EXPERIMENTAL
NEUROLOGY
111,65-73
(1991)
Adult Rat Retinal Glia in Vitro: Effects of in Viva Crush-Activation on Glia Proliferation and Permissiveness for Regenerating Retinal Ganglion Cell Axons M. B#HR Department
of Neurology,
University Spemannstr,35/Z,
Tiibingen, and Max-Planck 7400 Tiibingen, Federal
Press.
fiir Entwicklungsbiologie of Germany
Tiibingen,
axon-glia interactions after central nervous system (CNS) lesions and during regeneration have given new insights into the functions of adult CNS glia. Glia involved in axon ensheathment or myelination, like type II astrocytes and oligodendrocytes (for review see (37)), seem to be rather poor substrata for axonal growth (19, 42) and components of adult CNS myelin have been shown to constitute a nonpermissive environment for regenerating neurites (13, 14). Type I astrocytes have been shown to react by hypertrophy and formation of a glial scar after brain lesions (16,38), thus blocking axon regeneration (26). This aspect, however, seems to be a function of maturity since “immature” astrocytes have been shown to promote axon regeneration in the adult CNS (41). In contrast, glia contained in peripheral nerves used as grafts to reconnect lesioned CNS tracts have been shown to support axonal growth of adult CNS neurons (for review see (11)). We have recently demonstrated that isolated purified Schwann cells support axon regeneration from adult rat retinal explants in vitro (5, 7). An exudate derived from regenerating sciatic nerves that is presumably produced by Schwann cells contains a powerful neurotrophic activity for adult rat retinal ganglion cells (RGC) (44). These results suggest that the inability of CNS neurons to survive axotomy and to regenerate an axon may not be due to a cell-intrinsic defect, but rather may be determined by influences of the surrounding glial environment. Support of this hypothesis was found in recent experiments of adult rat retinal explants cultured on cortical astrocytes where no enhanced survival or neurite outgrowth of retinal ganglion cells was observed (6). Only “early” astrocytes, resembling immature astrocytes in uiuo, which still express laminin (23) were able to support axon growth from adult rat retinal explants (6). This suggests that during the time when CNS fiber tracts are established, astrocytes might be involved in promoting axon growth (43) and guiding axons to their
The effects of optic nerve crush on adult rat retinal glia activation were studied in vitro. In adult rats the optic nerves were crushed and the corresponding retinae were explanted 5 to 7 days later and cultured in vitro. The glial response of retinae with precrushed optic nerves was compared to the glial response of retinae without prior optic nerve crush. As a consequence of crush-axotomy more glial cells migrated out from retinal explants and covered significantly larger areas of the substratum than glia from noncrushed retinae. Migration of immunohistochemically distinguishable Vimentin-positive Mtiller cells and glial fibrillary acidic protein-positive astrocytes could be observed in both types of cultures. Astrocytes as well as Muller cells incorporated bromodeoxyuridine after explantation. In noncrushed retinal explants Thy l.l-immunopositive flat cells were much more frequent and the relative proportion of glial cells was much lower than in crush-activated cultures. In a second set of experiments the ability of adult rat retinal glia to support retinal ganglion cell regeneration was examined. Normal retinal explants (without optic nerve crush) which usually do not substantially regenerate axons were cultured on retinal glia from normal and crush-activated explants. Both glia preparations supported axon growth from retinal explants after 3 days in vitro. Neuritic growth was significantly better when retinal explants from normal adult rats were cultured on crush-activated retinal glia as compared to glia derived from noncrushed retinae. It is concluded that activated adult rat retinal glia, unlike adult glia found in other brain regions, support adult rat retinal ganglion cell regeneration in vitro. 0 1991 Academic
Znstitut Republic
Inc.
INTRODUCTION
During recent years several new functions have been added to the traditional role of glia, such as providing trophic support for neurons and promoting axonal growth (1, 6, 15, 17, 27-29, 41, 43). In particular the 65
All
Copyright Q 1991 rights of reproduction
0014.4886/91 $3.00 by Academic Press, Inc. in any form reserved.
66
M. B.&HR
target regions (30). This ability of astrocytes to support neuritic growth seems to be lost in the normal adult mammalian CNS. Cortical astrocytes from adult rat brain, however, that are activated in situ by neuronal degeneration (24) before cultivation seem to be able to support neuronal survival and axon growth. It has been shown that during retinal development Miiller cells, which are a kind of specialized glia for RGC (12), support retinal ganglion cell survival (2, 36). This dependence of retinal ganglion cells on Miiller cells in vitro (36) is lost at a time corresponding to the in uiuo situation when the adult retinotectal projection has been established. Unlike other parts of the CNS, however, retinal ganglion cells seem to be able to regenerate axons within the retina even in adulthood (20,28). This suggests that retinal glia, which in the rat does not contain type II astrocytes or oligodendrocytes (19, 45), might be able to support regeneration of adult CNS axons. We have recently demonstrated that retinal glia transiently increase their cell surface ganglioside expression after conditioning lesions applied to the optic nerve in uiuo (8). Gangliosides recognized by the monoclonal antibody “JONES” were reexpressed at levels found during embryonic development of the retinotectal projection (8). Optic nerve glia did not show a similar behavior, suggesting that retinal glia might have preserved its ability to support adult rat retinal ganglion cell regeneration by mediating the effects of neurite promoting or neurotrophic factors (8). The aim of the present study was to examine the effects of in uiuo crush activation on adult rat retinal glia and to examine whether adult rat retinal glia are able to proliferate in response to lesions. In addition, the ability of adult rat retinal glia to support neuritic growth from adult rat retinal explants in uitro was investigated. MATERIAL
AND
METHODS
Surgical Procedures Twenty adult Lewis rats were used in this study. To apply a conditioning lesion to the optic nerve 10 rats were deeply anesthetized with ip chloral hydrate (0.42 mglkg body weight). The optic nerve was crush-axotomized intraorbitally (“conditioning” lesion) at a distance of about 3 mm from the optic disc by means of fine watchmaker forceps (3-5). The efficacy of the optic nerve crush was determined by complete loss of the pupillary reaction on the contralateral eye. Care was taken to not alter retinal blood supply and in each crushed eye the retinal blood vessels were inspected through the lens. Only animals that showed a normal retinal blood circulation were used in this study. Explantation Techniques Five to seven days after optic nerve crush the corresponding retinae were dissected as described recently
(3). Briefly, the retinae were freed from connective and vitreal tissue and whole-mounted on semipermeable membrane filters (Millipore) in Hanks’ buffered salt solution with the ganglion cell layer upward. Flat mounted retinae were cut into eight triangular pieces with their tips centered in the former optic disc. Retinal explants were transferred to polylysine (PL; Sigma, 1 mg/ml in Hanks’)- and laminin (Lam; BRL, 10 pg/ml in Hanks’)coated petri dishes (Petriperm, Hereaus) and cultured with the ganglion cell layer attached to cellular or acellular substrates. Cultures were fed with 3 ml Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal calf serum every 3 days and incubated with 5% CO, at 37°C. Retinal Glia CultiuationlCocultiuation and Adult Retinal Explants
of Retinal Glia
Twenty retinal explants with (n = 10) or without (n = 10) conditioning lesions prior to explantation were scored every 3 days for retinal ganglion cell axon growth and glialflat cell migration. Five to eight days after explantation the retinal pieces were removed from the dishes, resulting in cultures of single cells that had migrated out from the explants. Two days later precrushed or noncrushed retinal explants derived from other retinae were placed on the retinal glia monolayers and cultured for 3-5 days (n = 20). Retinal glia or glia-retina cocultures were fixed in 4% paraformaldehyde in phosphate buffer at pH 7.4 (PFA) and processed for immunohistochemistry. Immunohistochemistry Retinal glia cultures (n = 80) or glia-retina (n = 20) cocultures were fixed by overnight incubation in PFA after different time points in culture. Then cultures were rinsed in DMEM and incubated with anti-Thy 1.1 monoclonal antibody (Serotec, 100 pg/ml in DMEM), anti-Vimentin (Boehringer and Sigma, 100 pglml in DMEM), anti-glial fibrillary acidic protein (GFAP; Dakopatts, 2 g/ml in DMEM), or anti-neurofilament antibody (SMI 31; Sternberger & Meyer Immunocytochemicals, 2 @g/ml in DMEM). For GFAP, Vimentin, and SMI 31 immunohistochemistry, specimens were permeabilized with methanol at -20°C for 10 min. Cultures were then rinsed again in DMEM until the pH had normalized; incubated for 45 min with anti-GFAP, anti-Vimentin, or anti-neurofilament antibodies; and rinsed and stained by fluorescein- or rhodamine-conjugated secondary antibodies (Cappel). Specimens were viewed with epifluorescence on an inverted microscope (Zeiss) and photographed on Kodak Tri-X-Pan or Ektachrome films. Retinal Glia Proliferation To determine retinal glia proliferation, bromodeoxyuridine (BrdU; Bioscience Products, 10 pmol) was ap-
ADULT
TABLE Retinal Ganglion Polylysine-Laminin Precrushed Days
in vitro
Number
3 6 9
10f 6 31 k 10 52 k 23
Cell
RAT
RETINAL
67
GLIA
1
RESULTS
Axon Growth Substrates
retina Length 140 + 53 208 f 63 1112 f 388
on
Normal Number 0 12 k 5 11 t6
Retinal Ganglion Cell Axon Growth on PolylysineLaminin Substrates retina Length 0 74k 13 375 k 223
Note. The table shows the total number and the length (mean diameter of the neuritic halo; values are means + standard deviation) of RGC neurites that have reextended from precrushed and normal adult rat retinal explants (n = 80). At all time points analyzed significantly more RGC axons extend from precrushed retinae compared to normal retinae (t test, P G 0.01).
