0021-972X/99/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 1999 by The Endocrine Society
Vol. 84, No. 10 Printed in U.S.A.
Leptin Response to Insulin in Humans Is Related to the Lipolytic State of Abdominal Subcutaneous Fat* J. S. FISHER, R. C. HICKNER†, S. B. RACETTE, E. F. BINDER, M. LANDT, W. M. KOHRT
AND
Departments of Internal Medicine (J.S.F., R.C.H., S.B.R., E.F.B., W.M.K.) and Pediatrics (M.L.), Washington University School of Medicine, St. Louis, Missouri 63110 ABSTRACT Insulin-induced leptinemia in humans appears to be blunted by insulin resistance. We therefore examined the relationship between insulin action and plasma leptin by monitoring regional and whole body lipolysis and plasma leptin levels in 15 premenopausal women (body fat range, 14 –59%) during a two-stage euglycemic clamp (insulin was infused 90 min each at 6 –10 and 12–20 mU/m2zmin). Microdialysis probes were placed in abdominal and femoral sc adipose tissue. Subjects were given a primed, constant infusion of a stable isotope tracer (2H5-glycerol), and plasma glycerol isotope enrichments were analyzed by mass spectrometry to determine glycerol kinetics. Although there was no mean change in plasma leptin during the
L
EPTIN, A peptide hormone secreted by adipocytes, may be a satiety-signaling factor in the control of energy balance (1). If so, leptin production and secretion should be acutely regulated by lipogenic factors such as insulin. However, reported effects of hyperinsulinemia on plasma leptin concentrations in humans vary widely. Some studies have found no change in plasma leptin expression during hyperinsulinemia (2– 6), although others have found increases of about 20% over 3– 4 h of hyperinsulinemia (2, 3, 7). Although insulin may or may not acutely increase plasma leptin in humans, this hormone clearly attenuates the decline in plasma leptin induced by fasting (2– 6, 8). Saad et al. have recently shown a positive relationship between maximal insulin-stimulated glucose disposal and the degree of increase in plasma leptin during hyperinsulinemia (8). This suggests that individual differences in insulin responsiveness could help explain the variable findings regarding leptinemic responses to hyperinsulinemia mentioned above (2–7, 9). If insulin is acting to influence the Received February 12, 1999. Revision received May 26, 1999. Accepted June 28, 1999. Address all correspondence and requests for reprints to: Dr. Wendy M. Kohrt, Department of Internal Medicine, Box 8113, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, Missouri 63110. E-mail:
[email protected]. * This work was supported by Diabetes Research and Training Center Grant DK-20579 and Biomedical Mass Spectrometry Resource Grant RR-00954. Technical support was provided by the Washington University Radioimmunoassay CORE Laboratory and General Clinical Research Center (MO1-RR-00036). Investigators were supported by institutional and individual training grants (AG00078 and DK09542 to R.C.H., AG00078 to S.B.R., NCMRR-HD-07434 and AG-00078 to J.S.F., and NIH Research Career Development Award AG00663 to W.M.K.). † Current address: Human Performance Laboratory, East Carolina University, Greenville, North Carolina 27858.
clamp (baseline, 16.6 6 4.5 ng/mL; final, 16.3 6 4.3 ng/mL), there was large interindividual variability in the changes in plasma leptin (range, 218% to 119%). Changes in plasma leptin during the clamp stages were correlated with abdominal dialysate glycerol concentrations (r 5 20.44; P , 0.05), but not femoral dialysate glycerol concentrations (r 5 20.15), the rate of appearance of glycerol in plasma (r 5 0.005), or plasma insulin levels (r 5 0.16). The results suggest that insulin-induced changes in plasma leptin are more related to the lipolytic state (i.e. low leptin response when lipolysis is high) of abdominal sc adipose tissue than that of other fat depots. (J Clin Endocrinol Metab 84: 3726 –3731, 1999)
plasma leptin concentration by specifically altering leptin production by adipocytes, insulin-induced changes in leptinemia should be more closely associated with insulin action on adipocytes per se than with the whole body insulin action studied by Saad et al. (8). We therefore examined the relationships between indexes of global and regional lipolysis (a relatively specific marker of adipocyte metabolism) and changes in plasma leptin during physiological hyperinsulinemia. Whole body lipolysis was assessed by stable isotope kinetics, and lipolysis in abdominal and femoral sc adipose tissue was evaluated by microdialysis. Subjects and Methods Subjects Fifteen premenopausal (mean 6 se, 33 6 2 yr) women with no history of diabetes or cardiovascular disease participated in the study. Mean body weight was 70 6 5 kg (range, 50 –118 kg), and body mass index was 26 6 2 kg/m2. Body fat was 32 6 4% (range, 14 –59%). All subjects had normal glucose tolerance and had had normal monthly menses during the previous year. Subjects gave informed consent before participating, and all procedures were approved by the institutional review board at Washington University School of Medicine.
