Pyruvate produces inotropic responses in the adult reperfused heart. Pyruvate oxidation and anaplerotic entry into the tricarboxylic acid (TCA) cycle via carboxylation are linked to the stimulation of contractile function. The goals of this study were to determine if these metabolic pathways operate and are maintained in the developing myocardium after reperfusion. Immature male swine (age: 10–18 days) were subjected to cardiopulmonary bypass (CPB). Intracoronary infusion of [2-13C]pyruvate (to achieve an estimated final concentration of 8 mM) was given for 35 min, starting either during weaning (group I) and after its discontinuation (group II) or without (control) CPB. Hemodynamic data were collected. 13C NMR spectroscopy was used to determine the fraction of pyruvate entering the TCA cycle via pyruvate carboxylation (PC) to total TCA cycle entry (PC plus decarboxlyation via pyruvate dehydrogenase). Liquid chromatography-mass spectrometry was used to determine total glutamate enrichment. Pyruvate infusion starting during the weaning of mechanical circulatory support improved maximum dP/dt (P < 0.05) but waiting to start the infusion until after the discontinuation of CPB did not. Glutamate fractional enrichment was confirmed by liquid chromatography-mass spectroscopy as adequate (>5%) to provide signal to noise in the NMR experiment in all groups. The ratio of pyruvate carboxylase to total pyruvate entry into the TCA cycle did not differ between groups (group I: 20 ± 4%, group II: 23 ± 7%, and control: 27 ± 7%). These data show that robust PC operates in the neonatal pig heart and is maintained during reperfusion under conditions that emulate CPB and reperfusion in human infants.
- cardiac metabolism
- pyruvate carboxylation
- pyruvate decarboxlyation
- citric acid cycle
- tricarboxylic acid cycle
cardiac dysfunction is an important problem in infants after the repair of congenital heart defects (14, 17). Studies in neonatal animal models have suggested that enhanced myocardial reliance on carbohydrate metabolism plays an important role in determining postischemic dysfunction (3, 29). The neonatal heart demonstrates a shift in substrate metabolism from predominantly carbohydrate to fatty acid oxidation within the first 1–2 wk after birth (2, 4, 18). The newborn myocardium also retains higher levels of glycogen stores during this transition period (33). However, the newborn heart depletes less glycogen than the adult heart during normothermic ischemia (36). Elevated lactate and glucose-6-phosphate levels in the newborn suggest that these end products rapidly accumulate and inhibit glycogenolysis and glycolysis (7, 35, 36). Inhibition of these pathways can be reduced by washout (35), such as occurs with cardioplegia, but this strategy should result in irreversible loss of carbon substrate. This will, in turn, deplete the heart of tricarboxylic acid (TCA) cycle intermediates and thereby reduce the maximal oxidative capacity and cardiac work in the postischemic period.
Exogenous pyruvate supplementation ameliorates TCA cycle intermediate depletion in the adult porcine myocardium after reperfusion (16). Pyruvate also acts as an inotropic agent for both the aerobic and reperfused adult heart through several mechanisms (19, 21, 37). As noted, this natural carbohydrate serves as a metabolic intermediate. Pyruvate also mitigates mitochondrial oxidative stress during reperfusion, presumably through the production of NADPH (20). Pyruvate dehydrogenase (PDH) oxidizes pyruvate to acetyl-CoA and releases NADH in the mitochondria. Acetyl-CoA then enters the TCA cycle as a two-carbon moiety and undergoes condensation with oxaloacetate (OAA) by citrate synthetase, forming citrate. OAA recycles through the TCA cycle, but the concentration of the intermediates can be depleted by constant efflux from the mitochondria in the normal myocardium. This efflux, termed “cataplerosis,” may be accentuated under stress conditions such as a very high work state or ischemia/reperfusion (25). The small size of the TCA cycle intermediate pool relative to the high rate of CO2 extraction creates cycle vulnerability and may limit oxidative capacity during stress conditions, where oxygen demand increases or supply decreases. TCA cycle intermediate replenishment through anaplerotic reactions must match cataplerosis under normal aerobic conditions or even exceed cataplerosis during reperfusion (10).
