Development Supplement 2, 1991, L31-139Printed in Great Britain © The Company of Biologists Limited 1991

131

Axon repulsion during peripheral nerve segmentation

ROGER J. KEYNES, KAREN F. JAQUES and GEOFFREY M. W. COOK

Department of Anatomy, Downing Street, Cambridge CB2 3DY, UK

Summary

The guidance of axons during embryonic development islikely to involve both adhesive and repulsive interactionsbetween growth cones and their environment. We arecharacterising the role and mechanism of repulsionduring the segmental outgrowth of motor and sensoryaxons in the somite mesoderm of chick embryos. Axonsare confined to the anterior half of each somite by theexpression in the posterior half of a glycoconjugatesystem (48xlO3Mr and 55xlO3Mr) that causes thecollapse of dorsal root ganglion growth cones when

applied in vitro. Enzymatic cleavage of this fraction withspecific combinations of endo- and exoglycosidasesremoves collapse activity, suggesting that carbohydrateresidues are involved in the execution of collapse. Asimilar activity is also detectable in normal adult greymatter, suggesting roles for repulsion beyond thedevelopment of spinal nerve segmentation.

Key words: growth cone, contact inhibition, neuralsegmentation.

Introduction

Although differential adhesion has been recognised forsome time as an important mechanism in neuraldevelopment, a parallel role for contact inhibition orrepulsion has been proposed only recently (for reviewssee Patterson, 1988; Walter et al. 1990a; Keynes et al.1991). Contact inhibition has been well known as aphenomenon in vitro since its original description byAbercrombie and Heaysman (1954). On mutual con-tact, chick heart fibroblasts were seen to undergo asequence of reactions comprising adhesion, localparalysis of motility in the region of contact, and finallya cytoskeletal contraction and withdrawal. Growthcones can display a remarkably similar response invitro: in cocultures of retinal and sympathetic explants,when a retinal growth cone contacts a sympatheticaxon, or vice versa, the growth cone stops moving andcollapses within a few minutes (Kapfhammer et al. 1986;Kapfhammer and Raper, 1987). As for interactingfibroblasts, fine strands of cytoplasm remain adherentto the axon during the collapse, and normal movementis then restored with the elaboration of another growthcone.

The full growth cone collapse seen in vitro has notbeen described in vivo (and would not necessarily beexpected, see below), but such observations do suggestthat analogous mechanisms may be involved in normaldevelopmental processes such as axon guidance. Thedevelopment of an in vitro assay for growth conecollapse (Raper and Kapfhammer, 1990) has alsoallowed a molecular analysis of the phenomenon to beundertaken. Candidate molecules have been isolated bythe application of the assay to two vertebrate systems,

the retinotectal projection (Stahl et al. 1990) and spinalnerve segmentation (Davies et al. 1990), and the lattersystem will form the subject of this review.

Spinal nerve segmentation

The generation of a segmented pattern of spinal nervesin higher vertebrate (chick) embryos provides a goodmodel system for the study of axon guidance, beingboth phenomenologically simple and experimentallyaccessible. The peripheral nervous system of the trunkdevelops alongside the mesodermal somites, whichappear in a segmented chain adjacent to the neural tube(Fig. 1). The somites first cleave from the unsegmentedmesoderm as epithelial rosettes, and subsequentlydifferentiate into dermomyotome and sclerotome(Keynes and Stern, 1988). The earliest components ofthe peripheral nervous system, migrating neural crestcells and outgrowing motor and sensory axons, areconfined to each anterior (cranial) half-sclerotomewhen they encounter the somite mesoderm (Fig. 1;Keynes and Stern, 1984; Keynes and Stern, 1988). Withthe exception of a few presumptive sensory neuronsthat later migrate or degenerate, they appear to avoidposterior half-sclerotome, and do so also after itssurgical displacement or enlargement in ovo (Keynesand Stern, 1984; Stern and Keynes, 1987; Stern et al.1991).

This positional restriction produces a segmentedpattern of spinal nerves along the anterior-posterioraxis, and probably serves to ensure a correct anatomicalrelationship between the axial peripheral nerves andthe flexible vertebral column (Keynes and Stern, 1988).

132 R. J. Keynes, K. F. Jaques and G. M. W. Cook

Neural Tube Dorsal Root Ganglion

Somite

Sympathetic Ganglion

In principle, it could result from attractive influences inanterior half-sclerotome and/or repulsive influences inposterior half-sclerotome (Keynes and Stern, 1984).Prior ablation of the somites (Lewis et al. 1981) orsclerotomes (Tosney, 1988) abolishes axonal segmen-tation but does not prevent axonal growth, suggesting aprominent role for repulsion by posterior half-sclero-tome in generating the segmental pattern.

