DeBerardinis et al. 10.1073/pnas.0709747104. |
Fig. 6. Glucose is the predominant lipogenic precursor during glioblastoma cell proliferation. To confirm the hypothesis that glucose was the predominant lipogenic precursor in SF188 cells, we cultured cells in medium containing 5 mM [1,6-13C2]glucose and 5 mM [2-13C]glucose. In this experiment, half of the total glucose-derived acetyl-CoA pool is labeled at C-2 (from [1,6-13C2]glucose), and one quarter is labeled at C-1 (from [2-13C]glucose). Fatty acid labeling at w-2 is derived from [1,6-13C2]glucose via [2-13C]-Ac-CoA, whereas labeling at w-1 is derived from [2-13C]glucose via [1-13C]-Ac-CoA. Therefore, the likelihood (p) that a fatty acid labeled at w-2 is also labeled at w-1 can be predicted by the equation:
p
= (fglc total)(f[2-13C]-glc) = (0.6)(0.25) = 0.15,where fglc total is the fraction of the Ac-CoA pool derived from all glucose (assumed to be 0.6 based on data in Fig. 3C) and f[2-13C]glc is the fraction of that pool derived from [2-13C]glucose. Cells were cultured in this manner, and lipids were analyzed by 13C NMR spectroscopy. In the spectrum, the doublet at w-2 accounted for 0.14 of the total w-2 area (SI Fig. 6), very close to the predicted value of 0.15, confirming that the majority of the lipogenic acetyl-CoA pool is derived from glucose metabolism.
SI Materials and Methods
Perfusion Experiments.
Sterilized 200-mm diameter microcarriers (Cultispher, HyClone) were mixed with freshly trypsinized SF188 cells at a ratio of 107 cells per gram. The cells were grown in the microcarriers for 8-9 days in polypropylene bottles; the medium was changed once daily initially and twice daily near the end of the growth period. Cell density inside the microcarriers was monitored with 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT). The cell-laden microcarriers were perfused inside a 20-mm screw-top NMR tube with the apparatus detailed previously (1). Medium containing 13C-labeled substrates was introduced to the culture as a bolus that was circulated through the system by using a peristaltic pump. To extend the time course, a continuous feed with continuous product removal was used, so that nutrients could be maintained at steady state. In the perfusion with 10 mM [1,6-13C2]-glucose (Figs. 1, 2A and 3A), the initial bolus was 100 ml and the continuous feed was 51 ml/h. In the two-stage perfusion experiment (Figs. 4 and 5), the culture was perfused with a 95-ml bolus of complete medium containing 4 mM [3-13C]-glutamine and 10 mM unlabeled glucose. After 3.5 h, a continuous 3-h feed of 21 ml/h was established that maintained a constant total glutamine concentration. Next, the culture was perfused with a 115 ml bolus containing 4 mM [3-13C]-glutamine and 10 mM [1,6-13C2]-glucose. After 3.5 h, a new feed of 32 ml/h was established to maintain the glucose concentration at »3.6 mM. During both experiments, 2-ml samples of circulating medium were removed at hourly intervals for off-line analysis.NMR Spectroscopy and Data Analysis.
