Journal of Neuroscience Research 30:531-539 (1991)
Establishment of Highly Enriched Type-2 Astrocyte Cultures and Quantitative Determination of Intense Glutamine Synthetase Activity in These Cells B.H.J. Juurlink and L. Hertz Departments of Anatomy (B.H.J.J.), pharmacology (L.H.), and Anaesthesia (L.H.), University of Saskatchewan, Saskatoon, Canada We report procedures that allow one to develop and maintain cultures highly enriched in rat neopallial type-2 astrocytes. Even after four weeks such cultures consist of more than 90% type-2 astrocytes, -5% 0-2A progenitors and fewer than 2% type-1 astrocytes. Their survival for more than 5 days requires the addition of conditioned medium from type-1 astrocyte cultures. The type-2 astrocytes have an intense glutamine synthetase activity whose basal level is sevenfold higher than in type-1 astrocytes. The glutamine synthetase activities of both the type-2 and type-1 astrocytes are increased after exposure to cortisol. Thus, type-2 astrocytes express the two quintessential astrocytic features: glial fibrillary acidic protein (previously reported by others) and glutamine synthetase.
1986; Levi et al., 1986; Trotter and Schachner, 1989) and from cerebrum (Aloisi et al., 1988; Behar et al., 1988; Ingraham and McCarthy, 1989). Much information is now available about type-l astrocytes because they have been available in highly enriched primary cultures for more than 15 years (Juurlink and Hertz, 1985). Less information is available about type-2 astrocytes, since highly enriched (>90%) long-term (>1 week) cultures of these cells have not yet been available. The type-2 astrocytes exhibit a number of properties that traditionally have been considered neuronal and not astrocytic. For example, type-2 astrocytes avidly accumulate y-aminobutyric acid (GABA) (Johnstone et al., 1986; Levi et al., 1986; Barres et al., 1990) but, in contrast to type-1 astrocytes (Schousboe et al., 1977; Bardakjian et al., 1979), they do not appear to metabolize it readily; rather, type-2 astrocytes release GABA Key words: glia, neopallium, 0-2A lineage, rat, tis- upon stimulation with glutamate or glutamate analogues sue culture (Gallo et al., 1989). Similar to GABAergic neurons and unlike type-1 astrocytes, type-2 astrocytes as well as their 0-2A progenitors synthesize GABA (Barres et al., INTRODUCTION 1990). Furthermore, type-2 astrocytes have a “neuLess than 10 years ago, Raff and his colleagues ronal” channel phenotype (Barres et al., 1988). (Raff et al., 1983a,b) demonstrated that there are two One criterion that classifies type-2 astrocytes as distinct astrocyte types in cultures established from optic astrocytic in nature is the presence of the intermediate nerve that differ from each other in lineage origins, plas- filament protein glial fibrillary acidic protein (GFAP). malemmal ganglioside composition, morphology, and GFAP was the first described molecular marker of astroresponse to epidermal growth factor. One, termed the cytes (Eng et al., 1971; Bignami and Dahl, 1977; type-1 astrocyte was a large flat cell that resembled in Schachner et al., 1977). A second characteristic of morphology and other properties the major population of astrocytes is the enzyme glutamine synthetase. Both astrocytes in cultures established from other brain re- immunohistochemical (Norenberg and Martinez-Hergions such as the cerebral hemispheres (e.g., Hertz et nandez, 1979) and molecular biological (Mearow et al., al., 1982). The other, termed the type-2 astrocyte arose from a glial precursor, the 0-2A progenitor, that could differentiate into either an oligodendrocyteor into a type- Received December 6, 1990; revised March 13, 1991;accepted March 2 astrocyte, depending upon culture environment 25, 1991. (reivewed in Raff, 1989). 0-2A progenitors and their Address reprint requests to Dr. B.H.J. Juurlink, Department of Anattype-2 astrocyte derivatives have also been described in omy, University of Saskatchewan, Saskatoon, Saskatchewan, Canada cultures established from cerebellum (Johnstone et al., S7N OWO. 0 1991 Wiley-Liss, Inc.
532
Juurlink and Hertz
1989) techniques have been used to demonstrate that the enzyme glutamine synthetase is restricted to astroglial cells within the central nervous system (CNS). Several recent reports, however, suggest that some glutamine synthetase activity is also present in oligodendrocytes (Warringa et al., 1988; Cammer, 1990; D'Amelia et al., 1990). Type-2-like astrocytes also have immunocytochemically detectable glutamine synthetase (Raff et al., 1983a; B .H.J. Juurlink, unpublished observations); however, no quantitative information is available regarding glutamine synthetase activity in this class of astrocyte, since relatively homogeneous cultures of type-2 astrocytes have only relatively recently been obtained (Ingraham and McCarthy, 1989), and long-term cultures of highly enriched type-2 astrocytes do not appear to be available. In this report, we describe procedures to obtain cultures highly enriched (>90%) in type-2 astrocytes that can be maintained at this purity for at least 4 weeks. Such type-2 astrocyte cultures were found to display intense glutamine synthetase activity.