plied to retinal tissue cultures at 6 days after explantation (n = 40). BrdU is incorporated into the DNA of dividing cells and can be detected by specific antibodies. Retinal cultures were incubated with BrdU for 12 h, fixed in PFA, and then incubated with anti-BrdU antibodies. To visualize anti-BrdU antibodies, specimens were incubated with fluorescein- or rhodamine-conjugated secondary antibodies (Cappel). For double-labeling immunohistochemistry of BrdU and GFAP, Vimentin (VIM) or Thy 1.1 standard procedures were used as described above. Electron Microscopic Studies For scanning electron microscopic examination of retinal ganglion cell axons extending on adult rat retinal glia, cultures were fixed in 2% glutaraldehyde in PBS and osmificated in 1% 0~0, in PBS. After dehydration in graded ethanols specimens were dried in CO, by the critical point method and finally coated with gold. After that areas near the explant’s tips were viewed in an autoscan scanning electron microscope with an angle of 45”.
The extent of RGC axon growth was determined at 3, 6, and 9 days after explantation. After 3 days (see Table 1) no axons extend from retinae without a prior conditioning lesion, whereas axon growth can be observed from precrushed retinal explants (3). At 6 days in vitro there is still considerable growth of RGC axons on the culture substratum in crushed retinae but only a few single axons have extended processes from noncrushed explants (Table 1). Table 1 shows that without a conditioning lesion no axons extend from retinal explants at 3 days and that at 6 and 9 days the number of RGC axons is always significantly lower than in precrushed retinae. Nevertheless, RGC axons extend from both precrushed or noncrushed retinal explants but the total number and length of axons extending from explants at corresponding time points is always significantly lower in noncrushed retinae than in crush-activated retinae. Retinal Glia Migration
To determine which retinal cell type is able to migrate or proliferate in normal retinae and retinae with a prior conditioning lesion to the optic nerve, cells that had migrated out from retinal explants were characterized by immunohistochemistry. In retina cultures with preactivated glia, cell migration was significantly better than in normal retinae (Table 2). Mainly VIM-positive Miiller cells and GFAP-positive astrocytes were found to proliferate and migrate out from retinal explants (Table 3 and Fig. 1). Only a few Miiller cells acquire GFAP in these cultures as determined by double labeling for VIM and GFAP (about 10% of the VIM-positive glia popula-
TABLE Migration
of Glia Precrushed
Morphometric Analysis To evaluate retinal glia migration, cultures were scored at 3,6, and 9 days after explantation. The extent of glia migration was determined by measuring the radial distance between the edges of retinal explants and the furthest front of cell migration (cellular halo). The same analysis was performed to determine retinal ganglion cell axon outgrowth from retinal explants (see also (4,5)). In cultures that had been processed for immunohistochemistry, the proportion of GFAP, Vimentin, or Thy 1.1 cells was determined by counting at least 500 cells in standard areas from more than two different explants after different time intervals in uitro.
and Proliferation
Days
in vitro
Glia
3 6 9
87.5 100 100
from
2 Retinal
Explants
retina Distance 70 k 32 237 + 70* 1491 f 532*
Normal Glia 25 75 100
retina Distance 30* 17 99 2 74* 716 + 424*
Note. The table shows the percentage of retinal explants (n = 80) with glia migration and the distance (values are means + standard deviation) of the outward rim of the glia halo from the explant’s edges at 3, 6, and 9 days in vitro (div). Glia migration starts earlier from precrushed explants compared to normal retinae with 87.5% retinal explants showing glia migration compared to only 25% in normal retinae at 3 div. At 6 and 9 div the mean distances of retinal glia to the explants are significantly higher in precrushed retinae than in normal retinae. *Shows significant differences, t test, P < 0.01.
68
M.
TABLE Immunohistochemical Migratory Precrushed Days
in 6 3
vitro
GFAP 11.5 11.3
VIM 60.7 45 ’
BAHR
3
Characterization Retinal Cells retina Thy 20.1 35
of
Normal 1.1
GFAP 7.8 5.5
VIM 43 35.5
retina Thy
1.1
Only noncrushed normal retinae were cultured on the retinal glia monolayer. Usually (see results Table 1) no axons extend from noncrushed retinae after 3 days in vitro on a PL-Lam substratum. In retinal glia-retina cocultures however, many retinal ganglion cell axons extend from normal retinae at 3 days in vitro on acti-
40.4 50.7
Note. The table shows the percentage of cells immunopositive for GFAP, VIM, or Thy 1.1 that had migrated out from retinal explants (at least 200 cells from different standard areas in two different experiments were counted in each case) at 6 or 3 days in uitro (div). At 6 div a higher percentage of retinal glia characterized by GFAP and VIM is present in cultures of precrushed retinae (about 70%) compared to normal retina cultures (less than 60%). At 3 div the relative number of glia decreases under both conditions (about 55% in precrushed and 40% in normal retina cultures) and the number of Thy l.l-positive cells increases from 20 to 35% in precrushed retinae and from 40 to 50% in normal retinae.
tion also shows GFAP staining at 9 days in uitro). After 6 days in vitro only fibroblasts seem to proliferate (Table 3). At very early stages in vitro a population of small rounded cells can be observed near the explants edges. These could be macrophages removing cellular debris. In normal retina explants significantly fewer explants show glial cell migration compared to “activated” retinae (Table 2). Immunohistochemical staining shows that besides VIM-positive glia a high percentage of Thy l.l-positive fibroblasts proliferate and migrate out from retinal explants in these cultures (Table 2 and Fig. 2). When crush-activated retinal glia cultures were incubated with BrdU for 12 h about 39% of the cells that had migrated out from the explants incorporated BrdU. Double-labeling experiments showed that about 10% of the BrdU-labeled cells were positive for GFAP and 27% showed VIM staining (Fig. 1). Therefore only about 40% of the cells present after 6 days in vitro seem to be of glial origin. The immunohistochemically defined glial cell population seems to consist of 25% astrocytes and about 75% Mtiller cells (see Table 3). In cultures from normal retinae only 12% of the BrdU-incorporating cells were positive for either GFAP or VIM. Therefore, proliferating cells from normal retinae seem to consist mainly of Thy l.l-positive flat cells. We did not see substantial accumulation of GFAP in VIM-positive Mtiller cells (less than 10%) during the time of observation which is in agreement with earlier reports on Miiller glia in vitro (9). Axon Regeneration from Normal Retinae on Retinal Glia The ability of retinal glia from normal and activated retinae to promote retinal ganglion cell axon growth was determined in retinal glia-adult retina cocultures.
FIG. 1. Double labeling of retinal glia with antibodies recognizing Vimentin and BrdU. This figure shows a population of activated retinal glial cells that have migrated out from retinal explants at 6 days in vitro (phase-contrast in a). About 60% of the cell population is positive for Vimentin at that time point (b). When retinal tissue cultures were incubated with BrdU for 12 h, most of the VIM-positive cells also incorporated BrdU (thin arrow in c), but VIM-negative flat cells (thick arrow in c), presumably fibroblasts, are also labeled with BrdU. Bar, 50 pm.
ADULT
RAT
RETINAL
GLIA
69
ina-retinal glia cocultures were processed for neurofilament and GFAP immunohistochemistry no special attraction to or avoidance of retinal astrocytes by regenerating ganglion cell axons could be observed (Fig. 3). However, when retinal ganglion cell axons reach the out-
FIG. 2. Double labeling of retinal glia with antibodies recognizing GFAP and Thy 1.1. A population of retinal glial cells derived from normal retinal explants is shown at 6 days in vitro (phase-contrast in a). Only a few cells can be labeled with GFAP in these cultures (b) and the majority of the cell population consists of Thy 1.1.positive flat cells (c), presumably fibroblasts. Bar, 50 pm.
vated glia (see Figs. 3 and 4). Significantly fewer axons emanate from retinal explants cultured on retinal glia derived from normal explants (9, 13) than from explants on activated glia (16, 40) (t test, P 5 0.01)). RGC axons visualized by anti-neurofilament antibodies and examined in scanning EM grow on Miiller cells, astrocytes, and fibroblasts as determined by scanning EM and immunohistochemistry (Figs. 3 and 4). When ret-
FIG. 3. Retinal ganglion cell axon growth on retinal glia monolayers. A coculture experiment of normal rat retinal explants on activated retinal glia is shown (phase-contrast in a). Retinal glia that had migrated out from retinal explants with prior optic nerve crush (“activated glia”) was used as a substratum for retinal explants. Only a few GFAP-positive astrocytes can be detected in these cultures at 9 days in vitro (arrow in b). RGC axons are visualized by neurofilament antibodies (c). The arrowhead in c shows the identical location as the arrow in (b), indicating that astrocytes are permissive for RGC axons in vitro, although no special preference of RGC axons can be seen in these experiments. Bar, 50 pm.
