expression of myelin proteins in the opossum optic nerve: late appearance of inhibitors implicates...

THE JOURNAL OF COMPARATIVE NEUROLOGY 37227-36 (1996)

Expression of Myelin Proteins in the Opossum Optic Nerve: Late Appearance

of Inhibitors Implicates an Earlier Non-Myelin Factor in Preventing

Ganglion Cell Regeneration

ROBERT E. MACLAREN Department of Human Anatomy, University of Oxford, Oxford OX1 3QX, UK

ABSTRACT The pattern of appearance of myelin-associated proteins in the visual system of the

Brazilian opossum Monodelphis dornestica is described. Whole mounts of optic nerve, chiasm, and optic tract were sectioned horizontally and incubated with antibodies to myelin basic protein (MBP), proteolipid protein (PLP), myelin-associated glycoprotein (MAG), “Rip,” and the neurite inhibitory protein (IN-11, followed by visualization with diaminobenzidine and a peroxidase-conjugated secondary antibody. PLP is first detectable 24 days after birth (P24) at the centre of the optic chiasm. MBP, MAG, Rip, and IN-1 appear first in the same area at P26. By P28 the distribution of all proteins is similar, occupying the entire chiasm, optic tracts, and prechiasmatic portion of the optic nerves. Protein expression progresses along the optic nerve to reach the lamina cribrosa by P34, coincident with the time of eye opening. A critical period in which the retinofugal pathway has a regenerative capacity has recently been observed in Monodelphis. This period ends at P12,2 weeks before the appearance of the myelin-associated inhibitory proteins MAG and IN-1. These results therefore suggest that regeneration in the developing retinofugal projection of the opossum is restricted by an earlier non-myelin factor, which is in contrast to current literature on the spinal cord.

Indexing terms: development, IN-1, Mortodelphis, oligodendroeyte, retina

0 1996 Wiley-Liss, Inc.

In recent years, much attention has focused on the role of myelin in the mammalian central nervous system (CNS) as an inhibitor of axonal growth and regeneration. Oligoden- drocytes have been shown to repel growth cones in vitro (Schwab and Thoenen, 1985; Schwab and Caroni, 1988; Fawcett et al., 19891, a phenomenon not observed in neuronal cultures from non-mammalian species where CNS regeneration is feasible (Bastmeyer et al., 1991). Two proteins of 35 kDa and 250 kDa, inhibitory to axonal growth, have been isolated from rat CNS myelin (Caroni and Schwab, 1988). A monoclonal antibody (IN-1) raised against the 250-kDa protein fraction has facilitated limited neurite outgrowth on myelin both in vitro (Savio and Schwab, 1989) and in vivo (Schnell and Schwab, 1990). More recently, myelin-associated glycoprotein (MAG) has also been implicated as a major inhibitor of neurite growth in vitro (Mukhopadhyay et al., 1994).

The importance of the role of myelin as an inhibitor to regeneration has also been underlined by observations of successful regeneration in neonatal mammals before the onset of myelination. Injury to developing spinal cord tracts

of the opossum (Xu and Martin, 1989) and chick (Hasan et al., 1993) have resulted in successful regrowth of CNS axons around the lesion. The end of this critical period of regenerative capacity coincides with the normal developmental appearance of myelin in these species (Shimizu et al., 1990; Ghooray and Martin, 1993); moreover, the critical period has been slightly prolonged by delaying the onset of myelination with X-irradiation (Keirstead et al., 1992). Myelination also heralds the end of a critical period in which late growing (neogenic) axons grow around a lesion of the hamster corticospinal tract (Kalil and Reh, 1979; Reh and Kalil, 1982). Furthermore, in vitro studies in the opossum have also shown a close correlation between the end of a critical period of spinal cord plasticity and the onset of myelination, which significantly also coincided with the onset of IN-1-dependent inhibitory activity (Varga et al., 1995).

Accepted February 13,1996 Address reprint requests to Robert E. MacLaren, Dept. of Human

Anatomy, Univ. of Oxford, South Parks Road, Oxford OX1 3QX, UK.

o 1996 WILEY-LISS, INC.

28

We have recently observed a similar critical postnatal period in the retinofugal pathway of the Brazilian opossum Monodelphis. Up to the age of P12 in this species, ganglion cells are able to regrow through a retinal lesion and renavigate as far as the optic tracts (MacLaren and Taylor, 1995). Since the course of these regrowing fibres involves traversing the visual pathway, which is normally heavily myelinated in the adult, one might assume that the end of this critical period coincides with the onset of myelination. The experiments described in this report use immunohistochemistry to reveal the developmental appearance of the myelin-associated inhibitory proteins recognized by the MAG and IN-1 antibodies. We compare spatially and temporally the appearance of these antigens together with those recognized by the myelin basic protein (MBP), proteolipid protein (PLP), and “Rip” (Friedman et al., 1989) antibodies in the retinofugal pathway o€ Monodelphis. Since, unlike in the spinal cord, there are no grey matter areas in the optic nerve, regrowing neurites have no choice but to grow through future myelinated areas. Similarly, the optic nerve differs from the spinal cord in that it myelinates uniformly as a solitary developing pathway, and not patchily according to the individual developmental states of separately matur- ing pathways (Ghooray and Martin, 1993). These factors make the developing retinofugal pathway an ideal system in which to investigate the relationship between the appearance of myelin-associated proteins and the cessation of regeneration in a CNS tract.

Some of this work has previously been presented in abstract form (MacLaren and Taylor, 1994).

