Maintenance of isolated oligodendrocytes in long-term culture

Maintenance of isolated oligodendrocytes in long-term culture

Brain Research, 200 (1980) 151-164 © Elsevier/North-Holland Biomedical Press 151 M A I N T E N A N C E OF ISOLATED O L I G O D E N D R O C Y T E S I...

5MB Sizes 0 Downloads 61 Views

Brain Research, 200 (1980) 151-164 © Elsevier/North-Holland Biomedical Press

151

M A I N T E N A N C E OF ISOLATED O L I G O D E N D R O C Y T E S IN L O N G - T E R M CULTURE *

SARA SZUCHET, KARI STEFANSSON, ROBERT L. WOLLMANN, GLYN DAWSON and BARRY G. W. ARNASON Departments of Neurology, Pathology, Pediatrics and Biochemistry, The University of Chicago, Chicago, IlL 60637 (U.S.A.)

(Accepted May 8th, 1980) Key words: oligodendrocytes - - glial cell -- galactocerebroside - - sulfatide -- 2',Y-cyclicnucleotide-

3'-phosphodiesterase - - cell fractionation

SUMMARY A new procedure for isolating oligodendrocytes from ovine white matter is described. The method separates oligodendrocytes into two bands on a linear sucrose gradient. Five criteria have been employed to classify the separated cells. It is shown by indirect immunofluorescence with specific antisera that 97 ~ of the cells from both bands carry galactocerebroside, a specific surface marker for oligodendrocytes, on their plasma membranes and 95 ~ of the cells retain myelin basic protein as distinct patches on their surfaces. Isolated cells conform ultrastructurally to current concepts of oligodendrocytes. The cells incorporate [aH]galactose into galactocerebroside and carrier free H235SO4 into sulfatide, specific markers for oligodendrocytes. The specific activity of 2',3'-cyclic nucleotide-3'-phosphodiesterase in the two cell fractions is comparable to that reported for isolated oligodendrocytes by others. It is concluded that conservatively, 95 ~ of the cells in both fractions are oligodendrocytes. Cells from both bands survive in culture for months. In vitro the cells extend two or more processes, contain 'gliosomes', and surround themselves with extensive sheet-like membranes; i.e. they exhibit the morphological characteristics ascribed to oligodendrocytes in explant cultures. Conservatively 9 0 ~ of cultured cells stain with an antimyelin basic protein serum. The staining is localized in the cytoplasm and processes. The cells also stain with antigalactocerebroside and antioligodendrocyte sera. Cells remain differentiated for up to 70 days in vitro as evidenced by their incorporation of [aH]galactose and H235SO4 into galactosyl and sulfogalactosylceramide, respectively.

* A preliminary account of this work was presented at the 8th Annual Meeting of the Society for Neuroscience, Atlanta, Ga., November 2-6, 1979.

152 INTRODUCTION Oligodendrocytes synthesize and maintain central nervous system (CNS) myelin 10. In situ, oligodendrocytes vary both in size and in the extent to which they stain with OsO4 (osmiophilia). On the basis of ultrastructural studies of the corpus callosum of young rats, Mori and Leblond 9 have subdivided oligodendrocytes into "light', 'medium' and 'dark'. The 'light' cell is characterized by a large, pale nucleus, a central nucleolus and abundant cytoplasm rich in microtubules. The 'dark' cell has a small nucleus, often eccentrically located, and scanty cytoplasm. The medium cell is considered to be a transitional form between 'light" and 'dark'. It has been proposed that these 3 cell types constitute sequential stages in the maturation of oligodendrocytes, the 'dark' cell being the mature formiC, ~. All 3 types are believed to participate in the assembly and maintenance of myelin but their respective roles in this process are not understood. Long-term cultures of pure oligodendrocytes offer a unique model system i\~r studying myelin synthesis and for determining the role, if any, of neural cells other than oligodendrocytes in the synthesis of myelin precursors and in the process of myelination. Procedures for oligodendrocyte purification have been developed earlier3,15,2~ but long-term viability of the cells obtained has proven elusive. We have developed a procedure for isolating viable oligodendrocytes from white matter of ovine brain 27 e.~. With this procedure oligodendrocytes separate into two distinct bands. Here we present evidence that: (a) cells from both bands are oligodendrocytes according to currently accepted immunological 5,13,17,~s, uttrastructural 9,1°, and biochemical 3,11,lr~. 19 criteria; and (b) oligodendrocytes isolated by our procedure can survive in vitro for extended periods of time, remain differentiated, and synthesize redundant membranes. MATERIALS Reagents. Hank's balanced salt solution without Ca z+, Mg ~÷, NaCOaH and without phenol red (Hanks' BSS), fetal bovine serum (FBS), horse serum (HS), Dulbecco's modified Eagle's medium (DMEM) (all from GIBCO, Long Island, N.Y.); Ficoll 400 (Pharmacia, Piscataway, N.J.): soybean trypsin inhibitor Type IIS (Sigma, St. Louis, Mo.); trypsin (Worthington, Freehold, N.J.). Brains. Ovine brains from 3 to 6-month-old lambs, obtained at the abattoir immediately after death, are immersed in cold, sterile Hanks' BSS containing antibiotics (see below) and brought to the laboratory within one hour of death. METHODS Cell isolation. A slight modification of our previously described procedure 27-29 was used. Solutions: (1) 2 ~ (w/v) Ficoll in Hanks' BSS ('H-2~F'); (2) 0.1 ~ trypsin --0.02 To EDTA in 'H-2 ~F'. The concentration of trypsin may have to be adjusted for different brands; (3) trypsin inhibitor in 'H-2~F'. The concentration of this solution was varied according to the supplier's specification of activity so as to inactivate all the trypsin used; and (4) 0.9 M, 1.0 M, 1.2 M sucrose in half-strength Hanks' BSS [( ×/2)HI.

