| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Endocrinology, Diabetes, and Metabolism (J.H.L., J.L.C., C.O., C.S.M.), Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts 02215; Harokopio University (N.Y., M.K.), Athens, Greece 17671; Division of Endocrinology (E.E.), Connecticut Children’s Medical Center, University of Connecticut School of Medicine, Hartford, Connecticut 06106; and Preventative Cardiology Program (R.S.), Division of Cardiology, Hartford Hospital, Hartford, Connecticut 06102
Address all correspondence and requests for reprints to: Christos S. Mantzoros, M.D., Division of Endocrinology, Diabetes, and Metabolism, Department of Internal Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, 330 Brookline Avenue, Stoneman 816, Boston, Massachusetts 02215. E-mail: cmantzor{at}bidmc.harvard.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
(RELM-), expressed in adipose tissue, heart, lung, and tongue, and RELMß, which is expressed in the intestine (3, 4). Initial studies in mice suggested that resistin mediates insulin resistance by antagonizing insulin action and modulating one or more steps in the insulin-signaling pathway (1, 5). However, conflicting animal data have since been presented regarding whether resistin production is increased (1, 5) or decreased (6, 7, 8, 9, 10) in obesity and decreased (1, 5, 11, 12, 13) or increased (6, 14) by thiazolidinediones (TZDs), drugs known to reduce insulin resistance. Importantly, the relevance and physiological role of resistin in humans remain unknown. Given the incomplete homology (59%) between human and mouse resistin (4) and the absence in humans of one of the three murine resistin isoforms, resistin in humans may have a different physiological role than that in mice. Studies of genetic variations in the resistin gene, including single-nucleotide polymorphisms, are controversial regarding the role of resistin in obesity and insulin sensitivity (15, 16, 17, 18, 19, 20). Moreover, resistin mRNA expression is very low in isolated human adipocytes and does not correlate consistently with insulin resistance or obesity (21, 22, 23), making the role of human resistin in insulin resistance unclear.
To gain insight into the predictors and physiological role of circulating resistin in humans, we conducted a series of cross-sectional and interventional studies. To evaluate whether resistin correlates with obesity, insulin resistance, fasting lipid profile, and/or leptin (another adipocyte-secreted hormone that has important associations with obesity and insulin resistance in rodents and leptin-deficient humans), we studied cross-sectionally anthropometric, metabolic, and hormonal predictors of serum resistin levels in 123 middle-aged women and 120 healthy young subjects. To assess whether resistin levels are higher in insulin-resistant states, we compared serum resistin levels in a separate study of 19 normal-weight healthy adolescents and 19 adolescents with insulin resistance, i.e. obese nondiabetic subjects and obese type 2 diabetic (T2DM) subjects. Finally, to determine whether, similar to studies in mice (1, 2), resistin levels are regulated by fasting (a low-insulin and low-leptin state) and/or leptin administration, we measured resistin levels in healthy subjects in response to fasting alone, fasting with physiologic-dose leptin administration, and pharmacological doses of leptin in the fed state.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
One hundred thirty Greek middle-aged women [age 49.4 ± 9.2 yr, body mass index (BMI) 30.9 ± 5.5 kg/m2], without known history of diabetes, and 120 Greek students (age 17.7 ± 1.7 yr, BMI 22.3 ± 3.6 kg/m2) were consecutively enrolled in two separate studies approved by the Ethics Committee at Harokopio University in Athens, Greece, and the Institutional Review Board of Beth Israel Deaconess Medical Center. Use of two studies (relatively older and overweight vs. younger and lean subjects) allowed us to determine whether there is a delayed effect of resistin on insulin resistance, i.e. a low-penetration phenomenon, and/or whether there is a role of resistin in the absence of obesity. Subjects (and parents of the students) gave written informed consent to participate in the studies. Subjects completed a self-administered questionnaire on demographic characteristics, general health status, smoking history, and current medications, including hormone replacement therapy and oral contraceptive treatment for the middle-aged women. The women were classified as premenopausal if they had regular monthly periods, perimenopausal if they had irregular periods or elevated FSH levels, and postmenopausal if they reported no periods for 6 or more months. Blood samples were collected from all subjects after an overnight fast (
10 h) between 0830 and 1030 h, and the sera were stored at -70 C. Analysis was restricted to 123 subjects for whom resistin data were available.
Anthropometry and body composition. Anthropometric and body composition measurements were performed in all study participants before breakfast, with the subject wearing light clothing without shoes. For all subjects, weight and height were measured to the nearest 0.5 kg and 0.5 cm, respectively, and the BMI was calculated as weight (kilograms) divided by height squared (square meters). Waist circumference (WC) and hip circumference were also measured by the same observer to a precision of 0.1 cm, and the waist-to-hip ratio (WHR) was calculated. Triceps, biceps, subscapular, and suprailiac skinfolds were measured twice by one observer on the right side of the body to a precision of 0.2 mm using a Lange skinfold caliper (Cambridge Scientific Instruments, Cambridge, MA) with the average of the two measurements reported. Body composition was assessed for all subjects by bioelectrical impedance analysis (BIA), using a single frequency bioimpedance analyzer (model 101, RJL Systems, Mt. Clemens, MI). Fat-free mass and percent fat mass (%FM) were calculated by anthropometric and BIA variables using the fatness-specific regression equations developed by Segal et al. (24) for the group of middle-aged women, and using the age- and gender-specific equations developed by Deurenberg et al. (25) for the students.
