Recent evidences suggest intriguing relationships between several gastrointestinal (GI) and adipose tissue secreted hormones with energy balance, insulin sensitivity, type 2 diabetes and cardiovascular disease. Ghrelin is an orexigenic hormone predominantly secreted by the stomach <sup>[1Kojima M, Hosoda H, Date Y, Nakanzato M, Matsuo H, Kangawa K. Ghrelin is a growth-horome-releasing acylated peptide from stomach. Nature 1999; 402: 656-60

Cliquez ici pour aller à la section Références]</sup>. Administration of physiological concentrations of ghrelin increases food intake in rodents <sup>[2Tschop M, Smiley DL, Heiman ML. Ghrelin induces adiposity in rodents. Nature 2000; 407: 908-13

Cliquez ici pour aller à la section Références]</sup> and increases the perception of hunger and the consumption of food in humans <sup>[3Wren AM, Seal LJ, Cohen MA, et al. Ghrelin enhances appetite and increases food intake in humans. J Clin Endocrinol Metab 2001; 86: 5992

Cliquez ici pour aller à la section Références]</sup>. Continuous or repeated ghrelin administration increases body weight, decreases fat oxidation and increases adiposity in rodents <sup>[2

Cliquez ici pour aller à la section Références, 4

Cliquez ici pour aller à la section Références et 5

Cliquez ici pour aller à la section Références]</sup>. Plasma ghrelin concentration decreases with weight gain and obesity <sup>[6Shiiya T, Nakazato M, Mizuta M, et al. Plasma ghrelin levels in lean and obese humans and the effect of glucose on ghrelin secretion. J Clin Endocrinol Metab 2002; 87: 240-4

Cliquez ici pour aller à la section Références]</sup> and in humans, fasting plasma ghrelin correlates negatively with indices of adiposity: body weight, % body fat, BMI, leptin and insulin <sup>[7Tschop M, Weyer C, Tataranni PA, Devanarayan V, Ravussin E, Heiman ML. Circulating ghrelin levels are decreased in human obesity. Diabetes 2001; 50: 707-9

Cliquez ici pour aller à la section Références]</sup>.

Leptin is an anorexigenic hormone predominantly produced by white adipose tissue (WAT) <sup>[8Fruhbeck G, Gomez-Ambrosi J, Muruzabal FJ, Burrell MA. The Adipocyte: a model for integration of endocrine and metabolic signaling in energy metabolism regulation. Am J Physiol Endo Metab 2001; 280: E827-E847

Cliquez ici pour aller à la section Références]</sup>. Through its effect on the central nervous system <sup>[9

Cliquez ici pour aller à la section Références et 10

Cliquez ici pour aller à la section Références]</sup>, recombinant leptin reduces food intake and body weight in both human and mice models <sup>[11

Cliquez ici pour aller à la section Références et 12

Cliquez ici pour aller à la section Références]</sup>. Moreover, leptin is reported to increase whole body glucose utilization, decrease hepatic glycogen stores and suppress the lipogenic capacity of white adipose (for review <sup>[13Faraj M, Lu HL, Cianflone K. Diabetes, lipids, and adipocyte secretagogues. Biochem Cell Biol 2004; 82: 170-90

Cliquez ici pour aller à la section Références]</sup>). Circulating leptin concentrations are elevated in obesity <sup>[14van Dielen FM, van't Veer C, Buurman WA, Greve JW. Leptin and soluble leptin receptor levels in obese and weight-losing individuals. J Clin Endocrinol Metab 2002; 87: 1708-16

Cliquez ici pour aller à la section Références]</sup> and hyperleptinemia in obese subjects is associated with insulin resistance and high cardiovascular events <sup>[15Ren J. Leptin and hyperleptinemia - from friend to foe for cardiovascular function. J Endocrinol 2004; 181: 1-10

Cliquez ici pour aller à la section Références]</sup>.

