Plasma Adiponectin Increases Postprandially in ... - Wiley Online Library

Plasma Adiponectin Increases Postprandially in Obese, but not in Lean, Subjects Patrick J. English,* Steven R. Coughlin,* Katharine Hayden,† Iqbal A. Malik,* and John P.H. Wilding*

Abstract ENGLISH, PATRICK J., STEVEN R. COUGHLIN, KATHARINE HAYDEN, IQBAL A. MALIK, AND JOHN P.H. WILDING. Plasma adiponectin increases postprandially in obese, but not in lean, subjects. Obes Res. 2003;11: 839 – 844. Objective: We investigated the acute responses of plasma adiponectin levels to a test meal in lean and obese subjects. Research Methods and Procedures: We studied 13 lean and 11 obese subjects after a 10-hour overnight fast. Glucose, insulin, and adiponectin concentrations were measured at baseline and 15, 30, 60, 120, and 180 minutes after a fixed breakfast. Results: At baseline, fasting adiponectin concentrations were lower in the obese group vs. the lean group [mean (95% confidence interval): 2.9 (2.1 to 4.1) ␮g/mL vs. 8.6 (6.5 to 11.3) ␮g/mL], but rose 4-fold postprandially in the obese group, reaching a peak at 60 minutes [baseline: 2.9 (2.1 to 4.1) ␮g/mL vs. 60 minutes: 12.1 (8.5 to 17.4) ␮g/mL; p ⬍ 0.0001] and remaining elevated for the remainder of the study. There were no postprandial changes in plasma adiponectin concentrations in lean subjects. Discussion: This increase of adiponectin concentrations in obese individuals might have important beneficial effects on postprandial glucose and lipid metabolism and might be viewed as a mechanism for maintaining normal glucose tolerance in those who are obese and insulin resistant. Key words: adiponectin, insulin resistance, heart rate variability, postprandial

Received for review December 31, 2002. Accepted in final form May 20, 2003. *The University of Liverpool Diabetes and Endocrinology Research Group, Clinical Sciences Centre, and †Department of Biochemistry, University Hospital Aintree, Liverpool, United Kingdom. Address correspondence to Dr. John P.H. Wilding, Diabetes and Endocrinology Research Group, Clinical Sciences Centre, University Hospital Aintree, Lower Lane, Liverpool L9 7AL, United Kingdom. E-mail: [email protected] Copyright © 2003 NAASO

Introduction Adiponectin is a complement-related 244 amino-acid protein produced exclusively by white adipose tissue (1,2). It belongs to the family of adipokines, which are a group of proteins secreted from adipose tissue that may modulate the effects of obesity on health. Plasma concentrations of adiponectin are decreased in obese individuals (3), an effect reversible by weight loss (4), and are inversely correlated with insulin resistance (5). Administration of adiponectin, returning it to normal physiological levels in rodent models of obesity and lipoatrophy, improves glucose and lipid metabolism (6). These data suggest that adiponectin is an important regulator of insulin sensitivity in humans. A further role for adiponectin as a vascular protection protein is implied by studies with cultured human aortic endothelial cells, showing that adiponectin accumulates in injured vessel walls (7) and dose-dependently inhibits tumor necrosis factor-␣-induced monocyte attachment (8,9), the phagocytic activity of macrophages, and lipopolysaccharide-induced TNF-␣ production by macrophages (10). Additionally, adiponectin concentrations are clearly decreased in the presence of type 2 diabetes (11) and vascular disease (8), the effect of these conditions being additive, and this provides a possible explanation for the association between obesity and both insulin resistance and increased incidence of vascular disease. Although insulin resistance alone is a predictor of vascular disease and mortality, recent evidence shows that increased prandial glucose excursions, independent of fasting glucose concentrations, are associated with increased mortality (12,13). Therefore, an individual with increased insulin resistance but normal prandial glucose excursions is at lower risk for vascular disease or death than an individual with increased insulin resistance and increased prandial glucose excursions. We speculated that, given its properties, adiponectin might play a role in this phenomenon. For this to be true, adiponectin should be prandially regulated, and this prandial regulation should be related to prandial glucose regulation. Although one study has looked at adiponectin concentration at baseline and 2 hours postprandially in a OBESITY RESEARCH Vol. 11 No. 7 July 2003

839

Adiponectin Increases after Feeding in Obese Subjects, English et al.

small number of subjects (11), no study has examined in detail the acute postprandial regulation of adiponectin secretion.

