| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Institute of Arteriosclerosis Research at the University of Muenster (M.K., A.B., N.P., H.S., G.A., U.W.), 48149 Muenster, Germany; Institute of Clinical Chemistry (A.v.E.), 8091 Zurich, Switzerland; Institute of Clinical Chemistry and Laboratory Medicine (M.F., G.A.), University of Muenster, 48129 Muenster, Germany; and University of Applied Sciences (U.W.), 48151 Muenster, Germany
Address all correspondence and requests for reprints to: Dr. Mario Kratz, Institute of Arteriosclerosis Research at the University of Muenster, Domagkstrasse 3, 48149 Muenster, Germany. E-mail: mkratz{at}uni-muenster.de.
Abstract
The recently discovered hormone leptin is primarily secreted by adipose tissue and serves as an internal signal indicating the size of body fat stores. The aim of the present study was to investigate the impact of the dietary fatty acid composition on serum leptin concentrations. Therefore, serum leptin levels were measured by RIA in healthy nonobese men (n = 30) and women (n = 25). First, all participants received a baseline high-fat diet, rich in saturated fat, for 2 wk and were then randomly assigned to one of three high-fat dietary treatments, which contained refined olive oil (rich in monounsaturated fatty acids, n = 19), rapeseed oil [rich in monounsaturated fatty acids and 
-linolenic acid (18:3n-3), n = 17], or sunflower oil (rich in n-6-polyunsaturated fatty acids, n = 19) as the principal source of fat for 4 wk. On the rapeseed oil diet, serum leptin concentrations increased slightly in men [+0.25 ng/ml, T(9) = -2.778, P = 0.021], but decreased distinctly in women [-4.70 ng/ml, T(6) = 5.083, P = 0.002]. Both the olive oil and the sunflower oil diet did not affect serum leptin concentrations. Thus, it is proposed that serum leptin levels were affected by the high amount of -linolenic acid in rapeseed oil. However, questions remain as to why this diet differently affected serum leptin in men and women.
LEPTIN WAS DISCOVERED in 1994 (1) and is a 146-amino-acid peptide, which is primarily secreted by adipose tissue and serves as an internal signal indicating the size of the body fat stores (2). Although leptin has been shown to be involved in processes such as puberty or reproduction, there is agreement that the major function of the hormone is the regulation of body weight by affecting appetite, energy expenditure, and thermogenesis (2). The initial hypothesis that obesity in humans results from a relative or absolute deficiency of leptin could not be confirmed. Paradoxically, most obese humans have high circulating concentrations of leptin, which are even raised in proportion to fat mass (3). Thus, it has been suggested recently that human obesity might represent a state of leptin resistance (4, 5).
Maffei et al. (3) observed a pronounced interindividual heterogeneity in the leptin concentration at a given body mass index (BMI). These differences indicate that factors other than body fat mass also contribute to the regulation of serum leptin concentrations. Among these, glucose and insulin have been found to up-regulate leptin expression in adipocytes, whereas GHs or T4 suppress leptin expression (reviewed in Ref. 6). Sex hormones, such as estrogens and possibly also testosterone, have also been shown to play major roles in the regulation of leptin production, which is also reflected by the fact that serum leptin concentrations are markedly higher in women than in men (2).
Serum leptin concentrations are also affected by diet. Energy restriction and fasting have been shown to reduce serum leptin levels more than could be predicted from the loss in body fat mass in rats (7) and humans (8, 9). Furthermore, first studies indicate that the fatty acid composition of the diet might have an impact on serum leptin levels. Cha and Jones (7) observed in rats that a diet rich in n-6- and n-3-polyunsaturated fatty acids (PUFA) led to higher serum leptin levels than a diet rich in both saturated fatty acids (SFA) and monounsaturated fatty acids (MUFA). In contrast, Reseland et al. (10) showed that a high intake of n-3-PUFA decreased leptin gene expression both in vitro using a human cell line and in vivo in rats. Our group recently observed that serum leptin levels decreased in hypertriglyceridemic patients after dietary SFA had been replaced by marine n-3-PUFA and MUFA (11). However, in this study, the diet also differed, in several other regards, from the habitual diets of these patients (for example it contained more vegetables and more complex carbohydrates). Also, this diet resulted in dramatic decreases in serum triglycerides and a small, but significant, drop in body weight; and thus, the changes in serum leptin levels could not be exclusively attributed to the changes in dietary fat quality. Thus, there is still a lack of knowledge, with regard to the impact of dietary fatty acid composition on serum leptin levels, particularly in humans.
