Preferential Utilization of Acetate by Astrocytes Is Attributable to Transport

MATERIALS AND METHODS

Astrocyte cultures. Primary cultures of astrocytes were prepared from cortices of 2- to 3-d-old rats (Frangakis and Kimelberg, 1984). Brain tissue was removed from pups. Cells were plated in 24-well trays or 100 mm plates (Falcon) at a density of 10–20,000 cells/cm². Cells were maintained in Eagle’s Medium supplemented with 10% fetal bovine serum, penicillin, and streptomycin at 36°C in a humidified incubator with 5% CO₂ and 95% air and were used in experiments after 3–4 weeks. At this point the cells had formed a confluent monolayer of high density; glial fibrillary acidic protein immunostaining indicated that >95% were astrocytes.

Preparation of synaptosomes. Adult rats were lightly anesthetized with CO₂ and decapitated. Brains were rapidly removed and cooled in 0.9% saline on ice. The cerebral cortex was dissected, weighed, and homogenized in 0.32 m sucrose. Synaptosomes that are relatively free of extrasynaptosomal mitochondria and glial contamination were prepared by the method of Löscher et al. (1985). Synaptosomes prepared in this way have been assayed for glutamine synthesis (Battaglioli and Martin, 1990). They produce glutamine at <5% of the rate found in astrocyte cultures (Waniewski, 1992), indicating low astrocytic contamination.

Measurement of [¹⁴C]acetate oxidation to¹⁴CO₂. To measure acetate metabolism, astrocytes or synaptosomes were incubated at 37°C in 24-well culture plates (Falcon) with 0.2 mm1-[¹⁴C]acetate and sodium salt (54.7 mCi/mmol; Sigma, St. Louis, MO) in oxygenated, glucose-containing HEPES-buffered HBSS (Martin and Shain, 1979). Individual wells were fitted with specially made Plexiglas adaptors and serum tube stoppers that held a small cup containing 50 μl benzethonium hydroxide (Sigma) to trap the¹⁴CO₂. The reaction was stopped by injecting 0.2 m citric acid through the stopper into the incubation medium. The tray was heated at 60°C for 15 min to volatilize CO₂ dissolved in the medium, and the trapped¹⁴CO₂ was analyzed by scintillation counting. Wells without cells and addition of citric acid to cells at the start of the incubation period gave the same results and served as blank controls.

Measurement of acetyl-CoA synthetase activity. Acetyl-CoA synthetase was measured by following the formation of [¹⁴C]acetyl-CoA from 1-[¹⁴C]acetate (0.6 mm) and coenzyme A (0.6 mm) in the presence of ATP (3 mm) and Mg²⁺ in potassium phosphate buffer (50 mm, pH 8.0) (Reijnierse et al., 1975). The product was separated from the substrate by collection on ion exchange filters (DE81; Whatman, Maidstone, UK; 2.4 cm) and washed with 1 ml of acetic acid (2%) (Roughan and Ohlrogge, 1994).

For assay, astrocytes were scraped from 100 mm tissue culture plates (Falcon) with a rubber policeman and collected in buffer on ice. Cells, cortical tissues, and synaptosomes were homogenized in Tris-buffered, 0.32 m sucrose containing MgCl₂ (27 mm), 1.25% Triton X-100, and bovine serum albumin (BSA) (4.5 mg/ml). Enzyme activity was measured by incubating homogenates at 37°C for 15, 30, 45, and 60 min and normalized to protein content (Lowry et al., 1951).

Analysis of transport. Acetate transport was measured under conditions similar to those for measuring radiolabeled lactate transport by the monocarboxylate transporter (Poole and Halestrap, 1993). The incubation medium was oxygenated HEPES-buffered balanced salt solution without glucose or additional anions other than chloride (in mm: 10 HEPES, 150 sodium chloride, 5 potassium chloride, 1 calcium chloride, and 1 magnesium chloride, pH 7.4, with 1 NaOH; Garcia et al., 1994).

