Neuroscience Letters, 138 (1992) 89-92 © 1992 Elsevier Scientific Publishers Ireland Ltd, All rights reserved 0304-3940/92/$ 05.00
89
NSL 08531
Mtiller (glial) cells in the retina of urodeles and anurans reveal different morphology by means of freeze-fracturing Hartwig Wolburg ~, Karin Berg-von der Emde ~ and Christiane Naujoks-Manteuffel b "Institute of Pathology, University of T~ibingen, Tiibingen ( FRG) and blnstitute of Brain Research, University of Bremen, Bremen ( FRG) (Received 11 December 1991; Revised version received 13 January 1992; Accepted 17 January 1992)
Key words: Glial (Mtiller) cell; Orthogonal array of particles; Amphibia; Freeze-fracturing Mfiller (glial) ceils of the retina of various species of amphibia (urodeles and anurans) were investigated by means of the freeze-fracture technique. This was done because Mtiller cells in anamniotes were believed to differ from those in mammals in that they should lack the so-called orthogonal arrays of particles (OAP) which are a characteristic feature of Miiller cells in mammalian retina. However, as we could demonstrate previously (Bergvonder Emde and Wolburg, Glia, 2 (1989) 458), fish retinal Mtiller cells also contain OAP in their membranes suggesting that OAP are a general marker of MiJller cells in all vertebrates. As demonstrated in this study, Mfiller cells of urodeles (Batrachoseps attenuatus and Pleurodeles waltlii) are OAP-positive, whereas two anurans (Rana esculenta and Xenopus laevis) do not reveal any OAP in their Mfiller cell membranes. Under phylogenetic aspects, it appears very interesting that frogs are as yet the only vertebrate group that deviates from all other vertebrates in terms of Mtiller cell membrane morphology.
Mfdler cels represent the main glial cell type in the vertebrate retina extending from the inner to the outer limiting membrane and contacting all neuronal elements. At the vitreal pole of the retina Mfiller cells form endfeet that seal up the neuronal milieu from the vitreous body. Here, the membrane is covered by a basal lamina, the inner limiting membrane. Mtiller cells and astrocytes which are believed to be related to each other are characterized by the occurrence of orthogonal arrays of particles (OAP) which can be visualized as yet exclusively by the freeze-fracturing technique. They accumulate predominantly at that cellular pole which contacts mesenchymal compartments such as blood vessels, the subpial space at the surface of the brain or - - in the retina - - the vitreous body. Since astrocytes are known to express OAP only in reptiles, birds and mammals [5, 7, 12] and not in fish and amphibia [3, 6, 20, 21], OAP- positive Mtiller cells were believed to be distributed correspondingly among the vertebrates. In contrast to that, we could demonstrate that the Mt~ller cells in the goldfish retina possess OAP and that their distribution resembles that in the mammalian retinal MOller cell [1]. This prompted us to ask whether there is a heterogeneity of OAP-occur-
Correspondence." H. Wolburg, Pathologisches [nstitut der Universit~it, LiebermeisterstaBe 8, W-7400 T~ibingen, FRG.
