The mechanism(s) of fatty acid uptake by liver cells is not fully understood. We applied new approaches to address long-standing controversies of fatty acid uptake and to distinguish diffusion and protein-based mechanisms. Using HepG2 cells containing an entrapped pH-sensing fluorescence dye, we showed that the addition of oleate (unbound or bound to cyclodextrin) to the external buffer caused a rapid (seconds) and dose-dependent decrease in intracellular pH (pHin), indicating diffusion of fatty acids across the plasma membrane. pHin returned to its initial value with a time course (in min) that paralleled the metabolism of radiolabeled oleate. Preincubation of cells with the inhibitors phloretin or triacsin C had no effect on the rapid pHin drop after the addition of oleate but greatly suppressed pHin recovery. Using radiolabeled oleate, we showed that its esterification was almost completely inhibited by phloretin or triacsin C, supporting the correlation between pHin recovery and metabolism. We then used a dual-fluorescence assay to study the interaction between HepG2 cells and cis-parinaric acid (PA), a naturally fluorescent but slowly metabolized fatty acid. The fluorescence of PA increased rapidly upon its addition to cells, indicating rapid binding to the plasma membrane; pHin decreased rapidly and simultaneously but did not recover within 5 min. Phloretin had no effect on the PA-mediated pHin drop or its slow recovery but decreased the absolute fluorescence of membrane-bound PA. Our results show that natural fatty acids rapidly bind to, and diffuse through, the plasma membrane without hindrance by metabolic inhibitors or by an inhibitor of putative membrane-bound fatty acid transporters.
- intracellular pH
- fatty acid metabolism
hepatocytes take up large amounts of fatty acids from plasma for oxidation and lipid synthesis. It has been estimated that >50% of long-chain dietary fatty acids bound to albumin can dissociate and bind to liver cells in one pass through the liver (32). The mechanism by which fatty acids enter these cells has been a subject of debate.
Early evidence supporting protein-mediated uptake of fatty acids came from in vitro studies showing saturation of uptake (44) and inhibition by antibody of plasma membrane-associated fatty acid binding protein as well as energy dependence and “Na+ linkage” of uptake (15, 43). Fatty acid transporter/CD36 and selected isoforms of fatty acid transport proteins have also been described as fatty acid membrane protein transporters (without elucidation of the mechanism) in the liver (18, 19, 25) and other cell types (5, 7, 14, 28). Many of these putative transporter proteins have also been found to be involved in metabolic processing of fatty acids (7, 8, 13, 34, 35). More recently, the involvement of caveolin in the uptake of oleate has been reported in HepG2 cells and adipocytes (36–38). However, because uptake was measured by cellular incorporation of radioactive oleate over different time periods (20 s–10 min), the results could not distinguish whether caveolin facilitates oleate translocation across the membrane or whether it might facilitate “vectoral transport” by promoting intracellular metabolism of lipids (6).
Other investigators provided evidence that entry of fatty acids into liver cells occurs by diffusion through the plasma membrane (32). It has been shown that the apparent saturation of binding of fatty acid to membranes in the presence of albumin can be explained by partitioning of fatty acid into the lipid bilayer without the involvement of a transport protein (10, 32).
The importance of understanding the mechanism(s) of fatty acid entry into the liver is illustrated by a recent report (33) of acute liver failure in two young men, in which the pathology was attributed to a defect in the transport of long-chain fatty acids across the plasma membrane. Such an interpretation argues that diffusion is ineffective and that protein-mediated transport of fatty acids is crucial for an adequate supply of fatty acids to the liver (4), supporting the broader hypothesis that “cells tightly regulate all aspects of long-chain fatty acids utilization beginning with cellular uptake and retention” (40), issues that are actively debated for adipocytes and other cell types (20, 21, 24).
In this study, we used fluorescence approaches to monitor the binding, transmembrane movement, and metabolism of fatty acids in human hepatoma-derived HepG2 cells, which are considered suitable and convenient models for hepatocytes (39). Our methods allowed real-time measurements of transport without separation procedures, primarily by continuous measurement of intracellular pH (pHin) (20, 22, 23). We also used a dual-fluorescence approach to simultaneously measure the adsorption of fatty acids to the plasma membrane and its transmembrane movement (23) to test whether a commonly used inhibitor of fatty acid uptake, phloretin, inhibits binding and/or transmembrane transport.