Adiponectin is a more recently described adipocytokine that is exclusively synthesized and secreted by WAT <sup>[16

Cliquez ici pour aller à la section Références et 17

Cliquez ici pour aller à la section Références]</sup>. Adiponectin is the only WAT derived hormone whose levels are down-regulated in obesity as human and mice models of obesity and/or insulin resistant have shown reduced circulating adiponectin concentrations and expression in WAT <sup>[16

Cliquez ici pour aller à la section Références, 17

Cliquez ici pour aller à la section Références et 18

Cliquez ici pour aller à la section Références]</sup>. Elevated concentration of adiponectin is associated with improved insulin sensitivity <sup>[19

Cliquez ici pour aller à la section Références et 20

Cliquez ici pour aller à la section Références]</sup>, reduced lipidemia <sup>[21Matsubara M, Maruoka S, Katayose S. Decreased plasma adiponectin concentrations in women with dyslipidemia. J Clin Endocrinol Metab 2002; 87: 2764-9

Cliquez ici pour aller à la section Références]</sup> and reduced inflammatory markers <sup>[22

Cliquez ici pour aller à la section Références et 23

Cliquez ici pour aller à la section Références]</sup>. In fact, low plasma adiponectin has been proposed as an independent predictor of the incidence of myocardial infarctions in men <sup>[24Pistola T, Girman CJ, Hotamisligil GS, Rifai N, Hu FB. Plasma adiponectin levels and risk of myocardial infarction in men. J Am Med Assoc 2004; 291: 1730-7

Cliquez ici pour aller à la section Références]</sup>.

Despite emerging evidence of a link between ghrelin, leptin and adiponectin with CVD and type 2 diabetes, little is known about their in vivo regulation in human. Particularly, little evidence exists on how the plasma levels of these hormones may be complementarily influenced by eating behaviours, physical fitness, components of energy balance (intake and expenditure) and use of oral contraceptive pills <sup>[25Diamanti-Kandarakis E, Baillargeon JP, Iuorno MJ, Jakubowicz DJ, Nestler JE. A modern medical quandary: polycystic ovary syndrome, insulin resistance, and oral contraceptive pills. J Clin Endocrinol Metab 2003; 88: 1927-32

Cliquez ici pour aller à la section Références]</sup> in non-obese women. As recent data suggests that dysregulation of the adipocytokine axis may begin early in life <sup>[26Nemet D, Wang P, Funahashi T, et al. Adipocytokines, body composition, and fitness in children. Pediatr Res 2003; 53: 148-52

Cliquez ici pour aller à la section Références]</sup> understanding the regulation of ghrelin, adiponectin and leptin by lifestyle factors, may provide a mechanism for their voluntary modifications. This would be of vital importance in the primary prevention of obesity, diabetes and CVD in later stages of life.

To further broaden our knowledge in this area, we thus examined the association between components of energy balance, use of oral contraceptive pills and other fitness and feeding behaviours with plasma concentrations of ghrelin, adiponectin and leptin in a well-characterized cohort of young non-obese women. Herein, we hypothesize that higher energy intake and/or lower energy expenditure and physical fitness may predict unfavourable hormonal profile characterized by higher ghrelin and lower leptin and adiponectin in this population.

Sixty-three non-obese young women (18-35 years old) were recruited by announcements to participate in this study at the Université de Montréal area (Montréal, Canada). All women signed a consent form approved by the ethics committee of the Université de Montréal. The ethnic make-up of our cohort consisted of 90% Caucasians, 5% Arabs and 5% African Americans. Exclusion criteria for participation were 1) pregnancy, 2) acute illness within a month, 3) known eating disorders, 4) diabetes mellitus, 5) hypertension and 6) medication that could affect cardiovascular function and/or metabolism. All women were tested in the follicular phase of the menstrual cycle.