rence between astrocytes and Mi~ller cells also in amphibians. Retinas of four amphibian species, salamanders and frogs (Pleurodeles, Batrachoseps, Rana and Xenopus) were dissected out after anesthesia with methane tricaine sulfonate (MS 222, Sandoz) and perfusion fixation with 2% glutaraldehyde (Paesel, FRG) buffered with 0.1 M cacodylate buffer. For preparation of ultrathin sections, routine methods were employed. Briefly, the tissue was osmicated in 1% OsO4, dehydrated in ethanol, embedded in Araldite (Serva, FRG) and sectioned on a LKB Nova ultramicrotome. For freeze fracturing, the glutaraldehyde-fixed specimens were impregnated in 30% glycerol for 1 h, mounted between two gold specimen holders, shock-frozen in nitrogen-slush (-210°C) and fractured in a Balzers freeze-fracture device (BAF 400D) at -150°C and about 5 x 10-6 mbar. The fracture faces were shadowed with platinum/carbon (2 nm, 45 °) and carbon (30 nm), and the replicas were cleaned in 12% sodium hypochlorite for 30 min and water. Sections and replicas were observed in a Zeiss EM 10 and Philips CM 12 electron microscope. Batrachoseps. Freeze-fracture replicas of Mtiller cell endfeet show a large density of intramembranous particles. Additionally, OAP and caveolae are found. The density of OAP is in the magnitude of 100-150//~m2. Single OAP mostly form rows of double or triplicate sub-
91
units, reaching a whole of 12-20 subunits. The caveolae are not equally distributed in the endfoot membrane but form groups of about 100. In a membrane area rich in caveolae the OAP density is decreased, although a direct association of both structures is not avoided (Fig. 1). Where the endfoot membrane has lost the direct contact with the basal lamina the OAP density decreases; but the decrease is not as abrupt as was observed in the goldfish [1] or rabbit [16, 18] Mtiller cell membrane (Fig. 2). Caveolae are not confined to the true endfoot region as well; however, as can be observed in ultrathin sections, their frequency is higher at the Mialler cell-vitreal interface than at the Miiller cell-neuropil interface. Pleurodeles, In principle, the membrane morphology of Miiller cells resembles that of Batrachoseps. However, the Mialler cells of both species differ in that Pleurodeles reveals a lower density of OAP (in the magnitude of 3050/pm 2) and nearly no OAP in the membranes outside the basal lamina contact. As described in Batrachoseps, we find a coexistence of OAP and caveolae in the same membrane. Rana and Xenopus. In Mialler cell membranes of Rana and Xenopus no OAP could be found. This was not due to seasonal variations nor to the youth of the animals, because we investigated adult frogs from different seasons. The endfeet terminating at the inner limiting membrane are very large (often more than 10 pm2). The membrane is homogeneously occupied with intramembraneous particles (Fig. 3). Where the contact between endfoot and basal lamina is lost the membrane architecture is not altered (Fig. 4). Caveolae are found, but very rarely in comparison to the urodeles described above (Fig. 3). The results presented here demonstrate a heterogeneity of Mtiller cell structure within the group of amphibia. This heterogeneity is unexpected and puzzling: formerly, Mtiller cells of anamniotes were believed to show the same features as the astrocytes in that they would lack any OAP (see Gotow and Hashimoto [4]);
however, the Mailer cells of the goldfish were shown to possess OAP in contrast to astrocytes [1]. Now assuming that all vertebrates contain OAP in their Mtiller cell membranes, we investigated different amphibia and found that anurans are the only exception among vertebrates studied so far as they really lack any OAP in glial cells (this study, [5, 6, 21]). To complete the puzzle one should emphasize that Franzini-Armstrong [2] described OAP in skeletal muscle cells of Rana. Provided that OAP in muscle and glial cells would represent the identical gene products, the information for the expression of OAP would be realized differentially in different cell types. Why OAP are not expressed in glial cells but in muscle cells (Rana) or in Mtiller cells but not in astrocytes (goldfish [1], urodeles [21]; own results from the optic nerve of Pleurodeles, unpublished) is an enigma not to be resolved before getting more information about the biochemical identity and the functions of these particle arrays. It was speculated that OAP could be related to the K+-conductance of the Mtiller cell, because both parameters are identically distributed across the surface of the cell [1, 13, 15, 18, 19]. Alternatively, also the distribution of anionic currents in glial cells was considered to be correlated with that of OAP [8]. In fact, sodium bicarbonate cotransporter sites were localized preferentially to the salamander Mtiller cell endfoot [14]. Retinal Mtiller cells of frogs do not only differ from those of salamanders and other vertebrates in that they do not possess any OAP; also their cytoskeleton as well as that of spinal cord radial glial cells do not possess considerable amounts of GFAP or vimentin [10, 11, 17, 22]. Frogs are highly derived amphibians and might have acquired a unique trait with the absence of OAP in the central nervous system, a character that is not found in other vertebrates. Investigations in caecilians and dipnoans would shed more light on the question whether frogs are really unique in lacking OAP in the membranes of retinal Mtiller cells.