R.E. MACLAREN

MATERIALS AND METHODS Monodelphis pups were bred from an in-house colony,

based on an established breeding protocol (Fadem et al., 1992). Litters of six to ten pups are born after 14 days gestation; the day of birth is designated as PO. The neonatal Monodelphis is a very immature animal in terms of CNS development: At birth, the opossum cerebral cortex is similar to that of a rat at embryonic day 13 (Saunders et al., 1989). Similarly, development of the opossum visual system is protracted, occurring entirely postnatally (Stone et al., 1994; Taylor and Guillery, 19941, rather than in utero, akin to eutherian mammals (Frost et al., 1979; Godement et al., 1984; Reese and Colello, 1992). Eye opening occurs at P34, and the late postnatal ages quoted in this paper should therefore be considered in light of the protracted retinofugal development that is normal for this species.

At various postnatal stages, pups were detached from the mother’s teats, anaesthetized with volatile ether, and per- fused transcardially with phosphate-buffered saline (PBS). Two pups from at least three separate litters were used for each postnatal stage. The optic nerves, chiasm, and lateral walls of the diencephalon were dissected out as a single entity and placed flat on a glass slide. By folding out the lateral walls of the diencephalon flat, the plane of both optic tracts became horizontal, so that the retinofugal pathway was now presented as an “X”-shape, with each component lying parallel to the glass slide. The whole preparation was frozen in Optimal Cutting Compound (OCT), and 10- to 12-pm horizontal sections were collected onto 5% gelati- nized slides from a cryostat. With optimal alignment, four to five complete eye-to-tract sections could be obtained.

Sections were air-dried for 20 minutes and fixed at 4°C for a further 20 minutes either in 5% acetic acid/95%

ethanol (MBP, MAG, and IN-1) or 4% paraformaldehyde (PLP and Rip). Sections were then rehydrated, washed 3 x 5 minutes in PBS, and blocked for 20 minutes with 5% bovine serum albumin (BSA, Sigma) in PBS. Incubation in primary antibody overnight at 4°C followed the protocols outlined below, with three 5-minute PBS washes before addition of secondary antibody for 1 hour at room temperature: 1. MBP (Bohringer Mannheim)-diluted 1:500 with 5%

BSA in PBS. Biotinylated horse anti-mouse IgG secondary diluted 1:200 with 5% BSA in PBS, followed by an avidin-biotin complex (ABC) reaction (Vector Laborato- ries).

2. MAG (Bohringer Mannheiml-diluted 1 5 0 with 5% BSA in PBS. Biotinylated horse anti-mouse IgG secondary diluted 1:200 with 5% BSA in PBS, followed by the ABC reaction as above.

3. IN-1 (A gift from Prof. Martin Schwab)--diluted 1:2 with 5% normal goat serum (NGS, Sigma) in PBS. Peroxidase-conjugated goat anti-mouse IgM secondary (Sigma) diluted 1:200 with 5% normal goat serum (NGS) in PBS.

4. PLP (A gift from Dr. Ray Colello: now available from Serotec)-diluted 1:1,000 with 1% BSA, 4% NGS, and 1% Triton X-100 (Sigma) in PBS. Peroxidase-conjugated goat anti-rabbit IgG secondary (Dako) diluted 1:200 with 5% NGS in PBS.

5 . Rip (A gift from Dr. Joel Black)-diluted 1 : l O with 1% BSA, 4% NGS, and 1% Triton in PBS. Peroxidase- conjugated goat anti-mouse IgG secondary (Jackson) diluted 1:200 in 5% NGS in PBS.

Visualization of the peroxidase reaction product was with 0.1% 3,3’-diaminobenzidine-4 HC1 (DAB, Sigma) dissolved in 0.1 M phosphate buffer, with 0.03% hydrogen peroxide (1 drop of 30% to 50 ml of above solution) as a catalyst. In all cases the DAB reaction was monitored carefully and stopped when appropriate contrast was achieved after 2 to 15 minutes at room temperature.

Specimens were then dehydrated in ascending alcohols, cleared in histoclear, and mounted under glass coverslips in DPX mounting medium.

RESULTS The temporal and spatial pattern of appearance of all

myelin antigens studied is virtually identical, and is typi- cally represented by the PLP series seen in Figure 1. The retinofugal pathway, from the pigmented optic nerve heads to the tracts, is seen in the left column (A, D, G, J, M), the chiasm is shown in the centre column (B, E, H, K, N), with the nerves uppermost, and a high-power view of the chiasm is seen in the right column (C, F, I, L, 0). The earliest age at which any myelin proteins could be detected in the retinofugal pathway of Monodelphis is P24. Here one or two processes labelled with PLP antibody can just be seen in the centre of the optic chiasm (Fig. 1C). At P26, PLP staining is heavier in the central chiasm (Fig. 1D-F), and the high- power view shows bipolar outlines characteristic of early oligodendrocyte morphology (Butt and Ransom, 1993). At this stage all the other antibodies start to appear in the same region (compare also Fig. 3A for IN-1). By P28 heavy staining can now be seen in the chiasm with PLP extending up into the tracts and bases of the optic nerves (Fig. 1G-I). At P28, the distribution of all antibody staining described is

MYELINATION IN A MAMMALIAN RETINOFUGAL PATHWAY 29

Fig. 1. Developmental appearance of the proteolipid protein (PLP) antigen in the opossum visual system. The left column (A,D,G,J,M) shows the progression of reactivity (arrows) from the chiasm to the optic nerve heads. A higher-power view of the chiasm is seen in the

centre column (B,E,H,K,N); the chiasm is outlined in B. A very-high- power view in the right column (C,F,I,L,O) shows a few strands of reactivity first detectable in the centre of the chiasm at P24. Scale bars = 1 mm in A,D,GJ,M, 250 km in B,E,H,K,N, and 25 pm in C,F,I,L,O.

