Energy Metabolism in Astrocytes and Neurons Treated with Manganese: Relation among Cell-Specific Energy Failure,Glucose Metabolism,and Intercellular Trafficking Using Multinuclear NMR-Spectroscopic Analysis

Abstract

A central question in manganese neurotoxicity concerns mitochondrial dysfunction leading to cerebral energy failure. To obtain insight into the underlying mechanism(s), the authors investigated cell-specific pathways of [1–¹³C]glucose metabolism by high-resolution multinuclear NMR-spectroscopy. Five-day treatment of neurons with 100-μmol/L MnCl₂ led to 50% and 70% decreases of ATP/ADP and phosphocreatine–creatine ratios, respectively. An impaired flux of [1-¹³C]glucose through pyruvate dehydrogenase, which was associated with Krebs cycle inhibition and hence depletion of [4–¹³C]glutamate, [2–¹³C]GABA, and [¹³C]glutathione, hindered the ability of neurons to compensate for mitochondrial dysfunction by oxidative glucose metabolism and further aggravated neuronal energy failure. Stimulated glycolysis and oxidative glucose metabolism protected astrocytes against energy failure and oxidative stress, leading to twofold increased de novo synthesis of [3–¹³C]lactate and fourfold elevated [4–¹³C]glutamate and [¹³C]glutathione levels. Manganese, however, inhibited the synthesis and release of glutamine. Comparative NMR data obtained from cocultures showed disturbed astrocytic function and a failure of astrocytes to provide neurons with substrates for energy and neurotransmitter metabolism, leading to deterioration of neuronal antioxidant capacity (decreased glutathione levels) and energy metabolism. The results suggest that, concomitant to impaired neuronal glucose oxidation, changes in astrocytic metabolism may cause a loss of intercellular homeostatic equilibrium, contributing to neuronal dysfunction in manganese neurotoxicity.

Keywords

Manganese neurotoxicity Energy metabolism Astrocytes Neurons Glucose NMR spectroscopy

Manganese, required for the activity of many enzymes, is essential for normal physiologic functioning (Underwood, 1981). Chronic exposure to manganese, however, leads to neurotoxicity (manganism) and is characterized by neuropsychiatric symptomatology that often resembles Parkinson disease (Huang et al., 1989; Mena et al., 1967). Neuropathologic changes are manifested mainly in the globus pallidus and include neuronal loss, gliosis, and Alzheimer type II astrocytosis (Pentschew et al., 1963; Yamada et al., 1986). The primary basis for manganese neurotoxicity remains unclear, but changes of both astrocytic and neuronal integrity indicate that several biochemical events may be involved. The accumulation of manganese in mitochondria (Wedler et al., 1989) suggests the possibility of mitochondrial derangements. As mitochondrial function is tightly coupled to cellular energy production, compromised brain energy metabolism leading to decreased oxidative phosphorylation may be a central feature of manganese neurotoxicity.

Astrocytes sequester manganese (Aschner et al., 1992; Tholey et al., 1990; Wedler et al., 1989) and are considered as the likely initial targets of manganese neurotoxicity (Henriksson and Tjalve 2000; Spranger et al., 1998). The glial localization of pyruvate carboxylase (PC), the key anaplerotic enzyme in the brain (Shank et al., 1985), and glutamine synthetase (GS) (Norenberg and Martinez-Hernandez, 1979), suggest an important role of these cells in brain energy metabolism. In this context, changes in astrocyte metabolism may either modify primary neuronal injury or induce secondarily neuronal dysfunction. In contrast to other organs, brain energy metabolism depends almost exclusively on the oxidation of glucose (Gjedde and Marrett, 2001). Via tricarboxylic acid (TCA) cycle activity, glucose metabolism also leads to the biosynthesis of neurotransmitters such as glutamate, aspartate, and γ-aminobutyric acid (GABA). The metabolic activities of the energy-producing pathways are under strict coordinated control and involve glycolysis, the TCA cycle, oxidative phosphorylation, and the pentose–phosphate pathway. So far, however, no data are available from simultaneous studies to correlate the mechanisms that link manganese-induced mitochondrial dysfunction and energy failure with glucose metabolism. The use of stable isotope-labeled precursors is a common technique used to study biochemical pathways. In particular, ¹³C-labeled substrates and the high chemical specificity of ¹³C-NMR permits the identification of ¹³C label in specific positions of given compounds and allows for the simultaneous calculation of substrate concentrations and multiple inter- and intracellular metabolic fluxes in the intact brain and cell cultures in a comprehensive manner, which is not possible with classic studies using ¹⁴C-labeled isotopes (Bachelard and Badar-Goffer, 1993; Cruz and Cerdan, 1999; Gruetter et al., 1994).

In the present study, we applied multinuclear NMR spectroscopy and an in vitro approach using primary astrocytes, neurons, and their cocultures to analyze metabolic pathways coupled to oxidative phosphorylation by simultaneous determination of high-energy phosphates, cell-specific de novo synthesis of metabolites derived from anaerobic and oxidative [1–¹³C]glucose metabolism, and the involvement of intercellular trafficking after manganese treatment.

MATERIALS AND METHODS

Cell cultures

Primary cell cultures of cortical astrocytes, cortical neurons, and cocultures of both cell types were obtained as described previously (Booher and Sensenbrenner, 1972; Zwingmann et al., 2000). Astrocytes were prepared from the brains of 1- or 2-day-old Sprague-Dawley rats and cultured in Dulbecco modified Eagle medium (DMEM) containing 10% fetal calf serum for 3 weeks. Neurons were prepared from 16- to 18-day-old rat embryos according to Richter-Landsberg (1988) and cultured in basal medium Eagle containing 0.5% fetal calf serum (to avoid astrocytic proliferation). For cocultures, astrocytes were first cultivated in DMEM until reaching confluence; thereafter, neuronal cells were added and cultivated as described for neurons. Approximately 10⁷ to 10⁸ cells were used for each NMR experiment. In cocultures, the ratio of astrocytes to neurons was found to be 4:6. All cultures were kept in an incubator with a humidified atmosphere of 5% CO₂ at 37°C. Astroglial and neuronal cultures were characterized by GFAP immunoreactivity and antineurofilament antibodies, respectively. The purity of the individual cells was found to be >95% for primary astrocytes and >90% for primary neurons.

Incubation and preparation of cell extracts

Confluent cell cultures were incubated every 24 h with fresh media containing 100-μmol/L MnCl₂ on 5 consecutive days. During the last 12 h, control cultures and manganese-treated cells were incubated for 12 h with DMEM containing 5-mmol/L [1–¹³C]glucose (Cambridge Isotope Laboratories, Andover, MA, U.S.A.). After removal of the incubation media (lyophilized for subsequent NMR measurements), the cells were extracted with perchloric acid (PCA) as described previously (Zwingmann et al., 2000). Because of the paramagnetic effect of manganese on T₁ relaxation times, di- and trivalent cations were complexed by incubation of the supernatant with iminodiacetic acid (Chelex 100; Sigma, St. Louis, MO, U.S.A.) for 20 minutes. After removing of the chelating resin by filtration, the supernatants were lyophilized.

Biochemical analysis

The enzymatic assays (determination of lactate concentrations) were performed according to the supplier's instructions (Roche Molecular Biochemicals, Mannheim, Germany) with measurement of the UV/VIS absorption of NADH at 340 nm. The protein content was quantified according to the Biuret reaction (Goa, 1953) using bovine serum albumin as a standard and VIS absorption at 540 nm. All biochemical assays were adapted for use in 96-well microtiter ELISA plates. Samples were measured in four concentrations in the plate, each with four replicates.

Nuclear magnetic resonance spectroscopy

The lyophilized samples were redissolved in 0.5 to 0.6 mL deuterium oxide (D₂O; Merck, Darmstadt, Germany) and centrifuged. Before the NMR analysis, the samples were neutralized with DCl and NaOH to allow unique chemical shift assignments and to prevent glutamine-carbamate formation in the media.

¹H-, ¹³C- and ³¹P-NMR spectra were recorded on Bruker spectrometers DRX600 or AMX360/AM 360 (operating for ¹H-NMR at frequencies of 600 MHz or 360 MHz, for ¹³C-NMR at 150.9 MHz or 90.5 MHz, and for ³¹P-NMR at 145.7 MHz). ¹H-NMR spectra were recorded with a 5-mm H,C,N inverse triple-resonance probe using 200 to 400 accumulations, a flip angle of 40°, a repetition time of 15 seconds, a spectral width of 7,183 Hz (DRX 600) or 3,623 Hz (AM/AMX 360), a data size of 16 K, and zero filling to 32 K. Chemical shifts were referenced to lactate (1.33 ppm). ¹³C-NMR spectra were recorded with a 5-mm ¹H/¹³C dual probe using 15,000 to 20,000 accumulations, a repetition time of 2.5 seconds, a flip angle of 27°, composite pulse decoupling with WALTZ-16, a spectral width of 47,619 Hz (DRX 600) or 20,833 Hz (AM/AMX 360), a data size of 32 K (16 K), and zero filling to 64 K (32 K). Chemical shifts were referenced to lactate (21.3 ppm). ³¹P-NMR spectra were recorded with a 5-mm HX probe using 5,000 to 8,000 accumulations, a flip angle of 80°, a repetition time of 3.5 seconds, a spectral width of 5,155 Hz, composite pulse decoupling with WALTZ-16, a data size of 16 K, and zero filling to 32 K. The resonance of phosphocreatine (−2.33 ppm) was used as internal shift reference.