A candidate molecular system for mediating axonrepulsion was provided by the observation that, in thechick embryo, the plant lectin, peanut agglutinin(PNA) binds the cells of posterior but not anterior half-sclerotome (Stern etal. 1986). Subsequent SDS-PAGEanalysis of PNA-binding material from embryonictrunks showed two components (48xlO3Mr and55 x 103 Mr) which correspond to the major bands seenin posterior but not anterior half-sclerotome in onedimensional separations. The same PNA-binding glyco-protein fraction causes a three-fold reduction overcontrol values for axon outgrowth when presented as asubstrate in vitro, confirming that it might indeed repelgrowing axons (Davies et al. 1990). Using the growthcone collapse assay (Fig. 2), Davies et al. (1990) havealso shown that detergent extracts of chick sclerotomesor trunks cause collapse of motor and sensory growthcones extending on a laminin substratum (Fig. 3). Thisactivity is eliminated from the extracts by the use of

Fig. 1. Spinal nerve segmentationin the chick embryo. (A) Thesomites have differentiated intodermatome, myotome andsclerotome and neural crest cells(heavy arrows) migrate from thedorsal neural tube into theanterior (left in the figure) half ofeach sclerotome. The dorsal rootganglion and spinal roots alsodevelop in anterior half-sclerotome. Reproduced fromKeynes and Stern (1985).(B) Segmental outgrowth of motoraxons in a whole-mounted 3-daychick embryo. Motor axons,stained with zinc iodide-osmiumtetroxide, have grown out fromthe neural tube (below) and areconfined to the anterior (left)halves of the sclerotomes. Thesegment boundaries are indicatedby arrowheads. Scale bar 50 fim.

immobilised PNA. Rabbit polyclonal antibodies, di-rected against the 48xlO3Mr and 55xlO3Mr com-ponents and applied to sections of chick embryo trunks,bind to posterior half-sclerotome in preference toanterior half-sclerotome. When immobilised on asuitable support, they can also be used to eliminatecollapse activity from detergent extracts of somitematerial (Davies et al. 1990).

These observations strongly suggest that the PNA-binding material is responsible for excluding outgrow-ing axons from the posterior half-sclerotome in vivo.Consistent with this possibility, the lectin receptor isfound to be associated with the surface of isolatedposterior half-sclerotome cells; on treatment withfluorescent PNA a characteristic ring reaction is seen,followed by patching and capping (Davies et al. 1990).Although axons are able to grow on substrates ofposterior cells in vitro, under these circumstances theymay avoid those regions of the cell surface bearing thelectin receptor (Stern et al. 1986). During normaldevelopment, such conditions for escape from thesurface-associated contact repellant would not exist,and the entire posterior half-sclerotome would providean exclusion zone for trespassing axons. If so, a criticaltest of this hypothesis will be to administer in ovosuitable antibodies against the 48xlO3A/r and/or55xlO3Mr components (that block, for example, the

Contact inhibition and axon guidance 133

DISSECTION

HOMOGENISE,•- DETERGENT

SOLUBILISE». CENTRIFUGE

100,000 g for 1h

PHOSPHOLIPIDin DETERGENT

CULTURE OFNEURONS

ON LAMININ

Fig. 2. Growth cone collapse assay based on the method of Raper and Kapfhammer (1990). Detergent (CHAPS) extractsof somites are mixed with phospholipids and the detergent is then removed by dialysis. The resulting liposomes,incorporating the inhibitory material, are added to explants of chick dorsal root ganglia (DRG) growing on a lamininsubstratum. Cultures are fixed one hour later and scored for the presence of growth cone collapse using phase contrastoptics. Reproduced from Keynes et al. (1991).

growth cone collapse in vitro), which should allow axonentry.

The full growth cone collapse detected in the assaydoes not necessarily imply that the same takes place invivo. Nor has it been seen in cocultures of axons and

posterior half-sclerotome cells (Tosney, 1987). When amotor axon exits the neural tube opposite a posteriorhalf-sclerotome, a single filopodial contact with aposterior half-sclerotome cell may result in localcollapse or paralysis of the growth cone and avoidance

134 R. J. Keynes, K. F. Jaques and G. M. W. Cook

80 n

control sclerotome

B

•I IFig. 3. Sclerotome extracts cause growth cone collapse.(A) A marked increase is seen in the percentage ofcollapsed growth cones in cultures treated with liposomesincorporating sclerotome extract ('sclerotome', 7 cultures)compared with those treated with plain liposomes('control', 7 cultures). (B) Characteristic spreadmorphology of DRG growth cones (treated with plainliposomes) extending on a laminin substratum. (C) Growthcone collapse one hour after addition of liposomesincorporating somite extract. B,C, magnification 700x.Reproduced from Davies et al. (1990).