The HepG2 cell line was purchased from the American Type Culture Collection. 2′,7′-Bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF)-AM and cis-parinarate (PA) were purchased from Molecular Probes (Eugene, OR). Culture media and plastic wares as well as enzyme-free cell dissociation solution were purchased from Fisher Scientific (Agawan, MA). Oleate was purchased from Nu-chek Prep (Elysian, MN); phloretin was purchased from Sigma. [9,10-3H]oleate and [1-14C]oleate were from NEN (Boston, MA). The purity of the fatty acids was >99% (as stated by the supplier). Triacsin C was purchased from BioMol (Plymouth Meeting, PA). Cyclodextrin complexed with oleate (36 mg methyl β-cyclodextrin and oleate/ml H2O, 4.6 mM oleate) was purchased from Sigma.
Fatty acids and inhibitor stock solutions.
A stock solution of oleate (9 mM) was prepared by dissolving K+-oleate in deionized water. [1-14C]oleate and [9,10-3H]oleate were used as tracers for metabolic products. Stock solutions of PA, triacsin C, and phloretin were prepared in DMSO.
HepG2 cells were grown in DMEM with 10% fetal bovine serum, with medium changed every 2–3 days. After 2 days at confluence, cells were incubated with 1 μM BCECF-AM for 25 min in the maintenance medium and washed three times with PBS at 37°C. Cells were then incubated with an enzyme-free cell-dissociation solution at 37°C for 5 min. This solution was then rapidly removed, and dissociated cells were suspended in Krebs-Ringers solution buffered with MOPS (5 mM glucose). After 30 min of incubation, the cell suspension was used directly. For each measurement, ∼4 × 106 cells (equivalent to 2.2 ± 0.3 mg protein) were used.
For selected experiments, triacsin C (5 μM) was added to cell cultures for 30 min and then washed off with KRB solution [containing (in mM) 118 NaCl, 5 KCl, 1.1 MgSO4, 2.5 CaCl2, 1.1 KH2PO4, 20 MOPS, and 5 glucose (pH 7.4)] before cells were dissociated from the culture dishes for experiments as described above. For experiments with phloretin pretreatment, the inhibitor (100 μM) was added to the cell culture for 30 min. Cells were then treated similarly to those treated with triacsin C except that the phloretin concentration was maintained at 100 μM in the subsequent washing and incubation solutions throughout the experiment. Such a constant presence of phloretin could be important for testing its function as a direct inhibitor of fatty acid transporter proteins.
Cells preloaded with the pH-sensitive fluorescent probe BCECF were suspended in 3.0 ml buffer at pH 7.4 in a stirred polystyrene cuvette (26). The sample temperature was maintained at 37°C by circulating water. Fluorescence measurements were carried out with a SPEX Fluoromax fluorimeter (Edison, NJ). The desired amounts of fatty acids were added to the cells in a cuvette with micropipettes. The changes in fluorescence intensity (sampling time: 2.0 s; band pass: 3 nm) were measured using excitation wavelengths of 439 and 505 nm and an emission wavelength of 535 nm for BCECF and 325-nm excitation and 409-nm emission for PA. For simultaneous measurement of the fluorescence of BCECF and PA, a sampling time of 5.0 s was used. There is a linear relationship between the fluorescence of BCECF and pHin in the range of pH values investigated in this study (26), and changes in the fluorescence of BCECF in HepG2 cells are assumed to quantitatively reflect changes in pHin. No change in the fluorescence of BCECF was found when oleate was added directly to BCECF in cell-free buffer.
As an alternative method of adding oleate to HepG2 cells, we repeated the fluorescence pHin assay by providing cells with oleate bound to β-cyclodextrin molecules. For this experiment, an aliquot of the stock solution containing 20–40 nmol oleate (oleate-β-cyclodextrin = 1:6) was added to HepG2 cells in a cuvette with stirring.
At desired time points, cellular metabolism was stopped by adding 0.1 ml H2SO4 (10 N) to cells. Cellular lipids were extracted by the Folch method (16). The organic fractions from three subsequent extractions were combined and dried under a stream of N2. With this method, 85–90% of the isotope was recovered. The dried lipid extract was then dissolved in CHCl3 and separated by thin-layer chromotography (TLC) using a solvent system of hexane-ethylether-acetic acid (80:20:1). After elution, the TLC plate was visualized in an iodine tank, and the identified lipid fractions were scraped off and quantified by scintillation counting. Unlabeled lipid standards were used as indicators to enhance the visibility.