70
M.
FIG. 4. Retinal ganglion cell axons grow on adult retinal glia. Many neurites visualized by anti-neurofilament antibodies regrow from normal retinal explants on retinal glia after 3 days in uitro (a). In b a corresponding area is shown in scanning EM (X7840 enlargement). The arrowheads indicate RGC axons that grow over surfaces of retinal glia (indicated as G). The axons do not extend onto the acellular PL-Lam substrate. Bar, 50 pm.
ward rim of the glia monolayer, they show a clear preference for the cellular substratum compared to the PLLam-coated surface of the petri dishes (Fig. 5). Scanning EM also shows that retinal ganglion cell axons prefer to grow on cell surfaces compared to the acellular PL-Lam culture substrate (Fig. 4). DISCUSSION
Within the retina of most mammals two different types of glial cells, Miiller cells and astrocytes (for re-
B;l;HR
view see (33)), can be distinguished. Miiller cells are radially oriented cells that span the whole retinal width with their processes (33). They usually do not express GFAP but rather express VIM (9,lO). Astrocytes found in the retina resemble those found in other brain regions (19) with a preference for a location near blood vessels. Several recent experiments indicated that retinal astrocytes might be immigrants from the optic nerve in the therefore correspond to rat (19, 45). Retinal astrocytes type I astrocytes found in the optic nerve (37) since type II astrocytes and oligodendrocytes are kept out of the rat retina (19). After lesioning the retina or the optic nerve, retinal astrocytes increase their GFAP-staining intensities (10). This has been shown to correlate with “reactive” astrogliosis observed in other brain regions after penetrating lesions (34, 38, 40). Mtiller cells as well as astrocytes (reactive astrogliosis) have been shown to accumulate GFAP after optic nerve lesions (10, 22, 31), but the number of retinal Miiller cells that are transformed seems to be rather low (9, 39). The aim of the present paper was to investigate the reaction of retinal astrocytes and Miiller cells after lesioning the optic nerve in situ. We have used a tissue culture model to study glia populations and fibroblasts from adult rat retinae that proliferate and migrate out after retina explantation on acellular substrates. Both glia populations from adult retinae were able to survive in tissue culture. Miiller cells and retinal astrocytes migrated out from adult rat retinal explants in vitro. With an earlier onset the extent of glia migration was significantly higher in retinal explants whose corresponding optic nerves were crush-axotomized in situ. Retinal Miilier cells that were VIM-positive and GFAPnegative did not substantially accumulate GFAP during the time of observation, which is consistent with earlier findings (9,39). In normal adult rat retinal explants, the cell population that migrates and proliferates after explantation consists to a high degree of Thy l.l-positive flat cells, presumably fibroblasts. Therefore our results suggest that RGC axotomy and/or degeneration influences the ability of adult rat retinal glia to survive, migrate, and proliferate in tissue culture. This is consistent with findings in adult rat cortical astrocytes that showed increased survival in vitro after degeneration of surrounding neuronal fiber tracts in uiuo (25). Therefore some signal of degenerating or regenerating CNS neurons seems to initiate proliferation or induce a transformation of adult rat glia into a more “vigorous” state (8, 21). This transformation might be reflected in increased GFAP expression (reactive astrocytes) and in changes in the composition of cell surface molecules (8). Nevertheless, glia in the adult mammal seem to have only a restricted potential to proliferate (25,34) because at 9 days in vitro the relative number of glia decreases in both glia preparations from crush-activated and normal
ADULT
RAT
RETINAL
GLIA
71
FIG. 5. Retinal ganglion cell axons respect cellular-acellular substrate borders. The outward rim (arrowheads in a) of retinal glial cells that had migrated out from adult rat retinal explants is shown (a, phase-contrast). When retinal explants from normal rats were cultured on these glia monolayers, regenerating RGC axons (visualized by neurofilament antibodies in b) respected the border to the acellular PL-Lam substrate. Bar, 50 pm.
retinae. There are controversial reports about glia proliferation in response to penetrating lesions in adult mammals. Glia identified with S-100 cell surface markers did not proliferate substantially as determined by [3H]thymidine incorporation (32). Other studies have shown an increased mitogenic activity for astrocytes accumulating around lesioned brain areas, but in general adult astrocytes seem to have a restricted potential to proliferate (25, 34). The present results of adult rat retinal glia in vitro show that glia from normal rats did not show substantial migration and proliferation after ex-
plantation. Glia derived from retinae whose optic nerves received a conditioning lesion in viuo were able to migrate over larger distances and to proliferate in vitro as determined by BrdU incorporation. This suggested that retinal glia might be able to support neurite growth from CNS neurons. To test this hypothesis coculture experiments of adult rat retina and retinal glia were performed. In this experimental model, retinal glia that had migrated out from retinal explants were tested for their ability to promote RGC axon regeneration. Activated glia that had migrated out from
72
M.
adult rat retinal explants were able to initiate RGC axon growth from normal retinal explants. This suggests that retinal glia can support RGC axon growth after optic nerve lesions (8) and a signal from retinal glia (that is not yet characterized) might potentiate other signals acting to allow RGC survival and initiate RGC axon outgrowth. This supports the findings of Lindsay (24) who demonstrated that glia derived from adult rat brain after kainate lesions were able to support neuronal survival and axon growth. This suggests that different “functional states” of adult astrocytes might be postulated: Penetrating CNS lesions seem to induce astroglial hypertrophy (10, 38) and this reactive astrogliosis seems to impede axon regeneration (26). Selective neuronal degeneration induced either by axotomy or by kainate injections (24) might lead to glial changes that are compatible with axon growth. The present results show that retinal ganglion cell axons which reextend from explants preferentially grow on surfaces of adult rat retinal glia. No special preference for either astrocytes or MUller cells could be observed in these experiments. Interestingly, RGC axons respect borders to the acellular substratum PL-Lam. This suggests that regenerating RGC axons were able to make a choice between cellular and acellular substrata. Whether regenerating RGC axons prefer glial cell surfaces or extracellular matrix molecules other than laminin (15) used to coat the culture dishes is not yet clear (7). We found that cellular substrates in general were preferred by those axons. For example RGC axons regenerating on nonactivated glia also grew on nonglial cells (data not shown) that consisted to a high degree of Thy l.l-positive flat cells. In these experiments no special preference for either glial or nonglial cells was observed. This seems to be different from embryonic CNS axons which show a clear preference for glial compared to nonglial cells (17, 35) in uitro. In conclusion, our experiments show that adult rat retinal glia is activated after lesioning the optic nerve in situ (8). This conditioning lesion activates Miiller cells and astrocytes which proliferate and migrate out from retinal explants in uitro. In addition adult rat retinal glia is able to initiate axon regeneration from normal retinal explants. Both, GFAP-positive adult astrocytes and VIM-positive Miiller cells proliferated and were permissive for RGC axon growth from retinal explants. This result is consistent with the finding that RGC axons (6, 18) and other CNS axons (18) extend axons on cortical astrocytes from embryonic or early postnatal rats in uitro. Our results, however, show that even adult activated glia might support axon regeneration of adult CNS neurons. The present results are also consistent with the hypothesis that the oligodendrocyte-type 2 astrocyte lineage (36, 41) might be responsible for inhibitory aspects of mature CNS glia since no glia of that cell lineage was present in our culture system.
BAHR
Future experiments are assigned to show which cell surface or extracellular matrix molecules promote RGC axon growth on adult glia and to determine whether activated adult rat retinal glia might also enhance RGC survival in vitro. Both enhanced survival and promotion of axon regeneration are essential prerequisites for effective CNS regeneration, The molecular mechanisms that determine the ability of different glia populations to promote or inhibit adult CNS regeneration are not understood yet. The present results suggest that glial subpopulations exist in adult rats and that some of these cells have preserved an ability to support regeneration in the adult mammalian CNS. ACKNOWLEDGMENTS The author thanks Professor J. Dichgans and Professor F. Bonhoeffer for supporting his work; Drs. B. Schlosshauer, T. Alsopp, and S. Thanos for critical comments on the manuscript; and Jiirgen Berger for technical assistance at the electron microscope.
REFERENCES 1.
ARD, M. D., AND R. P. BUNGE. 1988. Heparan sulfate proteoglycan and laminin immunoreactivity on cultured astrocytes: Relationship to differentiation and neurite growth. J. Neurosci. 8:
2844-2858. 2. ARMSON, P. F., M. R. BENNE?T,
AND T. R. RATU. 1987. Retinal ganglion cell survival and neurite regeneration: The change from Mtiller cell dependence to superior colliculi dependence during development. Deu. Brain Res. 32: 207-216.
3. BXHR, M., J. VANSELOW, ation Bruin
of adult rat ganglion Res. 73: 393-401.
AND S. THANOS. 1988. In vitro regenercell axons from retinal explants. Exp.
4. B~~HR, M., J. VANSELOW,
AND S. THANOS. rat ganglion cells to regrow axons in vitro fibroblast growth factor and gangliosides. 197-201.
5. BKHR, M., AND R. P. BUNGE.
6.
7. 8. 9.
10.
11.
