Adipose Lipoprotein Lipase Activity Research Articles

Role of insulin resistance in decreasing lipoprotein lipase activity in tumor-bearing rats.

The role of insulin resistance in the tumor-induced decrease in tissue lipoprotein lipase (LPL) activity was studied in vivo and vitro in methylcholanthrene-induced sarcoma-bearing rats. Intraperitoneal (i.p.) injection of 2U of regular insulin resulted in high-adipose LPL activity in control rats (CTR) of 122.0 +/- 42.4 U/mg tissue, but it had little effect on tumor-bearing rats (TBR), which showed a value of only 9.6 +/- 5.5 U/mg tissue (P = 0.002). When adjusted for serum insulin concentrations, adipose LPL activity remained significantly different between the TBR and CTR at 0.19 +/- 0.17 and 0.78 +/- 0.29 U/mg tissue, respectively (P = 0.02). Following the in vitro incubation with either 1.44 g/l glucose of 1 x 10-8 U insulin of adipose tissue fragments obtained from the TBR and the CTR, measurable LPL activity was maintained in the tissue from the CTR for 2 h but not in that from the TBR. These results suggest that the decreased LPL activities seen in the tumor-bearing state may be mediated by insulin resistance.

Measurement and seasonal variations of black bear adipose lipoprotein lipase activity

The black bear ( Ursus americanus) provides a unique model for the study of adipose physiology because it exhibits seasonal periods of rapid weight gain and weight loss without marked changes in its metabolic rate. To better understand fat cycling in this model, we obtained plasma and gluteal adipose tissue from five captive adult bears at approximately 20-day intervals from October 1 1992 through March 31 1993. The study included a predenning and denning period for each animal. Sampling during the predenning period followed a 12-h fast. Bears were anorectic while denning. Adipose LPL activities and plasma insulin concentrations were determined for each time point. Predenning LPL activities (4.83 ± 0.64 μmol/h/g) were significantly greater than those seen during the denning period (1.82 ± 0.65, p < 0.001). A biphasic pattern of fasting LPL activity was seen in four of the five bears during the predenning period. Fasting insulin concentrations showed no such pattern of variation during the study period (mean = 25.1 ± 1.36 pmol/l; range 1.1–6.0). We found no evidence of a linear relationship between LPL activity and insulin levels ( p = 0.139). Neither LPL activity nor insulin concentrations were related to changes in weight ( p = 0.257 and p = 0.7104, respecitively). We conclude that LPL activity can be measured in black bear adipose tissue and that fall (predenning) activities are significantly higher than those seen during the winter (denning period). Furthermore, the seasonal regulation of LPL involves some factor(s) in addition to insulin.

Effect of endotoxin and cytokines on lipoprotein lipase activity in mice.

Endotoxin (lipopolysaccharide [LPS]) stimulates the production of cytokines, which mediate many of the metabolic effects associated with infection. In LPS-sensitive C57B1/6 mice, LPS doses as low as 0.01 micrograms per mouse decreased adipose tissue lipoprotein lipase (LPL) activity by greater than 50%. In LPS-resistant C3H/HeJ mice, which do not produce cytokines in response to LPS, doses of LPS as high as 10 micrograms per mouse did not affect LPL activity in adipose tissue. In muscle of C57Bl/6 mice, LPL activity was decreased by 27% after 10 micrograms of LPS, whereas in C3H/HeJ mice there was no effect. These results indicate that the LPS-induced decrease in both adipose and muscle LPL activity is mediated by cytokines. Tumor necrosis factor (TNF), interleukin (IL)-1, leukemia-inhibiting factor (LIF), interferon alfa, and interferon gamma all decreased adipose tissue LPL activity in intact mice. In skeletal and cardiac muscle, only IL-1 and interferon gamma decreased LPL activity, whereas TNF, LIF, and interferon alfa had no effect. Inhibition of TNF activity blocked the increase in serum triglycerides that is characteristically observed after LPS but did not affect the ability of LPS to decrease adipose tissue LPL activity. Inhibition of IL-1 activity with IL-1 receptor antagonist partially inhibited the increase in serum triglycerides; however, the ability of LPS to decrease LPL activity in either adipose or muscle tissue was not affected. These data indicate that although TNF and IL-1 play a role in mediating the increase in serum triglyceride levels, these cytokines do not play a crucial role in the inhibition of either adipose or muscle LPL activity.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of estradiol on lipoprotein lipase activity and lipid availability in exercised male rats.