very similar (see Fig. 2 for detailed analysis). At P30 the density of the labelled fibres increases in the chiasm and tracts, whereas immunoreactivity continues further along the optic nerve (Fig. 1J-L). Similarly, a section at P32

shows an increase in density of immunolabelling at the chiasm and further progression along the optic nerves (Fig. 1M-0). By P34 the pattern of immunoreactivity is similar in distribution, but less dense than in the adult animal, in

30 R.E. MACLAREK

Fig. 2. A comparison of the distributions of different myelin antigens in adjacent sections of a single P28 specimen (views are similar to Fig. 1). At the gross level, the pattern of labelling appears similar for all the antigens (A,D,G,J), but a high-power view reveals subtle differences in the distribution of antigens in the somata and processes

(C,F,I,L). Arrows in C and F point to pre-ensheathing processes labelled by the myelin-associated glycoprotein (MAG) and myelin basic protein (MBP) antibodies. Scale bars = 1 mm in A,D,G,J, 250 Fm in B,E,H,K, and 25 Wm in C,F,I,L.

which staining is seen right up to the lamina cribrosa (Fig. 4). The appearance of myelin protein reactivity at the lamina cribrosa at P34 is coincident with eye opening in this species. There was a slight (*l day) variability between different litters, although between individual littermates the pattern of myelin protein reactivity was very similar.

The distribution of immunoreactivity for MAG, MBP, PLP, and Rip in a single P28 preparation can be seen in Figure 2 (the three columns correspond to the same regions as in Fig. 1). At this stage, staining extends from the chiasm into the optic tracts and prechiasmatic portion of the optic nerves. Literally frozen in time, these sections allow one to compare temporally the appearance of the myelin proteins. Taking into account the minor differences produced by

serial sections, these four proteins all extend to the same point along the nerves (Fig. ZA,D,G,J). This suggests that individual oligodendrocytes synthesize these separate proteins at the same time, or at least in rapid succession. In the centre column, the symmetrical reactivity of the inchoate chiasm can be seen, each antibody labelling equally the retinofugal projection from both eyes (Fig. BB,E,H,K).

In the right column of Figure 2, the high-magnification field from the centre of the chiasm reveals subtle differences in staining patterns. Both MAG and MBP reactivity are seen mainly in the somata of chiasm oligodendrocytes, giving a mottled appearance to this region, although occasional processes are also labelled (Fig. ZC,F-arrows). The PLP reactivity appears to be located almost entirely in the


oligodendrocyte processes, with no discernable cell bodies, leading to a crisscrossed appearance (Fig. 21). Rip activity labels mainly processes, but the occasional cell body can also be seen (Fig. 2L, but see also Fig. 5) .

The developmental appearance of the IN-1 antigen in the retinofugal pathway is shown in Figure 3. The absolute limit of early detectable activity is at P26, where two to three oligodendrocyte processes can faintly be seen in the midchiasm (Fig. 3A). These IN-1-labelled fibres have the

Fig. 4. Adult optic nerve head showing distribution of the IN-1 and MAG antibodies. Reactivity is seen in the optic nerve on the right, hut ends abruptly at the lamina cribrosa without entering the retina, in keeping with the known distribution of myelin in this region. The black area to the left is pigmented retina. Scale bars = 25 IJ-m in IN-1 and MAG panels.

Fig. 3. Developmental appearance of the myelin-associated inhibitory protein recognized by the neurite inhibitor protein (IN-1) antibody. The first IN-1 reactivity is detected at P26 in the centre of the chiasm ( A arrows). By P28, the whole outline of the chiasm is discernable, although reactivity has not yet spread into the nerves (B). In the adult (C) heavy IN-1 reactivity is present throughout, from the optic nerves (right) to the optic tracts (left). Scale bars = 10 IJ-m in A, 100 pm in B, and 2 mm in C.

characteristic bipolar morphology of early oligodendrocytes (arrows). By P28 the similarity to the pattern of the other antibodies becomes apparent, as the IN-1 antigen spreads throughout the chiasm (Fig. 3B). In the adult, the reactivity of the IN-1 antigen becomes as strong as the other myelin proteins, with individual fibres no longer detectable (Fig. 3C). Variations in the immunological protocol were tried at earlier stages with positive controls: Fluorescent secondary antibodies and avidin-biotin amplification all failed to show IN-1 activity anywhere in the retinofugal pathway before P26.

MAG and IN-1 reactivity can be seen at the adult retina-optic nerve junction in Figure 4. Here both antigens form a strong boundary behind the lamina cribrosa, corre- sponding to the rodent distribution of myelin in this area (Hildebrand et al., 1985). This also demonstrates the failure of either of these molecules to penetrate into the retina. Myelin is seen in the retina of some species (Schnitzer, 19851, but the absence of MAG or IN-1 demonstrated here shows that this is not true for Monodelphis. We might therefore reasonably conclude that these inhibitory proteins could have no role in repelling growth cones at the site of a lesion in the retina itself (MacLaren and Taylor, 1995). Similarly, the retina of the rat has previously been observed to be devoid of IN-1-sensitive inhibitory activity (Caroni and Schwab, 1989).

32 R.E. MACLAREN

oligodendrocyte could occasionally be seen to be in contact with two axons, these axons were always from the same eye; no oligodendrocyte was observed to be in contact with axons originating from both eyes. The clarity of Rip immunoreactivity for oligodendrocyte processes enabled specific retinofugal pathways to be delineated. This is illustrated in Figure 5B showing half of the Monodelphis chiasm at P30. Fibres from the eye leave the optic nerve (on) to enter the ipsilateral optic tract (ot) without approaching the midline (dotted line). This pattern is typical of marsupials (Taylor and Guillery, 1994). A high-power view of the chiasm/tract boundary in Figure 5C shows Rip-positive processes sur- rounding apparently defasciculated ipsilaterally projecting fibres (arrows). These intersect contralaterally projecting fibres from the opposite eye, to take up a position in the ventral portion of the optic tract. Note the relative unifor- mity of Rip labelling in these separate projections.

