ABSTRACT
We have previously reported that the cell-type distribution and pattern of expression of the surface antigen recognized by the monoclonal antibody 69A1, suggests that it may be involved during the period of nerve fibre outgrowth and the formation of fibre bundles in the rat (Pigott & Kelly, 1986). In this current study, we have examined the expression of the epitope recognized by antibody 69A1 in regions of the rat central nervous system in which it is possible to distinguish between neurones with axons that fasciculate to form clearly defined fibre tracts and neurones with non-fasciculating processes. We have also examined antibody 69Al labelling in several regions of the peripheral nervous system. We report that the 69A1 epitope is expressed on neurones with axons that fasciculate but is not found on neurones with short, non-fasciculating axons or on neurones without a morphologically identifiable axon. The antigen 69Al has been purified and shown to be immunochemically closely related or identical to the LI antigen.
INTRODUCTION
The association of nerve fibres in well-defined fascicu-li is a prominent feature of the organization of the nervous system. The most convincing evidence that fasciculi are established by adhesive interactions between juxtaposed growing nerve fibres comes from work on a related group of cell-surface glycoproteins termed cell-adhesion molecules or CAMs (Edelman, 1985) . A number of studies has shown that Fab’ fragments of antibodies against either neural CAM (N-CAM) or neurone-glia CAM (Ng-CAM, which is immunochemically identical to the molecules NILE and LI antigen, Bock, Richter-Lansberg, Faissner & Schachner, 1985; Sajovic, Kouvelas & Trenkner, 1986) reduce the thickness of neurite fasciculi emerging from neural explants in culture (Rutishauser, Gall & Edelman, 1978; Stallcup & Beasley, 1985; Fischer, Kunemunde & Schachner, 1986). The recent demonstration that antibodies to Ng-CAM are more effective in this respect (Hoffman et al. 1986) together with the finding that the expression of Ng-CAM/Ll/NILE is higher on nerve fibres than perikarya during the development of fibre tracts (Faissner et al. 1984b; Stallcup, Beasley & Levine, 1985; Thiery, Delouvee, Grumet & Edelman, 1985), suggest that Ng-CAM plays a more important role than N-CAM in this process.
It is not clear however whether the Ng-CAM/Ll/ NILE molecule is exclusively involved in the formation of fibre tracts or whether its role in axonal fasciculation is simply one aspect of a more general role in mediating a variety of adhesive interactions between neurones. Restriction of this molecule to neurones with axons that form into fibre bundles would however be evidence of a specific role in fasciculation.
The monoclonal antibody 69A1 recognizes an antigen which has been shown biochemically to be identical or closely related to the Ng-CAM/Ll/NILE molecule (Pigott & Kelly, 1984; Pigott & Kelly, 1986). In this current paper, we report the purification of the antigen 69A1 and show that it is also immunochemically indistinguishable from the LI antigen (Rathjen & Schachner, 1984). We have therefore used this antibody to study the neuronal cell-type distribution of the antigen in a number of regions of the developing rat nervous system in which it is possible to distinguish neurones with axons that fasciculate from those with non-fasciculating processes.
Materials and methods
Neural cultures
Cerebella and olfactory bulbs were obtained from postnatal rats less than 2 days old and retinal tissue from rats less than 12 h after birth. Dissociated cultures were prepared as follows (Dutton, Currie & Tear, 1981). Dissected tissue was incubated for 15 min at 37°C with 0·025% trypsin (Worthington) in calcium- and magnesium-free phosphate-buffered saline (CMF PBS) supplemented with 3 mg ml-1 bovine serum albumin (BSA) and 2 mg ml-1 glucose. En-zyme digestion was terminated by the addition of 80 μg ml-1 soya bean trypsin inhibitor (Sigma). The tissue was pelleted by centrifugation and was dissociated into a single-cell suspension in CMF PBS with 40μgml-1 DNase and 31 mM-MgSO4 by gentle trituration using a fire-polished Pasteur pipette. Cellular debris was removed by pelleting the cells at 150g through a cushion of CMF PBS containing 40mgml-1 BSA. The cells were cultured in multiwell plates on 13 mm diameter glass coverslips precoated with 500μgml-1 polyomithine and 10/tgml-1 laminin at a den-sity of 15X105 to 2·0x105 cells per coverslip. The culture medium was Ham’s F14 supplemented with 10 % heat-inactivated horse serum and 2 % chick embryo extract. For cerebellar and olfactory bulb cultures the potassium ion concentration was raised to 23 mM.