The effects of 17 beta-estradiol 3-benzoate (10 micrograms.0.1 ml sunflower oil-1 x 100 g body wt-1) and exercise on tissue lipid content and lipoprotein lipase (LPL) activity were determined in male rats. Estradiol administration significantly (P < 0.05) increased fatty acid contents of resting adipose, plasma, and white and red vastus muscle tissues and red vastus muscle triacylglycerol. Adipose and plasma fatty acids and red and white vastus muscle triacylglycerol were significantly higher in exercised estradiol-administered animals than in exercised oil-administered animals. Estradiol administration significantly reduced resting adipocyte LPL activity by 71% and increased myocardial LPL activity by 96%. After exercise, red vastus LPL activity was significantly increased by 76% in estradiol-administered animals compared with oil-administered animals. Ratios of red vastus to adipose LPL activity and myocardial to adipose LPL activity at rest and after exercise were significantly greater in estradiol-administered than in oil-administered animals. Estradiol administration significantly increased the ratio of white vastus to adipose LPL activity of exercised animals. These data indicate that estradiol increases the availability of lipid substrate to exercising muscle from multiple sources, including adipose, plasma, and intracellular muscle triacylglycerol. The absolute increases in muscle LPL activity, combined with a greater ratio of muscle to adipose LPL activity, lead to increased distribution of plasma triacylglycerol-derived fatty acids toward muscle.

Regulation of adipose tissue lipoprotein lipase gene expression by thyroid hormone in rats.

Lipoprotein lipase (LPL) is an important enzyme in lipid metabolism, and adipose LPL activity is increased in rats that are deficient in thyroid hormone. To examine the mechanism of thyroid hormone's effect on LPL, LPL gene expression was assessed in the epididymal fat pads of hypothyroid rats. When compared to control rats, LPL activity, mass, and synthetic rate in hypothyroid rats were increased; heparin-releasable LPL activity and mass were increased to 448% and 300% of control, respectively, and [35S]methionine incorporation into LPL was increased to 250% of control. The increases in LPL activity and mass were reversed by treatment of hypothyroid rats with triiodothyronine (T3). However, there was no change in the level of LPL mRNA when compared to the level of gamma-actin mRNA and no effect on LPL transcription using run-off assays. Isolated adipocytes were prepared from normal rats and exposed to 2 nM T3 in vitro for 24 h. The addition of T3 to cultures of adipocytes resulted in a decrease in LPL activity, mass, and [35S]methionine incorporation, but still no change in LPL mRNA level. To determine whether thyroid hormone regulated catecholamine responsiveness, adipocytes were prepared from hypothyroid and control rats, and the responses to epinephrine were compared. Although epinephrine inhibited [35S]methionine incorporation into LPL in control rat adipocytes, there was essentially no effect in hypothyroid rat cells. In addition, T3 treatment of the hypothyroid rats restored the responsiveness to epinephrine. Thus, thyroid hormone regulates LPL in rat adipose tissue posttranscriptionally, resulting in parallel changes in LPL synthetic rate, immunoreactive mass, and activity.(ABSTRACT TRUNCATED AT 250 WORDS)

A north-temperate migratory bird: a model for the fate of lipids during exercise of long duration

Changes in deposition and the utilization of lipids during one night of migratory activity (nocturnal physical activity) were investigated in dark-eyed Juncos (Junco hyemalis) held in large outdoor aviaries. During vernal migration (May), captive Juncos were sampled at the beginning and conclusion of one night of nocturnal restlessness. Comparisons of variables were drawn with control samples collected from birds in March. Measurements included body weight, fat stored in subcutaneous depots (adiposity), adipose and muscle lipoprotein lipase activity, and fat cell lipolysis. During the migratory period, body weight and adiposity were increased over levels measured in the March birds (p less than 0.01). On the other hand, neither body weight nor adiposity were significantly altered as a result of nocturnal physical activity and no significant changes were observed in adipose lipoprotein lipase activity. Fat cell lipolysis was lower at the beginning than at the end of nocturnal physical activity while, the opposite was observed for muscle lipoprotein lipase activity (p less than 0.05). These results suggest that the amount of work of one night of nocturnal physical activity modifies both muscle lipoprotein lipase activity and fat cell lipolysis in an interrelated fashion. This phenomenon could act to direct a steady supply of fatty acids to the site of energy utilization, i.e., flight muscles.