Quantification of metabolite concentrations

The concentrations of lactate were determined enzymatically (see Biochemical analysis). The pool sizes of other metabolites were determined from fully relaxed ¹H-NMR spectra using the known lactate concentration as internal standard or using (trimethylsilyl)propionic-2,2,3,3,d₄-acid (TSP) as external standard. ³¹P-NMR spectra were analyzed with reference to the phosphocreatine concentrations, calculated from ¹H-NMR spectra (nmol/mg protein for cell extracts, μmol/mg protein for media).

Calculation of fractional ¹³C-enrichments and amounts of ¹³C-labeled metabolites

The ¹³C-enrichments in C-3 of lactate were determined by the heteronuclear spin-coupling pattern in ¹H-NMR spectra as follows:

^{13} C ‐ enrichment = \frac{area (^{1} H ‐^{13} C) \times 100}{area (^{1} H ‐^{12} C) + area (^{1} H ‐^{13} C)}

(1)

where the sum (area [¹H − ¹²C] + area [¹H − ¹³C]) is equivalent to the pool size of lactate. The values were corrected for 1.1% natural abundance ¹³C. ¹³C-enrichments in individual carbons of amino acids are derived from ¹³C-NMR spectra using the known ¹³C-enrichment in lactate as described previously (Zwingmann et al., 2000):

E_{Met} (%) = \frac{A_{Met} - A_{n.a} (Met)}{A_{n.a.} (Met)} \times 1.1

(2)

where A_Met represents ¹³C carbon peak area of the metabolite, A_n.a. its natural abundance signal intensity, and 1.1 is the percentage factor of the ¹³C-isotope. The ¹³C signal intensities were corrected for nuclear Overhauser enhancement effects. The absolute amount of ¹³C in specified carbon positions (nmol/mg protein) is the product of the pool size times the fractional ¹³C-enrichment. Due to overlapping signals, glutathione and citrate cannot be integrated reliably by ¹H-NMR, and changes in ¹³C-label incorporation were deduced from relative signal intensities in ¹³C-NMR spectra (normalized by the corresponding signal areas obtained after incubation of cells with unlabeled glucose).

¹³C-isotopomer analysis

Theoretical labeling of metabolites from [1–¹³C]glucose. After incubation of cells with [1–¹³C]glucose, its metabolites are labeled in different carbon positions, depending on the relative contribution of the enzymatic pathways (Fig. 1; Zwingmann et al., 2000). Briefly, via the glycolytic pathway, [1–¹³C]glucose is transformed to [3–¹³C]pyruvate, [3–¹³C]alanine, and [3–¹³C]lactate. [3–¹³C]Pyruvate may enter the TCA cycle via the anaplerotic pathway (PC or malic enzyme) or the oxidative pathway (pyruvate dehydrogenase, PDH). Via PC, glutamate and glutamine are labeled at C-2 (and at C-3 after equilibration of oxaloacetate with fumarate). In the oxidative pathway, pyruvate enters the cycle as [2–¹³C]acetyl-CoA, and glutamate and glutamine are monolabeled at C-4. During the second TCA cycle turn, glutamate and glutamine are monolabeled after condensation of ¹³C-labeled oxaloacetate with unlabeled acetyl-CoA, but doubly labeled ([3,4–¹³C]- and [2,4–¹³C]) after condensation with [2–¹³C]acetyl-CoA. The order of labeled carbons is reversed in GABA (i.e., the labels in [2(4)–¹³C]glutamate end up in [4(2)–¹³C]GABA). In ¹³C-NMR spectra, monolabeled and doubly labeled isotopomers are observed as singlets and doublets, respectively.

FIG. 1.

Metabolic fate of ¹³C-label from [1–¹³C]glucose. Label distribution in glycolytic and tricarboxylic acid (TCA) cycle intermediates during metabolism of [1–¹³C]glucose. For label distribution in TCA cycle intermediates, only one TCA cycle turn from pyruvate to malate is considered; for glutamine, glutamate, and GABA, the label distribution after two TCA cycle turns is presented. One ¹³C-label position in lactate and alanine synthesized via lactate dehydrogenase (LDH) and alanine aminotransferase (ALAT), respectively, and in TCA cycle-related metabolites synthesized through pyruvate carboxylase (PC) or pyruvate dehydrogenase (PDH) during the first TCA cycle turn. Gray circles indicate positions of the ¹³C-label due to equilibration of oxaloacetate with fumarate and the ¹³C-label within the second TCA cycle turn via PDH. GABA, γ-aminobutyric acid; AAT, aspartate aminotransferase.

Contribution of pyruvate carboxylase versus pyruvate dehydrogenase. The relative contribution of the anaplerotic pathway was estimated by the ratio of C2/C4 in glutamate and glutamine.

Scrambling of the label in the TCA cycle. As PDH activity leads to [4–¹³C]glutamate and double-labeled [3,4–¹³C]glutamate during the first and the second TCA cycle turn, respectively, the labeling ratios of mono/double C-4 in glutamate reflect the number of turns that [2–¹³C]acetyl-CoA has made in the TCA cycle.

Enrichment in acetyl-CoA. If [4–¹³C]2-oxoglutarate remains in the TCA cycle up to [2–¹³C]oxaloacetate and reacts with unlabeled acetyl-CoA, monolabeled glutamate C-3 will appear. Double-labeled glutamate will result if [2–¹³C]acetyl-CoA and [2–¹³C]oxaloacetate are converted into glutamate. Thus, the ratio of mono- to double-labeled glutamate C-3 reflects the ¹³C-enrichment of the acetyl-CoA pool.

Statistical analysis

The NMR studies were carried out on four to six independent experiments. Data are expressed as mean ± SD values. Data from all experimental groups were analyzed and compared using ANOVA and post hoc Tukey test. Differences were considered significant when P < 0.05.

RESULTS

Cellular energy state

Analysis of ³¹P-NMR spectra of PCA extracts (Fig. 2) depicting high-energy phosphates such as nucleoside diand triphosphates (NDPs and NTPs), in particular ADP, ATP, and PCr, showed that 5-d treatment of astrocytes with 100-μmol/L MnCl₂ produced no change in NTP and PCr concentrations (Table 1). In cultured neurons, on the other hand, NTP/NDP and PCr/creatine ratios were 51% and 28% lower compared with control values, respectively (P < 0.001; creatine was calculated from ¹H-NMR spectra), indicating a decreased energy state. When astrocytes were cocultured with neurons (astrocytes:neurons = 4:6), PCr was lowered to 32% (P < 0.01), but NTP levels were unchanged. Considering the relative concentrations of PCr in astrocytes and neurons, the theoretical intermediate of PCr in the two cell types under control conditions and after manganese treatment (11.45 vs. 10.67 nmol/mg protein) showed a higher decrease of PCr in manganese-treated cocultures compared with neuron cultures (P < 0.001).

TABLE 1.

Effect of manganese on the concentration of high-energy phosphates

	Astrocytes		Cocultures		Neurons

	Control	Manganese	Control	Manganese	Control	Manganese
NTP (nmol/mg protein)	24.38 ± 2.01	28.39 ± 2.15	27.16 ± 1.35	23.72 ± 1.29	16.90 ± 1.42	11.51 ± 1.58*
NDP (nmol/mg protein)	7.47 ± 0.58	8.62 ± 0.38	7.86 ± 0.92	8.50 ± 0.72	4.24 ± 0.34	5.57 ± 0.70†
PCr (nmol/mg protein)	20.46 ± 2.29	23.84 ± 2.70	18.13 ± 2.31	12.39 ± 1.57*	5.42 ± 0.68	1.90 ± 0.21‡
Cr (nmol/mg protein)	18.72 ± 2.46	32.23 ± 4.52‡	15.35 ± 1.18	17.27 ± 1.40	10.91 ± 0.65	13.76 ± 1.03†
NTP/NDP (ratio)	3.27 ± 0.33	3.30 ± 0.35	3.49 ± 0.49	2.93 ± 0.32†	4.02 ± 0.60	2.06 ± 0.04‡
PCr/Cr (ratio)	1.10 ± 0.15	0.75 ± 0.13†	1.18 ± 0.14	0.72 ± 0.11‡	0.50 ± 0.05	0.14 ± 0.02‡
PCr/NTP (ratio)	0.85 ± 0.14	0.85 ± 0.11	0.67 ± 0.09	0.50 ± 0.04†	0.32 ± 0.06	0.17 ± 0.04‡

The concentrations of high-energy phosphates were calculated by integration of the respective signals in ³¹P-NMR spectra (Fig. 2) and 1H-NMR spectra (Cr) after 12-h incubation of astrocytes, neurons, and cocultures of both with media containing 5 mmol/L [1-¹³C]glucose under control conditions or after treatment with manganese (100 μmol/L MnCl₂, 5 d). The metabolites were quantified with reference to the PCr concentration, which was calculated from 1H-NMR spectra. Values are given as nmol/mg protein (or metabolite ratios) and represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and posthoc Tukey test (*P < 0.01, †P < 0.05 compared with controls, ‡P < 0.001).