The developmental timetable of major events in the Monodelphis retinofugal system is portrayed schematically in Figure 6. The time periods for each event have been estimated by compiling data from recent studies in this species (MacLaren and Taylor, 1994, 1995; Stone et al., 1994; Taylor and Guillery, 1994).

Fig. 5. Distribution of Rip immunoreactivity in the optic chiasm at P30. The clarity of Rip labelling for both cytoplasm and processes of oligodendrocytes can be seen in A, affecting equally the crossed fibres from both eyes. The anatomy of the opossum chiasm is seen in B, with oligodendrocyte processes labelling fibres passing from the optic nerve (on) to the ipsilateral optic tract (ot) without interacting with the chiasm midline (dotted line). The high-power view in C shows processes of the ipsilateral fibres (arrows) intersecting the crossed fibres to take up a position in the ventral part of the optic tract. Scale bars = 25 pm in A, 250 pm in B, and 50 pm in C.

The affinity of the Rip antibody for the soma, connecting stalks, and processes of oligodendrocytes can be seen in Figure 5A (P28). This pattern is very similar to that seen with Rip in Triton-treated sections of hamster optic tract (Jhaveri et al., 1992). Note also the crossed processes in this figure. Rip demonstrates quite clearly the symmetrical nature of the myelin protein pattern, affecting equally the crossing fibres from both eyes. Interestingly, although one

DISCUSSION Developmental localization of myelin proteins

PLP reactivity was first seen in the midchiasmatic region at P24. The slightly earlier appearance of PLP compared to the other myelin proteins is in contrast to previous observations in the rat, in which MBP precedes PLP (Hartman et al., 1982). The early visualization of PLP in the opossum, however, may not necessarily reflect cellular events. PLP shows a remarkable conservation of amino acid sequences across different species (Mikoshiba et al., 19911, which is not true for MBP (Sires et al., 1981). Thus, when using these rodent-derived antibodies in the opossum, the heavier PLP immunological staining may simply reflect better antibody affinity for the well-conserved PLP antigen. The relative concentrations of myelin membrane proteins are PLP-50%, MBP-30%, and MAG-1% (Agrawal et al., 1977; Hartman et al., 1982; Trapp, 1990). The higher concentra- tion of PLP may thus also be a factor leading to early visualization.

In this study, we assume that the appearance of myelin proteins in the retinofugal system heralds the onset of myelination. This has been noted in other species (Hilde- brand et al., 1993) and is supported here by the microscopic appearance of oligodendrocyte processes (Fig. 5A), which have an orientation and morphology consistent with ensheathment of axons by oligodendroglial processes (Remahl and Hildebrand, 1990). A similar immunological localization of myelin antigens in the developing rat brain demonstrated the appearance of PLP and MBP only in fully differentiated oligodendrocytes after ensheathment had started (Hartman et al., 1982). For these reasons, a chiasm- to-eye gradient of myelination could be predicted, although examination at the ultrastructural level would be necessary to identify a gradient defined by axonal wrapping and formation of compact lamellae.

After P26, the spread of immunoreactivity is away from the chiasm: rostrally towards the eye and caudally towards the superior colliculus. Thus the myelin proteins appear initially in the middle of this fibre tract and then spread in both directions. This is a pattern that is consistent with

MYELINATION IN A MAMMALIAN RETINOFUGAL PATHWAY

Conception Birth

33

Eye-opening

I I Myelinafion

Neurogenesis I I

EO E7 Po M P14 P21 P28 P35 Age of Monodelphis / days

Fig. 6. Schematic representation of developmental events in the retinofugal system of Monodelphis. The span of each neuron represents the developmental period of each stage, based on data compounded from various studies in this species (see Results). Note the end of the critical period of regeneration corresponds more closely to the onset of arborization than myelination.

current literature on oligodendrocyte development in the optic nerve. A glial progenitor (0-2A) cell has been described that can give rise to oligodendrocytes (Raff et al., 1983). This motile progenitor cell is thought to migrate into the optic nerve from the subventricular zone of the developing rat brain. The optic chiasm is closest to this zone, and the 0-2A cells populate this region first before moving along the nerves (Small et al., 19871, to differentiate into non-motile oligodendrocyte precursors (Miller et al., 1985). The oligodendrocyte precursors then undergo a sequence of programmed events, largely independent of external factors, leading to MBP transcription and translation (Zeller et al., 1985). We might tentatively conclude, therefore, that the optic chiasm is the region of the visual system in which myelin proteins first appear, because this is the region first populated by 0-2A progenitor cells.