1989. Ability of adult can be influenced by Neurosci. Lett. 96:
1989. Functional status influences the ability of Schwann cells to support adult rat retinal ganglion cell survival and axonal growth. Exp. Neural. 106: 27-40. BKHR, M., M. SCHACHNER, AND R. P. BUNGE. 1989. In vitro regeneration of adult rat retinal ganglion cell axons on astrocytes: Growth characteristics and molecular mechanisms. Sot. Neurosci. Abstr. 19: 331. BOHR, M., J. M. HOPKINS, AND R. P. BUNGE. 1989. Schwann cell promotion of adult rat retinal ganglion cell regeneration in vitro. Eur. J. Neurosci. (Suppl.) 2: 302. B~~HR, M., AND B. SCHLOSSHALJER. 1989. Jones ganglioside expression on retinal glia increases after transection of adult rat retinal ganglion cell axons. J. Neurocytol. 18: 553-563. BARTLETT, P. F., M. D. NOBLE, R. M. PRUSS, M. C. RAFF, S. RATTRAY, AND C. A. WILLIAMS. 1981. Rat neural antigen-2 (RAN-2): A cell surface antigen on astrocytes, ependymal cells, Miiller cells and leptomeninges defined by monoclonal antibody. Bruin Res. 204: 339-351. BIGNAMI, A., AND D. DAHL. 1979. The radial glia of Miiller in the rat retina and their response to injury: An immunofluorescence study with antibodies to glial fibrillary acidic (GFA) protein. Exp. Eye Res. 28: 63-69. BRAY, G. M., M. P. VILLEGAZ-PEREZ, M. VIDAL-SANZ, AND A. J. AGUAYO. 1987. The use of peripheral nerve grafts to enhance
ADULT neuronal survival, promote growth and permit nections in the central nervous system of adult 132: 5-19.
RAT
RETINAL
Astrocytes (S. Fedoroff demic Press, New York.
terminal reconrats. J. Erp. Biol.
12.
Biissow, H. 1980. The astrocytes in the retina head of mammals: A special glia for the ganglion Tissue Res. 206: 367-378.
13.
CARONI, P., AND M. E. SCHWAB. 1988. Two membrane protein fractions from rat central myelin with inhibitory properties for neurite growth and fibroblast spreading. J. CellBiol. 106: 12811288.
73
GLIA and A. Vernadakis,
Eds.),
Vol.
3. Aca-
28.
MCCAFFERY, C. A., T. R. RAJU, AND M. R. BENNET. 1984. Effects of cultured astroglia on the survival of neonatal rat retinal ganglion cells in vitro. Deu. Biol. 104: 441-448.
29.
MCDONNEL, cell axons
and optic nerve cell axons. Cell
P., AND M. BERRY. 1982. Regeneration in the mouse retina. Brain Res. 241:
of ganglion
362-365.
30.
MCLOON, S., L. K. MCLOON, S. L. PALM, AND L. T. FURCHT. 1988. Transient expression of laminin in the optic nerve of the developing rat. J. Neurosci. 8: 1981-1990.
14.
CARONI, P., AND M. E. SCHWAB. 1988. Antibody against melin associated inhibitor of neurite growth neutralizes nonpermissive substrate properties of CNS white matter. Neuron 1: 85-96.
31.
MILLER, M. N., AND M. OBERDORFER. 1981. Neuronal and neuroglial responses following retinal lesions in neonatal rats. J. Comp. Neural. 202: 493-504.
15.
DAVID, S. 1988. Neurite outgrowth from mammalian CNS neurons on astrocytes in vitro may not be mediated primarily by laminin. J. Neurocytol. 17: 131-144. ENG, L. F., P. J. REIER, AND J. D. HOULE. 1987. Astrocyte activation and fibrous gliosis: Glial fibrillary acidic protein immunostaining of astrocytes following intraspinal cord grafting of fetal CNS tissue. In Progress in Bruin Research (F. J. Seil, E. Herbert, and B. M. Carlson, Eds.1, Vol. 71. Elsevier Science Pub., Amsterdam/New York. FALLON, J. R. 1985. Preferential outgrowth of central nervous system neurites on astrocytes and Schwann cells as compared to nonglial cells in vitro. J. Cell Biol. 100: 198-207.
32.
MIYAKE, T., T. HATTORI, M. FUKUDA, AND T. KITAMURA. 1989. Reactions of S-100 positive glia after injury of mouse cerebral cortex. Brain Res. 489: 31-40. NEWMAN, E. A. 1986. The Miiller cell. In Astrocytes (S. Fedoroff and A. Vernadakis, Eds.), Vol. 1. Academic Press, New York.
16.
17.
18.
FAWCETT, J. W., J. ROKOS, axons in three-dimensional
AND I. BAKST. 1989. The growth of astrocyte cultures. Deu. Biol. 135:
33. 34.
NIETO-SAMPEDRO, M., P. R. SANETO, J. DE VELLIS, AND C. W. COTMAN. 1985. The control of glial populations in brain: Changes in astrocyte mitogenic and morphogenic factors in response to injury. Bruin Res. 343: 320-328.
35.
NOBLE, M., J. FOK-SEANG, AND J. COHEN. 1984. Glia are a unique substrate for the in vitro growth of central nervous system neurons. J. Neurosci. 4: 1892-1903.
36.
RAJU, T. R., AND M. R. BENNET. 1986. Retinal ganglion cell survival requirements: A major but transient dependence on Miiller glia during development. Brain Res. 383: 165-176.
37.
RAFF, M. C. 1989. Glial cell diversification Science 243: 1450-1455.
38.
REIER, P. J., L. J. STENSAAS, AND L. GUTH. 1983. The astrocytic scar as an impediment to regeneration in the central nervous system. In Spinal Cord Reconstruction (C. C. Kao, R. P. Bunge, and P. J. Reier, Eds.). Raven Press, New York. SCHERER, J., AND J. SCHNITZER. 1989. The rabbit retina: A suitable mammalian tissue for obtaining astroglia-free Miiller cell cultures. Neurosci. Lett. 97: 51-56.
449-458. 19.
20.
FFRENCH-CONSTANT, C., R. H. MILLER, J. F. BURNE, AND M. C. RAFF. 1988. Evidence that migratory oligodendrocyte-type 2 astrocyte (O-2A) progenitor cells are kept out of the rat retina by a barrier at the eye-end of the optic nerve. J. Neurocytol. 17: 1325. GOLDBERG, S., AND B. FRANK. 1980. Will central nervous systems in the adult mammal regenerate after bypassing a lesion? A study in the mouse and chick visual systems. Exp. Neural. 70: 675-689.
21.
JANECKO, K. 1989. Spatiotemporal patterns liferation in the rat brain injured at the development: A combined immunohistochemical graphic study. Brain Res. 485: 236-243.
22.
KARSCHING, A., H. WKSSLE, AND J. SCHNITZER. 1986. Immunohistochemical studies on astroglia of the cat retina under normal and pathological conditions. J. Comp. Neural. 249: 564-576.
23.
LIESI, P., D. DAHL, AND A. VAHERI. 1983. Laminin is produced by early astrocytes in primary culture. J. Cell Biol. 96: 920-924. LINDSAY, R. M. 1979. Adult rat brain astrocytes support survival of both NGF-dependent and NGF-insensitive neurones. Nature (London) 282: 80-82.
24.
25.
of the astroglial propostmitotic stage of and autoradio-
LINDSAY,R. M., P.C.BARBER, M.R.C. SHERWOOD,J.ZIMMER, AND G. RAISMAN. 1982. Astrocyte cultures from adult rat brain: Derivation, characterization and neuronotrophic properties of pure astroglial cells from corpus callosum. Brain Res. 243: 329343.
26.
LIUZZI, F. J., AND R. J. LASEK. 1987. Astrocytes block regeneration in mammals by activating a physiological pathway. Science 237: 642-645.
27.
MANTHORPE, M., J. S. RUDGE, AND S. VARON. 1986. Astroglial cell contributions to neuronal survival and neuritic growth. In
39.
in the rat optic
nerve.
40.
SCHIFFER, D., M. T. GIORDANA, A. MIGHELI, G. GIACCONE, S. PEZZOTTA, AND A. MAURO. 1986. Glial fibrillary acidic protein and Vimentin in the experimental glial reaction of rat brain. Brain Res. 374: 110-118.
41.
SMITH, G. S., R. H. MILLER, AND J. SILVER. 1986. Changing role of forebrain astrocytes during development, regenerative failure and induced regeneration upon transplantation. J. Camp. Neural. 251:23-43.
42.
SCHWAB, M. E., AND P. CARONI. 1988. Oligodendrocytes and CNS myelin are non-permissive substrates for neurite growth and fibroblast spreading in vitro. J. Neurosci. 8: 2381-2393. TOMASELLI, K. J., K. M. NEUGEBAUER, J. L. BIXBY, J. LILLIEN, AND L. F. REICHARDT. 1988. N-Cadherin and integins: Two receptor systems that mediate neuronal process outgrowth on astrocyte surfaces. Neuron 1: 33-43.
43.
44.
axonal stop 45.
THANOS, S., M. BKHR, Y.-A. BARDE, AND J. VANSELOW. 1988. Survival and axonal elongation of adult rat retinal ganglion cells: In vitro effects of lesioned sciatic nerve and brain derived neurotrophic factors. Eur. J. Neurosci. 1: 19-26. WATANABE, T., AND M. C. RAFF. 1988. Retinal astrocytes are immigrants from the optic nerve. Nature (London) 332: 834837.