Anaplerosis through pyruvate carboxylation (PC) to OAA provides a pathway for TCA cycle intermediate repletion. This anaplerotic pathway displays sensitivity to myocardial perfusion conditions. For instance, PC is depressed in the regionally hibernating mature porcine myocardium during low-flow ischemia (26). However, the contribution of PC to the TCA cycle in the developing myocardium in vivo is unknown and may be dependent on the maturation of enzyme systems such as PDH and pyruvate carboxylase. Furthermore, the specific role of pyruvate anaplerosis in replenishing the TCA cycle after global ischemia and reperfusion has not been examined even in the mature heart in vivo, which is subjected to the combined influence of multiple substrates, hormones, and neurological stimuli. Therefore, it is unknown whether this specific anaplerotic pathway represents a reasonable target for TCA cycle repletion during reperfusion in the neonatal heart.
Accordingly, the objectives of the present investigation were to 1) confirm a reliable method for the simultaneous estimation of the relative rates of PC and decarboxylation in the immature porcine heart in vivo, 2) determine if these relative rates were similar to published data obtained in the adult heart in vivo, and 3) determine if these two pathways were preserved during reperfusion in an animal model in vivo that emulates cardiopulmonary bypass (CPB) and ischemia in infants and children. Experiments were performed in immature male swine between the ages of 10 and 18 days subjected to these CPB conditions and with intracoronary infusion of sodium [2-13C]pyruvate. Previous studies have employed gas chromatography-mass spectrometry (GC-MS) methodology exclusively to determine the relative rates of pyruvate decarboxylation/PC (27). In this study, we used 13C detection by MRS and isotopomer analyses in addition to glutamate enrichment estimation by liquid chromatography-mass spectrometry (LC-MS).
MATERIALS AND METHODS
The use of pigs in our research was approved by the Children's Hospital and Regional Medical Center Animal Care Committee. All research adhered to the American Physiological Society's “Guiding Principles in the Care and Use of Animals.” For the experiments, male piglets between the ages of 10 and 18 days old and weighing between 9 and 13 kg were used. Sedation was achieved through an intramuscular injection of ketamine (33 mg/kg), atropine (0.02 mg/kg), and xylazine (2 mg/kg). Under propofol anesthesia (3mg/kg intravenous loading followed by 3 mg·kg−1·h−1 for maintenance), animals were intubated and ventilated with 40% oxygen. The femoral artery and vein were cannulated separately to monitor blood pressure, collect blood samples, and infuse propofol and a bolus of 5% dextrose-9% NaCl solution. Blood Po2, pH, and CO2 levels were measured by Radiometer ABL 820 (Radiometer America, Westlake, OH). After the performance of a median sternotomy, a flow probe was placed around the ascending aorta to measure cardiac output (TS420, Transonic Systems, Ithaca, NY). A Millar high-fidelity micromanometer (Millar Instruments, Houston, TX) was passed into left ventricle (LV) via the apex to measure LV pressure. The aorta and right atrial appendage were cannulated. The hemiazygous vein was ligated. A cannula was placed through the right atrial appendage and into the coronary sinus to direct coronary venous drainage extracorporeally into the jugular vein. A Transonic flow probe was placed around the coronary sinus shunt for continuous flow monitoring. Myocardial oxygen consumption (MV̇o2) was measured via the flow and oximetry data from the coronary sinus catheter (with frequent blood sampling to confirm the latter by blood gas and hemoximeter analyses).
Before the pyruvate infusion (control) or during ischemia (groups I and II), the left anterior descending coronary artery (LAD) was cannulated. A 24-gauge BD Saf-T-Infusion catheter (Becton Dickinson, Sandy, UT) was advanced retrograde starting just distal to the origin of the first branch from the LAD. Confirmation of intravascular positioning was made through aspiration of either cardioplegia or blood and injection of heparinized saline. The catheter was then carefully sutured in place.
In separate groups of pigs, sterile cannulae were placed in the ascending aorta and connected to a CPB circuit with a Stockert-Shiley roller pump and a hollow fiber oxygenator (Minimax X Plus, Medtronic), and animals were given heparin (350 U/kg). The circuit was primed with 500 ml of 10% Dextran 40 and 0.9% sodium chloride solution. The circuit exchanger and a heating pad initially maintained temperature at 37°C. The esophagus and radial temperature were monitored with thermo probes.