Dissociated cultures of dorsal root ganglia (DRG) and paravertebral sympathetic ganglia (SG) were established from newborn rats by a similar method, except that the tissue was incubated with trypsin at a concentration of 0·5 % for 30 min. For cultures of the myenteric plexus the tissue was additionally incubated for 45 min with 0·05 % collagenase (Worthington) prior to trypsin digestion. Cul-ture medium was Ham’s F14 supplemented with 10 % heat-inactivated horse serum. DRG and SG culture medium was additionally supplemented with NGF at a concentration of 25 ngml-1.
Immunocytochemistry
Unfixed cultures were double labelled with anti-tetanus toxin, to identify neurones (Mirsky et al. 1978) and 69A1 antibody as described previously (Pigott & Kelly, 1986). For double labelling with antibody 69A1 and anti-Thy 1, cultures were incubated simultaneously for 15 min with 500 μg ml-1 antibody 69A1 and a 1:100 dilution of F(ab’)2 fragments of rabbit anti-Thy-1 (gift of Dr R. Morris). Bound antibody was revealed as for antibody 69A1 and anti-tetanus toxin. Labelled cultures were fixed for 30min in 4 % paraformaldehyde in 0·1 M-phosphate buffer pH 7·4, rinsed in PBS and mounted in Citifluor (City University, Department of Chemistry, London). For double labelling with antibody 69A1 and anti-GABA antiserum, cultures were first fluorescently labelled with antibody 69A1. They were then fixed in 4 % paraformaldehyde and 0·05 % glutaraldehyde in 0·1 M-phosphate buffer, permeabilized with 0·2% Triton X-100 in PBS for 4 min and washed for three periods of 5 min in PBS containing 3 % goat serum (PBS/GS). The cultures were then incubated for 1 h with a 1:2000 dilution of a rabbit anti-GABA antiserum (gift of Dr P. Somogyi, Hodgson et al. 1985) in PBS/GS. After washing, the cultures were incubated for Ih with a 1:50 dilution of a goat antiserum to rabbit IgG conjugated to alkaline phosphatase (Sigma). Bound antibody was re-vealed by incubation for 30 min with 25 mM-Tris-maleate pH9 0 containing 0·15M-NaCl, 4mM-MgCl2, 0·4mgml-1 sodium alpha-naphthyl phosphate and lmgml-1 Fast Red TR salt. Cultures were viewed on a Zeiss Photomicroscope III equipped with epifluorescence illumination and photo-graphed on Kodak Tri-X film.
Immunohistochemistry
Paraformaldehyde-fixed, 20gm cryostat sections of devel-oping rat olfactory bulb and retina were obtained and labelled with antibody 69A1 as previously described (Pigott & Kelly, 1986).
Antigen 69A1 purification
Antigen 69A1 was purified from brains of 2- to 5-day postnatal rats. Crude membranes prepared from 100 rats as described by Hoffman et al. (1982) were extracted for 6 h in buffer containing 8g NaCl, 0·2g KC1, 0-2g KH2PO4 and 0·15 g Na2HPO4/litre, with 1 mM-EDTA and 0·5 % NP-40.
The immunoaffinity matrix was constructed by coupling 10 mg of Protein A-purified antibody 69A1 to 0·5 g CNBr-activated Sepharose 4B (Pharmacia). The solubilized membrane preparation was first incubated for 2h with mouse IgG coupled to Sepharose to remove components which might otherwise bind nonspecifically to the antibody matrix. Following removal of the IgG-Sepharose by centri-fugation, the solubilized membranes were incubated over-night at 4°C with the antibody 69A1 immunoabsorbent. Bound protein was recovered as described by Hoffman et al. (1986). Detergent was removed by incubation with Bio-Beads SM-2 (Bio-Rad). The sample was dialysed against H2O overnight and the purified antigen concentrated in an Amicon microconcentrator.
