On the Role of Glial Cells in the Mammalian Nervous System UPTAKE, EXCRETION, AND RIETABOLISM OF PUTATIVE NEUROTRANSRlITTERS BY CULTURED GLIAL TUMOR CELLS

SUMMARY The synthesis, uptake, and efilux of some putative neurotransmitters by cultured rat glial tumor cells were studied. Three glial cell clones were capable of pyridoxal-dependent synthesis of y-aminobutyrate both in cell-free homogenates and in intact cells in monolayers. Substantial catabolism of y-aminobutyrate was not found. Synthesis of taurine and isethionic acid, but not p-alanine, by one of the glioma lines was also shown. Glioma lines were found to take up glutamate and exhibited Na+-dependent uptake of y-amino-butyrate. The uptake of y-aminobutyrate consisted of a slow saturable component (KT 13 to 30 PM) and a rapid nonsaturable component. Both of these were antagonized by some structural analogs of y-aminobutyrate as well as by taurine, P-alanine, bicuculline, and low temperature. Similar kinetic parameters were found for three different glioma lines. obtained from only the nonsaturable component


SUMMARY
The synthesis, uptake, and efilux of some putative neurotransmitters by cultured rat glial tumor cells were studied. Three glial cell clones were capable of pyridoxal-dependent synthesis of y-aminobutyrate both in cell-free homogenates and in intact cells in monolayers.
Substantial catabolism of y-aminobutyrate was not found. Synthesis of taurine and isethionic acid, but not p-alanine, by one of the glioma lines was also shown.
Glioma lines were found to take up glutamate and exhibited Na+-dependent uptake of y-aminobutyrate.
The uptake of y-aminobutyrate consisted of a slow saturable component (KT -13 to 30 PM) and a rapid nonsaturable component.
Both of these were antagonized by some structural analogs of y-aminobutyrate as well as by taurine, P-alanine, bicuculline, and low temperature. Similar kinetic parameters were found for three different glioma lines.
Fibroblast-like cells obtained from rat brain cell cultures had only the nonsaturable component of yaminobutyrate uptake. Taurine uptake also consisted of two Na+-dependent and temperature-dependent components: a rapid saturable component (KT -10 to 17 PM) and a nonsaturable component which varied in magnitude between cell lines. These uptakes were antagonized by @-alanine but not by y-aminobutyrate.
The glioma lines also excreted the concentrated y-aminobutyrate into the extracellular milieu, but could maintain cell to medium concentration ratios of >80-fold.
In taurine efflux experiments, cell to medium concentration ratios in excess of 1500-fold could be maintained by one of the gliomas. The data are consistent with a possible role of central nervous system glial cells in the modulation of neuronal excitability via control of the levels of neuroactive substances in the extracellular milieu of neurons.
Galambos in 1961 (1) suggested that glial cells may have major functional roles in the central nervous system. Several * Present, address, Department of Zoology, University College, London WC-l, England. specific suggestions for such functions have been formed, including uptake of K+ ions (24) and neurotransmitters (5)(6)(7)(8) from synaptic clefts, as well as guides for neuronal migration (9). An ability of partially purified glial cells to specifically take up y-aminobutyrate, norepinephrine, and serotonin, albeit with lower affinity than did neurons, was suggested by the work of Henn and Hamberger (10). Bleecker (ll), using electron microscopic autoradiographic techniques, has found that glial cells in brain slices incubated with [3H]taurine, a putative neurotransmitter (12)(13)(14), appeared to have concentrated that compound.
Thus, some more complex role of glia in brain neurotransmitter control seemed likely.
In previous studies (15)(16)(17), we found that some rat glial tumor cell lines contained substantial activity of the enzyme glutamic acid decarboxylase (EC 4.1.1.15), although we were unable to find evidence for the pyridoxal-independent type II (18) activity.
In the present work we have investigated the ability of these tumor cells to synthesize some putative neurotransmitters, to concentrate them from the extracellular fluid and to release them. It was found that the glioma cell lines C-6, C-21, and NT-l behaved similarly in synthesis, uptake, and efflux of y-aminobutyrate and taurine. In addition, it was found that C-6 could synthesize significant amounts of taurine from cystine in the medium.
The three cell lines exhibited active, Na+-dependent concentration of y-aminobutyrate and taurine from the medium; the data suggested that there were separate transport channels for the two putative transmitters, although some overlapping functions were found. It is concluded that glial cells may play an important role in controlling the levels of neurotransmitters in the milieu of neurons. A preliminary report of this work has appeared (16). and then centrifuged at 1830 X (I for 10 min at 6". Supernatants were decanted and saved; tubes-were drained for 15 min at 6", and then weighed.