Research Methods and Procedures We compared the acute postprandial response of plasma adiponectin in lean and obese subjects to determine whether feeding alters plasma adiponectin concentrations that could influence postprandial glucose metabolism. The study was approved by the South Sefton Research Ethics Committee (project registration number EC.04.99) and was performed in accordance with the principles of the Declaration of Helsinki. Volunteers gave written informed consent, took no regular medication, and had a normal physical examination and electrocardiogram (ECG).1 All subjects were normotensive, euthyroid, and had normal fasting plasma glucose (⬍6.0 mM), urea, and electrolytes. We studied 13 lean [BMI (mean ⫾ SEM): 22.5 ⫾ 0.7 kg/m2) and 11 obese [BMI (mean ⫾ SEM): 41.6 ⫾ 3.6 kg/m2] subjects recruited from a weight management clinic and by advertisement. Subjects were asked to fast, drink only water, avoid smoking from 10:00 PM on the night before the study, and refrain from strenuous exercise or alcohol in the 24 hours preceding the study day. Protocol The study methods have been published previously (14), although one obese subject who was excluded from our previous study because of inadequate lithium heparin samples for ghrelin estimation was able to be included for the purposes of the present study because of sufficient EDTA plasma samples. Because our heart rate variability (HRV) data were not reported previously, the methods used for HRV determination are outlined here. Subjects voided urine immediately before anthropometric measurements, which were taken at 8:30 AM. They were weighed and measured wearing light clothing and without shoes. Waist circumference was measured at the level of the umbilicus and hip circumference at the level of the greater trochanters. Body fat percentage (BFP) was estimated by whole-body bioelectrical impedance analysis (Tanita Systems, Stokie, IL). A cannula was then inserted into a distal forearm or hand vein for collection of samples. Samples were collected into plastic EDTA, lithium heparin, and serum separator tubes. Serum samples were allowed to stand for 15 minutes before centrifugation at ⫺4 °C, whereas plasma samples were separated immediately. All samples were stored at ⫺80 °C until assayed. Male subjects 1 Nonstandard abbreviations: ECG, electrocardiogram; HRV, heart rate variability; BFP, body fat percentage; SNS, sympathetic nervous system; LF, low frequency; HF, high frequency; CI, confidence interval; HOMA-IR, homeostatic modeling assessment to estimate insulin resistance; AUC, area under the curve.