PCr, phosphocreatine; Cr, creatine.

FIG. 2.

³¹P-NMR spectra of perchloric acid extracts of manganese-treated (100-μmol/L MnCl₂, 5 d) or control astrocytes. NTP/NDP, nucleoside triphosphate/diphosphate; PCr, phosphocreatine; Pi, inorganic phosphate; (G)PE, glycero)phosphoethanolamine; (G)PC, (glycero)-phosphocholine.

Glycolytic metabolic flux of glucose

As a measure of the glycolytic flux, the consumption of [1–¹³C]glucose ([¹³C]Glc) and the de novo synthesis of [3–¹³C]lactate ([¹³C]Lac) were calculated from ¹H- and ¹³C-NMR spectra of media (Fig. 3) and PCA extracts (Tables 2 and 3).

TABLE 2.

Effect of manganese on extracellular [1-¹³C]glucose and [3-¹³C]lactate

	Astrocytes	Cocultures	Neurons
Extracellular [1-¹³C]glucose/uptake
T = 0 (μmol/mg protein)	10.87 ± 1.31	10.96 ± 1.44	8.61 ± 0.93
Control T = 12 h (μmol/mg protein)	6.04 ± 0.54	3.90 ± 0.66	6.54 ± 0.24
Manganese T = 12 h (μmol/mg protein)	2.68 ± 0.62	2.11 ± 0.43	7.24 ± 0.15
Uptake after manganese treatment (% of control)	170.54 ± 15.73*	126.14 ± 11.46†	66.79 ± 5.94*
Extracellular [3–1³C]lactate/release
Control T = 12 h (μmol/mg protein)	1.56 ± 0.11	2.46 ± 0.19	0.52 ± 0.07
Manganese T = 12 h (μmol/mg protein)	3.36 ± 0.42	3.81 ± 0.48	0.34 ± 0.06
Release after manganese treatment (% of control)	214.36 ± 19.81*	154.88 ± 8.20*	65.75 ± 7.92*
[3-¹³C]lactate release (% of glucose uptake)
Control	32.67 ± 3.94	34.94 ± 3.24	25.36 ± 2.86
Manganese	41.11 ± 3.74	42.95 ± 3.26	24.87 ± 2.41
Changes after manganese treatment (% of control)	125.65 ± 12.31†	123.28 ± 8.61	98.98 ± 14.46

Extracellular concentrations of [1-¹³C]glucose and [3-¹³C]lactate, synthesized de novo from [1-¹³C]glucose, were calculated from 1H and ¹³C-NMR spectra of lyophilized media before and after 12-h incubation of cell cultures with media containing 5 mmol/L [1-¹³C]glucose under control conditions or following treatment with manganese (100 μmol/L MnCl₂, 5 d); no unlabeled glucose was detected in 1H-NMR spectra. Values are given as μmol/mg protein or percentage of control after manganese treatment (in media based on the amounts of [1-¹³C]glucose/[3-¹³C]lactate, taken up from/released into the media during 12 h of incubation time) and represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and post hoc Tukey test (*P < 0.001; †P < 0.05 compared with controls; for differences between individual cell types see text).

TABLE 3.

Effect of manganese on intracellular [1-¹³C]glucose and [3-¹³C]lactate

	Astrocytes	Cocultures	Neurons
Intracellular [1-¹³C]glucose
Control T = 12 h (nmol/mg protein)	20.66 ± 1.35	15.27 ± 0.80	27.43 ± 2.29
Manganese T = 12 h (nmol/mg protein)	13.44 ± 1.07	2.52 ± 0.35	77.62 ± 7.37
Effect of manganese treatment (% of control)	65.21 ± 6.54*	16.49 ± 1.99*	283.58 ± 26.07*
Intracellular [3-¹³C]lactate
Control T = 12 h (nmol/mg protein)	5.58 ± 1.11	6.48 ± 1.22	2.96 ± 0.26
Manganese T = 12 h (nmol/mg protein)	9.35 ± 0.63	5.88 ± 0.37	1.27 ± 0.40
Effect of manganese treatment (% of control)	171.25 ± 25.06*	93.24 ± 14.91	43.03 ± 13.54*

Intracellular concentrations of [1-¹³C]glucose and [3-¹³C]lactate, synthesized de novo from [1-¹³C]glucose, were calculated from 1H- and ¹³C-NMR spectra of lyophilized media from PCA extracts after 12-h incubation of cell cultures with media containing 5 mmol/L [1-¹³C]glucose under control conditions or after treatment with manganese (100 μmol/L MnCl₂, 5 d); no unlabeled glucose was detected in 1H-NMR spectra. Values are given as μmol/mg protein or percentage of control after manganese treatment (in media based on the amounts of [1-¹³C]glucose/[3-¹³C]lactate, taken up from/released into the media during 12 h of incubation time) and represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and posthoc Tukey test (*P < 0.001 compared with controls; for differences between individual cell types see text).

FIG. 3.

¹³C-NMR spectra of lyophilized incubation media before and after incubation of astrocytes and neurons with media containing 5-mmol/L [1–¹³C]glucose. The spectra were minimized in size to show the ¹³C-signals of glucose and lactate (the major metabolite labeled with ¹³C).

Uptake of [1–¹³C]glucose and release of [3–¹³C]lactate. Table 2 shows the concentrations of extracellular [1–¹³C]glucose (T = 0 and 12 h) and [3–¹³C]lactate (T = 12 h). Under control conditions, the uptake of [¹³C]Glc and the release of [¹³C]Lac, which were calculated from the difference of their concentrations at T = 0 and 12 h, were highest in cocultures (Table 2; P < 0.001 between control values in astrocytes, neurons, and cocultures). Treatment of astrocytes and cocultures with 100-μmol/L manganese (5 d) increased the release of de novo synthesized [¹³C]Lac to 215% and 155% of control values (P < 0.001), respectively, which coincided with stimulated [¹³C]Glc uptake. In contrast, in neurons, a decrease in the consumption of [¹³C]Glc and the release of [¹³C]Lac to about 65% (P<0.001) of control values was observed 5 d after manganese exposure.

Intracellular [1–¹³C]glucose and [3–¹³C]lactate. Stimulated glycolytic flux of glucose in astrocytes after 5-d treatment with manganese was reflected also by 35% reduced intracellular [¹³C]Glc (P < 0.001; Table 3). [¹³C]Glc was metabolized almost completely when astrocytes were cocultivated with neurons (P < 0.01 compared with pure astrocytes). In contrast to astrocytes, however, intracellular [¹³C]Lac remained unchanged in manganese-treated cocultures. In neurons, a manganese-induced impaired glycolytic flux was associated with decreased intracellular [¹³C]Lac (P < 0.001) in parallel to an accumulation of nonmetabolized [¹³C]Glc to almost 300% of control.

Percentage ¹³C-enrichment in lactate. To determine the selectivity of [¹³C]Glc metabolism compared with other exogenous substrates, the percentage ¹³C-enrichments in lactate was calculated from ¹H-NMR spectra (Fig. 4). Treatment of astrocytes with manganese resulted in no obvious alterations of ¹³C-enrichments in intracellular lactate but increased selectively the release of lactate that had been synthesized de novo from [¹³C]Glc (P < 0.05; Fig. 4B). In neurons, the ¹³C-enrichments in intra- and extracellular lactate decreased to 70% to 75% of control (P < 0.01), whereas in cocultures the ¹³C-enrichment in extracellular lactate decreased to 75% of control (P < 0.01).

FIG. 4.

Percentage ¹³C-enrichment in intra- and extracellular lactate. (A) Segments of ¹H-NMR spectra of perchloric acid extracts obtained from astrocytes after 12-h incubation with 5-mmol/L [1–¹³C]glucose under control conditions or after treatment with manganese (100-μmol/L MnCl₂, 5 d). The methyl resonance of unlabeled lactate ([¹H-¹²C]) can be distinguished from the ¹³C-labeled isotopomer ([¹H-¹³C]) showing an additional doublet splitting arising from ¹H-¹³C-coupling (J = 128 Hz) and allowing the calculation of the percentage ¹³C-enrichment in lactate. (B) Effect of manganese on the ¹³C-enrichment in intra- and extracellular lactate. Values represent means ± SD of four to six individual experiments; ANOVA and post hoc Tukey test). Data between all experimental groups were analyzed using ANOVA and post hoc Tukey test (***P < 0.001; **P < 0.01, *P < 0.05 compared with controls; see text for differences between individual cell types). Lac, lactate; Ala, alanine; Thr, threonine.