A specific gravity of 1 g per ml for the pellet was assumed in order to obtain packed cell volume from these wet weights.
Pellets were then dissolved in 0.  These activities were completely inhibited by 1 mM aminooxyacetic acid, thus establishing the pyridoxal dependencies of these enzymes, similar to the activity normally found in brain. The specific activities for glioma cells were 35s or less of those of adult mouse or rat brain. Examination of glutamate decarboxylase activity in homogenates taken from various portions of a growth curve of C-6 glioma cells ( Fig. 1) showed that the development of the activity was regulated.
The specific activity increased about lo-fold as cultures became crowded, and a nearly parallel increase of r4C02 formation by homogenates was found.
Relationships between the various activities at low cell density were retained: (a) COz acid was usually somewhat lower than in those without it; and (c) the cells never developed an ability to form y-aminobutyrate in the presence of aminooxyacetic acid. Other synthetic pathways utilizing glutamic acid were also active during these in vitro assays. When the complete electrophoretogram from one assay of a C-6 homogenate was passed through a strip counter, eight distinct peaks of radioactivity were observed. Two of these (in addition to glutamate), amounting to <O.l% each of the total radioactivity, also were present in electrophoretograms of the boiled enzyme blank; these exceeded glutamate in their mobility toward the cathode and were not identified.
One radioactive peak (-10% as many counts as y-aminobutyrate) from the C-6 assays was found to co-electrophorese with aspartic acid, and may have been that amino acid, although further identification was not attempted.
Two other minor radioactive spots, not in the control, were detected on the cathodic side of the origin.
In the anodic direction, in addition to y-aminobutyrate, only one other radioactive area was found; that substance co-migrated with glutamine and contained 1.2 times as many r4C counts as the y-aminobutyrate spot. The presence of 3.97 mM glutamine in the growth medium had not apparently repressed the ability of C-6 cells to form glutamine from glutamic acid. A similarly active synthesis of glutamine was found in assays of L-929 cells, in which no y-aminobutyrate was formed. In these in vifro assays C-6 homogenates formed a total of 5 different labeled products of which 12oj, was y-aminobutyrate. This suggested that glutamate might be utilized by these cells in several ways, in addition to synthesis of y-aminobutyrate.

Synthesis and Excretion of y-Aminobutyrate by Living Cells
The ability of living cells to take up glutamate for synthesis of y-aminobutyrate and glutamine from glutamate added to the medium was examined in cultures of C-6 and L-929 cells (Table  II).
The glial cells utilized glutamate in the medium at a very rapid rate to form y-aminobutyrate.
Over 70% of the y-aminobutyrate formed was lost from the cells into the incubation medium, although a cell to medium concentration ratio of >48fold was present at the end of the experiment.
The sum of the intra-and extracellular contents of labeled y-aminobutyrate indicated that the apparent specific activity of glutamate decarboxylase in intact cells was similar to that found in homogenates. In contrast, L-929 cells formed no detectable intra-or TABLE II Synthesis of r-aminobutyric acid by living cells Glioma cells (4 dishes) or L-929 cells (3 dishes) growing in 55-cm2 surface area dishes at 89 and 135 rg of cell protein per cm*, respectively, were grown in medium supplemented with 0.2 mM glutamate for 24 hours and then were given fresh growth medium containing 0.2 mM [U-l%]glutamic acid (0.75 mCi per mmole, 3 rCi per dish) for 2 hours of incubation at 37" under otherwise usual growth conditions. Cells and media were treated as described under "Materials and Methods," and r-aminobutyrate production was quantitated by high voltage electrophoresis of pooled extracts from both cells and media. Electrophoretograms of L-929 cell and medium extracts had no counts above machine background in the r-aminobutyrate area.  (15). The electrophoretograms from both cell types showed minor peaks of radioactivity associated with stained areas other than glutamate and y-aminobutyrate.
However, L-929 cells contained >90y0 of their nonglutamate radioactivity as glutamine, for an apparent specific activity (assuming glutamine to be also uniformly labeled) of glutamine synthetase of 0.414 pmoles per hour per mg of cell protein; this, despite the fact that the growth medium contained 3.97 mM glutamine.
The ability of L-929 cells to form glutamine from glutamate in the medium showed that this failure to synthesize y-aminobutyrate was not due to an inability to take up its precursor.