PKIMARY ASTROCYTE CULTURES SHAKEN ON ROTARY SHAKER A'l' 37" C FOR 30 MIN
L REMOVE MEDIUM WITH SUSPENDED CELLS
I RINSE CULTURES AND ADD FRESH MEDIUM BUFFERED WITH GLYCEROPHOSPHATE SHAKE OVERNIGHT (- i a HR)
0-
i REMOVE SUSPENDED CELLS AND PLACE IN I 0 0 mm PETRI DISHES FOR A 3 HR INCUBATION AT 37" C IN A HUMIDIFIED AIR ATMOSPHERE
1 REMOVE MEDIUM AND GENTLY WASH PETRI DISHES TO REMOVE DEBRIS
I REMOVE THE LOOSELY ADHERENT 0-2A LINEAGE CELLS BY OENI'I,'? PIPE'TTING THE SURFACE OF THE DISHES WITH A BICARBONATE BUFFERED GROWTH MEDIUM CONTAINING 0 02% DN=c
I CENTRIFUGE CELLS AT 18Og FOR 10 MIN AND RESUSPEND IN GROW YH MEDIUM CONTAINING 0 02% DNaSe
L CENTRIFUGE CELLS AT 15Og FOR 15 MIN THROUGH A 3 5% BSA CUSHION
I RESUSPEND CELLS IN GROWTH MEDIUM CONTAINING 0 02% DNaSe AVD FILTER SEQUENTIALLY THROUGH NYLON MESHES OF PORE SIZES 3 5 p AND 15pm
I PLANT CELLS IN GROWTH MEDIUM CONTAINING 0 02% DNaSe ONTO POLYORNITHINE COATED SURFACES
i AFTER 15 MIN REMOVE MEDIUM AND ADD DNase FREE GROWTH MEDl UM
MATERIALS AND METHODS Culture Preparation The neopallia (i.e., those portions of the cerebral hemispheres that remain after removal of the olfactory bulbs, hippocampi, and basal ganglia) were aseptically separated from cerebral hemispheres obtained from young Wistar rat pups ranging in age from newborn (PO) to postnatal day 5 (P5).Meninges were removed, and the neopallia were diced into cubes of -1 mm3 and gently pushed through a nylon mesh with a pore size of 75 pm. The cells obtained from two pups were planted into three 75-cm2 (Falcon) tissue culture flasks, i.e., 30,000 viable [nigrosin-excluding] cells/cm2. In addition, some cells were planted onto glass coverslips. The cultures were placed into a humidified incubator maintained at 37°C containing a 5% C0,-95% air atmosphere. The culture medium was changed after 3 days and twice weekly thereafter. The growth medium consisted of a modified Eagle's MEM containing double concentrations of amino acids (except glutamine, which was maintained at 2 mM) and quadruple the concentrations of vitamins (Hertz et al., 1982) plus 20% (vh) horse serum. The horse serum was obtained from a horse maintained by the Department of Anatomy, University of Saskatchewan. 0-2A lineage cells were separated from type-1 astrocytes, essentially following the procedure of McCarthy and de Vellis (1980), as outlined in Figure 1. Primary flask cultures ranging in age from 10 days to 28 days in vitro were shaken at 200 gyrations per minute for 30 min and medium removed together with the loosely adherent
Fig. 1 . Flowchart demonstrating procedure for isolation of 02A lineage cells. For details, see Materials and Methods. cells, most of which were macrophages. The growth medium was then replaced with a modified MEM medium containing 15 mM glycerophosphate (pH 7.4) as the buffer, 1 mM NaHCO, instead of the usual 25 mM and 5% (viv) horse serum. The cultures were then placed in a large walk-in incubator maintained at 37°C and shaken overnight at 200 gyrations per minute. Since the capped tissue culture flasks are not completely impermeable to gases and the walk-in incubator contained an air atmosphere, the use of P-glycerophosphate together with the reduction of the NaHCO, concentration ensured that physiological pH could be maintained in the absence of a CO, environment. The next morning (i.e., after 18 hr of agitation), the medium with suspended cells was removed and placed into 100-mm tissue culture dishes and incubated for 3 hr at 37°C in a humidified incubator supplied with 100% air as the atmosphere. During this time, the macrophages and type- 1 astrocytes became firmly adherent to the substratum, while 0-2A lineage cells were only lightly adherent to the substratum. After 3 hr, the medium was removed, and the dishes were gently washed with a balanced salt solution to remove cellular debris. Thereafter, fresh medium containing 0.02% DNase was gently pipetted over the surface of the dishes to remove the 0-2A lineage cells from the substratum while leaving behind most of the macrophages and type-1 astrocytes.
-
Glutamine Synthetase in Type-2 Astroglia
The presence of DNase reduced the aggregation of removed cells caused by DNA released from damaged cells. The cell suspension was centrifuged at 180g for 10 min, resuspended in DNase containing medium, and centrifuged for 15 min at 150g through a 3.5% bovine serum albumin (BSA) cushion; this procedure reduced the amount of debris in the cell suspension. The resulting cell pellet was suspended in the usual NaHC0,-buffered growth medium containing 5% (v/v) horse serum plus 0.02% DNase, sieved through nylon meshes of 35 pm, followed by 15-pm pore sizes, and planted into either polyornithine-coated 35-mm tissue culture dishes or onto polyornithine coated glass coverslips at a cell density of 250 cells/mm2, i.e., -250,000 cells/35-mm culture dish. The reason for the sieving at this stage was that despite the use of DNase, some aggregates still formed that consisted of type-1 astrocytes, 0-2A lineage cells, and macrophages. Removal of these aggregates by sieving secured the absence of almost all the remaining non0-2A lineage cells resulting in cultures consisting almost exclusively of 0-2A lineage cells. After the initial removal of 0-2A lineage cells, some of the primary cultures which now consist mainly of type-1 astrocytes were subcultured either into 35-mm tissue culture dishes (Falcon) at a cell density of 250 cells/mm2or onto glass coverslips. The remainder of the primary cultures were used for the collection of conditioned medium.
Immunocytochemistry
533
Texas red-labelled donkey antirabbit IgG (Jackson Research Laboratories). Cells were visualized using a Zeiss fluorescence microscope with epifluorescent optics and appropriate excitation and barrier filters. To demonstrate GFAP in cells grown directly in 35-mm culture dishes, cultures were fixed in Zamboni’s fluid as above, permeabilized with 0.1% Triton X-100, incubated with rabbit anti-GFAP, washed with PBS, exposed to biotinylated goat antirabbit IgG (Sigma), washed, and exposed to extrAvidinTM-labeledalkaline phosphatase (Sigma). Alkaline phosphatase activity was visualized by using a-naphthyl phosphate as substrate and Fast red TR as the coupling agent.
Glutamine Synthetase Assay Four-week-old 0-2A lineage cell cultures (consisting of -90% type-2 astrocytes) as well as four week old secondary cultures of type- 1 astrocytes were assayed for glutamine synthetase activity by using the transferase reaction. Herzfeld’s (1973) modification of the assay described by Thorndike and Lehrer (1971) was used as previously described (Juurlink, 1982). One-half of the cultures was exposed during the last 48 hr before harvesting to 10 pM hydrocortisone succinate. Proteins were measured using the Bradford (1976) method with bovine gamma globulins serving as the standard.
