Motor neuron pathfinding following rhombomere reversals in the chick embryo hindbrain.

Motor neurons are segmentally organised in the developing chick hindbrain, with groups of neurons occupying pairs of hindbrain segments or rhombomeres. The branchiomotor nucleus of the trigeminal nerve occupies rhombomeres 2 and 3 (r2 and r3), that of the facial nerve r4 and r5, and that of the glossopharyngeal nerve r6 and r7. Branchiomotor neuron cell bodies lie within the basal plate, forming columns on either side of the ventral midline floor plate. Axons originating in rhombomeres 2, 4 and 6 grow laterally (dorsally) towards the exit points located in the alar plates of these rhombomeres, while axons originating in odd-numbered rhombomeres 3 and 5 grow laterally and then rostrally, crossing a rhombomere boundary to reach their exit point. Examination of the trajectories of motor axons in odd-numbered segments at late stages of development (19-25) showed stereotyped pathways, in which axons grew laterally before making a sharp turn rostrally. During the initial phase of outgrowth (stage 14-15), however, axons had meandering courses and did not grow in a directed fashion towards their exit point. When r3 or r5 was transplanted with reversed rostrocaudal polarity prior to motor axon outgrowth, the majority of axons grew to their appropriate, rostral exit point, despite the inverted neuroepithelial polarity. In r3 reversals, however, there was a considerable increase in the normally small number of axons that grew out via the caudal, r4 exit point. These findings are discussed with relevance to the factors involved in motor neuron specification and axon outgrowth in the developing hindbrain.     

Autonomous and nonautonomous functions for Hox/Pbx in branchiomotor neuron development

The vertebrate branchiomotor neurons are organized in a pattern that corresponds with the segments, or rhombomeres, of the developing hindbrain and have identities and behaviors associated with their position along the anterior/posterior axis. These neurons undergo characteristic migrations in the hindbrain and project from stereotyped exit points. We show that lazarus/pbx4, which encodes an essential Hox DNA-binding partner in zebrafish, is required for facial (VIIth cranial nerve) motor neuron migration and for axon pathfinding of trigeminal (Vth cranial nerve) motor axons. We show that lzr/pbx4 is required for Hox paralog group 1 and 2 function, suggesting that Pbx interacts with these proteins. Consistent with this, lzr/pbx4 interacts genetically with hoxb1a to control facial motor neuron migration. Using genetic mosaic analysis, we show that lzr/pbx4 and hoxb1a are primarily required cell-autonomously within the facial motor neurons; however, analysis of a subtle non-cell-autonomous effect indicates that facial motor neuron migration is promoted by interactions amongst the migrating neurons. At the same time, lzr/pbx4 is required non-cell-autonomously to control the pathfinding of trigeminal motor axons. Thus, Pbx/Hox can function both cell-autonomously and non-cell-autonomously to direct different aspects of hindbrain motor neuron behavior.     

Axon growth from limb motorneurons in the locust embryo: the effect of target limb removal on the pattern of axon branching in the periphery

Metathoracic limb buds have been unilaterally ablated from locust embryos at 25 to 30% of embryonic development and the effect of this operation on the axon morphology of the motorneuron fast extensor tibiae (FETi) observed at later embryonic stages. In control embryos this neuron sends a single axon out the main leg nerve, nerve 5, to the extensor tibiae muscle in the femur. In limb ablated embryos the axon of FETi is found in a wide variety of aberrant peripheral nerve pathways and projects to a wide range of foreign muscles. There is a degree of apparent selectivity, but no rigid hierarchy, in the choice of pathway and muscle made by FETi. A high degree of variability is found between one embryo and another in the extent and pattern of axon branching. The axon of FETi is generally found in pathways that correspond to nerves in control embryos but on occasion grows along novel routes. An anteriorly directed dendritic branch, seldom seen in control FETi neurons, is frequently seen in experimental FETis. These findings are discussed in terms of the rules for specific axon growth in normal development.     