Although there was no control in this intact cell assay for rapid catabolism of y-aminobutyrate by L-929 cells, which would have masked its synthesis, the in vifro assays with cell-free preparations in which were included 1 mM unlabeled y-aminobutyrate as a trap, strongly supported the conclusion that this was not the case. Similarly, the yaminobutyrate trap in cell-free assays with C-6 homogenates might have masked catabolism of that molecule, but the assay with intact cells showed that glioma cells contained little, if any, y-aminobutyrate transaminase activity. From these data on glutamate and y-aminobutyrate metabolism it was clear that glial tumor cells, but not fibroblasts, had the potential to control extracellular concentrations of two putative neurotransmitters: glutamate, a putative excitatory neurotransmitter, by rapid uptake and metabolism to COS, glutamine, y-aminobutyrate, and other products; and y-aminobutyrate, by synthesis and excretion.
These findings prompted investigation of the following: (a) possible extension of this potential for regulation of levels of neuroactive substances to other putative neurotransmitter molecules; and (5) ability of glial cells to concentrate y-aminobutyrate and other potential transmitters from the medium.

Synthesis of Other Putative Neurofransmitters
Glioma cells in culture were capable (Table III) of formation of taurine, another putative neurotransmitter, from cystine in the medium; this is consistent with the suggestion (30) that taurine and y-aminobutyrate may be formed by the same enzyme. The glioma cells also formed 2-hydroxyethanesulfonic acid (isethionic acid), the deamination product of taurine which has been found to exist in squid giant axons (31), and to be synthesized by heart muscle (28) and brain, but not by skeletal muscle or liver (32). Taurine and isethionate together accounted for 19% of the acid-soluble 8% compounds in the cells, and isethionic acid was formed from 23% of the taurine synthesized. Taurine formation had an apparent specific activity which was similar to that for y-aminobutyrate in intact cell assays. The cells did not, however, form /3-alanine, also an inhibitory neuroactive substance and the decarbosylation product of an excitatory amino acid, from aspartate added to the medium.
After incubation with [U-Wlaspartic acid, among the nine distinct radioactive spots on the electrophoretogram, greater than 54% of the counts remained in aspartic acid. However, substantial formations of glutamate and glutamine were found. This glutamate, however, unlike the result when glutamic acid itself was present in the medium (at much higher concentration), was not utilized for synthesis of significant amounts of y-aminobutyrate.
This may reflect a priority for the cells under these growth conditions to form glutamate for detoxification of ammonia by synthesis of glutamine; only when glutamate was Acid-soluble materials extracted from the cells were identified by high voltage electrophoresis as described under "Materials and Methods." Up to 100,000 cpm of 3% and 93,000 cpm of 1% were recovered from the electrophoretograms.
Radioactivity not accounted for in the table co-migrated with cysteic acid, and unknowns X and Ye when the precursor was cystine, and with aspartate (54%) and four other spots (< 1. 8 Provided in relatively large quantities did secondary reactions, such as y-aminobutyrate formation, take place. Clearly from these data, formation of @-alanine from aspartate occurred to only a minor extent. In addition, under normal circumstances of growth (i.e. without glutamate or aspartate in the culture medium), it appears unlikely that cells would make, store, or excrete significant quantities of ,&alanine or y-aminobutyrate. Similar distributions of radioactive products were obtained in repeat experiments with C-6 cells at 84 and 101 pg of cell protein per cm2 surface area.
These data showed that glioma cells may act to efficiently remove glutamate and aspartate from extracellular spaces and to synthesize y-aminobutyrate, taurine, and isethionic acid. All of these substances have effects on neuronal or muscle membranes, or both (12, 32). In light of data showing active uptake of choline by C-6 (33) as well as apparently active uptake of some neurotransmitter candidates, including y-aminobutyrate, by a purified "glial" cell fraction from brain (lo), it seemed reasonable that glial cells might be able to remove such substances from the milieu of neurons, and thereby modify local neuronal excitability.
Synthesis, uptake and excretory ability of glia for taurine and y-aminobutyrate might provide the capability of control, by satellite cells, of over-all neuronal escitability.

In&x and Ejj7ux of y-Aminobutyrate and Taurine
Uptake Time Courses-When the glial tumor cells were incubated with radioactive y-aminobutyrate or taurine, it was found that they were able to take up these materials against a concentration gradient.
The time courses of uptake of y-aminobutyrate and taurine by two of the glioma cell lines are shown in Fig. 2. Uptake of y-aminobutyrate ( Fig. 2A) by both cell lines showed a very rapid influx to 5 or 6 min followed by a linear uptake out to at least 30 min.
In contrast, taurine uptake t,ime course (Fig. 2B) showed a slight lag followed by a linear increase in internal taurine concentration out to 30 min. The uptake was more rapid with C-2,, reaching an apparent cell to medium concentration ratio of 61 at 30 min. The 30-min transmembrane concentration was 30-fold for C-6. Concentrative uptake against a gradient apparently had occurred with both glial tumor lines. Although many cell lines and tissues have been shown to concentrate taurine (33-39), concentrative uptake of y-aminobutyrate has been reported to be peculiar to brain (40). Because glial uptake of either substance from the extracellular milieu of neurons could have important effects 011 their function, these uptake phenomena were given further evaluation.