of this region. The direction that the growth conesubsequently takes along the anterior-posterior axis isnot random, however. Retrograde labelling of outgrow-ing motor axons, by the application of horseradishperoxidase to the anterior half-sclerotome, shows thatsuch axons grow consistently towards the nearest-available anterior half-sclerotome (Lim, T. M., Jaques,

K. F., Stern, C. D. and Keynes, R. J., 1991; Fig. 4).This directionality is most easily explained by thecoexistence of a chemotropic cue for axon guidance,emitted by cells in the anterior half-sclerotome (Fig. 5).

Molecular mechanisms

The relationship of the 48xlO3Mr and 55xlO3Mrcomponents to one another in the native state has yet tobe determined, and a full characterisation of thepeptide moieties will await their molecular cloning. Wehave investigated, meanwhile, the role of carbohydrategroups in executing growth cone collapse. One possi-bility would be that the carbohydrate group that bindsPNA directly, the disaccharide residue Gal/?l-3Gal-NAc, might also be directly responsible for executingcollapse, binding to an appropriate receptor on thegrowth cone. That this is not the case is shown by theobservation (Davies, J. A., Cook, G. M. W. andKeynes, R. J., unpublished observations) that lipo-somes incorporating the ganglioside GM1 (at a concen-tration of lmgml ), which bears the same disacchar-ide sequence at the non-reducing terminus (Fig. 6), donot elicit growth cone collapse, whether the gangliosideis in sialylated or non-sialylated form. On the otherhand, pretreatment of liposomes incorporating deter-gent extracts of chick embryo trunks with specificcombinations of endo- and exo-glycosidases doesremove collapse activity (Wajed, S., Howells, R.,Cook, G. M. W. and Keynes, R. J., unpublishedobservations). This indicates the involvement of glycanmoieties in mediating biological activity, either directly(for example, providing a ligand for a growth conereceptor) or indirectly (by maintaining the optimaltertiary structure of the collapse-inducing molecule).The fact that collapse activity survives heat treatment(incubation of extracts at 100 °C for 3min prior toincorporation into liposomes) may favour the formerpossibility.

Besides the PNA receptor, a variety of othermolecular differences between the anterior and pos-terior half-sclerotome have been detected (see Table 1for summary). In many cases it has been difficult toassign a critical functional role for these molecules,either as adhesion or repulsion systems for nerve cells,because their spatiotemporal expression patterns do notcorrelate precisely with neural crest migration and/orspinal axon outgrowth. To take one example, the70xl03Mr antigen described by Tanaka et al. (1989) isrestricted to the anterior half-sclerotome below uppercervical segmental levels, a pattern consistent with apossible role as an adhesion system; but against this, ithas no anterior-posterior polarity of expression in theupper cervical sclerotomes.

As regards possible repulsion systems in addition tothe peanut lectin-binding glycoconjugate, both chon-droitin sulphate proteoglycans and T-Cadherin arecandidates in view of their selective expression inposterior half-sclerotome. Tan et al. (1987) havecorrelated the expression in the somite mesoderm of a

Contact inhibition and axon guidance 135

NT

DM

Fig. 4. Longitudinal sectionthrough a stage 19 chick embryotrunk, following application ofhorseradish peroxidase (HRP)to the anterior (right) half-sclerotome in two (non-contiguous) segments.Retrograde transport of HRPfrom the site of applicationlabels motor neurons in theneural tube (NT). Thepopulation of motor neuronsprojecting into each somite isplaced symmetrically about theanterior half-sclerotome, and isout of phase with the somites bya distance of approximately onequarter somite. Upperarrowheads denote the anterior-posterior margins of the neuronpopulations, lower arrowheadsdenote the somite boundaries.DM, dermomyotome. Scale bar50 jum.

cytotactin-binding chondroitin sulphate proteoglycan(CTB, core protein 280xl03A/r) with neural crestmigration. During the early stages of crest migrationinto the sclerotome the proteoglycan is evenly distrib-uted in both sclerotome halves, and it later adopts aposterior location. Essentially similar expression pat-terns have been described using antibodies againstnative chondroitin sulphate and chondroitin-6-sulphate(Newgreen et al. 1990; Perris et al. 1991). The CTBproteoglycan also provides a poor substrate for crestattachment and migration in vitro. Because its polarisedexpression pattern arises comparatively late, it seemsunlikely to play a dominant role in repulsion fromposterior half-sclerotome, but (along with other chon-droitin sulphate proteoglycans) it may contribute to thisprocess, both in relation to the crest and also to axongrowth cones.

