The Journal of Nutrition Vol. 128 No. 5 May 1998, pp. 810-817

Exercise Training Down-Regulates Hepatic Lipogenic Enzymes in Meal-Fed Rats: Fructose versus Complex-Carbohydrate Diets^1,2,3

Russ Fiebig, Margaret A. Griffiths^{*, 4}, Mitchell T. Gore, David H. Baker^*, Lawrence Oscai, Denise M. Ney, and Li Li Ji⁵

Departments of Kinesiology and Nutritional Sciences, University of Wisconsin-Madison, Madison, WI; ^* Departments of Animal Science and Food and Human Nutrition, University of Illinois at Urbana-Champaign, Urbana, IL; and  Department of Kinesiology, University of Illinois at Chicago, Chicago, IL

	ABSTRACT

Abstract Introduction Methods Results Discussion References

The maximal activity and mRNA abundance of hepatic fatty acid synthase (FAS) and other lipogenic enzymes were investigated in rats meal-fed either a high fructose (F) or a high cornstarch (C) diet. The diet contained 50% F or C (g/100 g), casein (20%), cornstarch (16.13%), corn oil (5%), minerals (5.37%), vitamins (1%) and Solka-floc (2%). Female Sprague-Dawley rats (n = 44) were randomly divided into C or F groups that were meal-fed for 3 h/d; each group was subdivided into exercise-trained (T) and untrained (U) groups. Treadmill training was performed 4 h after the initiation of the meal at 25 m/min, 10% grade for 2 h/d, 5 d/wk, for 10 wk. Rats were killed 9 h after the meal and 27 h after the last training session. F-fed rats had significantly higher activities of all lipogenic enzymes assayed and mRNA abundance of FAS and acetyl-coenzyme A carboxylase (ACC) than C rats (P < 0.05). Concentrations of plasma insulin and glucose and liver pyruvate were not altered by F feeding. Proportions of the fatty acids 18:2 and 20:4 were lower, whereas those of 16:0 and 16:1 were higher, in livers of F than of C rats (P < 0.05). Training decreased FAS activity by 50% (P < 0.05), without affecting FAS mRNA level in C rats; this down-regulation was absent in the F rats. ACC mRNA abundance tended to be lower in CT than in CU rats (P < 0.075). L-Type pyruvate kinase activity was lower in FT than in FU rats (P < 0.05), whereas other lipogenic enzyme activities did not differ between T and U rats of each diet group. We conclude that hepatic lipogenic enzyme induction by high carbohydrate meal feeding may be inhibited by exercise training and that a fructose-rich diet may attenuate this training-induced down-regulation.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

The effect of chronic exercise on reducing body fat and obesity has been well established (Blair 1993). Previous work has focused primarily on the benefit of exercise in promoting fat oxidation during aerobic exercise (Saltin and Astrand 1993). However, there is evidence that exercise training may also down-regulate hepatic lipogenic enzymes, thereby reducing the availability of long-chain fatty acids required for synthesis of triglyceride (Askew et al. 1975, Winder et al. 1975). This antilipogenic effect of training is especially important when the diet is low in fat and high in simple sugars, because a diet rich in monosaccharides such as fructose has been shown to promote lipogenic enzyme induction and obesity (Goodridge 1987, Oscai 1982). The literature concerning down-regulation of lipogenic enzymes by exercise training is controversial because most of the previous studies used free-feeding animal models in which the influences of diet and exercise confounded each other. Furthermore, it is not clear whether the observed decreases of lipogenic enzyme activities were caused by a reduction of the de novo synthesis of enzyme protein or by an inactivation of the existing enzymes.

With the use of a food deprivation/refeeding model, we recently demonstrated that an acute bout of prolonged exercise significantly attenuated hepatic lipogenic enzyme inductions in rats deprived of food for 48 h and refed a high carbohydrate (CHO)⁶ and low fat diet (Griffiths et al. 1996). The most dramatic inhibition was that of fatty acid synthase (FAS, EC 2.3.1.85), the rate-limiting enzyme for hepatic lipogenesis, for which both maximal enzyme activity and mRNA abundance were down-regulated with exercise by as much as 50-70%. These changes were accompanied by decreased plasma insulin and elevated plasma glucagon concentrations in the exercised rats, as well as a decreased hepatic pyruvate content. This finding prompted us to hypothesize that animals involved in chronic endurance training while being fed a CHO-rich diet maintain a down-regulated lipogenic enzyme status. This hypothesis is based on the following three premises. First, endurance training decreases plasma concentration of insulin (Galbo 1983), which is an important hormone in the stimulation of gene expression of hepatic lipogenic enzymes (Paulauskis and Sul 1989). Second, training increases hepatic gluconeogenic capacity, resulting in a lower level of glycolytic intermediates (Coggan et al. 1995). Elevated glycolytic intermediates due to CHO feeding, especially fructose feeding, have been shown to stablize mRNA of FAS and other lipogenic enzymes, promoting increased enzyme synthesis (Vaulont and Kahn 1994). Finally, hepatic lipogenic enzymes are down-regulated at the transcriptional level by polyunsaturated fatty acids (PUFA) (Clarke and Jump 1996). We speculate that a high CHO diet and training may affect liver fatty acid composition and consequently lipogenic enzyme regulation.