Substrate Concentration EJects-The effects of extracellular substrate concentrations on the rates of y-aminobutyrate and taurine influxes were evaluated (Fig. 3) in the four cell lines of brain origin: C-6, C-21, and NT-l gliomas, and the uncloned RBF cell line (see "Materials and Methods"). For y-aminobutyrate, uptakes by C-6 and C-21 when Method a was used were the same as the results obtained with all the glioma lines with Method b. Differences between the glioma cell lines were negligible, and therefore the representative data obtained with NT-l cells are presented in Fig. 3A. y-Aminobutyric acid uptake was nearly linear in the range of 0.06 t.o 159 PM extracellular

concentrations.
However, consistent deviations from linearity (determined from the most accurate data at high substrate concentrations, dashed line in Fig. 3A) were found for the three glioma lines, but not for RBF cells. These findings led us to believe that two components of uptake were involved for the gliomas: a rapid and perhaps passive nonsaturable component and a much slower saturable influx.
When the linear component (picomoles of uptake per min per mg of protein per PM change in substrate concentration) from 100 to 150 pM concentration was subtracted from the data points, the lower curve of Fig. 3A was obtained, which described the kinetic behavior of the saturable uptake.
These derived data could be used to obtain reasonable approximations of the transport constant (Kr) and I',,, values via double reciprocal plots (Fig. 3A, inset). The data for both components of uptake for all the cell lines are given in Table IV where the similarity of all of these parameters for yaminobutyrate in the three gliomas is very apparent.
A saturable component of uptake was not obtained for the RBF line in which an apparent 2.3-fold concentration may have resulted from exchange diffusion (ti infra). Taurine uptake also apparently had two components in all four cell lines (Fig. 3B), although the saturable process contributed much more to total uptake for taurine than it did for yaminobutyrate at similar concentrations. It was found that C-6 cells accumulated taurine more rapidly than did NT-l, C-21 and RBF cells; since the data for the latter three cell lines were All assays were performed for 20 min at 37" in Solution 1 with cells grown in Linbro wells, with three wells at each substrate concentration and eight substrate concentrations (0.02 to 159 PM y-aminobutyrate and 0.02 to 200 pM taurine) as described for Method b under "Materials and Methods." Data for both substrates with NT-1 cells and for taurine with C-6 cells are those described in Fig. 3  essentially identical, only those for NT-1 cells and C-6 cells are given in Fig. 3B. When the nonsaturable components were subtracted, the derived data appeared to follow saturation kinetics in all cases (Fig. 3B, inset), and it was found that KT values for the four cell lines (Table IV) were essentially the same, while V,,,,p for C-6 may have been slightly lower than in the other cells. The data for C-2r cells agreed remarkably well with those for the other cell lines despite the wide variation in protein densities in those wells. In NT-l cells the KT values for y-aminobutyrate and taurine were essentially the same, although maximal transport rate (I',,,) was nearly 18 times greater for taurine.
Approximately the same relationship of V In&X values was found for the other two clonal lines (C-6, 17 times; C-21, 19 times as rapid for taurine), although in these lines the affinities for taurine were about three times greater than those for y-aminobutyrate.
When the internal to external concentrations were determined at various substrate concentrations (Fig. 4), it was found that the data were consistent with large nonsaturable and less active saturable components; that is, the apparent concentration gradient across cell membranes decreased markedly with increasing substrate concentration and approached a value of 1. In addition, the cell to medium concentration ratios for either substrate were sufficiently reproducible between the three glioma cell lines to allow averaging of the data from all of them, as shown in These ratios for three cell lines (C-6, C-2, and NT-l) were then averaged and this final average was plotted, with range in brackets.
Inset, detail of the curve and ranges at the lower concentrations.
Panel B, uptake data for taurine with C-6, C-21, NT-l, and RBF were treated as those for r-aminobutyrate above. Inset, detail of the curve and ranges at the lower concentrations.
Range brackets may not appear in both figures in a panel; where a point is presented in both figures in a panel without range brackets, all values fell within the radii of the datum shown. Fig. 4. The total uptake of y-aminobutyrate at an external concentration of 159 PM was nearly all due to the nonsaturable uptake process; the contributions of the saturable components were 8.2%, ll%, and 9% for C-6, C-21, and NT-l, respectively. In RBF cells a saturable process contribution at this substrate concentration was only 0.5q7, of total uptake, indicating that, if those cells contained the saturable component at all, it did not contribute significantly to y-aminobutyrate influx.
