On the Occurrence of Microtubules within Mature Astrocytes CEDRIC S. RAINE AND HENRYK WISNIEWSKI Department of Pathology (Neuropathology ), Albert Einstein College of Medicine, Bronx, N e w York 10461 Microtubules have been described to occur frequently within the subABSTRACT pial fibrous astrocytes of the adult spinal cord of three species. This finding is related directly to perfusion fixation by glutaraldehyde. Immersion fixation produced no or fewer tubules in this area. It is suggested that the presence of microtubules in these cells might reflect a specific skeletal function. Astrocytes from other locations within the central nervous system displayed fewer microtubules within their cytoplasm. During a recent study on experimental demyelination and remyelination (WiSniewski et al., '69; Raine et al., '69; Prineas et al., '69) it was observed that within the spinal cord, the subpial astrocytes and those penetrating the base of each spinal nerve root contained an abundance of microtubules in addition to bundles of filaments (Prineas et al., '69). The existence of microtubules within fibrous astrocytes has never been emphasized as a diagnostic criterion for the identification of this cell type (Farquhar and Hartman, '57; Schultz et al., '57; Palay et al., '62; Mugnaini and Walberg, '64; Wendell-Smith et al., '66; Kruger and Maxwell, '67). The validity of this observation forms the subject of the present report in which the ultrastructure of fibrous astrocytes from several locations within the central nervous system (CNS) of three species has been examined. hours, dehydrated through a graded series of ethyl alcohol, passed through propylene oxide and epon, and embedded flat in capsules and silicone rubber molds. Slices of cord fured by simple immersion in buffered glutaraldehyde alone, in buffered glutaraldehyde followed by post-osmication, and Dalton's chrome osmium alone, were taken from one rabbit and processed identically for comparison with glutaraldehyde perfused material. Sections 1 thick, stained with 1 % toluidine blue were examined light microscopically prior to sectioning for grids. Ultramicrotomy was carried out on Porter-Blum MT1 and Reichert OMU2 microtomes; thin sections were stained with lead citrate and uranyl acetate (Venable and Coggeshall, '65), and scanned in a Siemens Elmiskop 1A. RESULTS The subpial fibrous astrocytes of the spinal cord were intimately bounded along Twenty-four adult animals, 22 New their leptomeningeal surfaces by a 600 A Zealand albino rabbits, one cat, and one thick basement membrane (fig. 1). The dog were used for this study. Animals were cell surface beneath this membrane conanesthetized with intraperitoneal sodium tained periodic osmiophilic regions ("halfpentobarbitone prior to sacrifice. Fixation desmosomes" - Hirano et al., '69), and was achieved by arterial perfusion through between these superficial cells and their the left ventricle with 100 ml of 4% para- processes, frequent desmosome-like juncformaldehyde followed by 5 1 of 5% gluta- tions could be seen. Astrocytes in these reraldehyde in 0.1 m phosphate buffer at gions provided a tight continuous covering pH 7.3 (WiSniewski et al., '69). Following over the white matter, varying in thickness fixation, coronal sections of the brain from a single layer of processes to several 1-2 mm thick, and transverse sections of whole cells. Close examination revealed spinal cord at C2 and L6, were post-fixed that each of the astrocytic cell processes in Dalton's chrome-osmium at 4°C for 1-2 Received Nov. 7 , '69. Accepted Jan. 28, '70. MATERIALS AND METHODS ANAT. REC., 167: 303-308. 303 304 CEDRIC s. RAINE AND HENRYK W I ~ N I E W S K I Fig. 1 Rabbit spinal cord; subpial layers. Superficial astrocytic processes are covered by a basement membrane ( B M ) . Superficial cell membranes show periodic densifications and some processes have small desmosome-like junctions (arrows). Microtubules ( T ) can be discerned within astrocytic processes. Nonmyelinated fibers occur below. x 22,000. (less obviously the somata) contained compact bundles of filaments 90 A in diameter, each associated with accompanying discrete collections of 250 A microtubules (fig. 2). The latter were absent when osmium immersion fixation was used (fig. 3 ) . Microtubules were found within the subpial astrocytes of all three species studied and many displayed a central osmiophilic core with a diameter of approximately 40 A (figs. 2, 4 ) . Essentially, the dimensions of these microtubules were identical to those described in developing astrocytes (Peters and Vaughn, '67), and were ultrastructurally indistinguishable from normal neurotubules. The presence of the accompanying fibrillar bundles and desmosome-like junctions between cell processes facilitated the differentiation of small glial processes from non-myelinated nerve fibers (fig, 1 ) . In transverse section these filaments could be seen to contain a 20-25 A lumen (fig. 2). Microtubules, under the prevailing perfusion fixation conditions proved to be a common constituent of astrocytes throughout all the areas of the CNS studied. However, nowhere were microtubules as common and precisely arranged