NEUROLOGY
111,65-73
(1991)
Adult Rat Retinal Glia in Vitro: Effects of in Viva Crush-Activation on Glia Proliferation and Permissiveness for Regenerating Retinal Ganglion Cell Axons M. B#HR Department
of Neurology,
University Spemannstr,35/Z,
Tiibingen, and Max-Planck 7400 Tiibingen, Federal
Press.
fiir Entwicklungsbiologie of Germany
Tiibingen,
axon-glia interactions after central nervous system (CNS) lesions and during regeneration have given new insights into the functions of adult CNS glia. Glia involved in axon ensheathment or myelination, like type II astrocytes and oligodendrocytes (for review see (37)), seem to be rather poor substrata for axonal growth (19, 42) and components of adult CNS myelin have been shown to constitute a nonpermissive environment for regenerating neurites (13, 14). Type I astrocytes have been shown to react by hypertrophy and formation of a glial scar after brain lesions (16,38), thus blocking axon regeneration (26). This aspect, however, seems to be a function of maturity since “immature” astrocytes have been shown to promote axon regeneration in the adult CNS (41). In contrast, glia contained in peripheral nerves used as grafts to reconnect lesioned CNS tracts have been shown to support axonal growth of adult CNS neurons (for review see (11)). We have recently demonstrated that isolated purified Schwann cells support axon regeneration from adult rat retinal explants in vitro (5, 7). An exudate derived from regenerating sciatic nerves that is presumably produced by Schwann cells contains a powerful neurotrophic activity for adult rat retinal ganglion cells (RGC) (44). These results suggest that the inability of CNS neurons to survive axotomy and to regenerate an axon may not be due to a cell-intrinsic defect, but rather may be determined by influences of the surrounding glial environment. Support of this hypothesis was found in recent experiments of adult rat retinal explants cultured on cortical astrocytes where no enhanced survival or neurite outgrowth of retinal ganglion cells was observed (6). Only “early” astrocytes, resembling immature astrocytes in uiuo, which still express laminin (23) were able to support axon growth from adult rat retinal explants (6). This suggests that during the time when CNS fiber tracts are established, astrocytes might be involved in promoting axon growth (43) and guiding axons to their
The effects of optic nerve crush on adult rat retinal glia activation were studied in vitro. In adult rats the optic nerves were crushed and the corresponding retinae were explanted 5 to 7 days later and cultured in vitro. The glial response of retinae with precrushed optic nerves was compared to the glial response of retinae without prior optic nerve crush. As a consequence of crush-axotomy more glial cells migrated out from retinal explants and covered significantly larger areas of the substratum than glia from noncrushed retinae. Migration of immunohistochemically distinguishable Vimentin-positive Mtiller cells and glial fibrillary acidic protein-positive astrocytes could be observed in both types of cultures. Astrocytes as well as Muller cells incorporated bromodeoxyuridine after explantation. In noncrushed retinal explants Thy l.l-immunopositive flat cells were much more frequent and the relative proportion of glial cells was much lower than in crush-activated cultures. In a second set of experiments the ability of adult rat retinal glia to support retinal ganglion cell regeneration was examined. Normal retinal explants (without optic nerve crush) which usually do not substantially regenerate axons were cultured on retinal glia from normal and crush-activated explants. Both glia preparations supported axon growth from retinal explants after 3 days in vitro. Neuritic growth was significantly better when retinal explants from normal adult rats were cultured on crush-activated retinal glia as compared to glia derived from noncrushed retinae. It is concluded that activated adult rat retinal glia, unlike adult glia found in other brain regions, support adult rat retinal ganglion cell regeneration in vitro. 0 1991 Academic
Znstitut Republic
Inc.
INTRODUCTION
During recent years several new functions have been added to the traditional role of glia, such as providing trophic support for neurons and promoting axonal growth (1, 6, 15, 17, 27-29, 41, 43). In particular the 65
All
Copyright Q 1991 rights of reproduction
0014.4886/91 $3.00 by Academic Press, Inc. in any form reserved.
66
M. B.&HR
target regions (30). This ability of astrocytes to support neuritic growth seems to be lost in the normal adult mammalian CNS. Cortical astrocytes from adult rat brain, however, that are activated in situ by neuronal degeneration (24) before cultivation seem to be able to support neuronal survival and axon growth. It has been shown that during retinal development Miiller cells, which are a kind of specialized glia for RGC (12), support retinal ganglion cell survival (2, 36). This dependence of retinal ganglion cells on Miiller cells in vitro (36) is lost at a time corresponding to the in uiuo situation when the adult retinotectal projection has been established. Unlike other parts of the CNS, however, retinal ganglion cells seem to be able to regenerate axons within the retina even in adulthood (20,28). This suggests that retinal glia, which in the rat does not contain type II astrocytes or oligodendrocytes (19, 45), might be able to support regeneration of adult CNS axons. We have recently demonstrated that retinal glia transiently increase their cell surface ganglioside expression after conditioning lesions applied to the optic nerve in uiuo (8). Gangliosides recognized by the monoclonal antibody “JONES” were reexpressed at levels found during embryonic development of the retinotectal projection (8). Optic nerve glia did not show a similar behavior, suggesting that retinal glia might have preserved its ability to support adult rat retinal ganglion cell regeneration by mediating the effects of neurite promoting or neurotrophic factors (8). The aim of the present study was to examine the effects of in uiuo crush activation on adult rat retinal glia and to examine whether adult rat retinal glia are able to proliferate in response to lesions. In addition, the ability of adult rat retinal glia to support neuritic growth from adult rat retinal explants in uitro was investigated. MATERIAL
AND
METHODS
Surgical Procedures Twenty adult Lewis rats were used in this study. To apply a conditioning lesion to the optic nerve 10 rats were deeply anesthetized with ip chloral hydrate (0.42 mglkg body weight). The optic nerve was crush-axotomized intraorbitally (“conditioning” lesion) at a distance of about 3 mm from the optic disc by means of fine watchmaker forceps (3-5). The efficacy of the optic nerve crush was determined by complete loss of the pupillary reaction on the contralateral eye. Care was taken to not alter retinal blood supply and in each crushed eye the retinal blood vessels were inspected through the lens. Only animals that showed a normal retinal blood circulation were used in this study. Explantation Techniques Five to seven days after optic nerve crush the corresponding retinae were dissected as described recently
(3). Briefly, the retinae were freed from connective and vitreal tissue and whole-mounted on semipermeable membrane filters (Millipore) in Hanks’ buffered salt solution with the ganglion cell layer upward. Flat mounted retinae were cut into eight triangular pieces with their tips centered in the former optic disc. Retinal explants were transferred to polylysine (PL; Sigma, 1 mg/ml in Hanks’)- and laminin (Lam; BRL, 10 pg/ml in Hanks’)coated petri dishes (Petriperm, Hereaus) and cultured with the ganglion cell layer attached to cellular or acellular substrates. Cultures were fed with 3 ml Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal calf serum every 3 days and incubated with 5% CO, at 37°C. Retinal Glia CultiuationlCocultiuation and Adult Retinal Explants
of Retinal Glia
Twenty retinal explants with (n = 10) or without (n = 10) conditioning lesions prior to explantation were scored every 3 days for retinal ganglion cell axon growth and glialflat cell migration. Five to eight days after explantation the retinal pieces were removed from the dishes, resulting in cultures of single cells that had migrated out from the explants. Two days later precrushed or noncrushed retinal explants derived from other retinae were placed on the retinal glia monolayers and cultured for 3-5 days (n = 20). Retinal glia or glia-retina cocultures were fixed in 4% paraformaldehyde in phosphate buffer at pH 7.4 (PFA) and processed for immunohistochemistry. Immunohistochemistry Retinal glia cultures (n = 80) or glia-retina (n = 20) cocultures were fixed by overnight incubation in PFA after different time points in culture. Then cultures were rinsed in DMEM and incubated with anti-Thy 1.1 monoclonal antibody (Serotec, 100 pg/ml in DMEM), anti-Vimentin (Boehringer and Sigma, 100 pglml in DMEM), anti-glial fibrillary acidic protein (GFAP; Dakopatts, 2 g/ml in DMEM), or anti-neurofilament antibody (SMI 31; Sternberger & Meyer Immunocytochemicals, 2 @g/ml in DMEM). For GFAP, Vimentin, and SMI 31 immunohistochemistry, specimens were permeabilized with methanol at -20°C for 10 min. Cultures were then rinsed again in DMEM until the pH had normalized; incubated for 45 min with anti-GFAP, anti-Vimentin, or anti-neurofilament antibodies; and rinsed and stained by fluorescein- or rhodamine-conjugated secondary antibodies (Cappel). Specimens were viewed with epifluorescence on an inverted microscope (Zeiss) and photographed on Kodak Tri-X-Pan or Ektachrome films. Retinal Glia Proliferation To determine retinal glia proliferation, bromodeoxyuridine (BrdU; Bioscience Products, 10 pmol) was ap-
ADULT
TABLE Retinal Ganglion Polylysine-Laminin Precrushed Days
in vitro
Number
3 6 9
10f 6 31 k 10 52 k 23
Cell
RAT
RETINAL
67
GLIA
1
RESULTS
Axon Growth Substrates
retina Length 140 + 53 208 f 63 1112 f 388
on
Normal Number 0 12 k 5 11 t6
Retinal Ganglion Cell Axon Growth on PolylysineLaminin Substrates retina Length 0 74k 13 375 k 223
Note. The table shows the total number and the length (mean diameter of the neuritic halo; values are means + standard deviation) of RGC neurites that have reextended from precrushed and normal adult rat retinal explants (n = 80). At all time points analyzed significantly more RGC axons extend from precrushed retinae compared to normal retinae (t test, P G 0.01).