840

OBESITY RESEARCH Vol. 11 No. 7 July 2003

consumed a standard mixed meal containing 714 kcal with 56.8% calories as carbohydrate, 12.2% as protein, and 31% as fat. To correct for the lower energy requirements of women, female subjects received a smaller meal of 551 kcal, but with the same macronutrient composition. Blood samples for insulin, glucose, and adiponectin were collected at baseline and 15, 30, 60, 120, and 180 minutes after the meal. HRV Because there is evidence to suggest that ␤-adrenergic stimulation inhibits adiponectin gene expression in adipocytes (15), we considered it appropriate to use a measure of sympathetic nervous system (SNS) activity and assess its association with changes in adiponectin levels. HRV (the measure of the beat-to-beat variation in heart rate) (16,17) is a useful and safe tool to measure autonomic nervous system activity and provides comprehensive quantitative and qualitative evaluation of both parasympathetic and SNS activity (18,19); HRV was used to assess SNS activity in this study. HRV was measured at baseline and then between 45 and 60, 105 and 120, and 165 and 180 minutes postprandially. The room in which experiments were conducted was temperature controlled to 21 °C, and all HRV measurements were taken with subjects lying in a supine position, with a head-upward tilt of 30° and the room darkened to avoid any other stimulus interfering with measurements. Breathing was standardized to 12 breaths per minute using auditory command software and confirmed using a Pneumotrace respiration transducer (World Precision Instruments Ltd, Aston, United Kingdom). Fifteen minutes of continuous ECG data were recorded into the Acqknowledge software package (World Precision Instruments Ltd) using an MP100 ECG acquisition module (BIOPAC Systems Inc., Santa Barbara, CA) with three ECG limb leads. Spectral analysis of the ECG data was performed in Powermedic (Okiimura Taiwan CO. Ltd, Taiwan) using a fast Fourier transformation. Power spectral measurements were expressed as absolute units (ms2). Total power is defined as the sum of absolute, very low frequency (not shown), low frequency (LF), and high frequency (HF) power and is considered to be an indirect measure of cardiac autonomic tone. LF and HF were defined as the power in the spectral bands between 0.04 and 0.15 and 0.15 and 0.4Hz, representing sympathetic and parasympathetic activity and very low frequency as the power spectral band from 0 to 0.04Hz. The LF-to-HF ratio is considered an indirect measure of sympathovagal balance. Matsumoto and colleagues have worked to identify the specific power spectral peaks associated with thermoregulation, including postprandial thermogenesis (20 –22). These peaks theoretically represent the portion of SNS activity responsible for thermogenesis and have been identified as a band between 0.007and 0.035 Hz, which the authors call the VLO frequency component


Table 1. Baseline characteristics and summary measures of glucose, insulin, and adiponectin responses in lean and obese subjects

Age (years) [median (IQR)] Sex (male/female) BMI, kg/m2 [median (IQR)] Waist, cm [median (IQR)] BFP (mean ⫾ SD) Fasting glucose, mM (mean ⫾ SD) HOMA-IR Fasting adiponectin, ␮g/mL [mean (95% CI)] AUC insulin, mU 䡠 min/L [median (IQR)] AUC glucose, mM 䡠 min/L [median (IQR)] AUC adiponectin, ␮g 䡠 min/mL [median (IQR)]

Lean (n ⴝ 13)

Obese (n ⴝ 11)

p value for difference

28 (18.5) 10/3 22.5 (5.2) 78 (15) 19.6 ⫾ 2.4 4.2 ⫾ 0.8 0.98 ⫾ 0.21 8.6 (6.5 to 11.3) 2477 (3320) 163 (188) ⫺0.8 (8.3)

44 (15) 7/4 37.2 (24.4) 110 (45) 39.5 ⫾ 4.9 4.9 ⫾ 0.3 3.35 ⫾ 0.61 2.9 (2.1 to 4.1) 9890 (7976) 245 (97) 22.8 (14.2)

0.015 ⬍0.0001 ⬍0.0001 0.0009 0.07 0.0002 ⬍0.0001 ⬍0.0001 0.28 0.0007

IQR, interquartile range.

(20,22). They have shown these peaks to increase postprandially in young women, although the response was decreased in the obese subjects (22). Because this frequency component was that most likely to be altered by food ingestion, we included it as part of our analysis. Assays Assays were performed using commercially available methods. Adiponectin was determined using radioimmunoassay (Linco Research, Inc., St. Charles, MO), insulin by a chemiluminescence assay on the IMMULITE 2000 system (Diagnostic Products Corporation-UK, Llanberis, Gwynnedd, United Kingdom), and glucose by the glucose hexokinase method using the ADVIA 1650 system (Bayer UK Ltd, Newbury, United Kingdom). Statistical Analyses HRV indexes, adiponectin, and insulin concentrations were log-transformed for statistical analyses; therefore, data are presented as the geometric means and 95% confidence intervals (CIs) for those means. Comparisons between groups were made using paired and unpaired t tests or Mann-Whitney U tests as appropriate. We used StatsDirect statistical software version 1.9.7 for all statistical calculations (Buchan I. StatsDirect statistical software. http//:www. statsdirect.com. 2002. England: CamCode). We used the homeostatic modeling assessment to estimate insulin resistance (HOMA-IR). The area under the curve (AUC) values for adiponectin, insulin, and glucose responses were determined by calculating the total AUC using triangulation and then subtracting the area subtended by the baseline concen-