(__ Fig. 1. Freeze-fracture of the P-face of MUller cell endfoot of Batrachoseps revealing fields occupied by caveolae. Many OAP are visible. The arrowheads mark the coexistence of OAP and caveolae. × 100,000. Fig. 2. Mialler cell endfoot in the retina of Batrachoseps. The upper half shows the P-face of an vitreous-associated membrane with many OAP, the lower half the E-face of the endfoot membrane directed towards the retinal neuropil showing less OAP (encircled). x 80,000. Fig. 3. Freeze-fracture replica of a Mfiller cell endfoot in the retina of Rana. Below is the vitreous, in the upper half ganglion cell axons run between MOiler cell endfeet. The P-face of the lateral Mtiller cell endfoot membrane is characterized by a high density of intramembraneous particles and few caveolae, x 31,000. Fig. 4. P-face o f a paravitreous Miiller cell endfoot in the retina of Rana, lacking any OAP. Where the membrane deflects away from the contact with the basal lamina (row of asterisks), the architecture of the membrane does not alter. × 50,000.
92
Finally, an interesting detail concerns the coexistence of OAP and caveolae in the Mtiller cell endfoot membranes in salamanders. As Massa [9] pointed out, there seems to be an inverse relationship between the densities of OAP and caveolae. Accordingly, we could show in the goldfish that paravitreous Mtiller cell membranes contain OAP and no caveolae, and astrocytes reveal caveolae and no OAP [1]. However, in the urodele Mtiller cell we find relatively high densities of both OAP and caveolae at the same site (Fig. 1). The functional significance of this finding remains unclear. With support by the Deutsche Forschungsgemeinschaft. Mrs. E.-M. Schmid is thanked for photographic help. 1 Berg-von der Emde, K. and Wolburg, H., MOiler (glial) cells but not
2
3
4
5
6 7
8
9
astrocytes in the retina of the goldfish possess orthogonal arrays of particles, Glia, 2 (1989) 458-469. Franzini-Armstrong, C., Freeze-fracture of frog slow tonic fibers. Structure of surface and internal membranes, Tissue and Cell, 16 (1984) 647 664. Gotow, T. and Hashimoto, EH., Plasma membrane organization of astrocytes in elasmobranchs with special reference to the brain barrier system, J. Neurocytol., 13 (1984) 727 742. Gotow, T. and Hashimoto, EH., Orthogonal arrays of particles in plasma membranes of MiJller cells in the guinea pig retina, Glia, 2 (1989) 273-285. Kastner, R., Comparative studies on the astrocytic reaction in the lesioned central nervous system of different vertebrates, J. Hirnforsch., 28 (1987) 221 232. Korte, G.W. and Rosenbluth, J., Ependymal astrocytes in the fi-og cerebellum, Anat. Record, 199 (1981) 267 279. Landis, D.M.D., and Reese, T.S., Membrane structure in mammalian astrocytes: A review of freeze-fracture studies on adult, developing, reactive and cultered astrocytes, J. Exp. Biol., 95 (1981) 35-48. Landis, D.M.D., Weinstein, L.A. and Skordeles, C.J., Effects of" dexamethasone on the differentiation of membrane structure in cultered astrocytes, Glia, 4 (1991) 335-344. Massa, P., Plasmalemmal vesicles (caveolae) of fibrous astrocytes of the cat optic nerve, Am. J. Anat., 165 (1982) 69 81.