To determine the rate of fatty acid oxidation, [1-14C]oleate was used as a tracer and incubated with cells in T25 flasks sealed with a rubber septum with attached plastic center wells. A small piece of folded filter paper wet with 1 N NaOH was placed in the center well. At the end of the incubation, 0.5 ml of 9 N H2SO4 were injected into the reaction mixture to drive the CO2 into the gas phase, which was absorbed to the filter paper. After equilibrium at room temperature for 24 h, the center well was removed for scintillation counting. In our experimental time course, ∼0.1% of oleate (50 nmol added) was oxidized into CO2. Therefore, the net quantity of oleate being oxidized was considered to be negligible compared with the total amount added to the incubations.
Data are expressed as means ± SE. Results were analyzed using one-way ANOVA and Duncan's multiple-comparison test. Differences were considered statistically significant when P < 0.05.
Transport of oleate into cells and its metabolic fate.
For this experiment, the pH dye BCECF was first trapped in the cytosol of HepG2 cells. The combined membrane transport steps of adsorption and transmembrane movement were monitored by measuring the changes in BCECF fluorescence. The drop in the BCECF fluorescence intensity reflects a decrease in pHin (26), which occurred immediately after the addition of oleate and was complete within 20 ± 10 s (Fig. 1A).
The experiment was repeated for a range of concentrations of oleate, with each addition of oleate made to a fresh aliquot of cells. BCECF fluorescence decreased rapidly in each case with a maximal drop that was nearly linear with the oleate concentration (Fig. 1B). The addition of oleate complexed with β-cyclodextrin (methods) also produced dose-dependent decreases in pHin with the same kinetics as with oleate alone (not shown). The latter experiment lowered the concentration of unbound oleate in the aqueous phase and prevented aggregation of unbound fatty acids in the buffer. Cyclodextrin can release all of its bound oleic acid to model membranes within 0.5 s (K. B. Fontanini and J. A. Hamilton, unpublished observations), in contrast with albumin, which releases a small fraction of its fatty acid load over a longer time course under typical conditions used in cell assays.
Immediately after the rapid drop in pHin (Fig. 1A), pHin began to return over a slower time course to its basal value. This recovery of pHin could be a result of metabolic consumption of the added fatty acids and/or a proton leak caused by various possible mechanisms. To distinguish these possibilities, we followed the metabolic fate (esterification and oxidation) of [9,10-3H]oleate at various time points after its addition to cells (see methods), and we monitored pHin after the inhibition of activation of oleate. Oleate was incorporated into lipid esters on a time scale of minutes, with negligible partition into the β-oxidation pathway. Within 2 min, oleate was esterified into diglycerides, triglycerides, and, to a lesser extent, into phospholipids and cholesteryl esters (Fig. 1C). The amount of oleate recovered in diglycerides reached its maximum in 4 min and then gradually decreased. The amounts of oleate esterified into cholesteryl ester and phospholipids were relatively small and did not change significantly with incubation time (Fig. 1C). On the other hand, esterification of oleate into triglycerides increased rapidly with time. By 10 min, >80% of the added oleate was recovered in the triglyceride fraction. In the same time period, BCECF fluorescence returned almost to its initial level (Fig. 1A), implying a close association between the metabolic processing of added oleate and the recovery of pHin, as we (26) have previously reported in adipocytes. Similar results were found within a broad range of oleate concentrations (data not shown). The dose-dependent distribution of metabolic end products after 10 min (Fig. 1D) indicates a proportional increase in triglyceride synthesis with increased oleate concentrations, consistent with previous studies (26, 42, 45).
Effects of inhibitors on uptake of oleate.