Intracellular localization of myelin proteins Although PLP was chosen to illustrate the pattern of

myelin protein progression along the visual pathway, the distribution of MBP, MAG, and Rip is virtually identical at any one time (IN-1 is similar, but was not directly compared on adjacent sections in Fig. 2). This synchronization may be coincidental, but since all of these proteins are structural elements of a single myelin sheath (Mikoshiba et al., 1991; Berger and Frotscher, 1994), it may be that transcription of each individual gene is jointly controlled at a higher level. In contrast to this, the differences in distribution of proteins at the cellular level are quite marked (Fig. 2). MAG and MBP appear localized mainly in the cytoplasm, whereas PLP and Rip appear mainly in the processes. This could be

due to a number of factors. First, at a technical level, sections for both MAG and MBP antibodies were permeabi- lized with glacial acetic acid (see Materials and Methods), whereas Triton X-100 was used for PLP and Rip. Binding of the Rip antibody to the lipid-enriched myelinated processes of the oligodendrocyte has previously been shown to require the presence of Triton X-100 (Friedman et al., 1989). In the absence of detergent, binding is predominantly to the soma (J. Black, personal communication). Second, variations in cross-species conservation of antigens may also be a factor. Although little is known of the structure of the Rip epitope, PLP is known to be well conserved across species and is clearly demonstrated in the processes of the Monoddphis oligodendrocyte (Fig. 2). The structure of MBP and MAG is more variable (Mikoshiba et al., 1991; Hildebrand et al., 1993). Both undergo considerable post-translational modifications between synthesis and insertion into the plasma membrane (Trapp, 1990; Brophy et al., 1993). Significant cross-species variations in these modifications could result in the rodent-derived antibodies recognising intracellular but not peri-axonal membrane-bound isoforms of MBP and MAG in the opossum. Finally, a delay in the insertion of MAG and MBP into the membrane after translation may be a factor, but this delay would have to be substantial in order to give the high ratio of intracellular-to-process localization of the antigens observed (see Fig. 2: MAG and MBP vs. PLP and Rip). This is unlikely, since there is good evidence that translocation of MBP mRNA into the oligodendrocyte processes results in almost immediate insertion of newly translated MBP into the plasma membrane (Zeller et al., 1985; Trapp et al., 1987). MAG is synthesized on rough

34 R.E. MACLAREN

endoplasmic reticulum and transported in vesicles to the plasma membrane, but the delay between synthesis and membrane insertion is still only about 30 minutes (Colman et al., 1982).

In reality, any meaningful analysis of the differences in cellular localization of the proteins observed would require better knowledge of Rip biochemistry, uniform methods of tissue preparation, species-specific antibodies, and ribo- probe analysis.

IN-1 and regeneration in the opossum We have recently demonstrated a critical period in the

postnatal Monodelphis in which axons are able to regrow across a lesion in the retina to reinnervate targets in the optic tracts (MacLaren and Taylor, 1995). Transient periods also exist during the development of descending spinal pathways in other species, during which a limited degree of regeneration is feasible. The end of these periods corresponds roughly to the onset of spinal cord myelination (Reh and Kalil, 1982; Keirstead et al., 1992; Ghooray and Martin, 1993). In Monodelphis, however, regeneration ceases a t P12, whereas the earliest stages of myelination and appearance of inhibitory proteins are not until P24126. This lengthy delay suggests there may be other reasons for regeneration failure at earlier stages and raises some interesting questions.

At a technical level, IN-1 is a difficult antibody to use for immunohistochemistry. One might therefore argue that the inhibitory protein is present but not detected at earlier stages, but this is unlikely for several reasons. First, if one assumes that the inhibitory protein does actually appear earlier, by about P12, then one would have to accept as a coincidence that the threshold for detection is met at exactly the same time and anatomical location at which the other proteins appear at P26. Second, the morphology of the earliest oligodendrocytes at P26, progressing to a chiasmatic labelling that is light at P28 and heavy in the adult, is very suggestive of a developmental rather than technical series of events (Fig. 3). Third, the direct comparison to MAG in Figure 4 suggests that IN-1 reacts well in the adult. The pattern of activity in this region is in keeping with the known distribution of myelin and suggests that these antibodies do recognize oligodendrocyte antigens in Monodelphis. Hence we propose that the myelin-associated inhibitory protein recognised by IN-1 appears at approxi- mately the same time as the proteins recognized by the MAG, MBP, PLP, and Rip antibodies.

A previous report has suggested that IN-1 inhibitory activity is present in the rat optic nerve at PO, some days before the appearance of MBP at P5 (Caroni and Schwab, 1989). In that in vitro study, the presence of inhibitory molecules was deduced indirectly by the observation of enhanced fibroblast spreading in the presence of IN-1 on optic nerve explants. The observation is surprising, as it coincides with the population of the optic nerve, not by oligodendrocytes, but by the highly motile 02-A progenitor cells (Pringle et al., 19921, suggesting that these cells are the initial source of IN-1-sensitive inhibitory activity. Interestingly, this study did show that direct IN-1 staining on dissociated cells was not significant until P5, the time at which MBP staining is just starting, which is more in keeping with the observations in Monodelphis.

If the IN-1 immunoreactivity seen at P26 corresponds to the time at which inhibitory molecules appear in the visual system ofMonodeLphis, it does not correlate with the end of

the critical period of plasticity ending at P12. In studies of the lesioned spinal cord including other species, the end of a critical period of regeneration has often coincided with myelination (Keirstead et al., 1992; Ghooray and Martin. 1993) or myelin-associated inhibitory activity (Schwab et al., 1993; Kapfhammer and Schwab, 1994; Varga et al., 1995). The spinal cord, however, is a difficult area in which to look for correlation. There are many fibre tracts at different stages of development and myelinating at different rates; one tract may be permissive whilst its neighbour is inhibitory (Lahr and Stelzner, 1990). In the spinal cord of Didelphis for instance, MBP immunoreactivity starts at P15 and is not complete until P54 (Ghooray and Martin, 1993). Similarly, not much attention is paid to the spinal cord grey matter, which has areas devoid of IN-1 reactivity (Rubin et al., 1994), and could potentially support fibre outgrowth (Savio and Schwab, 1989). The ability of embryonic neurons to grow in heavily myelinated areas (Wictorin et al., 1990), challenges the role that myelin-associated inhibitory proteins might have in preventing regeneration of developing pathways.