153 Solutions were adjusted to pH 6.00 -¢- 0.05, sterilized by filtration through 0.22 #m filters and cooled to 4 °C except the trypsin solution which was kept at room temperature. Ovine white matter (W.M.), finely minced, was suspended in 200 ml of 'H2 ~ F ' containing 0.1~ trypsin + 0.02~ EDTA and incubated at 37 °C in a metabolic water bath for 1.8 min/g W.M. Trypsin was inactivated by addition of trypsin inhibitor. The softened tissue was washed twice with 'H-2 ~ F ' , resuspended in (×/2)H-0.9 M sucrose and disrupted by passage through a series of nylon and stainless steel screens of decreasing pore size from 350 #m down to 30 #m. The crude cell suspension was centrifuged at 2100 rpm (850 g) for 10 rain. During this step myelin floats to the top of the tube while the cells form a pellet (P1). P1 was suspended in 3-4 ml of (×/2)H-0.9 M sucrose, applied on a linear sucrose gradient from 1.0-1.2 M and centrifuged at 1500 rpm (431 g) for 20 min. Three bands separate on this gradient; these are labeled from top to bottom: Bands I, II and III. Band I contains red blood cells and astrocytes; cells from Bands II and III were used for the experiments described here. The cells in each band were separated, diluted 1/4 by slow addition of Hanks' BSS and centrifuged at 433 g for 10 min. The yield varies from one experiment to another and is particularly dependent on the age of the brains. Average yield is 3 × 106 cells/g wt W.M. Antisera. The antigalactocerebroside serum used was a generous gift from Dr. Maurice Rapport. The specificity of this rabbit antiserum is established21. A guinea pig anti-lamb myelin basic protein serum was obtained in our laboratory (Dr. A. B. Noronha) using a pure preparation of myelin basic protein (single band on acrylamide gel electrophoresis). The specificity of this antiserum for myelin was assessed by staining fixed tissue sections of lamb brains using the peroxidase-antiperoxidase techniquesaS. Rabbit antibovine glial fibrillary acidic protein (GFAP) was kindly supplied by Dr. L. F. Eng. Again, staining of tissue sections from lamb brains and sections from multiple sclerosis plaques which contain many fibrous astrocytes proved the serum to be specific for astrocytes. Antioligodendrocyte serum was obtained (Dr. R. Roos) by injecting freshly isolated cells in complete Freund's adjuvant into the foot pads of a guinea pig followed by a boost with cells in complete Freund's adjuvant given over the sternum. The antiserum, checked by indirect immunofluorescence or by the peroxidase-antiperoxidase method ~5 on lamb brain tissue sections was positive against oligodendrocytes up to a dilution of 1:500. No neurons or astrocytes stained, though some positive cells could not be definitely identified as either oligodendrocytes or astrocytes. Capillaries gave a positive reaction. The antiserum failed to react with rat, hamster, guinea pig or human brain tissue sections nor did it react with lamb liver, kidney or heart tissue sections. Immunocytochemical staining. Freshly isolated cells were spun down on glass coverslips at 300 g using a plate carrier (Dynotech Laboratories); cells stuck tightly to the coverslips. Staining with antisera was performed according to a modified indirect immunofluorescence techniqueZ5. We used the following dilutions: antimyelin-basic protein, 1:40; antigalactocerebroside, 1:100; antioligodendrocyte, 1:30; and antiGFAP, 1:100. Dilutions were made with 10 ~ normal goat serum. Pre-immune sera and sera from other normal animals at twice the concentrations employed for the