The data for taurine concentration ratios (Fig. 4B) similarly depended on external substrate concentration, although the degree of concentration, as well as the relative contributions of the saturable components, were greater for taurine than for y-aminobutyrate. At 209 j.hM extracellular taurine the average concentration ratio for all the cell lines was 4.74 when the C-6 data were included and 4.0 =t 2.6% when the C-6 data were excluded.
At that concentration saturable uptake contributed 49.4yo of total C-6 influx, but 71. 6,73.5,RBF,respectively. In experiments in which 4 mM unlabeled taurine was included with 20.7 pM [Wltaurine, the internal to external concentration ratios after 20 min were 0.52 for C-6 cells and 1.9 for C-21 cells, indicating for both lines a continuation of the trend illustrated in Fig. 4B. These data on concentration ratios confirmed the previous demonstrations of 2-component uptake systems for these amino acids; the saturable component in each instance was apparently capable of generating a considerable transmembrane concentration gradient.
Eflects of Blockers, Analogs, and Catecholamines-Evaluation of the specificity of the uptake mechanisms (Table V) showed that, with some minor exceptions, uptakes by the two glioma lines C-6 and C-2r were very similar.
Picrotoxin slowed yaminobutyrate uptake and stimulated taurine uptake by C-6 cells, while it had little effect on uptake of either substance by C-21 cells. Bicuculline had a modest effect on both transports in C-6, did not affect taurine uptake in C-2,, and appeared to reach a maximal effect on y-aminobutyrate uptake at 4Oa/, inhibition in C-21. Since, at this substrate concentration, in the TABLE V E$ects of blockers, analogs, and isoproterenol on r-aminobutyrate and taurine uptake by glioma cells in culture C-6 or C-21 glioma cells growing in Linbro wells were incubated in Solution 1 with the indicated test antagonist for 15 min before addition of either 15.9 pM y-amino-[1-Wlbutyrate or 20.65 PM [1,2J*C]taurine for 20-min uptake evaluation as described under "Materials and Methods." Protein content per cm* surface area of wells ranged from 68 to 103 rg for C-6 and from 40 to 82 rg for c-2,. Data are normalized from several different experiments and each datum represents the average of two or three wells, normalized for protein content.
At these substrate concentrations the saturable components for C-6 cells were 36.570 and 89.2'$!& of total uptake for r-aminobutyrate and taurine, respectively, and 36% and 94.7% of total uptake for the same compounds with C-2, cells.
Picrotoxin (100 PM absence of bicuculline, an estimated 36% of the total uptake was due to the saturable component, these data may indicate complete inhibition of the saturable component of y-aminobutyrate uptake in C-2r cells by bicuculline. Various aminobutyrate analogs were also found to affect uptake rates; the major findings were substantial inhibition of y-aminobutyrate influx by a-aminoisobutyrate and apparent stimulation of y-aminobutyrate transport in C-2, cells by D-or n-rY-amino-n-butyric acid. L-Isoproterenol, at a concentration above that necessary for maximal stimulation of cyclic adenosine monophosphate levels in these cells (41) had little if any effect on uptake of either amino acid by C-6 cells, but may have stimulated to a small degree the transport of at least taurine in C-21 cells. Uptakes, both saturable and nonsaturable, of y-aminobutyrate were markedly inhibited by 4 mM taurine, while 3 mM y-aminobutyrate appeared to have stimulated slightly the uptake of taurine. Thus, several data suggested that the transports for the two amino acids were occurring via different mechanisms.
It was interesting that fi-alanine caused marked inhibition of transport of both amino acids, apparently interfering with nonsaturable as well as saturable processes.
Tests for Exchange Diffusion-An apparent concentrative uptake of radioactive amino acids might occur by exchange diffusion with a large internal pool of the same compound (i.e. a one-for-one exchange of internal unlabeled molecule with an external labeled molecule).
Although previous data showed that under normal growth conditions C-6 cells did not synthesize y-aminobutyrate, because they were capable of doing so, those data could not rule out the possibility of an exchangeable pool within the cells. Such a pool was also possible for taurine since synthesis of that compound did occur in the cells under normal growth conditions. One test that has been used for evaluation of exchange diffusion (42-44) is preloading of cells with the unlabeled substrate before addition of labeled molecule to the exterior of the cells. It is reasoned that the size of the exchangeable pool would be thus increased and the rate of exchange diffusion would be enhanced by preloading.
When a similar preloading technique was used with C-6 cells (Table VI), it was found that preloading with either y-aminobutyrate or taurine decreased transport of the amino acid. This result was not consistent with exchange diffusion and, in fact, was consistent with the cells having a limited capacity for total content of these compounds.
