neural development suggests that the synthesis of GFAP may be related to the proliferation of radial glial cells from which astrocytes are derived. It is also feasible that GFAP provides structural support for the radial glial cell processes analogous to its role in differentiated astrocytes. GFAP was found to be present in rat sciatic nerves at birth and at all subsequent stages of development. These results indicate that some cellular elements in the rat sciatic nerve, such as Schwann cells, are capable of synthesizing GFAP which is immunochemically indistinguishable from its counterpart in the central nervous system. Thus it appears that GFAP is present both in the central and peripheral nervous system of the rat when the glial cells synthesizing GFAP are still undergoing differentiation.     

Cadmium induces alterations in the human spinal cord morphogenesis

The effects of cadmium on the central nervous system are still relatively poorly understood and its role in neurodegenerative diseases has been debated. In our research, cultured explants from 25 human foetal spinal cords (10–11 weeks gestational age) were incubated with 10 and 100 μM cadmium chloride (CdCl₂) for 24 h. After treatment, an immunohistochemical study [for Sglial fibrillary acidic protein (GFAP) and choline acetyltransferase (ChAT)], a Western blot analysis (for GFAP, β-Tubulin III, nerve growth factor receptor, Caspase 8 and poly (ADP-ribose) polymerase), and a terminal deoxynucleotidyl transferase biotin-dUTP nick end labelling (TUNEL) assay (for detection of apoptotic bodies) were performed. The treatment with CdCl₂ induced a significant and dose-dependent change in the ratio motor neurons/glial cells in the ventral horns of human foetal spinal cord. The decrease of the choline acetyltransferase-positive cells (motor neurons) and the reduction of β Tubulin III indicate that CdCl₂ specifically affects motor neurons of the ventral horns. While the number of motor neurons decreased for the activation of apoptotic pathways (as shown by the increased expression of Caspase 8, nerve growth factor receptor, and poly (ADP-ribose) polymerase), glial cells, both in the subependymal zone and in the gray matter of the ventral horns, increased (as shown by the increase of GFAP expression). These results provide the evidence that during human spinal cord development, CdCl₂ may affect the fate of neural and glial cells thus, being potentially involved in the etiopathogenesis of neurodegenerative diseases.     

Establishment of a Spinal Cord Injury Model in Adult Rats by an Electrocircuit-Controlled Impacting Device and its Pathological Observations

One of the crucial challenges in medicine is the treatment and rehabilitation of spinal cord injury (SCI). In this study, we established a stable and reproducible acute spinal cord injury model in adult rats. The SCI was inflicted by our self-innovated spinal cord impact device controlled by electrical circuit. The Basso, Beattie, and Bresnahan Locomotor Rating Scale (BBB) score, electrophysiology, histological, and immunohistochemical changes after SCI were observed. The BBB score of the injured rats began to increase from the 3rd day of SCI and reached at the score 7.2 ± 1.3 at the 28th day. The latency of cortical somatosensory evoked potentials (CSEP) was not observed 2 and 6 h after injury, but appeared 24 h after injury which was significantly prolonged. It recovered from day 3 gradually to 27.3 ± 2.7 ms on day 28. H&E staining showed that the structure of gray and white matter was disrupted after the SCI. The result also showed dramatic neuron degenerations, cellular swelling, and the proliferation of glial cells. The immunohistochemical analysis showed that the expression of neuron specific enolase (NSE) and neurofilament 200 (NF200) started lowering at 2 h and dropped to the bottom at 24 h. Their expression rebound from day 3 and yet to the original level at day 28 (P < 0.05). The number of cells expressing glial fibrillary acidic protein (GFAP) hiked from day 3, peaked at day 14, and began recovering from day 28 (P < 0.05). The changes of NSE, NF200, GFAP, and CSEP were significantly associated with the BBB score (P < 0.05). In conclusion, our self-innovated device can reproduce the injury model stably. The changes of NSE, NF, and GFAP after spinal cord injury reflect the characteristics of pathological change, which are closely associated with the functional recovery from the spinal cord injury.     