Oral glucose tolerance test Subjects reported after an overnight fast. After a baseline blood sample was taken, subjects consumed 75 g glucose. Blood samples were taken 30, 60, 90, and 120 min thereafter. Plasma glucose was determined by the glucose oxidase method (Beckman Coulter, Inc., Fullerton, CA). Normal glucose tolerance was defined as no blood glucose value more than 200 mg/dL and a 120 min value less than 140 mg/dL (10).
Body composition Body fat was measured by dual energy x-ray absorptiometry (QDR1000/w, Hologic, Inc., Waltham, MA), using version 5.64 of the enhanced whole body analysis program (11).
3726
VARIABLE LEPTINEMIC RESPONSE TO INSULIN
3727
Microdialysis
Hyperinsulinemic-euglycemic clamp
Inlet (30-cm) and outlet (15-cm) tubings of microdialysis probes (DL-3, Bioanalytical Systems, Inc., West Lafayette, IN) were separated by 3-cm of polyacrylonitrile dialysis membrane (id, 0.25 mm; od, 0.35 mm). After ethylene oxide gas sterilization, probes were rinsed in 20% ethanol for 20 min, then soaked overnight (;16 h) in 5% ethanol while being perfused with Ringer’s solution to remove glycerol from the dialysis membrane. Microdialysis probes were inserted with sterile technique into sc adipose tissue under local anesthesia. A 14-gauge catheter was inserted in the adipose tissue parallel to the skin, entering and exiting the skin about 6 cm apart. The microdialysis probe was threaded through the catheter, and the catheter was removed, leaving the probe embedded in the adipose tissue. Probes were placed in abdominal adipose tissue (two probes, ;3 cm on each side of the umbilicus) and femoral adipose tissue (two probes at the midthigh, ;3 cm apart) in each subject. Microdialysis probes were perfused (CMA model 102 microdialysis pump, Stockholm, Sweden) at 2.0 mL/min with Ringer’s solution containing 2.5 mmol/L glucose and 5 mmol/L ethanol. The glucose was added to the perfusate to minimize loss of glucose from the interstitial fluid. Ethanol was included in the perfusate to permit detection of changes in blood flow in the region surrounding the probe (12, 13). Probes were allowed to equilibrate for 1 h before collection of microdialysate samples to allow the trauma from probe insertion to subside (14, 15). Dialysate was collected every 10 min for baseline measures (20 mL) and every 15 min during insulin infusions (30 mL). Dialysates were stored at 4 C and analyzed within 48 h for glycerol (16) and ethanol (14). We have found that the variation for in vitro glycerol recovery is approximately 2% from probe to probe and from lot to lot. In vivo variability in recovery is dependent on both the individual probe and the site to site physiological differences within a given adipose tissue depot. In the current study, the mean within-subject, within-depot coefficients of variation for dialysate glycerol concentrations were 16.7 6 3.9% (basal), 15.3 6 2.9% (low dose insulin), and 13.5 6 2.6% (moderate dose insulin). Recovery is optimized at low flow rates and approaches 100% at flow rates below 0.3 mL/min. At these low flow rates, the time required to accumulate enough sample for assays would be prohibitive. For these experiments, we perfused the probes only at 2.0 mL/min. In previous experiments, we found that dialysate glycerol concentrations when probes were perfused at 2.0 mL/min were highly correlated (r 5 0.84; n 5 10; P 5 0.002) with calculated interstitial glycerol concentrations based on the no net flux method (data from Ref. 17). We therefore believe that dialysate glycerol concentrations are a good marker of interstitial glycerol concentrations. We have previously found glycerol recovery by the probes used in this study to be about 60% in abdominal sc adipose tissue of premenopausal women (17). However, we did not measure recovery in the current study, and we have previously found that glycerol recovery is reduced as skinfold thickness increases (unpublished data). To corroborate findings for unadjusted data, we also performed statistics on dialysate glycerol concentrations that were adjusted for differences in recovery. Recovery was computed with a regression equation that we developed from a group of women similar to that in the current study (17): recovery, 20.863 3 (mm skinfold thickness) 1 67.97% (unpublished equation). The observation that probe recovery is affected by skinfold thickness has been made in two research laboratories other than our own (18, 19). Several factors that are known to affect recovery, such as tortuosity (dependent in part on interstitial water content), blood flow, or effective membrane surface area, can change with increasing adipose tissue thickness. Therefore, factors related to tissue thickness could affect recovery over a probe located entirely within adipose tissue. The ethanol microdialysis technique for assessment of adipose tissue blood flow in humans has been validated with the xenon 133 clearance technique (20). The concentration of ethanol infused has been found to have no effect on lipolysis in sc adipose tissue (12). The ratio of ethanol concentration recovered in the dialysate to that in the perfusate (outflow/inflow ratio) is inversely proportional to blood flow in the tissue surrounding the probe (13).