T-Cadherin is expressed as soon as the sclerotomeA P J,

Concentrationol

Repulsionmolecule

M O T O R A X O N S

dissociates from the epithelial somite and, therefore,appears early enough to influence the pathway of neuralcrest cell migration from its initiation (Ranscht andBronner-Fraser, 1991). Its functional effects are un-known, however, and it is equally plausible that it actsas an intercellular adhesion molecule for the cells of theposterior half-sclerotome themselves. As for the PNA-receptor, an important test for the function of any ofthese molecules will be to interfere with their action invivo, either by the use of antibodies or with transgenictechniques.

Cranial nerve segmentation

If the segmented pattern of spinal nerves is generatedby segmentation and repulsion in the adjacent meso-derm, is the same true for the segmentation of cranial

Fig. 5. Model for axon guidance duringperipheral nerve segmentation. Boundariesbetween anterior (A) and posterior (P)half-sclerotomes are delineated by finedashed lines. Motor axons (arrows)emerging opposite the posterior (right inthe figure) part of P-half sclerotome turnposteriorly, while those opposite theanterior part turn anteriorly, in both casesreaching the nearest-available A-halfsclerotome. This directionality cannoteasily be explained solely on the basis ofrepulsion by P half-sclerotome, and maybe provided instead by a chemotropicinfluence diffusing from the adjacent Ahalf-sclerotomes. Axonal segmentationresults from repulsion by P-halfsclerotome, while growth in A half-sclerotome is permitted by the presencehere of suitable adhesion molecules.(Reproduced from Keynes et al. 1991).

Concentrationof

Chemoattractant

136 R. J. Keynes, K. F. Jaques and G. M. W. Cook

Gal 0 i-3-GalNAcei

NANA-a2-3-Gal

GlcCerGal = GalactoseGalNAc = N-Acetyl galactosamineNANA = N-Acetyl neuraminic acid (sialic acid)GlcCer = Glucosylceramide

Fig. 6. Structure of GM1.

nerves? There are interesting differences between themanifestations of segmentation in the trunk and cranialregions. In contrast to the developing spinal cord (Limet al. 1991) the neural epithelium of the developinghindbrain displays conspicuous segmentation (rhombo-meres; Lumsden and Keynes, 1989), while the paraxialmesoderm is overtly segmented in the trunk (somites)but less obviously so in the cranial paraxial mesoderm(somitomeres; Anderson and Meier, 1981; Jacobson,1988).

In the trunk, anterior-posterior reversals of theneural tube, before axon outgrowth, do not alter theperipheral axon trajectories in the somite mesoderm

(Keynes and Stern, 1984). We have recently testedwhether this also holds for the cranial region. Fig. 7shows the effects of anterior-posterior reversals of theearly chick embryo hindbrain on the outgrowth patternof the trigeminal and facial nerves. In normal embryosthe trigeminal motor nerve exits the hindbrain onlyfrom the 2nd rhombomere, while it recruits axons fromrhombomeres 2 and 3, and the trigeminal sensoryganglion also lies adjacent to the 2nd rhombomere;motor and sensory axons merge to grow directly intothe adjacent mesenchyme of the 1st branchial arch.Similarly, the facial motor nerve exits from the 4thrhombomere, recruiting from rhombomeres 4 and 5,the facial ganglion lies opposite the 4th rhombomere,and motor and sensory axons grow into the 2ndbranchial arch (Fig. 7, normal; Lumsden and Keynes,1989). When the reversal is such that the positions ofrhombomeres 2 and 4 relative to the adjacent meso-derm are directly transposed (Fig. 7, symmetricalrotation), the nerves (visualised by whole-mountneurofilament staining) grow separately into the archesimmediately adjacent, as in normal embryos, and showlittle tendency to alter their trajectories outside theneural tube (17/28 embryos). As a result, the 1st arch iscolonised inappropriately by axons from rhombomeres4 and 5, while the 2nd arch is colonised by axons fromrhombomeres 2 and 3. When, however, the reversalsimultaneously shifts rhombomeres 2 and 4 onesegment caudal to the normal positions of the nerve

Table 1. Distribution of molecules in avian sclerotome during the phase of neural crest migration and axonoutgrowth. Those expressed directly by neural crest cells and axons are excluded from the table

Molecular species Location Ref.