plied to retinal tissue cultures at 6 days after explantation (n = 40). BrdU is incorporated into the DNA of dividing cells and can be detected by specific antibodies. Retinal cultures were incubated with BrdU for 12 h, fixed in PFA, and then incubated with anti-BrdU antibodies. To visualize anti-BrdU antibodies, specimens were incubated with fluorescein- or rhodamine-conjugated secondary antibodies (Cappel). For double-labeling immunohistochemistry of BrdU and GFAP, Vimentin (VIM) or Thy 1.1 standard procedures were used as described above. Electron Microscopic Studies For scanning electron microscopic examination of retinal ganglion cell axons extending on adult rat retinal glia, cultures were fixed in 2% glutaraldehyde in PBS and osmificated in 1% 0~0, in PBS. After dehydration in graded ethanols specimens were dried in CO, by the critical point method and finally coated with gold. After that areas near the explant’s tips were viewed in an autoscan scanning electron microscope with an angle of 45”.
The extent of RGC axon growth was determined at 3, 6, and 9 days after explantation. After 3 days (see Table 1) no axons extend from retinae without a prior conditioning lesion, whereas axon growth can be observed from precrushed retinal explants (3). At 6 days in vitro there is still considerable growth of RGC axons on the culture substratum in crushed retinae but only a few single axons have extended processes from noncrushed explants (Table 1). Table 1 shows that without a conditioning lesion no axons extend from retinal explants at 3 days and that at 6 and 9 days the number of RGC axons is always significantly lower than in precrushed retinae. Nevertheless, RGC axons extend from both precrushed or noncrushed retinal explants but the total number and length of axons extending from explants at corresponding time points is always significantly lower in noncrushed retinae than in crush-activated retinae. Retinal Glia Migration
To determine which retinal cell type is able to migrate or proliferate in normal retinae and retinae with a prior conditioning lesion to the optic nerve, cells that had migrated out from retinal explants were characterized by immunohistochemistry. In retina cultures with preactivated glia, cell migration was significantly better than in normal retinae (Table 2). Mainly VIM-positive Miiller cells and GFAP-positive astrocytes were found to proliferate and migrate out from retinal explants (Table 3 and Fig. 1). Only a few Miiller cells acquire GFAP in these cultures as determined by double labeling for VIM and GFAP (about 10% of the VIM-positive glia popula-
TABLE Migration
of Glia Precrushed
Morphometric Analysis To evaluate retinal glia migration, cultures were scored at 3,6, and 9 days after explantation. The extent of glia migration was determined by measuring the radial distance between the edges of retinal explants and the furthest front of cell migration (cellular halo). The same analysis was performed to determine retinal ganglion cell axon outgrowth from retinal explants (see also (4,5)). In cultures that had been processed for immunohistochemistry, the proportion of GFAP, Vimentin, or Thy 1.1 cells was determined by counting at least 500 cells in standard areas from more than two different explants after different time intervals in uitro.
and Proliferation
Days
in vitro
Glia
3 6 9
87.5 100 100
from
2 Retinal
Explants
retina Distance 70 k 32 237 + 70* 1491 f 532*
Normal Glia 25 75 100
retina Distance 30* 17 99 2 74* 716 + 424*
Note. The table shows the percentage of retinal explants (n = 80) with glia migration and the distance (values are means + standard deviation) of the outward rim of the glia halo from the explant’s edges at 3, 6, and 9 days in vitro (div). Glia migration starts earlier from precrushed explants compared to normal retinae with 87.5% retinal explants showing glia migration compared to only 25% in normal retinae at 3 div. At 6 and 9 div the mean distances of retinal glia to the explants are significantly higher in precrushed retinae than in normal retinae. *Shows significant differences, t test, P < 0.01.
68
M.
TABLE Immunohistochemical Migratory Precrushed Days
in 6 3
vitro
GFAP 11.5 11.3
VIM 60.7 45 ’
BAHR
3
Characterization Retinal Cells retina Thy 20.1 35
of
Normal 1.1
GFAP 7.8 5.5
VIM 43 35.5
retina Thy
1.1
Only noncrushed normal retinae were cultured on the retinal glia monolayer. Usually (see results Table 1) no axons extend from noncrushed retinae after 3 days in vitro on a PL-Lam substratum. In retinal glia-retina cocultures however, many retinal ganglion cell axons extend from normal retinae at 3 days in vitro on acti-
40.4 50.7
Note. The table shows the percentage of cells immunopositive for GFAP, VIM, or Thy 1.1 that had migrated out from retinal explants (at least 200 cells from different standard areas in two different experiments were counted in each case) at 6 or 3 days in uitro (div). At 6 div a higher percentage of retinal glia characterized by GFAP and VIM is present in cultures of precrushed retinae (about 70%) compared to normal retina cultures (less than 60%). At 3 div the relative number of glia decreases under both conditions (about 55% in precrushed and 40% in normal retina cultures) and the number of Thy l.l-positive cells increases from 20 to 35% in precrushed retinae and from 40 to 50% in normal retinae.
tion also shows GFAP staining at 9 days in uitro). After 6 days in vitro only fibroblasts seem to proliferate (Table 3). At very early stages in vitro a population of small rounded cells can be observed near the explants edges. These could be macrophages removing cellular debris. In normal retina explants significantly fewer explants show glial cell migration compared to “activated” retinae (Table 2). Immunohistochemical staining shows that besides VIM-positive glia a high percentage of Thy l.l-positive fibroblasts proliferate and migrate out from retinal explants in these cultures (Table 2 and Fig. 2). When crush-activated retinal glia cultures were incubated with BrdU for 12 h about 39% of the cells that had migrated out from the explants incorporated BrdU. Double-labeling experiments showed that about 10% of the BrdU-labeled cells were positive for GFAP and 27% showed VIM staining (Fig. 1). Therefore only about 40% of the cells present after 6 days in vitro seem to be of glial origin. The immunohistochemically defined glial cell population seems to consist of 25% astrocytes and about 75% Mtiller cells (see Table 3). In cultures from normal retinae only 12% of the BrdU-incorporating cells were positive for either GFAP or VIM. Therefore, proliferating cells from normal retinae seem to consist mainly of Thy l.l-positive flat cells. We did not see substantial accumulation of GFAP in VIM-positive Mtiller cells (less than 10%) during the time of observation which is in agreement with earlier reports on Miiller glia in vitro (9). Axon Regeneration from Normal Retinae on Retinal Glia The ability of retinal glia from normal and activated retinae to promote retinal ganglion cell axon growth was determined in retinal glia-adult retina cocultures.
FIG. 1. Double labeling of retinal glia with antibodies recognizing Vimentin and BrdU. This figure shows a population of activated retinal glial cells that have migrated out from retinal explants at 6 days in vitro (phase-contrast in a). About 60% of the cell population is positive for Vimentin at that time point (b). When retinal tissue cultures were incubated with BrdU for 12 h, most of the VIM-positive cells also incorporated BrdU (thin arrow in c), but VIM-negative flat cells (thick arrow in c), presumably fibroblasts, are also labeled with BrdU. Bar, 50 pm.
ADULT
RAT
RETINAL
GLIA
69
ina-retinal glia cocultures were processed for neurofilament and GFAP immunohistochemistry no special attraction to or avoidance of retinal astrocytes by regenerating ganglion cell axons could be observed (Fig. 3). However, when retinal ganglion cell axons reach the out-
FIG. 2. Double labeling of retinal glia with antibodies recognizing GFAP and Thy 1.1. A population of retinal glial cells derived from normal retinal explants is shown at 6 days in vitro (phase-contrast in a). Only a few cells can be labeled with GFAP in these cultures (b) and the majority of the cell population consists of Thy 1.1.positive flat cells (c), presumably fibroblasts. Bar, 50 pm.
vated glia (see Figs. 3 and 4). Significantly fewer axons emanate from retinal explants cultured on retinal glia derived from normal explants (9, 13) than from explants on activated glia (16, 40) (t test, P 5 0.01)). RGC axons visualized by anti-neurofilament antibodies and examined in scanning EM grow on Miiller cells, astrocytes, and fibroblasts as determined by scanning EM and immunohistochemistry (Figs. 3 and 4). When ret-
FIG. 3. Retinal ganglion cell axon growth on retinal glia monolayers. A coculture experiment of normal rat retinal explants on activated retinal glia is shown (phase-contrast in a). Retinal glia that had migrated out from retinal explants with prior optic nerve crush (“activated glia”) was used as a substratum for retinal explants. Only a few GFAP-positive astrocytes can be detected in these cultures at 9 days in vitro (arrow in b). RGC axons are visualized by neurofilament antibodies (c). The arrowhead in c shows the identical location as the arrow in (b), indicating that astrocytes are permissive for RGC axons in vitro, although no special preference of RGC axons can be seen in these experiments. Bar, 50 pm.
70
M.