RESULTS Obtention of Highly Enriched 0 2 A Lineage Cell Cultures By the end of the first week of culture, numerous
The mouse monoclonal A2B5 IgM antibody (Eisenbarth et al., 1979) was collected from tissue culture medium used to grow the A2B5 hybridoma cells islands of large flat cells with associated small processobtained from the American Tissue Type Collection. The bearing cells were observed (Fig. 2A). The small cells medium was concentrated using polyethylene glycol, di- bound the A2B5 monoclonal antibody (Fig. 2B) and alyzed against phosphate-buffered saline (PBS) and fro- were negative for GFAP and galactocerebroside and zen at -80°C until ready for use. To demonstrate galac- hence we interpreted these to be 0-2A progenitor cells. tocerebroside, the BRD 1 mouse IgG monoclonal The large flat cells were GFAP positive (Fig. 2C) and antibody, first described by Ranscht et al., (1982), was galactocerebroside negative and did not bind A2B5 and, used. We thank Dr. Patrick M. Wood of the Miami hence were interpreted to be type-1 astrocytes. During Project to Cure Paralysis, Miami, for the gift of this the next week, there was rapid proliferation of both the antibody. To demonstrate A2B5 or BRDl binding, living type-1 astroglial cells and the 0-2A progenitor cells such cells were incubated at 4°C with the antibody in culture that the cultures at 2 weeks consisted of a confluent layer medium buffered with glycerol phosphate for 30 min, of type- 1 astrocytes and numerous process-bearing cells washed with three changes of PBS, and fixed in Zam- (Fig. 3). During the next several weeks the cultures beboni’s fluid [4% formaldehyde, prepared from paraform- came more dense. The process-bearing cells belong to aldehyde, containing 0.2% picric acid in phosphate the 0-2A lineage, as demonstrated by their binding of buffer (Zamboni and de Martino, 1967)l. The fixed cells the A2B5 antibody (Fig. 4A), some of which were also were washed with PBS, permeabilized with 0.1% Triton positive for GFAP (compare Fig. 4A and 4B). The imX-100 and incubated with a 1:100 dilution of rabbit anti- pression gained when examining such cultures with GFAP IgG (Dako), washed with PBS, and incubated in phase microscopy was that all the 0-2A lineage cells a mixture of a 1:lOO dilution of FITC-labelled goat an- reside on top of the type-1 astrocytes. However, this is timouse IgM (Sigma) or 1:lOO dilution of FITC-labeled true only for some of them; many 0-2A lineage cells are goat antimouse IgG (Sigma) and 1:800 dilution of situated between the layers of type-1 astrocytes, while
Figs. 2-4.
Glutamine Synthetase in Type-2 Astroglia
some reside directly on the substratum. This is reflected in the variable fluorescent intensity shown in Figure 4A, where the more superficial cells are more intensely fluorescent. The largest number of 0-2A lineage cells was observed in cultures established from P4 neopallium; hence, subsequent cultures were set up from rat pups of this age. Preliminary experiments determined that the optimal time for isolating 0-2A lineage cells were from cultures that were 10-14 days old. At this time, most of the 0-2A lineage cells were situated on top of the type-1 astrocytes. With older cultures, the type-1 astrocytic “mat” in which the majority of 0-2A lineage cells were embedded became too dense for easy isolation of 0-2A lineage cells. Attempts at 0-2A lineage isolation from these older cultures resulted in the sloughing of large pieces of this astrocytic “mat,” with the result that large amounts of debris entrapped any 0-2A lineage cells that separated from the primary culture. Critical steps in the isolation of highly purified 02A lineage cells were the preplating procedure and the sieving through 35-km and 15-km nylon meshes. During the preplating all the single cells attached to the substratum; after 3 hr, the non-0-2A lineage cells were firmly adherent to the substratum while the 0-2A lineage cells were only loosely adherent and could be readily washed off. The washed off cells had a tendency to aggregate, but this was inhibited by the presence of DNase. The major source of non-0-2A lineage cells was in small cell aggregates that had not become attached during the preplating; many of these could be eliminated by gently washing the preplate dishes and the rest were eliminated
535
with the sieving procedure. By following the above steps, cultures of highly purified 0-2A lineage cells were obtained (Fig. 5), where very few non-0-2A cells could be identified during the initial few days of culture. Any deviation from the procedures described above resulted in considerable contamination by non-0-2A cells, in particular type-1 astrocytes, which formed a major portion of the cells in the aggregates. The number of 0-2A lineage cells obtained from cultures set up from 1 rat pup varied somewhat from experiment to experiment but was generally 0.5-1.0 X lo6 cells. In order to maintain the cultures of 0-2A lineage cells, it is necessary to add conditioned medium obtained from type-1 astrocyte cultures to the growth medium. The optimal concentration of conditioned medium was determined to be 20%. The growth medium routinely used in cultures of 0-2A lineage cells that differentiated toward type-2 astrocytes consisted of modified MEM containing 5% horse serum, as well as 20% conditioned medium. If grown without the conditioned medium, the cells degenerated by day 5 of culture.