Early axonal contacts during development of an identified dendrite in the brain of the zebrafish

We have identified the initial synaptic contacts made onto the Mauthner (M) cell, an identified neuron that arises during early development of the zebrafish hindbrain. The contacts are made by a small bundle of pioneering trigeminal sensory axons onto the M cell soma before it forms dendrites. The sensory bundle is then partially enveloped by the M cell. The lateral dendrite appears at about the site of the contact, and eventually the trigeminal inputs are shifted to its trunk. As the dendrite elongates, other sensory contacts are made on its distal regions, sequentially from the acoustico-vestibular nerve and the lateral line nerves. To learn whether the earliest inputs induce the initial outgrowth of the M cell dendrite, we ablated the trigeminal neurons by laser irradiation before they contacted the M cell. Morphogenesis of the M cell, including its dendrite, appeared normal.     

In vivo dynamics of axon pathfinding in the Drosophila CNS: A time-lapse study of an identified motorneuron

We developed a system for time-lapse observation of identified neurons in the central nervous system (CNS) of the Drosophila embryo. Using this system, we characterize the dynamics of filopodia and axon growth of the motorneuron RP2 as it navigates anteriorly through the CNS and then laterally along the intersegmental nerve (ISN) into the periphery. We find that both axonal extension and turning occur primarily through the process of filopodial dilation. In addition, we used the GAL4-UAS system to express the fusion protein Tau-GFP in a subset of neurons, allowing us to correlate RP2's patterns of growth with a subset of axons in its environment. In particular, we show that RP2's sharp lateral turn is coincident with the nascent ISN. © 1998 John Wiley & Sons, Inc. J Neurobiol 37: 607–621, 1998     

Diffusible signals and fasciculated growth in reticulospinal axon pathfinding in the hindbrain

We have addressed the control of longitudinal axon pathfinding in the developing hindbrain, including the caudal projections of reticular and raphe neurons. To test potential sources of guidance signals, we assessed axon outgrowth from embryonic rat hindbrain explants cultured in collagen gels at a distance from explants of midbrain-hindbrain boundary (isthmus), caudal hindbrain, or cervical spinal cord. Our results showed that the isthmus inhibited caudally directed axon outgrowth by 80% relative to controls, whereas rostrally directed axon outgrowth was unaffected. Moreover, caudal hindbrain or cervical spinal cord explants did not inhibit caudal axons. Immunohistochemistry for reticular and raphe neuronal markers indicated that the caudal, but not the rostral projections of these neuronal subpopulations were inhibited by isthmic explants. Companion studies in chick embryos showed that, when the hindbrain was surgically separated from the isthmus, caudal reticulospinal axon projections failed to form and that descending pioneer axons of the medial longitudinal fasciculus (MLF) play an important role in the caudal reticulospinal projection. Taken together, these results suggest that diffusible chemorepellent or nonpermissive signals from the isthmus and substrate-anchored signals on the pioneer MLF axons are involved in the caudal direction of reticulospinal projections and might influence other longitudinal axon projections in the brainstem.     

Intrinsic properties guide proximal abducens and oculomotor nerve outgrowth in avian embryos

Proper movement of the vertebrate eye requires the formation of precisely patterned axonal connections linking cranial somatic motoneurons, located at defined positions in the ventral midbrain and hindbrain, with extraocular muscles. The aim of this research was to assess the relative contributions of intrinsic, population-specific properties and extrinsic, outgrowth site-specific cues during the early stages of abducens and oculomotor nerve development in avian embryos. This was accomplished by surgically transposing midbrain and caudal hindbrain segments, which had been pre-labeled by electroporation with an EGFP construct. Graft-derived EGFP+ oculomotor axons entering a hindbrain microenvironment often mimicked an abducens initial pathway and coursed cranially. Similarly, some EGFP+ abducens axons entering a midbrain microenvironment mimicked an oculomotor initial pathway and coursed ventrally. Many but not all of these axons subsequently projected to extraocular muscles that they would not normally innervate. Strikingly, EGFP+ axons also took initial paths atypical for their new location. Upon exiting from a hindbrain position, most EGFP+ oculomotor axons actually coursed ventrally and joined host branchiomotor nerves, whose neurons share molecular features with oculomotor neurons. Similarly, upon exiting from a midbrain position, some EGFP+ abducens axons turned caudally, elongated parallel to the brainstem, and contacted the lateral rectus muscle, their originally correct target. These data reveal an interplay between intrinsic properties that are unique to oculomotor and abducens populations and shared ability to recognize and respond to extrinsic directional cues. The former play a prominent role in initial pathway choices, whereas the latter appear more instructive during subsequent directional choices.     

