INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ADMINISTRATION OF 2,5-anhydro-D-mannitol (2,5-AM), a fructose analog, stimulates feeding behavior in rats by its metabolic actions in the liver (28), which generate a signal to eat that is carried by vagal afferents to the brain (25). Similar to fructose, 2,5-AM is metabolized in the liver to the 1,6-bisphosphate form, but unlike fructose, it is not metabolized further. As a result, 2,5-AM has a variety of effects on liver metabolism, including inhibition of glycolysis, glycogenolysis, and gluconeogenesis (23, 24), but its mechanism of action to stimulate food intake is not related to these substrate-level events. Instead, several lines of evidence indicate that the eating response after 2,5-AM treatment is due to the decrease in ATP that results from the trapping of phosphate in its phosphorylated forms (2,5-AM-1-phosphate and 2,5-AM-1,6-bisphosphate). The drop in liver ATP precedes the behavioral response as it should if it were a stimulus, and the time course of the eating response then tracks subsequent rises in ATP levels (14). Injection of phosphate blocks both the decrease in ATP and the increase in food intake (19). Administration to rats of other metabolic inhibitors that also decrease liver ATP content but by different mechanisms than does 2,5-AM also stimulates eating (21). The effect of 2,5-AM on liver ATP content and on feeding is modulated by changes in fatty acid metabolism. Administration of 2,5-AM and methyl palmoxirate, an inhibitor of fatty acid oxidation, produces a synergistic decrease in ATP and also stimulates feeding synergistically (12). In contrast, feeding rats a high-fat/low-carbohydrate (HF/LC) diet that promotes fatty acid oxidation attenuates the decrease in liver ATP and prevents the eating response after 2,5-AM treatment (20).

In addition to providing evidence that changes in liver ATP can control food intake, the modulation of the eating response to 2,5-AM by changes in fatty acid oxidation is also of interest because it may model the kinds of metabolic fuel interactions that have been long thought to play a role in the control of feeding behavior (8). In the present studies, we used isolated hepatocytes to 1) examine the mechanism(s) by which fat metabolism modulates the metabolic actions of 2,5-AM; 2) determine whether the metabolic effects seen in vivo are independent of hormonal and neural influences; and 3) evaluate the use of hepatocytes as a model to understand the hepatic metabolic control of food intake.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. Male Sprague-Dawley CD rats (200-300 g) from Charles River Laboratory (Wilmington, MA) were used in all experiments. They were housed individually in wire cages at 22°C on a 12:12-h light-dark cycle with lights on at 7 AM. They had free access to tap water and to either of two isocaloric diets (17) (custom-made by Dyets, Bethlehem, PA) for at least 2 wk before the experiments. One diet was formulated to be high carbohydrate/low fat (HC/LF; 63% carbohydrate, 13% fat, and 24% protein by calorie) and the other to be HF/LC (13% carbohydrate, 63% fat, and 24% protein by calorie).

Isolation and incubation of hepatocytes. Hepatocytes were isolated from fed rats during the morning hours by collagenase perfusion as described by Seglen (27). Isolated cells were washed twice with and resuspended in Hank's balanced salt solution (HBSS; omitting glucose). Cell viability assessed by trypan blue exclusion was routinely greater than 90%. Cells were then diluted to a density of 8 × 10⁶/ml. The protein content of the cell suspension was determined by the biuret method (15). Incubation was carried out in 25-ml Erlenmeyer flasks in a 37°C shaking water bath. Each flask contained 9 ml of HBSS (made in 2% fatty acid-free bovine albumin) with or without substrates [(in mM) 5 glucose, 1 oleate, or 0.5 palmitate]. When oleate or palmitate was the substrate, it was stirred in HBSS with bovine albumin at 4°C overnight to allow the formation of fatty acid-albumin complex to solubilize the fatty acid. 2,5-AM (Toronto Research Chemicals, North York, Ontario, Canada) was added as a 300-mM stock solution to the final concentrations indicated. Incubation was initiated by adding 1 ml of the cell suspension to the flask that was immediately flushed thoroughly with 5% CO₂ (balanced with O₂), sealed with a rubber cap, and incubated in the water bath for 30 min.

ATP assay. Hepatocytes were harvested from rats eating one of the two diets (n = 3 rats/group). At the end of the 30-min incubation, 1 ml of the incubation mixture was withdrawn and immediately mixed with ice-cold 0.2 ml of 1.8 M HClO₄. After being cooled on ice for 10 min, the mixture was centrifuged at 13,000 g for 5 min. The resultant supernatant was transferred and neutralized with 3.5 M K₂HPO₄ to ~pH 6.7. Centrifugation was repeated; the supernatant was saved and stored at 80°C until adenine nucleotides were analyzed by HPLC (26). Within 24 h after obtaining the extracts, HPLC analysis was carried out using a Rainin HPXL solvent delivery system (Rainin Instrument, Emeryville, CA) equipped with a 250 × 4.6-mm Primesphere 5  C18-HC column and a 30 × 4.6-mm guard column (Phenomenex, Torrance, CA). Peaks were detected by a Jasco UVIDEC-100-III UV spectrophotometer (Japan Spectroscopic), identified, and quantified by comparing with standards.

³¹P-NMR spectroscopy. Isolated hepatocytes from rats eating the HF/LC or HC/LF diets (n = 4 rats/group) were incorporated into agarose beads as described by Bental et al. (1). Approximately 1.5 ml of the beads suspended in 1 ml of DMEM were placed into a glass NMR tube (internal diameter = 10 mm). A filter (100-µm pore size) was inserted into the tube to keep the beads in place. The NMR tube along with a capillary tube containing a methylenediphosphonate standard was placed in the magnet (kept at 36°C) of a DMX-400 Avance Spectrometer (Bruker, Billerica, MA). The cells were superfused with DMEM equilibrated with 5% CO₂-95% O₂ and maintained at 36°C. The medium was introduced through an inflow line at a rate of 2.4 ml/min, and collection of the ³¹P spectra began. After the system was stabilized during the next 60-90 min (baseline period), the inflow line was switched to DMEM with 1 mM 2,5-AM for 70-80 min. Then it was switched back to DMEM without 2,5-AM for hepatocytes to recover (~2 h).

Oxygen consumption. Rate of oxygen consumption in hepatocytes from rats eating either the HF/LC or HC/LF diets (n = 3-4 rats/group) was measured in a chamber system equipped with an oxygen electrode (Instech Laboratories, Plymouth Meetings, PA). Hepatocytes were harvested from different groups of rats for experiments involving incubation of cells with substrate before addition of 2,5-AM (preincubation) and no preincubation. Under conditions of no preincubation, hepatocytes were incubated for 30 min with or without 2,5-AM along with either 5 mM glucose or 1 mM oleate as the substrate in the same way as described above. When preincubation was involved, cells were preincubated with a substrate (5 mM glucose or 0.5 mM palmitate) for 1 h before addition of 2,5-AM, and incubation continued for another 30 min. At the end of incubation, cells were centrifuged and resuspended to a density of 20 × 10⁶/ml in a fresh medium of the same composition used in the preincubation. The chamber containing 0.6 ml of the incubation medium was kept at 37°C by a circulator. After the medium was saturated with air, reaction was initiated with the addition of 80 µl of the cell suspension. The chamber was sealed, and oxygen content in the medium was monitored with a biological oxygen monitor (model 5300, YSI, Yellow Springs, OH). The initial linear decrease in oxygen content was used to calculate the rate of oxygen consumption.

Substrate oxidation. Cells from rats eating either the HF/LC or HC/LF diets (n = 4-8 rats/group) were incubated for 1 h as described above with various amounts of 2,5-AM plus 5 mM [U-¹⁴C]fructose (1 µCi; Sigma, St. Louis, MO) or 0.5 mM [U-¹⁴C]palmitate (1 µCi; DuPont NEN Products, Boston, MA) as the substrate. [¹⁴C]CO₂ generated by the hepatocytes was collected and quantified as described elsewhere (11). Total CO₂ production was determined and used as the measurement of substrate oxidation.

Materials. Unless otherwise indicated, all chemicals were purchased from Sigma.

Statistical analyses. Data were analyzed by a Student's t-test or two-way ANOVA, with repeated measures where applicable.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cellular ATP content. Addition of 2,5-AM to incubated rat hepatocytes decreased ATP content in a dose-dependent manner (Fig. 1). When no substrate (Fig. 1A) was present in the incubation medium, ATP decreased less [F(1,16) = 5.61, P < 0.05] in cells from rats fed the HF/LC diet ("HF/LC cells") than in those from rats fed the HC/LF diet ("HC/LF cells"), particularly at the higher concentrations of 2,5-AM [F(3,16) = 3.78, P < 0.05].