An additional test of the exchange diffusion mechanism involved use of the fact that the glioma cells were able to excrete both y-aminobutyrate and taurine into the extracellular medium. If the cells contained large exchangeable pools of y-aminobutyrate or taurine, these pools could be depleted by preliminary incubations without substrate prior to assay for uptake rate with the radioactive molecules.
For such preunloading studies cells were incubated in Solution 1 without either substrate for varying periods before fresh medium and labeled substrate were added for the usual uptake experiment.
It was reasoned that, if progressive release of y-aminobutyrate and taurine took place into the medium, this should have decreased the exchangeable pools with time and caused the apparent concentration rate due to exchange diffusion to progressively decrease. That this, in fact, did not happen is shown in Fig. 5. With neither cell line was there a change in the uptake rate of y-aminobutyrate with increasing time in the first incubation.
In contrast, there appeared to be a small, but definite increase in the rate of taurine uptake by both cell lines. These data strongly argue against exchange diffusion as the mechanism of uptake.
Effect of Metabolic Inhibitors on Taurine and y-Aminobutyrate Uptake-The data of Table VII show that there was little effect on uptake of prior incubation of C-21 or C-6 cells with sodium azide, 2,4-dinitrophenol, or ouabain. At these substrate concentrations, the saturable components of uptake (legend to Table VII) were sufficiently large so that, if the inhibitors affected only those components, an effect would be expected to be demonstrable.
Using 10 mM sodium azide in C-21 cells, there was no difference in the 20-min uptake of either amino acid when the prior incubation with the inhibitor was 5 to 25 min (data not shown).
It is difficult to explain concentrative uptake without an apparent requirement for energy; the limited effects of metabolic inhibitors on the uptake processes may reflect the availability of large stores of high energy compounds in these cells or the functioning of a chemiosmotic uptake mechanism, or both (45, 46). That taurine uptake showed some inhibition by dinitrophenol not found with y-aminobutyrate is consistent with these explanations (more rapid transport of taurine than y-aminobutyrate and a higher percentage of that transport via a saturable, concentrative, and therefore probably energy-requiring, process). When incubations were carried out in 6.9 meq per liter of sodium ion or at 0.7", transport of both amino acids was markedly curtailed.
This degree of inhibition could not, have occurred by interruption of only the saturable component; although it is not known which component remained, it was clear that low sodium and low temperature affected both uptake mechanisms very substantially.  VII Efects of metabolic and nonspecific uptake inhibitors on y-aminobutyrate and taurine uptake by glioma cells in culture C-6 or C-2, glioma cells growing in Linbro wells were incubated (Experiments 1 to 3) with the indicated inhibitor for 30 min, or (Experiment 4) in low sodium medium (Solution 2, 6.94 meq of Na+ per liter) for 15 min, or (controls) in Solution 1 for 15 min before addition of either r-amino[l-Wlbutyric acid (15.9 pM final concentration) or [l ,2-i4C]taurine (20.7 JAM final concentration) for 20-min uptake evaluation as described under "Materials and Methods." Some wells of C-6 (Experiment 5) were incubated in Solution 1 with either 6.0 (low substrate) or 207 (high substrate) JAM [r4C]taurine or 6.0 or 159 PM [14C]y-aminobutyrate for 20 min at either 37" or 0.7" (ice water bath) and then processed as usual. Protein content of wells ranged from 63 to 143 fig per cma surface area for C-6 cells and from 60 to 98 rg for C-21 cells. Data are normalized from several different experiments and each datum represents the average of two or three wells normalized for protein content. The saturable uptake of -y-aminobutyrate in control cultures contributed 38, 36.5, and 8.1% to total uptake at 6.0, 15.9, and 159 pM, respectively, in C-6 cells; and 32, 36, and 11% at the same concentrations in C-2i cells. Similar values for taurine uptake were 92. 3,89.2,and 49.4y,at 6.0,20.7,and 207 PM,respectively,and 96.8,94.7,and 71.6yo Table II had shown that efflux into the medium accounted for a large proportion of the y-aminobutyrate synthesized by intact C-6 glioma cells. We next evaluated whether putative transmitters which had been concentrated from the external milieu were similarly subject to efflux and whether such efflus had any relationship to P-adrenergic stimulation of cyclic adenosine 3' : 5'-monophosphate levels in these cells. Initially, the effect of time upon efflux of the neurotransmitters was studied.
C-6 glioma cells (Fig. 6A) lost y-aminobutyrate to the medium linearly for 30 min of the second incubation with a rate approsimately one-third that of total uptake (at this substrate concentration about 35% of y-aminobutyrate uptake was due to the saturable component).