Contactin 1 knockdown in the hindbrain induces abnormal development of the trigeminal sensory nerve in <Emphasis Type="Italic">Xenopus</Emphasis> embryos

In Xenopus tailbud embryos, the mandibular branch of trigeminal sensory nerve has a transient pathway innervating the cement gland. This pathway is settled by pioneer neurons in the trigeminal ganglion and along which extend later-growing axons from the trigeminal ganglion and the hindbrain. Axons in this branch express a neuronal recognition molecule, Contactin 1, from the initial stage of its outgrowth in early tailbud embryos and form a tightly joined, strongly Contactin 1-positive fascicle in the later stages. When the expression vector encoding the enhanced green fluorescent protein was electrotransfected into the brain neurons of early tailbud embryos, the fluorescence was detected in the hindbrain and the trigeminal nerve at late tailbud stages. Cotransfection of antisense vector caused knockdown of Contactin 1 concurrent with defasciculation and misguidance of the sensory axons in the trigeminal mandibular branch. The results suggest that Contactin 1 is required for the growing axon of hindbrain sensory neurons to recognize and follow the pathway settled by the pioneer neurons.     

Neuronal expression of peripherin,a type III intermediate filament protein,in the mouse hindbrain

Peripherin is a 57 kDa Type III intermediate filament protein associated with neurite extension, neuropathies such as amyotrophic lateral sclerosis, and cranial nerve and dorsal root projections. However, knowledge of peripherin expression in the CNS is limited. We have used immunoperoxidase histochemistry to characterise peripherin expression in the mouse hindbrain, including the inferior colliculus, pons, medulla and cerebellum. Peripherin immunolabelling was observed in the nerve fibres and nuclei that are associated with all cranial nerves [(CN) V–XII] in the hindbrain. Peripherin expression was prominent in the cell bodies and axons of the mesenchephalic trigeminal nucleus and the pars compacta region of nucleus ambiguus, and in the fibres that comprise the solitary tract, the descending spinal trigeminal tract and the trigeminal and facial nerves. A small proportion of peripherin positive fibres in CN VIII likely arise from cochlear type II spiral ganglion neurons. Peripherin positive fibres were also observed in the inferior cerebellar peduncle and folia in the intermediate zone of the cerebellum. Antibody specificity was confirmed by absence of labelling in hindbrain tissue from peripherin knockout mice. This study shows that in the adult mouse hindbrain, peripherin is expressed in discrete neuronal subpopulations that have sensory, motor and autonomic functions.     

Fiber pathways and positional changes in efferent perikarya of 2.5-to 7-day chick embryos as revealed with dil and dextran amiens

The differentiation of facial motoneurons and inner ear (octaval) efferents was examined in chicken embryos by applying Dil or dextran amines to the cut VII/VIII nerve (peripheral label) or to the basal/floor plate of rhombomeres 4/5 (central label). Central labeling found axons of these efferent neurons to leave the brain as early as 2.5 days of incubation. Peripheral labeling identified cell bodies ipsilaterally in rhombomeres 4 and 5 at 2.5 days. Central labeling at 3.5 days showed these fibers to have fully segregated into separate pathways to the facial nerve and the inner ear and that the octaval efferent axons had reached the otocyst wall. By 3.5 days many peripherally labeled octaval efferent somata were found in the floor plate and by 5 days they were found bilaterally. At 6 days, selective peripheral labeling of either the VIIth or VIIIthe nerve showed that the contralateral population consisted of octaval efferents and central label applied to the floor plate of rhombomeres 4/5 identified fibers that entered the octaval nerve via the facial root and entered the vestibular sensory epithelia. To gether these data suggest an initial mingling of two different motoneuron populations (facial and octaval) in rhombomeres 4/5 and a subsequent segregation by differential migration. Our data also find a much earlier arrival of octaval efferent axons at the otic vesicle than previously described and suggest a contralateral migration of many octaval efferents beginning shortly after their axons reach the facial nerve root. © 1993 John Wiley & Sons, Inc.     

Normal and abnormal pathfinding of facial nerve fibers in the chick embryo.