Criteria Used for Immunocytochemical Characterization of Highly Enriched 0 2 A Lineage Cell Cultures The criteria detailed by Raff et al., (1983b) were used in defining the cell populations found in the purified 02A lineage cells. Under the culture conditions described in the present work, none of the cells after 1 week of culturing on glass coverslips was positive for galactocerebroside; however, when cell were grown using other culture conditions many of the 02A progenitors differentiated into galactocerebroside positive cells. Small process bearing that were positive for GFAP were Fig. 2. One-week P4 neopallial culture grown on a glass cov- classified as type-2 astrocytes; of these, 95.5% also erslip and fixed with Zamboni’s fluid A: Small process-bearing bound the A2B5 monoclonal antibody (this proportion of cells can be seen associated with larger flat cells, phase-contype-2 astrocytes binding A2B5 is similar to that reported trast microscopy. B: Same field as in A, demonstrating the by Raff et al. (1983) for optic nerve derived type-2 aslocalization of A2B5 binding to the small process bearing cells . Small process-bearing cells that were negative trocytes) (i.e., 0-2A progenitor cells). C: Same field as in A and B, for GFAP and negative for galactocerebroside and that demonstrating that the localization of GFAP is to the large flat bound the A2B5 monoclonal antibody were classified as cells (i.e., type-1 astrocytes). Bar = 25 Fm. 02A progenitor cells. Large flat cells that were positive Fig. 3. Phase-contrast micrograph of a 2-week-old P4 neopal- for GFAP, negative for galactocerebroside, and did not lial culture grown on a plastic substratum, demonstrating the bind A2B5 were classified as type-1 astrocytes. It cannot confluence of type-1 astrocytes and large numbers of process- be excluded that 0 2 A lineage cells behave differently on bearing 0-2A lineage cells. Bar = 50 Fm. glass coverslips than on the plastic substratum of petri dishes. Since the biochemical studies were conducted Fig. 4. Four-week-old neopallial culture grown on a glass cov- using type-2 astrocytes grown directly on the plastic suberslip and fixed with Zamboni’s fluid. Arrows demonstrate stratum of petri dishes, we performed the cell type analidentical points in the field. A: Demonstration that A2B5 bindysis for the 3-week cultures on cells grown for 3 weeks ing is restricted to process-bearing cells. The more faintly stained cells lie beneath type-I astrocytes. B: Same field as in directly on the plastic substratum of the petri dishesA, demonstrating the localization of GFAP. Some of the A2B5 here only one antibody (anti-GFAP) was used in the binding cells can also be seen to be clearly positive for GFAP. classification of the cells. The following criteria were used in defining the cell populations: (1) small to meBar = 25 pm.
536
Juurlink and Hertz
Fig. 5. Phase-contrast micrograph of living 0-2A lineage cells 1 day after isolation growing on glass coverslips. Bar = 50 wm. Fig. 6. Phase-contrast micrograph of living 0-2A lineage cells 1 week after isolation growing on glass coverslips. Bar = 50
Fig. 7. Light micrograph of 3-week-old type-2 astrocytes grown in a tissue culture dish, demonstrating the localization of GFAP using the biotin-extrAvidin-alkaline phosphatase technique. Note that several of the type-2 astrocytes are quite large by this stage of culture. Bar = 50 pm.
CLm.
Fig. 8. Phase-contrast micrograph of living type-1 astrocyte culture 4 weeks after subculturing. Bar = 50 pm.
dium-sized process bearing cells that were positive for GFAP were classified as type-2 astrocytes; (2) small process bearing cells that were negative for GFAP were classified as 0 2 A progenitor cells; and (3) large flat cells that were positive for GFAP were classified as type-1 astrocytes. That this interpretation of 3-week cultures using only one antibody was valid was confirmed by doing an analysis on a small sample (215 cells) of cells
grown on coverslips and finding similar proportions of cell types; again, at this time, no galactocerebroside positive cells were seen. One difference from 1-week cultures was that, although all GFAP-negative processbearing cells were A2B5 positive (i.e., 0 2 A progenitors), only 65% of the type-2 astrocytes (small to medium process-bearing GFAP positive cells) in 3-week cultures bound the A2B5 monoclonal antibody.
Glutamine Synthetase in Type-2 Astroglia
537
TABLE I. Cellular Composition of Type-2 Astrocyte-Enriched Cultures Culture aee (weeks) 1' 3'
Cell typea 2A (%) 89.2 91.0
2
0.9d
f 4.6
0-2A (%)
1A (%)
Otherb (%)
1.7 f 2.3 5.5 f 2.8
n.d." 1.46 _t 1.5
3.2 2 1.2 2.1 f 1.5
aA total of 1,700 cells from two experiments were counted. bCells could not be identified with the antibodies used. 'Cells grown on polyornithine-coated glass coverslips.
dtS.D. eIn the random fields examined, type-1 astrocytes were not detected. 'Cells grown on polyornithine-coated plastic substratum of 35-mm tissue culture dishes.
Description of Highly Enriched 0 2 A Lineage Cell Cultures After 1 day of culture, almost all cells were characterized at the phase-contrast microscopic level by small somas with associated radiating processes (Fig. 5 ) . There was little change in the appearance of these cultures over the next week (Fig. 6). Only -30% of the 0-2A lineage cells were positive for GFAP 1 day after isolation, this increased to -90% by 1 week (Table I) and remained the same for the next few weeks. By 3 weeks of culture, non-0-2A lineage cells constitute -3.5% of the total cells in the cultures of which about half were type-1 astrocytes and the remainder were not identifiable with the antibodies used (Table I). 0-2A progenitors formed a significant fraction (-5%) of the cells in culture at all culture stages examined. During the first week of culture, all the type-2 astrocytes were characterized by the presence of small somas with thin radiating processes (Figs. 5,6). By 2 weeks of culture, some of the type-2 astrocytes (-2.5%) increased in size, displaying large, robust processes (Fig. 7), and by 3-4 weeks, the proportion of large type-2 astrocytes increased to 1I%. This increase in size of type-2 astrocytes with time has also been reported by Raff et al. (1983b). Aloisi et al. (1988) report that after I week of culture, many of the type-2 astrocytes assume a flat nonprocess-bearing morphology reminiscent of type- 1 astrocytes. In our cultures, the process-bearing morphology of type-2 astrocytes was maintained for up to 4 weeks. We did note that some of the large type-2 astrocytes eventually underwent degeneration. It was our impression that there was a continual slow recruitment of type-2 astrocytes from dividing 0-2A progenitors and that this appeared to be balanced by the degeneration of large type-2 astrocytes; however, we did not verify this impression experimentally.