28) was instilled antegrade into the aortic root. After 60 min of ischemia, the snare was released, and reperfusion was initiated. Swine were gradually rewarmed over 10 min to 36°C. Circulatory support was gradually decreased starting 20 min after cross-clamp release and discontinued 40 min after release. Defibrillation was performed if necessary.groups I and II. After 10 min of stable parameters had been achieved, the hypothermia protocol was initiated, and data were recorded continuously. Subsequently, cooling was initiated through the circuit, generally achieving a core temperature of 30°C within 20 min. The aorta snare was tightened, and cardioplegia solution [composed of (in μEq/l) 140 Na+, 45 K+, 3 Mg2+, 104 Cl−, 27 acetate, and 23 gluconate, adjusted to pH 7.8] (
For all groups, labeled sodium [2-13C]pyruvate was infused directly into the LAD for 35 min. The infusion was given to achieve an estimated final concentration of 8 mM. Calculation of the pyruvate infusion rate was based upon the mean LV coronary artery flow per body weight calculated in preliminary neonatal pig experiments and followed methodology similar to that used by Hermann et al. (11) in humans. This concentration is supraphysiological but near the range used clinically to improve contractile function (11).
The control group (n = 5) did not undergo CPB or ischemia-reperfusion. Control pyruvate infusion began after stabilization and coronary artery canalization. A graphical representation of the ischemia/reperfusion protocol is shown in Fig. 1. In group I (n = 5), the pyruvate infusion began 10 min after the release of the aortic cross-clamp. Thus, the pyruvate infusion in group I started while on full CPB support and continued during the gradual weaning and discontinuation of mechanical support. In group II (n = 4), the pyruvate infusion began 5 min after the cessation of CPB. Immediately upon completion of the pyruvate infusion, portions of the heart perfused by the LAD were rapidly excised and freeze clamped. Myocardial tissue was frozen in liquid nitrogen and stored at −70°C for further analyses.
Heart extracts were prepared using perchloric acid as previously described (12) for the determination of specific carbon glutamate labeling by 13C MRS and 13C glutamate enrichment by LC-MS. NMR spectra were obtained within a 500-MHz Brueker at the 13C resonance frequency using a 60° pulse angle and 4-s recycle delay and 16 K data points. Protons were decoupled with a Waltz decoupling scheme. Fourier-transformed spectra were fitted with commercial software (NMR NUTS), and areas for glutamate C2, C3, and C5 appearing as singlets were determined. The 13C NMR spectrum is complex due to detection multiple peaks caused by the natural abundance of 13C in metabolites occurring at high concentrations. Atlases identify frequency shifts for various peaks, including glutamate. However, the particulars of the ionic environment in the porcine myocardium can alter the chemical shift. To accurately identify peaks, we determined chemical shift assignments by spiking nonlabeled pig heart extracts with high concentrations of unlabeled glutamate.
[2-13C]pyruvate will yield [2-13C]OAA in a carboxylation reaction. This OAA should scramble “backwards” into the fumarate pool, yielding [2-13C]OAA or [3-13C]OAA. On the first turn of the cycle, this will yield [3-13C]glutamate or [2-13C]glutamate. On later turns, [1-13C]glutamate will be produced. Assuming that pyruvate recycling (OAA → pyruvate → acetyl-CoA) does not occur, the carboxylation reaction will never produce label in C5 or C4 of glutamate.
In a decarboxylation pathway, [2-13C]pyruvate will yield [1-13C]acetyl-CoA and produce [5-13C]glutamate in the first turn. After losing the C1 from α-ketoglutarate and scrambling in the succinate-fumarate pool, the next turn of the cycle will allow some [1-13C]glutamate. Through PDH (a decarboxylation pathway), [2-13C]pyruvate can yield only [5-13C]glutamate and potentially a limited amount of [1-13C]glutamate but never label in C2, C3, or C4. Accordingly, the percent contribution of citrate by PC was determined by area C2 + C3/(C2 + C3 + C5) and pyruvate decarboxylation by C5/(C2 + C3 + C5).
Isotopic enrichment of tissue extract glutamic acid was analyzed using MS in the form of butyl ester (1). For an improved signal-to-noise ratio and to remove tryptophane interference, a short reverse-phase LC-MS method was used. Agilent 1100 HPLC with Zorbax SB-3 C18 column (2.1 × 100 mm, 80-A pores, and 3.6-μm particles) was coupled to Esquire LC ion trap mass spectrometer operated in positive ionization mode. Water with 5% acetonitrile-1% acetic acid and acetonitrile-1% acetic acid were used as solvents A and B, respectively. A gradient from 10% solvent B to 64% solvent B in 9 min was followed by 5 min at 100% solvent B, and the column was equilibrated during 12 min at a flow rate of 200 μl/min. The glutamic acid retention time was 8.5 min. The internal standard phenylalanine eluted at 7.5 min and allowed the assessment of the amount of glutamic acid in the sample.