as within the subpial layers of the spinal cord. Those astrocytic processes seen to penetrate the spinal nerve root entry zones also contained bundles of both microtubules and filaments (Prineas et al., '69). Astrocytes situated within deeper layers of the spinal white matter, the anterior horns, and sub- MICROTUBULES WITHIN ASTROCYTES 305 Fig. 2 Rabbit spinal cord; subpial layers. A n astrocytic process is seen i n transverse section to contain filaments (some displaying a lumen), and several microtubules to the right. A central density can be seen i n one tubule (arrow). Desmosome-like contact at top. Glutaraldehyde perfusion. x 102,000. Fig. 3 Rabbit spinal cord; subpial layers. A n astrocytic process is seen to contain n o microtubules, only filaments. Osmium immersion. X 80,000. Fig. 4 Cat spinal cord; subpial layers. A n astrocytic process close to the soma contains many tubules, note central density (arrow), interspersed between filaments and mitochondria, Glutaraldehyde perfusion. x 40,000. 306 CEDRIC s. RAINE AND HENRYK WISNIEWSKI Fig. 5 Rabbit spinal cord; subependymal layem. A fibrous astrocyte contains scattered microtubules (arrows). Note tubules in adjacent astrocytic cell process above. x 30,000. ependymal layers contained scattered microtubules between bundles of filaments (fig. 5). The subpial astrocytic layer around the cerebral hemispheres was observed to comprise layers of fibrous astrocytic processes in which microtubules, though usually present, were not numerous. The cytoplasm of these cells was more electronlucent than their spinal cord counterparts, and contained glycogen granules. Oligodendrocytes with their denser cytoplasm and containing only microtubules, i.e., no filaments (Mugnaini and Walberg, ’64; Kruger and Maxwell, ’66; Raine and Bornstein, in press), were common and readily identifiable. Spinal cord tissue fixed by immersion in glutaraldehyde with or without subsequent post-osmication, while displaying many neurotubules, contained only scattered astrocytic microtubules between bundles of filaments. On the other hand, however, tissue fixed by osmic immersion revealed no astrocytic microtubules whatsoever (fig. 3). As a consequence of immersion fixation, general preservation was poorer and the astrocytes contained a “watery” cytoplasm, not evident after perfusion. DISCUSSION Under conditions of glutaraldehyde perfusion, microtubules are invariably found within fibrous astrocytes. There was, however, a marked variance in the microtubule populations encountered. Spinal cord subpial astrocytes contained more microtubules than astrocytes located elsewhere in the CNS. The higher frequency of microtubules within these subpial astrocytes compared with those from deeper layers of the cord, suggested either that there was a gradient of good fixation in these areas (i.e., the quality of good fixation decreased towards deeper layers), or more probably, that the subpial astrocyte is inherently endowed with a more elaborate system of microtubules. Such a system may serve some specific skeletal or physiological function in this area, more likely the former in view of the peripheral location of the cell and the fact that microtubules are known in many other instances to be an essential element in the maintenance MICROTUBULES WITHIN ASTROCYTES of form (Tilney and Byers, '69). Ultrastructurally, this cell type possessed both astrocytic (filaments) and oligodendrocytic (tubules) features. However, its proliferation activity in disease (Prineas et al., '69; Bunge et al., 'Sl), its fibrillar content, and its location, identified it as astrocytic. The "transitional" macroglial cell recognized by Bunge et al. ('61) and Bunge and Glass ('65), during their study on experimental remyelination, because of its subpial location and proliferative activity, was probably the same cell as that under discussion here. Its ability to partake in the process of remyelination was postulated by Bunge et al. ('61). In a more recent study on experimental remyelination in this region of the CNS, this cell type appeared not to be involved in the process of remyelination (Prineas et al., '69), and that function was carried out by reactive oligodendrocytes. The majority of reports on the ultrastructure of the subpial astrocytes of the CNS (using osmium or aldehyde fixation by immersion and perfusion) have failed to demonstrate the existence of microtubules in these cells (Lin et al., '60; Nelson et al., '61; Ramsey, '65; Waggener and Beggs, '67), although similar findings regarding the presence of a covering basement membrane and intercellular junctions were reported. Their occurrence, however, was briefly reported by Hirano et al. ('69), during a recent study on remyelination. Brightman and Reese ('69), have also demonstrated microtubules within astrocytes quite convincingly (fig. 4, loc. cit.) without drawing attention to the fact. Astrocytic microtubules are known to occur in the developing optic nerve (Peters and Vaughn, '67; Vaughn and Peters, '67) although the authors preferred to consider this organelle a rarity in the adult cell owing to its postulated transformation during development into filaments, and