Venous blood samples (fasting for at least 12 hours) were collected and immediately centrifuged at 3900G for 10 min at 4ºC. Lipid profile was performed the same day and other serum samples were stored at –80ºC for subsequent analysis. Analyses were done on the COBAS INTEGRA 400 (Roche Diagnostic, Montréal, Canada) analyzer for total cholesterol, HDL-cholesterol, triglycerides and glucose combined with specific cassettes containing in vitro diagnostic reagent system. LDL-cholesterol concentration was calculated by Friedewald equation using total cholesterol, HDL-cholesterol and triglycerides <sup>[27Schectman G, Patsches M, Sasse EA. Variability in cholesterol measurements: comparison of calculated and direct LDL cholesterol determinations. Clin Chem 1996; 42: 732-7

Cliquez ici pour aller à la section Références]</sup>. Insulin level was determined by electrochemiluminescence “ECLIA” adapted for Elecsys 1010 analyzer (Insulin Elecsys Ref. 12017547 kit). Homeostasis model assessment (HOMA) was calculated according to the formula of Matthews et al. <sup>[28Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412-9

Cliquez ici pour aller à la section Références]</sup>. Serum immunoreactive total ghrelin (Phoenix Pharmaceuticals, Belmont, CA, USA) adiponectin and leptin (Linco Research, St-Charles, MO, USA) levels were measured in duplicate with a commercial radioimmunoassay (RIA) procedure using ¹²⁵I-labeled bioactive human ghrelin, adiponectin and leptin as tracers and a rabbit polyclonal antibody raised against full-length peptides. For the different RIA kits, intra- and inter-assay percent coefficients of variation (%CV) were under 10% and 15% respectively (as specified by the manufacturers).

Body weight (kg) was measured using an electronic scale (BIM, Balance Industrielles Montréal Inc., Canada) to the nearest 20 g and standing height was measured using a wall stadiometer (Perspective Enterprises, Portage, Michigan, USA) to the nearest 0.1 cm. Body mass index (BMI) was calculated as body weight in kg over height m². Fat-free mass, total fat mass, percent body fat mass, central and peripheral fat mass were evaluated by DXA (Dual Energy x-ray Absorptiometry) using a LUNAR, Prodigy system, version 6.10.019 (General Electric Lunar Corporation, Madison, Wisconsin, USA). In test-retest analyses, ICC in 18 subjects was: 0.999 for fat mass and 0.998 for fat-free mass. Waist circumference was measured with a flexible plastic metric tape at the nearest 0.5 cm. Blood pressure was determined as the average of the last four readings of five (at 1/min) in the left arm after subjects rested quietly for 5 min using a Dinamap automatic machine (Welch Allyn Inc., San Diego).

A 24-hour dietary recall was used for evaluation of total daily energy intake. The recall was directed by a trained dietician. Portion sizes were evaluated using blank models of food serving size. Energy and macronutrients intakes were calculated based on the corrected 2001b Canadian Nutrient File (CNF) <sup>[29Canada H. Canadian Nutrient File. Ottawa, Ontario, Canada: Health Canada, 2001

Cliquez ici pour aller à la section Références]</sup>.

Eating behavior was assessed by the self administrated questionnaire of Stunkard and Messick: the Three-Factor Eating Questionnaire <sup>[30Stunkard AJ, Messick S. The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. J Phychosom Res 1985; 29: 71-83

Cliquez ici pour aller à la section Références]</sup>. This 51-item questionnaire measures three dimensions of human eating behavior. The first factor measures cognitive restrained eating (dietary restraint), that is the perception that one regularly and intentionally eats less than one desires. The second factor represents tendency toward disinhibition: an incidental inability to resist eating cues, or inhibition of dietary restraint and emotional eating. The third factor examines the subjective feeling of general hunger. Every dimension is represented by a score obtained by the sum of points of each item (0 or 1). The Three-Factor Eating Questionnaire has been validated as one accurate measure of cognitive concomitants of eating behavior <sup>[30

Cliquez ici pour aller à la section Références et 31

Cliquez ici pour aller à la section Références]</sup>.