-
Type 1
Type- 2
Fig. 9. Graph depicting glutamine synthetase activity in 4week-old purified type- 1 astrocyte cultures and in 4-week-old purified type-2 astrocyte cultures. Assays were performed on two separate batches of cultures. Cortisol at a concentration of 10 pM was present in one-half the cultures for 48 hr prior to harvest. The basal level of glutamine synthetase activity of type-2 astrocytes is significantly higher than that of type- 1 astrocytes (P
Establishment of Highly Enriched Type-2 Astrocyte Cultures and Quantitative Determination of Intense Glutamine Synthetase Activity in These Cells B.H.J. Juurlink and L. Hertz Departments of Anatomy (B.H.J.J.), pharmacology (L.H.), and Anaesthesia (L.H.), University of Saskatchewan, Saskatoon, Canada We report procedures that allow one to develop and maintain cultures highly enriched in rat neopallial type-2 astrocytes. Even after four weeks such cultures consist of more than 90% type-2 astrocytes, -5% 0-2A progenitors and fewer than 2% type-1 astrocytes. Their survival for more than 5 days requires the addition of conditioned medium from type-1 astrocyte cultures. The type-2 astrocytes have an intense glutamine synthetase activity whose basal level is sevenfold higher than in type-1 astrocytes. The glutamine synthetase activities of both the type-2 and type-1 astrocytes are increased after exposure to cortisol. Thus, type-2 astrocytes express the two quintessential astrocytic features: glial fibrillary acidic protein (previously reported by others) and glutamine synthetase.
1986; Levi et al., 1986; Trotter and Schachner, 1989) and from cerebrum (Aloisi et al., 1988; Behar et al., 1988; Ingraham and McCarthy, 1989). Much information is now available about type-l astrocytes because they have been available in highly enriched primary cultures for more than 15 years (Juurlink and Hertz, 1985). Less information is available about type-2 astrocytes, since highly enriched (>90%) long-term (>1 week) cultures of these cells have not yet been available. The type-2 astrocytes exhibit a number of properties that traditionally have been considered neuronal and not astrocytic. For example, type-2 astrocytes avidly accumulate y-aminobutyric acid (GABA) (Johnstone et al., 1986; Levi et al., 1986; Barres et al., 1990) but, in contrast to type-1 astrocytes (Schousboe et al., 1977; Bardakjian et al., 1979), they do not appear to metabolize it readily; rather, type-2 astrocytes release GABA Key words: glia, neopallium, 0-2A lineage, rat, tis- upon stimulation with glutamate or glutamate analogues sue culture (Gallo et al., 1989). Similar to GABAergic neurons and unlike type-1 astrocytes, type-2 astrocytes as well as their 0-2A progenitors synthesize GABA (Barres et al., INTRODUCTION 1990). Furthermore, type-2 astrocytes have a “neuLess than 10 years ago, Raff and his colleagues ronal” channel phenotype (Barres et al., 1988). (Raff et al., 1983a,b) demonstrated that there are two One criterion that classifies type-2 astrocytes as distinct astrocyte types in cultures established from optic astrocytic in nature is the presence of the intermediate nerve that differ from each other in lineage origins, plas- filament protein glial fibrillary acidic protein (GFAP). malemmal ganglioside composition, morphology, and GFAP was the first described molecular marker of astroresponse to epidermal growth factor. One, termed the cytes (Eng et al., 1971; Bignami and Dahl, 1977; type-1 astrocyte was a large flat cell that resembled in Schachner et al., 1977). A second characteristic of morphology and other properties the major population of astrocytes is the enzyme glutamine synthetase. Both astrocytes in cultures established from other brain re- immunohistochemical (Norenberg and Martinez-Hergions such as the cerebral hemispheres (e.g., Hertz et nandez, 1979) and molecular biological (Mearow et al., al., 1982). The other, termed the type-2 astrocyte arose from a glial precursor, the 0-2A progenitor, that could differentiate into either an oligodendrocyteor into a type- Received December 6, 1990; revised March 13, 1991;accepted March 2 astrocyte, depending upon culture environment 25, 1991. (reivewed in Raff, 1989). 0-2A progenitors and their Address reprint requests to Dr. B.H.J. Juurlink, Department of Anattype-2 astrocyte derivatives have also been described in omy, University of Saskatchewan, Saskatoon, Saskatchewan, Canada cultures established from cerebellum (Johnstone et al., S7N OWO. 0 1991 Wiley-Liss, Inc.
532
Juurlink and Hertz
1989) techniques have been used to demonstrate that the enzyme glutamine synthetase is restricted to astroglial cells within the central nervous system (CNS). Several recent reports, however, suggest that some glutamine synthetase activity is also present in oligodendrocytes (Warringa et al., 1988; Cammer, 1990; D'Amelia et al., 1990). Type-2-like astrocytes also have immunocytochemically detectable glutamine synthetase (Raff et al., 1983a; B .H.J. Juurlink, unpublished observations); however, no quantitative information is available regarding glutamine synthetase activity in this class of astrocyte, since relatively homogeneous cultures of type-2 astrocytes have only relatively recently been obtained (Ingraham and McCarthy, 1989), and long-term cultures of highly enriched type-2 astrocytes do not appear to be available. In this report, we describe procedures to obtain cultures highly enriched (>90%) in type-2 astrocytes that can be maintained at this purity for at least 4 weeks. Such type-2 astrocyte cultures were found to display intense glutamine synthetase activity.