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Cellular ATP contents in isolated hepatocytes after 30-min incubation with 2,5-anhydro-D-mannitol (2,5-AM) and with no substrate (A), 5 mM glucose (B), or 1 mM oleate (C). Values are means ± SE of hepatocytes from 3 rats fed either a high-carbohydrate/low-fat (HC/LF) diet or high-fat/low-carbohydrate (HF/LC) food. *Significant difference between HC/LF and HF/LC cells (P < 0.05).

³¹P-NMR spectroscopy. The changes in hepatocyte phosphate-carrying compounds after the addition of 2,5-AM were followed using ³¹P-NMR spectroscopy. Figure 2 shows a typical ³¹P-NMR spectrum collected from isolated hepatocytes incorporated in agarose beads under baseline conditions. After 1 mM of 2,5-AM was added into the superfusion medium, ATP contents in both HC/LF and HF/LC cells declined from baseline starting at 15 min (P < 0.05; t-test). Minimum levels were reached by 25 min and were sustained until 2,5-AM was removed from the superfusion medium (Fig. 3A). The maximum reduction in ATP was significantly larger in HC/LF cells (~55%) than that in HF/LC cells (~40%) [F(1,36) = 52.4, P < 0.001], and the magnitudes of these reductions were similar to those seen in incubated cells (Fig. 1B).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Typical 31P-NMR spectrum of isolated rat hepatocytes incorporated in agarose beads. PME, phosphomonoesters; MDP, methylene diphosphonate; NTP, triphosphate nucleotides.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Changes in ATP (A), PME (B), and Pi (C) in isolated hepatocytes during perfusion of 2,5-AM determined by 31P-NMR spectroscopy. Hepatocytes were harvested from rats fed either a HC/LF diet or HF/LC food. Values are means ± SE of percent changes from baseline of cells from 4 rats. *Significant difference between HC/LF and HF/LC cells (P < 0.05).

Oxygen consumption. As shown in Fig. 4, A and B, HC/LF cells consumed oxygen at a higher rate than did HF/LC cells when the substrate was either 5 mM glucose [F(1,8) = 10.8, P < 0.02] or 1 mM oleate [F(1,8) = 25.8, P < 0.01]. When 5 mM glucose was the substrate, addition of 2.4 mM 2,5-AM reduced the rate of oxygen consumption significantly [F(1,8) = 10.7, P < 0.02]. However, only the HC/LF cells showed a significant decrease [28%; t(4) = 3.62, P = 0.02], whereas the HF/LC cells did not (14%). When 1 mM oleate was the substrate, 2,5-AM decreased the rate of oxygen consumption in cells from both diet groups [F(1,8) = 120, P < 0.001]. However, oxygen consumption fell more in HC/LF cells (29%) than it did in HF/LC cells [16%; F(1,8) = 16.4, P < 0.01].

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Oxygen consumption in isolated hepatocytes. Oxygen consumption was measured after cells were incubated with 5 mM glucose and 2,5-AM for 30 min (A), incubated with 1 mM oleate and 2,5-AM for 30 min (B), preincubated with 5 mM glucose for 1 h and then with added 2,5-AM for another 30 min (C), and preincubated with 0.5 mM palmitate for 1 h and then with added 2,5-AM for another 30 min (D). Values are means ± SE of hepatocytes from 3 to 4 rats fed either a HC/LF diet or HF/LC food. *Significant difference between incubation with and without 2,5-AM (P < 0.05).