Values are reported as means ± SE in the text, figures, and tables. Data were analyzed with repeated-measures ANOVA within groups and single-factor ANOVA across groups (StatView 4.5, Abacus Concepts, Berkley, CA) as well as with Fisher's test and unpaired one- and two-tailed t-tests when appropriate. The criterion for significance was P < 0.05 for all comparisons.
Hemodynamic indexes were recorded throughout the experiment. These indexes included cardiac power, developed pressure, cardiac output, MV̇o2, and the minimum and maximum of the first derivative of LV pressure (LV dP/dtmin and dP/dtmax, respectively). Of note, all of the group I piglets and one-half of the group II piglets required defibrillation to restart myocardial contraction immediately upon reperfusion. Absolute baseline hemodynamic values (before the start of CPB) are shown in Table 1. The time of the baseline values is represented by time A in Fig. 1. Baseline cardiac functional measurements were similar among all groups. Final hemoglobin values (measured after pyruvate infusion) were similar between group I (5.1 ± 0.8 g/dl) and group II (3.7 ± 0.3 g/dl) and were significantly lower (P < 0.05 by paired t-test) than baseline due to hemodilution from the CPB crystalloid prime. The next functional comparison was made 45 min after reperfusion and is shown as time B in Fig. 1. At this time, both groups I and II had been off of CPB for a similar time but only group I had received pyruvate. Cardiac function (Table 2) was not altered by ischemia and reperfusion within either group at time B (assessed via paired t-test), although LV dP/dtmax trended higher in group I and lower in group II (Table 2). However, LV dP/dtmax was significantly higher in group I compared with group II at time B. With the exception of developed pressure, the remaining functional parameters trended higher in group I at time B, but the comparisons did not reach significance (Table 2).
The importance of the timing of pyruvate infusion was examined by comparing cardiac function immediately before (time B) and at the end of pyruvate infusion (time C) in group II. No functional parameters demonstrated improvement by the later pyruvate infusion in these piglets (Table 2).
MV̇o2 did not change significantly throughout these protocols and did not differ between groups at any time point. In summation, the functional data show that pyruvate can modestly elevate cardiac function after CPB but only if it is initiated during the weaning of circulatory support. The superior cardiac function in group I, which received early pyruvate infusion, was not due to increased MV̇o2.
LC-MS was performed primarily to confirm cardiac glutamate labeling in these experiments. Enrichment was determined by the abundance of glutamate labeled in one (M1, molecular weight: 260) or two carbons (M2, molecular weight: 261) relative to total glutamate, including unlabeled glutamate (M0, molecular weight: 259), and corrected for the natural 13C abundance. The relative distribution of the isotopomers is shown in Table 3. At least 9% of the glutamate pool was labeled with 13C in each experiment, confirming the viability of performing MRS in these samples. The maximum enrichment for a single sample was 46%.
Pyruvate entry into the TCA cycle.
As noted previously, pyruvate can enter the TCA cycle through either PC (anaplerosis) to yield OAA or decarboxylation via PDH to yield acetyl-CoA. [2-13C]pyruvate carboxylase labels glutamate carbons at position C2 and C3 and decarboxylation labels C5. A sample spectrum from this experiment is shown in Fig. 2 with glutamate carbons C2, C3, and C5 highlighted. As expected from the labeling strategy, no 13C was detected at C1 and C4. Having confirmed labeling at C2, C3, and C5, we were able to then calculate the percentage of pyruvate carbons entering the TCA cycle via anaplerosis. These calculations were performed by determining the relative peak areas of C2 and C3 to the total peak areas of all labeled glutamate carbons, where pyruvate carboxylase/(pyruvate carboxylase + pyruvate decarboxylation) = (C2 + C3)/(C2 + C3 + C5).
The validity of this computation was confirmed by performing a simulation in the TCaSIM provided by the Rogers NMR facility at the University of Texas Southwestern. The simulation showed that the relationship between pyruvate carboxylase/PDH and C2/C5 or C3/C5 remained linear over the range of interest.