10 Miller, R.H. and Liuzzi, E.J., Regional specialization of the radial glial cells of the adult frog spinal cord, J. Neurocytol., 15 (1986) 187 196. 11 Naujoks-Manteuffel, C. and Roth, G., Astroglial cells in a salamander brain (Salamandra salamandra) as compared to mammals: a glial fibrillary acidic protein immunohistochemistry study, Brain Res., 487 (1989) 397-401. 12 Neuhaus, J. and Wolburg, H., Heterogeneity of astrocytic membranes in the optic nerve and spinal cord of the lizard, Anolis carolinensis, Cell Tissue Res., 242 (1985) 185 190. 13 Newman, E.A., Distribution of potassium conductance in mammalian Mi~ller (glial) cells: a comparative study, J. Neurosci., 7 (1987) 2423 2432. 14 Newman, E.A. and Astion, M.L., Localization and stoichiometry of electrogenic sodium bicarbonate cotransport in retinal glial cells, Glia, 4 (1991) 424 428. 15 Nilius, B. and Reichenbach, A., Efficient K + buffering by mammalian retinal glial cells is due to cooperation of specialized ion channels, Pfl~gers Arch., 411 (1988) 654-660. 16 Richter, W., Reichenbach, A. and Reichelt, W., Orthogonal arrays of intramembranous particles in the Mtiller cells and astrocyte endt~)ot membrane of rabbit retina. Postnatal development and adulthood, J. Neurocytol., 19 (1990) 127- 139. 17 Szaro, B.G. and Gainer, H., Immunocytochemical identification of non-neuronal intermediate filament proteins in the developing Xenopu.g laevis nervous system, Dev. Brain Res., 43 (1988) 207 224. 18 Wolburg, H. and Berg, K., Heterogeneity of Miiller cell endfeet in the rabbit retina as revealed by freeze-fracturing, Neurosci. Lett., 82 (1987) 273 277. 19 Wolburg, H. and Berg, K., Distribution of orthogonal arrays of particles in the MUller cell membrane of the mouse retina, Glia, 1 (1988) 246 252. 20 Wolburg, H., Kastner, R. and Kurz-lsler, G., Lack of orthogonal particle assemblies and presence of tight junctions in astrocytes of goldfish. A freeze-fracture study, Cell Tissue Res., 234 (1983) 389 402. 21 Wujek, J.R. and Reier, P.J., Astrocytic membrane morphology: Differences between mammalian and amphibian astrocytes after axotomy, J. Comp. Neurol.. 222 (1984) 607 620. 22 Zamora, A.J. and Mutin, M.. Vimentin and glial fibrillary acidic protein filaments in radial gtia of the adult urodele spinal cord, Neuroscience, 27 (1988) 279 288.
89
NSL 08531
Mtiller (glial) cells in the retina of urodeles and anurans reveal different morphology by means of freeze-fracturing Hartwig Wolburg ~, Karin Berg-von der Emde ~ and Christiane Naujoks-Manteuffel b "Institute of Pathology, University of T~ibingen, Tiibingen ( FRG) and blnstitute of Brain Research, University of Bremen, Bremen ( FRG) (Received 11 December 1991; Revised version received 13 January 1992; Accepted 17 January 1992)
Key words: Glial (Mtiller) cell; Orthogonal array of particles; Amphibia; Freeze-fracturing Mfiller (glial) ceils of the retina of various species of amphibia (urodeles and anurans) were investigated by means of the freeze-fracture technique. This was done because Mtiller cells in anamniotes were believed to differ from those in mammals in that they should lack the so-called orthogonal arrays of particles (OAP) which are a characteristic feature of Miiller cells in mammalian retina. However, as we could demonstrate previously (Bergvonder Emde and Wolburg, Glia, 2 (1989) 458), fish retinal Mtiller cells also contain OAP in their membranes suggesting that OAP are a general marker of MiJller cells in all vertebrates. As demonstrated in this study, Mfiller cells of urodeles (Batrachoseps attenuatus and Pleurodeles waltlii) are OAP-positive, whereas two anurans (Rana esculenta and Xenopus laevis) do not reveal any OAP in their Mfiller cell membranes. Under phylogenetic aspects, it appears very interesting that frogs are as yet the only vertebrate group that deviates from all other vertebrates in terms of Mtiller cell membrane morphology.