The maturation of glial elements may have an important role in defining a critical period, because reactive astrocytes have also been shown to express factors that are inhibitory to axonal growth (Bovolenta et al., 1993). Furthermore, this inhibitory activity increases as development proceeds (Smith et al., 1986; Ajemian et al., 1994). The reactive gliosis could be in the optic nerve where astrocytes sur- round degenerating ganglion cell axons (Trimmer and Wunderlich, 19901, but more likely would have to be in the retina around the lesion site itself. In the latter case, most of the “gliosis” may be mediated by Muller cells, which react to injury in a similar way to astrocytes by increasing glial fibrillary acidic protein (GFAP) expression (Scherer and Schnitzer, 1991; MacLaren, 1996). Muller cells, however, have not yet been shown to express the specific inhibitory molecules seen on reactive astrocytes (McKeon et al., 1991; Tiveron et al., 1992; MacLaren, 1996), and are normally supportive for growth of ganglion cells during development (Araujo and Linden, 1993). Thus, the role of the glial reaction to retinal injury still remains uncertain.

In a study using the Rip antibody in the hamster, myelination in the optic tract occurred just 3-4 days after the end of a critical period of retinofugal plasticity (Jhaveri et al., 1992). Thus a rough correlation exists that is in contrast to these observations in Monodelphis. This apparent discordance m2y be explained by differences in the site of axotomy, leading to a fundamental difference in the two mechanisms of regeneration. In the hamster the axon is lesioned distally, near to the region of terminal arborization (So et al., 19811, whereas in Monodelphis the axon is lesioned close to the soma in the retina (MacLaren and Taylor, 1995). Thus in the hamster paradigm, the axoto- mized ganglion cell still projects to a distal site, and although lesioned, may still receive sufficient growth factors from its distal stump to remain viable. With viability of the soma ensured, resprouting from the lesioned axon can be attempted, the outcome of which then becomes dependent upon the effects of myelin-associated inhibitory proteins. In Monodelphis, however, retinal axotomy com- pletely separates the ganglion cell body from any distal trophic factors. If this occurs at a developmental stage in which the survival of the ganglion cell is critically dependent upon distal growth factors, cell death will occur and resprouting cannot even be attempted. The presence or


absence of myelin-associated molecules inhibitory to axonal sprouting in this case becomes irrelevant.

CONCLUSIONS The expression of myelin-associated inhibitory proteins

appears not to be the major factor halting regeneration in the developing visual system of the opossum. The end of this critical period correlates more closely with the time of arborization, when these neurons start to innervate distal targets. The susceptibility of a ganglion cell to axotomy, a t a time when it is critically dependent upon target-derived growth factors, may be the overriding factor in defining this critical period.

ACKNOWLEDGMENTS This study would not have been possible without the

generous gifts of the IN-1 from Prof. Martin Schwab and “Rip” from Dr. Joel Black. Dr. Ray Colello advised on the techniques of optic nerve sectioning and myelin immunohistochemistry described. I am also most appreciative of the help and advice given by the project supervisor Dr. Jeremy Taylor, who together with Prof. Ray Guillery, made con- structive comments on the manuscript. This research was supported by the Medical Research Council (UK) and the Wellcome Trust.

LITERATURE CITED Agrawal, H.C., B.K. Hartman, W.T. Shearer, S. Kalmbach, and F.L.

Margolis (1977) Purification and immunohistochemical localization of rat brain myelin proteolipid protein. J. Neurochem. 28:495-508.

Ajemian, A,, R. Ness, and S. David (1994) Tenascin in the injured rat optic nerve and in non-neuronal cells in vitro: potential role in neuronal repair. J. Comp. Neurol. 340:233-242.

Araujo, E.G., and R. Linden (1993) Trophic factors produced by retinal cells increase the survival of retinal ganglion cells in vitro. Eur. J. Neurosci. 511181-1188.

Bastmeyer, M., M. Beckmann, M.E. Schwab, and C.A. Stuermer (1991) Growth of regenerating goldfish axons is inhibited by rat oligodendrocytes and CNS myelin but not by goldfish optic nerve tract oligodendrocyte like cells and fish CNS myelin. J. Neurosci. 11(3):626640.

Berger, T., and M. Frotscher (1994) Distribution and morphological charac- teristics of oligodendrocytes in the rat hippocampus in situ and in uitro: an immunocytochemical study with the monoclonal Rip antibody. J. Neurocytol. 23.61-74.

Bovolenta, P., F. Wandosell, and M. Nieto-Sampedro (1993) Characteriza- tion of a neurite outgrowth inhibitor expressed after CNS injury. Eur. J. Neurosci. 5:454465.

Brophy, P., G. Boccaccio, and D. Colman (1993) The distribution of myelin basic protein mRNAs within myelinating oligodendrocytes. Trends Neurosci. I6(12):5 15-521.

Butt, A.M., and B.R. Ransom (1993) Morphology of astrocytes and oligodendrocytes during development in the intact rat optic nerve. J. Comp. Neurol. 338:141-158.

Caroni, P., and M.E. Schwah (1988) Two membrane protein fractions from rat central myelin with inhibitory properties for neurite growth and fibroblast spreading. J. Cell Biol. 106:1281-1288.

Caroni, P., and M.E. Schwab (1989) Codistrihution of neurite growth inhibitors and oligodendrocytes in rat CNS: appearance follows nerve fibre growth and precedes myelination. Dev. Biol. 136287-295.

Colman, D.R., G. Kreihich, A.B. Frey, and D.D. Sabatini (1982) Synthesis and incorporation of myelin polypeptide into CNS myelin. J. Cell Biol. 95598-608.