154

antisera were used as controls. As a first step cells were incubated with normal goat serum at room temperature for 15 rain after which time coverslips were blotted with a filter paper but not washed. This was followed by application of primary antiserum for 30 rain, also at room temperature. Cells were next washed 5 times with phosphate-saline buffer (PBS) and a second application of normal goat serum for 15 min followed. Fluorescein isothiocyanate conjugated goat antiserum to guinea pig immunoglobulins or rabbit immunoglobulins (Cappel Laboratories) was applied for 30 rain followed by 5 washes with PBS. Cells were mounted on glass slides and examined with a Leitz Ortholux Fluorescence Microscope. Microscopy. A drop of the cell suspension was mixed with a drop of FBS, smears made, fixed with methanol, stained with Giemsa (Fisher-Scientific) and examined by light microscopy. For electron microscopy cells from each band were centrifuged at 2100 rpm (845 g) for 10 rain to form pellets which were fixed in 2 % glutaraldehyde-2 '!~, paraformaldehyde in 0.1 M sodium phosphate buffer pH 7.2 for 15 h, post-fixed in 1 . 5 ~ OsO4, dehydrated with alcohols and embedded in Epon. Thin sections were stained with lead citrate and uranyl acetate. 2',Y-cyclic nucleotide-3'-phosphodiesterase(CNP). The method of Kurihara and Tsukada as modified by Sprinkle et al. 24 was employed. [3H-8]Adenosine T-cyclic monophosphate (Amersham, Searle) was used as substrate and was separated from the product by thin layer chromatography. Biosynthesis of complex carbohydrates. The method described by Dawson .) was used. Cells are cultured in the presence of H2aaSO4 and [3H]galactose for 60 h, harvested mechanically with a rubber policeman, sonicated (Model 185 W Sonifier, Heat Systems Ultrasonic, Plainview, Long Island, N.Y.), and an aliquot removed for protein determination by the method of Lowry 6. To the remaining suspension, chloroform-methanol (2:1 v/v) was added, the biphasic mixture was vortexed~

Fig. 1. Morphologic and immunocytochemical characterization of freshly isolated cells obtained from ovine white matter. A : representative preparation of cells from Band IIl stained with Giemsa. Magnification : 2000 ;<. B : photomicrograph of freshly isolated and unfixed cells treated with rabbit antigalactocerebroside serum followed by fluorescein isothiocyanate conjugated goat anti-rabbit immunoglobulin. Note patches of fluorescence on a background of faint homogeneous staining. The patchiness may in part reflect membrane damage. Magnification : 1250 ×. C: photomicrograph of freshly isolated and unfixed cells stained with guinea pig antimyelin basic protein followed by fluoresceinlabeled goat anti-guinea pig immunoglobulins. Staining is concentrated in a limited number of clumps on the cell surface. Magnification: 1250 ,:.

155 centrifuged at 600 g for 5 min, upper and lower phases separated and each evaporated to dryness. Authentic sulfogalactosylceramide was added to the lower phase and labeled sulfogalactosylceramide was isolated by silicic acid and thin layer chromatography. Gangliosides were first fractionated on superfine Sephadex G-25 using chloroform-methanol-water (120:60:9) as eluent, and then separated on thin layer chromatography. Cell culture. Cells were washed once with DMEM supplemented with 125 #g/ml of Amphotericin B, 20/~g/ml of Gentamycin and containing 20 ~o of either FBS or HS, centrifuged, resuspended in the same medium and cell concentration adjusted to 1.0 × 106 cells/ml. For long-term culture and for studies of incorporation of labeled precursors precoated plastic Petri dishes (Falcon, Cockeysville, Md.) were used. For morphological and immunochemical studies glass chamber slides (Lab-Tek, Naperville, I11.) were preferred. Plating density was 1.7 × 105 cells/sq, cm. Cells were incubated at 37 °C with 5 ~ CO2 and humidity at saturation. RESULTS