Mfdler cels represent the main glial cell type in the vertebrate retina extending from the inner to the outer limiting membrane and contacting all neuronal elements. At the vitreal pole of the retina Mfiller cells form endfeet that seal up the neuronal milieu from the vitreous body. Here, the membrane is covered by a basal lamina, the inner limiting membrane. Mtiller cells and astrocytes which are believed to be related to each other are characterized by the occurrence of orthogonal arrays of particles (OAP) which can be visualized as yet exclusively by the freeze-fracturing technique. They accumulate predominantly at that cellular pole which contacts mesenchymal compartments such as blood vessels, the subpial space at the surface of the brain or - - in the retina - - the vitreous body. Since astrocytes are known to express OAP only in reptiles, birds and mammals [5, 7, 12] and not in fish and amphibia [3, 6, 20, 21], OAP- positive Mtiller cells were believed to be distributed correspondingly among the vertebrates. In contrast to that, we could demonstrate that the Mt~ller cells in the goldfish retina possess OAP and that their distribution resembles that in the mammalian retinal MOller cell [1]. This prompted us to ask whether there is a heterogeneity of OAP-occur-
Correspondence." H. Wolburg, Pathologisches [nstitut der Universit~it, LiebermeisterstaBe 8, W-7400 T~ibingen, FRG.
rence between astrocytes and Mi~ller cells also in amphibians. Retinas of four amphibian species, salamanders and frogs (Pleurodeles, Batrachoseps, Rana and Xenopus) were dissected out after anesthesia with methane tricaine sulfonate (MS 222, Sandoz) and perfusion fixation with 2% glutaraldehyde (Paesel, FRG) buffered with 0.1 M cacodylate buffer. For preparation of ultrathin sections, routine methods were employed. Briefly, the tissue was osmicated in 1% OsO4, dehydrated in ethanol, embedded in Araldite (Serva, FRG) and sectioned on a LKB Nova ultramicrotome. For freeze fracturing, the glutaraldehyde-fixed specimens were impregnated in 30% glycerol for 1 h, mounted between two gold specimen holders, shock-frozen in nitrogen-slush (-210°C) and fractured in a Balzers freeze-fracture device (BAF 400D) at -150°C and about 5 x 10-6 mbar. The fracture faces were shadowed with platinum/carbon (2 nm, 45 °) and carbon (30 nm), and the replicas were cleaned in 12% sodium hypochlorite for 30 min and water. Sections and replicas were observed in a Zeiss EM 10 and Philips CM 12 electron microscope. Batrachoseps. Freeze-fracture replicas of Mtiller cell endfeet show a large density of intramembranous particles. Additionally, OAP and caveolae are found. The density of OAP is in the magnitude of 100-150//~m2. Single OAP mostly form rows of double or triplicate sub-
91
units, reaching a whole of 12-20 subunits. The caveolae are not equally distributed in the endfoot membrane but form groups of about 100. In a membrane area rich in caveolae the OAP density is decreased, although a direct association of both structures is not avoided (Fig. 1). Where the endfoot membrane has lost the direct contact with the basal lamina the OAP density decreases; but the decrease is not as abrupt as was observed in the goldfish [1] or rabbit [16, 18] Mtiller cell membrane (Fig. 2). Caveolae are not confined to the true endfoot region as well; however, as can be observed in ultrathin sections, their frequency is higher at the Mialler cell-vitreal interface than at the Miiller cell-neuropil interface. Pleurodeles, In principle, the membrane morphology of Miiller cells resembles that of Batrachoseps. However, the Mialler cells of both species differ in that Pleurodeles reveals a lower density of OAP (in the magnitude of 3050/pm 2) and nearly no OAP in the membranes outside the basal lamina contact. As described in Batrachoseps, we find a coexistence of OAP and caveolae in the same membrane. Rana and Xenopus. In Mialler cell membranes of Rana and Xenopus no OAP could