Concentrations of TCA cycle–related metabolites and organic osmolytes

The concentrations of amino acids were calculated from the 2.2- to 3.6-ppm region of ¹H-NMR spectra (Table 4). Glutamine declined to 45% in cocultures of astrocytes and neurons compared with nontreated cells (P < 0.001), whereas no changes were observed in pure astrocytes or neurons. The concentration of glutamate increased in manganese-treated astrocytes and cocultures (5 d, 100 μmol/L) to 180% and 170% of control, respectively (P < 0.01; not statistically significant between astrocytes and cocultures). Whereas manganese increased aspartate in astrocytes to 125% (P < 0.05), it remained unchanged in cocultures. Manganese-treated neurons manifested 30% to 35% decreased intracellular levels of glutamate and aspartate (P < 0.05). After 5-d treatment with manganese, a generalized increase of the TCA-cycle intermediate succinate to 140% to 185% of control levels (P < 0.01) was observed in the order cocultures>neurons>astrocytes (P < 0.001 neurons and cocultures vs. astrocytes). In addition, the concentrations of organic osmolytes were calculated (3.4 to 4.05 ppm). Taurine levels were increased in astrocytes to 250% of control (P < 0.001), but decreased in neurons (P < 0.05) and, in particular, in cocultures to 60% of control (P < 0.001; P < 0.01 cocultures vs. neurons). Hypotaurine, the direct precursor for taurine, and myo-inositol, an astrocytic marker molecule, were unchanged in pure astrocytes after 5-d manganese treatment, but decreased to 25% and 55% of control levels (P < 0.001) in cocultures.

TABLE 4.

Effect of manganese on the concentration of tricarboxylic acid cycle–related metabolites and organic osmolytes

	Astrocytes		Cocultures		Neurons

	Control	Manganese	Control	Manganese	Control	Manganese
Glutamine	96.66 ± 8.77	103.58 ± 5.23	125.80 ± 2.36	49.18 ± 5.09*	68.69 ± 4.83	76.90 ± 10.41
Glutamate	21.76 ± 1.64	39.70 ± 5.73*	35.68 ± 3.23	60.95 ± 5.48*	54.18 ± 5.17	35.99 ± 6.87†
GABA	n.d.	n.d.	8.67 ± 0.72	17.65 ± 4.09*	17.06 ± 2.68	15.74 ± 3.76
Aspartate	5.99 ± 0.61	7.42 ± 0.62‡	7.34 ± 0.71	7.61 ± 1.46	15.43 ± 1.13	11.12 ± 1.99‡
Succinate	3.20 ± 0.17	4.47 ± 0.42†	4.48 ± 0.41	8.32 ± 0.72*	3.61 ± 0.39	6.11 ± 0.68*
Taurine	77.67 ± 3.04	195.20 ± 18.11*	93.66 ± 8.55	55.67 ± 7.84*	75.73 ± 6.12	61.81 ± 4.96‡
Hypotaurine	135.57 ± 12.84	143.51 ± 13.27	65.70 ± 8.91	15.74 ± 1.86*	8.72 ± 0.09	9.68 ± 0.11
myo-Inositol	97.11 ± 6.05	88.58 ± 10.53	78.43 ± 8.70	42.41 ± 5.21*	n.d.	n.d.

The concentrations of the metabolites were calculated by integration of the respective signals in 1H-NMR spectra after 12-h incubation of cell cultures with media containing 5 mmol/L [1-¹³C]glucose under control conditions or after treatment with manganese (100 μmol/L MnCl₂, 5 d). The values are given as nmol/mg protein and represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and posthoc Tukey test (*P < 0.001; †P < 0.01, ‡P < 0.05 compared with controls; for differences between individual cell types see text).

GABA, γ = aminobutyric acid.

Cell-specific [1–¹³C]glucose metabolism

Figure 5 shows representative ¹³C-NMR spectra obtained from extracts of cocultures of astrocytes and neurons after 12-h incubation with 5-mmol/L [1–¹³C]glucose. Whereas ¹³C-labeled glutamate was hardly detected in astrocytes under control conditions, 5-d exposure of astrocytes to 100-μmol/L manganese led to increased synthesis of the isotopomer [4–¹³C]glutamate, formed in the PDH pathway, reflected by an increased total amount and ¹³C-enrichment to 400% and 220% of control, respectively (Tables 5 and 6). Consistent with unchanged glutamine concentrations (Table 4), the increased glutamate synthesis was not accompanied by subsequent incorporation of ¹³C-label into glutamine. When astrocytes were cocultured with neurons, manganese exposure resulted not only in stimulated PDH-mediated synthesis of [4–¹³C]glutamate, but in particular in increased flux through PC leading to increased [2–¹³C]glutamate and [3–¹³C]aspartate synthesis (P < 0.001). Despite stimulated glutamate synthesis in cocultures, however, manganese exposure led to decreased [¹³C]glutamine levels to 7% (C-4) and 26% (C-2) of control values (P < 0.001) (Table 5). A similar observation was made from ¹³C-NMR spectra of the incubation media. In control cells, the release of ¹³C-labeled glutamine within 12-h incubation with [1–¹³C]glucose was in the following order: cocultures (357.84 ± 48.73 nmol/mg protein), astrocytes (165.70 ± 15.92 nmol/mg protein), and neurons (which did not release ¹³C-labeled glutamine) (P > 0.001 between the individual cell types). After exposure to manganese, however, ¹³C-labeled glutamine was not detected in media of cocultures (data not shown). In contrast to astrocytes and cocultures, a decrease of mitochondrial [¹³C]glucose metabolism in manganese-treated neurons was associated with a 55% decrease in ¹³C-enrichment in [4–¹³C]glutamate, synthesized via PDH (P < 0.001), and reduced ¹³C-label incorporation from [¹³C]Glc into aspartate to 40% of control values (P < 0.001). The ¹³C-isotopomer pattern of GABA is interesting in that for neurons, the labeling pattern of GABA resembled that for ¹³C-labeled glutamate (i.e., manganese exposure selectively decreased de novo synthesis of [2–¹³C]GABA compared with [4–¹³C]GABA) (P < 0.01), which is metabolically derived from [4–¹³C]glutamate (Fig. 1). In manganese-treated cocultures, the ¹³C-enrichment and the absolute amount of ¹³C in [2–¹³C]GABA increased dramatically to almost 200% and 400% of control values, respectively.

TABLE 5.

Effect of manganese on ¹³C-label incorporation from [1-¹³C]glucose into tricarboxylic cycle–related amino acids: total amounts of ¹³C

	Astrocytes		Cocultures		Neurons

	Amount [¹³C] (nmol/mg protein)		Amount [¹³C] (nmol/mg protein)		Amount [¹³C] (nmol/mg protein)
	Control	Mn (% of control)	Control	Mn (% of control)	Control	Mn (% of control)
Gln C2	13.48 ± 0.83	116.54 ± 9.17	22.13 ± 1.45	26.21 ± 2.35*	0.44 ± 0.09	100.14 ± 8.96
Gln C3	7.91 ± 0.55	98.27 ± 9.28	23.46 ± 1.78	30.49 ± 6.07*	0.55 ± 0.12	107.78 ± 12.82
Gln C4	10.73 ± 0.69	122.26 ± 7.18	24.89 ± 2.27	7.21 ± 1.16*	0.57 ± 0.14	101.44 ± 8.84
Glu C2	2.93 ± 0.25	225.55 ± 21.11*	5.71 ± 0.47	316.90 ± 36.96*	7.64 ± 0.22	61.12 ± 6.08*
Glu C3	1.71 ± 0.18	371.86 ± 35.29*	7.84 ± 0.44	138.52 ± 12.88†	9.81 ± 0.29	57.01 ± 2.46*
Glu C4	2.53 ± 0.24	401.06 ± 22.07*	7.22 ± 0.87	252.99 ± 38.02*	13.15 ± 0.64	44.62 ± 4.23*
GABA C4	n.d.	n.d.	0.46 ± 0.11	134.52 ± 12.83†	2.82 ± 0.18	85.24 ± 5.99
GABA C3	n.d.	n.d.	3.95 ± 0.45	175.48 ± 21.07	2.71 ± 0.20	64.18 ± 5.32
GABA C2	n.d.	n.d.	1.86 ± 0.18	392.98 ± 31.56*	4.14 ± 0.33	64.22 ± 9.58†
Asp C3	0.15 ± 0.02	164.47 ± 22.62*	0.64 ± 0.05	272.79 ± 19.70*	2.16 ± 0.25	29.99 ± 0.80*
Asp C2	0.09 ± 0.01	176.50 ± 13.86*	n.d.	n.d.	1.91 ± 0.09	30.62 ± 3.37*

The total amounts of ¹³C (nmol/mg protein) in specific carbon positions of individual tricarboxylic cycle–related amino acids labeled from [1-¹³C]glucose were calculated from ¹³C-NMR spectra obtained from cell extracts after 12-h incubation with media containing 5 mmol/L [1-¹³C]glucose under control conditions of following treatment with manganese (Mn; 100 μmol/L MnCl₂, 5 d). Values are given as nmol/mg protein for the controls and as percentage of control after Mn treatment and represent means ± SD of four to six individual treatments. Data between all experimental groups were analyzed using ANOVA and posthoc Tukey test (*P < 0.001, †P < 0.01 compared with controls; for differences between individual cell types see text).

Gln, glutamine; Glu, glutamate; GABA, γ-aminobutyric acid; Asp, aspartate.

TABLE 6.