We, therefore, investigated the response of serum leptin concentrations to diets differing in their relative amounts of MUFA and n-6 and n-3 PUFA in healthy, young, nonobese male and female nonsmokers. The study had initially been designed to investigate the effect of the diets on low-density lipoprotein composition and oxidizability (12). Because leptin is closely related to glucose and insulin metabolism, we also investigated the effect of these diets on insulin, glucose, and glycosylated hemoglobin (HbA1c).
Subjects and Methods
Subjects
Of 700 students living under boarding school-like conditions in a third-level technical college, 115 nonsmoking volunteers were screened for participation. Inclusion criteria were: a BMI of less than 27 kg/m2, serum cholesterol concentrations less than 300 mg/dl, and triacylglycerol concentrations less than 300 mg/dl. Of the 115 volunteers, 1 was excluded because of diabetes mellitus; 3 because of hyperlipidemia; 5 because of thyroid disease; 2 because of intake of vitamin supplements; 4 because of hyperuricemia; and 25 because of allergy, intolerance, or aversion to foodstuffs contained in the study diets. Other exclusion criteria were: drug or substance abuse and malabsorption syndromes. Of the 75 students who qualified for participation in the study, 69 (35 male, 34 female), 18–43 yr old, were chosen for inclusion by drawing lots. Six subjects withdrew during the study because of intercurrent illness, and 5 withdrew because they were unwilling or unable to comply with the dietary regimen. Fifty-eight participants finished the study. Complete data sets for all parameters were available for 55 participants. The baseline characteristics of these 55 participants (30 male, 25 female) are shown in Table 1
. Twenty female participants (olive oil group, 8 out of 9; sunflower oil group, 7 out of 9; rapeseed oil group, 5 out of 7) who were taking oral contraceptives were instructed not to stop taking them and not to change to another pill. The participants were also asked not to change their regular lifestyles and their usual extent of physical activity throughout the study.
|
Design and diets
The study was conducted in a parallel design and consisted of two consecutive dietary periods for each subject. All participants consumed a baseline high-fat diet, rich in SFA, for 2 wk and were then randomly divided into three groups. This was done using tables of random digits, separate for men and women, which were generated by a professional biostatistician. Each group received a high-fat diet containing refined olive oil (10 men, 9 women), sunflower oil (10 men, 9 women), or rapeseed oil (10 men, 7 women), respectively, as the principal source of fat for 4 wk. These diets were identical in every respect, apart from the fatty acid composition. Venous blood samples were obtained at the beginning of the study (visit 1), after the baseline period (visit 2), after 2 wk of the study diets (visit 3), and at the end of the study (visit 4). All samples were drawn between 0645 h and 0845 h, after an overnight fast of at least 9 h. Serum was obtained immediately after venipuncture, and all samples were stored at -70 C until analysis.
The composition of the participants’ habitual diet and the study diets is shown in Table 2. All study diets consisted of conventional mixed foods that were freshly prepared. Menus were changed daily. The participants were served breakfast, lunch, and dinner from Monday morning to Friday noon. This food was immediately consumed in the school canteen, under the direct supervision of one of the authors (M.K.). On Friday afternoons, participants were given hampers containing their entire food supply for the weekend. All foodstuffs were weighed. Basic menus of the study diets were identical for all participants. All dietary items were low in fat, e.g. they contained lean meat, skimmed milk, and low-fat dairy products. This provided scope for the enrichment of these meals with the specific oils, which were provided in sauces, desserts, and salad dressings. A margarine was specially manufactured based on these oils. This margarine contained 20% water, 20% hard stock (coconut fat, palm kernel fat, and palm oil), and 60% refined olive oil, rapeseed oil, or sunflower oil, respectively. These margarines were supplemented with different amounts of vitamin E to compensate for the different amounts of vitamin E in the oils. We also used specially baked oil-enriched bread containing 10% oil (olive, rapeseed, or sunflower, respectively). To compensate for short-term differences in individual energy requirements, participants were provided, on request, with special bread rolls that were baked so as to contain exactly the same nutrient composition as that person’s study diet. By means of these rolls, energy balance was ensured without changing the composition of the diets.
|
.
|
Laboratory methods
Serum concentrations of leptin were measured by RIA (WAK Chemie, Homburg, Germany), and insulin was measured by enzyme-linked immunoassay (DAKO Corp., Arhus, Denmark). Measurements were performed in duplicate, and the complete set of samples of each participant was measured within the same series. Within-assay and between-assay coefficients of variation were 9.0% and 4.1% for leptin measurement, and 2.1% and 8.4% for insulin measurement, respectively. HbA1c (reference range for healthy individuals, 3.4–4.7%) was measured, from EDTA blood, by a Merck \|[amp ]\| Co., Inc.-Hitachi Scientific Instruments, Inc. L-9100 automated HPLC system (Merck \|[amp ]\| Co., Inc., Darmstadt, Germany). Glucose concentration in serum was analyzed, by the hexokinase method, on a Hitachi Scientific Instruments, Inc. 917 autoanalyzer (Roche Diagnostics, F. Hoffmann-La Roche, Basel, Switzerland). HbA1c and glucose are validated by regular analyses of reference samples supplied by the national German INSTAND proficiency testing program.