FIG. 4. Retinal ganglion cell axons grow on adult retinal glia. Many neurites visualized by anti-neurofilament antibodies regrow from normal retinal explants on retinal glia after 3 days in uitro (a). In b a corresponding area is shown in scanning EM (X7840 enlargement). The arrowheads indicate RGC axons that grow over surfaces of retinal glia (indicated as G). The axons do not extend onto the acellular PL-Lam substrate. Bar, 50 pm.
ward rim of the glia monolayer, they show a clear preference for the cellular substratum compared to the PLLam-coated surface of the petri dishes (Fig. 5). Scanning EM also shows that retinal ganglion cell axons prefer to grow on cell surfaces compared to the acellular PL-Lam culture substrate (Fig. 4). DISCUSSION
Within the retina of most mammals two different types of glial cells, Miiller cells and astrocytes (for re-
B;l;HR
view see (33)), can be distinguished. Miiller cells are radially oriented cells that span the whole retinal width with their processes (33). They usually do not express GFAP but rather express VIM (9,lO). Astrocytes found in the retina resemble those found in other brain regions (19) with a preference for a location near blood vessels. Several recent experiments indicated that retinal astrocytes might be immigrants from the optic nerve in the therefore correspond to rat (19, 45). Retinal astrocytes type I astrocytes found in the optic nerve (37) since type II astrocytes and oligodendrocytes are kept out of the rat retina (19). After lesioning the retina or the optic nerve, retinal astrocytes increase their GFAP-staining intensities (10). This has been shown to correlate with “reactive” astrogliosis observed in other brain regions after penetrating lesions (34, 38, 40). Mtiller cells as well as astrocytes (reactive astrogliosis) have been shown to accumulate GFAP after optic nerve lesions (10, 22, 31), but the number of retinal Miiller cells that are transformed seems to be rather low (9, 39). The aim of the present paper was to investigate the reaction of retinal astrocytes and Miiller cells after lesioning the optic nerve in situ. We have used a tissue culture model to study glia populations and fibroblasts from adult rat retinae that proliferate and migrate out after retina explantation on acellular substrates. Both glia populations from adult retinae were able to survive in tissue culture. Miiller cells and retinal astrocytes migrated out from adult rat retinal explants in vitro. With an earlier onset the extent of glia migration was significantly higher in retinal explants whose corresponding optic nerves were crush-axotomized in situ. Retinal Miilier cells that were VIM-positive and GFAPnegative did not substantially accumulate GFAP during the time of observation, which is consistent with earlier findings (9,39). In normal adult rat retinal explants, the cell population that migrates and proliferates after explantation consists to a high degree of Thy l.l-positive flat cells, presumably fibroblasts. Therefore our results suggest that RGC axotomy and/or degeneration influences the ability of adult rat retinal glia to survive, migrate, and proliferate in tissue culture. This is consistent with findings in adult rat cortical astrocytes that showed increased survival in vitro after degeneration of surrounding neuronal fiber tracts in uiuo (25). Therefore some signal of degenerating or regenerating CNS neurons seems to initiate proliferation or induce a transformation of adult rat glia into a more “vigorous” state (8, 21). This transformation might be reflected in increased GFAP expression (reactive astrocytes) and in changes in the composition of cell surface molecules (8). Nevertheless, glia in the adult mammal seem to have only a restricted potential to proliferate (25,34) because at 9 days in vitro the relative number of glia decreases in both glia preparations from crush-activated and normal
ADULT
RAT
RETINAL
GLIA
71
FIG. 5. Retinal ganglion cell axons respect cellular-acellular substrate borders. The outward rim (arrowheads in a) of retinal glial cells that had migrated out from adult rat retinal explants is shown (a, phase-contrast). When retinal explants from normal rats were cultured on these glia monolayers, regenerating RGC axons (visualized by neurofilament antibodies in b) respected the border to the acellular PL-Lam substrate. Bar, 50 pm.
retinae. There are controversial reports about glia proliferation in response to penetrating lesions in adult mammals. Glia identified with S-100 cell surface markers did not proliferate substantially as determined by [3H]thymidine incorporation (32). Other studies have shown an increased mitogenic activity for astrocytes accumulating around lesioned brain areas, but in general adult astrocytes seem to have a restricted potential to proliferate (25, 34). The present results of adult rat retinal glia in vitro show that glia from normal rats did not show substantial migration and proliferation after ex-
plantation. Glia derived from retinae whose optic nerves received a conditioning lesion in viuo were able to migrate over larger distances and to proliferate in vitro as determined by BrdU incorporation. This suggested that retinal glia might be able to support neurite growth from CNS neurons. To test this hypothesis coculture experiments of adult rat retina and retinal glia were performed. In this experimental model, retinal glia that had migrated out from retinal explants were tested for their ability to promote RGC axon regeneration. Activated glia that had migrated out from
72
M.
adult rat retinal explants were able to initiate RGC axon growth from normal retinal explants. This suggests that retinal glia can support RGC axon growth after optic nerve lesions (8) and a signal from retinal glia (that is not yet characterized) might potentiate other signals acting to allow RGC survival and initiate RGC axon outgrowth. This supports the findings of Lindsay (24) who demonstrated that glia derived from adult rat brain after kainate lesions were able to support neuronal survival and axon growth. This suggests that different “functional states” of adult astrocytes might be postulated: Penetrating CNS lesions seem to induce astroglial hypertrophy (10, 38) and this reactive astrogliosis seems to impede axon regeneration (26). Selective neuronal degeneration induced either by axotomy or by kainate injections (24) might lead to glial changes that are compatible with axon growth. The present results show that retinal ganglion cell axons which reextend from explants preferentially grow on surfaces of adult rat retinal glia. No special preference for either astrocytes or MUller cells could be observed in these experiments. Interestingly, RGC axons respect borders to the acellular substratum PL-Lam. This suggests that regenerating RGC axons were able to make a choice between cellular and acellular substrata. Whether regenerating RGC axons prefer glial cell surfaces or extracellular matrix molecules other than laminin (15) used to coat the culture dishes is not yet clear (7). We found that cellular substrates in general were preferred by those axons. For example RGC axons regenerating on nonactivated glia also grew on nonglial cells (data not shown) that consisted to a high degree of Thy l.l-positive flat cells. In these experiments no special preference for either glial or nonglial cells was observed. This seems to be different from embryonic CNS axons which show a clear preference for glial compared to nonglial cells (17, 35) in uitro. In conclusion, our experiments show that adult rat retinal glia is activated after lesioning the optic nerve in situ (8). This conditioning lesion activates Miiller cells and astrocytes which proliferate and migrate out from retinal explants in uitro. In addition adult rat retinal glia is able to initiate axon regeneration from normal retinal explants. Both, GFAP-positive adult astrocytes and VIM-positive Miiller cells proliferated and were permissive for RGC axon growth from retinal explants. This result is consistent with the finding that RGC axons (6, 18) and other CNS axons (18) extend axons on cortical astrocytes from embryonic or early postnatal rats in uitro. Our results, however, show that even adult activated glia might support axon regeneration of adult CNS neurons. The present results are also consistent with the hypothesis that the oligodendrocyte-type 2 astrocyte lineage (36, 41) might be responsible for inhibitory aspects of mature CNS glia since no glia of that cell lineage was present in our culture system.
BAHR
Future experiments are assigned to show which cell surface or extracellular matrix molecules promote RGC axon growth on adult glia and to determine whether activated adult rat retinal glia might also enhance RGC survival in vitro. Both enhanced survival and promotion of axon regeneration are essential prerequisites for effective CNS regeneration, The molecular mechanisms that determine the ability of different glia populations to promote or inhibit adult CNS regeneration are not understood yet. The present results suggest that glial subpopulations exist in adult rats and that some of these cells have preserved an ability to support regeneration in the adult mammalian CNS. ACKNOWLEDGMENTS The author thanks Professor J. Dichgans and Professor F. Bonhoeffer for supporting his work; Drs. B. Schlosshauer, T. Alsopp, and S. Thanos for critical comments on the manuscript; and Jiirgen Berger for technical assistance at the electron microscope.
REFERENCES 1.
ARD, M. D., AND R. P. BUNGE. 1988. Heparan sulfate proteoglycan and laminin immunoreactivity on cultured astrocytes: Relationship to differentiation and neurite growth. J. Neurosci. 8:
2844-2858. 2. ARMSON, P. F., M. R. BENNE?T,
AND T. R. RATU. 1987. Retinal ganglion cell survival and neurite regeneration: The change from Mtiller cell dependence to superior colliculi dependence during development. Deu. Brain Res. 32: 207-216.
3. BXHR, M., J. VANSELOW, ation Bruin
of adult rat ganglion Res. 73: 393-401.
AND S. THANOS. 1988. In vitro regenercell axons from retinal explants. Exp.
4. B~~HR, M., J. VANSELOW,
AND S. THANOS. rat ganglion cells to regrow axons in vitro fibroblast growth factor and gangliosides. 197-201.
5. BKHR, M., AND R. P. BUNGE.
6.
7. 8. 9.
10.
11.
