axon-glial relationships in the anterior medullary velum of the adult rat

Journal of Neurocytology 24, 965-983 (1995)

Axon-glial relationships velum of the adult rat

in the anterior medullary

M . B E R R Y , 1. M . I B R A H I M , 2 J. C A R L I L E , 1 F. R U G E , 1 A . D U N C A N 1 and A . M . B U T T 2

1Division of Anatomy and Cell Biology and 2Sherrington School of Physiology, UMDS, Guy's and St. Thomas's Hospitals, London, UK

Received 12 June 1995; revised 23 August 1995; accepted 30 August 1995

Summary

The anterior medullary velum is a thin sheet of CNS tissue which roofs the rostral part of the Wth ventricle and contains fascicles of myelinated fibres which, in part, arise from the nucleus of the IVth cranial nerve. This study used histochemical, immunohistochemical, and intracellular dye-injection techniques to describe cellular interrrelationships in the velum in whole-mounts and in sections. Rip antibody-stained whole mounts provided a unique description of both oligodendrocyte units (defined as an oligodendrocyte and the complement of myelinated internodal segments it forms), and consecutive myelin sheaths along the same axon. A broad range of unit morphologies was categorised into four arbitrary groups, according to classical criteria, which comprised small cells supporting the short, thin myelin sheaths of 15-30 small diameter axons (Type I), through intermediate types (II &III), to the largest cells forming the long, thick myelin sheaths of 1-3 large diameter axons. Rip antibody and ferric ion-ferrocyani~de staining, together with intracellular dye injection, revealed oligodendrocyte process branching patterns and their mode of engagement of myelin sheaths, nodes of Ranvier, and the spatial disposition of the outer cytoplasmic rims of myelin sheaths. The latter formed a conspicuous spiral ridge on the exterior surface of myelin sheaths which connected with the paranodal loops at each heminode. Large bundles of axons decussated through the velum, the bulk of which were IVth nerve fibres which constituted the IVth nerve rootlet. The PNS/CNS transitional zone of the IVth nerve was located 0.25-0.50 mm along the root, where astrocytic end-feet defined an abrupt margin, convex towards the periphery, where the heminodes of central and peripheral myelin were apposed, and where the basal lamina tubes of the Schwann cell units were discontinued. The basal processes of ependymal cells lining the ventricular wall of the velum, passed between axon bundles before abutting on the basal lamina of the pia. Many of these processes branched and ran along the axonal bundles. A monolayer of microglia occupied a subependymal stratum in which the non-overlapping dendritic territories of each cell formed a regular mosaic throughout the velum without any obvious interaction with either axons or other glial cells. Astrocytes were also uniformly distributed; their fine processes made up a dense lattice amongst axons, often running parallel and within the fibre bundles; stouter ones had terminal end-feet which undercoated the basal lamina of both the glia limitans externa and the blood vessels in the velum.

Introduction

The anterior medullary velum of the rat develops from the roof plate of the neural tube which does not produce neurons and restricts axon entry (Silver et al., 1993). In the rat, it is a thin sheet of CNS tissue, 15- 150 ~tm in depth located in the brain stem roofing in the rostral part of the IVth ventricle, caudal to the inferior colliculi, and underlying the cerebellum. It extends from the caudal entrance of the mid-brain aqueduct, behind the subcollicular circumventricular organ, undercover of the rostral vermal cerebellar lobules of the lingula, lobus centralis and culmen, to the superior recess of the IVth ventricle. It attaches laterally to the superior cerebellar peduncles on the dorsal surface of the pons. In most species, the velum

* To whom correspondence should be addressed.

contains the fibres of the trochlear (IVth) nelvce which originate from each mid-brain nucleus, enter the velum laterally, decussate in the midline, and exit in the IVth nerve rootlets which are attached at the pial surface of the lateral margins of the velum (Warwick, 1964; Gacek, 1974; Kerns, 1980; Garcia et al., 1983; McConnell et al., 1984; Fraher et al., 1988; Derouiche et al., 1994). The calibre of the axons coursing through the velum is wide ranging, including large myelinated and fine unmyelinated fibres (Ibrahim et al., 1995). Axons are either grouped in bundles, or widely dispersed and this arrangement facilitates the study of relationships between glia and axons to a degree which is almost unparalleled anywhere eJse in the

0300-4864/95 �9 1995 Chapman and Hall

Axon-glial relations in the anterior medul la ry ve lum

CNS. The anterior medul la ry ve lum contains most forms of CNS glia but only occasional neuronal cell bodies.

The ve lum can be p repared as a whole moun t after detachment f rom the brain stem and cerebellum, to provide a unique almost two dimensional preparat ion for investigating interactions be tween axons and glia in the CNS (McConnell et al., 1984; Derouiche et al., 1994; Butt et aI., 1995; Ibrahim et al., 1955), as well as interactions be tween central and per ipheral glial at the exit point of the IVth nerve rootlet. This paper defines axon-glial relations in the ve lum of the normal adult rat and complements the quantitat ive (Ibrahim et al., 1995) and immunohis tochemical (Butt et al., 1995) descriptions of this structure.

Materials and methods

Animals

The vela of 110 adult Wistar female rats, weighing between 200-250g, were used for immunohistochemistry, silver, osmium tetroxide, Toluidine Blue and ferric ion-ferrocyanide staining. A further 20 vela were prepared for iontophoretic dye injection of glia.

Immunohistochemistry

Animals were perfusion-fixed with 4% paraformaldehyde in phosphate buffered saline (PBS) under hypnovel (Roche)/ hypnorm (Janssen) anaesthesia. Brain stems were removed and placed in 4% paraformaldehyde overnight. Ten specimens were then placed in 10% sucrose in PBS for 2 h, then in 20% sucrose overnight, frozen into tissue tech and 10 gm sections cut on a cryostat. A further 15 brain stems were dehydrated in a graded series of alcohols, infiltrated with polyester wax at 37~ (Steedman, 1957), and 7 gm sections were cut from three groups of five vela in the coronal, sagittal and horizontal planes. In five more animals, the brain stem was trimmed to facilitate the cutting of longitudinal sections through the IVth nerve rootlet exit point on each side. Sections were cut on a cooled chuck and ribbons floated onto slides subbed in a 1% gelatin solution and

967

allowed to dry before staining. After dewaxing sections through a descending series of alcohols, they were allowed to stand for 15 min in Dulbecco's phosphate buffered saline (DPBS) containing 0.1% Tween-20 before the primary antibody was added. For whole mounts, 40 vela were dissected from the brain stems, separated from the cerebella, and placed in 4% paraformaldehyde overnight and transferred to 1.5% triton in PBS for 24 h before incubation with primary antibodies.