Effect of manganese on ¹³C-label incorporation from [1-¹³C]glucose into tricarboxylic cycle–related amino acids: percentage of ¹³C-enrichments

	Astrocytes		Cocultures		Neurons

	Percentage ¹³C-enrichment		Percentage ¹³C-enrichment		Percentage ¹³C-enrichment
	Control	Mn (% of control)	Control	Mn (% of control)	Control	Mn (% of control)
Gln C2	13.97 ± 0.79	108.48 ± 6.97	17.59 ± 1.23	67.53 ± 8.98*	0.69 ± 0.12	84.15 ± 9.92
Gln C3	8.19 ± 0.32	91.55 ± 8.71	18.64 ± 1.13	77.45 ± 7.29†	0.86 ± 0.12	90.71 ± 10.14
Gln C4	12.13 ± 0.84	114.12 ± 11.27	19.76 ± 1.42	18.60 ± 3.69*	0.88 ± 0.15	84.77 ± 5.25
Glu C2	13.49 ± 0.88	124.28 ± 12.43	16.05 ± 1.42	184.71 ± 6.77*	14.21 ± 1.79	82.18 ± 6.30‡
Glu C3	7.89 ± 0.89	207.21 ± 44.32*	16.08 ± 2.51	81.77 ± 7.44	18.22 ± 1.93	77.00 ± 7.95†
Glu C4	11.71 ± 1.89	222.81 ± 38.34*	20.22 ± 1.60	147.94 ± 6.29*	24.38 ± 1.82	60.57 ± 11.29*
GABA C4	n.d.	n.d.	5.30 ± 0.84	68.59 ± 17.89‡	11.14 ± 0.67	94.33 ± 5.50
GABA C3	n.d.	n.d.	18.42 ±2.18	88.94 ± 21.41	16.07 ± 2.16	70.44 ± 8.49
GABA C2	n.d.	n.d.	21.43 ± 0.28	196.54 ± 22.15*	24.52 ± 2.78	70.68 ± 8.11†
Asp C3	2.51 ± 0.33	133.62 ± 14.66	8.83 ± 1.21	211.88 ± 49.05*	13.97 ± 0.90	42.18 ± 5.16*
Asp C2	1.41 ± 0.15	142.71 ± 11.19‡	3.43 ± 0.89	97.29 ± 27.52	12.46 ± 1.44	42.86 ± 5.12*

The percentage ¹³C-enrichments in specific carbon positions of individual tricarboxylic cycle–related amino acids labeled from [1-¹³C]glucose were calculated from ¹³C-NMR spectra obtained from cell extracts after 12-h incubation with media containing 5 mmol/L [1-¹³C]glucose under control conditions or after treatment with manganese (Mn; 100 μmol/L MnCl₂, 5 d). Values are given as percentage ¹³C-enrichment for the controls and as percentage of control after Mn treatment and represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and posthoc Tukey test (*P < 0.001; †P < 0.01, ‡P < 0.05 compared with controls; for differences between individual cell types see text).

Gln, glutamine; Glu, glutamate; GABA, γ-aminobutyric acid; Asp, aspartate.

FIG. 5.

¹³C-NMR spectra of perchloric acid extracts obtained from cocultures of astrocytes and neurons after 12-h incubation with 5-mmol/L [1–¹³C]glucose under control conditions and after 5-d treatment with 100-μmol/L MnCl₂. The bottom spectrum was obtained from cocultures incubated with a medium containing [¹²C]glucose and shows resonances resulting from natural abundance ¹³C (n.a.) in the metabolites. Asp, aspartate; Gln, glutamine; Glu, glutamate; GLT, glutathione; GABA, γ-aminobutyric acid.

Other ¹³C-enriched metabolites. Whereas treatment of astrocytes with 100-μmol/L manganese resulted in a more than fourfold elevation of [¹³C]glutathione (P < 0.001), its signal intensity was reduced to 25% and 70% of control levels in cocultures and neurons, respectively (P < 0.01; P < 0.01 between cocultures and neurons; Fig. 6). The conversion of [1–¹³C]Glc into the TCA cycle intermediate citrate increased in astrocytes and neurons to 140% and 180%, respectively (P < 0.001), and more than 12-fold in cocultures (P < 0.001 between the individual cell types).

FIG. 6.

Effect of manganese on synthesis of glutathione and citrate. The ¹³C-signal intensities in glutathione and citrate, labeled from [1–¹³C]glucose, were calculated by integration of the respective signals in ¹³C-NMR spectra after 12-h incubation of astrocytes, neurons, and cocultures of both with media containing 5-mmol/L [1–¹³C]glucose under control conditions or after treatment with manganese (100-μmol/L MnCl₂, 5 d). Values are given as percentage of control and represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and post hoc Tukey test (***P < 0.001; **P < 0.01 compared with controls; see text for differences between individual cell types).

Oxidative (pyruvate dehydrogenase) versus anaplerotic (pyruvate carboxylase) glucose metabolism. The relative contribution of the oxidative (PDH) versus the anaplerotic (PC) pathway was calculated by the ratio of ¹³C-4/¹³C-2 in glutamate and glutamine (Fig. 7A). Under control conditions, the relative flux of label through PC was highest in astrocytes, and only marginal in neurons. The greater increase of ¹³C-label incorporation into C-4 compared with C-2 of glutamate in astrocytes after manganese treatment, however, indicates a relative almost twofold activation of the oxidative pathway (PDH) (P < 0.001). In contrast, manganese treatment caused a relative decrease of oxidative [¹³C]Glc metabolism in cocultures and neurons, resulting in a decreased ratio of [4–¹³C]glutamate to [2–¹³C]glutamate, to 80% and 75% of control levels, respectively (P < 0.05). In addition, manganese caused a decrease in the ratio of [4–¹³C]glutamine (PDH) to [2–¹³C]glutamine (PC), to 27% of control values in cocultures.

FIG. 7.

Relative contribution of pyruvate dehydrogenase (PDH) and pyruvate carboxylase (PC) ¹³C-enrichment in acetyl-CoA and tricarboxylic acid (TCA) cycling ratio. (A) The relative contribution of the oxidative (PDH) compared with the anaplerotic pathway (PC) has been deduced from the ratio of ¹³C-4/¹³C-2 in glutamate and glutamine. (B) The ratio of double/monolabeled [3–¹³C]glutamate reflects the ¹³C-enrichment in acetyl-CoA entering the TCA cycle. The ratio of double/monolabeled [4–¹³C]glutamate reflects the number of TCA cycle turns that the label has passed through before [4–¹³C]glutamate is synthesized. Values represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and post hoc Tukey test (***P < 0.001; ** P < 0.01, *P < 0.05 compared with controls; see text for differences between individual cell types).

¹³C-enrichment in acetyl-CoA and TCA cycling ratio. Changes in ¹³C-label incorporation into C-3 of glutamate, glutamine, or GABA in all cell types may be due to an altered percentage ¹³C-enrichment in [2–¹³C]acetyl-CoA entering the TCA. The increased ratio of double- to mono-labeled [3–¹³C]glutamate in manganese-treated astrocytes is consistent with a twofoldincreased ¹³C-enrichment in acetyl-CoA (P < 0.001, Figs. 7B and 8). The reductions of double-labeled glutamate C-3 in neurons and cocultures, in contrast, indicate decreased formation of [2–¹³C]acetyl-CoA by 40% and 60%, respectively (P < 0.001). Concomitant with an increased ¹³C-enrichment in acetyl-CoA in manganese-treated astrocytes, a substantial amount of the ¹³C-label passed through the TCA cycle for more than one turn before glutamate was formed, as revealed by a twofold increase in the ratio of double/monolabeled glutamate C-4 (P < 0.001; Fig. 7B). Consistent with impaired mitochondrial oxidative metabolism in neurons, manganese treatment resulted in reduced cycling ratio of [2–¹³C]acetyl-CoA within the TCA cycle to 70% of control (P < 0.01), which was further decreased in cocultures (P < 0.01).

FIG. 8.

¹³C-NMR spectra of perchloric acid extracts obtained from astrocyte cultures after 12-h incubation with 5-mmol/L [1–¹³C]glucose under control conditions and after 5-d or 12-h treatment with 100-μmol/L MnCl₂. For peak assignments see Fig. 6 The multiple pattern of [3–¹³C]glutamate and [3–¹³C]glutamine is shown in the expanded plots of the respective chemical shift region (the experimental spectra are smoothed by an exponential multiplication corresponding to a line broadening of 0.5 Hz).