Fadem, B.H., G.L. Trupin, E. Maliniak, J.L. VandeBerg, and V. Hayssen (1992) Care and breeding of the gray short-tailed opossum (Monodelphzs domestics). Lab. Anim. Sci. 32f41t405-409.

Fawcett, J.W., J. Rokos, and I. Bakst (1989) Oligodendrocytes repel axons and cause growth cone collapse. J. Cell Sci. 9293-100.

Friedman, B., S. Hockfield, J.A. Black, K.A. Woodruff, and S.G. Waxman (1989) In situ demonstration of mature oligodendrocytes and their processes: an immunocytochemical study with a new monoclonal antibody, Rip. Glia 2380-390.

Frost, D.O., K.F. So, and G.E. Schneider (1979) Postnatal development of retinal projections in Syrian hamsters: a study using autoradiographic and anterograde degeneration techniques. Neuroscience 4: 1649-1677.

Ghooray, G.T., and G.F. Martin (1993) The development of myelin in the spinal cord of the North American opossum and its possible role in loss of rubrospinal plasticity. A study using myelin basic protein and galactoce- rebrocide immuno-histocbemistry. Dev. Brain Res. 7267-74.

Godement, P., J. Salaiin, and M. Imbert (1984) Prenatal and postnatal development of retinogeniculate and retinocollicular projections in the mouse. J. Comp. Neurol. 230:552-575.

Hartman, B.K., H.C. Agrawal, D. Agrawal, and S. Kalmbach (1982) Develop- ment and maturation of central nervous system myelin: comparison of immunohistochemical localization of proteolipid proten and hasic protein in myelin and oligodendrocytes. Proc. Natl. Acad. Sci. U.S.A. 79:42174220.

Hasan, S.J., H.S. Keirstead, G.D. Muir, and J.D. Steeves (1993) Axonal regeneration contributes to repair of injured hrainstem-spinal neurons in embryonic chick. J. Neurosci. 13(2):492-507.

Hildebrand, C., S. Remahl, and S.G. Waxman (1985) Axo-glial relations in the retina-optic nerve junction of the adult rat: electron-microscopic observations. J. Neurocytol. 14597-617,

Hildebrand, C., S. Remahl, H. Person, and C. Bjartmar (1993) Myelinated nerve fibres in the CNS. Prog. Neurobiol. 40:319-384.

Jhaveri, S., R.S. Erzurumlu, B. Friedman, and G.E. Schneider (1992) Oligodendrocytes and myelin formation along the optic tract of the developing hamster: an immunohistochemical study using the Rip antibody. Glia 61138-148.

Kalil, K., and T. Reh (1979) Regrowth of severed axons in the neonatal central nervous system: establishment of normal connections. Science ZO511158-1160.

Kapfhammer, J.P., and M.E. Schwab (1994) Inverse patterns of myelination and GAP-43 expression in the adult CNS: neurite growth inhibitors as regulators of neuronal plasticity? J. Comp. Neurol. 340: 194-206.

Keirstead, H.S., S.J. Hasan, G.D. Muir, and J.D. Steeves (1992) Suppression of the onset of myelination extends the permissive period for the functional repair of embryonic spinal cord. Proc. Natl. Acad. Sci. U.S.A. 89: 1164-1 168.

Lahr, S.P., and D.J. Stelzner (1990) Anatomical studies of dorsal column axons and dorsal root ganglion cells after spinal cord injury in the newborn rat. J. Comp. Neurol. 293:377-398.

MacLaren, R.E. (1996) Development and role of retinal glia in regeneration of ganglion cells following retinal injury. Br. J. Ophthalmol. 5 4 5 8 4 6 4 .

MacLaren, R.E., and J.S.H. Taylor (1994) Myelination in the optic nerve of the opossum: correlations to a transient regenerative capacity of retinal ganglion cells. Soc. Neurosci. Abstr. 201467.

MacLaren, R.E., and J.S.H. Taylor (1995) A critical period for axon regrowth through a lesion in the developing mammalian retina. Eur. J. Neurosci. 10:2111-2118.

McKeon, R.J., R.C. Schreiber, J.S. Rudge, and J. Silver (1991) Reduction of neurite outgrowth in a model of glial scarring following CNS injury is correlated with the expression of inhibitory molecules on reactive astrocytes. J. Neurosci. 11(11):3398-3411.

Mikoshiba, K., H. Okano, T. Tamura, and K. Ikenaka (1991) Structure and function of myelin protein genes. Annu. Rev. Neurosci. 14:201-217.

Miller, R.H., S. David, R. Patel, E.R. Abney, and M. Raff (1985) A quantitative immunohistochemical study of macroglial cell development in the rat optic nerve: in vivo evidence for two distinct astrocyte lineages. Dev. Biol. 111:35-41.

Mukhopadhyay, G., P. Doherty, F.S. Walsh, P.R. Crocker, and M.T. Filbin (1994) A novel role for myelin-associated glycoprotein as an inhibitor of axonal regeneration. Neuron 13:757-767.

Pringle, N.P., H.S. Mudhar, E.J. Collarini, and W.D. Richardson (1992) PDGF receptors in the rat CNS: during late neurogenesis, PDGF alpha-receptor expression appears to be restricted to glial cells of the oligodendrocyte lineage. Development 115535651.

RaR, M.C., R.H. Miller, and M. Noble (1983) A glial progenitor cell that develops in uitro into an astrocyte or an oligodendrocyte depending on the culture medium. Nature 303:390-396.

Reese, B.E., and R.J. Colello (1992) Neurogenesis in the retinal ganglion cell layer of the rat. Neuroscience 46:419-429.