The percentage of pyruvate carboxylase relative to total pyruvate entering the TCA cycle was 20 ± 4% for group I, 23 ± 7% for group II, and 27 ± 7% for control. These data confirm that PC occurs in the heart during these experiments. Furthermore, this anaplerotic pathway provides a significant proportion of the carbons entering the TCA cycle as pyruvate. However, no significant differences in the relative proportions entering through carboxylation or decarboxylation were detected among the groups.
The present study performed in vivo is the first examination of PC in the intact heart using NMR spectroscopy and isotopomer analyses. PC replenishes OAA and has been previously characterized in perfused rat hearts and in the porcine heart in vivo through the use of 13C substrates and assessment of mass isotopomers by GC-MS (27). Several investigators have also confirmed anaplerotic flux with [13C]acetate, proprionate, or aspartate and NMR analysis in isolated perfused hearts, although PC was not evaluated (8, 23, 27). NMR detection of 13C intermediates lacks the sensitivity of the GC-MS technique but more specifically identifies positional labeling. Glutamate occurs at substantially higher tissue concentrations than TCA cycle intermediates and, through transaminase, exists in a bidirectional flux with α-ketoglutarate. Accordingly, spectra are obtained by 13C NMR, and the glutamate labeling pattern serves as a surrogate for the TCA cycle intermediate α-ketoglutarate. Significant technical challenges exist when performing these labeling experiments in vivo. Foremost, and unlike the situation in the isolated perfused heart, the labeled pool is diluted by endogenous substrate. This dilution limits glutamate pool labeling to 31% maximum, as detected by LC-MS, and markedly decreases the signal-to-noise ratio in the NMR experiment compared with isolated perfused hearts. This technical limitation is obviated in part by increasing the number of NMR acquisitions in a single experiment.
The Developing Heart
This study also offers the first assessment of the anaplerotic contribution to the TCA cycle in the neonatal heart. Several investigators have established the high glycolytic capacity and preference for carbohydrates as the oxidative substrate in the newborn heart (1, 13, 18). In most species, the substrate preference shifts more toward oxidation of free fatty acids within the first weeks after birth (2, 4, 18). The stimulation of this neonatal metabolic transition appears to involve the activation or disinhibition of carnitine palmitoyl-transferase I, the primary regulator of fatty acid oxidation, rather than the direct inhibition of PDH, which is responsible for the final step in the oxidation of glucose or lactate to acetyl-CoA (5). The newborn porcine heart also maintains a high capacity for malate/aspartate and α-glycerophosphate shuttles (30–32). These shuttles oxidize cytosolic NADH and are critical in maintaining the maximal glycolytic flux to PDH. In this highly glycolytic environment, one might expect an enhanced preference for flux through PDH rather than through pyruvate carboxylase. In contrast, our study showed a robust pyruvate contribution to OAA through PC. Surprisingly, our estimates of the relative rates of PC and decarboxylation are very close to those provided by Panchal et al. (27) in substantially older pigs (27). Pyruvate under near normal physiological and substrate conditions in situ provides 4–6% of citrate through anaplerosis in the aerobic adult porcine myocardium. This contribution represents 15% of all pyruvate entering the TCA cycle. The consistency of these data implies that despite enhanced glycolysis and pyruvate flux in the newborn, the capacity for PC does not change with development.
The present data indicate that pyruvate given shortly after reperfusion increases the function of the neonatal heart after CPB and ischemia. In this study, we used supraphysiological doses of pyruvate (∼8 mM) in the coronary artery. Bunger et al. (6) generated dose-contractile response curves by varying pyruvate concentrations in the perfusate for isolated hearts. Maximal inotropic responses occurred at pyruvate concentrations in the perfusate between 5 and 10 mM under both fully aerobic and ischemia-reperfusion conditions (6). Accordingly, studies in mature swine showed no hemodynamic or inotropic responses to pyruvate supplied in modest doses, achieving ∼1 mM elevation in arterial pyruvate (26, 27, 34). In contrast, Ochiai (24) showed a modest increase in hemodynamic indexes in swine after the elevation of arterial pyruvate by >3 mM. Similarly, intracoronary pyruvate infusions at 3 and 6 mM elevated the cardiac index in patients with congestive heart failure (11).