Purified antigen 69Al was analysed by gradient SDS-PAGE (6—12 %) and silver stained as described by Ansorge (1985). Molecular weight markers were: ovalbumin (Mr 45X103, 45K), bovine albumin (66K), phosphorylase b (97K), β-galactosidase (116K) and myosin (205K).
Antibody production
Rabbit antibodies to the 200K polypeptide were produced by fractionating the antigen 69A1 by SDS-PAGE, briefly staining the gel with toluidine blue (2-3 min) and excising the relevant band. The excised band was then rinsed in PBS, crushed and combined with adjuvant prior to injec-tion. Initial injections were of 100 μg protein followed by three further injections of 50μg. The IgG fraction was purified by (NH4)2SO4 precipitation and ion-exchange chromatography (Johnstone & Thorpe, 1982).
Antibody characterization
Crude membranes were prepared from mature rat brains as described by Hoffman et al. (1982). Proteins were fraction-ated by gradient (6-12 %) SDS-PAGE with 100 pg per lane and transferred to nitrocellulose using a semi-dry blotter (Dako) for 2h at 0·8 V cm-2 The nitrocellulose was stained briefly with Ponceau-S (Serva), both to check the fidelity of transfer and as an aid to cutting the strips. Strips were incubated overnight in PBS with 10 % horse serum (PBS/HS) at 4°C and then incubated with primary anti-body for 2h at room temperature. Antibodies used were: rabbit anti-69Al (1:2000), preimmune serum (1:500), rabbit anti-Ll (1:250, Rathjen & Schachner, 1984. Gift of Dr M. Schachner) and rabbit anti-D2/N-CAM (1:1000, Jorgensen, Delouvee, Thiery & Edelmann, 1980. Gift of Dr O. S. Jorgensen). Following four washes of 15 min in PBS/HS, the strips were incubated for 1 h with a 1:1000 dilution of a goat anti-rabbit IgG conjugated to peroxidase (Tago). Bound peroxidase was revealed by incubation with 0’5mgml-1 diaminobenzidine in 50rriM-Tris/HCl pH7·6 with 0·005% H2O2.
Results
Cerebellum
Most of the cells in these cultures were identified as neurones since they were labelled by tetanus toxin. Of these, about 95% were 69A1 positive and were morphologically similar, having a cell body less than lOjum in diameter and two or three fine neurites (Fig. 1). The remaining 5 % of neurones were 69A1 negative and after 2 to 3 days in culture were characterized by a larger cell body (10 to 15 gm in diameter) and the presence of only one, occasionally two, prominent, thick processes (Fig. 2). When cul-tures were incubated with anti-GABA antiserum and antibody 69A1, all of the GABA-positive neurones were found to be 69A1 negative (Fig. 3).
Double immunolabelling of cerebellar cells after 1 day in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-tetanus toxin. The vast majority of cells were labelled by antibody 69A1 and were identified as neurones since they were also labelled by anti-tetanus toxin. Bar, 20,μm.
Double immunolabelling of cerebellar cells after 7 days in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-tetanus toxin. Three neurones positive for the anti-tetanus toxin with a prominent thick process are unlabelled by antibody 69A1 although fine 69Al-positive neurites are apparent extending across the field of view. Bar, 20 μm.
Double immunolabelling of cerebellar cells after 7 days in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-tetanus toxin. Three neurones positive for the anti-tetanus toxin with a prominent thick process are unlabelled by antibody 69A1 although fine 69Al-positive neurites are apparent extending across the field of view. Bar, 20 μm.