Multiwell (24) culture plates containing astrocytes were individually corked and transferred to a temperature-controlled water bath at 4°C. Cells were rapidly rinsed (3–5 sec) with five washes of buffer and then either incubated immediately with acetate or allowed to preincubate for 15 min at 4°C. The preincubation medium was removed with three washes before adding [¹⁴C]acetate. When tested, transport inhibitors were added to the cells in the medium containing the 1-[¹⁴C]acetate. To stop the reaction, the cells were washed three times with fresh ice-cold buffer. The background uptake was determined by incubating cells for 0 min. This was done by washing with fresh ice-cold buffer immediately after adding the 1-[¹⁴C]acetate. The cells were then treated with 300 μl of 1 m NaOH to stop metabolism and prevent evolution of acetic acid or volatile acidic metabolites. After sitting overnight at 37°C, solubilized cells were transferred to scintillation vials, and the wells were rinsed two times with 450 μl of water. An aliquot was taken to measure the protein content (Bradford, 1976) and the remainder was analyzed for ¹⁴C by scintillation counting. The 0 min values for uptake were then subtracted from the 15 sec time points in each experiment, and the difference was expressed in moles of acetate uptake per min × milligrams of protein.

Synaptosomes were incubated with 1-[¹⁴C]acetate for the indicated times at 4 or 37°C, diluted into ice-cold medium containing α-cyano-4-hydroxycinnamate (CHC) (10 mm), and immediately centrifuged through silicone oil (SF1250, General Electric, Waterford, NY) at room temperature using the inhibitor–stop procedure (Poole and Halestrap, 1993) to terminate uptake and prevent efflux. When the centrifugation step was done in a cold room, the oil viscosity was too high to allow the synaptosomes through. The 0 min points were obtained by adding an aliquot of the synaptosome suspension directly to the centrifuge tubes containing 1-[¹⁴C]acetate and CHC and immediately starting the centrifuge for each sample. The incubation medium and oil layers were removed by vacuum aspiration. The synaptosomal pellets were dissolved overnight in 1 m NaOH and neutralized with 1 m HCl, and ¹⁴C was determined by scintillation counting.

[¹⁴C]Glutamate uptake was measured in parallel with 1-[¹⁴C]acetate uptake using the same synaptosomal suspensions.

Determination of sucrose space.[¹⁴C(U)]Sucrose (540 mCi/mmol; Amersham, Arlington Heights, IL) was purified by TLC (Haldorsen, 1977) and used to estimate the entrapped fluid volume in the synaptosomal pellet. Synaptosomes in suspension at five different dilutions were incubated at 4°C with [¹⁴C(U)]sucrose and centrifuged through silicone oil with procedures identical to those used for the 0 min incubation of synaptosomes with [¹⁴C]acetate.

Protein analysis. Protein was measured by the Coomassie blue dye-binding assay (Bradford, 1976) or by the method of Lowry et al. (1951) with BSA as a standard. Estimates of protein obtained with the Lowry method were approximately two times higher than those obtained with the Bradford method in the same synaptosomal suspensions.

Statistics. Experimental treatment groups were compared by one-way ANOVA. When a significant F value was found at the 0.05 level, individual comparisons were made a posteriori byt tests with the Bonferroni correction.

RESULTS

DISCUSSION

The present study demonstrates first that astrocytes produce¹⁴CO₂ from 1-[¹⁴C]acetate much more rapidly than do synaptosomes. The difference in acetate utilization was not attributable to the selective cellular localization of acetyl-CoA synthetase, because acetyl-CoA synthetase activity is greater in synaptosomes than in astrocytes. The preferential utilization of acetate by astrocytes is attributable to much more rapid uptake of acetate by astrocytes than synaptosomes. Transport of acetate appears to be mediated by a carrier having many properties in common with the proton-coupled monocarboxylate transporter. Synaptosomes do not have detectable acetate uptake other than that caused by diffusion under any of the conditions used to measure uptake by astrocytes. The metabolism of acetate to CO₂ was significantly inhibited by the specific monocarboxylate transport inhibitor CHC in both preparations.