RMR was measured after a 12 hour fast by indirect calorimetry. Subjects were instructed not to exercise for at least 24 h prior to RMR testing. Concentrations of CO₂ and O₂ were measured using the ventilated hood technique with a SensorMedics Delta Track II (Datex-Ohmeda, Helsinki, Finland) and were used to determine RMR, carbohydrate and fat oxidation rates <sup>[32Cunningham JJ. Calculation of energy expenditure from indirect calorimetry: assessment of the Weir equation. Nutrition 1990; 6: 222-3

Cliquez ici pour aller à la section Références]</sup>. Subjects were instructed to: 1) fast and drink water only for 12 h before testing, 2) consume no alcohol and restrain from smoking for 24 h before testing, 3) restrain from physical activity for 24 h before testing, 4) keep physical activity to a minimum the morning of the test. Measurements were performed while subjects were lying in a supine position, without speaking or sleeping and with minimum movements. Measurements were performed during 40 minutes, the first 10 minutes were considered as an acclimatization period and the 30 last minutes were used for analyses. The gas analyzers were calibrated before every measurement for pressure and gas concentrations. The ICC for RMR determined using test-retest condition was 0.956 in our laboratory in a pilot study of 19 volunteers.

PAEE was measured by tri-axial accelerometer RT3 (TriTrac) (Stayhealthy, Monrovia, California, USA). The RT3 technique is a validated tool against the gold standard doubly labelled water that is used to measure physical activity objectively in free living individuals <sup>[33

Cliquez ici pour aller à la section Références et 34

Cliquez ici pour aller à la section Références]</sup>. The RT3 measures acceleration in the anterior-posterior (x), medio-lateral (y) and vertical (z) axis and summarizes that information as a vector magnitude. The vector was calculated as the square-root of the sum of the squared accelerations for each direction. Activity counts are given for each direction. Thereafter, activity calories per minutes were calculated with the following formula: ((activity counts/10) x (body weight x 1.692))/10000. The frequency response for the measurement of acceleration is 1 Hz and data are recorded every minute. The RT3 was worn on the right hip of the subject for three consecutive days representative of normal daily physical activity: two week-days and one weekend-day. The average of the 3 measured days is reported.

Physical fitness was assessed by measurement of aerobic capacity (VO_{2 peak}) on an ergocycle Ergoline 900 (Bitz, Germany), Ergocard (Medi Soft, Dinant, Belguim) with a cardiopulmonary exercise test station. The system was calibrated before every measurement for barometric pressure, relative humidity and gas concentrations with primary standard gasses. Gas volumes were calibrated using a 2 L syringe. Aerobic capacity was tested by a progressive test starting at 60 W with an augmentation of 40 W every three minutes. Subjects were asked to maintain a constant speed and the level of resistance on the wheel was adjusted in order to preserve a constant power output. O₂ and CO₂ were measured by a direct system using a face mask. VO_{2 peak} was achieved (average maximal oxygen consumption for 30 sec) when the power output could no longer be maintained. Heart rate was monitored during all tests using a POLAR heart rate monitor S610 (Polar Electro Oy, Kempele, Finland). A test-retest reliability trial (n = 19) for VO₂ (L/min) was performed prior to data collection and yielded an inter-assay coefficient constant (ICC) of 0.956.

Statistical analyses were performed by SPSS for Windows (versions 11.5), and significance was achieved at P ≪ 0.05. Data are presented as means ± standard deviation (SD). To examine statistical difference between women who were using oral contraceptive pills versus those who were not using the pills, one-tailed t-test was used for parametric variables and Mann-Whitney rank sum test was used for non-parametric variables. Pearson correlations were used to examine relationships between the measured parameters. Forward stepwise linear regression was used to examine the independent predictors of plasma ghrelin, adiponectin and leptin. Independent predictors considered in each model were non-collinear variables that correlated with the dependent variable of interest. Since the 3 examined hormones have strong documented association with weight and body composition, all regression analysis were performed with correction for BMI for ghrelin or fat mass for adipose-tissue derived hormones leptin and adiponectin.