Substrate oxidation. The rate of fructose oxidation (measured as the production of ¹⁴CO₂) was considerably higher [F(1,22) = 186, P < 0.001] and more sensitive to inhibition by 2,5-AM [F(3,22) = 18.7, P < 0.001] in HC/LF as opposed to HF/LC cells (Fig. 5A). Without the addition of 2,5-AM, HC/LF cells oxidized palmitate (Fig. 5B) at a lower rate than did HF/LC cells [t(1,14) = 6.75, P < 0.001]. However, with the addition of 2,5-AM, HC/LF, but not HF/LC, cells increased palmitate oxidation [F(3,32) = 78, P < 0.001]. Only high concentrations of 2,5-AM (>1 mM) inhibited palmitate oxidation in HF/LC cells.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Substrate oxidation in isolated hepatocytes. Cells were incubated with either 5 mM [U-14C]fructose (A) or 0.5 mM [U-14C]palmitate (B) and with 2,5-AM for 1 h. Production of [14C]CO2 was used as a measure of the oxidation of the substrates. Values are means ± SE of hepatocytes from 4 to 8 rats fed either a HC/LF diet or HF/LC food. *Significant difference between HC/LF and HF/LC cells (P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Rats that are fed a HF/LC diet that promotes fatty acid oxidation do not increase food intake and show an attenuated decrease in liver ATP after injection of 2,5-AM compared with those fed a HC/LF diet (20). The purpose of these experiments was to determine the mechanism by which feeding a HF/LC diet makes rats resistant to the effect of 2,5-AM treatment on liver ATP content. The results provide an explanation for this effect of diet on the hepatic metabolic response to 2,5-AM and, by implication, on the behavioral response as well. These findings also demonstrate the utility of in vitro experiments in understanding the metabolic and behavioral responses to 2,5-AM treatment.

2,5-AM decreased hepatocyte ATP content in a dose-dependent manner and, as seen in the liver in vivo (13, 18), this reduction in ATP was greater in HC/LF cells than in HF/LC cells. ³¹P-NMR spectroscopy of hepatocytes showed a time course for changes in ATP after 2,5-AM treatment similar to that seen in whole liver under in vivo conditions (7), with a pronounced drop by 15 min followed by recovery toward normal levels 2-3 h after 2,5-AM was removed from the perfusate. As observed in vivo (18), 2,5-AM increased hepatocyte PME and did so more in HC/LF cells compared with HF/LC cells. Hepatocyte-free P_i decreased when 2,5-AM was added to the perfusate, a result in keeping with observations made on the liver in vivo (18). PME in livers of rats treated with 2,5-AM are composed primarily of phosphorylated 2,5-AM (18). Because 2,5-AM is phosphorylated at the one and then six position but not further metabolized, the decrease in ATP after 2,5-AM treatment is largely the result of phosphate being trapped in the phosphorylated forms of the analog (18). It is likely that the drop in hepatocyte ATP was also due to such phosphate trapping given the inverse changes in ATP and PME over time. It would appear, therefore, that the resistance to 2,5-AM-induced ATP depletion in hepatocytes and livers from rats fed the HF/LC diet results because 2,5-AM is phosphorylated less readily in these cells and tissue than it is in those from rats fed the HC/LF food.

The decrease in 2,5-AM phosphorylation in livers and hepatocytes of rats fed the HF/LC diet appears to result from a shift in the use of metabolic fuels from carbohydrates to fat. Compared with the HC/LF diet, the HF/LC diet used in the present experiments enhances the utilization of fat (20, 22). The findings presented here confirm this shift in the use of fat and carbohydrate fuels by showing that isolated hepatocytes from rats fed the HF/LC diet oxidized palmitate at a higher rate and fructose at a lower rate than do hepatocytes from rats fed the HC/LF diet. Several mechanisms may underlie the reduced phosphorylation of 2,5-AM in rats fed the HF/LC diet. Because oxidation of fatty acids inhibits glycolysis in hepatocytes (2, 10), one would expect that 2,5-AM, an analog of fructose, would be phosphorylated at a lower rate because it shares the first steps of the glycolytic pathway with fructose and glucose. 2,5-AM had a proportionately smaller effect on oxygen consumption in cells from rats fed the HF/LC compared with the HC/LF diet, which is consistent with a decreased flux of glycolytic intermediates through the tricarboxylic acid cycle in animals consuming the HF/LC diet.

In addition to trapping phosphate, 2,5-AM treatment may lower liver ATP production from carbohydrate by inhibiting glycolysis (9, 24). One might expect, given their greater utilization of carbohydrate, that hepatocytes from rats fed the HC/LF diet would be more vulnerable to ATP depletion by 2,5-AM than those from rats eating the HF/LC diet. Inhibition of glycolysis alone should have little effect on ATP production from fatty acids. This may explain the resistance to 2,5-AM-induced ATP depletion by hepatocytes and livers (6) from rats fed the HF/LC diet even though P_i was reduced to the same extent by 2,5-AM as it was in rats fed the HC/LF food. In hepatocytes from rats eating the HC/LF diet, the magnitude of the reduction in the oxidation of fructose was close to the dilution factor of this substrate by 2,5-AM especially at low concentrations, presumably because 2,5-AM and fructose compete for the same transport and metabolic pathways. Apparently, such competition is less a factor when there is relatively low carbohydrate utilization, as is the case of hepatocytes from rats fed the HF/LC diet.