consequently de-emphasized its presence in the mature cell. The present study shows that within the mature CNS of the three species studied, microtubules were always evident, and were more common within subpial astrocytes than those from other sites within the CNS. 307 Microtubules have not been considered a normal component of astrocytic cytoplasm by previous workers who have contributed to macroglial cell classification (Farquhar and Hartman, '57; Schultz et al., '57; Palay et al., '62; Mugnaini and Walberg, '64; Wendell-Smith et al., '66; Kruger and Maxwell, '66). The present study demonstrates that like most mammalian cells, the astrocyte also contains microtubules. In spite of their paucity (or absence, in the case of osmium immersion) when perfusion fixation was not used, axoplasmic microtubules (neurotubules) could usually be found. It would appear therefore that astrocytic microtubules are more sensitive to the type of fixation employed and might indeed be different chemically from neurotubules. ACKNOWLEDGMENTS The authors thank Dr. Robert D. Terry for his advice throughout this study. The technical assistance of Karen Berkman, Cyrilla Ho, Larry Gonzalez and Sidney Gravney is gratefully acknowledged. This work was supported by grants NS 02255 and NS 03356 from the National Institutes of Health, and grant 600A-2 from the National Multiple Sclerosis Society. LITERATURE CITED Brightman, M. W.,and T. S. Reese 1969 Junctions between intimately apposed cell membranes in the vertebrate brain. J. Cell Biol., 40: 648-677. Bunge, M. B., R. P. Bunge and H. Ris 1961 Ultrastructural study of remyelination in an experimental les.ion in adult cat spinal cord. J. Biophys. Biochem. Cytol., 10: 67-94. Bunge, R. P., and P. M. Glass 1965 Some observations on myelin-glial relationships and on the etiology of the cerebrospinal fluid exchange lesion. Ann. N. Y. Acad. Sci., 122: 15-28. Farquhar, M. G., and J. F. Hartman 1957 Neuroglial structure and relationship as revealed by electron microscopy. J. Neuropath. Exp. Neurol., 16: 18-39. Hirano, A., H. M. Zimmerman and S. Levine 1969 Electron microscopic observations of peripheral myelin in a central nervous system lesion. Acta Neuropath. (Berlin), 12: 348-365. Kruger, L., and D. S. Maxwell 1966 Electron microscopy of oligodendrocytes in normal rat cerebrum. Am. J. Anat., 118: 411435. 1967 Comparative fine structure of vertebrate neuroglia: teleosts and reptiles. J. Comp. Neur., 129: 115-142. Lin, H A ,D. Duncan and R. S. Alexander 1960 A note on the subpial cytoplasm of the central 308 CEDRIC S . RAINE AND HENRYK WISNIEWSKI nervous system: a n electron microscopic study. Tex. Rep. Biol. Med., 18: 620-624. Mugnaini, E., and F. Walberg 1964 Ultrastructure of neuroglia. Ergebn. Anat. Entwickl. Gesch., 37: 194-236. Nelson, E., K. Blinzinger and H. Hager 1961 Electron microscopic observations on subarachnoid and perivascular spaces of the Syrian hamster brain. Neurology, 1I: 285-295. Palay, S. L., S. M. McGee-Russell, S. Gordon and M. A. Grillo 1962 Fixation of neural tissues for electron microscopy by perfusion with solutions of osmium tetroxide. J. Cell Biol., 12: 385-410. Peters, A., and J. E. Vaughn 1967 Microtubules and filaments in the axons and astrocytes of early postnatal rat optic nerves. J. Cell Biol., 32: 113-119. Prineas, J., C. S. Raine and H. Wisniewski 1969 A n ultrastructural study of experimental demyelination and remyelination. 111. Chronic EAE in the central nervous system. Lab. Invest., 21: 472484. Raine, C. S., and M. B. Bornstein 1970 Experimental allergic encephalomyelitis: an ultrastructural study of experimental demyelination in vitro. J. Neuropath. Exp. Neurol., in press. Raine, C . S., H. Wisniewski and J. Prineas 1969 A n ultrastructural study of experimental demyelination and remyelination. 11. Chronic EAE in the peripheral nervous system. Lab. Invest., 21: 316-328. Ramsey, H. J. 1965 Fine structure of the surface of the cerebral cortex of human brain. J. Cell Biol., 26: 323-333. Schultz, R. L., E. A. Maynard and D. C. Pease 1957 Electron microscopy of neurons and neuroglia of cerebral cortex and corpus callosum. Am. J. Anat., 100: 369407. Tilney, L. G., and B. Byers 1969 Studies on the microtubules in Heliozoa. V. Factors controlling the organization of microtubules in the axonenal pattern in Echinosphaerium (Actinosphaertum) nucleofilum. J. Cell Biol., 43: 148165. Vaughn, J. E.,and A. Peters 1967 Electron microscopy of the early postnatal development of fibrous astrocytes. Am. J. Anat., 121: 131-152. Venable, J. H., and R. Coggeshall 1965 A simplified lead citrate stain for use in electron microscopy. J. Cell Biol., 25: 407-408. Waggener, J. D.,and J. Beggs 1967 The membranous coverings of neural tissues: an electron microscopy study. J. Neuropath. Exp. Neurol., 26: 412-426. Wendell-Smith, C. P., M. J. Blunt and F. Baldwin 1966 The ultrastructural characterization of macroglial cell types. J. Comp. Neur., 127: 219-240. WiSniewski, H., J. Prineas and C. S. Raine 1969 An ultrastructural study of experimental demyelination and remyelination. I. Acute experimental allergic encephalomyelitis in the peripheral nervous system. Lab. Invest., 21: 105118.