To reduce background of the indirect immunolabelling, all antisera were diluted with DPBS containing 1% bovine serum albumin (BSA). Incubation times ranged from 1-24 h. Polyclonal rabbit (Dakopatts) or monoclonal mouse (Sigma) anti-GFAP antibodies were used at 1:250 and 1:100, respectively, as markers for astrocytes. Polyclonal rabbit anti-mouse laminin (Bethesda Research Laboratories) at 1:100 was used as a marker for the basal laminae of the glia limitans externa, blood vessels and Schwann cell-axon units. A range of markers were used to label oligodendrocytes; polyclonal rabbit anti-carbonic anhydrase II (CAII, prepared by Dr N Gregson, UMDS, UK) at 1:200; mouse monoclonal antibody Rip (Friedman et aI., 1989) which recognises an unknown oligodendrocyte specific epitope (supplied by Dr B. Friedman, Regeneron Pharmaceuticals, Tarrytown, USA), used as undiluted supernatant; polyclonal rabbit anti-proteolipid (PLP, Serotec, Oxford, UK) was used at I : 100 as a marker for CNS myelin. The monoclonal mouse anti-neurofilament antibody RT97, raised against 200 kDa component of the neurofilament triplet (prepared by ProfeSsor B. Anderton, Institute of Psychiatry, De Crespigny Park, UK), was used at 1:5 to label axons. Microglia were detected using Griffonia simplicifolia (B4, Sigma), which binds to the terminal ~-D-galactosyl residues of microglia plasma membranes, at a dilution of 1 : 10. Labelling was visualized using either immunoperoxidase staining, peroxidase anti- peroxidase (PAP) staining, or fluorescein/rhodamine con- jugated secondary antibodies.

Immunoperoxidase staining was carried out using the ABC Vector stain Elite Kit (Vector Laboratories, Buringame, CA). Vela were prepared as whole mounts by first washing in PBS and quenching endogenous peroxidase by incubation with 0.3% H202 in PBS for 30 rain. After rinsing in PBS, vela were incubated for 30 min in 1.5% goat serum diluted in PBS, in order to reduce non-specific binding. After an overnight

Fig. 1. Anatomy of the anterior medullary velum. (A, B) Whole mounts of the rat velum showing axons stained with silver (A) and their myelin sheaths stained with osmium (B); a high magnification of (B) is shown in Fig. 3A. The entry points of the IVth nerve fibres at the rostral lateral margins (*) lie either anterior or posterior to the IVth nerve roots (arrows), the axons of which come from the contralateral nucleus. The upper edge is rostral and incisions were made in the mid-line caudally, and occasionally anteriorly, to flatten the velum. (C) A mid-sagittal Toluidine Blue-stained section through the isthmus between the inferior colliculi (IC). The anterior medullary velum (arrowheads) passes from the posterior aspect of the subcollicular circumventricular organ (small arrow) in the midbrain aqueduct (A) to terminate posteriorly (*) in the superior recess of the IVth (IV) ventricle, undercover of the dorsal (DLC) and ventral (VLC) lobules of the lobus centralis and the l in~ta (L) of the cerebellum. The darkly stained myelin sheaths of the axon bundles are seen in the velum, the most prominent of which are the decussating IVth nerve fibres. (D, E) Semithin parasagittal Toluidine Blue-stained sections show details of the relation of the velum with the cerebellum. The pial surface of the anterior medullary velum is separated from that of the cerebellum by loose connective tissue and blood vessels (D). The ependymal surface of the anterior medullary velum (arrows in D, E) faces the IVth ventricle. Between these two membranes, are myelinated and non-myelinated axons, together with astrocy~es (arrow heads in E), oligodendrocytes (open arrow in E) and microglia (see Fig. 2B). (Abbreviations in C-E: DC-dorsal culmen, VC- ventral culmen, GL-granular layer, PCL-Purkinje cell layer, and ML-molecular layer of the cerebellum. Magnification: A- C x30, D x 110, E x 450).


incubation at 4~ with primary antibody, diluted with 1% BSA/PBS (in order to reduce the background of the indirect immunolabelling), specimens were treated with 1:100 dilution of biotinylated goat anti-rabbit IgG (Vector) for l h, followed by a 1 h incubation with a biotin-avidin peroxidase complex (Vector). Finally, the tissue was treated for 2-5 min with 0.5 mg m1-1 diamino-benzidine (DAB, Sigma) in PBS containing 0.01% H202. All the above stages were separated by PBS washes. The vela were finally washed, dried onto glass slides, and mounted in citifluor.

For peroxidase anti-peroxidase (PAP) staining, vela were washed and the endogenous peroxidase quenched by incubation with 0.3% H202 in PBS for 30min. They were then rinsed in PBS and incubated in a 1.5% serum from the animal in which the secondary antibody was raised in order to reduce non-specific binding. After an overnight incubation at 4 ~ C with the primary antibody, diluted with 1% B SA/ PBS, the vela were treated with a 1 : 100 dilution of either rabbit Ig's to mouse Ig's or swine Ig's to rabbit Ig's (Dakopatts) for I h. They were then incubated with PAP mouse/rabbit, a soluble complex of horseradish peroxidase (HRP) and rabbit/mouse anti-HRP (Dakopatts) and devel- oped in DAB before mounting in citifluor.

Light microscopy Ten whole mounts were stained for myelin sheaths with 2% osmium tetroxide in PBS, and ten were stained for axons using Kiernan's physical developer method (Kiernan, 1981). Twenty more animals were perfused with 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4), and the brain stems left in fixative for 2 h, subsequently washed and post-fixed in phosphate buffered 1% OsO4, dehydrated in an ascending series of ethanols and propylene oxide, and finally embedded in TAAB resin polymerized at 60~ for 24h. Semithin sagittal, coronal, longitudinal and horizontal sections of vela were stained with Toluidine Blue.

Ferric ion-ferrocyanide staining of nodes of Ranvier We used a modification of the method by Waxman and Quick (1978), with adaptations by Walkley and Pierok (1986). Five animals were perfused with 5% glutaraldehyde and 4% paraformaldehyde in 0.15 M sodium cacodylate (pH 7.2). Vela were immersion fixed in the same solution for a further 2.5h, and washed three times 0.15M sodium cacodylate buffer (pH 7.2) for 5 min at room temperature. Dissected vela were osmicated in 1% OsO4 in 0.15M cacodylate buffer (pH 7.2). After washing in distilled water, vela were incubated at 37~ in 0.02 M aqueous ferric chloride for 100 min, washed in distilled water, and then immersed in 2% acidified potassium ferrocyanide (pH 2)

969

with HC1 for 40 min. After washing in distilled water and overnight immersion in 0.15 M cacodylate buffer, vela were dehydrated, embedded in epon resin and sections cut at 1- 2 ~tm.

Intracellutar iontophoretic dye injection Dye injection enabled individual glial cell morphologies to be visualized in the whole tissue. Veta were mounted flat onto black millipore paper and placed in a brain slice chamber perfused with artificial CSF at 32 ~ C. Glial cells were impaled using a high resistance (80-120Mf~) microelectrode back-filled with lysinated rhodamine dextran (LRD, 4% in 0.5M KC1) and iontophoretically injected for 90s using 400ms pulses of 5-10 ~A depolarising current at 1 Hz, as described previously (Butt et al., 1994a). After injection, vela were immersion fixed in 4% paraformaldehyde for 30 min, washed in PBS and whole mounted in citifluor. All filled cells were examined using both a conventional epifluores- cence and a Bio-Rad MRC 600 laser scanning confocal microscope (see Figs 5A-C; 8A-F). The latter instrument has an argon ion laser as the excitation source and a YHS filter to visualise the LRD injected cells.