Thus, in this study, we investigated the effect of 10 wk of treadmill training on the activities and mRNA abundance of hepatic FAS and other lipogenic enzymes in rats fed two types of high CHO diets, i.e., fructose, which is particularly effective in inducing lipogenic enzymes (Goodridge 1987), versus cornstarch, a complex CHO. To simulate the food deprivation/refeeding effect observed in our early study (Griffiths et al. 1996), we used a meal-feeding regimen in rats, and training sessions were scheduled at a time period in which maximal transcription of the FAS gene was observed (Iritani et al. 1992). Possible mechanisms underlying the interactive effects of diet and training were also explored.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Animals.  Female Sprague-Dawley rats (n = 44) were purchased from Harlan Sprague Dawley, Indianapolis, IN. Female rats were used because male rats reduce food intake and body weight in response to endurance training, which could potentially confound interpretation of training-induced changes of lipogenic enzymes, whereas female rats increase food intake with training without losing body weight (Schaible and Scheuer 1985). All rats were individually housed in temperature-controlled rooms (22°C) with a reverse 12-h dark:light cycle (0700-1900 h, dark; 1900-0700 h, light). During the first 2 wk, all rats received the cornstarch diet and were trained to eat their daily amount of food in 3 h; the 3-h meal feeding per day was continued throughout the study. All rats had free access to water. At the end of wk 2, all rats were acclimated to treadmill running for 1 wk.

Dietary treatment.  Starting the day after their arrival, all rats were started on the meal feeding regimen using a semipurified high CHO diet. Fifty percent of the diet (by weight) consisted of a basal mix, which contained the required amounts of amino acids, fats, carbohydrates, vitamins, and minerals for the rats to maintain normal metabolic functions; the other 50% of the diet was cornstarch (Table 1). The rats were given 30 g of the diet for 3 h/d; at the end of the 3 h, the food was removed from the cages until the next morning. The daily feeding was started 1 h into the dark (active) cycle. Body weights of all rats were measured weekly throughout the study, before the daily feeding. The weight was then used to determine the amount of food given to each individual rat each day. All rats were able to adjust to this meal feeding regimen and had adequate food intakes (7-8 g/100 g body weight) after 2 wk. Rats maintained normal growth throughout the experimental period.

After 3 wk of this dietary regimen, the rats were divided into two dietary groups; one continued on the aforementioned cornstarch-based diet for the rest of the experiment (C), whereas the other was meal-fed a high fructose (F) diet (Table 1). The C and F diets were isocaloric and isonitrogenous. Each rat received 8 g food/(100 g body weight·d). Each dietary group of rats was further divided into two groups, exercise trained (CT and FT, n = 11) and untrained (CU and FU, n = 11). The division into groups was done by rank ordering their body weights from the highest to the lowest and then randomly assigning a number from 1 to 4 to each rat. There was no difference in the body weights among the four resulting groups (Table 2).

Exercise training.  After adjusting to the semipurified diet and meal feeding routine for 2 wk, all rats were acclimated to treadmill running on a Quinton (Seattle, WA) rodent treadmill for 1 wk. The rats started running at 15.5 m/min, 0% grade for 10 min/d, 5 d/wk during the wk 1; the treadmill speed and the length of running time were increased daily so that by the end of wk 4, they were able to run at 20 m/min and 0% grade for 120 min. During the following weeks, the treadmill speed and grade were gradually increased every 2 wk so that beginning wk 9 all rats were running at 25.0 m/min and 10% grade for 2 h/d, 5 d/wk. This exercise intensity was maintained through the end of the 10-wk training period.

To exert maximal inhibitory effect of exercise on hepatic lipogenic enzyme induction, the exercise training sessions were scheduled during 4-6 h after the beginning of the meal; the rationale for this comes from previous studies showing that FAS mRNA transcription rate was greatest at 4 h after refeeding, and FAS mRNA concentration reached its peak at 8 h (Griffiths et al. 1996, Iritani et al. 1992)

CT and FT rats trained for 2-3 d followed by 1 d of rest, then trained for 2-3 d followed by another rest day, throughout the training period. On the rest days, the CU and FU rats ran on the treadmill at 15-17 m/min for 10 min. This allowed the control animals to experience the sounds and stress of the treadmill without being trained.

Tissue preparation.  At the end of the 10-wk exercise training program, all rats were killed by decapitation in the resting state ~27 h after the end of their last training or handling session. On the day of killing, 2-3 rats from each treatment group were given their usual C or F meal diet, beginning in a staggered fashion, for 3 h, and killed at 15-min intervals ~9 h after the start of the meal. Previous work indicated that the activities of hepatic lipogenic enzymes do not vary significantly over the course of a day during meal-feeding regimens (Armstrong et al. 1976).