PKIMARY ASTROCYTE CULTURES SHAKEN ON ROTARY SHAKER A'l' 37" C FOR 30 MIN
L REMOVE MEDIUM WITH SUSPENDED CELLS
I RINSE CULTURES AND ADD FRESH MEDIUM BUFFERED WITH GLYCEROPHOSPHATE SHAKE OVERNIGHT (- i a HR)
0-
i REMOVE SUSPENDED CELLS AND PLACE IN I 0 0 mm PETRI DISHES FOR A 3 HR INCUBATION AT 37" C IN A HUMIDIFIED AIR ATMOSPHERE
1 REMOVE MEDIUM AND GENTLY WASH PETRI DISHES TO REMOVE DEBRIS
I REMOVE THE LOOSELY ADHERENT 0-2A LINEAGE CELLS BY OENI'I,'? PIPE'TTING THE SURFACE OF THE DISHES WITH A BICARBONATE BUFFERED GROWTH MEDIUM CONTAINING 0 02% DN=c
I CENTRIFUGE CELLS AT 18Og FOR 10 MIN AND RESUSPEND IN GROW YH MEDIUM CONTAINING 0 02% DNaSe
L CENTRIFUGE CELLS AT 15Og FOR 15 MIN THROUGH A 3 5% BSA CUSHION
I RESUSPEND CELLS IN GROWTH MEDIUM CONTAINING 0 02% DNaSe AVD FILTER SEQUENTIALLY THROUGH NYLON MESHES OF PORE SIZES 3 5 p AND 15pm
I PLANT CELLS IN GROWTH MEDIUM CONTAINING 0 02% DNaSe ONTO POLYORNITHINE COATED SURFACES
i AFTER 15 MIN REMOVE MEDIUM AND ADD DNase FREE GROWTH MEDl UM
MATERIALS AND METHODS Culture Preparation The neopallia (i.e., those portions of the cerebral hemispheres that remain after removal of the olfactory bulbs, hippocampi, and basal ganglia) were aseptically separated from cerebral hemispheres obtained from young Wistar rat pups ranging in age from newborn (PO) to postnatal day 5 (P5).Meninges were removed, and the neopallia were diced into cubes of -1 mm3 and gently pushed through a nylon mesh with a pore size of 75 pm. The cells obtained from two pups were planted into three 75-cm2 (Falcon) tissue culture flasks, i.e., 30,000 viable [nigrosin-excluding] cells/cm2. In addition, some cells were planted onto glass coverslips. The cultures were placed into a humidified incubator maintained at 37°C containing a 5% C0,-95% air atmosphere. The culture medium was changed after 3 days and twice weekly thereafter. The growth medium consisted of a modified Eagle's MEM containing double concentrations of amino acids (except glutamine, which was maintained at 2 mM) and quadruple the concentrations of vitamins (Hertz et al., 1982) plus 20% (vh) horse serum. The horse serum was obtained from a horse maintained by the Department of Anatomy, University of Saskatchewan. 0-2A lineage cells were separated from type-1 astrocytes, essentially following the procedure of McCarthy and de Vellis (1980), as outlined in Figure 1. Primary flask cultures ranging in age from 10 days to 28 days in vitro were shaken at 200 gyrations per minute for 30 min and medium removed together with the loosely adherent
Fig. 1 . Flowchart demonstrating procedure for isolation of 02A lineage cells. For details, see Materials and Methods. cells, most of which were macrophages. The growth medium was then replaced with a modified MEM medium containing 15 mM glycerophosphate (pH 7.4) as the buffer, 1 mM NaHCO, instead of the usual 25 mM and 5% (viv) horse serum. The cultures were then placed in a large walk-in incubator maintained at 37°C and shaken overnight at 200 gyrations per minute. Since the capped tissue culture flasks are not completely impermeable to gases and the walk-in incubator contained an air atmosphere, the use of P-glycerophosphate together with the reduction of the NaHCO, concentration ensured that physiological pH could be maintained in the absence of a CO, environment. The next morning (i.e., after 18 hr of agitation), the medium with suspended cells was removed and placed into 100-mm tissue culture dishes and incubated for 3 hr at 37°C in a humidified incubator supplied with 100% air as the atmosphere. During this time, the macrophages and type- 1 astrocytes became firmly adherent to the substratum, while 0-2A lineage cells were only lightly adherent to the substratum. After 3 hr, the medium was removed, and the dishes were gently washed with a balanced salt solution to remove cellular debris. Thereafter, fresh medium containing 0.02% DNase was gently pipetted over the surface of the dishes to remove the 0-2A lineage cells from the substratum while leaving behind most of the macrophages and type-1 astrocytes.
-
Glutamine Synthetase in Type-2 Astroglia
The presence of DNase reduced the aggregation of removed cells caused by DNA released from damaged cells. The cell suspension was centrifuged at 180g for 10 min, resuspended in DNase containing medium, and centrifuged for 15 min at 150g through a 3.5% bovine serum albumin (BSA) cushion; this procedure reduced the amount of debris in the cell suspension. The resulting cell pellet was suspended in the usual NaHC0,-buffered growth medium containing 5% (v/v) horse serum plus 0.02% DNase, sieved through nylon meshes of 35 pm, followed by 15-pm pore sizes, and planted into either polyornithine-coated 35-mm tissue culture dishes or onto polyornithine coated glass coverslips at a cell density of 250 cells/mm2, i.e., -250,000 cells/35-mm culture dish. The reason for the sieving at this stage was that despite the use of DNase, some aggregates still formed that consisted of type-1 astrocytes, 0-2A lineage cells, and macrophages. Removal of these aggregates by sieving secured the absence of almost all the remaining non0-2A lineage cells resulting in cultures consisting almost exclusively of 0-2A lineage cells. After the initial removal of 0-2A lineage cells, some of the primary cultures which now consist mainly of type-1 astrocytes were subcultured either into 35-mm tissue culture dishes (Falcon) at a cell density of 250 cells/mm2or onto glass coverslips. The remainder of the primary cultures were used for the collection of conditioned medium.
Immunocytochemistry
533
Texas red-labelled donkey antirabbit IgG (Jackson Research Laboratories). Cells were visualized using a Zeiss fluorescence microscope with epifluorescent optics and appropriate excitation and barrier filters. To demonstrate GFAP in cells grown directly in 35-mm culture dishes, cultures were fixed in Zamboni’s fluid as above, permeabilized with 0.1% Triton X-100, incubated with rabbit anti-GFAP, washed with PBS, exposed to biotinylated goat antirabbit IgG (Sigma), washed, and exposed to extrAvidinTM-labeledalkaline phosphatase (Sigma). Alkaline phosphatase activity was visualized by using a-naphthyl phosphate as substrate and Fast red TR as the coupling agent.
Glutamine Synthetase Assay Four-week-old 0-2A lineage cell cultures (consisting of -90% type-2 astrocytes) as well as four week old secondary cultures of type- 1 astrocytes were assayed for glutamine synthetase activity by using the transferase reaction. Herzfeld’s (1973) modification of the assay described by Thorndike and Lehrer (1971) was used as previously described (Juurlink, 1982). One-half of the cultures was exposed during the last 48 hr before harvesting to 10 pM hydrocortisone succinate. Proteins were measured using the Bradford (1976) method with bovine gamma globulins serving as the standard.