Synthesis and regulation of lipoprotein lipase in the hippocampus.

Lipoprotein lipase (LPL) expression was determined in adult rat hippocampus and compared to enzyme expression in other brain regions. Hippocampus LPL mRNA levels were at least 2.5-fold higher than those detected in the cerebral cortex, cerebellum, and remaining brain regions. Enzyme mass and activity levels in the hippocampus were also increased to a similar degree. De novo synthesis of LPL in the hippocampus was confirmed by [35S]methionine-labeling of the tissue and identification of a 57 kDa protein obtained by immunoprecipitation. Addition of an excess amount of bovine LPL completely prevented the immunoprecipitation of this protein. The effect of nutritional modulations on brain LPL activity was determined after a 12-h fast. While no significant changes were observed in other regions of the brain, hippocampus LPL activity in fasted rats increased by 60% compared to the fed control group. Simultaneously, fasting reduced adipose LPL activity by 60%. Intraperitoneal injection of ACTH over a 5-day period had no effect on hippocampus LPL activity, while adipose LPL levels increased 2.3-fold and heart LPL levels decreased 1.4-fold. We conclude that LPL is synthesized, active and regulated in a tissue-specific manner in the adult rat hippocampus.

Two different mechanisms are involved in nutritional regulation of lipoprotein lipase in guinea-pig adipose tissue.

Lipoprotein lipase activity in adipose tissue responds rapidly to changes in the physiological state. To study what mechanisms are involved in the regulation, guinea pigs were fasted and the decrease in adipose-tissue lipoprotein lipase activity was compared with the decreases in mRNA and lipase synthesis. The mRNA pattern (three species) did not change. There was a close parallelism between the abundance of lipase mRNA and relative lipase synthesis (immunoprecipitable 35S-labelled lipoprotein lipase as fraction of total [35S]protein after pulse-labelling with [35S]methionine). Total protein synthesis decreased on fasting, compounding the decrease in relative lipase synthesis. Lipoprotein lipase mRNA changed similarly in fat-pads and in isolated adipocytes, whereas lipase activity changed more in the pads, indicating disproportionally large changes in extracellularly located lipase. In old guinea pigs the decreases in lipoprotein lipase activity and lipase synthesis were comparable, but in young animals the change in lipase activity was substantially larger than the change in lipase synthesis. Refeeding of fasted young guinea pigs with glucose resulted in a rapid increase in lipoprotein lipase activity, but there was only a small change in lipase mRNA. Old animals responded slowly to refeeding. The results indicate that in older animals the major mechanism for regulation of adipose lipoprotein lipase activity is a relatively slow change in lipase mRNA, whereas in younger animals an additional, more rapid, regulation is exerted on the transport and turnover of the enzyme.

Effect of tumor necrosis factor (TNF) on lipid metabolism in the diabetic rat. Evidence that inhibition of adipose tissue lipoprotein lipase activity is not required for TNF-induced hyperlipidemia.

Tumor necrosis factor (TNF) administration produces an increase in plasma triglycerides that may be due to inhibition of adipose lipoprotein lipase activity and/or a stimulation of hepatic lipogenesis. We now report that TNF administration to insulinopenic diabetic rats increases serum triglycerides (2 h, 2.4-fold; 17 h, 4.3-fold). Adipose tissue lipoprotein lipase activity was markedly decreased in diabetic animals compared with controls and was not further inhibited by TNF. Incorporation of tritiated water into fatty acids in the liver was increased 45% 1-2 h after TNF and 87% at 16-17 h. These results indicate that the TNF-induced increase in circulating lipid levels can occur in the absence of a TNF-induced inhibition of adipose tissue lipoprotein lipase activity. Moreover, the clearance from the circulation of triglycerides in chylomicrons was similar in control and TNF-treated animals; these results provide further evidence that the removal of triglyceride-rich lipoproteins is not altered in the TNF-treated animals. Our data suggest that the TNF-induced stimulation of hepatic lipid synthesis may play an important role in the increase in serum triglycerides. In addition, TNF administration to diabetic animals leads to an elevation in serum glucose levels (73% at 17 h) without a change in serum insulin levels. Thus, TNF stimulation of hepatic lipogenesis is independent of changes in insulin.

Exercise termination effects on food intake, plasma insulin, and adipose lipoprotein lipase activity in the Osborne-Mendel rat

Thirteen-week-old male, Osborne-Mendel rats were exercised for 6 weeks on a motorized treadmill. Exercise depressed weight gain and cumulative light cycle food intake while cumulative dark cycle and 24-hour total food intake were unaffected. Rats in sedentary and exercise groups were killed 24 hours after the last bout of exercise to assess the effects of chronic exercise and at 48, 60, 72, and 84 hours to determine the effects of exercise termination. Compared to sedentary controls, exercise decreased plasma insulin, epididymal and retroperitoneal depot weight and cell size, and retroperitoneal lipoprotein lipase (LPL) activity. Forty-eight hours after exercise, plasma insulin concentration increased to sedentary levels. By 60 hours, dark cycle food intake was increased above and adipose LPL activity was comparable to sedentary levels. At 84 hours postexercise termination, dark cycle food intake, plasma triglyceride, and epididymal LPL activity per depot and per cell were significantly greater than sedentary values. Exercise termination resulted in a preparatory response for rapid lipid deposition probably arising from increased food intake, plasma insulin, and enhanced LPL activity within 84 hours following termination of exercise.

Effect of adrenalectomy and glucocorticoid replacement on development of obesity.

Female obese and lean Zucker rats were adrenalectomized (ADX) or sham-operated at 4 wk of age. ADX animals were given daily injections of 0.01, 0.05, 0.50, 1.0, or 2.0 mg hydrocortisone/100 g body wt for 30 days. ADX rats gained less weight than sham-operated controls. Obese ADX rats at the lowest dose (0.01) had a net positive energy gain but lost body fat. As steroid dose increased, obese rats deposited more fat and less protein. Doses of 0.01 and 0.05 mg produced rats that were less fat than sham-operated controls, whereas doses of 0.50, 1.0, and 2.0 mg produced rats of comparable body fat composition. Obese rats were consistently fatter and had a significantly smaller percentage body protein than lean rats at each dose. Body fat elevation was reflected by heavier parametrial and retroperitoneal fat depots and larger fat cells at all doses except the lowest. Compared with sham-operated controls, lean and obese rats at the two lowest replacement doses (0.01, 0.05) exhibited significantly decreased plasma insulin and triglyceride levels and significantly elevated brown adipose tissue protein content and citrate synthase (CS) activity. Obese rats at these doses had significantly reduced adipose tissue lipoprotein lipase (LPL) activity in the retroperitoneal depot and lower food intake. Furthermore, these obese rats had adipose depot weights, cell sizes, LPL activity, and plasma insulin, glucose, and triglyceride comparable to that of lean sham-operated controls. As steroid dose increased (0.5, 1.0, 2.0), plasma insulin and triglyceride and food intake markedly increased only in obese rats. Adipose tissue LPL activity appeared unaffected by dose. Brown adipose tissue protein content and CS activity significantly decreased as dose increased in both lean and obese rats. At all doses of replacement obese rats were more responsive to steroid than were lean rats. Obese rats receiving 0.01 mg had comparable fat depot weights, cell sizes, and plasma insulin and triglyceride as lean rats receiving 50 times as much steroid per day (0.50 mg). These results suggest glucocorticoids play an important role in the early development of obesity in the Zucker rat and support the hypothesis that obese rats are more responsive to glucocorticoids than are lean rats.

Modulation of adipose lipoprotein lipase by thyroid hormone and diabetes. The significance of the low T3 state.