All rats were killed by decapitation as approved by the Institutional Review Board of the University of Illinois at Urbana-Champaign. The mixed arteriovenous blood was collected in a conical tube that contained 100 µL heparin (6 g/L) and was chilled on ice. The blood sample was centrifuged at 500 × g for 15 min and blood plasma was stored at 80°C until hormone and plasma glucose assays were performed. After blood collection, the abdominal cavity was opened and the liver quickly excised, weighed and frozen in liquid N₂. The liver samples were either stored in liquid N₂ or at 80°C until processing and assay. At a later time, one lobe of the liver was thawed in a phosphate buffer (50 mmol/L potassium phosphate, 1 mmol/L EDTA and 1 mmol/L dithiothreitol, pH 8.0) and homogenized at 0-4°C with a Potter-Elvehjm (Thomas, Swedesboro, NJ) homogenizer. The homogenate was centrifuged with a Beckman (Palo Alto, CA) model L8-55M ultracentrifuge at 105,000 × g for 60 min to obtain liver cytosol, which was stored at 80°C until lipogenic enzyme assays were performed. The other portion of the frozen liver tissue was used for measurement of liver metabolites and RNA. After the liver was excised, the deep portion of the vastus lateralis muscle was removed and frozen in liquid N₂. At a later time, the frozen muscle was weighed and then placed into 0.1 mol/L phosphate buffer (pH 7.4) at 0-4°C (wt/vol, 1 g:10 mL). These samples were minced and homogenized at 0-4°C with a motor-driven Potter-Elvehjem homogenizer. The homogenate was frozen at 80°C and used for the assay of citrate synthase (CS, EC 4.1.3.7).

Enzyme activity.  Maximal activities of hepatic lipogenic enzyme FAS, L-type pyruvate kinase (L-PK, EC 2.7.1.40), ATP-citrate lyase (ACL, EC 4.1.3.8), malic enzyme (ME, EC 1.1.1.40), and glucose 6-phosphate dehydrogenase (G6PDH, EC 1.1.1.8) were measured spectrophotometrically in liver cytosol according to previously described methods (Griffiths et al. 1993). Acetyl coenzyme-A carboxylase (ACC, EC 6.4.1.2) activity was determined using ¹⁴C-labeled CO₂ incorporation into acetyl-CoA according to Inoue and Lowenstein (1972). CS activity was determined according to Shepherd and Garland (1969). The concentration of plasma glucose was determined using Sigma Diagnostics Glucose Kit 510-DA (Sigma Chemical, St. Louis, MO). The amount of pyruvate in the liver was measured spectrophotometrically according to Bucher et al. (1965). The concentration of plasma insulin was determined with RIA by using assay kits (TKIN, Diagnostic Products, Los Angeles, CA). Protein content was determined by the Bradford method (Bradford 1976) with bovine serum albumin as the standard.

Northern blot and slot blot.  Total RNA was isolated from frozen livers by the method of Chomczynski and Sacchi (1987). Nucleic acid concentration was estimated spectrophotometrically by absorbance at 260 nm. RNA quality was determined by gel electrophoresis and ethidium bromide staining. To verify the specificity of the cDNA probe, 10-15 mg total RNA per lane was loaded onto an 8 g/L agarose gel containing formaldehyde. The gel was then soaked in 10X standard stock solution (SSC; 0.15 mol/L NaCl, 15 mmol/L sodium citrate) to remove residual formaldehyde and in 0.05 mol/L NaOH (20 min) to facilitate transfer of the RNA to the nylon membrane. The RNA was transferred by capillary action with 20X SSC overnight onto a nylon filter and covalently bound by UV crosslinking.

The cDNA probes for FAS, ACC and 18S rRNA were labeled using random primer extension (Feinberg and Vogelstein 1983) by a random primer labeling kit (Megaprime, Amersham, Arlington Heights, IL ) using [³²P] dCTP (Amersham). Northern and slot blots were prehybridized at 42°C for 3 h in a solution consisting of 500 g/L formamide, 5X Denhardt's solution (1 g/L polyvinylpyrrolidone, 1 g/L bovine serum albumin and 1 g/L Ficoll), 0.75 mol/L NaCl, 0.05 mol/L NaH₂PO₄, 5 mmol/L EDTA and 1 g/L SDS. Radiolabeled probes were added at a level of 33 MBq/L hybridization solution and allowed to hybridize overnight. The stringency washes consisted of two 20-min washes with 1X SSC:5 g/L SDS and two 20-min washes with 0.5X SSC:5 g/L SDS. All washes were performed at 42°C. Filters were wrapped in plastic while still damp and exposed to film at 80°C. After autoradiography, the probe was removed from filters with a solution of 50% formamide, 2X SSPE at 65°C for 60 min. After verification of probe removal, the filters were then rehybridized to another probe and exposed to film for autoradiography.

For quantification of relative RNA abundance, slot blots were performed on the samples. Preliminary slot blots were performed over a range of RNA amounts to determine the levels at which linear autoradiographic responses would result. All samples were blotted onto the same filter, immobilized by UV crosslinking and allowed to hybridize to the labeled probe. Slot blot data were normalized to 18S rRNA signal intensity to control blotting efficiency. Quantification of the ACC, FAS and 18S rRNA signals was achieved by use of a scanning densitometer (Biorad model GS-670, BioRad, Richmond, CA). ACC and FAS values were expressed relative to the density of their respective 18S rRNA.