Identification of freshly isolated cells A representative Giemsa-stained smear of freshly isolated cells from Band III is shown in Fig. 1A. The cells vary somewhat in size but the density and smooth contour of the nuclei are readily appreciated. Cells from Band II and Band III looked basically alike. Immunological identification Several workersS,1a,17,18 have suggested that galactocerebroside is a specific surface marker for oligodendrocytes. We stained freshly isolated cells from Band II and from Band III with antiserum against galactocerebroside. 97 ~ of cells in both bands stained positively. The staining appeared as bright patches scattered over the surface of the cells (Fig. 1B). Weak staining between patches was also observed. No staining was seen when control sera were used. Freshly isolated cells were also examined for the presence of myelin remnants (see above) using an antimyelin basic protein serum. 95 ~o of the cells stained. Again, the staining pattern was patchy; the patches were larger but fewer than those observed with antigalactocerebroside (Fig. 1C). Cells also stained with an antioligodendrocyte serum but did not fluoresce with anti-GFAP, a specific marker for astrocytesT,2L Ultrastructural characterization The majority of cells from both bands were well preserved. Cytoplasmic and nuclear staining ranged from light to very dark (Figs. 2A and B). The cells varied in the amount of cytoplasm they contained; some had a rim of cytoplasm surrounding a centrally located nucleus; others had extensive cytoplasm at one pole with the nucleus at the other. The pattern of chromatin clumping was similar to that seen in oligodendrocytes in situ. Microtubules were relatively difficult to find in the isolated cells. Since

156 we detected a large number of microtubules in oligodendrocytes of starting white matter, it is likely that they dissociated during isolation. We have adopted a set of minimal criteria in order to identify oligodendrocytes. M a n y cells had myelin continuous with (not merely adherent to) the plasma m e m b r a n e (Fig. 3): this by itself identified such cells as oligodendrocytes. When

Fig. 2. Cross-sections of pellets of cells from a sucrose gradient. Magnification : 3800 ,~. A : cells from Band II. B : cells from Band II1. Observe variability of nuclear and cytoplasmic osmiophilia in both cell subpopulations.

157

Fig. 3. Electronmicrograph of a group of cells from Band III at high magnification (24,000 x ). Cells are osmiophilic; the nuclear and plasma membranes appear intact. Single arrows point to membrane fragments assumed to represent sites from which cell processes have been removed during isolation. Double arrows show loops of myelin attached to one of the cells. attached myelin was not seen, at least two of the following criteria had to be satisfied simultaneously before a cell was called an oligodendrocyte: multiple layers o f r o u g h endoplasmic reticulum, large numbers o f microtubules, intense osmiophilia and fragments or loops o f plasma m e m b r a n e projecting f r o m the cell surface (presumed to

158 be sites f r o m which cell processes were removed during isolation) 3°. Astrocytes were identified by the presence o f 7-9 nm cytoplasmic filaments and lightly stained cytoplasm 9,1°. Applying these criteria to 4 cell preparations we identified 65 ~: 14 ~o o f cells in Band II and 91 ~_ 3 ~ o f cells in Band I I I as oligodendrocytes. Analysis of the frequency distributions revealed that the n u m b e r of microtubules and rough endoplasmic reticulum was approximately the same in oligodendrocytes o f Bands II and III. We found a slightly larger n u m b e r o f cells bearing attached myelin in Band 11 than in Band I l I but the extent o f osmiophilia and the occurrence o f m e m b r a n e fragments extending f r o m the plasma m e m b r a n e was significantly higher for oligodendrocytes in Band I I I than in Band 11. Cells which failed to meet the criteria listed above were considered unidentified. These latter cells had light cytoplasm, little endoplasmic reticulum, scattered ribosomal clusters and no distinguishing nuclear or cytoplasmic characteristics. When the same classification system was applied to starting white matter fixed by immersion, 70 ~ o f cells were found to be oligodendrocytes, 13 ~ were astrocytes and l 7 o/~,could not be classified. Biochemical identification

A distinctive feature o f oligodendrocytes is their ability to synthesize sulfogalactosylceramide (sulfatide), a c o m p o n e n t o f myelin 2,5,14. We tested the ability o f isolated cells from Bands II and III to make sulfatide. When cells from these bands are grown as m o n o l a y e r cultures in the presence o f [3H]galactose and H235SO4 they incorporate the isotopes preferentially into galactosylceramide and sulfagalactosylceramide, respectively (Table 1). The cells also synthesize gangliosides, particularly those o f the G m