Reh, T., and K. Kalil (1982) Development of the pyramidal tract in the hamster. 11. An electron microscopic study. J. Comp. Neurol. 205177238,

36

Remahl, S., and C. Hildebrand (1990) Relation between axons and oligodendroglial cells during initial myelination. I. The glial unit. J. Neurocytol. 19: 3 13-328.

Rubin, B.P., I. Dusart, and M.E. Schwab (1994) A monoclonal antibody (IN-1) which neutralizes neurite growth inhibitory proteins in the rat CNS recognizes antigens localized in CNS myelin. J. Neurocytol. 23,209- 217.

Saunders, N.R., E. Adam, M. Reader, and K. Mdlgird (1989) Monodelphis dornestica (grey short-tailed opossum): an accessible model for studies of early neocortical development. Anat. Embryol. (Berl.) 280,227-236.

Savio, T., and M.E. Schwab (1989) Rat CNS white matter, hut not gray matter, is nonpermissive for neuronal cell adhesion and fiber outgrowth. J. Neurosci. 9(4):1126-1133.

Scherer, J., and J. Schnitzer (1991) Intraorbital transection of the rabbit optic nerve: consequences for ganglion cells and neuroglia in the retina. J. Comp. Neurol. 312,175-192.

Schnell, L., and M.E. Schwab (1990) Axonal regeneration in the rat spinal cord produced by an antibody against myelin-associated neurite growth inhibitors. Nature 343:269-272.

Schnitzer, J. (1985) Distribution and immunoreactivity of glia in the retina of the rabbit. J. Comp. Neurol. 240:128-142.

Schwab, M.E., and P. Caroni (1988) Oligodendrocytes and CNS myelin are nonpermissive substrates for neurite growth and fibroblast spreading in vitro. J. Neurosci. 8(7/:2382-2393.

Schwab, M.E., and H. Thoenen (1985) Dissociated neurons regenerate into sciatic but not optic nerve explants in culture irrespective of neuro- trophic factors. J. Neurosci. 5(9):2415-2423.

Schwab, M.E., C.E. Bandtlow, Z. Varga, and J. Nicholls (1993) Developmen- tal expression of myelin-associated neurite growth inhibitors correlates with the loss of regeneration after spinal cord lesions in the opossum. SOC. Neurosci. Abstr. 19:283.19.

Shimizu, l., R.W. Oppenheim, M. O’Brien, and A. Schneiderman (1990) Anatomical and functional recovery following spinal cord transection in the chick embryo. J. Neurobiol. 21(6/:91%937.

Sires, L.R., S. Hruby, E.C. Alvord, I . Hellstrom, and J.E. Casnellie (1981) Species restrictions of a monoclonal antibody reacting with residues 130 to 137 in encephalitogenic myelin basic protein. Science 214:87-89.

Small, R.K., P. Riddle, and M. Noble (1987) Evidence for migration of oligodendrocyte-type-2 astrocyte progenitor cells into the developing rat optic nerve. Nature 328:155-157.

R.E. MACLAREN

Smith, G.M., R.H. Miller, and J. Silver (1986) Changing role of forebrain astrocytes during development, regenerative failure and induced regeneration upon transplantation. J. Comp. Neurol. 251:23-43.

So, K.F., G.E. Schneider, and S. Ayes (1981) Lesions ofthe brachium of the superior colliculus in neonate hamsters: correlation of anatomy with behavior. Exp. Neurol. 72379-400.

Stone, E., C. Jensen, J. Iqhal, J. Elmquist, C.D. Jacobson, and D.S Sakaguchi (1994) Postnatal neurogenesis in the developing retina of the Brazilian opossum. SOC. Neurosci. Abstr. 20:1322.

Taylor, J.S.H., and R.W. Guillely (1994) Early development of the opticchi- asm in the Gray Short-Tailed Opossum, Monodelphis dornestica. J. Comp. Neurol. 350:109-121.

Tiveron, M.C., E. Barhoni, F.B. Rivero, A.M. Gormley, P.J. Seeley, F. Grosveld, and R. Morris (1992) Selective inhibition of neurite outgrowth on mature astrocytes by Thy-1 glycoprotein. Nature 355:745-748.

Trapp, B.D. (1990) Myelin-associated glycoprotein location and potential functions. Ann. N.Y. Acad. Sci. 605:29-43.

Trapp, B.D., T. Moench, M. Pulley, E. Barbosal, and G. Tennekoon (1987) Spatial segregation of mRNA encoding myelin-specific proteins. Proc. Natl. Acad. Sci. U S A . 84:7773-7777.

Trimmer, P.A., and R.E. Wunderlich (1990) Changes in astrological scar formation in rat optic nerve as a function of development. J. Comp. Neurol. 296,359-378.

Varga, Z.M., C.E. Bandtlow, S.D. Erulkar, M.E. Schwab, and J.G. Nicholls (1995) The critical period for repair of CNS of neonatal opossum (Monodelphis domestica) in culture: correlation with development of glial cells, myelin and growth-inhibitory molecules. Eur. J. Neurosci. 7110/:2119-2 129.

Wictorin, K., P. Brundin, B. Gustavii, 0. Lindvall, and A. Bjorklund (1990) Reformation of long axon pathways in adult rat central nervous system by human forebrain neuroblasts. Nature 347,556558,

Xu, X.M., and G.F. Martin (1989) Developmental plasticity of the rubrospinal tract in the North American Opossum. J. Comp. Neurol. 279:36% 381.

Zeller, N.K., T.N. Behar, M.E. Dubois-Dalcq, and R.A. Lazzarini (1985) The timely expression of myelin basic protein gene in cultured rat brain oligodendrocytes is independent of continuous neuronal influences. J. Neurosci. 5(1 1):2955-2962.

expression of myelin proteins in the opossum optic nerve: late appearance of inhibitors implicates...

Documents