Hepatic fatty acid composition.  Hepatic fatty acid composition was analyzed according to Monsma and Ney (1993). Briefly, ~1 g of frozen liver tissue was homogenized in chloroform/methanol (2:1, v/v) using a Polytron (Tekmar, Cincinnati, OH) homogenizer (setting #6) for 40 s. The homogenate was placed in a centrifuge tube, flushed with N₂, capped and stored at 4°C overnight. The homogenate was filtered through filter paper (Whatman #40, Maidstone, UK) into screw-cap culture tubes to remove cell debris. To separate the organic and aqueous phases, 4 mL KCl (7.4 g/L) was added, mixed and centrifuged at 500 × g for 15 min. The aqueous phase was aspirated off and the organic phase was washed three times by adding 4 mL of chloroform/methanol/H₂O (3:48:47, v/v/v/), mixing and centrifuging at 500 × g for 10 min and aspirating the aqueous phase. The organic phase was dried under a gentle stream of N₂. Immediately, 5 mL of chloroform/methanol (2:1) was added to the lipid extract, flushed with N₂ and stored at 20°C overnight.

Lipids were dried down under N₂ and 400 µL of 0.5 mol/L NaOH in methanol with 50 µg C17 added as an internal standard. Samples were flushed with N₂, capped, mixed using a vortex mixer, wrapped in aluminum foil and boiled for 10 min. After cooled to room temperature, 500 µL BF₃/methanol (wt/wt) reagent (Supelco, Bellefonte, PA) was added and boiled for another 10 min. The samples were washed in 1.5 mL of pentane and 3 mL methanol/H₂O (6:39), mixed by hand for 3 min and centrifuged at 250 × g for 10 min at 4°C. The lower phase was discarded and the wash was repeated three times. After the final wash, the lower phase was discarded and sodium sulfate was added in excess to adsorb any residual H₂O. Samples were flushed with N₂ and stored at 20°C until analysis.

Fatty acid methyl esters (FAME) were analyzed with a Varian 3400 gas chromatograph equipped with a flame-ionization detector, a Varian 1093 SPI temperature programmable injector (Palo Alto, CA), a Supelcowax 10 fused silica column (30 m × 0.32 mm i.d. 0.25 µm film, Supelco), and an In-Board Data Handling System (IBDH, Sugarland, TX). The injector temperature was programmed from 50 to 250°C at a rate of 100°C/min and held for 15 min during the analysis. The initial column temperature was held for 2 min at 50°C, programmed to 160°C at a rate of 30°C/min, then programmed to 190°C at a rate of 3°C/min and finally programmed to 227°C at a rate of 4°C/min and held for 15 min for a total analysis time of 40 min. The carrier and make-up gasses were helium at 2 and 30 cm³/min, respectively. The detector temperature was 300°C. Fatty acids were identified by comparing the retention times with those of known standards (Nu-Chek-Prep, Elysian, MN) and expressed as the weight percent distribution of FAME.

Statistics.  Two-way ANOVA was used to determine significant differences (P < 0.05) in the means of the various treatment groups. The main effects in the 2 × 2 design were diet (C, F) and exercise training (T, U). After an overall F was found to be significant, a post hoc least significant difference (LSD) test (SYSTAT, Evanston, IL) was used to evaluate the significance of differences between treatment groups.

	RESULTS

Abstract Introduction Methods Results Discussion References

Body weight, liver weight and food intake.  There were no significant differences in body weights of the treatment groups either at the beginning of the training period, or on the day of killing before and after the meal (Table 2). The mean food intake was not different among treatment groups during the 10 wk of training. Liver weights, expressed both as absolute values and relative to body weight, were significantly greater in the F-fed rats than in the C-fed rats (P < 0.05; Table 2). However, there were no differences in liver weights between CT and CU or FT and FU.

Concentrations of plasma insulin, glucose and liver pyruvate.  Plasma insulin concentration did not differ significantly between trained and untrained rats or between C- and F-fed rats (Table 2). Plasma glucose and liver pyruvate concentrations also were not affected by diet or training.

Activities of lipogenic enzymes.  The activity of FAS was more than doubled in F-fed vs. C-fed untrained rats (P < 0.05; Fig. 1). Further, FAS activity was >50% lower in the CT versus CU rats (P < 0.05), but did not differ in FT and FU rats. Thus, FT rats had more than twofold higher FAS activity than CT rats (P < 0.05). ACC activity was significantly higher in F-fed than in C-fed rats (P < 0.05, main effect; Fig. 1). Training had no significant effect on ACC activity in either C or F rats.

View larger version (22K): [in this window] [in a new window]   Fig 1. Maximal activity of hepatic fatty acid synthase [FAS, nmol NADPH/( min· mg protein)] and acetyl-CoA carboxylase [ACC, nmol malonyl CoA/(min·mg protein)] in trained (T) and untrained (U) rats meal-fed a high cornstarch (C) or high fructose (F) diet. Each bar represents the mean ± SEM, n = 11. Bars for an enzyme not sharing a letter are significantly different, P < 0.05.