Development of the facial nerve was studied in normal chicken embryos and after surgical disruption of ingrowing sensory facial nerve fibers at 38-72 h of incubation. Disruption of facial nerve fibers by otocyst removal often induced a rostral deviation of the facial nerve and ganglion to the level of the trigeminal ganglion. Cell bodies of the geniculate ganglion trailed their deviating neurites and occupied an abnormal rostral position adjacent to the trigeminal ganglion. Deviating facial nerve fibers were labeled with the carbocyanine fluorescent tracer DiI in fixed tissue. Labeled fibers penetrated the cranium adjacent to the trigeminal ganglion, but they did not follow the trigeminal nerve fibers into the brain stem. Rather, after entering the cranium, they projected caudally to their usual site of entrance and proceeded towards their normal targets. This rostral deviation of the facial nerve was observed only after surgery at 48-72 h of incubation, but not in cases with early otocyst removal (38-48 h). A rostral deviation of the facial nerve was seen in cases with partial otocyst removal when the vestibular nerve was absent. The facial nerve followed its normal course when the vestibular nerve persisted. We conclude that disruption of the developing facial pathway altered the routes of navigating axons, but did not prevent pathfinding and innervation of the normal targets. Pathfinding abilities may not be restricted to pioneering axons of the facial nerve; later-developing facial nerve fibers also appeared to have positional information. Our findings are consistent with the hypothesis that navigating axons may respond to multiple guidance cues during development. These cues appear to differ as a function of position of the navigating axon.     

Vitamin A deficiency results in the dose-dependent acquisition of anterior character and shortening of the caudal hindbrain of the rat embryo

The developing nervous system is particularly vulnerable to vitamin A deficiency. Retinoid has been proposed to be a posteriorizing factor during hindbrain development, although direct evidence in the mammalian embryo is lacking. In the present study, pregnant vitamin A-deficient (VAD) rats were fed purified diets containing varying levels of all-trans-retinoic acid (atRA; 0, 0.5, 1.5, 6, 12, 25, 50, 125, or 250 microg/g diet) or were supplemented with retinol. Hindbrain development was studied from embryonic day 10 to 12.5 ( approximately 6 to 40 somites). Normal morphogenesis was observed in all embryos from groups fed 250 microg atRA/g diet or retinol. The most caudal region of the hindbrain was the most sensitive to retinoid insufficiency, as evidenced by a loss of the hypoglossal nerve (cranial nerve XII) in embryos from the 125 microg atRA/g diet group. Further reduction of atRA to 50 microg/g diet led to the loss of cranial nerves IX, X, XI, and XII and associated sensory ganglia IX and X in all embryos as well as the loss of hindbrain segmentation caudal to the rhombomere (r) 3/4 border in a subset of embryos. Dysmorphic orthotopic otic vesicles or immature otic-like vesicles in both orthotopic and caudally ectopic locations were also observed. As the level of atRA was reduced, a loss of caudal hindbrain segmentation was observed in all embryos and the incidence of otic vesicle abnormalities increased. Perturbations in hindbrain segmentation, cranial nerve formation, and otic vesicle development were associated with abnormal patterning of the posterior hindbrain. Embryos from VAD dams fed between 0.5 and 50 microg atRA/g diet exhibited Hoxb-1 protein expression along the entire neural tube caudal to the r3/r4 border at a time when it should be restricted to r4. Krox-20 protein expression was expanded in r3 but absent or reduced in presumptive r5. Hoxd-4 mRNA expression was absent in the posterior hindbrain, and the rostral limit of Hoxb-5 protein expression in the neural tube was anteriorized, suggesting that the most posterior hindbrain region (r7/r8) had been deleted and/or improperly patterned. Thus, when limiting amounts of atRA are provided to VAD dams, the caudal portion of the hindbrain is shortened and possesses r4/r5-like characteristics, with this region finally exhibiting r4-like gene expression when retinoid is restricted even more severely. Thus, regions of the anterior hindbrain (i.e., r3 and r4) appear to be greatly expanded, whereas the posterior hindbrain (r5-r8) is reduced or absent. This work shows that retinoid plays a critical role in patterning, segmentation, and neurogenesis of the caudal hindbrain and serves as an essential posteriorizing signal for this region of the central nervous system in the mammal.     