It was not determined whether an equilibrium situation could be reached with an excess y-aminobutyrate concentration inside the cells, although the cell to medium ratio at 507, efflux was still in excess of 300. In contrast, these cells, after loss of about 1570 of concentrated taurine, did not allow any further net efflux of that compound. Cell to medium concentration ratios averaged 1535 throughout the 32.min test of taurine efflux. Similar results were obtained with C-21 cells (Fig. 6B) except that y-aminobutyrate was extruded more rapidly than with C-6 and net excretion of taurine was also more rapid initially, but then decreased with time. y-Aminobutyrate escretion by C-21 reached a plateau at an excretion of about 75% of the concentrated molecules, at which 1777 point the cell to medium concentration ratio was 84 and the external concentration was 0.703 PM. Previously determined uptake capabilities did not account entirely for the plateau, as the net excretion rate between 10 and 20 min in Fig. 6B was 22 pmoles per min per mg of cell protein, while, at that external concentration, y-aminobutyrate uptake would have been 2.08 pmoles per min per mg of protein (65% saturable).
That equilibrium was reached with a transmembrane gradient of more than @fold may indicate that a control mechanism exists for the maintenance of either intracellular or extracellular levels of y-aminobutyrate.
Similar calculations of transmembrane gradient and uptake rates for taurine (Fig. 6B, lower curve) show a maximum external concentration of 0.703 PM; precisely that found with y-aminobutyrate.
The internal concentration at 7 min in the second incubation was 332 &M taurine, or 472-fold greater than the extracellular concentration; after 32 min of the second incubation the internal concentration was 344 PM and the cell-medium ratio was 729. At an extracellular concentration of 0.703 PM the taurine uptake rates would have been 26 pmoles per min per mg of protein (saturable) and 0.58 pmoles per min per mg of protein (nonsaturable).
The rate of decrease in net efflux from 22 to 32 min in the second incubation was estimated from the curve to be 0.3447, per min of the total uptake from the first incubation or 26.8 pmoles per min per mg of protein.
Thus, the decrease in net efflux of taurine seen in Fig.  6B, is entirely consistent with simple reuptake from the medium. The cells not only maintained, but worked to increase, tissuemedium ratios of >300.
In order to study the process of efflux in more detail, the effects of various washes and incubation media were evaluated (Table  VIII).
The basal excretions of y-aminobutyrate and taurine are shown in Experiment 1. Inclusion of 10 PM L-isoproterenol in the medium for the second incubation resulted in a small (26%) decrease in y-aminobutyrate efflux and a 2770 increase in the equilibrium level of taurine efflux. Incubation in substrate-free medium (Experiment 3) resulted in a 44% inhibition of y-aminobutyrate efflux, indicating that exchange diffusion (as previously defined) may have contributed significantly to loss of that amino acid at 3 mM extracellular concentrations.
In contrast, taurine efflux was unaffected by a change in substrate concentration in the external medium. Incubation in low Na+ media without substrate (Experiment 4) resulted in little change (from Experiment 3) in y-aminobutyrate efflux. With taurine efflux, however, incubation in low Na+ medium profoundly inhibited the loss; the differences shown in Table VIII are understatements of the actual effect since normal reuptake of excreted taurine was essentially eliminated in low Na+ medium.
Thus, taurine, but not y-aminobutyrate, efflux by C-2i glioma cells appeared to be a Na+-dependent process. In other efflux experiments with C-6 cells, it was found that isoproterenol affected the rate of efflux neither of accumulated y-aminobutyrate, nor of total counts from cells grown for four generations in [UJ4C]glucose. DISCUSSION These data show that rat glial tumor cells are capable of: (a) functioning to remove glutamic and aspartic acids from extracellular milieu; (b) synthesis of the neuroactive molecules yaminobutyric acid, taurine, isethionic acid, glutamate, and glutamine; (c) concentrative uptake of y-aminobutyrate and taurine by apparently different mechanisms and probably separate pathways; and ( Relevance of Glioma Cells to Glial Function-The significance of these data to brain function depends on the degree to which normal glial function is mirrored in these glioma cells. That the glioma cells actively transport y-aminobutyrate, a property thought to be exclusive for brain (40), suggests that this may be a differentiated function retained by these cells. Data of Henn and Hamberger (10) apparently demonstrate such an uptake phenomenon in partially purified glial cells. That the y-aminobutyrate uptake kinetics of the three glioma cell lines are virtually identical, although the lines originated from different tumors and have different growth characteristics and biochemical capabilities, lends support to the claim that this function is characteristic of normal glial cells. In line with these arguments are the apparent inhibition of y-aminobutyrate uptake by bicuculline, an alkaloid known to inhibit y-aminobutyrate action on its receptors (53, 59), and the absence of the saturable uptake may be important to brain function, that capability appears to be ubiquitous rather than specific for neural tissue; the presence of at least two components of taurine uptake in RBF cells, neuroblastoma cells (33), and some other cultured cells has been