For 31P spectroscopy, acquisition parameters were 60o pulses, repetition time of 1,000 ms, 4,096 points, and a spectral width of 15,000 Hz. Free induction delays were zero filled to 8,192 points and apodized with exponential multiplication (15-Hz line broadening). Resonance areas were corrected for partial saturation with T1 values determined previously (2). During 13C spectroscopy, spectra were acquired with 60o pulses, repetition time of 1,200 ms, 4,096 points, 900 scans, and a spectral width of 25,000 Hz. Bilevel WALTZ-16 decoupling was used to produce nuclear Overhauser enhancement (N.O.E.) and to collapse 1H-13C. The free-induction decays were zero-filled to 16,384 points, and apodized with exponential multiplication. Line broadening of 3 Hz was used for most spectra, with line broadening of 0.2 Hz for quantitation of [4-13C]glutamate and [3,413C2]glutamate in Fig. 5 and SI Fig. 6. NMR spectra were analyzed by using NUTS 1D (Acorn NMR).For the perchloric acid experiment (Fig. 2B), cells were grown to 70% confluency in two T225 flasks. The culture medium was removed and replaced with medium containing 10 mM D-[2-13C]glucose for 6 h. Subsequently, the medium was removed, and the cells were rinsed in cold PBS. They were extracted with 4 ml of 10% ice-cold percholric acid, and the flask was washed further with 4 ml of ice-cold double-distilled water. Precipitated proteins were removed by centrifugation at 4,000 ´ g for 10 min, and the supernatant was neutralized with 8 M potassium hydroxide. The precipitate was removed by centrifugation, and the supernatant was frozen, lyophilized, and rehydrated with 0.45 ml of deuterium oxide. Insoluble material was removed by centrifugation, the pH was adjusted to 7.0 with 1 M KOD or DCl, and the sample was analyzed in a 5-mm NMR tube fit with susceptibility plugs (Doty Scientific). Samples were analyzed with a 9.4 T Varian spectrometer in a 5-mm probe. To analyze extracellular medium from this experiment, a 20-mm NMR tube was used, and NMR spectroscopy was performed under fully relaxed conditions with no N.O.E.
Lipid Biochemistry.
To calculate lipid synthesis rate by using cells cultured in D-[U-14C6]glucose, the radioactivity of 14C-labeled lipids was compared to a standard curve generated with D-[U-14C6]glucose to determine the molar amount of label in the extracted lipids. This value was extrapolated to the total glucose pool to determine a total rate of lipid synthesis from glucose. After 13 individual cultures, we obtained an average of 1.46 mmol per 109 cells per hour. To determine efficiency of recovery of labeled lipids, a parallel extraction was performed using [dipalmitoyl-1-14C]phosphatidylcholine (Perkin-Elmer). We recovered 60% of the labeled phosphatidylcholine. The lipid synthesis rate was corrected to account for incomplete recovery, yielding the final value of 2.4 mmol per 109 cells per hour.The 13C-labeled lipids were obtained after 103 h of culture in medium containing 10 mM D-[U-13C6]glucose. These cells were trypsinized, washed in PBS, and sonicated in chloroform/methanol (1:1). After three sequential organic extraction steps, pooled chloroform fractions were evaporated under nitrogen. Dry extracts were reconstituted in a ternary mixture of 400 ml of CDCl3/320 ml of CD3OD/160 ml of EDTA (0.2 M in D2O, pH 7.0). Spectra were collected on the CDCl3-rich lower phase containing lipids. NMR analysis was performed on a Bruker Advance 400 wide-bore 9.4 T instrument equipped with a 5-mm 1H/13C dual probe. The 13C spectra were recorded under approximately fully relaxed conditions and broadband proton decoupling. N.O.E was assumed to be the same for all protonated carbons. Carbon spectra were acquired under the following conditions: pulse flip angle, 45o; repetition time, 1.5 s; spectral width, 25 KHz; number of data points, 64,000; number of scans, »25,000. Free-induction decays were zero-filled to 131,072 points and apodized with exponential multiplication. Line broadening of 0.2 Hz was used. The CDCl3 signal at 77 ppm was used as the chemical shift reference. Spectra were analyzed by using the NUTS 1D software.
Gas Chromatography-Mass Spectrometry (GC-MS).
The content of 15NH3 in the medium was determined by converting NH3 to aminobutyrate in the presence of oxobutyrate and glutamate dehydrogenase. The 15N enrichment in aminobutyrate was determined by conversion to its t-butyldimethylsilyl derivative and analysis by GC-MS. The atom percent excess (A.P.E.) of aminobutyrate was determined by monitoring m/z at 173/172. The content of [15N]alanine in the medium was determined by converting alanine to its t-butyldimethylsilyl derivative and monitoring m/z at 233/232. The total ammonia/alanine content was multiplied by the A.P.E. at each time point to calculate a rate of 15NH3 production and of [15N]alanine production from L-[a-15N]glutamine.1. Mancuso A, Beardsley NJ, Wehrli S, Pickup S, Matschinsky FM, Glickson JD (2004) Biotechnol Bioeng 87:835-48.
2. Mancuso A, Zhu A, Beardsley NJ, Glickson JD, Wehrli S, Pickup S (2005) Magn Reson Med 54:67-78.