The relationship among intracellular metabolism of oleate and membrane transport and mechanisms for recovery of pHin were investigated further by preincubating cells with metabolic inhibitors. The first inhibitor we used was triacsin C, a potent inhibitor of acyl CoA synthetase (29), which prevents the activation of fatty acids (formation of acyl CoA esters) required for both β-oxidation and esterification. If the pHin recovery is caused mainly by metabolism but not by proton leak, pHin should recover slower in the presence of triacsin C. As shown in Fig. 2A, the addition of 75 nmol of oleate (25 μM final concentration) to the triacsin C-treated cells resulted in an instantaneous drop in pHin similar to that observed in cells not pretreated with triacsin C, implying no effect of triacsin C on the transport of oleate across the plasma membrane. As predicted, pHin did not recover within the subsequent 10-min observation period. Analysis of the metabolic end products in cells thus treated using [9,10-3H]oleate showed that pretreatment with triacsin C largely blocked the esterification of oleate, leaving the majority of added oleate as unesterified free fatty acids after a 10-min incubation (Fig. 2B).
We next tested phloretin, a small molecule that has been widely used as an inhibitor of putative plasma membrane-bound fatty acid transporters to block the efflux of fatty acids from cells (14, 30). If phloretin blocks the movement of fatty acids across the plasma membrane, it would be expected to reduce or abolish the changes in pHin mediated by exogenous oleate. For these experiment, cells were preincubated with phloretin (100 μM) for 30 min, and the same concentration of phloretin was present in all the subsequent washing and incubation solutions. The addition of oleate to phloretin-treated cells resulted in an instantaneous drop in pHin, similar to that found when oleate was added to untreated cells. The recovery of pHin, however, was much slower in phloretin-treated cells than in the control (Fig. 2A) but faster than that in triacsin C-treated cells (Fig. 2A). Analysis of the partitioning of radiolabeled oleate shows that esterification of oleate was largely blocked by phloretin compared with the control (although to lesser extent than for triacsin C), with the majority of added oleate recovered as free fatty acid (Fig. 2B).
Uptake of PA.
For fatty acid molecules to be transported into cells (by any mechanism), they must first adsorb to the cell membrane. The instantaneous drop in pHin in HepG2 cells after the addition of oleate in the external medium implies that absorption was very rapid but did not measure it directly. To perform a more direct measurement of adsorption, we studied the fluorescent response when HepG2 cells were exposed to PA, a long-chain fatty acid with multiple conjugated double bonds that is metabolized very slowly (17, 47). PA dissolved in aqueous buffer does not fluoresce, but in a hydrophobic environment it fluoresces intensely. Therefore, PA serves as an indicator for the adsorption of fatty acids to cell membranes without complications from metabolic events. A combination of the measurement of BCECF and PA fluorescence, which can be done simultaneously in the same cell suspension, monitors both adsorption of fatty acid molecules to the cellular membrane and their translocation through the membrane (23).
Dual-fluorescence measurements of the binding of PA to the membrane and changes in pHin are illustrated in Fig. 3. The addition of PA to the external buffer of HepG2 cells correlated with an instantaneous increase in PA fluorescence intensity (Fig. 3A) and a simultaneous decrease in BCECF fluorescence (Fig. 3B). The changes in fluorescence remained stable over time (>10 min), consistent with the low metabolic activity of this fatty acid (17, 47). Each 20 nmol addition of PA to the same cell preparation resulted in an increase in PA fluorescence and a decrease in BCECF fluorescence. However, the magnitude of the changes in PA fluorescence decreased slightly as the total amount of exogenous PA increased. This suggests that either the saturation of plasma membrane, leading to reduced PA adsorption (decreased partitioning into the membrane), or a nonlinear relationship between PA fluorescence and its concentration in the membrane.
To address this issue further, we monitored the PA fluorescence over a wide range of concentrations. Fluorescence traces as a function of time after the addition of three selected PA doses are shown in Fig. 4A, with each trace representing a separate addition of PA to cells. Most of the increase in fluorescence occurred very rapidly, within the time resolution of our measurement (5 s). This rapid increase was followed by a slower change occurring over 10–30 s, and equilibrium was reached more slowly with higher PA concentrations (areas between the double-dashed lines, Fig. 4A). When plotted against the concentration of added PA (Fig. 4B), the maximal fluorescence is nonlinear, with a steeper slope for added PA below 75 nmol.
The simultaneous measurement of intracellular BCECF fluorescence in each sample (Fig. 4C) revealed a rapid decrease in pHin, as seen for oleate (Fig. 1A). Over a wide range of concentrations of PA, the maximal decrease in fluorescence was nearly linear with increasing amounts of added PA (Fig. 4D), as for oleate (Fig. 1B). Compared with oleate, the most notable difference was that PA caused a persistent depression of pHin, with no return to the initial values within the 5-min period monitored (Fig. 4C). The large increase in PA fluorescence upon the addition to the cell suspension demonstrated insertion of the acyl chain of PA into a hydrophobic (lipid) environment.