Modulation of Matrix Metalloproteinases Activity in the Ventral Horn of the Spinal Cord Re-stores Neuroglial Synaptic Homeostasis and Neurotrophic Support following Peripheral Nerve Injury

Modulation of extracellular matrix (ECM) remodeling after peripheral nerve injury (PNI) could represent a valid therapeutic strategy to prevent maladaptive synaptic plasticity in central nervous system (CNS). Inhibition of matrix metalloproteinases (MMPs) and maintaining a neurotrophic support could represent two approaches to prevent or reduce the maladaptive plastic changes in the ventral horn of spinal cord following PNI. The purpose of our study was to analyze changes in the ventral horn produced by gliopathy determined by the suffering of motor neurons following spared nerve injury (SNI) of the sciatic nerve and how the intrathecal (i.t.) administration of GM6001 (a MMPs inhibitor) or the NGF mimetic peptide BB14 modulate these events. Immunohistochemical analysis of spinal cord sections revealed that motor neuron disease following SNI was associated with increased microglial (Iba1) and astrocytic (GFAP) response in the ventral horn of the spinal cord, indicative of reactive gliosis. These changes were paralleled by decreased glial aminoacid transporters (glutamate GLT1 and glycine GlyT1), increased levels of the neuronal glutamate transporter EAAC1, and a net increase of the Glutamate/GABA ratio, as measured by HPLC analysis. These molecular changes correlated to a significant reduction of mature NGF levels in the ventral horn. Continuous i.t. infusion of both GM6001 and BB14 reduced reactive astrogliosis, recovered the expression of neuronal and glial transporters, lowering the Glutamate/GABA ratio. Inhibition of MMPs by GM6001 significantly increased mature NGF levels, but it was absolutely ineffective in modifying the reactivity of microglia cells. Therefore, MMPs inhibition, although supplies neurotrophic support to ECM components and restores neuro-glial transporters expression, differently modulates astrocytic and microglial response after PNI.     

Formation of the peripheral nervous system during tail regeneration in urodele amphibians: ultrastructural and immunohistochemical studies of the origin of the cells.

In the regenerating newt tail, epimorphic regeneration--which recapitulates morphologically normal embryonic development--proceeds along a rostrocaudal differentiation gradient. Innervation of the new myomeres results from the spinal roots of segments rostral to the amputation plane and from ventral roots emerging from the lateroventral region of the regenerating spinal cord, in which motor neurons are differentiating. Electron microscopy and an indirect immunofluorescence study with anti-glial fibrillary acid protein (GFAP) confirm that the ventrolateral part of the regenerated ependymal tube gives rise to cells of the ventral root sheath and the spinal ganglia. Anti-GFAP and anti-neurofilament antibodies showed that ependymoglial cells and Schwann cells may play a role in neuronal pathfinding by helping guide and stabilize pioneering axons as they extend toward the myomeres. The carbohydrate epitope NC-1 is expressed in the spinal cord, in sheath cells of the spinal ganglia and in the non-myelin-forming Schwann cells of the peripheral nervous system. L1, a Ca++ independent neural cell adhesion molecule, was detected in the axonal compartments of the regenerating spinal cord, on immature and/or non-myelin-forming Schwann cells within the peripheral nervous system (PNS), and on nerve fibers within the regenerate. These immunohistochemical observations collectively support the hypothesis that Schwann cells already present in the blastema could be involved in organizing neural pathways.     

Expression of vimentin and glial fibrillary acidic protein in human developing spinal cord

Summary Expression of intermediate filament proteins was studied in human developing spinal cord using immunoperoxidase and double-label immunofluorescence methods with monoclonal antibodies to vimentin and glial fibrillary acidic protein (GFAP). Vimentin was found in the processes of radial glial cells in 6-week embryos, while GFAP appeared in vimentin-positive astroglial cells at 8–10 weeks. GFAP and vimentin were present in approximately equal amounts in differentiating astrocytes in 23-week spinal cord. In 30-week fetuses, astrocytes reacted strongly for GFAP, while both the reaction intensity and the number of vimentin-positive cells fluctuated predominantly in the grey matter. No clear-cut transition from vimentin to GFAP was noticed during the development of astrocytes. The majority of ependymal cells in 23-week fetuses contained vimentin but only a few of them reacted for GFAP. The expression of vimentin continued during the whole development of the ependymal layer, in contrast to the reactivity for GFAP which disappeared between the 30th week and term.     

The molecular cloning of glial fibrillary acidic protein in <Emphasis Type="Italic">Gekko japonicus</Emphasis> and its expression changes after spinal cord transection

The glial fibrillary acidic protein (GFAP) is an astrocyte-specific member of the class III intermediate filament proteins. It is generally used as a specific marker of astrocytes in the central nervous system (CNS). We isolated a GFAP cDNA from the brain and spinal cord cDNA library of Gekko japonicus, and prepared polyclonal antibodies against gecko GFAP to provide useful tools for further immunochemistry studies. Both the real-time quantitative PCR and western blot results revealed that the expression of GFAP in the spinal cord after transection increased, reaching its maximum level after 3 days, and then gradually decreased over the rest of the 2 weeks of the experiment. Immunohistochemical analyses demonstrated that the increase in GFAP-positive labeling was restricted to the white matter rather than the gray matter. In particular, a slight increase in the number of GFAP positive star-shaped astrocytes was detected in the ventral and lateral regions of the white matter. Our results indicate that reactive astrogliosis in the gecko spinal cord took place primarily in the white matter during a short time interval, suggesting that the specific astrogliosis evaluated by GFAP expression might be advantageous in spinal cord regeneration.     