Short-term effects of manganese on neuronal and glial energy metabolism. Since manganese is a cofactor or activator for several enzymes, we investigated whether manganese directly regulates metabolic pathways in astrocytes and neurons during 12-h incubation with [1–¹³C]glucose. Figure 8 shows ¹³C-NMR spectra of PCA extracts obtained from astrocytes, treated for 12 h or 5 d with 100-μmol/L manganese, after 12-h incubation with [1–¹³C]glucose. After 12-h treatment, both astrocytes and neurons showed decreased ATP concentrations to approximately 70% of control values without changes in PCr concentrations (P < 0.01). No changes in [1–¹³C]glucose uptake and metabolism (i.e., glycolytic activity and lactate synthesis), TCA cycle activity, or glutathione synthesis were observed in neurons after 12-h exposure to manganese (data not shown). After 5-d treatment, PCr decreased in neurons, and this decrease was associated with strongly impaired glucose utilization (see results under Glycolytic Metabolic Flux of Glucose). Figure 9 compares the changes induced after 12-h and 5-d manganese exposure in astrocytes (percent of control). ATP levels recovered in astrocytes, but this was accompanied by a less stimulating effect of manganese on glucose utilization, TCA cycle flux, and [4–¹³C]glutamate synthesis after 5-d compared with 12-h exposure (P < 0.05). Whereas glutamine formation was unchanged in astrocytes treated for 5 d with manganese, de novo synthesis of [4–¹³C]glutamine (through PDH), but not of [2–¹³C]glutamine (through PC), selectively decreased to almost 40% of control values after 12-h treatment (P < 0.001; P < 0.001 compared with 5-d treatment; Fig. 9B), despite a marked increase in [4–¹³C]glutamate; [4–¹³C]glutamine release was completely inhibited after 12 h. The short-term effect of manganese was also associated with a 170% increased and a 45% decreased flux of [1–¹³C]glucose through PDH versus PC (ratio of ¹³C-4/¹³C-2) contributing to de novo synthesis of glutamate and glutamine, respectively. In addition, after 12-h treatment with manganese, a dramatic (10-fold compared with 5-d treatment) increase of ¹³C-labeled aspartate in astrocytes is consistent with an activation of flux through PC (data not shown in Fig. 9). The relative flux of ¹³C-label through the PC pathway contributing to glutamine synthesis from glutamate as well as glutathione and citrate synthesis decreased after 5-d compared with 12-h manganese treatment (P < 0.001; P < 0.01 compared with 5-d treatment; Figs. 9B and 9C). In addition, a 75% higher ratio of double- to mono-labeled carbon C-4 in the [4–¹³C]glutamate residue of glutathione compared with glutamate (data not shown) show that glutathione was synthesized subsequently to stimulated PDH activity (i.e., after the ¹³C-label exceeded several TCA cycle turns).

FIG. 9.

Short-term effects of manganese in neurons and astrocytes. The changes in ¹³C-labeled metabolites related to energy metabolism (A and B) and of metabolites related to mitochondrial integrity and antioxidant defense (C) after 12-h and 5-d treatment of astrocytes with manganese, respectively, were calculated from ³¹P and ¹³C-NMR spectra. Changes in ATP, PCr, [1–¹³C]glucose, taurine, and succinate are based on their absolute concentrations; changes in [4–¹³C]glutamate and [2–¹³C]acetyl-CoA as well as in [4–¹³C]glutamine and in the ratio of ¹³C-4/¹³C-2 in glutamate and glutamine are based on their percentage ¹³C-enrichments, and changes in glutathione and citrate are based on their ¹³C-signal intensity in ¹³C-NMR spectra. Values are given as percentage of control and represent means ± SD of four to six individual experiments. Data between all experimental groups were analyzed using ANOVA and post hoc Tukey test (***P < 0.001; **P < 0.01, compared with controls; see text for differences between 12-h and 5-d treatment). Asp, aspartate; Gln, glutamine; Glu, glutamate; GSH, glutathione; EDTA, ethylenediamine tetraacetic acid.

DISCUSSION

Manganese neurotoxicity may involve mitochondrial dysfunction leading to decreased energy production. Previous studies, however, have provided little direct evidence for an altered brain energy metabolism and have not distinguished between the potentially different susceptibility of neural cells to manganese-induced metabolic derangements.

By means of multinuclear NMR investigations, allowing simultaneous monitoring of cellular energy state, [1–¹³C]glucose metabolism and intercellular metabolic trafficking, we have shown for the first time that manganese selectively impairs neuronal glucose oxidation leading to reduced antioxidant capacity and disruption of mitochondrial energy production. Astrocytes, in contrast, increased their oxidative and anaerobic glucose metabolism to protect against energy failure and oxidative stress but failed to evade other derangements, such as the distortion of glutamine synthesis, leading to a disturbed intercellular metabolic homeostasis and increased risk of manganese-induced neuronal energy failure.

A selective energy failure in cultured neurons was reflected by a rapid decrease of ATP and a dramatic depletion of PCr. Astrocytes were able to restore initially reduced ATP levels, consistent with previous results (Hazell et al., 1999). Interestingly, when neurons were cocultivated with astrocytes, the decrease of PCr was even higher than the theoretical intermediate between the individual cell types. Whereas neurons were not able to increase their glycolytic activity, astrocytes showed stimulated anaerobic metabolism of external [1–¹³C]glucose (with NAD⁺ regeneration by subsequent lactate synthesis) that enabled, at least in part, cytosolic ATP synthesis. Unlike [3–¹³C]lactate synthesis, changes in the total amount of lactate were not significant in manganese-treated astrocytes or neurons, as observed earlier (Hazell et al., 1999). The analysis of the percentage ¹³C-enrichments in [3–¹³C]lactate (i.e., the selective changes in de novo synthesized [3–¹³C]lactate derived from the pyruvate pool of glycolytic origin compared with exogenous substrates other than [1–¹³C]glucose), however, directly reflects cell-specific changes in glycolytic activity. Our results are consistent with increased lactate in brain tissue in parallel with an ATP depletion after striatal injection of MnCl₂ (Brouillet et al., 1993), but give supplementary evidence for a cell-specific occurrence of both features. Nevertheless, brain energy production depends almost exclusively on mitochondrial glucose oxidation. Pyruvate dehydrogenase, transforming [1–¹³C]glucose into [2–¹³C]acetyl-CoA and hence into the TCA-cycle intermediate [4–¹³C]α-ketoglutarate, is active in both neurons and astrocytes and plays the dominant role in regulation of oxidation and mitochondrial energy production in the brain. Only in astrocytes can [1–¹³C]glucose be transformed into [2–¹³C]α-ketoglutarate via the astrocyte-specific PC, the main anaplerotic enzyme in the brain (Shank et al., 1985). ¹³C-NMR analysis showed a dramatic inhibition of neuronal oxidative glucose metabolism via PDH, decreased ¹³C-enrichment in acetyl-CoA entering the TCA cycle and reduced TCA cycle turnover, leading to deterioration of the neuronal energy state and impaired de novo synthesis of neurotransmitter amino acids (i.e., glutamate, aspartate, and GABA). In contrast, in astrocytes a strongly stimulated uptake and oxidation of [1–¹³C]glucose through PDH, which coincided with increased ¹³C-enrichment in acetyl-CoA and amplified ¹³C-cycling ratio within the TCA cycle, seems to be the most prominent mechanism to avoid manganese-induced cellular energy failure. Our results suggest that limited availability of external glucose for mitochondrial metabolism leads to an inability of neurons to antagonize an energy depletion by oxidative metabolism, and that limited flux through PDH further aggravates the energy failure.

Since astrocytes sequester manganese (Aschner et al., 1992; Tholey et al., 1990; Wedler et al., 1989), it has been suggested that these cells are likely initial targets of manganese toxicity (Henriksson and Tjalve, 2000; Spranger et al., 1998). The prevention of energy failure by stimulated glucose oxidation and glycolytic activity in manganese-treated astrocytes does not necessarily contradict this hypothesis. The accumulation of succinate and citrate in all cell types suggests an inhibition of intermediary metabolism at the level of succinate dehydrogenase (Hietanen et al., 1981) and of aconitase (Zheng et al., 1989). Similar to a direct effect of manganese on oxidative phosphorylation (Gavin et al., 1992) and complex I activity (Galvani et al., 1995), an inhibition of these TCA cycle enzymes ultimately leads to decreased availability of reduced equivalents for oxidative phosphorylation and hence to an energy deficit. Interestingly, accumulation of succinate and citrate as well as the decrease of [¹³C]acetyl-CoA entering the TCA cycle were highest when neurons were cocultivated with astrocytes. Therefore, manganese-induced changes in astrocytes may either modify a primary neuronal injury or induce secondarily neuronal dysfunction.

Besides neuronal glucose oxidation, neurons rely on the supply of substrates synthesized and released by astrocytes (Hertz et al., 1999). Several in vitro studies (Pellerin et al., 1998; Waagepetersen et al., 1998; Zwingmann et al., 2000) have proposed an astrocyte–neuron “lactate shuttle” for the maintenance of neuronal energy and neurotransmitter metabolism. The different effects of manganese on the percentage ¹³C-enrichments in intraand extracellular lactate are consistent with the presence of metabolically compartmentalized pools of cytosolic pyruvate in neurons and astrocytes (Bouzier et al., 1998; Cruz et al., 2001; Zwingmann et al., 2001), which may be modified under pathological conditions. Compared with pure astrocytes, however, manganese-induced stimulation of glycolytic flux was less after cocultivation with neurons, and in view of the decreased ¹³C-enrichment in extracellular lactate, astrocytic ability to regulate glycolytic synthesis of releasable lactate seems to be uncoupled from substrate requirements in manganese-treated neurons when cocultivated with astrocytes. Nevertheless, the normal rate of glycolysis in astrocytes does not exceed approximately 15 nmol glucose·min·⁻¹mg⁻¹ protein (Marrif and Juurlink 1999), a value almost reached in our manganese-treated cocultures, and neuronal oxidative metabolism cannot rise more rapidly than dictated by the subtype LDH5 isoenzyme prevailing in astrocytes (Gjedde and Marrett, 2001). Although the support of neuronal function with lactate as a sole energy substrate has been reported in several encephalopathies (Nemoto et al., 1974; Schurr et al., 1997), astrocyte-derived lactate may be insufficient to restore the energy deficit in neurons suffering from impaired oxidative metabolism of glucose.

Several positron emission tomography studies have shown that neurons regulate their oxidative metabolism using their own rather than astrocytic-generated pyruvate (Gjedde and Marrett, 2001). Since neurons lack anaplerotic activity, however, they cannot maintain de novo synthesis of neurotransmitter amino acids and permanently require new carbons from astrocytes. In contrast to lactate, glial PC activity and glutamine synthesis are significant in brain interplay between glutamatergic action and glucose metabolism (Gruetter et al., 2001). Manganese may play an important role in glial metabolism (Tholey et al., 1987) owing to the localization of PC and GS (Norenberg and Martinez-Hernandez, 1979) in these cells, with both metalloenzymes requiring manganese.

In astrocytes, manganese treatment resulted in abnormal stimulation of PDH versus PC activity and intracellular accumulation of de novo synthesized [4–¹³C]glutamate. From the manganese-induced accumulation of glutamate in these cells, one might expect an increased flux through GS, but glutamine synthesis remained unchanged and even decreased after brief exposure despite a more than fivefold elevation of [4–¹³C]glutamate. Despite cocultivation of astrocytes with neurons stimulating both synthesis and release of glutamine, which is consistent with increased GS messenger RNA and activity (Mearow et al., 1990; Wu et al., 1988), de novo synthesized glutamine was hardly detectable after manganese treatment. These findings are in disagreement with reports of increased GS expression in manganese-exposed rat brain (Zheng et al. 1999) and cultured astrocytes (Chen and Liao, 2002). The complexity of GS regulation provides several explanations for this paradoxical effect. For example, changes in the synthesis rate may not necessarily coincide with changes in protein expression. In addition, depending on the concentration, manganese may both activate and inhibit GS, and maximal GS activity occurs at a defined cation/ATP ratio (Tholey et al. 1987). Although the distortion of glutamine synthesis may conserve astrocytic energy consumption, this may have profound consequences for the normal metabolic balance between astrocytes and neurons. First, as glutamine synthesized in astrocytes serves as an important neuronal energy substrate and precursor of glutamate (Hertz et al., 1999), a limited supply of glutamine to neurons would additionally impair neuronal oxidative metabolism and deplete the neurotransmitter pool of glutamate as well. Second, the inability to convert glutamate to glutamine may represent one mechanism contributing to impaired reuptake of glutamate into astrocytes after manganese exposure (Hazell and Norenberg, 1997), making neuronal cells more prone to excitotoxicity. An impaired reuptake of neuronal glutamate is consistent with the relative greater decrease of glutamine metabolically derived from [4–¹³C]glutamate (PDH) compared with that derived from [2–¹³C]glutamate formed exclusively via astrocytic PC. Furthermore, the reduction in the total amount of [4–¹³C]glutamine was higher than the increase of [4–¹³C]glutamate (23.1 vs. 11.05 nmol/mg protein), further indicating impaired trafficking via the glutamine-glutamate cycle. Similarly, the accumulation of [2–¹³C]GABA in cocultures, which is metabolically derived from neuronal [4–¹³C]glutamate rather than from astrocytic glutamine, may be the result of impaired reuptake via the GABA shunt, an important pathway for the reentry of GABA into the TCA cycle (Hassel et al., 1998).

Several studies using various paradigms of oxidative stress have shown that reactive oxygen species can inactivate GS (Brand et al., 1999; Minana et al., 1997; Oliver et al., 1990; Schor, 1988), which would be consistent with manganese-induced oxidative stress involving formation of reactive oxygen species and free radicals (Chen and Liao, 2002; Spranger et al., 1998). The reduced label incorporation from [1–¹³C]glucose into the [4–¹³C]glutamate residue of neuronal glutathione, a major intracellular antioxidant (Dringen, 2000), after prolonged manganese exposure indicates a likely impairment of the neuronal antioxidant system in parallel with energy deficits. Astrocytes, in contrast, showed a rapid, more than 10-fold increase in the synthesis of glutathione from [1–¹³C]glucose, which likely reflects a protective response against formation of reactive oxygen species. The higher ratio of double- to mono-labeled ¹³C-4 in the glutamate residue of glutathione compared with glutamate revealed glutathione synthesis occurring consecutively to manganese-induced stimulation of PDH (i.e., after the carbon skeleton of glutamate has passed several TCA cycle turns). However, the initial dramatic stimulation of PDH activity and hence glutathione synthesis slowed down in astrocytes, and the even greater decrease of [2–¹³C]acetyl-CoA and [¹³C]glutathione in cocultures compared to pure neurons, together with the deterioration of the cellular energy state, suggests a failure of astrocytes to maintain energy metabolism and antioxidant defense mechanisms after sustained manganese exposure. Interestingly, manganese exerted a similar effect on the concentration of taurine in neurons and cocultures, which, like glutathione, is a free radical scavenger and can protect against excitotoxicity (Saransaari and Oja, 2000) and plays an important role as an organic osmolyte in astrocytes (Olson and Li, 2000). Interestingly, hypotaurine, the precursor for taurine, and myo-inositol, which both participate in cell volume regulation in astrocytes (Brand et al., 1993; Isaacks et al., 1999), were strongly decreased in these cells when cocultivated with neurons and may reflect effects of manganese on the control of ion and volume regulation in astrocytes in the presence of neurons.

In this study, a selective impairment of neuronal oxidative metabolism of glucose and astrocytic substrates was shown by ¹³C-isotopomer analysis of [1–¹³C]glucose metabolism, which labels TCA cycle metabolites in different carbon positions depending on the contribution of cell-specific and intercellular pathways. We have presented data indicating that manganese produces differential changes in glucose utilization in astrocytes and neurons that may play an important role in determining the relative vulnerability of neuronal cells in manganese neurotoxicity (Fig. 10). Inhibition of neuronal PDH impedes the ability of neurons to compensate for mitochondrial dysfunction by raising oxidative glucose metabolism and further aggravates neuronal energy failure. Although astrocytes maintain their normal energy state and antioxidant mechanisms by upregulation of glucose metabolism through increased glycolysis, PDH and Krebs cycle activity, astrocytic dysfunction in the presence of neurons (e.g., reduced glutamine synthesis) may cause a homeostatic dysequilibrium between both cell types, leading to further potentiation of the primary neuronal derangement and, conceivably, the eventual development of an excitotoxic state. These self-protective metabolic responses may also facilitate astrocyte survival and the development of reactive gliosis in the face of concomitant neuronal dysfunction. Our results provide a cell type-specific basis for some possible mechanisms associated with energy failure in manganese neurotoxicity. Future studies are necessary to characterize the sequence of metabolic events. Further NMR investigations using ¹³C-labeled substrates involved in energy and neurotransmitter metabolism as well as the cellular antioxidant defense system, combined with in vitro, ex vivo, and in vivo approaches, may provide additional information concerning the primary basis and significance of energy failure in manganese neurotoxicity.

FIG. 10.

Relation between cell-specific glucose metabolism and neuronal vulnerability in manganese neurotoxicity. Mitochondrial dysfunction in neurons (1; possibly related to oxidative stressmediated changes) is associated with impaired cellular energy metabolism. Such abnormalities may lead to energy failure (2). Inhibition of neuronal pyruvate dehydrogenase (PDH; 3) hinders the ability of neurons to compensate for mitochondrial dysfunction by increasing oxidative glucose metabolism and further aggravates neuronal energy failure. In addition, metabolic changes in astrocytic glucose metabolism, such as the inhibition of glutamine synthesis, may cause a homeostatic dysequilibrium between neurons and astrocytes (4), leading to further potentiation of the primary neuronal dysfunction, and conceivably activation of an excitotoxic state (5).

References

Aschner

Gannon

Kimelberg

(1992) Manganese uptake and efflux in cultured rat astrocytes. J Neurochem 58:730–735

Bachelard

Badar-Goffer

(1993) NMR spectroscopy in neurochemistry. J Neurochem 61:412–429

Booher

Sensenbrenner

(1972) Growth and cultivation of dissociated neurons and glial cells from embryonic chick, rat and human brain in flask cultures. Neurobiology 2:97–105

Bouzier

Goodwin

de Gannes

Valeins

Voisin

Canioni

Merle

(1998) Compartmentation of lactate and glucose metabolism in C6 glioma cells: a ¹³C and ¹H NMR study. J Biol Chem 273:27162–27169

Brand

Richter-Landsberg

Leibfritz

(1993) Multinuclear NMR studies on the energy metabolism of glial and neuronal cells. Dev Neurosci 15:289–298

Brand

Leibfritz

Richter-Landsberg

(1999) Oxidative stress-induced metabolic alterations in rat brain astrocytes studied by multinuclear NMR spectroscopy. J Neurosci Res 58:576–585

Brouillet

Shinobu

McGarvey

Hochberg

Beal

(1993) Manganese injection into the rat striatum produces excitotoxic lesions by impairing energy metabolism. Exp Neurol 120:89–94

Chen

Liao

(2002) Oxidative stress involves in astrocytic alterations induced by manganese. Exp Neurol 175:216–225

Cruz

Cerdan

(1999) Quantitative ¹³C NMR studies of metabolic compartmentation in the adult mammalian brain. NMR Biomed 12:451–462

10.