Effects of phloretin on uptake of PA.
A previous study (17) reported that phloretin markedly decreases PA fluorescence when added to adipocytes and attributed the findings to the inhibition of transport of PA across the plasma membranes. We attempted to determine whether such effects occur with HepG2 cells and to elucidate additional details of the entry of fatty acids into these cells. First, we varied the concentration of PA, keeping the concentration of phloretin fixed at 100 μM, a commonly used concentration in previous tests of transport mechanisms (2). With phloretin, there was a dose-dependent increase in PA fluorescence with increasing PA concentration, as in the case of control cells without added phloretin. However, phloretin greatly reduced the maximal fluorescence at each PA concentration (data not shown), as reported previously (17).
To delineate the possible effects of phloretin on the binding and transmembrane movement of PA, we used a dual-fluorescence approach to monitor changes in the fluorescence of PA and BCECF simultaneously in the same cell preparation. In these experiments, the PA concentration was fixed at 40 μM and the concentration of phloretin was varied (Fig. 5A). PA fluorescence was reduced by 50% by 50 μM phloretin and by ∼90% when phloretin was increased to 200 μM. These results closely replicated the reported findings with adipocytes (17). However, the changes in pHin detected by BCECF were unaffected by phloretin in a large concentration range (0–300 μM; Fig. 5B). Thus the decrease in PA fluorescence was not correlated with decreased adsorption or the transmembrane movement of PA. To test the hypothesis that effects of phloretin on PA fluorescence were due to direct interactions leading to quenching, we added PA (50 μM) to the cells first and then added phloretin (200 μM) to the same cell suspension. As shown in Fig. 6, after the expected rapid and robust increase in fluorescence when PA was added to the cells, phloretin caused an instant and large decrease in PA fluorescence. Subsequent delivery of additional PA only caused small increase in fluorescence, similar to the traces of high phloretin concentrations shown in Fig. 5A.
To enhance our understanding of the mechanisms by which fatty acids enter HepG2 cells and to differentiate contributions of transport in the plasma membrane and intracellular metabolism, we studied the uptake of a common dietary fatty acid (oleate) and a natural but uncommon long-chain fluorescent fatty acid that is very poorly metabolized (PA) (47). Both fatty acids caused an immediate and rapid decrease in pHin after their addition to HepG2 cells (Figs. 1–5). The kinetics of the pHin decrease (and recovery) was the same when oleate was added in the unbound form or complexed to β-cyclodextrin. The latter protocol reduced the concentration of unbound fatty acid to very low concentrations and addresses the criticism that micromolar concentrations of unbound fatty acids alter the mechanism of diffusion through the lipid bilayer (11) or mask the contribution of protein-mediated fatty acid transport across the plasma membrane (1, 41).
The maximal decrease in pHin occurred within 20 ± 10 s, after which pHin gradually returned to its initial level in cells treated with oleate but essentially remained unchanged in cells treated with PA. The rapid pH drop is similar to that for adipocytes (26); pancreatic β-cells showed slower acidification even though the flip-flop mechanism was validated by the use of analogs (22). Studies with multiple fluorescent probes will be undertaken to determine whether the adsorption of fatty acid into the extracellular leaflet of the plasma membrane or the transmembrane movement of FA is rate limiting.
The recovery of pHin could be a result of metabolic consumption of intracellular fatty acids and/or a proton leak caused by various mechanisms (48). However, because the decrease in unesterified oleate (Fig. 1C) paralleled the recovery of pHin (Fig. 1A) and correlated with a simultaneous increase in the esterification of oleate (Fig. 1C), our results support the hypothesis that pHin recovery reflects the metabolic disposition of added oleate, as also found in adipocytes (26). The lack of pHin recovery with PA (Fig. 3) or in experiments in which oleate metabolism was inhibited (Fig. 2) also supports this hypothesis. Changes in pHin in adipocytes and β-cells caused by poorly metabolized analogs of fatty acids also recover very slowly (9, 27). Furthermore, neither model membranes nor cell membranes (adipocytes and HepG2 cells) are leaky to H+ in the presence of long-chain fatty acids, as demonstrated by the sustained drop in pHin when the metabolism of oleate was inhibited by triacsin C (Fig. 2).