We examined 63 young women with a broad range of body weights (albeit non-obese), physical and behavioural characteristics, the baseline characteristics of which are presented in tables I and II. Family history frequency of type II diabetes (19%), obesity (25%), dyslipidemia (2%) or hypertension (18%) was reported in this population. Forty six percent of the women used oral contraceptives. As compared to non-users of oral contraceptives, women who used oral contraceptive pills had significantly higher fasting plasma insulin, insulin resistance, triglyceride levels (table III). Moreover, despite lack of any effect of oral contraceptive pills on RMR, users of oral contraceptive pills were utilizing more fat and less carbohydrate as the preferential energy substrate at fasting (data not shown). There was however no significant effect of oral contraceptive pills on plasma concentrations of ghrelin, adiponectin, leptin or any other measured parameter.

To test the hypotheses that plasma concentrations of ghrelin, leptin and adiponectin may be influenced by components of energy balances, physical fitness and eating behaviours in young women, we examined the association between these parameters and the three hormones of interest (table IV). As previously published by our group <sup>[35St Pierre DH, Karelis AD, Cianflone K, et al. Relationship between ghrelin and energy expenditure in healthy young women. J Clin Endocrinol Metab 2004; 89: 5993-7

Cliquez ici pour aller à la section Références]</sup>, ghrelin was negatively correlated with 24-hour energy intake (r = -0.26, P = 0.04) and RMR (r = -0.28, P = 0.03). Ghrelin was also negatively correlated with plasma insulin (r = -0.262; P = 0.038) and HOMA (r = -0.252 P = 0.046) while positively correlated with plasma HDL-cholesterol level (r = 0.318; P = 0.011). In a stepwise forward regression analysis model, using insulin (or HOMA), HDL-cholesterol, RMR and 24-hour energy intake as independent variables, and after correcting for BMI, only plasma HDL-cholesterol and RMR remained independent predictors of plasma ghrelin, explaining 17.6% of its variation (Power 93%) (table V).

Plasma concentration of leptin was positively correlated with RMR (r = 0.33, P = 0.007) and eating disinhibition (r = 0.27, P = 0.03) while negatively correlated with physical fitness (VO_{2 peak}) (r = -0.34, P = 0.006) and with plasma adiponectin (r = -0.253; P = 0.045) (table IV). Leptin was also positively correlated with plasma insulin (r = 0.419; P ≪ 0.001) and HOMA (r = 0.437; P ≪ 0.001) and as would be anticipated, with all indices of adiposity even within this cohort of non-obese women; BMI (r = 0.604; P ≪ 0.001), total fat mass (r = 0.726; P ≪ 0.001), central fat mass (r = 0.712; P ≪ 0.001) and waist circumference (r = 0.546; P ≪ 0.001). In a stepwise forward regression analysis to predict plasma leptin where total fat mass, insulin (or HOMA), RMR, disinhibition, VO_{2 peak} and adiponectin were entered as independent variables, total fat mass explained up to 52.7% of the variation in leptin, with the other variables adding no significant power to the prediction equation (Power=100%) (table V). Similar conclusion was obtained if central fat mass was entered in the regression analysis instead of total fat mass (prediction power of 50.7%).

Finally, adiponectin was correlated negatively with energy intake per 24 hours (r = -0.38, P = 0.002) and positively with PAEE (r = 0.31; P = 0.019). In contrast to leptin, the plasma concentration of this adipose-tissue hormone correlates negatively with all indices of adiposity; BMI (r = -0.284; P = 0.024), total fat mass (r = -0.309; P = 0.014), central fat mass (r = -0.352; P = 0.005) and waist circumference (r = -0.452; P = 0.001). There was also a positive correlation between adiponectin and plasma HDL-cholesterol (r = 0.360; P = 0.004) and total cholesterol (r = -0.282; P = 0.025) and a negative one with leptin (r = -0.25, P = 0.049). A forward stepwise regression analysis was used to predict plasma adiponectin where the independent variables used were: 24-hour energy intake, PAEE, total fat mass, waist circumference and HDL. Up to 43% of the variance in plasma adiponectin concentration was predicted by a linear combination of 24-hour energy intake, waist circumference and PAEE even after correcting for total fat mass (Power 100%) (table V, figure 1).