TABLE I Biosynthesis" o f complex carbohydrates by lamb oligodendroeytes in tissue culture

Cells labeled with 5/tCi [aH]Gal and 150/~Ci of H2zaSO4are grown as monolayer cultures in modified Eagle's medium supplemented with 10% fetal bovine serum for 60 h. Most of the cells extruded processes during this time. Abbreviations: Gal, galactose; Cer, ceramide; Lac, lactoside; Tetrahex, tetrahexoside; LCFA, long-chain fatty acid; aHOFA, a-hydroxyl fatty acid; G, ganglioside. Compound

GalCer LCFA GalCer aHOFA LacCer Sulfo-GalCer Tetrahex-Cer GM3 GM2 GM1 Gma/GD3 "GDIb"

Band I1

Band I11

. :'H ;

~asS/' /;~H / (cpm/mg protehl)

.':~5S /

2200 1660 300 1350 200 800 120 130 2000 1400

---1750

---3400 ---------

-------

3200 3000 480 2500 220 1000 130 140 1920 1440

159 /GDa region (Table I) which have been reported t2 to be components ofoligodendrocyte membrane. High activity of CNP enzyme is customarily taken as another marker for oligodendrocytesll,l~,ls, 19. The specific activity of CNP expressed as moles of 2'AMP formed per min/mg protein found in our cells immediately after isolation was 1.9 ± 0.4 and 2.0 4- 0.6 for cells from Band II and Band III, respectively. Starting white matter had an activity of 11 4- 3. These values are in good agreement with those obtained by others 11.

Culture of cellsfrom Bands H and 111 The in vitro behavior of the cells depended on experimental conditions and in particular on the media used. Even in the same medium we observed variations from one preparation to the next. Sometimes the cells attached within 24 h and started to extend processes; at other times, the cells formed aggregates which floated for 4-6 days and then attached, either singly or as a group. In either case, the ultimate

Fig. 4. Phase-contast micrograph of live oligodendrocytes grown on plastic after 15 days in culture. Arrow points to cell with 4 processesflaring up into membranous sheets. This cellfits the conceptualized model of an oligodendrocyte (see text). Magnification: 625 ×.

160

k

t

r

Fig. 5 Morphological features of cultured oligodendrocytes from Bands I1 and 111. A, B and ( : 21 days in culture. Cells were grown in DMEM-20 ~ HS on glass chamber slides. Medium was removed, cells washed 3 times with PBS, fixed with methanol for 2-4 rain and stained with Giemsa. A: typical appearance of cultures. Small arrows point to bipolar cells; large arrows show multipolar cells; double arrows indicate cell aggregates. Magnification: 280 ~,~. B: groups of cells from A at higher magnification : 1100 -. Small arrows points to gliosomes; large arrow indicates network of processes. C : another group of cells from A at higher magnification: 1100 :~. Small arrow shows cells resembling 'mossy" oligodendrocytes (see text); large and double arrows indicate sheet-like membrane at the foot of a process and along processes, respectively. D: phase-contrast micrograph of an unfixed cell grown on a plastic surface for 14 days. Note extensive membranes. Magnification: 1000 .

e v o l u t i o n o f the c u l t u r e s a p p e a r e d

to be the same. V i a b l e cells p e r c u l t u r e did n o t

c h a n g e a p p r e c i a b l y f r o m d a y 1 to 10. A t t a c h e d cells w h e n s t a i n e d with G i e m s a after 24 h in c u l t u r e w e r e m o r p h o l o g i c a l l y i n d i s t i n g u i s h a b l e f r o m freshly i s o l a t e d cells, b u t a f t e r a w e e k in c u l t u r e altered m o r p h o l o g y was e v i d e n t . N o t e w o r t h y was the v a r i a b i l i t y in cell s h a p e a n d size, a n d in the n u m b e r o f p r o c e s s e s as well as t h e i r w i d t h

161 and length. The most remarkable feature of these cells is their ability to synthesize extensive sheet-like membranes. These are found at the end of processes in monopolar, bipolar and multipolar cells or surrounding the cell body. Fig. 4 illustrates the appearance, after 15 days in culture, of a multipolar cell with processes flaring up into membranous sheets. This cell is reminiscent of the conceptualized model for oligodendrocytes proposed by Bunge (cf. Fig. 6, ref. 1). Other characteristics of cultured oligodendrocytes are presented in Fig. 5 where Giemsa-stained cells, after 21 days in culture, are shown. Note bipolar as well as multipolar cells (Fig. 5A). Often cells form small groups interlaced with processes which exhibit the bead-like swellings known as 'gliosomes' (Fig. 5B) first described by Wolfgram and Rose 81 as being peculiar to oligodendrocytes in explant culture. Cells resembling the mossy oligodendrocytes described by Ramon-Moliner20 and found by him only in young animals are seen in Fig. 5C. Extensive membranes sometimes covering surface areas 10-20 times that of the freshly isolated cells are also frequently observed as illustrated in Fig. 5D. To determine whether cultured cells retain the surface properties of freshly isolated cells we stained cells after 11-15 days in culture with antigalactocerebroside and antioligodendrocyte sera. 75 ~ of the cells reacted positively with each antisera. At 21 days in culture more than 90700 of the cells and their processes stained with antimyelin basic protein serum but in contrast to freshly isolated cells, the fluorescence was cytoplasmic, i.e. only fixed cells stained (K.S. et al., unpublished observations). We have also stained cells after 7, 14 and 29 days in culture with anti-GFAP serum. None of the cells gave a positive reaction with this serum. Cells kept in culture for up to 90 days incorporate labeled [aH]galactose and H2~5SO4 into galactocerebrosides and sulfatide. This incorporation reaches a maximum 10-12 times that measured for freshly isolated cells at 20-30 days in culture and then gradually decreases to a relatively low level at 70-90 days in culture8. DISCUSSION In this communication we describe a method for isolation of oligodendrocytes which survive in vitro for extended periods.

Identification of fresh(F isolated cells Five criteria were employed to classify cells from Bands II and III. First, on ultrastructural examination 65 ~ of cells in Band II and 91 ~o of cells in Band III were identified as oligodendrocytes when rigid criteria for classification were used. Second, we demonstrated that 97 ~ of cells from both bands carry galactocerebroside on their surface as evidenced by their positive staining with antigalactocerebroside serum. Galactocerebroside is a specific marker for oligodendrocytess,la,17,1s. Third, we showed by indirect immunofluorescence that 95 ~ of the cells contain patches of myelin on their surface since they stain for the basic protein of myelin. This finding points to the presence of myelin fragments on the cells. These could be continuous with the plasma membrane of the cells; alternatively and/or additionally myelin debris could have adsorbed to cell surfaces during the isolation procedure. The former has been de-

162 monstrated directly by electron microscopy (Fig. 3) but some contribution by the latter to the immunofluorescence results cannot be excluded. Fourth, the specific activity of CNP 12,15,19 found in our cell preparations agrees with that reported by others ~l for freshly isolated oligodendrocytes. Fifth, cells from Bands II and Ill express one of the functions of differentiated oligodendrocytes, namely, the synthesis of sulfatides. Taking these various criteria in total, we conclude that more than 95 o~; of freshly isolated cells from both Bands lI and Ill are in fact oligodendrocytes. A number of cells from Band lI which varied from 16 '~ to 48 ",~ (average 35 o~,) in the 4 preparations analyzed failed to meet our ultrastructural guidelines for classification as oligodendrocytes even though they satisfied the immunological criteria used. This points, in our opinion, to the relative lack of sensitivity of ultrastructural analysis when applied to isolated cells stripped of their processes. It is to be noted that synthesis of sulfogatactosylceramide, expressed as cpm/mg protein, by our cells shortly after isolation is higher than that reported by other workers for isolated bovine and lamb oligodendrocytes 14 and for two mouse oligodendroglioma cell lines 2. Note also that the essentially equimolar ratio of long-chain fatty acids to a-hydroxy fatty acids in galactosylceramide (Table i) is close to that found in myelin and in whole brain. These cells also synthesize gangliosides. Amongst the latter the GD series predominates: GD gangliosides have been reported as present in oligodendrocyte plasma membranes 12. Thus, active synthesis of gangliosides may relate to the extensive development of plasma membranes observed morphologically.

Properties of maintained oligodendroeytes Morphologically, oligodendrocytes from Bands I[ and lII maintained in vitro exhibit most of the characteristics ascribed to these cells in explant cultures 4,2°,3~. We find polar and multipolar cells which have gliosomes and sheet-like membranes; we even find cells which resemble 'mossy' oligodendrocytes (Fig. 5C). Furthermore, maintained cells continue to be recognized by antioligodendrocyte sera and retain galactocerebroside on their membranes. Cells kept in culture incorporate [~H]galactose and H235SO4 into galactosylceramide and sulfogalactosylceramides, respectively, two galactolipids specific for oligodendrocytes. The ceils also synthesize myelin basic protein. Thus, traits associated with oligodendrocytes can be recognized in cells which have been kept in vitro for 2-3 months, and we conclude that among growing cells in culture at this time, oligodendrocytes predominate. Clearly, direct demonstration of myelin synthesis by maintained cells would provide the most telling proof of cell identity but this has not been achieved to date and may require interactions between oligodendrocytes and other cell types not present in our preparations. Experience with peripheral nerve cultures suggests that this is likely to be the case. Differences and similarities between cells from Band H and Band HI As reported elsewhere, cells in Band III are larger on average than those in Band I129. The cells in these bands also differ to a limited extent in their ability to synthesize galactocerebroside and sulfatides (Table I). We have not been able to define rigid

163 ultrastructural differences between cells f r o m Bands I I and I I I and their relation to the light, m e d i u m and dark subpopulations described by M o r i and Leblond 9 remains uncertain. Nevertheless, cells in Band I I I are darker on the average (i.e. more osmiophilic) than those in Band II. Morphological differences between ceils f r o m Bands II and I l I in culture have yet to be found. ACKNOWLEDGEMENTS We are grateful to Drs. A.B.C. N o r o n h a , M. R a p p o r t and R. R o o s for gifts o f specific antisera. To Mr. P.E. Polak, Ms. A n n Speckman and Mr. S.M. Kernes we express our thanks for excellent technical assistance. This research was supported by grants f r o m : N a t i o n a l Multiple Sclerosis Society, nos. R G 1223-A-2 (S.S.) and RGl130-B-15 (B.G.W,A.); N I N C D S , no. NS13526-03 (B.G.W.A.), U S P H S - H D - 0 6 4 2 6 , H D 0 9 4 0 2 and NS-00029 (G.D.). S.S. was the recipient o f a Senior Fellowship f r o m the National Multiple Sclerosis Society and K.S. acknowledges a Fellowship f r o m the Icelandic Science Foundation.

R EFERENCES 1 Bunge, R. P., Structure and function of neuroglia: some recent observations. In F. O. Schmitt (Ed.), The Neurosciences: Second Study Program, Rockefeller University Press, New York, 1970, pp. 732-797. 2 Dawson, G., Synthesis of myelin glycosphyngolipids (galactosylceramide and galactosyl (3-0Sulfate) ceramide (sulfatide)) by cloned cell lines derived from mouse neurotumors, J. biol. Chem., 252 (1977) 2777-2779. 3 Fewster, M. E., Ihrig, T. J. and Blackstone, S. C., The preparation and characterization of isolated oligodendroglia from bovine white matter, Brain Research, 18 (1973) 263-271. 4 Korinkova, P. and Lodin, Z., A transitional differentiation of glial cells of cultured corpus callosum caused by dibutyril cyclic adenosive monophosphate, Neuroscience, 2 (1977) 1103-1114. 5 Lisak, R. P., Abramsky, O., Dorfman, S. H., George, J., Manning, G. C., Plesaure, D. E., Saida, T. and Silberberg, D. H., Antibodies to galactocerebroside bind to oligodendroglia in suspension culture, J. neurol. Sci., 40 (1979) 65-73. 6 Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall R, J., Protein measurement with the Folin phenol reagent, J. biol. Chem., 193 (1951) 265-275. 7 Ludwin, S. K., Kosek, J. C. and Eng, L. F., The topographical distribution of S-100 and GFA proteins in the adult rat brain: an immunohistochemical study using horseradish peroxidaselabeled antibodies, J. comp. Neurol., 65 (1976) 197-263. 8 Mack, S. R. and Szuchet, S., Differentiated functions of oligodendrocytes in long-term culture. Abstr., Amer. Soc. Neurochemistry, llth Annual Meet., Houston, Texas, 1980, 104 pp. 9 Mori, S. and Leblond, C. P., Electron microscopic identification of three classes of oligodendrocytes and a preliminary study of their proliferative activity in the corpus callosum of young rats, J. comp. Neurol., 139 (1970) 1-30. 10 Peters, A., Palay, S. L. and Webster, H. F., The Fine Structure of the Nervous System, W. B. Saunders, Philadelphia, Pa., 1976, pp. 232-263. 11 Pleasure, D., Abramsky, O., Silberberg, D., Qwinn, B. Parkis, J. and Saida, T., Lipid synthesis by an oligodendroglial fraction in suspension culture, Brain Research, 134 (1977) 377-382. 12 Poduslo, S. E., The isolation and characterization of a plasma membrane and a myelin fraction derived from oligodendroglia of calf brain, Neurochemistry, 24 (1975) 647-654. 13 Poduslo, S. E., Studies on isolated, maintained oligodendroglia: biochemistry metabolism, and in vitro myelin synthesis. In J. Palo (Ed.), Myelination and Demyelination, Plenum Press, New York, N.Y., 1978, pp. 71-94.