be found. This was not due to seasonal variations nor to the youth of the animals, because we investigated adult frogs from different seasons. The endfeet terminating at the inner limiting membrane are very large (often more than 10 pm2). The membrane is homogeneously occupied with intramembraneous particles (Fig. 3). Where the contact between endfoot and basal lamina is lost the membrane architecture is not altered (Fig. 4). Caveolae are found, but very rarely in comparison to the urodeles described above (Fig. 3). The results presented here demonstrate a heterogeneity of Mtiller cell structure within the group of amphibia. This heterogeneity is unexpected and puzzling: formerly, Mtiller cells of anamniotes were believed to show the same features as the astrocytes in that they would lack any OAP (see Gotow and Hashimoto [4]);
however, the Mailer cells of the goldfish were shown to possess OAP in contrast to astrocytes [1]. Now assuming that all vertebrates contain OAP in their Mtiller cell membranes, we investigated different amphibia and found that anurans are the only exception among vertebrates studied so far as they really lack any OAP in glial cells (this study, [5, 6, 21]). To complete the puzzle one should emphasize that Franzini-Armstrong [2] described OAP in skeletal muscle cells of Rana. Provided that OAP in muscle and glial cells would represent the identical gene products, the information for the expression of OAP would be realized differentially in different cell types. Why OAP are not expressed in glial cells but in muscle cells (Rana) or in Mtiller cells but not in astrocytes (goldfish [1], urodeles [21]; own results from the optic nerve of Pleurodeles, unpublished) is an enigma not to be resolved before getting more information about the biochemical identity and the functions of these particle arrays. It was speculated that OAP could be related to the K+-conductance of the Mtiller cell, because both parameters are identically distributed across the surface of the cell [1, 13, 15, 18, 19]. Alternatively, also the distribution of anionic currents in glial cells was considered to be correlated with that of OAP [8]. In fact, sodium bicarbonate cotransporter sites were localized preferentially to the salamander Mtiller cell endfoot [14]. Retinal Mtiller cells of frogs do not only differ from those of salamanders and other vertebrates in that they do not possess any OAP; also their cytoskeleton as well as that of spinal cord radial glial cells do not possess considerable amounts of GFAP or vimentin [10, 11, 17, 22]. Frogs are highly derived amphibians and might have acquired a unique trait with the absence of OAP in the central nervous system, a character that is not found in other vertebrates. Investigations in caecilians and dipnoans would shed more light on the question whether frogs are really unique in lacking OAP in the membranes of retinal Mtiller cells.
(__ Fig. 1. Freeze-fracture of the P-face of MUller cell endfoot of Batrachoseps revealing fields occupied by caveolae. Many OAP are visible. The arrowheads mark the coexistence of OAP and caveolae. × 100,000. Fig. 2. Mialler cell endfoot in the retina of Batrachoseps. The upper half shows the P-face of an vitreous-associated membrane with many OAP, the lower half the E-face of the endfoot membrane directed towards the retinal neuropil showing less OAP (encircled). x 80,000. Fig. 3. Freeze-fracture replica of a Mfiller cell endfoot in the retina of Rana. Below is the vitreous, in the upper half ganglion cell axons run between MOiler cell endfeet. The P-face of the lateral Mtiller cell endfoot membrane is characterized by a high density of intramembraneous particles and few caveolae, x 31,000. Fig. 4. P-face o f a paravitreous Miiller cell endfoot in the retina of Rana, lacking any OAP. Where the membrane deflects away from the contact with the basal lamina (row of asterisks), the architecture of the membrane does not alter. × 50,000.