Double immunolabelling of cerebellar cells after 3 days in culture. (A,C) Anti-GABA immunoperoxidase labelling; (B,D) antibody 69A1 immunofluorescence. The GABA-positive neurones (arrowed) are unlabelled by antibody 69A1. The slight fluorescence of the cell bodies is glutaraldehyde-induced autofluorescence. Bars, 10μm.
Double immunolabelling of cerebellar cells after 3 days in culture. (A,C) Anti-GABA immunoperoxidase labelling; (B,D) antibody 69A1 immunofluorescence. The GABA-positive neurones (arrowed) are unlabelled by antibody 69A1. The slight fluorescence of the cell bodies is glutaraldehyde-induced autofluorescence. Bars, 10μm.
Olfactory bulb
These cultures contained a greater diversity of mor-phologically distinguishable neurones, but as in cer-ebellar cultures the predominant type (70-80 %) had a small cell body with one to three processes. In contrast to the small neurones in cerebellar cultures, almost all of these neurones were 69A1 negative (Fig. 4). Faint labelling was, however, occasionally detectable on the processes of a small number of these neurones. The remaining 20 to 30% of neurones in these cultures had large cell bodies (15-25 μm in diameter) and one or two thick pro-cesses. A proportion of these large neurones was labelled by antibody 69A1. Initially, both the cell body and all processes were labelled (Fig. 5), but, after several days, labelling was in most cases pre-dominantly restricted to a single process, the other processes and cell body being only faintly labelled or unlabelled (Fig. 6). An accurate count of the number of these labelled neurones was not feasible owing to the great difficulty of tracing the long, convoluted, labelled fibres back to unlabelled cell bodies. We estimate, however, that between 5-10% of the total number of neurones in these cultures were 69A1 positive.
Double immunolabelling of olfactory bulb cells after 1 day in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-tetanus toxin. As in cerebellar cultures, the vast majority of cells are small neurones, however, these neurones are not labelled by antibody 69AL Bar, 20 μm.
Double immunolabelling of a large olfactory bulb neurone after 1 day in culture. (A) Phase contrast; (B) antibody 69Al; (C) anti-tetanus toxin. Bar, 10μm.
Double immunolabelling of olfactory bulb cells after 3 days in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-tetanus toxin. A large neurone is prominently labelled by antibody 69A1 on a single, fine process which appears to emerge from the ‘base’ of the cell (large arrow). Although the neurone is clearly labelled by anti-tetanus toxin this same process is only faintly labelled. The small bipolar neurone (small arrow) is unlabelled by antibody 69Al. Bar, 20μm.
Double immunolabelling of olfactory bulb cells after 3 days in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-tetanus toxin. A large neurone is prominently labelled by antibody 69A1 on a single, fine process which appears to emerge from the ‘base’ of the cell (large arrow). Although the neurone is clearly labelled by anti-tetanus toxin this same process is only faintly labelled. The small bipolar neurone (small arrow) is unlabelled by antibody 69Al. Bar, 20μm.
Retina
Antibody 69A1 labelling in these cultures was restric-ted to a minor population of neurones that had large cell bodies (20 to 40 aurn in diameter) and were also labelled by anti-Thy-1. These neurones were initially labelled by antibody 69A1 both on the cell body and processes (Fig. 7), but after several days, labelling was prominently restricted to a single long (>100μm) process (Fig. 8). The remaining neurones in these cultures were typically multipolar small cells which were both 69A1 negative and Thy-1 negative.
Double immunolabelling of a retinal ganglion cell after 1 day in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-Thy-1. Bar, 10μm.
Double immunolabelling of retinal cells after 3 days in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-Thy-1. A large cell is labelled by both antibodies although 69A1 labelling is predominant on the process rather than the cell body. Bar, 10μm.
Peripheral nervous system
All neurones in DRG, SG and MP cultures were found to be intensely labelled by antibody 69A1 (Figs 9-11). As with other neurones, labelling was initially present on cell bodies and neurites but became particularly prominent on neurites after 2-3 days in vitro. In the case of DRG neurones, labelling was marked on growth cones (Fig. 9).
Antibody 69A1 labelling of the neurites of a dorsal root ganglion cell. Labelling is apparent both on the neurites and the growth cones. Bar, 20μm.