Normal and abnormal pathfinding of facial nerve fibers in the chick embryo

Development of the facial nerve was studied in normal chicken embryos and after surgical disruption of ingrowing sensory facial nerve fibers at 38–72 h of incubation. Disruption of facial nerve fibers by otocyst removal often induced a rostral deviation of the facial nerve and ganglion to the level of the trigeminal ganglion. Cell bodies of the geniculate ganglion trailed their deviating neurites and occupied an abnormal rostral position adjacent to the trigeminal ganglion. Deviating facial nerve fibers were labeled with the carbocyanine fluorescent tracer Dil in fixed tissue. Labeled fibers penetrated the cranium adjacent to the trigeminal ganglion, but they did not follow the trigeminal nerve fibers into the brain stem. Rather, after entering the cranium, they projected caudally to their usual site of entrance and proceeded towards their normal targets. This rostral deviation of the facial nerve was observed only after surgery at 48–72 h of incubation, but not in cases with early otocyst removal (38–48 h). A rostral deviation of the facial nerve was seen in cases with partial otocyst removal when the vestibular nerve was absent. The facial nerve followed its normal course when the vestibular nerve persisted. We conclude that disruption of the devloping facial pathway altered the routes of navigating axons, but did not prevent pathfinding and innervation of the normal targets. Pathfinding abilities may not be restricted to pioneering axons of the facial nerve; later-developing facial nerve fibers also appeared to have positional information. Our findings are consistent with the hypothesis that navigating axons may respond to multiple guidance cues during development. These cues appear to differ as a function of position of the navigating axon. © 1992 John Wiley & Sons, Inc.     

Ventral midline cells are required for the local control of commissural axon guidance in the mouse spinal cord.

Specialized cells at the midline of the central nervous system have been implicated in controlling axon projections in both invertebrates and vertebrates. To address the requirement for ventral midline cells in providing cues to commissural axons in mice, we have analyzed Gli2 mouse mutants, which lack specifically the floor plate and immediately adjacent interneurons. We show that a Dbx1 enhancer drives tau-lacZ expression in a subpopulation of commissural axons and, using a reporter line generated from this construct, as well as DiI tracing, we find that commissural axons projected to the ventral midline in Gli2(-/-) embryos. Netrin1 mRNA expression was detected in Gli2(-/-) embryos and, although much weaker than in wild-type embryos, was found in a dorsally decreasing gradient. This result demonstrates that while the floor plate can serve as a source of long-range cues for C-axons in vitro, it is not required in vivo for the guidance of commissural axons to the ventral midline in the mouse spinal cord. After reaching the ventral midline, most commissural axons remained clustered in Gli2(-/-) embryos, although some were able to extend longitudinally. Interestingly, some of the longitudinally projecting axons in Gli2(-/-) embryos extended caudally and others rostrally at the ventral midline, in contrast to normal embryos in which virtually all commissural axons turn rostrally after crossing the midline. This finding indicates a critical role for ventral midline cells in regulating the rostral polarity choice made by commissural axons after they cross the midline. In addition, we provide evidence that interactions between commissural axons and floor plate cells are required to modulate the localization of Nr-CAM and TAG-1 proteins on axons at the midline. Finally, we show that the floor plate is not required for the early trajectory of motoneurons or axons of the posterior commissure, whose projections are directed away from the ventral midline in both WT and Gli2(-/-) embryos, although they are less well organized in Gli2(-/-)mutants.     

Slit-mediated repulsion is a key regulator of motor axon pathfinding in the hindbrain

The floor plate is known to be a source of repellent signals for cranial motor axons, preventing them from crossing the midline of the hindbrain. However, it is unknown which molecules mediate this effect in vivo. We show that Slit and Robo proteins are candidate motor axon guidance molecules, as Robo proteins are expressed by cranial motoneurons, and Slit proteins are expressed by the tissues that delimit motor axon trajectories, i.e. the floor plate and the rhombic lip. We present in vitro evidence showing that Slit1 and Slit2 proteins are selective inhibitors and repellents for dorsally projecting, but not for ventrally projecting, cranial motor axons. Analysis of mice deficient in Slit and Robo function shows that cranial motor axons aberrantly enter the midline, while ectopic expression of Slit1 in chick embryos leads to specific motor axon projection errors. Expression of dominant-negative Robo receptors within cranial motoneurons in chick embryos strikingly perturbs their projections, causing some motor axons to enter the midline, and preventing dorsally projecting motor axons from exiting the hindbrain. These data suggest that Slit proteins play a key role in guiding dorsally projecting cranial motoneurons and in facilitating their neural tube exit.     