The responses of other hepatic lipogenic enzymes L-PK, ACL, ME and G6PDH to diet and exercise are shown in Table 3. Diet significantly affected L-PK, because FU and FT rats had 82 and 67% higher L-PK activities than CU and CT rats, respectively (P < 0.05). L-PK activity of FT rats was 18% lower than that of FU rats (P < 0.05), but CT and CU rats did not differ significantly. The activity of ACL was more than doubled, whereas ME activity was two- to threefold higher in F-fed rats than in C-fed rats, regardless of training status (P < 0.01). However, there was no effect of exercise training on either of the enzymes. G6PDH activity was also doubled as a result of fructose feeding (P < 0.01), but no training effect was observed.

To verify that the training regimen was effective, we measured CS activity in the deep vastus lateralis muscle as a mitochondrial marker for oxidative potential. CS activity was 70% greater (P < 0.01) in the trained than in the untrained rats, regardless of diet (Table 3).

mRNA abundance.  Northern analysis was performed on hepatic total cellular RNA. For FAS, major transcripts were detected corresponding to ~9.0 and 8.6 kb as we previously reported (Griffiths et al. 1996) (Fig. 2a). A Northern blot for ACC mRNA is shown in Fig. 2b. A major transcript of 9.6 kb was found, which was consistent with previous findings (Bai et al. 1986). Meal-feeding a high fructose diet resulted in significantly higher levels of FAS mRNA compared with cornstarch feeding, with F-fed rats having a fivefold higher FAS mRNA abundance than C-fed rats (P < 0.01; Fig. 3). Relative FAS mRNA abundance was not affected by exercise training in either F or C groups. ACC mRNA abundance also was significantly affected by diet (P < 0.01; Fig. 3). FU rats had 33% higher (P < 0.05) ACC mRNA levels than CU rats, whereas FT rats had almost twofold higher ACC mRNA levels than CT rats (P < 0.01). A diet × training interaction was found for ACC mRNA levels (P < 0.05), with the post-hoc test revealing a trend for lower ACC mRNA levels in CT than in CU rats (P < 0.075).

View larger version (102K): [in this window] [in a new window]   View larger version (103K): [in this window] [in a new window]   Fig 2. Typical Northern blots of hepatic fatty acid synthase (FAS; panel a) and acetyl-CoA carboxylase (ACC; panel b) mRNA isolated from trained (T) and untrained (U) rats meal-fed a high cornstarch (C) or high fructose (F) diet. 18S ribosomal RNA was used as a reference to normalize blotting efficiency of FAS and ACC mRNA. FU, fructose untrained; FT, fructose trained; CU, cornstarch untrained; and CT, cornstarch trained.

View larger version (25K): [in this window] [in a new window]   Fig 3. Relative abundance of hepatic fatty acid synthase (FAS) and acetyl-CoA carboxylase (ACC) mRNA in trained (T) and untrained (U) rats meal-fed a high cornstarch (C) or high fructose (F) diet. mRNA abundance in CU rats is defined as 100%. Each bar represents the mean ± SEM, n = 11. Bars for an enzyme not sharing a letter are significantly different, P < 0.05. For ACC, there was a significant (P < 0.05) interaction between diet and training.

Hepatic fatty acid profiles.  Fructose feeding resulted in significantly higher liver proportions of 16:0, 16:1 and 18:1 fatty acids, whereas those of 18:0, 18:2 and 20:4 were significantly lower (Table 4). Thus FU rats had 22, 125 and 83% higher liver levels of 16:0, 16:1 and 18:1, respectively, than CU rats (P < 0.05), expressed as a percentage of total FAME. In contrast, FU rats had 33, 38 and 40% lower proportions of 18:0, 18:2 and 20:4 in the liver, respectively, than CU rats (P < 0.05). Exercise training did not significantly alter the fatty acid profile in liver in either of the diet groups.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

The main finding of this study was that endurance training significantly down-regulated hepatic FAS activity in rats meal-fed a high cornstarch diet and L-PK activity in rats meal-fed a high fructose diet. These diet-specific training effects on lipogenic enzyme activities extended our earlier work, which showed a prominent inhibition of these enzymes by an acute bout of prolonged exercise in food-deprived and refed rats (Griffiths et al. 1996). Together, our studies have provided strong evidence that physical exercise is a powerful means to reduce the lipogenic potential in liver of rats fed a CHO-rich diet.

In comparison with previous training studies by other investigators and by us (Askew et al. 1975, Barakat et al. 1987, Griffiths et al. 1993, Winder et al. 1975), this study had a unique experimental design, which provided us with more definitive information about the effects of diet and exercise on lipogenic enzymes. First, rats were meal-fed instead of having free access to food. This dietary regimen permitted a better control of the various hormonal and metabolic factors known to influence lipogenic enzymes. Second, a purified diet instead of a commercial diet was used so that the effect of fructose could be compared with that of complex CHO in the form of cornstarch. Third, exercise training was conducted 4-6 h after a meal when dietary induction of lipogenic enzymes is in its most active phase. Furthermore, this study used female rats in contrast to male rats used in most previous studies. Because female rats maintained their food intake and body weight throughout the training period, potential confounding influences of energy and nutritional intake on lipogenic enzyme induction were ruled out.