reported (37-39) in addition to its uptake by slices of brain from newborn and adult rats (34, 60). Additional differentiated functions which have been shown for one or more of these glioma lines include synthesis of S-100 protein (19), presence of the enzymes adenosine 2':3'-cyclic monophosphate 3'-phosphohydrolase (61), a putative marker for myelin-forming cells (62) ; catechol 0-methyltransferase and .aonoamine oxidase (63), synthesis of glial fibrils (20), and of acid mucopolysaccharides (64) and the presence of the catecholamine-mediated stimulation of 3':5'-cyclic AMP accumulation (41). The cell lines C-6 and C-21 are presently referred t,o by many authors as astrocytoma cells, although the original tumor cells were thought to resemble more closely oligodendroglia (20). That origin could explain the presence of the phosphohydrolase activity as well as the ability to concentrate taurine (1 I). In addition, we have found3 that neither these glioma cells nor RBF cells contain measurable amounts of the glial fibrillary acidic protein found in brain astrocytes (65) and in some of the cells in cultures of fetal rat brain cells.3 Thus, these glial tumor cells may be models of differentiated oligodendrocytes which have retained some of the functions of that cell type in the intact nervous system. Assuming this to be true, the impact of these data upon brain function may be assessed.
Functions of Glial Cells in Intact Nervous System-The data of this report lead to the conclusion that glia might affect the levels of neuroactive substances in the extracellular spaces of the nervous system. Glial control over the extracellular milieu could occur at several different levels of functional complexity. As has been noted previously (2, 3), the addition of glial cell volume to that of the extracellular space in brain (by the presence of purely passive glial cell membranes) would dilute substantially the concentration of neuroactive substances around neurons. To be effective, a disposal mechanism would also be required, although this could be as simple as diffusion into capillaries. A clearly more effective glial buffering of the extracellular fluid would be provided by an ability of the glial cell to concentrate neuroactive substances, and to maintain a transmembrane gradient, capabilities which have been suggested by other authors (5,10,11,38,66). Data of this report show these abilities for glioma cells with y-aminobutyrate and taurine, both of which are present in brain tissue in millimolar concentrations (32,34,52,60,67,68) and have been considered by many authors to be more or less likely candidates as neurotransmitters (an excellent review for taurine is Ref. 69). As a test of the ability of cells with these characteristics to influence brain function, the following should be noted. The concentration of taurine in cerebrospinal fluid, which may be representative of extracellular fluid, has been measured at 5 to 35 PM (69). In this range the saturable component of taurine uptake for NT-1 cells comprised 95 to 92% of total taurine uptake.
If the same concentration range of y-aminobutyrate were encountered by NT-l cells, saturable uptake would comprise 50 to 22% of the total uptake.
The Kr values for the saturable processes fall within this range for both amino acids with all the glioma lines. Clearly, these uptake phenomena may function physiologically to remove these neuroactive substances from the immediate milieu of neurons. 3 A. Bignami and B. K. Schrier, unpublished data.

1779
Further increments in complexity of glial function seem to require that the cells respond to a stimulus.
Such a response would demand the function of a sensor, a transducer, and an effector; in the simplest case all three functions might reside within the same macromolecule.
Examples of such stimulus responses, with increasing complexity, would be: (a) maintenance by glial cells of a particular extracellular concentration(s) of one or more neuroactive substances.
(b) In response to neuronal activity or changed extracellular levels of a neuroactive substance(s), or both, glia might modify their influx or efflux rate for that or some other neuroactive substance.
This response would function to complement or counteract the effects of the local milieu on neuronal membranes.
(c) Several spatially related glial cells might respond to local or general hyperor hypoactivity of neurons (through sensing of K+ ions, neuroactive substances, electrical, or other changes in their domains) by the uptake or excretion (or both) of neuroactive substances within those domains in order to modify in concert the state of over-all excitability.
(d) Glial cells might communicate with each other concerning over-all excitability of the nervous system, and then take appropriate damping or activating responses. The data of this report do not define the precise level at which glial cells function.
However, it is perhaps not unrealistic to assume that cells which comprise approximately 90% of the nervous system cells might be able to exert considerable control over the status of that system.
Probably the ultimate possibility for glial function would be that glial cells control brain function absolutely, including memory, etc., with neurons only as messengers, and all information storage and processing occurring in glial cells. While this possibility seems very remote at present, there are, as far as we know, no data which preclude its demonstration at some future time. It is reasonable to conclude that glial function may contribute substantially to the working of the nervous system.