Anterior (A) half Posterior (P) half

Lectin receptorsPeanut agglutininJacalin

Extracellular matrix moleculesCytotactin/tenascin/Jl

Chondroitin sulphate/CTB proteoglycan

Heparan sulphate proteoglycanKeratan sulphate proteoglycanHyaluronanLamininFibronectin

Cell adhesion moleculesN-CAM

EnzymesButyrylcholinesterase

Others70xl03Mr antigen

T-CadherinHNK-1 epitope (quail)Differences resolved on 2D gels

A (later stages) P (early stages)Uniform distribution at early stages

P (later stages)Uniform distribution in both halvesVariable, A/P distribution depending on epitopeUniform distribution in both halvesUniform distribution in both halvesUniform distribution in both halves

Uniform distribution in both halves

A

A (lower cervical segments and below)Uniform distribution in upper cervical segments

PA

Variable according to region and stage

1,2,3,43

5,67,8

5,8,99998,10,11

6,10,11,12

12,13

14

1516817

References:1 Stern et al. 1986; 2 Bagnall and Sanders, 1989; 3 Davies et al. 1990; 4 Schroeter et al. 1990; 5 Tan et al. 1987; 6 Mackie et al. 1988; 7 Stern

et al. 1989; 8 Newgreen et al. 1990; 9 Perris et al. 1991; 10 Rickmann et al. 1985; " Krotoski et al. 1986; 12 Duband et al. 1987; 13 Tosney etal. 1986; 14 Layer et al. 1988; 15Tanaka et al. 1989; 16 Ranscht and Bronner-Fraser, 1991; 17 Norris et al. 1989.

Contact inhibition and axon guidance 137

ASYMMETRICALROTATION NORMAL

SYMMETRICALROTATION

Fig. 7. Effects of anterior-posterior reversal of the embryonic chick hindbrain (at stages 9-12) on the pattern of outgrowthof the trigeminal (V) and facial (VII) cranial nerve trunks. Above, diagrams showing the rhombomere positions followingreversal. The normal pattern is shown in the middle; the positions marked X indicate the levels of trigeminal (upper) andfacial (lower) nerve roots in the paraxial mesoderm, which also correspond to the rhombomere levels at which these nervesexit and enter the hindbrain (shaded). Following 'symmetrical' rotation (right), the rhombomere exit/entry levels aredirectly transposed and continue to lie adjacent to position X in the paraxial mesoderm. After 'asymmetrical' rotation (left)the rhombomere exit/entry levels are shifted one segment posterior to level X in the paraxial mesoderm. Below, Whole-mounted, neurofilament-stained embryos. Right, normal embryo, showing positions of trigeminal (V) and facial (VII)nerves, growing into their respective branchial arches. A nerve-free zone separates the two main trunks. Left, embryo afterasymmetrical rotation; axons have grown in the characteristic peripheral patterns of V and VII (arrowheads), and a largetrunk (arrow) unites the two cranial nerves immediately adjacent to the neural tube.

roots in the paraxial mesoderm (Fig. 7, asymmetricalrotation), in only 1/26 embryos is the normal patternobserved. In the remainder, axons turn along theanterior-posterior axis immediately adjacent to theneural tube, so merging and fasciculating with axons ofthe adjacent cranial nerve, before entering the archmesenchyme (Fig. 7).

The simplest interpretation of this result is that tissueadjacent to rhombomeres 2 and 4 in normal embryos ispermissive for axon outgrowth, whereas that oppositerhombomeres 3 and 5 is non-permissive. Positive cuesmay be provided by the cranial neural crest cells (whichhad migrated from the neural tube before reversal in

the majority of grafted embryos). It is also possible thata system of paraxial exclusion zones for nerve cellsexists in the cranial mesoderm as well as the trunkmesoderm. If so, it will be important to establish therelationship of such regions to the somitomeres(Anderson and Meier, 1981; Jacobson, 1988), whosesignificance is somewhat uncertain (Keynes and Lums-den, 1990).

Comparison with other systems

Recent progress in the analysis of other developmental

138 R. J. Keynes, K. F. Jaques and G. M. W. Cook

systems where growth cone repulsion may be involvedhas been reviewed elsewhere (Keynes et al. 1991). Onesystem, which bears particular comparison with periph-eral nerve segmentation, concerns the development ofan ordered retinotectal projection. In the normal avianprojection, nasal and temporal retinal axons projectrespectively to posterior and anterior optic tectum, andtheir accurate targeting requires them to respondappropriately to directional cues on the tectal surface.The nature of at least one of these cues appears to berepulsive for a subset of retinal axons.