The apparent inhibition of uptake of fatty acids by phloretin has often been interpreted as support of protein-mediated fatty acid transport across the plasma membrane (2, 17, 42, 43), even though phloretin has a broad range of effects, including inhibition of glucose and anion transport as well as changes in membrane leakiness (3, 12, 31, 46). For example, a study (17) with adipocytes concluded that phloretin blocked uptake of PA and considered that as strong evidence against the diffusion mechanism. Phloretin is almost routinely added to assays of fatty acid uptake that require separation procedures with the stated purpose of preventing loss of radiolabeled fatty acids from cells during washing and cell separation (2). In our assay without separation procedures, phloretin had no effect on the initial pHin drop after the addition of oleate; however, it had a marked inhibitory effect on the subsequent recovery of pHin and the esterification of added oleate. Phloretin, therefore, did not interfere with the diffusion of oleate across the plasma membrane, and our experiments revealed no evidence for protein involvement in this transport step.
Experiments with PA (Figs. 3–6) tested our flip-flop hypothesis and further clarified the effects of phloretin. By monitoring the changes in fluorescence of PA and BCECF simultaneously, we observed both binding of fatty acid to the plasma membrane and pHin changes in real time. Upon the addition of PA to HepG2 cells, its fluorescence increased immediately and rapidly to a stable value, demonstrating incorporation of the hydrocarbon chain of PA into a hydrophobic environment. The slower component at higher PA concentrations most likely reflects the formation of aggregates that must dissolve before PA binds to the membrane. There is no evidence that PA fluorescence senses the translocation of PA across a bilayer or that the slow component might reflect flip-flop, as instantaneous drops in pHin were also seen at low PA concentrations when only a fast component in PA fluorescence was detected. The nonlinear changes in fluorescence of PA (Fig. 4B), almost reaching a plateau at the highest concentrations (150–200 μM), could reflect decreased partitioning of PA into the plasma membrane or self-quenching of the PA fluorescence. The latter is more plausible, because each addition of PA produced a simultaneous and rapid decrease in BCECF fluorescence, and the dose dependency of the pHin drop was almost linear within a broad range of PA concentrations.
Phloretin had a profound effect on PA fluorescence, appearing to inhibit its binding to HepG2 cells, as previous reported in adipocytes (17). However, we proved that this is not the case because phloretin did not prevent the diffusion of PA across the plasma membrane. It is likely that binding of phloretin to the cell membrane altered the fluorescence of PA in the membrane environment and quenched the PA fluorescence, as evidenced by the fact that phloretin substantially attenuated the fluorescence of PA that was already bound to the cells (Fig. 6). The same effects of phloretin on PA fluorescence can be replicated in protein-free phospholipid bilayers or by aqueous PA alone (data not shown) and therefore are not likely related to membrane proteins in HepG2 cells. In view of the interference of phloretin with PA fluorescence, its inhibition of oleate metabolism, and its variety of other cellular effects, phloretin is not a good probe for elucidating the mechanisms of fatty acid entry into cells. On the other hand, the results of experiments with PA (with or without phloretin) provide direct evidence of the rapid binding of fatty acids to the plasma membrane and its immediate diffusion across the membrane by the flip-flop mechanism.
In summary, this study monitored the uptake of fatty acids into HepG2 cells in real time without the complications of separation procedures. The fluorescence traces make it possible to separate the events of membrane transport from those of metabolism. Our conclusions regarding the mechanism of fatty acid transport in membranes differ from some other studies that attribute most, if not all, transport to membrane proteins but are in agreement with previous conclusions stating that fatty acids diffuse into liver cells (49, 50). The rapid diffusion of fatty acid into cells ensures that cells will be supplied with fatty acids, but the utilization of fatty acids depends on intracellular trafficking and metabolism. Uptake must be observed in a very short time interval to distinguish membrane transport from the metabolic contributions to uptake, and most studies have not applied such methods. Furthermore, our results indicate that data obtained for fatty acid uptake with inhibitors present must be interpreted with caution, as such molecules can have complex biophysical and metabolic effects.
This work was supported by National Institutes of Health Grants RO1-DK-59261 (to W. Guo) and RO1-HL-67188 (to J. A. Hamilton).
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 © 2006 the American Physiological Society