RESULTS Obtention of Highly Enriched 0 2 A Lineage Cell Cultures By the end of the first week of culture, numerous
The mouse monoclonal A2B5 IgM antibody (Eisenbarth et al., 1979) was collected from tissue culture medium used to grow the A2B5 hybridoma cells islands of large flat cells with associated small processobtained from the American Tissue Type Collection. The bearing cells were observed (Fig. 2A). The small cells medium was concentrated using polyethylene glycol, di- bound the A2B5 monoclonal antibody (Fig. 2B) and alyzed against phosphate-buffered saline (PBS) and fro- were negative for GFAP and galactocerebroside and zen at -80°C until ready for use. To demonstrate galac- hence we interpreted these to be 0-2A progenitor cells. tocerebroside, the BRD 1 mouse IgG monoclonal The large flat cells were GFAP positive (Fig. 2C) and antibody, first described by Ranscht et al., (1982), was galactocerebroside negative and did not bind A2B5 and, used. We thank Dr. Patrick M. Wood of the Miami hence were interpreted to be type-1 astrocytes. During Project to Cure Paralysis, Miami, for the gift of this the next week, there was rapid proliferation of both the antibody. To demonstrate A2B5 or BRDl binding, living type-1 astroglial cells and the 0-2A progenitor cells such cells were incubated at 4°C with the antibody in culture that the cultures at 2 weeks consisted of a confluent layer medium buffered with glycerol phosphate for 30 min, of type- 1 astrocytes and numerous process-bearing cells washed with three changes of PBS, and fixed in Zam- (Fig. 3). During the next several weeks the cultures beboni’s fluid [4% formaldehyde, prepared from paraform- came more dense. The process-bearing cells belong to aldehyde, containing 0.2% picric acid in phosphate the 0-2A lineage, as demonstrated by their binding of buffer (Zamboni and de Martino, 1967)l. The fixed cells the A2B5 antibody (Fig. 4A), some of which were also were washed with PBS, permeabilized with 0.1% Triton positive for GFAP (compare Fig. 4A and 4B). The imX-100 and incubated with a 1:100 dilution of rabbit anti- pression gained when examining such cultures with GFAP IgG (Dako), washed with PBS, and incubated in phase microscopy was that all the 0-2A lineage cells a mixture of a 1:lOO dilution of FITC-labelled goat an- reside on top of the type-1 astrocytes. However, this is timouse IgM (Sigma) or 1:lOO dilution of FITC-labeled true only for some of them; many 0-2A lineage cells are goat antimouse IgG (Sigma) and 1:800 dilution of situated between the layers of type-1 astrocytes, while
Figs. 2-4.
Glutamine Synthetase in Type-2 Astroglia
some reside directly on the substratum. This is reflected in the variable fluorescent intensity shown in Figure 4A, where the more superficial cells are more intensely fluorescent. The largest number of 0-2A lineage cells was observed in cultures established from P4 neopallium; hence, subsequent cultures were set up from rat pups of this age. Preliminary experiments determined that the optimal time for isolating 0-2A lineage cells were from cultures that were 10-14 days old. At this time, most of the 0-2A lineage cells were situated on top of the type-1 astrocytes. With older cultures, the type-1 astrocytic “mat” in which the majority of 0-2A lineage cells were embedded became too dense for easy isolation of 0-2A lineage cells. Attempts at 0-2A lineage isolation from these older cultures resulted in the sloughing of large pieces of this astrocytic “mat,” with the result that large amounts of debris entrapped any 0-2A lineage cells that separated from the primary culture. Critical steps in the isolation of highly purified 02A lineage cells were the preplating procedure and the sieving through 35-km and 15-km nylon meshes. During the preplating all the single cells attached to the substratum; after 3 hr, the non-0-2A lineage cells were firmly adherent to the substratum while the 0-2A lineage cells were only loosely adherent and could be readily washed off. The washed off cells had a tendency to aggregate, but this was inhibited by the presence of DNase. The major source of non-0-2A lineage cells was in small cell aggregates that had not become attached during the preplating; many of these could be eliminated by gently washing the preplate dishes and the rest were eliminated
535
with the sieving procedure. By following the above steps, cultures of highly purified 0-2A lineage cells were obtained (Fig. 5), where very few non-0-2A cells could be identified during the initial few days of culture. Any deviation from the procedures described above resulted in considerable contamination by non-0-2A cells, in particular type-1 astrocytes, which formed a major portion of the cells in the aggregates. The number of 0-2A lineage cells obtained from cultures set up from 1 rat pup varied somewhat from experiment to experiment but was generally 0.5-1.0 X lo6 cells. In order to maintain the cultures of 0-2A lineage cells, it is necessary to add conditioned medium obtained from type-1 astrocyte cultures to the growth medium. The optimal concentration of conditioned medium was determined to be 20%. The growth medium routinely used in cultures of 0-2A lineage cells that differentiated toward type-2 astrocytes consisted of modified MEM containing 5% horse serum, as well as 20% conditioned medium. If grown without the conditioned medium, the cells degenerated by day 5 of culture.