In an elegant analysis, Bonhoeffer and colleagues(Walter et al. 1987a,b) presented growing retinal axonswith substrata of alternating stripes of anterior andposterior tectal membranes, and found that temporalretinal axons remain exclusively on anterior stripeswhile nasal axons show no anterior-posterior prefer-ence. The preference of temporal axons for anteriormembranes was abolished by pretreatment of posteriormembranes with heat or protease (Walter et al. 19876),or phospholipase C (Walter et al. 19906), suggestingthat it arises because of repulsion by a component ofposterior membranes rather than preferential attractionby anterior membranes. Consistent with this possibility,posterior tectal membranes also cause the collapse oftemporal, but not nasal, retinal growth cones (Cox etal.1990).

Stahl et al. (1990) have now identified a 33xlO3Mrglycoprotein as the candidate repulsion molecule.Liposomes incorporating detergent-solubilised materialfrom embryonic chick brain membranes carry the33xlO3Mr component and replicate the stripe-guidingand collapse-inducing properties of posterior tectalmembranes. Only this component of the liposomes isalso concentrated in posterior rather than anteriortectal membranes, shows sensitivity to phospholipase Cand is subject to developmental regulation.

An interesting property of this system is that theresponding (temporal) axons habituate on continuedexposure to the repulsion stimulus: that is, temporalaxons are able to grow normally on a substratumcomposed exclusively of posterior membranes (Walteret al. 1987a). Walter et al. (1990a) have recentlyproposed a model for axon guidance in the tectum inwhich habituation of the growth cone plays animportant role. During spinal nerve segmentation, bycontrast, habituation appears not to take place and,indeed, would not be anticipated. Unlike the tectum,the posterior half-sclerotome must exclude all axons.Consistent with this, axon extension in vitro issubstantially reduced on a substrate of repellantmaterial (see above).

Conclusions

Growth cone guidance mechanisms involving contactinhibition may prove to be widespread during ver-tebrate embryonic development. In the periphery,regions such as developing limb girdle mesoderm(Tosney and Landmesser, 1985), peri-notochordal

mesenchyme (Tosney and Oakley, 1990; Tosney, 1991)and epidermis (Verna, 1985) are also inhibitory togrowing axons. In the CNS, midline systems such as theoptic chiasm (Godement et al. 1990) and floor plate(Kuwada and Bernhardt, 1990) have been suggested torepel certain axons while allowing others to decussate,and growth cone collapsing activity is found indetergent extracts of embryonic chick brain (Raper andKapfhammer, 1990). Lastly, growth cone-repellingactivities are present in both the white matter (Schwab,1990) and grey matter (Keynes et al. 1990) of the adultbrain, and may contribute to the failure of axonregeneration following injury. To establish the extent towhich these various systems are related, and to devisestrategies for modifying their action, now depends onmolecular cloning of the agents involved and, ulti-mately, detailed comparison of their structure andfunction.

We are grateful to T. M. Lim for Fig. 4. This work issupported by grants from Action Research for the CrippledChild, The Medical Research Council and The WellcomeTrust. G.M.W.C. is a Member of the External Scientific Staffof the Medical Research Council.

References

ABERCROMBIE, M. AND HEAYSMAN, J. E. M. (1954). Observationson the behaviour of cell contacts in tissue culture:II. Monolayering of fibroblasts. Expl Cell Res. 6, 293-306.

ANDERSON, C. B. AND MEIER, S. (1981). The influence of themetameric pattern in the mesoderm on migration of cranialneural crest cells in the chick embryo. Devi Biol. 85, 385-402.

BAGNALL, K. M. AND SANDERS, E. J. (1989). The binding patternof peanut lectin associated with sclerotome migration and theformation of the vertebral axis in the chick embryo. Anat.Embryo!. 180, 505-513.

Cox, E. C , MOLLER, B. AND BONHOEFFER, F. (1990). Axonalguidance in the chick visual system: posterior tectal membranesinduce collapse of growth cones from the temporal retina.Neuron 4, 31-37.

DAVIES, J. A., COOK, G. M. W., STERN, C. D. AND KEYNES, R. J.(1990). Isolation from chick somites of a glycoprotein fractionthat causes collapse of dorsal root ganglion growth cones.Neuron 4, 11-20.

DUBAND, J.-L., DUFOUR, S., HATTA, K., TAKEICM, M., EDELMAN,G. M. AND THIERY, J. P. (1987). Adhesion molecules duringsomitogenesis in the avian embryo. J. Cell Biol. 104, 1361-1374.