1989. Ability of adult can be influenced by Neurosci. Lett. 96:
1989. Functional status influences the ability of Schwann cells to support adult rat retinal ganglion cell survival and axonal growth. Exp. Neural. 106: 27-40. BKHR, M., M. SCHACHNER, AND R. P. BUNGE. 1989. In vitro regeneration of adult rat retinal ganglion cell axons on astrocytes: Growth characteristics and molecular mechanisms. Sot. Neurosci. Abstr. 19: 331. BOHR, M., J. M. HOPKINS, AND R. P. BUNGE. 1989. Schwann cell promotion of adult rat retinal ganglion cell regeneration in vitro. Eur. J. Neurosci. (Suppl.) 2: 302. B~~HR, M., AND B. SCHLOSSHALJER. 1989. Jones ganglioside expression on retinal glia increases after transection of adult rat retinal ganglion cell axons. J. Neurocytol. 18: 553-563. BARTLETT, P. F., M. D. NOBLE, R. M. PRUSS, M. C. RAFF, S. RATTRAY, AND C. A. WILLIAMS. 1981. Rat neural antigen-2 (RAN-2): A cell surface antigen on astrocytes, ependymal cells, Miiller cells and leptomeninges defined by monoclonal antibody. Bruin Res. 204: 339-351. BIGNAMI, A., AND D. DAHL. 1979. The radial glia of Miiller in the rat retina and their response to injury: An immunofluorescence study with antibodies to glial fibrillary acidic (GFA) protein. Exp. Eye Res. 28: 63-69. BRAY, G. M., M. P. VILLEGAZ-PEREZ, M. VIDAL-SANZ, AND A. J. AGUAYO. 1987. The use of peripheral nerve grafts to enhance
ADULT neuronal survival, promote growth and permit nections in the central nervous system of adult 132: 5-19.
RAT
RETINAL
Astrocytes (S. Fedoroff demic Press, New York.
terminal reconrats. J. Erp. Biol.
12.
Biissow, H. 1980. The astrocytes in the retina head of mammals: A special glia for the ganglion Tissue Res. 206: 367-378.
13.
CARONI, P., AND M. E. SCHWAB. 1988. Two membrane protein fractions from rat central myelin with inhibitory properties for neurite growth and fibroblast spreading. J. CellBiol. 106: 12811288.
73
GLIA and A. Vernadakis,
Eds.),
Vol.
3. Aca-
28.
MCCAFFERY, C. A., T. R. RAJU, AND M. R. BENNET. 1984. Effects of cultured astroglia on the survival of neonatal rat retinal ganglion cells in vitro. Deu. Biol. 104: 441-448.
29.
MCDONNEL, cell axons
and optic nerve cell axons. Cell
P., AND M. BERRY. 1982. Regeneration in the mouse retina. Brain Res. 241:
of ganglion
362-365.
30.
MCLOON, S., L. K. MCLOON, S. L. PALM, AND L. T. FURCHT. 1988. Transient expression of laminin in the optic nerve of the developing rat. J. Neurosci. 8: 1981-1990.
14.
CARONI, P., AND M. E. SCHWAB. 1988. Antibody against melin associated inhibitor of neurite growth neutralizes nonpermissive substrate properties of CNS white matter. Neuron 1: 85-96.
31.
MILLER, M. N., AND M. OBERDORFER. 1981. Neuronal and neuroglial responses following retinal lesions in neonatal rats. J. Comp. Neural. 202: 493-504.
15.
DAVID, S. 1988. Neurite outgrowth from mammalian CNS neurons on astrocytes in vitro may not be mediated primarily by laminin. J. Neurocytol. 17: 131-144. ENG, L. F., P. J. REIER, AND J. D. HOULE. 1987. Astrocyte activation and fibrous gliosis: Glial fibrillary acidic protein immunostaining of astrocytes following intraspinal cord grafting of fetal CNS tissue. In Progress in Bruin Research (F. J. Seil, E. Herbert, and B. M. Carlson, Eds.1, Vol. 71. Elsevier Science Pub., Amsterdam/New York. FALLON, J. R. 1985. Preferential outgrowth of central nervous system neurites on astrocytes and Schwann cells as compared to nonglial cells in vitro. J. Cell Biol. 100: 198-207.
32.
MIYAKE, T., T. HATTORI, M. FUKUDA, AND T. KITAMURA. 1989. Reactions of S-100 positive glia after injury of mouse cerebral cortex. Brain Res. 489: 31-40. NEWMAN, E. A. 1986. The Miiller cell. In Astrocytes (S. Fedoroff and A. Vernadakis, Eds.), Vol. 1. Academic Press, New York.
16.
17.
18.
FAWCETT, J. W., J. ROKOS, axons in three-dimensional
AND I. BAKST. 1989. The growth of astrocyte cultures. Deu. Biol. 135:
33. 34.
NIETO-SAMPEDRO, M., P. R. SANETO, J. DE VELLIS, AND C. W. COTMAN. 1985. The control of glial populations in brain: Changes in astrocyte mitogenic and morphogenic factors in response to injury. Bruin Res. 343: 320-328.
35.
NOBLE, M., J. FOK-SEANG, AND J. COHEN. 1984. Glia are a unique substrate for the in vitro growth of central nervous system neurons. J. Neurosci. 4: 1892-1903.
36.
RAJU, T. R., AND M. R. BENNET. 1986. Retinal ganglion cell survival requirements: A major but transient dependence on Miiller glia during development. Brain Res. 383: 165-176.
37.
RAFF, M. C. 1989. Glial cell diversification Science 243: 1450-1455.
38.
REIER, P. J., L. J. STENSAAS, AND L. GUTH. 1983. The astrocytic scar as an impediment to regeneration in the central nervous system. In Spinal Cord Reconstruction (C. C. Kao, R. P. Bunge, and P. J. Reier, Eds.). Raven Press, New York. SCHERER, J., AND J. SCHNITZER. 1989. The rabbit retina: A suitable mammalian tissue for obtaining astroglia-free Miiller cell cultures. Neurosci. Lett. 97: 51-56.
449-458. 19.
20.
FFRENCH-CONSTANT, C., R. H. MILLER, J. F. BURNE, AND M. C. RAFF. 1988. Evidence that migratory oligodendrocyte-type 2 astrocyte (O-2A) progenitor cells are kept out of the rat retina by a barrier at the eye-end of the optic nerve. J. Neurocytol. 17: 1325. GOLDBERG, S., AND B. FRANK. 1980. Will central nervous systems in the adult mammal regenerate after bypassing a lesion? A study in the mouse and chick visual systems. Exp. Neural. 70: 675-689.
21.
JANECKO, K. 1989. Spatiotemporal patterns liferation in the rat brain injured at the development: A combined immunohistochemical graphic study. Brain Res. 485: 236-243.
22.
KARSCHING, A., H. WKSSLE, AND J. SCHNITZER. 1986. Immunohistochemical studies on astroglia of the cat retina under normal and pathological conditions. J. Comp. Neural. 249: 564-576.
23.
LIESI, P., D. DAHL, AND A. VAHERI. 1983. Laminin is produced by early astrocytes in primary culture. J. Cell Biol. 96: 920-924. LINDSAY, R. M. 1979. Adult rat brain astrocytes support survival of both NGF-dependent and NGF-insensitive neurones. Nature (London) 282: 80-82.
24.
25.
of the astroglial propostmitotic stage of and autoradio-
LINDSAY,R. M., P.C.BARBER, M.R.C. SHERWOOD,J.ZIMMER, AND G. RAISMAN. 1982. Astrocyte cultures from adult rat brain: Derivation, characterization and neuronotrophic properties of pure astroglial cells from corpus callosum. Brain Res. 243: 329343.
26.
LIUZZI, F. J., AND R. J. LASEK. 1987. Astrocytes block regeneration in mammals by activating a physiological pathway. Science 237: 642-645.
27.
MANTHORPE, M., J. S. RUDGE, AND S. VARON. 1986. Astroglial cell contributions to neuronal survival and neuritic growth. In
39.
in the rat optic
nerve.
40.
SCHIFFER, D., M. T. GIORDANA, A. MIGHELI, G. GIACCONE, S. PEZZOTTA, AND A. MAURO. 1986. Glial fibrillary acidic protein and Vimentin in the experimental glial reaction of rat brain. Brain Res. 374: 110-118.
41.
SMITH, G. S., R. H. MILLER, AND J. SILVER. 1986. Changing role of forebrain astrocytes during development, regenerative failure and induced regeneration upon transplantation. J. Camp. Neural. 251:23-43.
42.
SCHWAB, M. E., AND P. CARONI. 1988. Oligodendrocytes and CNS myelin are non-permissive substrates for neurite growth and fibroblast spreading in vitro. J. Neurosci. 8: 2381-2393. TOMASELLI, K. J., K. M. NEUGEBAUER, J. L. BIXBY, J. LILLIEN, AND L. F. REICHARDT. 1988. N-Cadherin and integins: Two receptor systems that mediate neuronal process outgrowth on astrocyte surfaces. Neuron 1: 33-43.
43.
44.
axonal stop 45.
THANOS, S., M. BKHR, Y.-A. BARDE, AND J. VANSELOW. 1988. Survival and axonal elongation of adult rat retinal ganglion cells: In vitro effects of lesioned sciatic nerve and brain derived neurotrophic factors. Eur. J. Neurosci. 1: 19-26. WATANABE, T., AND M. C. RAFF. 1988. Retinal astrocytes are immigrants from the optic nerve. Nature (London) 332: 834837.