Statistical analysis
Statistical analyses were performed using the Statistical Package for the Social Sciences (version 10, SPSS, Inc., Chicago, IL). Leptin, insulin, glucose, and HbA1c were found to be approximately normally distributed, as confirmed by checking normal plots and histograms of the data and performing Kolmogorov-Smirnov tests. Because serum leptin concentrations differed widely between men and women, these examinations were done separately for men and women for this parameter.
Effects of the diets on serum leptin concentrations were assessed by repeated-measures ANOVA, with gender and diet group as between-subject factors. The serum leptin concentrations of visits 2–4 were used as the three levels of the innersubject factor (time), and the BMI values of visits 2–4 were used as a nonconstant covariate. The significant interaction effect found was then analyzed post hoc by Student’s t tests. For these tests, the differences between serum leptin of visit 2 and visit 4 were adjusted for the differences between BMI values of visit 2 and visit 4. The effect of the diets on serum concentrations of insulin, glucose, and HbA1c were also first analyzed by repeated-measures ANOVA, with gender and diet group as between-subject factors. Any differences found were then analyzed post hoc by paired Student’s t tests. The difference in serum leptin concentrations of visits 1 and 2, again adjusted for the change in BMI between visit 1 and visit 2, were also compared by Student’s t tests. Pearson’s correlation coefficient was calculated for the relation between the changes in serum insulin and the changes in serum leptin. All tests were two-tailed, and the level of significance was set to P less than 0.05.
Results
During the baseline diet, the serum leptin concentrations decreased by 0.57 ± 0.97 ng/ml (mean ± SD) (or 18%) in men and by 2.65 ± 4.08 ng/ml (or 18%) in women (Fig. 1). After adjustment for the changes in BMI, however, only the reduction in men remained statistically significant (T(29) = -3.692, P = 0.001).
|
]. By contrast, the rapeseed oil diet distinctly lowered serum leptin levels in women [-4.70 ng/ml, T(6) = 5.083, P = 0.002]. Also, a direct comparison of men and women, with regard to their leptin responses to the rapeseed oil diet, showed a highly significant difference [T(15) = -6.443, P < 0.001]. Serum leptin concentrations were not significantly affected by the diets rich in olive oil and sunflower oil, either in men or in women. Also, men and women did not differ in their leptin response to these diets.
|
). The baseline diet led to a small (but statistically significant) reduction in HbA1c [-0.05; 95% confidence interval, -0.01 to -0.09; T(54) = 2.554, P = 0.014] and lower fasting serum insulin levels [-7 pM; 95% confidence interval, -1 to -13 pM; T(54) = 2.397; P = 0.020]. Also, there was a nonsignificant trend toward lower fasting glucose concentrations in serum during the baseline diet [-0.08 mM; 95% confidence interval, -0.18 to +0.01 mM; T(54) = 1.766, P = 0.083].
|
Discussion
The present study showed that a diet rich in MUFA and -linolenic acid, compared with a SFA-rich diet, distinctly reduced serum leptin concentrations in women and slightly increased serum leptin levels in men. However, the latter effect was only small and supposedly of marginal clinical significance, particularly when the inconsistent nature of the changes in this group are taken into account. Nevertheless, these findings raise the interesting hypothesis that serum leptin concentrations in men and women respond differently to diet.