The present study was performed to assess the potential relationship between the low T3 syndrome and hypothyroidism. Comparative studies were performed on the relative effects of diabetes and insulin on heparin-releasable adipose lipoprotein lipase (LPL) in the intact and hypothyroid rat. Hypothyroidism for 10 days (Tx) significantly increased adipose LPL activity (5.8 +/- 0.2 mu eq/g/h) compared with the activity (3.6 +/- 0.4 mu eq/g/h) in the normal group. Diabetes for 72 h (streptozocin, STZ, 10 mg/100 g body wt, i.p.) significantly reduced (P less than 0.005) adipose LPL activity in the Tx model. However, despite the suppressant effect of diabetes (43 +/- 11%), the enzyme activity remained equivalent to the normal group. Insulin stimulated adipose LPL in the Tx-diabetic group. The enzyme demonstrated a synergistic response to insulin and hypothyroidism. Subsequent studies were performed in the intact diabetic rat, a low T3 state. Adipose LPL activity was reduced to a similar degree by diabetes (79 +/- 2%) irrespective of the serum T3 concentration. Furthermore, the magnitude of the adipose LPL stimulation by insulin was not modulated by the endogenous serum T3. However, co-treatment of the diabetic group with T3 and insulin blunted the adipose LPL response to insulin. These various modulations in adipose LPL activity were associated with significant but opposite changes in serum triglyceride levels in both the hypothyroid and intact rat. These studies demonstrate that hypothyroidism counteracts the suppressant effect of diabetes on heparin-releasable rat adipose LPL activity and magnifies the enzyme response to insulin.(ABSTRACT TRUNCATED AT 250 WORDS)

Activité lipoprotéine-lipasique de différents dépôts adipeux, et ses relations avec la taille des adipocytes chez la vache tarie en cours d'engraissement, ou en début de lactation

Twelve dry and eight lactating cows were slaughtered and the whole body was chemically analysed. Adipose tissue lipid and adipose cell size varied greatly with body condition and also with anatomical site (perirenal, omental, intermuscular, rump subcutaneous, leg subcutaneous). Lipoprotein lipase activity per adipose cell was very low in lactating cows, but very high in dry cows during fattening. Activity tended to be higher in omental adipose tissue and lower in subcutaneous adipose tissue. Total adipose lipoprotein lipase activity per cow was positively related with energy balance and negatively related with plasma non-esterified fatty acids. There was no correlation with body fatness or adipose cellularity.

Evidence for independent genetic regulation of heart and adipose lipoprotein lipase activity.

The relationship between the genes controlling heart and adipose lipoprotein lipase in fasted animals has been studied. 32 inbred mouse strains were tested for variations in heart or adipose specific activity and thermolability. The survey revealed that specific activity of heart and adipose lipoprotein lipase varied as much as 3-fold and 20-fold, respectively. In thermolability, up to a 2-fold variation was observed in the lipase in each tissue. The correlation coefficient between variations in heart and adipose lipase was apparently not significant for both parameters studied. Additional studies were performed in two strains, BALB/c and C57BL/6, along with the recombinant inbred set derived from them. The two strains did not show genetic variation for lipoprotein lipase thermolability, although the inactivation rate of heart lipase was higher than that of adipose lipase. However, BALB/c and C57BL/6 displayed significant differences in their levels of lipoprotein lipase specific activity. Thus, strain C57BL/6 showed higher heart activity when compared to BALB/c, whereas the latter showed higher adipose lipase activity when compared to C57BL/6, i.e. an inverse relationship. The specific activity levels of heart and adipose lipoprotein lipase in the recombinant inbred strains derived from BALB/c and C57BL/6 exhibited independent inheritance. Thus, in adipose tissue, a single major gene seems to control the variation observed, while the inheritance pattern of heart activity could imply involvement of more than one gene. Moreover, two out of the seven recombinant strains showed distinct recombinant phenotypes, indicating that separate unlinked genes control the variations found in heart and adipose activity. We conclude that the expression of heart and adipose lipoprotein lipase activity is under independent genetic control.

Lipoprotein lipase activity and lipolysis after swim training in obese Zucker rats.

Obese and lean Zucker rats, 7 wk old, were swim trained or kept sedentary. Another group of obese rats was food restricted and exercised. Half the rats were killed after training for 8 wk, the remainder were retired and killed after an additional 8 wk. Neither treatment decreased adipocyte size in obese rats. Although basal lipolysis per cell was elevated in obese rats, their adipocytes were insensitive to epinephrine at 15 and 23 wk of age. Exercise training did not affect lipolysis. At all ages, adipose lipoprotein lipase (LPL) capacity was higher in obese relative to lean rats. In obese rats, swim training and exercise plus food restriction increased adipose and gastrocnemius LPL activity and depressed plasma insulin and triglyceride levels. All effects of exercise were transient. Thus, exercise training improved some of the metabolic disturbances in the Zucker obese rat but did not normalize adipocyte size, LPL activity, or lipolysis.

