Biochemistry 1992,31, 1 1 1 59-1 1165
11159
'H NMR Observation of Redox Potential in Liver? Youngran Chungt and Thomas Jue*l* Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 0651 1 Received December 3, 1991; Revised Manuscript Received August 10, 1992
ABSTRACT: lH N M R spectral editing techniques can select the distinct signals of lactate, pyruvate, @-hydroxybutyrate,and acetoacetate and provide a unique way to monitor the biochemical processes in vivo. These metabolite levels reflect the near-equilibrium dehydrogenase activity and therefore the cellular redox state. The quantitative comparison between the 1H N M R and biochemical assay data is in excellent agreement. Lactate/pyruvate and fl-hydroxybutyrate/acetoacetate ratios, obtained from normalized lH N M R spectra, respond directly to changes in the cytosolic and mitochondrial redox states. Because N M R is noninvasive, our results set the groundwork for implementing these techniques to observe tissue redox states in vivo.
The cellular redox potential plays a central role in regulating bioenergetics and metabolism. It is characterized by the ratio of free NAD/NADH' (oxidized and reduced states, respectively, of nicotinamide adenine dinucleotides) and is commonly measured by fluorescence (Chance & JBbsis, 1959; Chance et al., 1962) or biochemical assay methods (Williamson et al., 1967). The optical technique excites NADH at 366 nm and monitors the fluorescence at 450 nm. Biochemical assay assesses indirectly the cytosolic and mitochondrial NAD/ NADH by measuring lactate/pyruvate and /3-hydroxybutyrate/acetoacetate ratios, respectively (Williamson et al., 1967). These substrates react with the near-equilibrium dehydrogenase reactions involving pyridine dinucleotides and reflect the free NAD/NADH pool. Current techniques to measure cellular redox state, however, are invasive and are limited. Both fluorescence and biochemical assay methods require invasive probes and lack sampling specificity. Fluorescence can only measure surface tissue, and biochemical assay requires destructivemanipulation or detailed accounting of contributions between the local tissue region and the distal sampling point. Recent NMR advances have suggested an alternative method. Sincelactate(R0thmanet al., 1984; Jueet al., 1985) and pyruvate signals in vivo (Jue et al., 1988) are detectable with lH NMR spectral editing techniques, tissue redox state is then potentially observable also. For lactate, 'H NMR spectral editing technique manipulates the &modulation in a spin-echo pulse sequence to select the CH3 signal and to suppress the endogenous background resonances (Rothman et al., 1984; Jue et al., 1985). For pyruvate, an anti-editing strategy discriminates against the signals of coupled spin systems (Jue et al., 1988). We undertook a perfused liver study to establish the feasibility of utilizing the 'H NMR editing strategy to measure f This work was supported by National Science Foundation Grant PCM 840270. Preliminary findings were presented at meetings of the Society of Magnetic Resonance in Medicine in 1987, 1988, and 1989. * To whom correspondence should be addressed. Telephone: (916) 752-4569. FAX: (916)752-3516. t Present address: Department of Biological Chemistry, University of California, Davis, CA 95616. 1 Abbreviations: NAD, oxidized &nicotinamide adenine dinucleotide; NADH, reduced @-nicotinamideadenine dinucleotide; LDH, lactate dehydrogenase;TSP, sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4;B/A, j3-hydroxybutyrate/acetoacetate;BHB, @-hydroxybutyrate;BDH, B-hydroxybutyrate dehydrogenase; L/P, lactate/pyruvate.
-
0006-2960/92/043 1 1 1 159$03.00/0
lactate/pyruvate and to demonstrate that these editing methods will also select the signals of o-hydroxybutyrateand acetoacetate, which have similar spin systems as lactate and pyruvate, respectively. The NMR data are in excellent agreement with the standard biochemical assay results and underscore the potential to measure cellular redox states in vivo.
MATERIALS AND METHODS Rat Liver Perfusion. The rat liver perfusion method was previously described (Jue et al., 1988). Male Sprague-Dawley rats were fasted 24 h before the experiment. Initially, the isolated liver was washed for 25 min with nonrecirculating Krebs bicarbonate buffer at pH 7.35-7.45 to remove endogenous lactate and pyruvate before recirculation. The isolated liver was inserted into a 20-mm NMR tube, which was then placed in a Bruker AM 360 wideborevertical magnet equipped with 20 mm 'HqX} probe. The probe characteristics and the perfusion method have been well established in the laboratory. The perfusate was oxygenated by a membrane oxygenator exposed to 95% 02/5% COz. Fifty feet of silastic tubing in the oxygenator was coiled around a temperature jacket. The liver perfusate temperature was maintained at 36-37 OC.The flow rate was 5 mL min-1 (g of liver)-l to ensureproper oxygen delivery in this hemoglobin-free perfusate. Red blood cells and albumin were omitted in the perfusate to minimize the NMR spectral interferences. The perfused liver is, nevertheless,viable (FrBhlich et al., 1973; Sies, 1978; Suganoet al., 1978). The recirculating perfusate volume was 150 mL and contained 6%D2O for field-frequency lock. The isotope effect on the lactate dehydrogenase reaction at this concentration of D20 was negligible. Up to 20%D20, the in vitro assay of LDH extracted from the in situ rat liver showed no significant isotope-related difference in the V,,,,,. During all experiments, the liver viability was monitored by 02consumption, stable perfusate pH, a high ATP/Pi ratio spectra, and LDH leakage into the perfusate. 0 2 from the 31P consumption was greater than 1 pmol min-l (g wet weight)-' for the fasted liver (Gores et al., 1986). Glucose assay of the perfusate sampled periodically during the experiment yielded a linear production rate of 1 .O f 0.1 pmol min-I (g wet weight of liver)-', which is a key indicator of the metabolic integrity of the fasted liver (Ross et al., 1967). With our perfused liver system, the LDH leakage is less than 10 units/h into the perfusate, well within the established viability criterion 0 1992 American Chemical Society
11160 Biochemistry, Vol. 31, No. 45, 1992 (Lemasters et al., 1983; Sies, 1978). Each liver was also inspected for gross edema. Any data from liver with a wet/ dry weight ratio exceeding 3.8 were discarded (Tischler et al., 1977a). Wet liver weight was determined after removing any nonhepatic tissue and blotting with filter paper. lH N M R Spectral Editing. Homonuclear editing techniques to select the 'H signal of C&H lactate (Jue et al., 1985) and an anti-editing strategy to extract the C&H pyruvate signal (Jue et al., 1988) in the lH NMR spectra of the perfused liver have been previously described. NMR spectra were collected with a 20-mm 'HdXJ probe on an AM 360 Bruker spectrometer; 4K data points and 5-kHz spectral width were used. An inversion recovery pulse sequence was used to determine the TISof lactate and pyruvate in the solution at 37 OC, and the data were processed by the Bruker TI fitting routine. All peaks were referenced to the H20 signal as 4.76 ppm at 25 OC, which was calibrated against TSP. The lH NMR signal intensities were calibrated against a known concentration of infused substrate. A separate set of data was obtained by spectrophotometric assay of the corresponding perfusate, sampled at periodical intervals throughout the experiment. These two data sets were then compared. Redox measurement requires only a ratio of lactate/ pyruvate or &hydroxybutyrate/acetoacetate, not the absolute concentrations of each substrate. To scale the appropriate NMR signal intensities requires a normalization factor, obtained from 10 mM pyruvate and lactate solution signals, under experimental pulsing conditions. The ionic composition, pH, and temperature of the analytical solution were the same as the Krebs bicarbonate buffer used during the liver experiments. The intensity ratio of edited lactate and pyruvate signals was then calibrated against the corresponding fully relaxed lH resonances. Similar experiments gave the normalization factor for the B/A ratio. Metabolite Assay. Aliquots of perfusate were withdrawn at 10-min intervals in the course of NMR experiments and were immediately deproteinized with ice-cold 1 N perchloric acid. Acidic samples were neutralized and assayed enzymatically for pyruvate (Czok & Lamprecht, 1974), lactate (Noll, 1984), ff-hydroxybutyrate (Williamson & Mellanby, 1974), acetoacetate (Mellanby & Williamson, 1974), and glucose (Bergmeyer et al., 1974). All the chemicals and enzymes were purchased from Sigma and used without further purification. Hepatic LDH was extracted from in situ rat liver. Male Sprague-Dawley rats were anesthetized with sodium pentobarbital (50 mg/kg body weight), and the liver was immediately removed from the animal. Approximately 1 g of liver tissue was homogenized in 5 volumes of 0.1 M phosphate buffer, and the homogenate was centrifuged at 18OOOg at 4 OC for 30 min (Hyldgaard-Jensen & Valenta, 1970). LDH activity in the diluted supernatant was then measured in the reaction mixture containing saturating concentrations of pyruvate (0.63 mM) and NADH (0.18 mM) as the substrates (Bergmeyer & Bernt, 1974). For the LDH K, measurement of pyruvate and NADH, pyruvate and NADH concentrations in the reaction mixture were varied, and corresponding LDH activities were measured. Estimate of Percentage Signal in Liver us Perfusate. A point H 2 0 sample (