Criteria Used for Immunocytochemical Characterization of Highly Enriched 0 2 A Lineage Cell Cultures The criteria detailed by Raff et al., (1983b) were used in defining the cell populations found in the purified 02A lineage cells. Under the culture conditions described in the present work, none of the cells after 1 week of culturing on glass coverslips was positive for galactocerebroside; however, when cell were grown using other culture conditions many of the 02A progenitors differentiated into galactocerebroside positive cells. Small process bearing that were positive for GFAP were Fig. 2. One-week P4 neopallial culture grown on a glass cov- classified as type-2 astrocytes; of these, 95.5% also erslip and fixed with Zamboni’s fluid A: Small process-bearing bound the A2B5 monoclonal antibody (this proportion of cells can be seen associated with larger flat cells, phase-contype-2 astrocytes binding A2B5 is similar to that reported trast microscopy. B: Same field as in A, demonstrating the by Raff et al. (1983) for optic nerve derived type-2 aslocalization of A2B5 binding to the small process bearing cells . Small process-bearing cells that were negative trocytes) (i.e., 0-2A progenitor cells). C: Same field as in A and B, for GFAP and negative for galactocerebroside and that demonstrating that the localization of GFAP is to the large flat bound the A2B5 monoclonal antibody were classified as cells (i.e., type-1 astrocytes). Bar = 25 Fm. 02A progenitor cells. Large flat cells that were positive Fig. 3. Phase-contrast micrograph of a 2-week-old P4 neopal- for GFAP, negative for galactocerebroside, and did not lial culture grown on a plastic substratum, demonstrating the bind A2B5 were classified as type-1 astrocytes. It cannot confluence of type-1 astrocytes and large numbers of process- be excluded that 0 2 A lineage cells behave differently on bearing 0-2A lineage cells. Bar = 50 Fm. glass coverslips than on the plastic substratum of petri dishes. Since the biochemical studies were conducted Fig. 4. Four-week-old neopallial culture grown on a glass cov- using type-2 astrocytes grown directly on the plastic suberslip and fixed with Zamboni’s fluid. Arrows demonstrate stratum of petri dishes, we performed the cell type analidentical points in the field. A: Demonstration that A2B5 bindysis for the 3-week cultures on cells grown for 3 weeks ing is restricted to process-bearing cells. The more faintly stained cells lie beneath type-I astrocytes. B: Same field as in directly on the plastic substratum of the petri dishesA, demonstrating the localization of GFAP. Some of the A2B5 here only one antibody (anti-GFAP) was used in the binding cells can also be seen to be clearly positive for GFAP. classification of the cells. The following criteria were used in defining the cell populations: (1) small to meBar = 25 pm.
536
Juurlink and Hertz
Fig. 5. Phase-contrast micrograph of living 0-2A lineage cells 1 day after isolation growing on glass coverslips. Bar = 50 wm. Fig. 6. Phase-contrast micrograph of living 0-2A lineage cells 1 week after isolation growing on glass coverslips. Bar = 50
Fig. 7. Light micrograph of 3-week-old type-2 astrocytes grown in a tissue culture dish, demonstrating the localization of GFAP using the biotin-extrAvidin-alkaline phosphatase technique. Note that several of the type-2 astrocytes are quite large by this stage of culture. Bar = 50 pm.
CLm.
Fig. 8. Phase-contrast micrograph of living type-1 astrocyte culture 4 weeks after subculturing. Bar = 50 pm.
dium-sized process bearing cells that were positive for GFAP were classified as type-2 astrocytes; (2) small process bearing cells that were negative for GFAP were classified as 0 2 A progenitor cells; and (3) large flat cells that were positive for GFAP were classified as type-1 astrocytes. That this interpretation of 3-week cultures using only one antibody was valid was confirmed by doing an analysis on a small sample (215 cells) of cells
grown on coverslips and finding similar proportions of cell types; again, at this time, no galactocerebroside positive cells were seen. One difference from 1-week cultures was that, although all GFAP-negative processbearing cells were A2B5 positive (i.e., 0 2 A progenitors), only 65% of the type-2 astrocytes (small to medium process-bearing GFAP positive cells) in 3-week cultures bound the A2B5 monoclonal antibody.
Glutamine Synthetase in Type-2 Astroglia
537
TABLE I. Cellular Composition of Type-2 Astrocyte-Enriched Cultures Culture aee (weeks) 1' 3'
Cell typea 2A (%) 89.2 91.0
2
0.9d
f 4.6
0-2A (%)
1A (%)
Otherb (%)
1.7 f 2.3 5.5 f 2.8
n.d." 1.46 _t 1.5
3.2 2 1.2 2.1 f 1.5
aA total of 1,700 cells from two experiments were counted. bCells could not be identified with the antibodies used. 'Cells grown on polyornithine-coated glass coverslips.
dtS.D. eIn the random fields examined, type-1 astrocytes were not detected. 'Cells grown on polyornithine-coated plastic substratum of 35-mm tissue culture dishes.
Description of Highly Enriched 0 2 A Lineage Cell Cultures After 1 day of culture, almost all cells were characterized at the phase-contrast microscopic level by small somas with associated radiating processes (Fig. 5 ) . There was little change in the appearance of these cultures over the next week (Fig. 6). Only -30% of the 0-2A lineage cells were positive for GFAP 1 day after isolation, this increased to -90% by 1 week (Table I) and remained the same for the next few weeks. By 3 weeks of culture, non-0-2A lineage cells constitute -3.5% of the total cells in the cultures of which about half were type-1 astrocytes and the remainder were not identifiable with the antibodies used (Table I). 0-2A progenitors formed a significant fraction (-5%) of the cells in culture at all culture stages examined. During the first week of culture, all the type-2 astrocytes were characterized by the presence of small somas with thin radiating processes (Figs. 5,6). By 2 weeks of culture, some of the type-2 astrocytes (-2.5%) increased in size, displaying large, robust processes (Fig. 7), and by 3-4 weeks, the proportion of large type-2 astrocytes increased to 1I%. This increase in size of type-2 astrocytes with time has also been reported by Raff et al. (1983b). Aloisi et al. (1988) report that after I week of culture, many of the type-2 astrocytes assume a flat nonprocess-bearing morphology reminiscent of type- 1 astrocytes. In our cultures, the process-bearing morphology of type-2 astrocytes was maintained for up to 4 weeks. We did note that some of the large type-2 astrocytes eventually underwent degeneration. It was our impression that there was a continual slow recruitment of type-2 astrocytes from dividing 0-2A progenitors and that this appeared to be balanced by the degeneration of large type-2 astrocytes; however, we did not verify this impression experimentally.
-
Type 1
Type- 2
Fig. 9. Graph depicting glutamine synthetase activity in 4week-old purified type- 1 astrocyte cultures and in 4-week-old purified type-2 astrocyte cultures. Assays were performed on two separate batches of cultures. Cortisol at a concentration of 10 pM was present in one-half the cultures for 48 hr prior to harvest. The basal level of glutamine synthetase activity of type-2 astrocytes is significantly higher than that of type- 1 astrocytes (P