For these studies, functionally intact and relatively pure preparations were required to allow direct comparison of astrocytes and neurons. Two different preparations were used: rat cortical astrocytes in primary cell culture and rat brain cortical synaptosomes. Neuronal cell cultures are not of sufficient density to permit reliable detection of acetate transport, and the presence of astrocytes in such cultures would compromise the goal of the study. Contamination with glia and extrasynaptosomal mitochondria is minimal in the synaptosomal preparation used (Dodd et al., 1981; Löscher et al., 1985). Our previous studies confirmed that the astrocytic contamination was low because the rate of glutamine production in these synaptosomes (Battaglioli and Martin, 1990) is <5% of that of cultured astrocytes (Waniewski, 1992). It was important also to eliminate free mitochondria, because they may accumulate and metabolize acetate.

Earlier studies demonstrated that there was much more incorporation of acetate carbon into glutamate, glutamine, and TCA cycle intermediates in astrocytes than in neurons (Sonnewald et al., 1993). The current study demonstrates that acetate metabolism to CO₂ is much more rapid in astrocytes than synaptosomes. The metabolism of other energy substrates to CO₂ has been compared in similar preparations (McKenna et al., 1993). The rate of¹⁴CO₂ production from [¹⁴C]acetate by astrocytes in the present study, 26 nmol/hr per milligram of protein with 0.2 mm acetate is faster than the rate reported for any of the tested substrates at 1 mm. These included malate, lactate, glucose, glutamine, and 3-hydroxybutyrate (McKenna et al., 1993).

Both astrocytes and synaptosomes have significant and comparable levels of acetyl-CoA synthetase activity. The similarity of the activity in these two preparations to that found in whole cortex homogenates argues against the specific cellular compartmentation of this enzyme. Other investigators have reported high activities for acetyl-CoA synthetase in mammalian brain synaptosomes (Tuček, 1967b; Szutowicz and Łysiak, 1980), but this is the first direct comparison of acetyl-CoA synthetase activity in astrocytes and synaptosomes. Thus, the preferential utilization of acetate by astrocytes is not explained by differences in the cellular distribution of acetyl-CoA synthetase.

Research on acetate transport in the nervous system has focused primarily on movement across the blood–brain barrier; previous studies suggest that uptake occurs by passive diffusion (Oldendorf, 1973;Terasaki et al., 1991). Such nonspecific entry would not likely be differentially expressed in neurons versus astrocytes. Our findings of time dependence, saturability, and inhibition by selective inhibitors provide evidence that intact astrocytes accumulate acetate by a specific carrier mechanism. This carrier displays many of the properties described for the proton-coupled monocarboxylate transporter. Uptake is transaccelerated by preloading cells with acetate or l-lactate, and the pharmacology of inhibition is similar to that of the monocarboxylate transporters studied in erythrocytes, liver, skeletal muscle, and heart cells (Poole and Halestrap, 1993), and the cloned transporters MCT1 and MCT2 (Garcia et al., 1994).

Fluoroacetate is toxic because it is metabolized to fluorocitrate, which inhibits the aconitase step in the TCA cycle (Peters, 1963). Fluoroacetate has been shown to selectively inhibit the TCA cycle of astrocytes over neurons (Keyser and Pellmar, 1994), but the mechanism for its selectivity is not known. The chemical similarity of acetate and fluoroacetate and the ability of fluoroacetate to effectively inhibit acetate uptake suggests that it may also be accumulated preferentially by astrocytes.

CHC is an effective inhibitor of [¹⁴C]acetate uptake but does not inhibit completely. One possible explanation for this is that acetate enters astrocytes by an additional pathway. The partial inhibition produced by DIDS and ethacrynic acid suggests that some of the acetate may enter by another carrier such as an anion exchanger.