Uterine and adipose lipoprotein lipase activity in hormone-treated and pregnant rats.

We report for the first time the presence of lipoprotein lipase (LPL) activity in the uteri of adult rats. This uterine LPL activity was significantly elevated by administration of the estrogen, estradiol benzoate (EB), but not by treatment with the antiestrogen, nafoxidine. Both EB and nafoxidine, however, greatly reduce both parametrial and retroperitoneal adipose tissue LPL activities. Uterine LPL activity also changed as a function of the duration of pregnancy, increasing severalfold between days 14 and 18 and then decreasing by about 20% on day 21. These data suggest the existence of a hormone-dependent system of providing lipids for storage and/or utilization by metabolic and reproductive tissues as needed.

Rat adipocyte lipoprotein lipase activity is inhibited by theophylline and stimulated by inosine through adenosine-independent mechanisms

Incubation of rat adipose tissue or isolated rat adipocytes with high (50 mM) but not with low concentrations (0.5 mM) of theophylline results in a decrease of lipoprotein lipase (LPL) activity. This effect is not altered by the addition of adenosine deaminase, indicating that the decrease of adipose LPL activity by theophylline is not due to the competition of theophylline with adenosine. On the contrary, incubation of isolated fat cells with adenosine (0.1 – 100 μM) results in an increase of the intracellular form of LPL activity. As this effect is also observed in cells incubated with adenosine deaminase (40 mU/ml) or with inosine (0.1 – 100 μM) but not in cells incubated with the adenosine analog N 6-phenylisopropyladenosine, it is concluded that the increase in the intracellular form of LPL found after incubation with adenosine is not due to adenosine per se but to inosine generated from the breakdown of endogenous adenosine by adenosine deaminase.

The Influence of Dietary Fat and Meal Frequency on Lipoprotein Lipase and Hormone-Sensitive Lipase in Rat Adipose Tissue

Activities of NaCl-inactivated lipoprotein lipase (LPL) and protamine-resistant hormone-sensitive lipase (HSL) in adipose tissue, accumulation of carcass fat, and serum triglycerides (TG) were determined in meal-fed (MF) and ad libitum-fed (AD) rats. At each feeding frequency, diets provided total fat as 15 or 30% of calories and polyunsaturated fatty acids (PUFA) as 2.5 or 11% of calories. The average energy intake of MF rats was 67% that of AD rats. Total weight gained by MF rats was only 60% that of the AD rats. Significantly greater activities of LPL, HSL, and LPL:HSL in adipose of MF rats suggested a greater capacity for fat accumulation which was not realized at the limited energy intake. In AD rats, the percentage of body fat was significantly correlated with LPL:HSL and with serum TG, suggesting that the relative enzyme activities and fat deposition may be influenced by the concentration of circulating TG. Mean body fat of rats receiving the 30% fat diet was significantly greater than that of rats fed 15% fat. Both serum TG and adipose LPL activity were significantly reduced when the diet contained high levels of PUFA.

Alterations of lipid metabolism in healthy volunteers during long-term ethanol intake.

Nine young, healthy male volunteers were given ethanol (75 g/day) for 5 weeks. The ethanol was divided into five daily doses and taken so that blood ethanol levels never exceeded 0.04% (w/v). During the latter part of the ethanol intake period, there was a significant, transient increase of plasma triglyceride (TG) concentrations followed by reduction to normal levels. A three-fold increase of lipoprotein lipase activity (LLA) occurred in biopsy specimens of adipose tissue. An increase of alpha-lipoprotein concentrations, which correlated significantly with the decrease in plasma TG levels and the increase in adipose LLA, was also observed during the ethanol intake period. No changes were observed in plasma cholesterol and beta-lipoprotein levels. A transient, three-fold increase of TG concentrations occurred in liver biopsy specimens. Ultrastructural and cytochemical examinations of the biopsy specimens showed hyperplasia of the smooth endoplasmic reticulum, and increased canallicular activity of gamma-glutamyl transferase (gamma-GT) activity in most subjects towards the end of and after the ethanol intake period. Serum gamma-GT levels also increased significantly.

Weight, food intake, and body composition: effects of exercise and of protein deficiency.

Weight, food intake, and body composition: effects of exercise and of protein deficiency.