92
Finally, an interesting detail concerns the coexistence of OAP and caveolae in the Mtiller cell endfoot membranes in salamanders. As Massa [9] pointed out, there seems to be an inverse relationship between the densities of OAP and caveolae. Accordingly, we could show in the goldfish that paravitreous Mtiller cell membranes contain OAP and no caveolae, and astrocytes reveal caveolae and no OAP [1]. However, in the urodele Mtiller cell we find relatively high densities of both OAP and caveolae at the same site (Fig. 1). The functional significance of this finding remains unclear. With support by the Deutsche Forschungsgemeinschaft. Mrs. E.-M. Schmid is thanked for photographic help. 1 Berg-von der Emde, K. and Wolburg, H., MOiler (glial) cells but not
2
3
4
5
6 7
8
9
astrocytes in the retina of the goldfish possess orthogonal arrays of particles, Glia, 2 (1989) 458-469. Franzini-Armstrong, C., Freeze-fracture of frog slow tonic fibers. Structure of surface and internal membranes, Tissue and Cell, 16 (1984) 647 664. Gotow, T. and Hashimoto, EH., Plasma membrane organization of astrocytes in elasmobranchs with special reference to the brain barrier system, J. Neurocytol., 13 (1984) 727 742. Gotow, T. and Hashimoto, EH., Orthogonal arrays of particles in plasma membranes of MiJller cells in the guinea pig retina, Glia, 2 (1989) 273-285. Kastner, R., Comparative studies on the astrocytic reaction in the lesioned central nervous system of different vertebrates, J. Hirnforsch., 28 (1987) 221 232. Korte, G.W. and Rosenbluth, J., Ependymal astrocytes in the fi-og cerebellum, Anat. Record, 199 (1981) 267 279. Landis, D.M.D., and Reese, T.S., Membrane structure in mammalian astrocytes: A review of freeze-fracture studies on adult, developing, reactive and cultered astrocytes, J. Exp. Biol., 95 (1981) 35-48. Landis, D.M.D., Weinstein, L.A. and Skordeles, C.J., Effects of" dexamethasone on the differentiation of membrane structure in cultered astrocytes, Glia, 4 (1991) 335-344. Massa, P., Plasmalemmal vesicles (caveolae) of fibrous astrocytes of the cat optic nerve, Am. J. Anat., 165 (1982) 69 81.
10 Miller, R.H. and Liuzzi, E.J., Regional specialization of the radial glial cells of the adult frog spinal cord, J. Neurocytol., 15 (1986) 187 196. 11 Naujoks-Manteuffel, C. and Roth, G., Astroglial cells in a salamander brain (Salamandra salamandra) as compared to mammals: a glial fibrillary acidic protein immunohistochemistry study, Brain Res., 487 (1989) 397-401. 12 Neuhaus, J. and Wolburg, H., Heterogeneity of astrocytic membranes in the optic nerve and spinal cord of the lizard, Anolis carolinensis, Cell Tissue Res., 242 (1985) 185 190. 13 Newman, E.A., Distribution of potassium conductance in mammalian Mi~ller (glial) cells: a comparative study, J. Neurosci., 7 (1987) 2423 2432. 14 Newman, E.A. and Astion, M.L., Localization and stoichiometry of electrogenic sodium bicarbonate cotransport in retinal glial cells, Glia, 4 (1991) 424 428. 15 Nilius, B. and Reichenbach, A., Efficient K + buffering by mammalian retinal glial cells is due to cooperation of specialized ion channels, Pfl~gers Arch., 411 (1988) 654-660. 16 Richter, W., Reichenbach, A. and Reichelt, W., Orthogonal arrays of intramembranous particles in the Mtiller cells and astrocyte endt~)ot membrane of rabbit retina. Postnatal development and adulthood, J. Neurocytol., 19 (1990) 127- 139. 17 Szaro, B.G. and Gainer, H., Immunocytochemical identification of non-neuronal intermediate filament proteins in the developing Xenopu.g laevis nervous system, Dev. Brain Res., 43 (1988) 207 224. 18 Wolburg, H. and Berg, K., Heterogeneity of Miiller cell endfeet in the rabbit retina as revealed by freeze-fracturing, Neurosci. Lett., 82 (1987) 273 277. 19 Wolburg, H. and Berg, K., Distribution of orthogonal arrays of particles in the MUller cell membrane of the mouse retina, Glia, 1 (1988) 246 252. 20 Wolburg, H., Kastner, R. and Kurz-lsler, G., Lack of orthogonal particle assemblies and presence of tight junctions in astrocytes of goldfish. A freeze-fracture study, Cell Tissue Res., 234 (1983) 389 402. 21 Wujek, J.R. and Reier, P.J., Astrocytic membrane morphology: Differences between mammalian and amphibian astrocytes after axotomy, J. Comp. Neurol.. 222 (1984) 607 620. 22 Zamora, A.J. and Mutin, M.. Vimentin and glial fibrillary acidic protein filaments in radial gtia of the adult urodele spinal cord, Neuroscience, 27 (1988) 279 288.