We observed a marked overall decrease in serum leptin concentrations of the female participants allocated to the rapeseed oil diet (49% within 6 wk) despite only minor changes in their body weight, and supposedly also body fat content and body fat distribution. A considerable part of this decrease occurred during the baseline diet, and because this diet also significantly lowered serum leptin levels of the men in our study, the baseline diet apparently had some leptin-lowering potency. The habitual diet and the baseline diet differed considerably in food pattern, with the baseline diet containing far more vegetables, fruit, and whole-meal products, reflected by a distinctly higher fiber intake and a lower content of mono- and disaccharides. Carbohydrate absorption is slowed by fiber (14), which is also reflected by an inverse relationship between total fiber intake and HbA1c, at least in diabetics (15). Thus, the higher fiber intake during the baseline diet might have attenuated the postprandial glucose and insulin response, and thus the 24-h blood glucose and insulin concentrations. This hypothesis is consistent with the observed decrease in HbA1c, the decreased fasting insulin level, and the observed trend toward lower fasting glucose concentrations at visit 2. Because both insulin and glucose up-regulate ob gene expression in and leptin secretion by adipocytes in vitro and in vivo (16, 17, 18, 19, 20, 21, 22, 23), serum leptin concentrations might have been lowered as a consequence of reduced serum levels of these stimuli. Furthermore, a positive correlation between HbA1c as an indicator for longer-term serum glucose levels, and serum leptin concentrations has been reported (20), indicating that glucose and insulin play major roles in the regulation of serum leptin concentrations. Though this is an appealing hypothesis, it must be stated clearly that we have not designed this study to investigate the effect of the baseline diet on serum leptin concentrations; and thus, hypotheses based on observations made on this diet have to be tested in well-controlled studies using an appropriate design.
The decrease in serum leptin concentrations in the female participants given the rapeseed oil diet might also, in part, be attributable to effects of this diet on glucose and insulin metabolism. On the one hand, this diet was also characterized by a high fiber content and a low glycemic index. Given that it was recently reported that rats fed a high-glycemic index diet had higher basal glucose and leptin levels than rats fed a low-glycemic index diet despite comparable body weights (24), the high fiber content and the low glycemic index might have contributed to the observed lowering of serum leptin concentrations. On the other hand, the three oil diets were similar with regard to fiber content and glycemic index. Thus, the reduction in serum leptin in the women given the rapeseed oil diet is likely to be the result of another, more specific characteristic of this diet. A possible explanation might be provided by the fact that unsaturated (and particularly, n-3-PUFA) fatty acids improve glucose tolerance and insulin sensitivity, compared with SFA, leading to a decrease in 24 h-glucose and insulin levels (reviewed in Refs. 25 and 26), and thus to a decreased stimulus for adipocyte leptin production. This might indeed apply to the women given the rapeseed oil diet in this study, where fasting serum insulin concentrations dropped to a considerable (albeit not statistically significant) degree. Also, the fall in serum leptin and the fall in serum insulin concentrations correlated strongly in this group, which was not the case for any other group. The lower leptin levels on the rapeseed oil diet in women might also be attributable to direct effects of -linolenic acid (18:3n-3) or its longer-chain derivatives on ob gene expression, as suggested by findings in rats (27, 28, 29) and humans (10). Consistent with our findings, Reseland et al. also reported that, on a diet rich in fiber and n-3-PUFA but low in SFA and cholesterol, i.e. a diet comparable with our rapeseed oil diet, BMI-adjusted plasma leptin concentrations decreased in normal-weight human subjects. Although it remained unclear, from their study, to which specific nutrients the decline in leptin levels was primarily attributable, the inability of the olive oil diet and the sunflower oil diet to reduce serum leptin concentrations in the women of our study suggests that the high content of -linolenic acid in the rapeseed oil diet could have been the crucial factor.
Questions remain as to why this reduction of serum leptin concentrations did not also occur in the men given the rapeseed oil diet. One reason for this discrepancy might be that the reduction in serum leptin concentrations on the baseline diet in men limited the scope for further reductions. Also, it stands to reason that this difference might have been caused by disparities in sex hormone metabolism. It was reported by Ludwig et al. (31) that, in women, serum concentrations of leptin vary during the menstrual cycle, with a maximum in the luteal phase and a minimum in the follicular phase. We have not measured sex hormones and thus did not control for the menstrual cycle of the participating women. Thus, it is likely that, within a treatment group, women were assessed at different stages of their cycle. It is, therefore, extremely unlikely that the effect of the rapeseed oil diet on serum leptin levels has been confounded by menstrual cycle, particularly because the women of the olive oil group and the sunflower oil group had very similar serum leptin concentrations at different visits. However, the difference between men and women, with regard to their leptin responses to the rapeseed oil diet, might well be attributable to effects of the dietary fatty acids on serum sex hormone concentrations or their interaction with leptin metabolism (2). Unfortunately, current knowledge on this potential interaction is limited.