Finally, it is important to note that there was a positive correlation between HOMA and either total fat mass (r = 0.382, P = 0.002) or central fat mass (r = 0.408, P = 0.001) in this population. Moreover, there was no correlation between PAEE and RMR in our study.

Type 2 diabetes and cardiovascular disease are leading causes of death in adult men and women in the Canadian Population <sup>[36Filate WA, Johansen HL, Kennedy CC, Tu JV. Regional variations in cardiovascular mortality in Canada. Can J Cardiol 2003; 19: 1241-8

Cliquez ici pour aller à la section Références]</sup>. The GI tract and adipose tissue-derived hormones examined in the present study, ghrelin, leptin and adiponectin, have been associated with energy balance, insulin resistance, type 2 diabetes and cardiovascular disease in many human studies. However, little is know about how their plasma concentrations may be influenced by lifestyle habits in young non-obese population. This if known, may facilitate the voluntarily regulation of ghrelin, leptin and adiponectin and aid in the primary prevention of diabetes and cardiovascular disease. Thus, we examined the association between ghrelin, leptin and adiponectin and eating habits, components of energy balance (intake and expenditure) and physical fitness in a population of well characterized cohort of young non-obese women.

Ghrelin is involved in the regulation of endocrine functions and stimulation of food intake and appetite <sup>[1

Cliquez ici pour aller à la section Références, 2

Cliquez ici pour aller à la section Références et 3

Cliquez ici pour aller à la section Références]</sup>. Plasma concentrations of ghrelin are reduced in states of positive short- and long-term energy balance, as ghrelin is reduced after food intake and in obese subjects, respectively <sup>[6

Cliquez ici pour aller à la section Références, 7

Cliquez ici pour aller à la section Références, 37

Cliquez ici pour aller à la section Références et 38

Cliquez ici pour aller à la section Références]</sup>. It is therefore suggested that the paradoxical low concentrations of ghrelin, a promoter of appetite and food intake, in obese subjects is secondary to ghrelin down regulation in a state of positive energy balance (i.e. weight gain) <sup>[6Shiiya T, Nakazato M, Mizuta M, et al. Plasma ghrelin levels in lean and obese humans and the effect of glucose on ghrelin secretion. J Clin Endocrinol Metab 2002; 87: 240-4

Cliquez ici pour aller à la section Références]</sup>. There was no correlation between ghrelin and indices of adiposity in our study, likely secondary to the narrow range of weight selected (non-obese). Yet, ghrelin was negatively associated with RMR in non-obese women independent of body weight. Therefore, we hypothesize that reduced ghrelin levels in obese subjects may not be solely secondary to positive energy balance but to the increased RMR that accompanies the increase in body size.

Moreover, the results in this study raise the possibility that circulating plasma ghrelin may be regulated by voluntary long-term behaviour modification that increases RMR, such as aerobic and resistance training (for review <sup>[39Speakman JR, Selman C. Physical activity and resting metabolic rate. Proc Nutr Soc 2003; 62: 621-34

Cliquez ici pour aller à la section Références]</sup>). In the only study that examined the effect of aerobic exercise (12 months intervention) on plasma ghrelin in human (albeit in post-menopausal overweight sedentary women), there was no effect of exercise per se on plasma ghrelin, independent of its effect on reducing body weight <sup>[40Foster-Schubert KE, McTiernan A, Frayo RS, et al. Human plasma ghrelin levels increase during a one-year exercise program. J Clin Endocrinol Metab 2005; 90: 820-5