Proliferation and migration of glial precursor cells in the developing rat spinal cord

The mechanisms that control the production and differentiation of glial cells during development are difficult to unravel because of displacement of precursor cells from their sites of origin to their permanent location. The two main neuroglial cells in the rat spinal cord are oligodendrocytes and astrocytes. Considerable evidence supports the view that oligodendrocytes in the spinal cord are derived from a region of the ventral ventricular zone (VZ). Some astrocytes, at least, may arise from radial glia. In this study a 5-Bromo-2′-deoxyuridine (BrdU) incorporation assay was used to identify proliferating cells and examine the location of proliferating glial precursor cells in the embryonic spinal cord at different times post BrdU incorporation. In this way the migration of proliferating cells into spinal cord white matter could be followed. At E14, most of the proliferating cells in the periventricular region were located dorsally and these cells were probably proliferating neuronal precursors. At E16 and E18, the majority of the proliferating cells in the periventricular region were located ventrally. In the white matter the number of proliferating cells increased as the animals increased in age and much of this proliferation occurred locally. BrdU labelling showed that glial precursor cells migrate from their ventral and dorsal VZ birth sites to peripheral regions of the cord. Furthermore although the majority of proliferating cells in the spinal cord at E16 and E18 were located in the ventral periventricular region, some proliferating cells remained in the dorsal VZ region of the cord.     

Estrogen receptor immunoreactivity in Schwann-like brain macroglia.

Olfactory ensheathing cells, tanycytes, pituicytes, pineal glia, retinal Müller cells, and Bergmann glia of normal male rats express concomitantly estrogen receptor, low-affinity neurotrophin receptor, antigen O4, and GFAP, markers characteristic of nonmyelinating Schwann cells. These cells were able to survive and proliferate when cultured from adult tissue, promoted neurite outgrowth, and could guide and ensheath growing neurites. We called this distinct group of growth-promoting central nervous system (CNS) macroglia aldynoglia (Greek: to make grow). Its proliferative and growth-promoting properties seem to be retained during the whole lifetime of the organism in those CNS loci where normal function depends on continuous axon renewal. Aldynoglia plasticity seems totally or partially lost with age where and when it is no longer critical, as in the case of adult cortical and spinal cord radial glia. The concomitant expression of estrogen receptor and low-affinity neurotrophin receptor may promote Schwann-like plasticity of glial cells.     

Collagen scaffolds modified with collagen-binding bFGF promotes the neural regeneration in a rat hemisected spinal cord injury model

Nerve conduit is one of strategies for spine cord injury(SCI)treatment.Recently,studies showed that biomaterials could guide the neurite growth and promote axon regeneration at the injury site.However,the scaffold by itself was difficult to meet the need of SCI functional recovery.The basic fibroblast growth factor(bFGF)administration significantly promotes functional recovery after organ injuries.Here,using a rat model of T9 hemisected SCI,we aimed at assessing the repair capacity of implantation of collagen scaffold(CS)modified by collagen binding bFGF(CBD-bFGF).The results showed that CS combined with CBD-bFGF treatment improved survival rates after the lateral hemisection SCI.The CS/CBD-bFGF group showed more significant improvements in motor than the simply CS-implanted and untreated control group,when evaluated by the 21-point Basso-Beattie-Bresnahan(BBB)score and footprint analysis.Both hematoxylin and eosin(H&E)and immunohistochemical staining of neurofilament(NF)and glial fibrillary acidic protein(GFAP)demonstrated that fibers were guided to grow through the implants.These findings indicated that administration of CS modified with CBD-bFGF could promote spinal cord regeneration and functional recovery.     

ErbB transmembrane tyrosine kinase receptors are expressed by sensory and motor neurons projecting into sciatic nerve.

Adult spinal cord motor and dorsal root ganglion (DRG) sensory neurons express multiple neuregulin-1 (NRG-1) isoforms that act as axon-associated factors promoting neuromuscular junction formation and Schwann cell proliferation and differentiation. NRG-1 isoforms are also expressed by muscle and Schwann cells, suggesting that motor and sensory neurons are themselves acted on by NRG-1 isoforms produced by their peripheral targets. To test this hypothesis, we examined the expression of the NRG-1 receptor subunits erbB2, erbB3, and erbB4 in rat lumbar DRG and spinal cord. All three erbB receptors are expressed in these tissues. Sciatic nerve transection, an injury that induces Schwann cell expression of NRG-1, alters erbB expression in DRG and cord. Virtually all DRG neurons are erbB2- and erbB3-immunoreactive, with erbB4 also detectable in many neurons. In spinal cord white matter, erbB2 and erbB4 antibodies produce dense punctate staining, whereas the erbB3 antibody primarily labels glial cell bodies. Spinal cord dorsal and ventral horn neurons, including alpha-motor neurons, exhibit erbB2, erbB3, and erbB4 immunoreactivity. Spinal cord ventral horn also contains a population of small erbB3+/S100beta+/GFAP- cells (GFAP-negative astrocytes or oligodendrocytes). We conclude that sensory and motor neurons projecting into sciatic nerve express multiple erbB receptors and are potentially NRG-1 responsive.     