tration of the hormone over the duration of the study. These AUC values were used as the summary measures of the size of the adiponectin, glucose, and insulin responses included in the regression analyses. We initially performed simple linear regression analyses, correlating the AUC adiponectin with the baseline adiponectin, HOMA-IR, BMI, waist circumference, BFP, AUC insulin, age, AUC glucose, AUC LF, AUC LF/HF, and AUC VLO because these all were factors considered to be potentially important in determining the adiponectin response. Those factors that were significant predictors of the adiponectin response using simple linear regression were then included in a multiple linear regression model. Because baseline adiponectin correlates with insulin resistance and HOMA-IR, we performed multiple regression analysis with and without baseline adiponectin concentrations to see whether this had an important effect on results. It was also determined that we should correct for BMI, glucose, and insulin in a final regression model, regardless of the strength of their individual predictive value because of their potential importance in determining the adiponectin response.

Results At baseline, fasting adiponectin concentrations were lower in the obese group vs. the lean group (Table 1) [mean (95% CI): 2.9 (2.1 to 4.1) ␮g/mL vs. 8.6 (6.5 to 11.3) ␮g/mL] and were negatively correlated with HOMA-IR (r ⫽ ⫺0.59, p ⫽ 0.0026), BMI (r ⫽ ⫺0.52, p ⫽ 0.01), waist circumference (r ⫽ ⫺0.60, p ⫽ 0.0018), and age (r ⫽ ⫺0.47, p ⫽ 0.021). After the meal, adiponectin concentrations rose 4-fold in the obese group, reaching a peak at 60 OBESITY RESEARCH Vol. 11 No. 7 July 2003

841


Figure 1: Postprandial adiponectin concentrations in lean and obese subjects. Concentrations rose 4-fold postprandially in the obese subjects, but remained unaltered in the lean subjects. 䡺, obese subjects; E, lean subjects.

minutes [baseline: 2.9 (2.1 to 4.1) ␮g/mL vs. 60 minutes: 12.1 (8.5 to 17.4) ␮g/mL; p ⬍ 0.0001] and remaining elevated for the remainder of the study (Figure 1). There were no postprandial changes in plasma adiponectin concentrations in lean subjects. The increase in adiponectin concentrations was partly predicted by age (r ⫽ 0.57, p ⫽ 0.0035), HOMA-IR (r ⫽ 0.48, p ⫽ 0.018), baseline adiponectin concentrations (r ⫽ ⫺0.84, p ⬍ 0.0001) and AUC glucose (r ⫽ 0.43, p ⫽ 0.034), but by none of the other factors examined. When these factors were examined using multiple linear regression, the baseline adiponectin concentration was the only significant predictor of the subsequent adiponectin response (p ⫽ 0.0003). This remained significant after correction for the size of the insulin response and BMI (p ⫽ 0.0006; multiple correlation coefficient: R ⫽ 0.87, R2 ⫽ 76.5%). If baseline adiponectin concentrations were left out of the multiple regression equation, then age

was the only significant predictor of the adiponectin response (p ⫽ 0.0098), and, similarly, this retained significance after correction for BMI and the insulin response (p ⫽ 0.0225, multiple correlation coefficient R ⫽ 0.72, R2 ⫽ 52.1%). The much-improved strength of the correlation when baseline adiponectin concentrations were used suggests that this is the single most important predictive factor for the adiponectin response. There was no difference between the groups in LF, LF/ HF, or VLO at baseline. LF, LF/HF, and VLO power increased after the meal in both groups (Table 2), although there was no significant difference between the groups in terms of the size of this response and no correlation with the size of the adiponectin response. None of the other parameters measured changed significantly after the meal (data not shown).

Discussion These results are intriguing. The negative correlation between adiponectin concentrations and measures of adiposity and insulin resistance has been well described, but the postprandial increase in obese subjects, absent in lean volunteers, has not previously been demonstrated. This increase is certainly convincing because it was both large (a 4-fold increase, on average) and occurred in 8 of 11 subjects by 60 minutes. The cause and purpose of this increase, however, remain to be elucidated. A systematic measurement artifact affecting results from the obese group might give rise to these findings. One factor that might cause this would be interference with the assay as a result of exaggerated postprandial hypertriglyceridemia. The manufacturers of the assay are unaware of any such interaction or of any other hormone or substance that might interfere with the assay in this way. It is unlikely, however, that a variable causing assay interference would so specifically affect the obese group postprandially.

Table 2. Baseline HRV parameters and the AUC for their responses to feeding in lean and obese subjects HRV parameters LF baseline (ms2) LF/HF baseline VLO baseline (ms2) AUC LF (ms2 h) AUC LF/HF (hours) AUC VLO (ms2 h)

Lean

Obese

p value for difference

1058 (120 to 9368) 0.83 (0.12 to 5.66) 865 (60 to 12,296) 2.4 ⫻ 109 (13.8 ⫻ 106 to 4.3 ⫻ 1011) 1.2 (0.0 to 313) 2.0 ⫻ 109 (2.9 ⫻ 106 to 1.3 ⫻ 1012)

726 (199 to 2652) 1.03 (0.09 to 12.08) 612 (150 to 2494) 6.9 ⫻ 108 (2.4 ⫻ 107 to 2.0 ⫻ 1010) 4.5 (0 to 977) 3.8 ⫻ 108 (1.2 ⫻ 107 to 1.2 ⫻ 1010)

0.35 0.66 0.46 0.20 0.28 0.17

Data were log transformed for statistical analyses and are, therefore, presented as geometric means and 95% confidence intervals for the mean. The very broad confidence intervals should be noted. ms2 represents absolute unit(s).

842



Figure 2: Postprandial insulin response in lean and obese subjects. Peak insulin levels were three times higher in the obese vs. lean subjects, and the AUC for the insulin response was increased 4-fold. 䡺, obese subjects; E, lean subjects.

Figure 3: Postprandial glucose response in lean and obese subjects. There was no significant difference between the lean and the obese groups, despite markedly different measures of insulin resistance. 䡺, obese subjects; E, lean subjects.

Despite the lack of a formal statistical correlation, it is tempting to speculate that the larger increase in plasma insulin seen postprandially (Figure 2) in the obese subjects at least partly mediates the increase in adiponectin concentrations. Studies on human adipose tissue demonstrate an up-regulation of adiponectin gene expression by insulin consistent with this concept (23). In related studies, ␤-adrenergic stimulation of 3T3-L1 adipocytes led to a down-regulation of adiponectin gene expression (15), although studies on peptide release have not been reported. It is possible that in obese subjects with increased sympathetic nervous stimulation of adipose tissue, downregulation of adiponectin expression and reduced plasma adiponectin concentrations occur, leading to insulin resistance and hyperinsulinemia. Initially hyperinsulinemia may partly counteract the effects of ␤-adrenergic stimulation of the adipocytes. Postprandially, epinephrine concentrations decrease, whereas plasma norepinephrine turnover in adipose tissue increases (24). We would, therefore, expect decreased ␤-adrenergic stimulation of adipocytes (with lower epinephrine concentrations) and an increase in plasma insulin concentrations, thus stimulating adiponectin production. This response would be exaggerated in those who are obese and insulin resistant and might lead to the increase in plasma adiponectin concentrations seen in our study. This could be viewed as a mechanism for maintaining normal glucose tolerance in those who are obese and insulin resistant (Figure 3). There are, however, difficulties with this explanation. First, Yu et al. showed that adiponectin concentrations were suppressed by 20% in humans and 50% in rats with a hyperinsulinemic euglycemic clamp (25), suggesting that hyperinsulinemia decreases adiponectin secretion and/or production from the adipocyte. These data seem robust and

have the advantage of being performed in vivo, in contrast to those studies conducted by Halleux et al. (23), Bogan and Lodish (26), and Fasshauer et al. (27). They have also highlighted the discrepancies in the results of experiments that investigated the effect of insulin on adiponectin expression and secretion (23,26,27), making an explanation of our findings even more difficult. Secondly, we did not show any convincing differences in the SNS response between the two groups, which might imply a withdrawal of sympathetic stimulation of adipose tissue in the obese subjects. However, measurement of HRV may be insufficiently sensitive to detect changes in SNS activity that are specific to adipose tissue; measurement of norepinephrine turnover (23) has been previously used to assess this postprandially and might be an appropriate technique for future studies. It is possible, of course, that a further, unexamined factor is involved in increasing adiponectin secretion and/or production postprandially in the obese group. There are a large number of gut hormones and peptides that change acutely postprandially and that could potentially influence the adiponectin response to feeding; however, few of them have been studied in terms of their effects on adiponectin secretion. Identifying those with differential responses in lean and obese subjects might suggest a further candidate for examination, but that hormone or peptide currently eludes us, and it is, therefore, clear that reconciling our findings with those of previous studies requires further work, both to verify the findings and to investigate the possible underlying mechanisms in more detail. In summary, this preliminary study has shown that plasma adiponectin concentrations increase acutely after a mixed meal in otherwise healthy obese subjects. Given the properties of adiponectin, this increase might have imporOBESITY RESEARCH Vol. 11 No. 7 July 2003

843


tant beneficial effects on postprandial glucose and lipid metabolism, although more work is needed to elucidate the mechanisms underlying this response and to investigate whether it is attenuated in the progression through impaired glucose tolerance to type 2 diabetes.

Acknowledgments

14.

15.

There was no outside funding/support for this study. References 1. Maeda K, Okubo K, Shimomura I, Funahashi T, Matsuzawa Y, Matsubara K. cDNA cloning and expression of a novel adipose specific collagen-like factor, apM1 (AdiPose Most abundant Gene transcript 1). Biochem Biophys Res Commun. 1996;221:286 –9. 2. Scherer PE, Williams S, Fogliano M, Baldini G, Lodish HF. A novel serum protein similar to C1q, produced exclusively in adipocytes. J Biol Chem. 1995;270:26746 –9. 3. Hu E, Liang P, Spiegelman BM. AdipoQ is a novel adiposespecific gene dysregulated in obesity. J Biol Chem. 1996;271: 10697–703. 4. Yang WS, Lee WJ, Funahashi T, et al. Weight reduction increases plasma levels of an adipose-derived anti-inflammatory protein, adiponectin. J Clin Endocrinol Metab. 2001;86: 3815–9. 5. Weyer C, Funahashi T, Tanaka S, et al. Hypoadiponectinemia in obesity and type 2 diabetes: close association with insulin resistance and hyperinsulinemia. J Clin Endocrinol Metab. 2001;86:1930 –5. 6. Yamauchi T, Kamon J, Waki H, et al. The fat-derived hormone adiponectin reverses insulin resistance associated with both lipoatrophy and obesity. Nat Med. 2001;7:941– 6. 7. Okamoto Y, Arita Y, Nishida M, et al. An adipocyte-derived plasma protein, adiponectin, adheres to injured vascular walls. Horm Metab Res. 2000;32:47–50. 8. Ouchi N, Kihara S, Arita Y, et al. Novel modulator for endothelial adhesion molecules: adipocyte-derived plasma protein adiponectin. Circulation. 1999;100:2473– 6. 9. Ouchi N, Kihara S, Arita Y, et al. Adiponectin, an adipocyte-derived plasma protein inhibits endothelial NF-kappaB signaling through a cAMP-dependent pathway. Circulation. 2000;102:1296 –301. 10. Yokota T, Oritani K, Takahashi I, et al. Adiponectin, a new member of the family of soluble defense collagens, negatively regulates the growth of myelomonocytic progenitors and the functions of macrophages. Blood. 2000;96:1723–32. 11. Hotta K, Funahashi T, Arita Y, et al. Plasma concentrations of a novel, adipose-specific protein, adiponectin, in type 2 diabetic patients. Arterioscler Thromb Vasc Biol. 2000;20: 1595–9. 12. Bonora E, Muggeo M. Postprandial blood glucose as a risk factor for cardiovascular disease in type II diabetes: the epidemiological evidence. Diabetologia. 2001;44:2107–14. 13. The DECODE Study Group. Glucose tolerance and mortality: comparison of WHO and American Diabetes Association

844


16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

diagnostic criteria. The DECODE study group. European Diabetes Epidemiology Group. Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe. Lancet 1999; 354:617–21. English PJ, Ghatei MA, Malik IA, Bloom SR, Wilding JP. Food fails to suppress ghrelin levels in obese humans. J Clin Endocrinol Metab. 2002;87:2984 –7. Fasshauer M, Klein J, Neumann S, Eszlinger M, Paschke R. Adiponectin gene expression is inhibited by beta-adrenergic stimulation via protein kinase A in 3T3–L1 adipocytes. FEBS Lett. 2001;507:142– 6. Young JB, Macdonald IA. Sympathoadrenal activity in human obesity: heterogeneity of findings since 1980. Int J Obes Relat Metab Disord. 1992;16:959 – 67. Macdonald IA. Advances in our understanding of the role of the sympathetic nervous system in obesity. Int J Obes Relat Metab Disord. 1995;19(Suppl 7):S2–7. Akselrod S, Gordon D, Ubel FA, Shannon DC, Berger AC, Cohen RJ. Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to-beat cardiovascular control. Science. 1981;213:220 –2. Akselrod S, Gordon D, Madwed JB, Snidman NC, Shannon DC, Cohen RJ. Hemodynamic regulation: investigation by spectral analysis. Am J Physiol. 1985;249(4 Pt 2):H867–75. Matsumoto T, Miyawaki T, Ue H, Kanda T, Zenji C, Moritani T. Autonomic responsiveness to acute cold exposure in obese and non-obese young women. Int J Obes Relat Metab Disord. 1999;23:793– 800. Matsumoto T, Miyawaki C, Ue H, Yuasa T, Miyatsuji A, Moritani T. Effects of capsaicin-containing yellow curry sauce on sympathetic nervous system activity and diet-induced thermogenesis in lean and obese young women. J Nutr Sci Vitaminol (Tokyo). 2000;46:309 –15. Matsumoto T, Miyawaki C, Ue H, Kanda T, Yoshitake Y, Moritani T. Comparison of thermogenic sympathetic response to food intake between obese and non-obese young women. Obes Res. 2001;9:78 – 85. Halleux CM, Takahashi M, Delporte ML, et al. Secretion of adiponectin and regulation of apM1 gene expression in human visceral adipose tissue. Biochem Biophys Res Commun. 2001; 288:1102–7. Patel J, Coppack SW, Rawesh A, Miles JM, Eisenhofer G. Regional and total body changes in sympathoadrenal activity in response to eating. Int J Obes Relat Metab Disord. 2000; 24(Suppl 2):S156 –7. Yu JG, Javorschi S, Hevener AL, et al. The Effect of thiazolidinediones on plasma adiponectin levels in normal, obese, and type 2 diabetic subjects. Diabetes. 2002;51:2968 – 74. Bogan JS, Lodish HF. Two compartments for insulin-stimulated exocytosis in 3T3–L1 adipocytes defined by endogenous ACRP30 and GLUT4. J Cell Biol. 1999;146:609 –20. Fasshauer M, Klein J, Neumann S, Eszlinger M, Paschke R. Hormonal regulation of adiponectin gene expression in 3T3–L1 adipocytes. Biochem Biophys Res Commun. 2002; 290:1084 –9.