The transport of lactate has been compared previously in astrocytes and neurons. Lactate efflux from astrocytes is seven times faster than from neurons (Walz and Mukerji, 1988). Although efflux was not blocked by CHC, the authors concluded that lactate efflux occurred by proton–lactate cotransport. Transport measured by lactate-induced acidification has shown that cultured neurons and astrocytes have the same K_m (0.38 mm), but neurons have an 80% higher V_max (Nedergaard and Goldman, 1993). CHC inhibited [¹⁴C]lactate transport in neuronal cultures (Dringen et al., 1993) and C6 glioma cells but not in astrocyte cultures (Dringen et al., 1995). Lactate transport by rat astrocytes recently has been shown to be inhibited by CHC and to have kinetic properties similar to those described here for acetate (Bröer et al., 1997). Thus, it appears that lactate transport is not restricted to either astrocytes or neurons.

The role of monocarboxylate transport in the nervous system is not well understood. Tsacopoulos and Magistretti (1996) have proposed a mechanism in which astrocytes metabolize glucose to lactate. Neurons accumulate lactate and metabolize it for use in the TCA cycle. If acetate and lactate are carried on the same transporter, it is not clear why neurons would accumulate lactate but not acetate. Although the evidence is indirect, our transacceleration experiments indicate that astrocytes also transport l-lactate. Two monocarboxylate transporters have been cloned (Garcia et al., 1994) that differ in organ distribution and have slightly different substrate and inhibitor affinities. Recent evidence indicates that MCT1 is expressed in astrocytes and MCT2 in neurons (Bröer et al., 1997). Additional members of this family may exist with differing substrate specificities (e.g., a lactate transporter in neurons that does not transport acetate). Under physiological conditions, the monocarboxylate transporters may be regulated by the transacceleration mechanism. Cells that produce lactate may have a greater capacity to accumulate acetate by heteroexchange with lactate. Cells that rapidly metabolize lactate may not be capable of such transacceleration. Similarly, cells that metabolize pyruvate via the TCA cycle may not transaccelerate acetate uptake. Mouse cortical astrocytes and neurons are capable of rapidly oxidizing pyruvate to CO₂ (Fitzpatrick et al., 1988). Such a process does not likely explain why pyruvate did not transaccelerate acetate uptake in the present study because metabolism was minimized by the low incubation temperature. Pyruvate may not be accumulated by astrocytes.

Several lines of evidence suggest that acetate transport may be important in the nervous system. Extracellular acetate in the brain may be derived from at least three sources. Transport from the peripheral circulation is possible because acetate is produced by anaerobic bacteria in the gut (Wolin and Miller, 1983) and is normally present at about 0.2–0.3 mm in the blood. Blood levels are influenced by diet and the microbial species in the intestine. Acetate and other minor short-chain fatty acids generated in the gut are used as an energy source and account for 5–10% of total caloric use in humans (McNeil, 1984).

Ethanol is metabolized to acetate. Plasma acetate levels increased from 0.3 to 1 mm for 4 hr after a single intravenous dose of 0.5 gm/kg ethanol in humans (Orrego et al., 1988). Although brain acetate levels have not been measured after alcohol consumption,¹⁴C from intraperitoneally injected [¹⁴C]ethanol is incorporated into glutamate, glutamine, and GABA in hamster brain (Roach and Reese, 1972).

Finally, acetylcholine (ACh) is rapidly metabolized by extracellular cholinesterases to acetate and choline in the brain. If cholinergic neurons are inefficient in recapturing acetate lost during ACh release, they would tend to consume acetyl-CoA to maintain their supply of ACh. A cycle may exist, like the glutamine cycle, that allows for the efficient reutilization by astrocytes of acetate lost from cholinergic neurons.

In conclusion, evidence is provided here for a previously undescribed metabolic process in the nervous system. Acetate is selectively accumulated from the extracellular space by a monocarboxylate transporter-like process in astrocytes, in which it is metabolized to CO₂, providing a potential source of energy in the brain. Incorporation of [¹³C] from acetate into glutamine is used to study “glial” metabolism in the nervous system (Badar-Goffer et al., 1990; Cerdan et al., 1990; Chapa et al., 1995;Sonnewald et al., 1996). The present results provide support for this approach by demonstrating that the biochemical basis for the selective utilization of acetate is attributable to much greater uptake by astrocytes than by neurons.