Keratin 8 of simple epithelia is expressed in glia of the goldfish nervous system

The intermediate filament protein composition in glial cells of goldfish optic nerve differs from that found in glial cells of the goldfish spinal cord and brain. Brain and spinal cord glial cells contain glial fibrillary acidic protein (GFAP), whereas glial cells in the optic nerve contain ON3. The ON3 protein of the goldfish optic nerve was recently identified as the goldfish equivalent to the mammalian type II keratin 8 protein. In addition to the ON3 protein, the goldfish optic nerve also contains a 48-kDa protein. Immunoblotting experiments suggest that this protein is equivalent to the mammalian type I keratin 18 protein, which typically pairs with keratin 8 to form filaments. We show that these proteins are not specific to the optic nerve. The ON3 and 48-kDa proteins of the goldfish optic nerve share common antigenic properties with the predominant keratin pair expressed in the goldfish liver. These proteins are also expressed at low levels in the goldfish brain and spinal cord. In addition RNase protection assays and Northern blots indicate that the mRNA for the ON3 protein in optic nerve is identical to the message found in other goldfish tissues. The expression of ON3 was also examined in cultured glial cells from goldfish spinal cord and optic nerve and cultured fibroblast cells. Analysis of intermediate filament protein expression in cultured glial cells taken from goldfish spinal cord demonstrated the absence of GFAP in these cells and the expression of ON3. This protein was also the predominant intermediate filament protein of cultured optic nerve glial cells and fibroblasts. The differences in the expression of intermediate filament proteins in mammals and lower vertebrates are discussed. In addition, we discuss how the expression of a simple epithelial keratin pair in glial cells of the goldfish optic nerve may be associated with this system's capacity for continuous growth and regeneration.     

Ependyma: phylogenetic evolution of glial fibrillary acidic protein (GFAP) and vimentin expression in vertebrate spinal cord

The phylogenetic evolution was studied of both glial fibrillary acidic protein (GFAP) and vimentin expression in the ependyma of the adult vertebrate spinal cord. Eleven species from different vertebrate groups were examined using different fixatives and fixation procedures to demonstrate any differences in immunoreactivity. GFAP expression in the ependymal cells showed a clear inverse relation with phylogenetic evolution because it was more elevated in lower than in higher vertebrates. GFAP positive cells can be ependymocytes and tanycytes, although depending on their structural characteristics and distribution, the scarce GFAP positive ependymal cells in higher vertebrates may be tanycytes. Ependymal vimentin expression showed a species-dependent pattern instead of a phylogenetic pattern of expression. Vimentin positive ependymal cells were only found in fish and rats; in fish, they were tanycytes and were quite scarce, with only one or two cells per section being immunostained. However, in the rat spinal cord, all the ependymocytes showed positive immunostaining for vimentin. The importance of the immunohistochemical procedure, the cellular nature of GFAP positive ependymal cells and the relationship between tanycytes and ependymocytes are discussed, as well as GFAP and vimentin expression.     

Commissure formation in the embryonic CNS of Drosophila.

In the ventral nerve cord of Drosophila most axons are organized in a simple, ladder-like pattern. Two segmental commissures connect the hemisegments along the mediolateral and two longitudinal connectives connect individual neuromeres along the anterior-posterior axis. Cells located at the midline of the developing CNS first guide commissural growth cones toward and across the midline. In later stages, midline glial cells are required to separate anterior and posterior commissures into distinct axon bundles. To unravel the genes underlying the formation of axon pattern in the embryonic ventral nerve cord, we conducted a saturating ethylmethane sulfonate mutagenesis, screening for mutations which disrupt this process. Subsequent genetic and phenotypic analyses support a sequential model of axon pattern formation in the embryonic ventral nerve cord. Specification of midline cell lineages is brought about by the action of segment polarity genes. Five genes are necessary for the establishment of the commissures. In addition to commissureless, the netrin genes, and the netrin receptor encoded by the frazzled gene, two gene functions are required for the initial formation of commissural tracts. Over 20 genes appear to be required for correct development of the midline glial cells which are necessary for the formation of distinct segmental commissures.