This study indicated that exercise training can elicit a dramatic suppression of hepatic FAS activity in rats meal-fed a high complex CHO diet. FAS activity was decreased by >50% in CT versus CU rats (Fig. 1). This finding was in agreement with an earlier report by Askew et al. (1975) in which trained male rats fed a nonpurified diet displayed an ~30% reduction of liver FAS activity. The different magnitude of training responses might be explained by the longer exercise duration and greater training intensity in this study (2 h/d, 5 d/wk for 10 wk) than the aforementioned study (1 h/d, 5 d/wk for 6 wk). The differences between the two studies in animal gender (female vs. male), diet (purified vs. nonpurified) and feeding pattern (meal vs. free feeding) might also play a role, but their influences on the outcome are not clear. Several other previous studies revealed no training effect on hepatic FAS activity using either male rats (Barakat et al. 1987, Winder et al. 1975) or female rats (Griffiths et al. 1993). However, rats in these studies all had free access to food, making it difficult to separate training vs. dietary effects on this enzyme that is highly sensitive to diet (Goodridge 1987). The current data were also comparable to our previous study in which there were 70 and 63% reductions of liver FAS activity after an acute exercise bout in rats starved/refed for 24 and 48 h, respectively (Griffiths et al. 1996). The high cornstarch diets used in these two studies were essentially identical; however, rats were exercised to exhaustion in the previous study, whereas in this study, rats were recovered from the last exercise bout for 27 h. Thus trained rats clearly had a decreased lipogenic potential even in the resting state, 9 h after consuming a high starch meal. This finding is extremely important because it provides an additional explanation for the lower plasma triglyceride and body fat observed in the trained state (Blair 1983).

The reason for this training-induced down-regulation of hepatic FAS cannot be clearly defined at present. Gene expression of FAS is regulated primarily by pretranslational mechanisms (Goodridge 1987). Insulin increases the rate of transcription of the FAS gene, whereas glucagon and cAMP cause it to decrease (Paulauskis and Sul 1989, Wilson et al. 1986). In addition, high CHO diets, particularly a high fructose diet, may induce FAS synthesis as a result of enhanced mRNA stability, mediated by high concentrations of glycolytic intermediates in the liver (Vaulont and Kahn 1994). We demonstrated that decreased hepatic FAS activity and mRNA abundance after an acute exercise bout were accompanied by a lower plasma insulin and a higher plasma glucagon concentration, as well as a lowered liver pyruvate concentration (Griffiths et al. 1996). However, these acute hormonal and metabolic changes that attenuate mRNA transcription and/or mRNA stability were not found in the trained rats (Table 2). Because plasma and liver samples were obtained in resting rats, one plausible explanation is that significant alterations of plasma hormonal profile and local metabolites in favor of inhibiting FAS activity occurred during each training session in C-fed rats, but recovered to the observed levels after exercise (Galbo 1983).

In contrast to the acute exercise effect, this study revealed no significant difference in relative adundance of FAS mRNA between the trained and untrained rats regardless of diets (Fig. 3). Because the diurnal fluctuation of FAS mRNA levels during meal feeding and training is unknown at present, the discrepancy between enzyme activity and mRNA abundance may be explained by the following: 1) translational and post-translational mechansims were responsible for the down-regulated FAS activity without changing mRNA levels and 2) mRNA levels were suppressed transiently after each training session, but returned to the resting levels due to the influence of meal feeding. Regarding the second possibility, it is noteworthy that mRNA abundance of ACC tended to be down-regulated by training (P < 0.075). Endurance training generally causes little change in resting plasma insulin concentration (Galbo 1983, Gyntelberg et al. 1977), whereas greater plasma insulin levels have been found in meal-fed rats than in either free-feeding or starved/refed rats (Romsos and Leveille 1974). Hyperinsulinemia after each meal feeding conceivably had a stimulatory effect on FAS mRNA transcription, whereas FAS enzyme synthesis lagged behind and remained down-regulated.