Regulation of SC1/DM-GRASP during the migration of motor neurons in the chick embryo brain stem

The hindbrain of the chick embryo contains three classes of motor neurons: somatic, visceral, and branchial motor. During development, somata of neurons in the last two classes undergo a laterally directed migration within the neuroepithelium; somata translocate towards the nerve exit points, through which motor axons are beginning to extend into the periphery. All classes of motor neuron are immunopositive for the SC1/DM-GRASP cell surface glycoprotein. We have examined the relationship between patterns of motor neuron migration, axon outgrowth, and expression of the SC1/DM-GRASP mRNA and protein, using anterograde or retrograde axonal tracing, immunohistochemistry, and in situ hybridization. We find that as motor neurons migrate laterally, SC1/DM-GRASP is down-regulated, both on neuronal somata and axonal surfaces. Within individual motor nuclei, these lateral, more mature neurons are found to possess longer axons than the young, medial cells of the population. Labelling of sensory or motor axons growing into the second branchial arch also shows that motor axons reach the muscle plate first, and that SC1/DM-GRASP is expressed on the muscle at the time growth cones arrive. 1994 John Wiley & Sons, Inc.     

Both neural crest and placode contribute to the ciliary ganglion and oculomotor nerve

The chick ciliary ganglion is a neural crest-derived parasympathetic ganglion that innervates the eye. Here, we examine its axial level of origin and developmental relationship to other ganglia and nerves of the head. Using small, focal injections of DiI, we show that neural crest cells arising from both the caudal half of the midbrain and the rostral hindbrain contribute to the ciliary as well as the trigeminal ganglion. Precursors to both ganglia have overlapping migration patterns, moving first ventrolaterally and then rostrally toward the optic vesicle. At the level of the midbrain/forebrain junction, precursors to the ciliary ganglion separate from the main migratory stream, turn ventromedially, and condense in the vicinity of the rostral aorta and Rathke's pouch. Ciliary neuroblasts first exit the cell cycle at early E2, prior to and during ganglionic condensation, and neurogenesis continues through E5.5. By E3, markers of neuronal differentiation begin to appear in this population. By labeling the ectoderm with DiI, we discovered a new placode, caudal to the eye and possibly contiguous to the trigeminal placode, that contributes a few early differentiating neurons to the ciliary ganglion, oculomotor nerve, and connecting branches to the ophthalmic nerve. These results suggest for the first time a dual neural crest and placodal contribution to the ciliary ganglion and associated nerves.     

Beyond the neckless phenotype: influence of reduced retinoic acid signaling on motor neuron development in the zebrafish hindbrain

Retinoic acid (RA) has been identified as a key signal involved in the posteriorization of vertebrate neural ectoderm. The main biosynthetic enzyme responsible for RA signaling in the hindbrain and spinal cord is Raldh2. However, neckless/raldh2-mutant (nls) zebrafish exhibit only mild degrees of anteriorization in the neural ectoderm, compared to full vitamin A deficiency in amniotes and the Raldh2-/- mouse. Here we investigated the role of RA during neuronal development in the zebrafish hindbrain and anterior spinal cord using DEAB, an inhibitor of retinaldehyde dehydrogenases. We show that the nls hindbrain and spinal cord are not fully devoid of RA, since blocking Raldh-mediated RA signaling leads to a more severe hindbrain phenotype than in nls. The anteroposterior distribution of branchiomotor neurons in the facial and more posterior nuclei depends on full RA signaling throughout early and late gastrula stages. In contrast, inhibition of RA synthesis after gastrulation reduces the number of branchiomotor neurons in the vagal nucleus, but has no effect on anteroposterior cell fates. In addition, blockage of RA-mediated signaling not only interferes with the differentiation of branchiomotor neurons and their axons in the hindbrain, but also affects the development of the posterior lateral line nerve.     

Microtubules and filaments in developing axons and optic stalk cells

Fibrous structures have been studied in the developing optic nerve of chick embryos. The first ganglion cell axons (3-day embryos) were of moderate size, with both neurofilaments and microtubules. Subsequently (4- and 5-day embryos), very small axons were also present. In thesc embryos and in the 4-day hatched chick, the density of microtubules fell within the same range for all but the very small axons, which tended to have more microtubules per unit area. Filaments similar to those previously thought to represent neurofilaments in other parts of the embryonic nervous system were present in the early optic stalk cells, calling into question the reliability of identifying early nerve cells on the basis of 'neurofilaments'.