Antibody 69A1 labelling of a sympathetic chain neurone after 1 day in culture. Bar, 20μm.
Double immunolabelling of myenteric plexus cells after 1 day in culture. (A) Phase contrast; (B) antibody 69A1; (C) anti-tetanus toxin. All the anti-tetanus- toxin-positive neurones are also labelled by antibody 69A1. Bar, 20μm.
Immunohistochemistry
Sections of the olfactory bulb were examined at postnatal day 0 (PO), P5 and P12. Labelling of fibres in the lateral olfactory tract was clearly apparent at P0 (Fig. 12C) and to a lesser extent at P5, but by P12 this staining was absent. In transverse sections of the P0 olfactory bulb there was evidence for the labelling of fibres in the area immediately below the mitral cell body layer (Fig. 12A). The granule cells were un-labelled at all ages studied, as were the plexiform layers. Prominent staining was found in the nerve fibre layer formed by the axons of the olfactory receptor cells (Fig. 12A) although the glomeruli in which these axons terminate were unlabelled.
(A) Antibody 69A1 labelling of a transverse section of the postnatal day 0 olfactory bulb. Staining is visible in the olfactory nerve layer and also on fibre bundles coursing through the granule cell layer. Abbreviations: of, olfactory nerve fibre layer; gl, glomeruli; ep, external plexiform layer; mb, mitral cell body layer; gr, granule cell layer; pv, periventricular zone. Bar, 50μm. (B) Transverse section of the postnatal day-0 olfactory bulb incubated with a control antibody of the same subclass (IgG 2a). Bar, 400μm. (C) Antibody 69A1 labelling of a longitudinal section of the postnatal day-0 olfactory bulb. Staining is visible in the olfactory nerve fibre layer (of) and the lateral olfactory tract (lot). Bar, 200μmt. (D) Antibody 69A1 labelling of the postnatal day-5 retina. Staining is visible only on the bundles of ganglion cell axons. Abbreviations: on, outer nuclear layer; op, outer plexiform layer; in, inner nuclear layer; ip, inner plexiform layer; gc, ganglion cell layer; of, optic nerve fibre layer. Bar, 200μm. (E) Section of the postnatal day-5 retina incubated with control antibody. Bar, 50μm.
(A) Antibody 69A1 labelling of a transverse section of the postnatal day 0 olfactory bulb. Staining is visible in the olfactory nerve layer and also on fibre bundles coursing through the granule cell layer. Abbreviations: of, olfactory nerve fibre layer; gl, glomeruli; ep, external plexiform layer; mb, mitral cell body layer; gr, granule cell layer; pv, periventricular zone. Bar, 50μm. (B) Transverse section of the postnatal day-0 olfactory bulb incubated with a control antibody of the same subclass (IgG 2a). Bar, 400μm. (C) Antibody 69A1 labelling of a longitudinal section of the postnatal day-0 olfactory bulb. Staining is visible in the olfactory nerve fibre layer (of) and the lateral olfactory tract (lot). Bar, 200μmt. (D) Antibody 69A1 labelling of the postnatal day-5 retina. Staining is visible only on the bundles of ganglion cell axons. Abbreviations: on, outer nuclear layer; op, outer plexiform layer; in, inner nuclear layer; ip, inner plexiform layer; gc, ganglion cell layer; of, optic nerve fibre layer. Bar, 200μm. (E) Section of the postnatal day-5 retina incubated with control antibody. Bar, 50μm.
Sections of the P0, P3, P12, P21 and mature retina were examined and at all ages labelling was found only in the nerve fibre layer which is composed of the axons of the ganglion cells (Fig. 12D). The ganglion cell bodies were unlabelled as were all other cell bodies and processes.
Antigen 69A1 purification
Antigen 69Al, purified from detergent lysates of neonatal rat brain membranes and fractionated by SDS-PAGE, appeared as two major bands with apparent relative molecular masses of 200 and 140x 103 (200K, 140K; Fig. 13). A less-prominent band was visible at 180K and a faint band was sometimes visible at 70K. The yield of antigen from 100 brains was approximately 500ug protein.