GODEMENT, P., SALAON, J. AND MASON, C. A. (1990). Retinalaxon pathfinding in the optic chiasm: divergence of crossed anduncrossed fibers. Neuron 5, 173-186.

JACOBSON, A. G. (1988). Somitomeres: mesodermal segments ofvertebrate embryos. Development 104 Supplement, 209-220.

KAPFHAMMER, J. P., GRUNEWALD, B. E. AND RAPER, J. A. (1986).The selective inhibition of growth cone extension by specificneurites in culture. J. Neurosci. 6, 2527-2534.

KAPFHAMMER, J. P. AND RAPER, J. A. (1987). Collapse of growthcone structure on contact with specific neurites in culture. J.Neurosci. 7, 201-212.

KEYNES, R. AND LUMSDEN, A. (1990). Segmentation and the originof regional diversity in the vertebrate central nervous system.Neuron 4, 1-9.

KEYNES, R. J., JOHNSON, A. R. AND COOK, G. M. W. (1991).Contact inhibition of growth cone motility during neuraldevelopment and regeneration. Semin. Neurosci. 3, 321-330.

KEYNES, R. J., JOHNSON, A. R., PICART, C. J., DUNIN-BORKOWSKI,O. M. AND COOK, G. M. W. (1990). A glycoprotein fractionfrom adult chicken grey matter causes collapse of CNS and PNSgrowth cones in vitro. Abstr. Soc. Neurosci. 16, 77.6.

Contact inhibition and axon guidance 139

KEYNES, R. J. AND STERN, C. D. (1984). Segmentation in thevertebrate nervous system. Nature 310, 786-789.

KEYNES, R. J. AND STERN, C. D. (1985). Segmentation and neuraldevelopment in vertebrates. Trends Neurosci. 8, 220-223.

KEYNES, R. J. AND STERN, C. D. (1988). Mechanisms of vertebratesegmentation. Development 103, 413-429.

KROTOSKI, D. M., DOMINGO, C. AND BRONNER-FRASER, M. (1986).

Distribution of a putative cell surface receptor for fibronectinand laminin in the avian embryo. J. Cell Biol. 103, 1061-1071.

KUWADA, J. Y. AND BERNHARDT, R. R. (1990). Axonal outgrowthby identified neurons in the spinal cord of zebrafish embryos.Expl Neurol. 109, 29-34.

LAYER, P. G., ALBER, R. AND RATHJEN, F. G. (1988). Sequential

activation of butyrylcholinesterase in rostral half somites andacetylcholinesterase in motoneurones and myotomes precedinggrowth of motor axons. Development 102, 387-396.

LEWIS, J., CHEVALLIER, A., KIENY, M. AND WOLPERT, L. (1981).

Muscle nerve branches do not develop in chick wings devoid ofmuscle. J. Embryo!, exp. Morph. 64, 211-232.

LIM, T.-M., JAQUES, K. F., STERN, C. D. AND KEYNES, R. J.(1991). An evaluation of myelomeres and segmentation of thechick embryo spinal cord. Development 113, 227-238.

LUMSDEN, A. AND KEYNES, R. (1989). Segmental patterns ofneuronal development in the chick hindbrain. Nature 337,424-428.

MACKIE, E. J., TUCKER, R. P., HALFTER, W., CHIQUET-EHRISMANN,

R. AND EPPERLEIN, H. H. (1988). The distribution of tenascincoincides with pathways of neural crest cell migration.Development 102, 237-250.

NEWGREEN, D. F., POWELL, M. E. AND MOSER, B. (1990).

Spatiotemporal changes in HNK-1/L2 glycoconjugates on avianembryo somite and neural crest cells. Devi Biol. 139, 100-120.

NORRIS, W. E., STERN, C. D. AND KEYNES, R. J. (1989). Molecular

differences between the rostral and caudal halves of thesclerotome in the chick embryo. Development 105, 541-548.

PATTERSON, P. H. (1988). On the importance of being inhibited, orsaying no to growth cones. Neuron 1, 263-267.

PERRIS, R., KROTOSKI, D., LALLIER, T., DOMINGO, C , SORRELL, J.M. AND BRONNER-FRASER, M. (1991). Spatial and temporalchanges in the distribution of proteoglycans during avian neuralcrest development. Development 111, 583-599.

RANSCHT, B. AND BRONNER-FRASER, M. (1991). T-Cadherinexpression alternates with migrating neural crest cells in thetrunk of the avian embryo. Development 111, 15-22.

RAPER, J. A. AND KAPFHAMMER, J. P. (1990). The enrichment of aneuronal growth cone collapsing activity from embryonic chickbrain. Neuron 4, 21-29.