Clamp studies were performed at the Washington University General Clinical Research Center after an overnight fast, as previously described (21). Baseline blood and dialysate samples were obtained 30, 20, 10, and 0 min before the start of insulin infusion. The hyperinsulinemic-euglycemic clamp consisted of two 90-min stages of primed, constant insulin infusion. In leaner subjects (,32% body fat), insulin was infused at 10 and 20 mU/m2zmin, whereas in more obese subjects (.35% body fat) insulin infusion rates were 6 and 12 mU/m2zmin. We anticipated that these infusion rates would result in similar insulin concentrations in lean and obese subjects (22). Blood samples were taken every 5 min during insulin infusions for determination of blood glucose. Blood glucose was maintained at approximately 5 mmol/L by modulating the infusion rate of 20% dextrose. Blood samples were obtained at 15-min intervals during the clamp for determination of substrate and hormone concentrations. Samples were stored at 280 C for determination of glycerol (16), insulin (23), and catecholamines (24) and at 220 C for determination of leptin (25).
Glycerol kinetics A primed (1.5 mmol/kg), constant (;0.1 mmol/kgzmin) infusion of H5-glycerol (99%; Tracer Technologies, Newton, MA) was started 90 min before the start of the first insulin infusion and continued until the end of the clamp. A blood sample taken before the start of the isotope infusion was used to determine background isotope enrichment. The actual isotope delivery rate was determined for each infusion by assay of the infusate enrichment. Analysis of 2H5-glycerol was performed using a modification of the negative ion chemical ionization (NCI) gas chromatography-mass spectrometry method (26). Plasma (50 mL) was added to 500 mL 3 mol/L perchloric acid, incubated at 4 C for 20 min, and centrifuged for 10 min at 2000 3 g. The supernatant was evaporated under nitrogen and reconstituted with 150 mL heptafluorobutyric acid-ethyl acetate (3:1) to form tris-heptafluorobutyryl ester derivatives, incubated at 70 C for 10 min, and dried under nitrogen. After reconstitution in 100 mL ethyl acetate, the NCI spectrum of the heptafluorobutyric acid derivative was obtained (gas chromatography-mass spectrometry, model 5988A, Hewlett-Packard Co., Palo Alto, CA). A DB-17 column (30 m; id, 0.25 mm; 0.25-mm film thickness; J and W Scientific, Folsom, CA) was used with a helium flow rate of 0.5 mL/min and a split ratio of 20:1. The column temperature was set at 100 C for 1 min and was increased 45 C/min to 280 C. The mass spectrometer conditions were: source temperature, 120 C; injector port and detector temperature, 250 C; emission, 300 mA; and ion source pressure, 0.5– 0.6 torr. Methane was the Cl2 reactant gas, and the concentration of M2 ions (nominal mass, 680 and 685 for unlabeled and labeled glycerol, respectively) was monitored. The rate of appearance (Ra) of glycerol over the last 30 min of each clamp stage was calculated using nonsteady state equations (27), assuming a volume of distribution for glycerol of 270 mL/kg (28). The coefficient of variation for day to day reproducibility of isotopic enrichment was 7.1 6 2.2% (n 5 6). 2
Leptin analysis The concentration of leptin in plasma samples was determined by RIA (25) with reagents from Linco Research, Inc. (St. Charles, MO). The antibody does not detect human insulin, proinsulin, glucagon, pancreatic polypeptide, or somatostatin, and intra- and interassay coefficients of variation are less than 8% (25).
Statistics Differences between baseline measures and measures during insulin infusions were evaluated by repeated measures ANOVA followed by paired t tests with Bonferroni adjustments to compare means at the different stages. Basal measures for insulin, glycerol, and ethanol outflow/inflow ratio were computed as the mean of the four baseline measurements from the half-hour before the beginning of the insulin infusions (230, 220, 210, and 0 min). Values for the low and moderate insulin clamp stages were computed as the mean value for the final half-hour of each stage. Changes in leptin during the clamp stages were
3728
FISHER ET AL.
JCE & M • 1999 Vol 84 • No 10
determined from the 0, 90, and 180 min values and expressed as percentages of baseline measures. Correlations between variables were determined by regression analysis. For all statistical procedures, a was 0.05. Data are presented as the mean 6 se.
Results Hormone and glycerol concentrations
Plasma leptin did not change during either stage of the hyperinsulinemic clamp (Table 1). Femoral dialysate glycerol fell to about 80% and 60% of basal level during the first (P , 0.05 vs. basal) and second (P , 0.05 vs. both basal and low dose) insulin infusions, respectively (Table 1). Abdominal dialysate glycerol concentration was suppressed to approximately 65% and 50% of baseline during the two clamp stages (P , 0.05, basal vs. low and moderate, low vs. moderate). Glycerol Ra was suppressed to approximately 50% and 40% of the basal rate during the clamp (P , 0.05, basal vs. both insulin stages). There were no differences in plasma catecholamine concentrations between the basal state and the insulin infusions. Blood flow
There were slight decreases in ethanol outflow/inflow ratios during the insulin infusions, indicative of an increased blood flow (Table 1). The abdominal outflow/inflow ratio did not change from baseline to the first clamp stage, but in the second stage, it was reduced compared to baseline and first stage values (P , 0.05). Femoral outflow/inflow decreased from baseline during the first stage (P , 0.05) and was further reduced in the second stage (P , 0.05). Correlations
The change in leptin during insulin infusions was inversely related (r 5 20.44; P , 0.05) to the abdominal dialysate glycerol concentration for both clamp stages combined (changes in plasma leptin during a stage are plotted against the abdominal dialysate glycerol concentration for that stage; Fig. 1A). Changes in leptin were correlated with abdominal dialysate glycerol during the first clamp TABLE 1. Hormone, glycerol, and blood flow measurements Basal
Plasma leptin (ng/mL) Glycerol conc. (mmol/L) Abdominal dialysate Femoral dialysate Glycerol Ra in plasma (mmol/kg fat masszmin) Insulin (mU/mL) Epinephrine (pg/mL) Norepinephrine (pg/mL) Ethanol outflow/inflow ratioc Abdominal Femoral
Insulin infusions Low dose
16.6 6 4.5 16.5 6 4.5 124 6 12 125 6 17 11.8 6 1.4
79 6 13a 104 6 16a 5.7 6 0.9a
Moderate dose
16.3 6 4.3 65 6 12a,b 74 6 10a,b 4.9 6 0.7a
4.6 6 1.0 16.8 6 1.1a 30.6 6 2.1a,b 27.3 6 4.1 29.3 6 5.2 31.2 6 5.0 187 6 29 204 6 29 214 6 28 0.52 6 0.04 0.48 6 0.03 0.45 6 0.03a,b 0.57 6 0.03 0.50 6 0.03a 0.44 6 0.03a,b
Each value is the mean 6 SE. a and b Difference compared to basal (a) or low dose (b) conditions (P , 0.05). c Decreased outflow/inflow ratio within femoral or abdominal probes corresponds to increased blood flow.
FIG. 1. Correlation between abdominal dialysate glycerol (A) or femoral dialysate glycerol (B) and the percent change in plasma leptin during low dose and moderate dose insulin infusions. *, Significant correlation (P , 0.05).
stage (r 5 20.55; P , 0.05), but the correlation was not statistically significant in the during the second stage (r 5 20.39). Changes in leptin were unrelated to femoral dialysate glycerol (Fig. 1B), glycerol Ra (Fig. 2), or plasma insulin (Fig. 3). Changes in leptin during the complete clamp (all 180 min of insulin infusion; Fig. 4) were correlated with abdominal dialysate glycerol concentrations from both the first clamp stage (r 5 20.57; P , 0.05) and the second clamp stage (r 5 20.57; P , 0.05). When dialysate glycerol concentrations were mathematically adjusted to account for reduced recovery with increasing skinfold thickness, the correlations between changes in leptin and dialysate glycerol for both stages combined were relatively unchanged (abdominal: r 5 20.45; P , 0.05; femoral: r 5 20.26; P 5 NS). Correlations between adjusted abdominal dialysate glycerol concentrations and changes in leptin for each stage separately were modest (first stage: r 5 20.47; second stage: r 5 20.45), but not statistically significant. The reported findings that skinfold thickness affects recovery (18, 19) do not invalidate the conclusions of this manuscript. The mathematical correction for changes in di-
VARIABLE LEPTINEMIC RESPONSE TO INSULIN
3729
FIG. 2. Correlation between glycerol Ra and the percent change in plasma leptin during low dose and moderate dose insulin infusions.
FIG. 4. Correlation between the percent change in leptin over the complete clamp and abdominal dialysate glycerol concentrations during the first stage (low dose insulin infusion; A) and the second stage (moderate dose insulin infusion; B) of the clamp. *, Statistically significant correlation (P , 0.05).
FIG. 3. Correlation between plasma insulin and the percent change in plasma leptin during low dose and moderate dose insulin infusions.
alysate concentration based on skinfold thickness does add additional variability to the data. However, the lack of substantial difference in the correlation coefficients for the relationships using uncorrected (r 5 20.44) and corrected (r 5 20.45) abdominal dialysate concentrations suggests that the uncorrected data are robust. We analyzed the relationships between changes in leptin and changes in abdominal dialysate glycerol. We found a significant relationship for the change in abdominal dialysate glycerol during the low dose infusion vs. the change in leptin over the entire clamp (r 5 20.60; P , 0.05), but not for changes in dialysate glycerol and plasma leptin within stages. We have plasma leptin values for nine of the subjects taken 90 min before the start of the insulin infusions. The decline in leptin during this 90-min baseline period (26.8 6 2.7%) was not related to the baseline abdominal adipose tissue dialysate glycerol concentration (r 5 0.02; P . 0.96). There was no relationship between the relative changes in
ethanol outflow/inflow ratios and relative changes in dialysate glycerol concentrations for either abdominal (r 5 20.03) or femoral (r 5 20.004) adipose tissue. Although the baseline plasma leptin concentration was highly correlated with body fat mass (r 5 0.97; P , 0.05), changes in leptin during the complete clamp were not significantly related to body fat mass (r 5 20.44; P . 0.10). Discussion
There appears to be wide variability in the changes in leptin expression that occur during hyperinsulinemia in humans (2–7, 9). Recently, Saad et al. reported that the amount of change in plasma leptin during physiological hyperinsulinemia was directly related to the whole body glucoregulatory action of insulin and was inversely related to the body fat content (8). We sought to determine whether changes in leptin during physiological hyperinsulinemia were related to the whole body and regional antilipolytic effects of insulin. We found considerable variability in the leptinemic response to physiological insulin infusions. Plasma leptin rose in some subjects during insulin infusions, whereas leptin levels in others continued to fall, as if they were fasting. Plasma leptin levels tended to increase the most in those subjects whose abdominal sc adipose tissue was the most sensitive to the antilipolytic effects of insulin.
3730
FISHER ET AL.
We found that the leptinemic response to insulin infusion was inversely related to a measure of lipolysis (dialysate glycerol concentration) in abdominal sc adipose tissue, but not to measures of femoral or whole body lipolysis. It has been reported that human abdominal sc fat secretes 2–3 times more leptin in vitro and contains more leptin messenger ribonucleic acid than omental fat (29). Furthermore, the fasting plasma leptin concentration has been found to be related to total sc fat and abdominal sc fat, but not to visceral fat (30). It appears that sc fat exerts more influence on plasma leptin expression than visceral fat, and our data suggest that insulin-associated changes in plasma leptin in women may be more related to the lipolytic state of abdominal sc fat than other (i.e. femoral) sc fat depots. The apparent inverse association of the level of lipolysis in abdominal sc fat with changes in plasma leptin levels is consistent with the hypothesis that leptin is a lipostatic messenger. We propose that when adipocytes are in the lipogenic state and lipolysis is low, leptin production is highest. Conversely, when adipocytes are most lipolytic, leptin production is lowest. In vivo and in vitro evidence supports this hypothesis. For example, isoprenaline, a b-adrenergic agonist that stimulates lipolysis, reduced leptin production by 20% after 90 min of iv infusion in humans (31). Furthermore, the isoprenaline-induced reduction in plasma leptin was reversed within 15 min of the cessation of isoprenaline infusion (31). Similarly, when lipolysis was stimulated by cAMP in human visceral adipocytes in vitro, leptin secretion was decreased (32). In addition, incubation of rat adipocytes with dibutyryl cAMP or the b-agonists isoproterenol and isoprenaline led to decreased leptin secretion (33, 34). On the other hand, maximal suppression of lipolysis by acipimox, a nicotinic acid analog, caused an approximately 5% increase in plasma leptin when infused for 3 h in humans, whereas, in contrast, leptin fell by about 7% during the saline control infusion (35). The current data and previous studies (31–35) suggest that leptin secretion may be acutely regulated at least partially by the same pathways that regulate lipolysis in adipocytes. It may be that the absolute levels of lipolysis in adipose tissue are more related to regulation of leptin levels than are changes in lipolysis. For example, if lipolysis and leptin secretion are regulated by the same pathways, it may be that there is a threshold level of signal necessary to both reduce lipolysis to a given level and stimulate leptin production. During fasting, when lipolysis levels are high in all subjects, the signals inhibiting lipolysis may be too low to have an effect on leptin production. The plasma leptin concentration follows a well known, diurnal pattern, peaking a few hours after the evening meal and reaching its lowest level before refeeding after an overnight fast (36). The fasting plasma leptin concentration has consistently been shown to be highly related to body fat content in humans (2, 30). Citing the high fasting leptin levels in obese subjects, some investigators have suggested that resistance to the anorectic activity of leptin develops during obesity (5). However, it is also possible that an attenuated rise in plasma leptin throughout the day, not the baseline leptin level, underlies the apparent disruption of leptin signaling in obese subjects. For example, it has been shown that the lep-
JCE & M • 1999 Vol 84 • No 10
tinemic response to insulin was lowest in the most obese subjects, who were insulin resistant (8). Saad et al. (8) suggested that the rise in plasma leptin in response to insulin may be attenuated or absent in insulin-resistant individuals. Fasting leptin levels have been found to be nearly 4 times higher (30.5 vs. 8.2 ng/mL) in obese subjects than those in normal weight subjects (37), but when the subjects were fed, there was only a 2-fold difference (34.6 vs. 15.8). Leptin almost doubled from nadir to peak in normal weight subjects, compared to only about a 10% change in obese subjects (37). It may be possible that a reduced effect of insulin on the plasma leptin concentration in obese subjects is responsible for an inappropriate rise in plasma leptin during the day. Although Saad et al. found an inverse relationship between adiposity and the rise in plasma leptin during hyperinsulinemia (8), we found no such relationship. It is likely that our low correlation between body composition and changes in leptin is based on methodological differences between our study and that of Saad et al. Although we examined subjects with a broader range of body fat than Saad et al. (7– 69 kg fat in women vs. ;5– 48 kg in men), we did not infuse insulin for nearly as long (3 vs. 8 h). In addition, we compared leptin levels to preinfusion levels, whereas Saad et al. compared leptin levels at the end of insulin infusions to leptin levels after a saline infusion. All of the current subjects had normal glucose tolerance, so any relationship between adiposity and the change in leptin was probably blunted by a restricted range of insulin resistance. However, there was a relationship between insulin action on lipolysis (i.e. absolute glycerol concentration) in abdominal sc adipose tissue and the change in leptin during insulin infusions. The results of the current study and others (8, 37) imply that interventions, such as dietary restriction or exercise training, that would increase insulin action would also augment the rise in leptin throughout the day or during hyperinsulinemia. Support for this conclusion was provided by the report of a 25% rise in the plasma leptin level in obese women during a 3-h euglycemichyperinsulinemic clamp after, but not before, a 6-day fast (3). One research group has reported gender differences in the response of leptin to hyperinsulinemia (2). For example, after 3 h of insulin infusion, women, but not men, had a 20% increase in plasma leptin levels (2). It is unknown whether this is a real gender-based difference or whether the difference was based on sample variability, because other studies have found 1) no increase in plasma leptin during hyperinsulinemia in women (3, 4) 2) an increase in plasma leptin during hyperinsulinemia in men (8), or 3) similar increases in men and women (7). Although it is known that the relationship between body fat and fasting plasma leptin is steeper in women than in men (2), we believe that it is premature to suggest that women have a greater leptinemic response to insulin than men. Our subjects were not studied during a standardized time of the menstrual cycle, but without solid evidence for gender differences in the leptin response to insulin (above) we have no reason to believe that this would affect the results. Furthermore, in a previous study, we found that estrogen replacement in postmenopausal women had no effect on fasting leptin levels (38). However, we must acknowledge that our findings for women may not be generalizable to men.
VARIABLE LEPTINEMIC RESPONSE TO INSULIN
Without a saline infusion control, we were only able to compare plasma leptin levels during insulin infusions to preinfusion levels. Over 3 h, leptin levels would be expected to fall about 20% in fasting subjects infused with saline (7). Thus, it is probable that even though leptin levels fell during insulin infusion in some subjects, leptin levels would have declined further during an insulin-free infusion. We would expect that correction of individual changes in leptin for saline control values would increase the statistical power to examine relationships between leptin and lipolysis, but we cannot be sure. The addition of a saline control in future studies would allow a more accurate examination of the relationship between lipolysis and leptin. Our findings suggest that the changes in plasma leptin during hyperinsulinemia may be related to the relative lipolytic or lipogenic state of certain sc adipose tissue depots. It appears that mechanisms for regulation of leptin expression in sc abdominal adipose tissue may share some of the pathways that regulate lipolysis. Acknowledgments We thank Sharif Abdelhamid and Jan Crowley for technical assistance.
References 1. Campfield LA, Smith FJ, Guisez Y, Devos R, Burn P. 1995 Recombinant mouse OB protein: evidence for a peripheral signal linking adiposity and central neural networks. Science. 269:546 –549. 2. Kennedy A, Gettys TW, Watson P, Wallace P, Ganaway E, Pan Q, Garvey WT. 1997 The metabolic significance of leptin in humans: gender-based differences in relationship to adiposity, insulin sensitivity, and energy expenditure. J Clin Endocrinol Metab. 82:1293–1300. 3. Anderson PH, Kristensen K, Pedersen SB, Hjølland E, Schmitz O, Richelsen B. 1997 Effects of long-term total fasting and insulin on ob gene expression in obese patients. Eur J Endocrinol. 137:229 –233. 4. Ryan AS, Elahi D. 1996 The effects of acute hyperglycemia and hyperinsulinemia on plasma leptin levels: its relationships with body fat, visceral adiposity, and age in women. J Clin Endocrinol Metab. 81:4433– 4438. 5. Dagogo-Jack S, Fanelli C, Paramore D, Brothers J, Landt M. 1996 Plasma leptin and insulin relationships in obese and nonobese humans. Diabetes. 45:695– 698. 6. Segal KR, Landt M, Klein S. 1996 Relationship between insulin sensitivity and plasma leptin concentration in lean and obese men. Diabetes. 45:988 –991. 7. Utriainen T, Malmstro¨m R, Ma¨kimattila S, Yki-Ja¨rvinen H. 1996 Supraphysiological hyperinsulinemia increases plasma leptin concentrations after 4 h in normal subjects. Diabetes. 45:1364 –1366. 8. Saad M, Khan A, Sharma A, et al. 1998 Physiological insulinemia acutely modulates plasma leptin. Diabetes. 47:544 –549. 9. Vidal H, Auboeuf D, De Vos P, Staels B, Riou JP, Auwerx J, Laville M. 1996 The expression of ob gene is not acutely regulated by insulin and fasting in human abdominal subcutaneous adipose tissue. J Clin Invest. 98:251–255. 10. American Diabetes Association. 1998 Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care. 20:1183–1194. 11. Kohrt WM. 1998 Preliminary evidence that DEXA provides an accurate assessment of body composition. J Appl Physiol. 84:372–377. 12. Galitzky J, Lafontan M, Nordenstro¨m J, Arner P. 1993 Role of vascular a-2 adrenoceptors in regulating lipid mobilization from human adipose tissue. J Clin Invest. 91:1997–2003. 13. Hickner RC, Rosdahl H, Borg I, Ungerstedt U, Jorfeldt L, Henriksson J. 1991 Ethanol may be used with the microdialysis technique to monitor blood flow changes in skeletal muscle: dialysate glucose concentration is blood flow dependent. Acta Physiol Scand. 143:355–356.