Antigen analysis
As determined by Western blot analysis an antiserum to the 200K component of the antigen 69Al reacted strongly with polypeptides at 200K, 180K and 140K and less strongly with a band at 70K (Fig. 14A). An antiserum to the LI antigen also reacted with poly-peptides of 200K, 180K and 140K but the 70K band was not visible; an additional band of 100K was however apparent (Fig. 14B). Polypeptides at 180K and 140K were also recognized by an antiserum to D2/N-CAM (Fig. 14C).
Western blots of crude membrane fractions from mature rat brain. (A) Rabbit anti-69Al; (B) rabbit anti-Ll; (C) rabbit anti-D2/N-CAM; (D) pre-immune serum.
Discussion
We have shown that in both primary cell cultures and tissue sections, antibody 69A1 labels only a subpopu-lation of neurones. We argue below that the common feature that distinguishes these neurones from un-labelled neurones is that they have axons that fascicu-late to form well-defined fibre layers or tracts.
Granule cells, the most numerous neurones in the cerebellum, have axons that fasciculate to form a prominent nerve fibre layer in the outer part of the cerebellar cortex. We have previously demonstrated that these axons are 69A1 positive in tissue sections of the developing cerebellum (Pigott & Kelly, 1986). The granule cells have a small cell body and are initially bipolar before the cell body migrates into the deeper layers of the cortex (Palay & Chan-Palay, 1973). The finding that the 69Al-positive neurones in culture also have a small cell body, a typically bipolar morphology and are by far the most numerous cell-type present, suggests that these are granule cells. The small proportion of neurones that were un-labelled by 69A1 antibody in cerebellar cultures were distinguished from the labelled neurones, not only by their larger size and distinctive morphology, but also by being specifically labelled by an anti-serum to the neurotransmitter GABA. The inhibitory intemeurones (stellate, basket and Golgi cells) and Purkinje cells of the cerebellum are GABAergic (Hokfelt & Ljungdahl, 1972), but, since Purkinje cells do not normally survive in culture, it seems likely that the GABA-positive, 69A1-negative neurones in our cultures were solely inhibitory inter-neurones. These neurones are scattered throughout the cerebellar cortex and have processes that do not fasciculate to form a distinct layer, such as formed by the granule cell axons, but intermingle with the processes and perikarya of neurones in the vicinity (Palay & Chan-Palay, 1973).
The olfactory bulb was selected as a comparison with the cerebellum since it too contains a predomi-nant population of small neurones (at the time of isolation approximately 80 % of the total number of neurones, Mair, Gellman & Gesteland, 1982). How-ever, these neurones either have relatively short axons, which do not form into an organized fibre layer, or are lacking a morphologically distinguish-able axon (Sheperd, 1972). The remaining neurones, the mitral and tufted cells, have much larger cell bodies (up to 30 μm in the case of mitral cells) and have axons that do fasciculate, forming the lateral olfactory tract. In cultures of the olfactory bulb there was a predominant population of 69Al-negative neurones with cell bodies less than 10μm in diameter and a smaller population of 69Al-positive large neurones. From their morphology, it is likely that the 69Al-negative neurones are intemeurones (mainly granule cells but also some short axon cells and periglomerular cells) and the 69Al-positive neurones are mitral and tufted cells. Further evidence that this is indeed the case comes from the studies of tissue sections which show that the lateral olfactory tract, formed by the mitral and tufted cell axons (Sheperd, 1972), was 69A1 positive but the granule cell bodies and processes w’ere unlabelled.