RICKMANN, M., FAWCETT, J. W. AND KEYNES, R. J. (1985). The

migration of neural crest cells and the growth of motor axonsthrough the rostral half of the chick somite. J. Embryol. exp.Morph. 90, 437-455.

SCHROETER, S., LASKY, C. J., OAKLEY, R. A., ERICKSON, C. A.

AND TOSNEY, K. W. (1990). Evidence for the delineation ofneural crest migration pathways by inhibitory boundaries. Abstr.

• Soc. Neurosci. 16, 140.7.SCHWAB, M. E. (1990). Myelin-associated inhibitors of neurite

growth and regeneration in the CNS. Trends Neurosci. 13,452-456.

STAHL, B., MOLLER, B., VON BOXBERG, Y., COX, E. C. ANDBONHOEFFER, F. (1990). Biochemical characterization of a

putative axonal guidance molecule of the chick visual system.Neuron 5, 733-743.

STERN, C. D., ARTINGER, K. B. AND BRONNER-FRASER, M. (1991).

Tissue interactions affecting the migration and differentiation ofneural crest cells in the chick embryo. Development 113,207-216.

STERN, C. D. AND KEYNES, R. J. (1987). Interactions betweensomite cells: the formation and maintenance of segmentboundaries in the chick embryo. Development 99, 261-272.

STERN, C. D., NORRIS, W. E., BRONNER-FRASER, M., CARLSON, G.

J., FAISSNER, A., KEYNES, R. J. AND SCHACHNER, M. (1989).

Jl/tenascin-related molecules are not responsible for thesegmented pattern of neural crest cells or motor axons in thechick embryo. Development 107, 309-319.

STERN, C. D., SISODIYA, S. M. AND KEYNES, R. J. (1986).

Interactions between neurites and somite cells: inhibition andstimulation of nerve growth in the chick embryo. J. Embryol.exp. Morph. 91, 209-226.

TAN, S. S., CROSSIN, K. L., HOFFMAN, S. AND EDELMAN, G. M.

(1987). Asymmetric expression in somites of cytotactin and itsproteoglycan ligand is correlated with neural crest celldistribution. Proc. natn. Acad. Sci. U.S.A. 84, 7977-7981.

TANAKA, H., AGATA, A. AND OBATA, K. (1989). A new membrane

antigen revealed by monoclonal antibodies is associated withmotoneuron axonal pathways. Devi Biol. 132, 419-435.

TOSNEY, K. W. (1987). Growth cone interactions with anterior andposterior sclerotome. Abstr. Soc. Neurosci. 13, 130.6.

TOSNEY, K. W. (1988). Proximal tissues and patterned neuriteoutgrowth at the lumbosacral level of the chick embryo: partialand complete deletion of the somite. Devi Biol. 127, 266-286.

TOSNEY, K. W. (1991). Cells and cell-interactions that guide motoraxons in the developing chick embryo. BioEssays 13, 17-23.

TOSNEY, K. W. AND OAKLEY, R. A. (1990). Perinotochordalmesenchyme acts as a barrier to axon advance in the chickembryo: implications for a general mechanism of axon guidance.Expl Neurol. 109, 75-89.

TOSNEY, K. W. AND LANDMESSER, L. T. (1985). Development ofthe major pathways for neurite outgrowth in the chick hindlimb.Devi Biol. 109, 193-214.

TOSNEY, K. W., WATANABE, M., LANDMESSER, L. AND

RUTISHAUSER, U. (1986). The distribution of NCAM in the chickhindlimb during axon outgrowth and synaptogenesis. Devi Biol.114, 437-452.

VERNA, J. M. (1985). In vitro analysis of interactions betweensensory neurons and skin: evidence for selective innervation ofdermis and epidermis. J. Embryol. exp. Morph. 86, 53-70.

WALTER, J., ALLSOPP, T. E. AND BONHOEFFER, F. (1990a). A

common denominator of growth cone guidance and collapse?Trends Neurosci. 13, 447-452.

WALTER, J., HENKE-FAHLE, S. AND BONHOEFFER, F. (19876).

Avoidance of posterior tectal membranes by temporal retinalaxons. Development 101, 909-913.

WALTER, J., KERN-VEITS, B., HUF, J., STOLZE, B. AND

BONHOEFFER, F. (1987a). Recognition of position-specificproperties of tectal cell membranes by retinal axons in vitro.Development 101, 685-696.

WALTER, J., MOLLER, B. AND BONHOEFFER, F. (19906). Axonal

guidance by an avoidance mechanism. J. Physiol. (Paris) 84,104-110.