Cliquez ici pour aller à la section Références]</sup>. This study however did not measure the change in RMR nor present the change in body composition, if ever, in response to the exercise intervention. Moreover, it is documented that weight loss which reduces RMR does indeed increase circulating ghrelin <sup>[41

Cliquez ici pour aller à la section Références et 42

Cliquez ici pour aller à la section Références]</sup>, which is in agreement with our hypothesis. This has to be however distinguished from weight loss induced by gastric bypass surgery which is associated with suppressed plasma ghrelin secondary to bypassing much of the stomach <sup>[41

Cliquez ici pour aller à la section Références et 43

Cliquez ici pour aller à la section Références]</sup>, the primary site of ghrelin secretion <sup>[1Kojima M, Hosoda H, Date Y, Nakanzato M, Matsuo H, Kangawa K. Ghrelin is a growth-horome-releasing acylated peptide from stomach. Nature 1999; 402: 656-60

Cliquez ici pour aller à la section Références]</sup>. However, it is impossible to distinguish an independent effect of RMR on plasma ghrelin in these studies (i.e. independent of the effect of change in body weight on ghrelin). Finally, although in our study there was no correlation between ghrelin and PAEE, PAEE did not correlate with RMR either. It is important to point out however that PAEE in our study is a measure of total daily physical activity and not of any specific training program that could have changed RMR. Prospective studies are therefore needed to examine whether plasma ghrelin is indeed regulated by modifying RMR, particularly while holding body weight constant.

Plasma ghrelin was also predicted by HDL levels in our study, an association that has been previously reported in human studies <sup>[44Purnell JQ, Weigle DS, Breen P, Cummings DE. Ghrelin levels correlate with insulin levels, insulin resistance, and high-density lipoprotein cholesterol, but not with gender, menopausal status, or cortisol levels in humans. J Clin Endocrinol Metab 2003; 88: 5747-52

Cliquez ici pour aller à la section Références]</sup>. Although it is not yet established how HDL-cholesterol may influence plasma ghrelin concentrations, a role of HDL-cholesterol in the transport of circulating ghrelin has been postulated <sup>[45Beaumont NJ, Skinner VO, Tan TMM, et al. Ghrelin can bind to a species of high density lipoprotein associated with paraoxonase. J Biol Chem 2003; 278: 8877-80

Cliquez ici pour aller à la section Références]</sup>. As weight loss is known to increase plasma HDL-cholesterol <sup>[46Poobalan A, Aucott L, Smith WCS, et al. Effects of weight loss in overweight/obese individuals and long-term lipid outcomes - a systematic review. Obesity Reviews 2004; 5: 43-50

Cliquez ici pour aller à la section Références]</sup>, it would be interesting to examine whether the putative actions of exercise and/or weight loss on ghrelin are related to their effect on HDL-cholesterol.

Although both are adipose-tissue derived hormones, the results in our study point to different mechanisms regulating plasma concentrations of leptin and adiponectin in young women. Plasma leptin correlated with RMR, physical fitness, and food disinhibition. However, the relation between leptin and all other energy balance components was abolished once fat mass, whether total or central, was accounted for. This suggests that behavioral modifications attempting to modify plasma leptin would succeed only if fat mass was changed in the process. Another important point to note is that, under normal circumstances, leptin is proposed to signal the abundance of adipose store to the hypothalamus to limit energy intake, increase energy expenditure and decrease body weight leading to favourable metabolic outcomes of increased glucose utilization and insulin sensitivity <sup>[11

Cliquez ici pour aller à la section Références, 15

Cliquez ici pour aller à la section Références, 47

Cliquez ici pour aller à la section Références, 48

Cliquez ici pour aller à la section Références et 49

Cliquez ici pour aller à la section Références]</sup>. Yet, hyperleptinemia in obese subjects is associated with hyperphagia, insulin resistance, hyperlipidemia, and hypertension <sup>[15Ren J. Leptin and hyperleptinemia - from friend to foe for cardiovascular function. J Endocrinol 2004; 181: 1-10