Results

The anterior medul la ry ve lum was easily detached from the cerebellum as an intact sheet of CNS tissue (Fig. 1A, B). The most prominent axons contained in the ve lum where those of the IVth nerve. Occasional neurons were also found, together with ependymal cells, microglia, astrocytes and oligodendrocytes.

Axons The distribution of axons in the ve lum was best demonst ra ted in whole mounts , where the course of IVth nerve fibres and other central axons was clear to see (Fig. 1A-C). Fibres of the trochlear (IVth) nerve, originating from each mid-brain nucleus, formed prominent decussating bundles which either directly crossed the velum, or arched caudally f rom their lateral points of entry and after crossing the midline, ran rostrally to exit in the IVth nerve rootlet. The latter was usual ly sited caudal to, sometimes co-localized with, and occasionally rostral to the point of entry of each IVth nerve central fibre bundle into the velum. Many other fibres also crossed in the velum, the bulk of which probably consti tuted the cerebellar

Fig. 2. Relations of axons with the ependymal layer. (A, B) RT97 immunostaining of unipolar neurons (large arrows) and their axons (medium arrows) in a transverse section (A, ependyma indicated by small arrows) and whole mount (B) of the velum; B was stained using the PAP technique. (C) Ferric-ion ferrocyanide staining of nodes in semi-thin osmium stained sections through an axon bundle in the velum. (D) Transverse section through an axon bundle showing immunofluorescence RT97 staining of axon profiles of different diameters. (E) Horizontal Toluidine Blue stained section through the basal aspects of ependymal cells (polygonal profiles) showing their close association with myelinated axon bundles (nodes of Ranvier are indicated by arrows). (F) Whole mount of velum showing the differential reactivity of ependymal cells with the anti-GFAP antibody. (Asterisk in A and D is the molecular layer of the cerebellum. Magnification in A-F x450).

Axon-glial relations in the anterior medullary velum

commissure, interconnecting the deep cerebellar den- tate, emboliform and globose nuclei, and some axons of the brachium conjunctivum which run in the lateral margins of the velum. Osmium staining revealed than many of these axons were myelinated (Fig. 1B, C). The velum was divisible into three regions (Ibrahim et al., 1995), namely a thin rostral zone containing small diameter decussating axons and also fibres running in a rostro-caudal plane, a middle area in which bundles of IVth nerve fibres cross directly to the contralateral root exit point, and a caudal region which also contained decussating IVth nerve fibres and other CNS axons (Fig. 1A, B), the former were the caudally looping fibres described above. Fibres in the rostral region were either unmyelinated or had thin sheaths of < 4 ~tm external diameter, whereas those in the middle IVth tract region were mostly myelinated and had large diameters of 4-5 ~tm. The caudal region contained unmyelinated and myelinated fibres of varying diameter from the submicron to 15~tm. Axons throughout the velum coursed between pia and ependyma, and were closely related to the basal processes of ependymal ceils, astrocyte processes, oligodendrocytes, and microglia (Figs 1D, E, 2D). Nodes of Ranvier of myelinated axons were clearly visible in semi-thin ferric ion-ferrocyanide (Fig. 2C), Toluidine Blue (Fig. 2E), and osmium-stained material (Fig. 3A).

Neuronal perikarya

A few neurons were present in the velum, usually in the middle and caudal regions and often adjacent to the attached lateral margins. They were all unipolar cells with large somata ranging from 15-30~tm in diameter (Fig. 2A, B). They probably represent displaced trigeminal sensory neurons from the nearby mesencephalic nuclei in the dorsolateral lips of the brainstem to which the velum is attached laterally.

The CNS/PNS junction of the IVth nerve

The junction between CNS and PNS was located along

971

the IVth nerve rootlet some 0.25-0.5mm from the point of exit from the velum, and was usually convex towards the periphery (Fig. 3C-F). At the junction, there was an abrupt demarcation between central and peripheral myelin, as indicated by PLP labell[ing (Fig. 3D), and between GFAP-labelled astrocytes in the CNS (Fig. 3E) and laminin-labelled tubes of Schwann cells in the PNS (Fig. 3F). Staining of axons with RT97 was not uniform across the junction, axons staining well in the CNS but poorly in the PNS; except at presumptive nodes and at random areas between nodes (Fig. 3C).

Ependyma

Ependymal cells were polygonal in outline m whole mounts and formed a regular mosaic over the ventricular surface (Fig. 2E, F). Some ependymal cells were completely GFAP-positive, whereas others had basal processes only which were positive, and some were entirely unstained(Figs 2F, 4B-D). Epen- dymal cell morphology varied in relation to the thickness of velum. In the thinnest areas, the ependyma was formed by a relatively simple cuboidal epithelium (Fig. 1E, 2D, E), in which the basal processes of the ependymal cells were a few microns long and abutted without branching onto the basal lamina of the glia limitans externa (Fig. 4B-E). In thicker areas, the complexity of the basal processes increased, branching and ramifying within, and usually parallel to, the axon bundles. Processes terminating as end feet at the gila limitans of the pia and on blood vessels were often seen within the thicker regions (Fig. 4B-D). Between the basal processes, myelinated and unmyelinated fibres ran in association with oligodendrocyte and astrocyte cell bodies and processes.

Microglia

Typical microglia were distributed uniformly throughout the velum as a monolayer of cells (Fig. 3B). Their process fields did not overlap but formed a regular mosaic. Microglial processes also ramified subepen- dymally amongst the axon bundles, but showed no

Fig. 3. Relations of axons with microglia and across the CNS/PNS junction of the IVth nerve rootlet. (A) Osmium stained myelin sheaths of varying diameter in a whole mount of the velum similar to that shown in Fig. lB. Nodes of Ranvier are clear (arrows) and thus the lengths of consecutive internodal myelin segments are measurable (Ibrahim et al., 1995). (B) Shows the distribution of microglia in a whole mount of the velum labelled with the biotinylated lectin Griffonia simplicifolia and visualized using the PAP technique and DIC microscopy. Note how the distribution of microglial processes (arrowheads) appears to be uninfluenced by the presence of axon bundles. (C-F) Immunofluorescent labelling of the CNS/PNS junction in the IVth nerve rootlet using RT97 for axons (C), PLP for CNS myelin (D), GFAP for astrocytes (E), and laminin (F) for the basal lamina of both the glia limitans externa (arrows) and Schwann cell tubes (star) in the PNS (CNS is marked by an asterisk). The characteristics of RT97 staining of axons changes across the junction (between large arrows in C). The incomplete staining in the PNS may be attributable to poor access of the primary antibody to the axons; the larger bright spots of fluorescence (small arrows) may thus correspond to sites of nodes and the finer banding where antibody reacts with Schmidt-Lantermann clefts. The CNS/PNS boundary is clearly delineated by FLIP (D)~ GFAP (E), and Iaminin (F). (D-F are sections from the same rootlet; E and F are double immunofluorescence labelling of the same section. Magnification in A-F x450).