The majority of neurones in the retina are small intemeurones which either have axons which ramify locally or lack a morphologically distinguishable axon. These processes contribute to the formation of the apparently unordered inner and outer plexiform layers (Rodeick, 1973). Only the axons of the ganglion cells fasciculate, forming bundles on the inner aspect of the retina that pass to the optic disc and coalesce to form the optic nerve. Antibody 69A1 labelling was confined to the nerve fibre layer in sections of the retina and to a minor population of large neurones in retinal cultures. The 69Al-positive cells in culture were identified as ganglion cells since they were also labelled by an antibody to the cell-surface glycoprotein Thy-1 which has been used as a specific marker for these cells both in vivo and in vitro (Beale & Osborne, 1982; Barnstaple & Drager, 1984). Although there is some evidence that Thy-1 may also be present on a number of bipolar and amacrine cells, in addition to ganglion cells (Perry, Morris & Raisman, 1984), the size and morphology of the Thy-1-positive neurones in our cultures suggests that all of these were ganglion cells. The finding that all other neurones in retinal cultures were 69A1 negative is again consistent with the proposal that neurones that have axons that are not organized into defined fibre tracts do not express the 69Al epitope in immunochemically detectable amounts.
In both the retina and the olfactory bulb we have studied the pattern of labelling over a period that covers the major phases of neurite outgrowth that lead to the formation of the plexiform layers. At no time was labelling found in the plexiform layers, thereby discounting the possibility that the 69A1 epitope may be expressed transiently on all neurones during the period of neurite outgrowth.
Fasciculation of axons is a common feature of neurones in the developing peripheral nervous sys-tem and all neurones in cultures of DRG, SG and MP were found to be intensely 69A1 positive.
These studies have also shown that the distribution of antibody 69A1 labelling on individual neurones is not uniform. In vitro, labelling was predominant on neurites rather than cell bodies and was frequently further restricted to a single, long, often unbranched, process. The characteristic morphology of these pro-cesses suggests that they correspond to axons. This is supported by studies of tissue sections in which the nerve fibre layer of the retina, formed by the ganglion cell axons, was labelled but the outer plexiform layer, into which the dendrites of these cells project, was not. Similarly, the lateral olfactory tract, formed by axons of the mitral and tufted cells, was labelled but the external plexiform layer, into which their den-drites project, was negative.
The results of the analysis of the purified antigen 69A1 and the immunochemical evidence presented in this current paper, are consistent with our previous data suggesting that the antigen 69A1 is identical, or closely related, to the cell-surface glycoprotein known variously as LI, NILE or Ng-CAM. Antisera to both the 200K component of immunopurified antigen 69Al and immunopurified LI, detect poly-peptides of identical relative molecular masses at 200K, 180K and 140K. Under the conditions employed in this study the high molecular weight components were consistently resolved into two sep-arate components at 200K and 180K, the 180K component comigrating with the high molecular weight band of N-CAM. Reports from other labora-tories have described components of LI and Ng-CAM in this region appearing either as diffuse (Rathjen & Schachner, 1984) or as two or three closely spaced bands (Faissner et al. 19846; Grumet, Hoffman & Edelman, 1984). The additional band detected at 100K by the anti-Ll antiserum is probably due to the presence of immunoreactivity against a proteolytic degradation product (Faissner, Kruse, Nieke & Schachner, 19846). Similar activity may be absent from the anti-69Al antiserum since this was raised against the isolated 200K component.
Our current studies have shown that the antigen 69A1 is immunochemically indistinguishable from the LI antigen and that the epitope recognized by the monoclonal antibody 69A1 is expressed on neurones that have axons that fasciculate, but not on neurones with non-fasciculating processes. We are not aware of any previous reports of the restriction of the Ng-CAM/Ll/NILE molecule to specific classes of neurones, although there is some evidence that it may not be present on all neurones at all stages of development (Stallcup et al. 1985; Fushiki & Schachner, 1986). Our findings suggest that either the Ng-CAM/Ll/NILE molecule itself is differentially expressed on neurones or that the molecule is hetero-genous with respect to the expression of the 69A1 epitope. Such heterogeneity has previously been reported for an epitope designated L2 found on the N-CAM molecule and may also exist on the LI antigen (Kruse et al. 1984).
ACKNOWLEDGEMENTS
The authors are extremely grateful for the gifts of antisera used in this study. This work was supported by the Wellcome Trust.