Another interesting observation was that, despite distinctly lower serum leptin concentrations at the end of the study, the women in the rapeseed oil group did not consume more food than before the study. Their average daily energy intake before the study was 10.81 MJ/d and, on average, 9.10 MJ/d in the last 4 wk of the study, despite a halved serum leptin concentration. Because one of the major functions of leptin is the suppression of appetite, these data suggest that leptin sensitivity might have been increased concomitant to the decreased serum level in these subjects. Although no data from human studies on this subject exist, it may well be that the fatty acid composition of the diet affects leptin signaling in the brain, either by effects on membrane fluidity and/or by direct effects on the expression of the leptin receptor and/or possibly also by effects on factors involved in leptin receptor signaling (reviewed in Ref. 32). On the other hand, a potential effect of the decreased serum leptin concentration on appetite might have been overlooked because of a time lag effect, given that significant decreases in leptin were not observed until visit 4, and after that point we have not assessed energy intake of the participants. Nevertheless, our data raise the hypothesis that n-3-PUFA or the ratio of n-3-PUFA to other fatty acid classes might have a special role in leptin sensitivity.
What are the implications of our findings? Jeffrey Flier (33) recently proposed that evolution is likely to have favored a leptin dose-response curve that limits leptin action at high blood leptin levels. Limited leptin action at high blood levels could best be described as leptin resistance and would lead to a more sufficient energy storage in periods of abundance and would increase the chance of survival in succeeding periods of starvation. A consequence of this hypothesis is that, in subjects with leptin levels that are high relative to body fat mass, any further increase in serum leptin concentration is more likely to result in inadequate leptin action. Thus, the risk of further weight gain and obesity would likely be higher for these subjects. This consideration is supported by a finding by Chessler et al. (34), who observed that relatively higher plasma leptin concentrations are associated with greater subsequent weight gains in Japanese Americans. Thus, decreased serum leptin concentrations, as seen in the women given the rapeseed oil diet in this study, might reduce the long-term risk of obesity in these subjects. A finding in Pima Indians indicates that low (instead of high) serum leptin concentrations predispose for weight gain (35). Seemingly, this finding is not consistent with our above hypothesis. It should be noted, however, that the subjects of this study were already morbidly obese at baseline (BMI of about 35 kg/m2), with accordingly high serum leptin concentrations. Thus, normal mechanisms of energy homeostasis are likely to have been abrogated in these subjects.
Besides its role in energy metabolism, leptin has also been shown to play important roles in other physiologic pathways. For example, high serum leptin concentrations have been implicated in the development of hypertension (reviewed in Ref. 36), coronary heart disease (37), and diabetes mellitus (38). Thus, a lower serum leptin concentration might also be beneficial, with regard to the development of these common diseases.
Taken together, in this study, we showed that, in women, a diet rich in MUFA and -linolenic acid (containing rapeseed oil as the principal source of fat) distinctly lowered serum leptin levels and apparently also increased leptin sensitivity, compared with a diet rich in SFA. This leptin- lowering potency was not shared by diets rich in monounsaturated fat (based on olive oil) or n-6-PUFA (based on sunflower oil). Thus, it is proposed that the reduction in serum leptin concentrations is a result of the relatively high amount of -linolenic acid (18:3n-3) in the rapeseed oil diet. However, uncertainties remain as to why this diet did not also lower serum leptin concentrations in the men of our study. Also, the impact of these findings is limited by the fact that serum leptin concentrations were measured ex post. Thus, the specific hypotheses based on our observations should be confirmed in future studies.
Acknowledgments
We are indebted to Dr. B. Jacobs, B. Berning, J. Harmsen, and particularly E. Gramenz and G. Klapdor for excellent technical assistance; to Drs. R. Schmidt, R. Junker, and G. Berger for performing the venipunctures; to the Bildungszentrum der Bundesfinanzverwaltung for their generous cooperation; to E. Ostermann and the Camphill Werkstaetten, Steinfurt, for supplying the oil-enriched bread and cake; to the Homann Company, Dissen, for supplying the specially manufactured margarine; and last, but not least, to the study subjects for participation.
Footnotes
This work was supported by grants from the Central Marketing Agency of the German Agricultural Industry, the German Union for the Promotion of Oil- and Protein-containing Plants, and the Broekelmann Oelmühle Company (Hamm, Germany).
Abbreviations: BMI, Body mass index; HbA1c, glycosylated hemoglobin; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids.
Received March 28, 2002.
Accepted July 25, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  K. K. Koh, S. M. Park, and M. J. Quon Leptin and Cardiovascular Disease: Response to Therapeutic Interventions Circulation, June 24, 2008; 117(25): 3238 - 3249. [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Bravata, L. Sanders, J. Huang, H. M. Krumholz, I. Olkin, C. D. Gardner, and D. M. Bravata Efficacy and Safety of Low-Carbohydrate Diets: A Systematic Review JAMA, April 9, 2003; 289(14): 1837 - 1850. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