Axon-glial relations in the anterior medullary velum

specialized relationship with either axons, their myelin sheaths, or with other glia.

Astrocytes

Astrocytes were uniformly distributed throughout the velum, their processes coursing within the subependymal region to form an intricate interwoven network (Fig. 4A). Within fibre bundles, astrocytic processes ran both parallel and orthogonal to the axons (Fig. 4B-D). Many endfeet of astrocyte processes contributed to the glia limitans externa insinuated between the cell bodies and basal processes of ependymal cells in the thinner regions of the velum. Blood vessels (Fig. 4B-E) in the velum were also surrounded by end feet (Fig. 4D, arrow), although many lacked a complete laminin positive basal lamina. Some of the GFAP positive processes interdigitated with apical ependymal processes to contribute to the lining of the IVth ventricle, but most of these belonged to displaced ependymal cells whose collateral branches ran with those of definitive astrocytes within the axon bundles.

To determine the whole-cell morphology of single cells, astrocytes iontophoretically filled with LRD were analysed (Fig. 5). All had a stellate morphology, but within axon bundles astrocytes exhibited a polarized morphology and extended processes predominantly parallel with axons (Fig. 5B). Other cells exhibited a characteristic 'starburst' morphology, with dicho- tomously branching processes extending randomly in all directions from a centrally located cell body (Fig. 5C). Many of the radially orientated astrocyte processes had terminal expansions at their sub-pial end-feet, as illustrated clearly in Fig. 5A (small arrows). Most filled astrocytes had both stem-like and velate processes, the latter were membrane expansions usually of the soma and proximal processes, but often from the distal processes also (Fig. 5A, large arrows). Fine processes and blebs also extended from radial and velate processes (Fig. 5A); these latter may correspond to the proximal parts of perinodal processes described both in electron microscopical (Butt et al., 1994c) and immunohistochemical studies (Butt et al. 1995), although the fine

973

terminal regions of nodal collaterals are probably unresolvable at the light microscope level (Butt et al., 1994a, b).

Oligodendrocytes

Oligodendrocytes were best demonstrated with CAII and Rip antibodies. Rip stained entire oligodendrocyte units (Figs 6B, 7A-D), defined as (see Bunge, 1968) the soma of an oligodendrocyte, its processes and all the internodal myelin segments formed by the cell. Labelling by CAII was confined to the cell bodies and proximal processes (Fig. 6E). Oligodendrocytes were less frequent than astrocytes and most plentiful in the heavily myelinated fibre bundles, where they appeared scattered throughout the subependy- real layer (Fig. 6D), intermingled with astrocytes (Fig. 6C).

In Rip-stained whole mounts, the relations of myelinated axons to oligodendrocytes was most obvious in thin regions of the velum where axons were widely dispersed. Here, oligodendrocyte processes were seen to engage axons anywhere along the internodal segments (Fig. 7B). Frequently, a darker Rip-stained ribbon spiralled over the sheath away from the point of contact (Fig. 7C); this was assumed to be the Rip-positive outer cytoplasmic rim of the myelin sheath adjacent to the external mesaxon. Nodes were clearly visible in the Rip-stained material (Fig. 6A; see ferric iron-ferrocyanide and osmium stained vela in Figs 2C, D and 3A). Paranodal regions (Fig. 6A), points of engagement of myelin sheaths by oligodendrocyte processes (Fig. 7B), the outer cytoplasmic rim of the sheaths (Fig. 7B), and occasional transverse bands along the myelin segments (Figs 6A, 7B) were more heavily Rip stained than the remaining parts of the myelin sheaths. A less heavily stained reticular meshwork of cytoplasmic intersections was visible in most sheaths, as described by Butt and colleagues (1995). Occasionally, the branches of a single oligodendrocyte process appeared to be associated with consecutive sheaths across a node of the same axon (Fig. 7B), although it was not clear whether these engaged the sheaths, or were en passant contacts.

Oligodendrocytes have been classically subdivided

Fig. 4. Axon-astrocyte relations in the anterior medullary velum. (A) Whole mount of the velum immunostained with anti- GFAP using the PAP technique. Astrocytes (some cell bodies indicated by arrows) had both radially orientated processes and processes running parallel with the axon bundles (arrow heads). In transverse section (B, D, E), immunofluorescent staining with anti-GFAP shows astrocyte processes and one astrocyte cell body (arrow in D) and the basal processes of ependymal cells running along the axon bundles. In parasagittal section (C), GFAP-positive ependymal cells are evident, the basal processes of which delineate fibre bundles in which many axon associated GFAP-positive processes are cut transversely and thus appear as groups of punctate profiles (small arrows). Double immunofluorescence labelling with anti-GFAP (D) and anti-laminin (E) shows the apical processes of astrocytes and ependymal cells contributing to the glia limitans externa of the velum. At the cerebellar surface, GFAP-positive end-feet of Bergmann glia processes (D) and the basal lamina (E) of the glia limitans externa are also visible. (Magnification in A-E x400).

974 B E R R Y , I B R A H I M , C A R L I L E , R U G E , D U N C A N a n d B U T T

Fig. 5. Fluorescence confocal photomicrographs of LRD-filled astrocytes. In A and C, astrocytes extend processes randomly in all directions from a centrally located cell body. The astrocyte in B extends processes preferentially along fibre bundles parallel with the axons. All cells display both velate (large arrows in A) and stem-like processes, some of which have terminal expansions associated with the glia limitans externa of the velum (small arrows in A). (Magnification: A x250; B and C x320).

Axon-glial relations in the anterior medullary velum 975

into types I-IV (del Rio-Hortega, 1928; Bunge, 1968; Stensaas & Stensaas, 1968), according to soma size, number of processes and the diameter of internodal myelin segments supported by the unit. These morphological subtypes could be distinguished in the anterior medullary velum (Fig. 7A-D). The smallest cells had four or more fine primary processes which branched randomly to myelinate 15-30 small diameter axons (Fig. 7A, B), and resembled types I and II of classical nomenclature; heminodes were commonly associated with these units. Cells corre- sponding to types III and IV had large cell bodies, with thick primary processes which rarely branched and myelinated a small number, usually less than five fibres, with external sheath diameters ranging from 4 to 15 ~tm (Fig. 7C, D). Rarely, type IV units were observed to form a single long myelin sheath usually over fibres of large diameter (see Butt et al., 1995). The diameters of most axons of a given oligodendrocyte unit were approximately the same (Fig. 7B, D), for example, the range of myelin sheath diameters spanned from the submicron to 4 ~tm in type I or II units, and 4 and 15 ~tm in type III and IV units (Fig. 7A, C).

It was difficult to distinguish many of the details of oligodendrocyte morphology in the dense fibre bundles in the middle and caudal part of Rip-stained vela. In these cases, intracellular iontophoretic injection of LRD was used to resolve whole-cell morphology of single oligodendrocyte units (Fig. 8). A wide range of morphologies was found using this technique, similar to those revealed by Rip staining, and a good impression of the variation in both myelin sheath length and thickness is given by these units. Figure 8A-C illustrates three oligodendrocyte units with 4-5 myelin sheaths of varying diameter and internodal length. In the units in Fig. 8A and B, myelin sheath external diameter varied from 10-15 ~tm and 1- 5 ~tm, respectively, with lengths up to 390 ~tm long in each unit. The sheaths in these units displayed prominent external cytoplasmic tongues which spiralled over the surface of the myelin sheaths up to the paranodal loops; a possible reconstruction of the spiral outer cytoplasmic ridge along the myelin sheath of one half internodal segment is illustrated in Fig. 8A. In Fig. 8C, the oligodendrocyte myelinated four small diameter axons with fibre diameters of 2-3 ~m and internodal myelin lengths ranging from 220- 390 ~tm. The units' illustrated in Figure 8E and F myelinated parallel arrays of 5-8 axons with sheaths of a similar, small diameter, < 2~tm, and uniform internodal lengths.

Discussion

Axon/oligodendrocyte unit interrelationships are more accessible in the velum than elsewhere in the

CNS, without serial sectioning (Fraher, 1978), or the teasing of fibre bundles (McDonald & Ohlrich, 1971; Murray & Blakemore, 1980). Qualitative and quantitative analyses are possible of both individual units and consecutive sheaths along the same axon (Ibrahim et al., 1995). It is now established that the processes of oligodendrocytes myelinate the internodal segments of several axons (Bunge et al., 1962; Bunge, 1968; Peters, 1964; Hirano & Dembitzer, 1967; Hirano, 1968; Butt & Ransom, 1989, 1993; Remahl & Hildebrand, 1990a,b; Bjartmar et al., 1994; Butt et al., 1994a,b). The classical studies of Del Rio-Hortega (1928) and Penfield (1932), supported by more recent findings reported heterogeneity of oligodendrocytes (Stensaas & Stensaas, 1968; Remahl & Hildebrancl, 1990a; Hildebrand et al., 1993; Bjartmar et al., 1994; Butt et al., 1995), suggests that there are different types of oligodendrocytes arbitrarily classed as type I-IV according to both the diameter of fibres rnyelinated and the number of myelin sheaths in each unit. The findings of the present study confirm this view and show that units range from type I, which comprise cells with small somata having several branched processes myelinating many internodal segments, through to type IV units which have cells with large somata and few processes (sometimes only one) which myelinate large diameter axons. Our results add weight to the suggestion of Waxman and Sims (1984) that sheath diameter can vary widely within a given unit. Internodal myelin sheath length, defined by dye injection of oligodendrocytes in the velum in this study, and by osmium staining in that of Ibrahim and colleagues (1995), appears to have a similar variance to that of sheath diameter. Butt and colleagues (1995) have demonstrated that most type III and IV units may be phenotypically different from I and II units in not expressing CAII. All units appeared to be Rip, transferrin, MOG, PLP, MBP and CNPase positive (Butt et al., unpublished observations). Ibrahim and colleagues (1995) also indicated that sheath diameter/ length relationships of type I/II units may be different from those of type III/IV units.

The outer oligodendrocyte cytoplasmic tongue of the myelin sheath was clearly visible in dye-filled and Rip-stained cells as a spiral ridge ori~nating at the point of contact of a process and extending to the paranodal region on the outer surface of the myelin sheath. Scanning electromicrographs (Berry & Butt 1995, and our unpublished data with J. Sievers and U. Mangold) show the presence of half spiral ridges on the exposed hemi-surfaces of axons in the velum. Occasional reverse spirals of tongues of oligodendrocyte cytoplasm are seen in dye-filled preparations superimposed on that of the outer cytoplasmic rim, which might represent either Schmidt-Lantermann clefts, or the inne r cytoplasmic tongue of the myelin sheath. The Rip-stained myelin sheaths had a

976 B E R R Y , I B R A H I

network of cytoplasmic interconnections (Butt et al., 1995) which are probably equivalent to the reticula- tions of Stensaas and Stensaas (1968) and are likely to correspond to Schmidt-Lantermann clefts (Blakemore, 1969). Most published views of CNS myelin (e.g. Hirano & Dembitzer, 1967; Bunge, 1968; Butt & Ransom, 1989) assume that the outer strip of glial cytoplasm runs linearly along the axon, parallel with the long axis and continuous with the outer- most paranodal cytoplasmic loops at each heminode. During maturation of internodal myelin segments, Fraher (1978) makes the point that the outer and inner cytoplasmic rims of the unrolled oligodendrocyte myelin sheaths are both convex to the long axis of the axon, and thus a considerable number of turns of the sheath end at abnodal levels. The presence of spiral outer cytoplasmic tongues in our Rip-stained and dye-injected oligodendrocyte units also indicates that during myelination the inner lamellae grow faster longitudinally than the outer ones, and that some of the outer myelin lamellae never extend to the node in the mature sheath. It is, however, possible that the presence of the spiral of the outer cytoplasmic rim is a feature of developing sheaths only and thus in the adult their appearance may be

M, C A R L I L E , R U G E , D U N C A N a n d B U T T

confined to internodal segments undergoing active remyelination, although the evidence for myelin sheath turnover in adult CNS is conjectural (Hildebrand et al., 1993).

There were rarely more than four primary processes in an oligodendrocyte unit. Thus, the number of internodal segments supported by each unit is a function of serial asynchronous dichotomous branching of the primary processes. The observation that a single Rip-labelled process contacted adjacent myelin segments suggests contiguous myelination by oligodendrocyte units, which may be a feature of non- parallel arrays of axons, like those traversing the velum. However, continuity of myelination by the same unit across nodes was not observed in dye- injected ologodendrocytes in either the anterior medullary velum or the optic nerve (Butt & Ransom, 1989, 1993; Ransom et al., 1991; Butt et al, 1994a, b), and it is thus possible that apparent serial contacts of adjacent myelin segments were en passent, rather than engagements of adjacent sheaths. Oligodendrocyte processes also appeared to contact internodal myelin segments anywhere along the sheath, including the paranodal regions (see also Butt et al., 1995, Fig. 7E), suggesting that nodal siting may not be a function of

Fig. 6. Axon-oligodendrocyte relations. (A) Myelin sheaths of varying diameter immunolabelled with Rip, a monoclonal antibody which recognises an unknown epitope on oligodendrocytes (Friedman et al., 1989). Note the reticulated staining pattern of the myelin sheaths, internodal banding (large arrows), and differential staining of nodes (small arrows). (B) Rip- labelled oligodendrocyte units in the rostral region of the velum. Rip provided complete labelling of the oligodendrocyte unit. Primary processes branch to be distributed to the internodal myelin segments of the widely dispersed axons. (C, D) Double immunofluorescence transverse section of the velum stained with anti-GFAP and anti-CAII, respectively. Oligodendrocyte cell bodies are situated within the subependymal region (ependyma indicated by arrows) amongst the astrocyte processes. (E) CAII-labelled oligodendrocytes in a whole mount of the velum showing the distribution of their primary and secondary processes into the axon bundles. (A, B and E were visualized using the PAP technique. Magnification: A, C, D x 450, B and E x 230). Fig. 7. Oligodendrocyte units in whole mounts of anterior medullary velum immunolabelled with Rip. (A, B) Multipolar oligodendrocytes support numerous myelin sheaths for small calibre fibres. (C, D) Large oligodendrocytes with fewer, stouter processes (arrows in D) engage large diameter myelin sheaths. (A) and (B) correspond to Types I and II of classical nomenclature, whilst (C) and (D) correspond to Types III and IV (Del Rio-Hortega, 1928; Bunge, 1968; Stensaas & Stensaas, 1968). Arrowheads in (B) show increased RIP staining at points of engagement of oligodendrocyte processes with internodal myelin segments. This cell appears to extend processes to consecutive internodal myelin segments (arrowheads), either side of a node (arrow). Spiralling of the outer cytoplasmic glial tongue is clear in larger myelin sheaths (arrows in C). Note that the number of axons engaged by each unit is inversely proportional to sheath diameter, but that diameters vary within individual units (C). (All panels are DIC microscopy of PAP preparations. Magnification in A-D, x 1200). Fig. 8. Fluorescence photomicrographs of LRD-filled oligodendrocytes. (A, B) Oligodendrocytes supporting five myelin sheaths. In (A) the myelin sheaths vary in diameter from 10-15 ~tm and have internodal lengths up to 380 ~m. In (B) three myelin sheaths run parallel and have internodal lengths of between 220-290 ~tm and sheath diameters of 3-4 ~tm, a fourth runs tangentially from bottom left to upper right and has an internodal length of 390 ~m and a sheath diameter of 3 ~tm, and a fifth runs tangentially from upper left to lower right and has a diameter of ~ 1 ~tm and internodal length of 180 ~tm. Note the dye filled paranodal loops (small arrows) and spirals of the outer cytoplasmic glial rim (large arrows); the external appearance of the upper left sheath is depicted in the inset drawn to scale. (C-E) Oligodendrocytes myelinating small diameter axons; the oligodendrocyte in (C) has four myelin sheaths of diameter 3-4 ~m and internodal lengths of 220-390 ~tm; (D) details the cell body and connecting processes. In (E) and (F) the oligodendrocytes myelinate parallel arrays of axons with a marked uniformity in diameter and internodal length. (In C an astrocyte filled by a separate impalement is visible. Magnification: A x 400; B x 500; D x 800; C and F x 200).

980 B E R R Y , I B R A H I

equidistant extension of myelin lamellae from the oligodendrocyte process. However, the precise point at which processes engaged myelin sheaths could not always be identified and, as already mentioned, some processes could make en passant contacts. We are currently studying the development of axon-glial relations in the velum to address these questions.

The environment of the fibres of the central trajectory of the IVth nerve has special significance since these fibres, like those of the ventral spinal roots (Ramon y Cajal, 1928; Risling et al., 1983; Linda et aI., 1992) regenerate after injury. The regenerative response of the IVth nerve fibres in the periphery (Murphy et al., 1990; Iannuzzelli et al., 1994, 1995) and in the velum (McConnell et al., 1984; Book et al., 1995; Derouiche et al., 1994) have been reported in the cat and rat. In the rat, IVth nerve fibres, and no others, regenerate after injury in the velum. Regenerating axon sprouts loop away from the margins of a mid-line lesion and enter the ipsilateral IVth nerve rootlet (McConnell et al., 1984). Their path to the rootlet is thus an aberrant one and this work shows that it is likely to be a putative growth inhibitory environment in which reactive astrocytes, demyelinating oligodendrocytes and degenerating CNS myelin (Schwab et at., 1993). are plentiful. It has been assumed that regenerating IVth nerve fibres in the velum respond to specific Schwann cell trophic signals emanating from the CNS/PNS junctional zone of the root. Our definition of the morphology and siting of the junction along the rootlet agrees with that of Fraher and colleagues (1988), and also with descriptions of the junctional complex in other dorsally placed roots (Berthold & Carlstedt, 1977; Fraher & Sheehan, 1987). Differential RT97 staining across the junction may identify different relationships between the myelin sheaths of oligodendrocytes and Schwann cells. We are currently defining the changes in the stability of the zone following degeneration and subsequent regeneration of IVth nerve axons to determine if Schwann cells migrate into the velum after injury to organise the regenerative response.

Since the velum is derived from the roof plate of the neural tube (Silver et al., 1993) neurogenesis is unlikely to occur in this membrane, and this probably accounts for the few neurons observed in the adult velum. They were found in the attached margins and were always pseudo-unipolar, probably representing trigeminal primary sensory neurons displaced from the nearby mesencephalic nuclei in the dorsolateral lips of the brainstem. Fibre ingrowth into the roof plate is also a rare event in the embryo and is probably restrained by the presence of growth inhibitory keratan sulphate proteoglycans, secreted by astrocytes in the dorsal midline of the neural tube (Snow et al., 1990). In the area of the presumptive ariterior medullary velum, proteoglycans may be absent when the IVth nerve

M, C A R L I L E , R U G E , D U N C A N a n d B U T T

decussation is established (Silver, 1993; Silver et al., 1993).

The glial cells of the velum may also derive from exogenous loci. Microglia are probably transformed from monocytes which enter the CNS parenchyma through blood vessels prenatally (Ling & Wong, 1993). The source of astrocytes in the velum is probably from endogenous stem cells which, like those of the subependyma of the cerebrum are probably not bipotential (Luskin et al., 1993; Luskin & McDermott, 1994), but instead exhibit lineage restrictions; one line producing astrocytes only, whilst the exclusive pro- geny of another may be oligodendrocytes. There is, however, some evidence that oligodendrocytes may accumulate in the velum by migration from the brain stem (Curtis et al., 1988; Reynolds & Wilkin, 1988) although, more recently, the mouse monoclonal GD3 marker used in those studies has been shown to label both microglia and pro-oligodendrocytes (Wolswijk, 1994). Many ependymal cells were GFAP-positive. They possessed basal processes forming both end- feet at the glia limitans externa and at blood vessels, and also had branches which ramified within axon bundles. Ependymal cells thus resemble astrocytes in many respects, suggesting origins from a common lineage.

The processes of astrocytes subserve barrier, trans- port and putative axonal support functions (Berry & Butt, 1995), and it is likely that only one phenotype exists in the velum, as in the optic nerve (Butt et al., 1994a, b). Specific nodal contacts of astocytes processes were not identified in this study, although Butt and colleagues (1995) reported structures in Rip/GFAP double stained preparations of the velum which were similar to perinodal processes described by Butt and colleagues (1994c) in electromicrographs of HRP-filled astrocytes of the mouse optic nerve. Oligodendrocyte/astrocyte spatial interrelationships were not as regular as those described in other CNS tracts (Suzuki & Raisman, 1992, 1994). Rows of oligodendrocytes interspersed with solitary astrocytes were not present, but instead both types of glia were scattered randomly throughout the subependymal layer amongst axons and the processes of microglia and ependymal cells. Astrocyte processes encircled oligodendrocyte somata and were closely related to their radial processes, and also the myelin sheaths (Butt et al., 1995).

The processes of microglia in the monolaye~ of the velum formed no specialized relationships with other glia or axon bundles, although they may stimulate oligodendrocytes to myelinate axons (Hamilton & Rome, 1994), and also participate in myelin turnover (Persson & Berthold, 1991) account- ing for the intracellular inclusions, thought to be myeloid bodies, seen in microglia elsewhere in the CNS, which have surrounding phosphatase


activity (Hi ldebrand & Skoglund, 1971; Hi ldebrand, 1982).

Acknowledgements

We thank Kevin Fitzpatrick and Sarah Smith for the

981

pho tography and Ken Appleby and his staff for biological services. The work was suppor ted by a grant f rom the Special Trustees of St. Thomas 's Hospital.

References

BERRY, M. & BUTT, A. M. (1995) Structure and function of glia in the central nervous system. In Greenfield's Neuro- pathology (editors GRAHAM, D. I. & LANTOS, L.) in press. Sevenoaks, Kent: Edward Arnold.

BERTHOLD, C.-H. & CARLSTEDT, T. (1977) Observations on the morphology at the transition between the peripheral and central nervous system in the cat. II. General organisation of the transitional region in the S1 dorsal rootlets. Acta Physiologica Scandinavia. 446 (Suppl.), 23-42.

BJARTMAR, C., HILDEBRAND, C. & LOINDER, K. (1994) Morphological heterogeneity of rat oligodendrocytes: electron microscope studies on serial sections. Glia 11, 235-44.

BLAKEMORE, W. F. (1969) Schmidt-Lantermann incisures in the central nervous system. Journal of Ultrastructure Research 29, 496-8.

BOOK, A. A., FISCHER, I., YU, X-J., IANNUZZELLI, P. G. &

MURPHY, E. H. (1995) MAP-2 and a phosphorylated isoform of MAP-IB are distinctly altered in trochlear motoneurons following PNS and CNS lesion. Experimental Neurology, in press.

BUNGE, R. P. (1968) Glial cells and the central myelin sheath. Physiological Reviews 48, 197-251.

BUNGE, M. B., BUNGE, R. P. & PAPPAS, G. D. (1962)

Electron microscopic demonstration of connections between glia and myelin sheaths in the developing mammalian nervous system. Journal of Cell Biology 12, 448-53.

BUTT, A. M., & RANSOM, B. R. (1989) Visualisation of oligodendrocytes and astrocytes in the intact rat optic nerve by intracellular injection of Lucifer Yellow and horseradish peroxidase. Glia 2, 470-5.

BUTT, A. M. & RANSOM, B. R. (1993) M o r p h o l o g y of

astrocytes and oligodendrocytes during development in the intact rat optic nerve. Journal of Comparative Neurology 338, 141-58.

BUTT, A. M., COLQUHOUN, K. & BERRY, M. (1994a)

Confocal imaging of glial cells in the intact rat optic nerve. Glia 10, 315-22.

BUTT, A. M., COLQUHOUN, K., TUTTON, M. & BERRY, M.

(1994b) Three-dimensional morphology of astrocytes and oligodendrocytes in the intact mouse optic nerve. Journal of Neurocytology 23, 469-85.

BUTT, A. M., DUNCAN, A. & BERRY, M. (1994c) Ast rocyte associations with nodes of Ranvier: ultrastructural analysis of HRP-filled astrocytes in the mouse optic nerve. Journal of Neurocytology 23, 486-99.

BUTT, A. M., IBRAHIM, M., RUGE, F. & BERRY, M. (1995)

Biochemical subtypes of oligodendrocyte in the anterior medullary velum of the adult rat as revealed by the monoclonal antibody Rip. Glia 14, 185-97.

CURTIS, R., COHEN, J., FOK-SEANG, J., HANLEY, M. R., GREGSON, N. A., REYNOLDS, R. & WILKIN, G. P. (1988)

Development of macroglial cells in rat cerebellum. I. Use of antibodies to follow early in vivo development and migration of oligodendrocytes. Journal of Neurocytology 17, 43-54.

DEL RIO-HORTEGA, P. (1928) Tercera aportaci6n al con- ocimiento morfologica e interpretaci6n funcional de la oligodendroglia. Memorias de la Real Sociedad iSspanola de Historia Natural 14, 5-122.

DEROUICHE, A., BERRY, M. & SIEVERS, J. (1994) Regenera- t ion of axons into the trochlear rootlet after anterior medullary lesions in the rat is specific for ipsilateral IVth nerve motorneurones. Journal of Comparative Neurology 341, 340-50.

FRAHER, J. P. (1978) Quantitative studies on the maturation of central and peripheral parts of individual ventral monorneurons axons. II. Internodal length. Journal of Anatomy 127, 1-15.

FRAHER, J. P. & SHEEHAN, M. M. (1987) The CNS-PNS

transitional zone of the rat cervical dorsal roots during development and at maturity. A morphological and morphometric study. Journal of Anatomy 152, 189-203.

FRAHER, J. P., SMIDDY, P. F. & O'SULLIVAN, V. R. (1988) The central-peripheral transitional regions of cranial nerves. Trochlear and abducent nerves. Journal of Anatomy 161, 115-23.

FRIEDMAN, B., HOCKFIELDS, S., BLACK, J. A., WOOD-

RUFF, K. A. & WAXMAN, S. G. (1989) In situ d e m o n -

s t ra t ion of mature oligodendrocytes and their processes: an immunocytochemical study with a new monoclonal antibody, Rip. Glia 2, 380-90.

GACEK, R. R. (1974) Localisation of neurons supplying the extraocular muscles in the kitten using horseradish peroxidase. Experimental Neurology 44, 381-403.

GARCIA, J. L. L., SEGADE, L. A. G. & NUNEZ, J. M. S. (1983) Localisation of motorneurons supplying the extracellular muscles of the rat using horseradish peroxidase and fluorescent double labelling. Journal of Anatomy 137, 247--61.

HAMILTON, S. P. & ROME, L. H. (1994) Stimulation of in vitro myelin synthesis by microglia. Glia 11, 326-35.

HILDEBRAND, C. (1982) Electron microscopic identification of Gomori-positive rings in normal spinal cord white matter. Acta Neuropathologica 56, 29-34.

HILDEBRAND, C. & SKOGLUND, S. (1971) Histochemical studies of adult and developing feline cord white matter. Acta Physiologica Scandinavia. 364(Suppl). 145-73

HILDEBRAND, C., REMAHL, S., PERSSON, H. & BJART- MAR, C. (1993) Myelinated nerve fibres in the CNS. Progress in Neurobiology 40, 319-84.

HIRANO, A. (1968) A confirmation of oligodendroglial

982 B E R R Y , I B R A H I M , C A R L I L E , R U G E , D U N C A N a n d B U T T

origin of myelin in the adult rat. Journal of Ceil Biology 38, 637-40.

HIRANO, A. & DEMBITZER, H. M. (1967) A structural analysis of the myelin sheath in the central nervous system. Journal of Cell Biology 34, 555-67.