164 14 Poduslo, S. E. and McKhann, G. M., Synthesis ofcerebrosides by intact oligodendroglia maintained in culture, Neurosci. Lett., 5 (1977) 159-163. 15 Poduslo, S. E. and Norton, W. T., Isolation and some chemical properties of oligodendroglia from calf brain, J. Neurochem., 19 (1972) 727-736. 16 Privat, A. and Fulcrand, J., Neuroglia - - from the subventricular precursor to the mature cell. In S. Federoff and L. Hertz (Eds.), Cell, Tissue and Organ Culture in Neurobiology, Academic Press. New York, N.Y., 1978, pp. 11 -37. 17 Raft, M. C., Fields, K. L., Hakomori, S. I., Mirsky, R., Pruss, R. M. and Winter, J., Cell-type specific markers for distinguishing and studying neurons and major classes of glial cells in culture, Bruin Research, 174 (I 979) 283--308. 18 Raft, M. C., Mirsky, R., Fields, K. L., Lisak, R. P., Dorfman, S. H., Silberberg, D. H., Greoson, N. A., Leibowitz, S. and Kennedy, M. C., Galactocerebroside is a specific cell-surface antigenic marker for oligodendrocytes in culture, Nature (Lond.), 274 (1978) 813-816. 19 Raine, C. S., Poduslo, S. E. and Norton, W. T., The ultrastructure of purified preparations of neurons and glial ceils, Brain Research, 27 (1971) 11-24. 20 Ramon-Moliner, E., A study on neuroglia, J. comp. NeuroL, 110 (1958) 157 -171. 21 Rapport, M. M., Graf, L., Autilio, L. A. and Norton, W. T., Immunochemical studies of organ and tumor lipids, J. Neurachem., 1 l (1964) 855-864. 22 Schachner, M., Hedley-Whyte, E. T., Hsu, D. W., Schoonmaker, G. and Bignami, A., Ultrastructural localization of glial fibrillary acidic protein in mouse cerebellum by immunoperoxidase labeling, J. Cell Biol., 75 (1977) 67-73. 23 Snyder, D. S., Raine, C. S., Farooq, M. and Norton, W. T., Exploration of new methods for bulk isolation of oligodendroglia. Abstr., Amer. Soc. Neurochenl., 9th Annual Meeting, Wash., D.C., 1977, pp. 64. 24 Sprinkle, T. J., Zaruba, M. E. and McKhann, G. M., Radioactive measurement of 2',3'-cyclic nucleotide 3'-phosphodiesterase activity in the central and peripheral nervous system and in extraneural tissue, Analyt. Biochem., 88 (1978) 449-456. 25 Sternberger, L. A., In lmmunoehemistry, 2rid Edn., John Wiley, New York, N.Y., 1979. 26 Sturrock, R. R., Light microscopic identification of immature glial cells in semi-thin sections of the developing mouse corpus callosum, J. Anat. (Lond.), 122 (1976) 521-537. 27 Szuchet, S., Arnason, B. G. W. and Polak, P. E., A new method for oligodendrocyte isolation, Biophys. J., 21 (1978) 51a. 28 Szuchet, S., Arnason, B. G. W. and Polak, P. E., Separation of ovine oligodendrocytes into two bands on a linear sucrose gradient, J. neurosci. Meth., (1980) in press. 29 Szuchet, S. and Stephansson, K., In vitro behavior of oligodendrocytes. In S. Fedoroff and L. Hertz (Eds.), Advances in Cellular Neurobiology, Academic Press, New York, N.Y., 1980, in press. 30 Szuchet, S., Wollmann, R. L. and Arnason, B. G. W., UItrastructural characterization of isolated oligodendrocytes, J. Neuropath. exp. Neurol., 37 (1978) 349. 31 Wolfgram, F. and Rose, A., The morphology of neuroglia in tissue culture with comparisons to histological preparations, J. Neuropath. exp. NeuroL, 16 (1957) 514-53 I.