Chronic meal feeding of a high F diet resulted in a prominent up-regulation of all lipogenic enzyme activities in rat liver. The differences between F- and C-fed rats for each individual enzyme ranged from two- to fivefold, identical to those found in the free-feeding rats reported in our previous study (Griffiths et al. 1993). Consistent with the enzyme activity, the mRNA abundance for FAS was elevated by more than fivefold with fructose feeding, and ACC mRNA was also increased significantly (Fig. 3). These data were consistent with the well-known effect of fructose feeding on hepatic lipogenesis and agreed with the contention that dietary induction of major lipogenic enzymes by CHO is mediated by a pretranslational mechanism (Goodridge 1987). Alterations of lipogenic enzyme status with fructose feeding were accompanied by significant increases in absolute and relative liver weights in rats without alterations in body weight. Because dietary compositions and food intake per 100 g body weight were similar in the two studies, it appears that meal feeding and free feeding have similar effects on dietary induction of hepatic lipogenic enzymes as long as the intakes of fructose and other nutrients are the same. Fructose feeding caused a relatively small induction of ACC activity in the liver compared with other lipogenic enzymes, possibly due to the fact that ACC is also regulated by mechanisms other than protein synthesis, such as allosteric and covalent modification, and polymerization (Pape et al. 1988). The current data do not clearly explain why fructose feeding could up-regulate lipogenic enzymes more than complex CHO feeding. Plasma concentrations of insulin and glucose, which play an important role in the dietary regulation of hepatic lipogneic enzymes (Paulauskis and Sul 1989; Vaulont and Kahn 1994), did not differ in F- and C-fed rats. Concentration of liver pyruvate, a marker of glycolytic intermediates that have been proposed to stablize mRNA (Vaulont and Kahn 1994), was also unaffected by fructose feeding. Thus, although there is convincing evidence that increased local metabolites are largely responsible for the fructose induction of lipogenic enzymes, this conclusion is derived primarily from starved and refed animal models (Noguchi et al. 1985). Alternative mechanisms of a potent fructose effect with meal feeding may still exist. This study unexpectedly revealed that a high F diet could decrease proportions of the PUFA, linoleate (18:2) and arachidonate (20:4), in liver (Table 4). Because PUFA suppress lipogenic enzyme gene expression in rat liver (Clark and Jump 1996, Clarke et al. 1990), it is possible that an additional mechanism for a fructose-induced lipogenesis is to lower hepatic PUFA levels, thereby relieving their inhibition on lipogenic enzymes.

Fructose feeding severely attenuated the down-regulation of FAS activity due to training as seen in the C-fed rats. A 50% reduction of FAS activity was observed in the trained C-fed rats, whereas in the F-fed rats, training elicited a small and nonsignificant effect (Fig. 1). These differential training responses due to fructose feeding were also noted in our previous study in which F-fed rats showed only 10% (P > 0.05) inhibition of FAS activity in response to an acute bout of exercise, compared with a 70% inhibition in C-fed rats, after 24-h refeeding (Griffiths et al. 1996). These findings suggest that during meal feeding, a high fructose diet can protect hepatic FAS from exercise-induced down-regulation.

In contrast to FAS, a significant down-regulation of L-PK activity was seen in the F-fed rats but not the C-fed rats. This result was in agreement with our previous study, in which an acute bout of exercise dramatically decreased L-PK activity only in F-fed rats (Griffiths et al. 1996). L-PK plays an important role in directing hepatic triose phosphates toward the lipogenic rather than the gluconeogenic pathway, and fructose feeding stimulates L-PK gene expression (Goodridge 1987). A decrease of L-PK activity with training is expected to reduce the availability of lipogenic precursors resulting from fructose metabolism.

In summary, this study revealed that endurance training performed 4 h after the daily meal can markedly suppress hepatic FAS enzyme activity while not altering resting levels of FAS mRNA. A high fructose diet may attenuate this training effect on FAS, because it promotes both enzyme activity and mRNA abundance for FAS and ACC via pretranslational mechanisms. However, fructose induction of L-PK activity is down-regulated by training possibly due to post-translational mechanisms. The inhibitory effect of training on ACC mRNA is relatively small and does not alter its enzyme activity. Thus, our data indicate that endurance training is effective in inhibiting key lipogenic enzymes in meal-fed rats. Exercise-mediated down-regulation of hepatic lipogenic enzymes may be important in the prevention of hypertriglyceridemia and obesity when the main dietary energy is from non-fat sources.

	ACKNOWLEDGMENTS

The cDNA for FAS and -actin were a kind gift from Stuart Smith, Oakland Children's Hospital, Oakland, CA. The cDNA for ACC was generously provided by Ki-Han Kim, Department of Biochemistry, Purdue University, Lafayette, IN. The cDNA for GAPDH was provided by Ye-Shih Ho, Wayne State University, Detroit, MI. We thank Thomas Griffiths, Jan Novakofski, Xing-xian Yu and Mike Grahn for their assistance in the study.