3731
14. Hickner RC, Rosdahl H, Borg I, Ungerstedt U, Jorfeldt L, Henriksson J. 1998 The ethanol technique of monitoring blood flow changes in rat skeletal muscle: implications for microdialysis. Acta Physiol Scand. 146:87–97. 15. Bolinder J, Hagstro¨m E, Ungerstedt U, Arner P. 1989 Microdialysis of subcutaneous adipose tissue in vivo for continuous glucose monitoring in man. Scand J Clin Lab Invest. 49:465– 474. 16. Weiland O. 1974 Methods for determination of metabolites. In: Bergmeyer HU, ed. Methods of enzymatic analysis. New York, London: Verlag Chemie Weinheim, Academic Press; 1404 –1409. 17. Hickner RC, Fisher JS, Kohrt WM. 1997 Regional differences in interstitial glycerol concentration in subcutaneous adipose tissue of women. Am J Physiol. 273:E1033–E1038. 18. Stallknecht, B. Estimation of interstitial concentrations of metabolites and hormones in adipose tissue by means of microdialysis-method evaluation. PhD Thesis, University of Copenhagen,Copenhagen; 11–1996. 19. Rosdahl H, Hamrin K, Ungerstedt U, Henriksson J. 1998 Metabolite levels in human skeletal muscle and adipose tissue studied with microdialysis at low perfusion flow. Am J Physiol. 274:E936 –E945. 20. Fellander G, Linde B, Bolinder J. 1996 Evaluation of the microdialysis ethanol technique for monitoring of subcutaneous adipose tissue blood flow in humans. Int J Obes. 20:220 –226. 21. Kohrt WM, Kirwan JP, Staten MA, Bourey RE, King DS, Holloszy JO. 1993 Insulin resistance in aging is related to abdominal obesity. Diabetes. 42:273–281. 22. Campbell PJ, Carlson MG, Nurjhan N. 1994 Fat metabolism in human obesity. Am J Physiol. 266:E600 –E605. 23. Morgan DR, Lazarow A. 1963 Immunoassay of insulin: two antibody system. Diabetes. 12:115–126. 24. Shah SD, Clutter WE, Cryer PE. 1985 External and internal standards in the single isotope derivative (radioenzymatic) assay of plasma norepinephrine and epinephrine in normal humans and persons with diabetes mellitus or chronic renal failure. J Lab Clin Med. 106:624 – 629. 25. Ma Z, Gingerich RL, Santiago JV, Klein S, Smith CH, Landt M. 1996 Radioimmunoassay of leptin in human plasma. Clin Chem. 42:942–946. 26. Gilker CD, Pesola GR, Matthews DE. 1992 A mass spectrometric method for measuring glycerol levels and enrichments in plasma using 13C and 2H stable isotopic tracers. Anal Biochem. 205:172–178. 27. Steele R. 1959 Influence of glucose loading and of injected insulin on hepatic glucose output. Ann NY Acad Sci. 82:420 – 430. 28. Beylot M, Martin C, Beaufrere B, Riou JP, Mornex R. 1987 Determination of steady-state and non-steady-state glycerol kinetics in humans using deuterium-labeled tracer. J Lipid Res. 28:414 – 419. 29. Van Harmelen V, Reynisdottir S, Eriksson P, Tho¨rne A, Hoffstedt J, Lo¨nnqvist F, Arner P. 1998 Leptin secretion from subcutaneous and visceral adipose tissue in women. Diabetes. 47:913–917. 30. Roemmich JN, Clark PA, Berr SS, et al. 1998 Gender differences in leptin levels during puberty are related to the subcutaneous fat depot and sex steroids. Am J Physiol. 275:E543–E551. 31. Pinkney JH, Coopack SW, Mohamed-Ali V. 1998 Effect of isoprenaline on plasma leptin and lipolysis in humans. Clin Endocrinol (Oxf). 48:407– 411. 32. Halleux CM, Servais I, Reul BA, Detry R, Brichard SM. 1998 Multihormonal control of ob gene expression and leptin secretion from cultured human visceral adipose tissue: increased responsiveness to glucocorticoids in obesity. J Clin Endocrinol Metab. 83:902–910. 33. Slieker LJ, Sloop KW, Surface PL, et al. 1996 Regulation of expression of ob mRNA and protein by glucocorticoids and cAMP. J Biol Chem. 271:5301–5304. 34. Hardie LJ, Guilot N, Trayhurn P. 1996 Regulation of leptin production in cultured mature white adipocytes. Horm Metab Res. 28:685– 689. 35. Hennes MMI, Dua A, Maas DL, Sonnenberg GE, Krakower GR, Kissebah AH. 1997 Relationships of plasma leptin levels to changes in plasma free fatty acids in women who are lean and women who are abdominally obese. Obes Res. 5:442– 446. 36. Shoeller DA, Cella LK, Sinha MK, Caro JF. 1997 Entrainment of the diurnal rhythm of plasma leptin to meal timing. J Clin Invest. 100:1882–1887. 37. Sinha MK, Opentanova I, Ohannesian JP, et al. 1996 Evidence for free and bound leptin in human circulation. J Clin Invest. 98:1277–1282. 38. Kohrt WM, Landt M, Birge Jr SJ. 1996 Serum leptin levels are reduced in response to exercise training, but not hormone replacement therapy, in older women. J Clin Endocrinol Metab. 81:3980 –3985.