Hepatocytes from rats consuming the HF/LC diet oxidized fatty acids at a higher rate than did cells from animals eating the HC/LF food (see Ref. 3). 2,5-AM inhibited fatty acid oxidation in hepatocytes from rats fed the HF/LC diet, but only at concentrations higher than 1 mM and then only marginally. Such reduced sensitivity to 2,5-AM inhibition is in line with reduced capability for carbohydrate metabolism. What little effect there was on fatty acid oxidation may have resulted because high concentrations of 2,5-AM depleted ATP enough to limit slightly the synthesis of fatty acyl CoA from fatty acids, coenzyme A, and ATP (the first step of fatty acid oxidation) and to reduce the flow of fatty acyl CoAs into mitochondria for oxidation and consequent production of CO₂.

In striking contrast to the response of hepatocytes from rats given the HF/LC diet, 2,5-AM markedly increased palmitate oxidation in cells from rats eating the HC/LF diet. This finding is consistent with and substantiates results of in vivo experiments in rats fed high-carbohydrate food, which show that 2,5-AM treatment enhances fat oxidation at a whole body level, as evidenced by decreased respiratory quotient, and in the liver, as indicated by elevated plasma ketone body concentrations (16, 20). Inhibition of glycogenolysis and glycolysis by 2,5-AM (9, 24) seemed to drive these cells to oxidize more fatty acids. On the other hand, glycolysis in HC/LF cells, even though inhibited, still occurred at higher rates than in HF/LC cells as indicated by relatively higher rates of fructose oxidation. This might supply the precursor (pyruvate) to replenish the pool of intermediates of the tricarboxylic acid cycle to sustain fatty acid oxidation even at high concentrations of 2,5-AM. Switching to fat metabolism provides an alternative way for the liver to produce energy and to recover promptly from the energy deficit caused by 2,5-AM.

Addition of 5 mM glucose into the incubation medium did not alter the effect of 2,5-AM on ATP content in hepatocytes from rats fed the HC/LF diet. Because hepatocytes retain up to 80% of glycogen content during the isolation procedure (5), it is likely that exogenous glucose had little effect on phosphorylation of 2,5-AM as endogenous glucose (from glycogen breakdown) was already available. In contrast, addition of glucose to incubating hepatocytes from rats fed the HF/LC diet largely abolished their resistance to ATP depletion by 2,5-AM (Fig. 1, A vs. B). Oxygen consumption of cells from rats fed the HF/LC diet was lower than that of hepatocytes from rats eating the HC/LF food and did not decrease as much after 2,5-AM treatment. However, after 1-h incubation with 5 mM glucose, hepatocytes from the two diet groups consumed oxygen at the same rate and decreased that rate to the same extent with the addition of 2.4 mM 2,5-AM to the incubation medium (Fig. 4C). It seems that the shift in metabolic fuel use induced in rats by feeding the HF/LC diet is readily reversible by incubating with glucose, even for a short period of time. On the other hand, incubation with a fatty acid as the substrate did not necessarily have the same effect on hepatocytes from rats fed the HC/LF diet, suggesting that adopting a fat fuel economy may take longer than adopting a carbohydrate fuel economy. This may reflect the fact that there is relatively little capacity to store carbohydrate than there is to store fat and that absorbed carbohydrate must be metabolized immediately to prevent hyperglycemia.

The diets used in the present studies affected fuel metabolism in hepatocytes in vitro much like they do in the liver in vivo (6). Similarly, the metabolic responses of hepatocytes in vitro to 2,5-AM paralleled in many respects the responses seen after administration of the analog in vivo. These findings suggest that hepatocytes in vitro are a useful model for understanding the effects of diet and 2,5-AM on liver metabolism under in vivo conditions. It is difficult to link metabolic responses of isolated cells in vitro with behavior. However, the close correspondence between the in vivo and in vitro metabolic responses to 2,5-AM and diet suggests that studies of hepatocytes in vitro might provide insight into the hepatic metabolic stimulus or stimuli that control food intake independently of neural and hormonal factors.