	FOOTNOTES

Manuscript received 30 September 1996. Initial reviews completed 5 December 1996. Revision accepted 23 December 1997.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Armstrong M. K., Romsos D. R., Leveille G. A. Time sequence of lipogenic changes in adipose tissue of rats when converted from ad libitum feeding to meal-eating. J. Nutr. 1976; 106:884-891
• Askew E. W., Barakat H., Kuhl G. L., Dohm G. L. Response of lipogenesis and fatty acid synthetase to physical training and exhaustive exercise in rats. Lipids 1975; 10:491-496 [Medline]
• Bai D. H., Pape M. E., Casillas F. L., Luo X. C., Dixon J. E., Kim K. H. Molecular cloning of cDNA for acetyl-coenzyme A carboxylase. J. Biol. Chem. 1986; 261:12395-12399 [Abstract/Free Full Text]
• Barakat H. A., Dohm G. G., Shukla N., Marks R. H., Kern M., Carpenter J. W., Mazzeo R. S. Influence of age and exercise training on lipid metabolism in Fischer-344 rats. J. Appl. Physiol. 1987; 67:1638-1642 ^{[Abstract/Free Full Text]}
• Blair S. N. Evidence for success of exercise weight loss and control. Ann. Intern. Med. 1983; 119:702-706
• Bradford M. A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principles of protein-dye binding. Anal. Biochem 1976; 72:248-254 [Medline]
• Bucher, T., Czok, R., Lamprecht, W. & Latzko, E. (1965) Pyruvate. In: Methods of Enzymatic Analysis (Bergmeyer H.-U., ed.), pp. 253-259. Academic Press, New York, NY.
• Chomczynski P., Sacchi N. Single-step method of RNA isolation by acid guanidium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 1987; 162:152-159
• Clarke S. D., Armstrong M. K., Jump D. B. Dietary polyunsaturated fats uniquely suppress rat liver fatty acid synthase and S14 mRNA content. J. Nutr. 1990; 120:225-232
• Clarke S. D., Jump D. B. Polyunsaturated fatty acid regulation of hepatic gene transcription. J. Nutr. 1996; 126:1105S-1109S
• Coggan A. R., Swanson S. C., Mendenhall L. A., Habash D. L., Kien C. L. Effect of endurance training on hepatic glycogenolysis and gluconeogenesis during prolonged exercise in men. Am. J. Physiol. 1995; 268:E375-E383 [Abstract/Free Full Text]
• Feinberg A. P., Vogelstein B. A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 1983; 132:6-13 [Medline]
• Galbo H. Hormonal and Metabolic Adaptation to Exercise. Thieme-Stratton, New York, NY. Goodridge, A. G. (1987) Dietary regulation of gene expression: enzymes involved in carbohydrate and lipid metabolism. Annu. Rev. Nutr. 1983; 7:157-185 ^[Medline]
• Griffiths M. A., Baker D. H., Novakofski J. E., Ji L. L. Effects of exercise training on diet-induced lipogenic enzymes and body composition in rats. J. Am. Coll. Nutr. 1993; 12:155-161 [Abstract]
• Griffiths M. A., Fiebig R., Gore M. T., Baker D. H., Esser K., Oscai L., Ji L. L. Exercise down-regulates hepatic lipogenic enzymes in food-deprived and refed rats. J. Nutr. 1996; 126:1959-1971
• Gyntelberg F., Rennie M. J., Hickson R. C., Holloszy J. O. Effect of training on the response of plasma glucagon to exercise. J. Appl. Physiol. 1977; 43:302-305 [Abstract/Free Full Text]
• Inoue H., Lowenstei J. M. Acetyl coenzyme A carboxylase from rat liver. Purification and demonstration of different subunits. J. Biol. Chem. 1972; 247:4825-32 [Abstract/Free Full Text]
• Iritani N., Nishimoto N., Katsurada A., Fukuda H. Regulation of hepatic lipogenic enzyme gene expression by diet quantity in rats fed a fat-free, high carbohydrate diet. J. Nutr. 1992; 122:28-36
• Monsma C. C., Ney D. M. Interrelationship of stearic acid content and triacylglycerol composition of lard, beef tallow and cocoa butter in rats. Lipids 1993; 28:539-547 [Medline]
• Noguchi T., Inoue H., Tanaka T. Transcriptional and post-transcriptional regulation of L-type pyruvate kinase in diabetic rat liver by insulin and dietary fructose. J. Biol. Chem. 1985; 260:14393-14397 [Abstract/Free Full Text]
• Oscai L. B. Dietary induced severe obesity: a rat model. Am. J. Physiol. 1982; 242:R212-R215
• Pape M. E., Lopez-Casillas F., Kim K. H. Physiological regulation of acetyl CoA carboxylase expression: effect of diet, diabetes, and lactation on acetyl CoA carboxylase mRNA. Arch. Biochem. Biophys. 1988; 271:104-109
• Paulauskis J. D., Sul H. S. Hormonal regulation of mouse fatty acid synthase gene transcription in liver. J. Biol. Chem. 1989; 264:574-577 [Abstract/Free Full Text]
• Romsos D. E., Leveille G. A. Effect of meal frequency and diet composition on glucose tolerance in the rat. J. Nutr. 1974; 104:1503-1512
• Saltin, B. & Astrand, P. O. (1993) Free fatty acids and exercise. Am. J. Clin. Nutr. 57 (suppl.): 752S-758S.
• Schaible T. F., Scheuer J. Cardiac adaptations to chronic exercise. Prog. Cardiovasc. Dis. 1985; 27:297-234 [Medline]
• Sheperd D., Garland P. B. Citrate synthase. Methods Enzymol. 1969; 8:11-16
• Vaulont S., Kahn A. Transcriptional control of metabolic regulation genes by carbohydrates. FASEB J. 1994; 8:28-35 [Abstract]
• Wilson S. B., Back D. W., Morris S. M. Jr., Swierczynski J., Goodridge A. G. Hormonal regulation of lipogenic enzymes in chick embryo hepatocytes in culture. J. Biol. Chem. 1986; 261:15179-15182 [Abstract/Free Full Text]
• Winder W. W., Booth F. W., Fitts R. H., Holloszy J. O. Effect of exercise on response of liver lipogenic enzymes to a high fructose diet. Proc. Soc. Exp. Biol. Med. 1975; 148:1150-1154 [Medline]

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences