NMR IN BIOMEDICINE NMR Biomed. 2002;15:1–5 DOI:10.1002/nbm.725
Rapid Paper
Tricarboxylic acid cycle of glia in the in vivo human brain S. BluÈml,1,2* A. Moreno-Torres,1,2 F. Shic,1,2 C.-H. Nguy1,2 and B. D. Ross1 1 2
Huntington Medical Research Institutes, Pasadena, California, USA Rudi Schulte Research Institute, Santa Barbara, California, USA
Received 7 June 2001; Revised 30 July 2001; Accepted 2 August 2001
ABSTRACT: In the brain, acetate is exclusively oxidized by glia. To determine the contribution of glial metabolism to the tricarboxylic acid cycle (TCA), 1-13C-acetate was infused in six studies in three normal adult subjects and one epileptic receiving valproic acid for seizure control. Ten grams of 99% 1-13C labeled acetate were infused intravenously as a 3.3% w/v solution over 60 min, during which in vivo 13C MR spectra of the brain were acquired. As expected, 13C label rapidly enriched cerebral bicarbonate, glutamate and glutamine C5. The mean rate of acetate oxidation calculated from steady-state 13C enrichment of bicarbonate in fasted normal subjects was 0.13 0.03 mmol/ g/min (n = 4), approximately 20% of the total cerebral TCA cycle rate. Copyright 2002 John Wiley & Sons, Ltd. KEYWORDS: proton decoupled
13
C MRS; acetate metabolism; glutamine synthesis; glial TCA cycle
INTRODUCTION The normal adult human brain derives almost all of its ATP from the aerobic metabolism of glucose in the tricarboxylic acid (TCA) cycle. Carbon 13 (13C) magnetic resonance spectroscopy (MRS) studies of cerebral energy metabolism in vivo in normal1–4 and diseased humans5,6 employing 13C glucose (Glc) as the substrate, reflect predominantly the metabolism of neurons. Acetate (Ac), an alternative fuel metabolized to acetyl-CoA only in the glial compartment,7 has been used in combination with 13C glucose to separate glial from neuronal metabolism in cell preparations, tissue slices and in vivo in experimental animals.8,9 13C MRS after 13C labeled acetate infusion therefore has the potential to investigate normal human glial metabolism and to elucidate brain diseases which originate in glia. Using 13C MRS after intra-venous (i.v.) 1-13C acetate infusion we determined glial acetate oxidation rate in three normal human subjects (five examinations) and in *Correspondence to: S. Blu¨ml, Huntington Medical Research Institutes, 660 South Fair Oaks Avenue, Pasadena, CA 91105, USA. Email:
[email protected] Abbreviations used: TR, repetition time; TCA, tricarboxylic acid cycle; Glc, glucose; Ac, acetate; EAc, 13C fractional enrichment of Ac; Glu, glutamate; Gln, glutamine; EGlu, 13C fractional enrichment of Glu; EGln, 13C fractional enrichment of Gln; EHCO3, 13C fractional enrichment of HCO3 ; VTCA, TCA-cycle rate; VgAc, rate of acetate oxidation; Vxg, glial ketoglutarate/glutamate exchange rate; Vin, total flux of 13C label into HCO3 pool; Vout, total flux out of the HCO3 pool; Vtot, total flux of 12C and 13C bicarbonate in and out of the HCO3 pool; VnGlc, VgGlc, rates of glucose oxidation in the neuronal and glial TCA-cycle. Copyright 2002 John Wiley & Sons, Ltd.
a patient in whom glial glutamine metabolism was modified by the anti-epileptic drug, valproate.
METHODS Human subjects, infusion protocols and of Brain
13
C MRS
A total of six 13C MRS examinations was performed in four adult subjects, three of whom were normal controls. All but one subjects were fasted overnight prior to intravenous (i.v) administration of 10 g 1-13C Ac, 99% enriched (Cambridge Isotope Laboratories, Andover, MA), as a 3.3% w/v pyrogen-free solution, over 60 min. The patient, a 20 yrs-old female with epilepsy, was receiving valproate (2 g/day) and tegretol (0.3 g/day) for seizure control. One of the normal subjects was examined twice in the fasted state and another in both the fasted and fed states (Table 1). Two of the three normal subjects had previously undergone 13C MRS after 1-13C glucose administration, for determination of whole-brain metabolic rate for glucose using methods described in Mason et al.2 Plasma samples to determine total acetate concentration and 13C fractional enrichment of acetate were collected at 10–20 min intervals and stored frozen until analysis. Plasma acetate concentrations were determined enzymatically. 10 Fractional enrichment (EAc) was determined from the relative peak areas of 12C and 13C in 1H MR spectra of ethanol-deproteinized plasma samples NMR Biomed. 2002;15:1–5
¨ ML ET AL. S. BLU
2
Table 1. Summary of results in individual subjects
Diagnoses
Fed/ fasted
Control Control Control Control
1 2 2 3
Mean Standard deviation Control 1 Epileptic patient
At t 15 min
At t 60 min EAc (%)
EHCO3 (%)
VAca (mmol/min/g)
EGln5 (%)
697 443 258 310
75 68 63 63
6.9 4.3 3.9 4.6
0.18 0.10 0.12 0.13
5.5 3.5 5.0 2.5
427 196 290 478
67 6 41 67
4.9 1.4 2.3 3.2
0.13 0.03 0.04 0.07
4.1 1.4 2.0 4.5
Sex/age
[Ac0] (mmol/dl)
[Ac]
EAc (%)
[Ac]
Fasted Fasted Fasted Fasted
M/30 M/26 M/26 M/61
60 33 35 53
469 429 145 376
60 65 37 58
Fed Fasted
34 18 M/30 F/20
45 13 38 73
355 145 748 272
55 12 74 58
Assuming a total TCA cycle rate of 0.7 mmol/min/g as reported by Mason et al.2 and as measured in this laboratory (unpublished) in adult controls. [Ac0] = resting plasma Ac concentration; [Ac] = plasma Ac concentration after 1-13C Ac infusion; EAc = 13C fractional enrichment (%) of plasma Ac; EHCO3 = steady-state 13C fractional enrichment of bicarbonate; EGln5 = maximum 13C fractional enrichment of glutamine C5.
a
which had been dried under a stream of nitrogen and resuspended in D2O. MRI, 1H MRS and 13C MRS were performed on a clinical GE 1.5 T MR scanner. 13C MR spectra were acquired from the occipital region of the brain, before and every 5 min during and after 1-13C acetate infusion, as described in detail4,11 using a repetition time, TR, of 2 s. Approval of the Internal Review Board of Huntington Memorial Hospital and the FDA (IND 59,950) as well as informed consent of the subjects, was obtained.
Data processing of
13
C MR spectra of brain
In vivo brain spectra were processed off-line on a Sun SPARCstation 2 using the SA/GE software package (GE, Waukesha, Wi). Analyses were operator independent. Spectra averaged to 20 min acquisition blocks and zerofilled to 8192 data points, were Fourier transformed and phase corrected. A 2 Hz lorentzian-to-gaussian lineshape transformation was applied. To correct for the underlying broad baseline a low-pass gaussian filter (150 Hz) was applied, producing a spectrum that contained only broad resonances. This was then subtracted from the original spectrum. In in vivo spectra, Ac1 and Glu5 resonances cannot be separated and the term Ac1/Glu5 refers to their combined signal. Peak amplitudes for acetate C1 (Ac1) plus glutamate C5 (Glu5), glutamine C5 (Gln5), and bicarbonate (HCO3) were analyzed and time courses of label accumulation in Ac1/Glu5, Gln5 and HCO3 were constructed. 13C fractional enrichments EAc1/Glu5, EGln5, and EHCO3 were calculated from: EMetabolite Spost =Spre 1:1%
1
where Spre and Spost are the signal intensities of the resonances in the baseline scan and in scans after infusion start, and 1.1% is the natural abundance of 13C. Copyright 2002 John Wiley & Sons, Ltd.
Determination of the rate of acetate metabolism The sites of production of HCO3 from 1-13C acetate and from unlabeled precursors, predominantly glucose, in neurons and astrocytes are shown in Fig. 1(A). Acetate enters the TCA cycle directly. In the first turn, 1-13C label is transferred from glial a-ketoglutarate to glial glutamate C5 position (gGlu5) with the glial ketoglutarate/glutamate exchange rate (Vx) and from there to other metabolites. Release of 1-13C label as carbon dioxide (CO2 = bicarbonate, HCO3 ) occurs only in a subsequent turn of the TCA cycle. During the second turn of the glial TCA cycle, 13C accumulation is also expected at the C1 position of glutamate and glutamine. Unlabeled bicarbonate arises in glia by the action of pyruvate dehydrogenase on substrates other than acetate as they enter the TCA cycle, as well as from activity of the neuronal TCA cycle. The rate of acetate oxidation in glial cells, (VgAc), can be determined from the steady-state fractional enrichment of bicarbonate (EHCO3), as follows. At steady state, metabolite pools which accumulate 13 C label other than bicarbonate (here predominately glutamate C5 and glutamine C5, see results) have constant 13 C fractional enrichment and the total influx of 13C label into the glial and neuronal TCA cycle, Vin, is equal to the total influx into the bicarbonate pool [Fig. 1(A,B)]: Vin EAc VgAc
Vtot
VgAc 1:1%
2
The total outflux of 13C label from the bicarbonate pool is given by: Vout Vtot EHCO3
3
where Vtot is the total rate of unlabeled and 13C labeled bicarbonate influx/outflux. Inserting eqn (2) in eqn (3) gives: VgAc Vtot
EHCO3
1:1%=
EAc
1:1%
4
The total bicarbonate production Vtot is given by Vtot = 3 NMR Biomed. 2002;15:1–5
ASSESSMENT OF HUMAN GLIAL METABOLISM IN VIVO
3
Figure 1. (A) The transfer of 13C label from 1-13C Ac to TCA cycle intermediates and exchange partners is illustrated. Only relevant intermediates are shown. Glc, glucose; Ac, acetate; Glu, glutamate; Gln, glutamine; CO2, carbon dioxide; Pyr, pyruvate; KG, ketoglutarate; VgAc, rate of Ac oxidation; VgGlc, glial Glc oxidation rate; VnGlc, neuronal Glc oxidation rate; Vx, neuronal/glial a-KG/glutamate exchange rate; Vgln, glutamine synthesis rate. (B) In¯ux/out¯ux of 13C label into/out from the bicarbonate (HCO3 ) pool at steady state. EAc = 13C fractional enrichment of Ac; EHCO3 = 13C fractional enrichment of bicarbonate; Vtot = total rate of 12C and 13C label entering/exiting the bicarbonate pool
VnGlc 3 VgGlc 2 VgAc, where VnGlc and VgGlc are the rates of glucose oxidation in the neuronal and glial compartments [Fig. 1(A)]. Since VgAc VnGlc, Vtot is approximately three times the total neuronal/glial TCA cycle rate (Vtca).
RESULTS Figure 2(A) shows the time course of plasma acetate concentrations and 13C fractional enrichment after 1-13C Ac administration over 60 min in a representative control subject. Accompanying the approximately 10-fold increase in plasma acetate within the first 10–15 min of infusion, on average a 60% 13C fractional enrichment of acetate (EAc) was observed between 15 and 60 min (Table 1). 13C-enriched resonances of Ac1/Glu5 and Gln5 were observed within 20 min and H13CO3 was identified within 30 min [Fig. 2(B)]. Between 30 and 80 min of Ac infusion the resonances of Ac1/Glu5 and Gln5 were essentially constant. The steady-state enrichment of HCO3 estimated from the time course of fractional 13C enrichments of HCO3 [Fig. 2(C)] and VAc calculated for each examination are shown in Table 1. Assuming an overall TCA-cycle rate of 0.7 mmol/min/g for the intact Copyright 2002 John Wiley & Sons, Ltd.
human brain (unpublished 1-13C glucose MRS studies in this laboratory consistent with results reported by Mason et al.2), the average rate of Ac oxidation, VgAc, of fasted normal subjects was 0.13 0.03 mmol/min/g (n = 4). As expected, due to the exclusive presence of glutamine synthetase and direct metabolism of 1-13C acetate in glia, Gln5 enrichment was rapid reaching a maximum of 4.1 1.4% in fasted control subjects 55–80 min after infusion start. In the patient receiving valproate 13C accumulation in Gln5 was markedly increased, and accumulation of H13CO3 was markedly decreased. In addition, both in the patient treated with valproate and in the single fed subject, VgAc was lower than in fasted controls and the Gln5 enrichment in the fed control subject was reduced (2% Table 1).
DISCUSSION While several investigators have defined the cerebral metabolic rate of acetate in experimental animals, using 13 C MRS in vitro or in vivo,8,9 thereby distinguishing between glial and neuronal metabolism, to the best of our knowledge, these are the first studies demonstrating the NMR Biomed. 2002;15:1–5
¨ ML ET AL. S. BLU
4
Figure 2. (A) Plasma acetate concentrations and 13C fractional enrichment in a representative control. (B) Spectra obtained from one of the normal subjects before and 30±55 and 55±80 min after intravenous 1-13C Ac infusion. In controls, the signal intensity of Gln5 in the spectra acquired after 55±80 min is 0.7 times that of HCO3. In an equivalent spectrum from the patient, the Gln5 amplitude is twice as intense as HCO3 (lower trace). Ac1 and Glu5 resonances could not be separated; the term Ac1/Glu5 refers to the combined signal. (C) 13C fractional enrichment of HCO3 versus time with superimposed ®t curve in a representative control
potential of 13C MRS after 1-13C acetate to monitor metabolism of normal and abnormal human brain in vivo. The rate of cerebral acetate oxidation in the glia was determined from the steady-state 13C fractional enrichment of bicarbonate (EHCO3). Although some Gln5 produced from acetate will be transferred from the glia to neurons, deaminated there to Glu5, and oxidized in the neuronal TCA-cycle, all 13C label accumulating in 13CO2 had to enter through the glial TCA cycle, as the neurons do not metabolize acetate.7 Therefore, at steady state, the rate of 13CO2 production is equal to the rate of acetate influx into the glial TCA cycle. Unlabeled substrates, predominantly glucose, are oxidized both in glia and neurons and contribute to the cerebral HCO3 production. The steady-state fractional enrichment of HCO3 (EHCO3) is therefore a measure of the relative influx of 13C and 12C label, and the rate of acetate oxidation can be determined from EHCO3. The computation of EHCO3 is complicated by the fact that steady states were not reached in our preliminary studies. Instead, to determine EHCO3, an exponential function was fitted to the time course of 13C fractional enrichment in each subject [Fig. 2(C)]. This Copyright 2002 John Wiley & Sons, Ltd.
approach is valid if we can assume that all 13C label entering the glial TCA cycle is being transferred to bicarbonate and not trapped in other pools, predominantly Gln5 and Glu5. Visual inspection showed that Gln5 and Glu5 peak amplitudes are approximately constant between about 30 and 80 min after start of the infusion, and the fit procedure was applied only to data points acquired during this interval. The rate of glial acetate oxidation determined in this way was 20% (0.13 mmol/ min/g assuming a total TCA cycle rate of 0.7 mmol/min/ g) of the total cerebral TCA cycle rate in fasted control subjects. Detailed studies of the impact of the acetate supply on its metabolic rate in the brain are necessary before this technique can be applied for systematic studies of brain diseases. However, the potential of this method is illustrated by a strikingly altered pattern of 13C enrichment observed in a single patient treated with the antiepileptic drug valproate. Valproic acid causes a relative inhibition of the hepatic urea cycle with increased ammonia delivery to the brain.12 Ammonia is metabolized by converting glutamate into glutamine by NMR Biomed. 2002;15:1–5
ASSESSMENT OF HUMAN GLIAL METABOLISM IN VIVO
glutamine synthetase, an enzyme exclusively located in the glial compartment, and the assessment of increased 13 C accumulation in Gln (as observed in our patient) would be a suitable method to explore the impact of valproic acid on cerebral metabolism. We also speculate that the activity of acetyl-CoA synthetase, the enzyme which most likely controls the rate of entry of acetate (and other more physiological fuels) to the TCA-cycle, can be studied by applying this technique in patients and in fed and fasted controls. In earlier studies it was demonstrated that after i.v. 1-13C glucose infusion, labeling of Glu1,2,3,4, Gln1,2,3,4, Asp2,3, NAA2,3, lactate3, alanine3, and HCO3 can be observed and new insights into the pathophysiology of brain diseases can be obtained.4,5 Using 1-13C acetate and the same MR technique we are now able to investigate glial metabolism. This study is highly suggestive that 13C MRS after 1-13C acetate infusion is feasible in humans and an appropriate tool to investigate putative abnormalities in glial metabolism in human brain disorders. Acknowledgements This work was supported by Cambridge Isotope Laboratories Inc, Andover, MA and the Rudi Schulte Research Institute, Santa Barbara, CA. The authors wish to thank Lea Lim RN and Alexander Lin for expert assistance, and Ms Darlette Luke, Fairview University Hospital, Minneapolis, for preparation of the 1-13C acetate. We are grateful to Dr Keiko Kanamori, Ph.D. for her insights.
Copyright 2002 John Wiley & Sons, Ltd.
5
REFERENCES 1. Beckmann N, Mueller S. Natural-abundance 13C spectroscopic imaging applied to humans. J. Magn. Reson. 1991; 93: 186–195. 2. Mason GF, Gruetter R, Rothman DL, Behar KL, Shulman RG, Novotny EJ. Simultaneous determination of the rates of the TCA cycle, glucose utilization, a-ketoglutarate/glutamate exchange, and glutamine synthesis in human brain by NMR. J. Cereb. Blood Flow Metab. 1995; 15: 12–25. 3. Gruetter R, Seaquist ER, Kim S, Ugurbil K. Localized in vivo 13CNMR of glutamate metabolism in the human brain: Initial results at 4 Tesla. Devl. Neurosci. 1998; 20: 380–388. 4. Bluml S, Hwang J-H, Moreno A, Ross BD. Novel peak assignments of in vivo 13C MRS in human brain at 1.5 T. J. Magn. Reson. 2000; 143: 292–298. 5. Bluml S, Moreno A, Hwang J-H, Ross BD. 1-13C glucose magnetic resonance spectroscopy of pediatric and adult brain disorders. NMR Biomed. 2001; 14: 19–32. 6. Moreno A, Ross BD, Bluml S. Direct determination of the Nacetyl-L-aspartate synthesis rate in the human 13C MRS and [1-13C] glucose infusion. J. Neurochem. 2001; 77: 347–350. 7. Muir D, Berl S, Clarke DD. Acetate and fluoroacetate as possible markers for glia metabolism in vivo. Brain. Res. 1986; 380: 336– 340. 8. Cruz F, Cerdan S. Quantitative 13C NMR studies of metabolic compartmentation in the adult mammalian brain. NMR Biomed. 1999; 12: 451–462. 9. Bachelard H (ed). Advances in Neurochemistry, Vol. 8. Plenum Press: New York, 1997. 10. Bartelt U, Kattermann R. Enzymatic determination of acetate in serum. J. Clin. Chem. Clin Biochem. 1985; 23: 879–881. 11. Bluml S. In vivo quantitation of cerebral metabolites concentration using natural abundance 13C MRS at 1.5 T. J. Magn. Reson. 1999; 136: 219–225. 12. Goodman AG, Gilmann LS. The Phamacological Basis of Therapeutics. Macmillan, New York, 1985; 461–472.
NMR Biomed. 2002;15:1–5