The elevation in contractile function observed during pyruvate infusion in group I is consistent with results of other studies performed in the heart ex vivo and in vivo using similar doses. The operative mechanisms probably vary according to conditions of perfusion and oxygen supply. Pyruvate inhibits free radical formation during ischemia, thereby mitigating oxidative stress caused by their release during reperfusion (20, 22) (24). In the present experiments, we supplied pyruvate starting 10 min after reperfusion in group I and later in group II. Thus, the impact on free radical formation and release should have been minimized. Under aerobic or reperfusion conditions similar to those existing during pyruvate supplementation in our study, pyruvate is know to elevate or maintain cytosolic ATP phosphorylation potential in part through the modulation of the mitochondrial NADH-to-NAD ratio (15, 16). Pyruvate can also enhance carbon substrate delivery to the TCA cycle during aerobic conditions, either through the formation of acetyl-CoA or pyruvate carboxylase. Butyrate similarly enhances both NADH production and acetyl-CoA delivery to the TCA cycle but does not affect anaplerosis. However, butyrate infusion does not alter phosphorylation potential or inotropy in the mature swine heart (24), casting doubt that enhanced NADH production and acetyl-CoA delivery are the primary effectors of cardiac function. As pyruvate increases citrate content during reperfusion (16), anaplerosis through PC may play an important role in reestablishing the TCA cycle intermediate pool, increasing substrate oxidative capacity, and maintaining postischemic contractile function. This mechanism may be more important in the neonatal heart, which retains the high glycolytic capacity during ischemia and propensity for carbon substrate washout during cardioplegia instillation and subsequent reperfusion. The PDH and acyl-CoA synthetase reactions lead to ATP synthesis through the production of two-carbon acetyl-CoA. However, these pathways provide inefficient mechanisms for replenishing TCA cycle intermediates as they do not contribute the basic four-carbon skeleton. In contrast, pyruvate carboxylase forms the basic four-carbon moiety, OAA, thus replenishing the TCA cycle intermediate pool.
The principal objectives of the present study were directed toward analyses of relative PC and pyruvate decarboxylation in the reperfused immature myocardium. Therefore, each experimental group received [13C]pyruvate at supraphysiological doses. Accordingly, the study design did not include groups without labeled pyruvate infusion. We found that PC relative to PDH flux did not differ between groups I and II. Despite this metabolic similarity, the pyruvate infusion in group II did not elevate cardiac function over preinfusion parameters. As we purposefully did not include a group lacking pyruvate infusion for comparison over the same time course, we cannot explicitly state that pyruvate failed to alter cardiac function. Pyruvate may have prevented time-dependent decay in some function parameters. Pilot studies using similar, although slightly different, protocols did show functional stability for at least 1 h after reperfusion. Despite this major limitation in the study design, the potential differences between early and late pyruvate infusion require further examination. Early pyruvate infusion occurred during the transition from CPB (no ventricular loading) to the normally loaded circulation with a substantially higher energy demand. Conceivably, TCA cycle replenishment is necessary to maintain cardiac function during that critical period. The pool may be irreversibly depleted if pyruvate is given later, leading to the poor functional response. The quantity of glutamate enrichment does not necessarily reflect TCA cycle intermediate enrichment, as the α-ketoglutarate-to-glutamate flux does not approach the rate of TCA cycle flux (9). Accordingly, glutamate enrichment determined by LC-MS in these experiments provides only qualitative confirmation of pool labeling and does not offer the quantitative assessment of CAC flux. The higher rate of [13C]glutamate enrichment in control animals implies that ischemia alters overall pyruvate entry into the TCA cycle or affects cycling between α-ketoglutarate. Conceivably, pyruvate more effectively replenishes TCA cycle intermediates during reperfusion if provided during this transition to ventricular loading. Absolute concentrations and 13C enrichments of TCA cycle intermediates were not measured in this study and represent a limitation. These measures are being performed in ongoing and separate studies.
This work was funded by National Heart, Lung, and Blood Institute (NHLBI) Grant R01-HL-60666 (to M. A. Portman), a fellowship award from the American Heart Association (to O. M. Hyyti), NHLBI Training Grant T32-HL-07828 and a Thompson Family Research Grant (to A. K. Olson), and a grant from the McMillen Foundation (to G. A. Cohen).
A portion of this research was performed using the Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the Department of Energy's Office of Biological and Environmental Research located at Pacific Northwest National Laboratory.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 by the American Physiological Society