Cliquez ici pour aller à la section Références]</sup>. This leads to the hypothesis that, like in the case of the insulin pathway, most obese subjects are leptin resistant <sup>[50Caro JF, Sinha MK, Kolaczynski JW, Zhang PL, Considine RV. Leptin: the tale of an obesity gene. Diabetes 1996; 45: 1455-62

Cliquez ici pour aller à la section Références]</sup>. In the present study, although the young women examined were non-obese, higher leptin was also associated with decreased insulin sensitivity and physical/aerobic capacity. This raises the possibility that, even in the absence of an obesity state, leptin sensitivity may be decreased with increased adiposity in humans.

On the other hand, in contrast to leptin, correcting for fat mass did not abolish the correlation of adiponectin with the components of energy balance: energy intake or PAEE. In fact we report here, for the first time, that 36% of circulating adiponectin may be predicted by energy intake and physical activity independent of fat mass or central fat mass. This novel finding suggests that adiponectin concentrations may be voluntarily augmented by increasing physical activity and/or to a greater extent by reducing food intake. Higher adiponectin is associated with improved insulin sensitivity <sup>[19

Cliquez ici pour aller à la section Références et 20

Cliquez ici pour aller à la section Références]</sup>, reduced lipidemia <sup>[21Matsubara M, Maruoka S, Katayose S. Decreased plasma adiponectin concentrations in women with dyslipidemia. J Clin Endocrinol Metab 2002; 87: 2764-9

Cliquez ici pour aller à la section Références]</sup> and reduced inflammatory markers <sup>[22

Cliquez ici pour aller à la section Références et 23

Cliquez ici pour aller à la section Références]</sup>. Reduced adiponectin levels have been shown to be an independent predictor of the development of diabetes in prospective studies with predisposed adults <sup>[51

Cliquez ici pour aller à la section Références et 52

Cliquez ici pour aller à la section Références]</sup> and children <sup>[53Cruz M, Garcia-Macedo R, Garcia-Valerio Y, et al. Low adiponectin levels predict type 2 diabetes in Mexican children. Diabetes Care 2004; 27: 1451-3

Cliquez ici pour aller à la section Références]</sup>. Efforts to increase plasma concentration of this potential anti-atherogenic anti-diabetic hormone should therefore be encouraged.

Some limitations of this study should however be taken into consideration when analyzing the data. Considering the cross sectional design of our study, the correlative nature of this study does not allow for a cause and effect relationship to be assumed. Underreporting of food intake is a documented limitation of 24-hour dietary recall but is usually less important in young non-obese women <sup>[54Tran KM, Johnson RK, Soultanakis RP, Matthews DE. In-person vs telephone-administered multiple-pass 24-hour recalls in women: validation with doubly labeled water. J Am Diet Assoc 2000; 100: 777-83

Cliquez ici pour aller à la section Références]</sup>. PAEE is more accurately measured with doubly labelled water than with questionnaires and insulin sensitivity with euglycemic hyperinsulinemic clamp than with HOMA thus our results should be confirmed with these gold standard methods. Present results only take into consideration total adiponectin and ghrelin levels, further investigations should examine the importance of the presence of high and low molecular weight adiponectin multimers and the presence of acylated vs non-acylated ghrelin circulating isoforms.

In conclusion, higher energy intake and lower physical activity, are independent predictors of an lower adiponectin profile, whereas lower RMR independently predicts higher circulating ghrelin in young non-obese women. This hormonal profile could promote increased appetite and energy intake and, thus promote complications such as type 2 diabetes mellitus. Further studies, particularly prospective, are needed to examine whether ghrelin and adiponectin can be voluntarily regulated by modifying either arm of the energy balance equation, which could promote favourable metabolic outcome irrespective of whether weight loss is achieved in the process.