Cruz

Villalba

Garcia-Espinosa

Ballesteros

Bogonez

Satrustegui

Cerdan

(2001) Intracellular compartmentation of pyruvate in primary cultures of cortical neurons as detected by (13)C NMR spectroscopy with multiple (13)C labels. J Neurosci Res 66:771–781

11.

Dringen

(2000) Metabolism and functions of glutathione in brain. Prog Neurobiol 62:649–671

12.

Galvani

Fumagalli

Santagostino

(1995) Vulnerability of mitochondrial complex I in PC12 cells exposed to manganese. Eur J Pharmacol 293:377–383

13.

Gavin

Gunter

(1992) Mn²⁺ sequestration by mitochondria and inhibition of oxidative phosphorylation. Toxicol Appl Pharmacol 115:1–5

14.

Gjedde

Marrett

(2001) Glycolysis in neurons, not astrocytes, delays oxidative metabolism of human visual cortex during sustained checkerboard stimulation in vivo. J Cereb Blood Flow Metab 21:1384–1392

15.

Goa

(1953) A micro biuret method for protein determination. Scand J Clin Lab Invest 5:218–333

16.

Gruetter

Novotny

Boulware

Mason

Rothman

Shulman

Prichard

Shulman

(1994) Localized ¹³C NMR spectroscopy in the human brain of amino acid labeling from D-[1–¹³C]glucose. J Neurochem 63:1377–1385

17.

Gruetter

Seaquist

Ugurbil

(2001) A mathematical model of compartmentalized neurotransmitter metabolism in the human brain. Am J Physiol Endocrinol Metab 281:E100–E112

18.

Hassel

Johannessen

Sonnewald

Fonnum

(1998) Quantification of the GABA shunt and the importance of the GABA shunt versus the 2-oxoglutarate dehydrogenase pathway in GABAergic neurons. J Neurochem 71151:1511–1518

19.

Hazell

Norenberg

(1997) Manganese decreases glutamate uptake in cultured astrocytes. Neurochem Res 22:1443–1447

20.

Hazell

Desjardins

Butterworth

(1999) Increased expression of glyceraldehyd-3-phosphate dehydrogenase in cultured astrocytes following exposure to manganese. Neurochem Int 35:11–17

21.

Henriksson

Tjalve

(2000) Manganese taken up into the CNS via the olfactory pathway in rats affects astrocytes. Toxicol Sci 55:392–398

22.

Hertz

Dringen

Schousboe

Robinson

(1999) Astrocytes: glutamate producers for neurons. J Neurosci Res 57:417–428

23.

Hietanen

Kilpio

Savolainen

(1981) Neurochemical and biotransformational enzyme responses to manganese exposure in rats. Arch Environ Contam Toxicol 10:339–345

24.

Huang

Chu

Wang

Tsai

Tzeng

Wolters

Calne

(1989) Chronic manganese intoxication. Arch Neurol 46:1104–1146

25.

Isaacks

Bender

Kim

Shi

Norenberg

(1999) Effect of osmolality and anion channel inhibitors on myo-inositol efflux in cultured astrocytes. J Neurosci Res 57:866–871

26.

Marrif

Juurlink

(1999) Astrocytes respond to hypoxia by increasing glycolytic capacity. J Neurosci Res 57:255–260

27.

Mearow

Mill

Freese

(1990) Neuron-glial interactions involved in the regulation of glutamine synthetase. Glia 3:385–392

28.

Mena

Marin

Fuenzalida

Cotzias

(1967) Chronic manganese poisoning: clinical picture and manganese turnover. Neurology 17:128–136

29.

Minana

Kosenko

Marcaida

Hermenegildo

Montoliu

Grisolia

Felipo

(1997) Modulation of glutamine synthesis in cultured astrocytes by nitric oxide. Cell Mol Neurobiol 17:433–445

30.

Nemoto

Hoff

Severinghaus

(1974) Lactate uptake and metabolism by brain during hyperlactatemia and hypoglycemia. Stroke 5:48–53

31.

Norenberg

Martinez-Hernandez

(1979) Fine structural localization of glutamine synthetase in astrocytes of rat brain. Brain Res 161:303–130

32.

Oliver

Starke-Reed

Stadtman

Liu

Carney

Floyd

(1990) Oxidative damage to brain proteins, loss of glutamine synthetase activity, and production of free radicals during ischemia/reperfusion-induced injury to gerbil brain. Proc Natl Acad Sci U S A 87:5144–5147

33.

Olson

(2000) Osmotic sensitivity of taurine release from hippocampal neuronal and glial cells. Adv Exp Med Biol 483:213–218

34.

Pellerin

Pellegri

Bittar

Charnay

Bouras

Martin

Stella

Magistretti

(1998) Evidence supporting the existence of an activity-dependent astrocyte-neuron lactate shuttle. Dev Neurosci 20:291–299

35.

Pentschew

Ebner

Kovatch

(1963) Experimental manganese encephalopathy in monkeys: a preliminary report. J Neuropathol Exp Neurol 22:488–499

36.

Richter-Landsberg

(1988) Nerve growth factor-inducible, large external (NILE) glycoprotein in developing rat cerebral cells in culture. Cell Tissue Res 252:181–190

37.

Saransaari

Oja

(2000) Taurine and neural cell damage. Amino Acids 19:509–526

38.

Schor

(1988) Inactivation of mammalian brain glutamine synthetase by oxygen radicals. Brain Res 456:17–21

39.

Schurr

Payne

Miller

Rigor

(1997) Glia are the main source of lactate utilized by neurons for recovery of function posthypoxia. Brain Res 774:221–224

40.

Shank

Bennett

Freytag

Campbell

(1985) Pyruvate carboxylase: an astrocyte-specific enzyme implicated in the replenishment of amino acid neurotransmitter pools. Brain Res 329:364–367

41.

Spranger

Schwab

Desiderato

Bonmann

Krieger

Fandrey

(1998) Manganese augments nitric oxide synthesis in murine astrocytes: a new pathogenetic mechanism in manganism? Exp Neurol 149:277–283

42.

Tholey

Bloch

Ledig

Mandel

Wedler

(1987) Chick brain glutamine synthetase and Mn²⁺–Mg²⁺ interactions. Neurochem Res 12:1041–1047

43.

Tholey

Megias-Megias

Wedler

Ledig

(1990) Modulation of Mn₂⁺ accumulation in cultured rat neuronal and astroglial cells. Neurochem Res 15:751–754

44.

Underwood

(1981) Trace metals in human and animal health. J Hum Nutr 35:37–48

45.

Waagepetersen

Bakken

Larsson

Sonnewald

Schousboe

(1998) Metabolism of lactate in cultured GABAergic neurons studied by ¹³C nuclear magnetic resonance spectroscopy. J Cereb Blood Flow Metab 18:109–117

46.

Wedler

Ley

Grippo

(1989) Manganese(II) dynamics and distribution in glial cells cultured from chick cerebral cortex. Neurochem Res 14:1129–1135

47.

Scully

de Vellis

(1988) Induction of glutamine synthetase in rat astrocytes by co-cultivation with embryonic chick neurons. J Neurochem 50:929–935

48.

Yamada

Ohno

Okayasu

Okeda

Hatakeyama

Watanabe

Ushio

Tsukagoshi

(1986) Chronic manganese poisoning: a neuropathological study with determination of manganese distribution in the brain. Acta Neuropathol 70:273–278

49.

Zheng

Ren

Graziano

(1989) Manganese inhibits mitochondrial aconitase: a mechanism of manganese neurotoxicity. Brain Res 799:334–342

50.

Zheng

Zhao

Slavkovich

Aschner

Graziano

(1999). Alterations of iron homeostasis following chronic exposure to manganese in rats. Brain Res 833:125–132

51.

Zwingmann

Richter-Landsberg

Brand

Leibfritz

(2000) NMR spectroscopic study on the metabolic fate of [3–(13)C]alanine in astrocytes, neurons, and cocultures: implications for glianeuron interactions in neurotransmitter metabolism. Glia 32:286–303

52.

Zwingmann

Richter-Landsberg

Leibfritz

(2001) ¹³C isotopomer analysis of glucose and alanine metabolism reveals cytosolic pyruvate compartmentation as part of energy metabolism in astrocytes. Glia 34:200–212

Energy Metabolism in Astrocytes and Neurons Treated with Manganese: Relation among Cell-Specific Energy Failure,Glucose Metabolism,and Intercellular Trafficking Using Multinuclear NMR-Spectroscopic Analysis

Abstract

Keywords

MATERIALS AND METHODS

Cell cultures

Incubation and preparation of cell extracts

Biochemical analysis

Nuclear magnetic resonance spectroscopy

Quantification of metabolite concentrations

Calculation of fractional 13C-enrichments and amounts of 13C-labeled metabolites

13C-isotopomer analysis

Statistical analysis

RESULTS

Cellular energy state

Glycolytic metabolic flux of glucose

Concentrations of TCA cycle–related metabolites and organic osmolytes

Cell-specific [1–13C]glucose metabolism

DISCUSSION

References

Calculation of fractional ¹³C-enrichments and amounts of ¹³C-labeled metabolites

¹³C-isotopomer analysis

Cell-specific [1–¹³C]glucose metabolism