IANNUZZELLI, P. G., WANG, X. H., WANG, Y. & MURPHY, E. H. (1994) Axotomy induced changes in cytochrome oxidase activity in the cut trochlear nucleus. Brain Research 637, 367-72.

IANNUZZELLI, P. G., MURRAY, M. & MURPHY, E. H. (1995) Regenerative axonal sprouting in the cat trochlear nerve. Journal of Comparative Neurology 354, 229-40.

IBRAHIM, M., BUTT, A. M. & BERRY, M. (1995) The relationship between myelin sheath diameter and internodal length in axons of the anterior medullary velum of the adult rat. Journal of the Neurological Sciences, in press.

KERNS, J. M. (1980) Postnatal differentiation of the rat trochlear nerve. Journal of Comparative Neurology 189, 291- 306.

KIERNAN, J. A. (1981) Histological & Histochemical Methods: Theory & Practice, pp. 272--3. Oxford: Pergamon Press.

LINDA, H., CULLHEIM, S. & RISLING, M. (1992) A light and electron microscopic study of intracellulary HRP-labelled lumbar montoneurons after intramedullary axotomy in the adult cat. Journal of Comparative Neurology 318, 188-208.

LING, E. A. & WONG, W-C (1993) The origin and nature of ramified and amoeboid microglia: a histrorical review and current concepts. Glia 7, 91-8.

LUSKIN, M. B. & MCDERMOTT, K. (1994)Divergent lineages for oligodendrocytes and astrocytes originating in the neonatal forebrain subventricular zone. Glia 11, 211-26.

LUSKIN, M. B., PARNAVELAS, J. & BARFIELD J. (1993) Neurons, astrocytes and oligodendrocytes of the rat cerebral cortex originate from separate progen{for cells: an ultrastructural analysis of clonally related cells. Journal of Neuroscience 13, 1930-50.

MCCONNELL, P., BERRY, M., REES, E. L. & SIEVERS, J. (1984) The injury response of nerve fibres in the anterior medullary velum of the adult rat. Brain Research 323, 257- 68.

MCDONALD, W. I. & OHLRICH, G. D. (1971) Quantitative anatomical measurements on single isolated fibres from the cat cord. Journal of Anatomy 110, 191-202.

MURPHY, E. H., BROWN, J. A. IANNUZZELLI, P. G. & BAKER, R. (1990) Regeneration and soma size changes following axotomy of the trochlear nerve. Journal of Comparative Neurology 292, 524-36.

MURRAY, J. A. & BLAKEMORE, W. F. (1980) The relationship between internodal length and fibre diameter in the spinal cord of the cat. Journal of the Neurological Sciences 48, 29-41.

PENFIELD, W. (1932) Neuroglia: normal and pathological. In Cytology and Cellular Pathology of the Nervous System, VoI. 2. New York: Hoeber.

PERSSON,~H. & BERTHOLD, C.-H. (1991) Cellular relations of paranodal Marchi-positive bodies in the normal rabbit studied with monoclonal antibodies against partially degraded CNS myelin fragments. Journal of Neurocytology 20, 277-89.

PETERS, A. (1964) Observations on the connexions between myelin sheaths and glial cells in the optic nerves of young rats. Journal of Anatomy 98, 125-34.

RAMON Y CAJAL, S. (1928) Degeneration and Regeneration of the Nervous System. London: Oxford University Press.

RANSOM, B. R., BUTT, A. M. & BLACK, J. (1991) Ultrastructural identification of HRP-injected oligodendrocytes in the intact rat optic nerve. Glia 4, 37-45.

REMAHL, S. & HILDEBRAND, L. (1990a) Relations between axons and oligodendroglial cells during initial myelination. I. The glial unit. Journal of Neurocytology 17, 313-28.

REMAHL, S. & HILDEBRAND, C. (1990b)Relations between axons and oligodendroglial cells during initial myelination. II. The individual axon. Journal of Neurocytology 19, 883-98.

REYNOLDS, R. & WILKIN, G. P. (1988) Development of macroglial cells in rat cerebellum. II. An in situ immunohistochemical study of oligodendroglial lineage from precursor to mature myelinating cell. Development 102, 409-25.

RISLING, M., CULLHEIM, S. & HILDEBRAND, C. (1983) Reinnervation of the ventral root L7 from ventral horn neurons following intramedullary axotomy in adult cats. Brain Research 280, 15-23.

SCHWAB, M. E., KAPFHAMMER, J. P. & BANDTLOW, C. E. (1993) Inhibitors of neurite growth. Annual Review of Neuroscience 16, 565-95.

SILVER, J. (1993) Glia-neuron interactions at the midline of the developing mammalian brain and spinal cord. Perspectives in Developmental Neurobiology 4, 427-36.

SILVER, J., EDWARDS, M. A. & LEVITT, P. (1993) Immuno- cytochemical demonstration of early appearing astroglial structures that form boundaries and pathways along axon tracts in the fetal brain. Journal of Comparative Neurology 15, 415-36.

SNOW, D. M., STEINDLER, D. A. & SILVER, J. (1990) Molecular and cellular characterization of the glial roof plate of the spinal cord and optic tectum: a profile for a proteoglycan in the development of an axon barrier. Developmental Biology 138, 359-76.

STEEDMAN, H. F. (1957) Polyester wax. A new ribboning embedding medium for histoloy. Nature 179, 1345.

STENSAAS, L. J. & STENSAAS, S. S. (1968) Astrocytic neuroglial cells, oligodendrocytes and microgliacytes in the spinal cord of the toad. I. Light microscopy. Zeitschrift fur ZeUforschung und mikroskopische Anatomie 84, 473-89.

SUZUKI, M. & RAISMAN, G. (1992) The glial framework of central white matter tracts: segmented rows of continuous interfascicular oligodendrocytes and solitary astrocytes give rise to a continuous meshwork of transverse and longtitudinal processes in the adult rat fimbria. Glial 6, 222-35.

SUZUKI, M. & RAISMAN, G. (1994) Multifocal pattern of postnatal development of the macroglial framework of the rat fimbria. Glia 12, 294-308.

WALKLEY, S. U. & PIEROK, A. L. (1986) Ferric ion-ferrocyanide staining in ganglioside storage disease establishes that meganeurites are of axon hillock origin and distinct from axonal spheroids Brain Research 382, 379-86.

WARWICK, R. (1964) Oculomotor organisation. In The Oculomotor System (edited by BENDER, M.B.) pp. 173- 204. New York: Harper and Rowe.

Axon-glial relations in the anterior medul la ry ve lum 983

WAXMAN, S. G. & QUICK, D. C. (1978) Intra-axonal ferric ion ferrocyanide staining of nodes of Ranvier and initial segments in central myelinated fibers. Brain Research 144, 1-10.

WAXMAN, S. G. & SIMS, T. J. (1984) Specificity in central

myelination: evidence for local regulation of myelin thickness. Brain Research 292, 179-85.

WOLSWIJK, G. (1994) GD3 cells in the adult rat optic nerve are ramified microglia rather than 0-2A adult progenitor cells. Glia 10, 244-9.

axon-glial relationships in the anterior medullary velum of the adult rat

Documents