To validate the BIA and anthropometry measurements, 60 of the middle-aged women also underwent a dual-energy x-ray absorptiometry (DEXA) scan using a Lunar DPX densitometer (Lunar Corp., Madison, WI; software version 4.7e) for measurement of total body and trunk fat mass (expressed in kilograms and percent body weight). Estimates of fat mass (FM) and %FM obtained from the two methods correlated strongly with each other (FM by BIA and DEXA: Pearson’s coefficient, r = 0.91, P < 0.001; %FM by BIA and DEXA: r = 0.88, P < 0.001). Moreover, FM and %FM by BIA and DEXA correlated with BMI (both FM and %FM by BIA and BMI: r = 0.94, P < 0.001; both FM and %FM by DEXA and BMI: r = 0.84, P < 0.001). In addition, trunk fat, a measurement of central adiposity, obtained by DEXA correlated highly with WC (r = 0.90, P < 0.001) and WHR (r = 0.54, P < 0.001). Thus, in our analysis, we used BMI and FM calculated by BIA as markers of overall adiposity, and WC and WHR as markers of central obesity. Secondary analysis was also performed using DEXA-derived measurements of FM and central obesity.
Study of lean subjects vs. obese, insulin-resistant, nondiabetic and diabetic subjects. In a separate cross-sectional study, 19 normal-weight healthy adolescents and 19 obese, insulin-resistant adolescents (10 nondiabetic subjects and nine T2DM subjects) were enrolled from the endocrinology clinic at Connecticut Children’s Medical Center (Hartford, CT), in a study approved by the Connecticut Children’s Medical Center Institutional Review Board. Subjects and their parents gave written informed consent to participate in the study. Subjects were classified as obese if their BMI was over the 95th percentile for age and gender (26). All of the diabetic adolescents were classified as having T2DM by criteria established by the National Diabetes Data Group (27). The mean BMI was 22.2 ± 2.1 kg/m2 for the normal-weight subjects; 40.3 ± 7.3 kg/m2 for the obese, nondiabetic subjects; and 43.5 ± 10.5 kg/m2 for the obese T2DM subjects. There was no significant difference in age among the three groups: 14.5 ± 1.6 yr, 12.9 ± 1.1 yr, and 14.8 ± 1.2 yr, respectively. The male:female ratios were similar, i.e. 10:9, 4:6, and 5:4, respectively. Fasting blood samples were obtained between 0800 and 1000 h, and the plasma was stored at -70C until analyses of resistin and insulin concentrations were performed. Of the subjects with T2DM, three were on insulin, four were on metformin, and the other two were not on pharmacotherapy. For the seven subjects on pharmacotherapy, insulin administration the evening before blood sampling was decreased to one third the normal dose and metformin was held the morning of blood sampling.
Interventional studies
Two interventional studies were performed to investigate the effect of fasting and/or leptin administration on resistin levels. The Beth Israel Deaconess Medical Center Institutional Review Board approved the studies, and all subjects gave written informed consent to participate in their respective study. Clinical-quality recombinant methionyl human leptin (r-metHuLeptin) was supplied by Amgen, Inc. (Thousand Oaks, CA) and was administered under an investigator-initiated investigational new drug application.
Administration of physiological doses of r-metHuLeptin in the fasted state. Six normal-weight men (age 22.0 ± 2.1 yr, BMI 24.2 ± 1.2 kg/m2) were screened for any medical problems and admitted to the General Clinical Research Center under three different conditions: baseline fed state, 48-h fasting with placebo administration, and 48-h fasting with physiologic-dose r-metHuLeptin administration. The three admissions were scheduled at least 7 wk apart from each other to allow for return of body weight to baseline, avoidance of residual effects of administered r-metHuLeptin, and recovery of hematocrit. Subjects were admitted to the General Clinical Research Center the evening before d 1 for each admission. Fasting blood samples for resistin and leptin measurements were obtained at 0800 h on d 1 and 3 of each admission.
During the baseline fed state, subjects were placed on an isocaloric diet to maintain their admission body weight, with four standardized meals per day: 20% of calories from breakfast (0800 h), 35% from lunch (1300 h), 35% from dinner (1800 h), and 10% from a snack (2200 h). During both fasting studies, subjects received only caffeine-free and calorie-free liquids, NaCl (500 mg/d), KCl (40 meq/d), and a standard multivitamin with minerals daily. R-metHuLeptin was administered sc at a dose of 0.04 mg/kg·d on the first day and 0.1 mg/kg·d on the second day, which were designed to achieve physiological serum leptin levels similar to the fed state. The total daily dose of r-metHuLeptin was divided into four equal doses given every 6 h, starting at 0800 h on the first day after baseline blood samples were obtained. During the fasting with placebo administration admission, buffer solution was administered sc according to the same schedule as r-metHuLeptin.
Administration of pharmacological doses of r-metHuLeptin in the fed state. Five normal-weight healthy men (BMI < 25 kg/m2), five obese healthy men (BMI > 29 kg/m2), and five normal-weight healthy women (BMI < 25 kg/m2) were admitted to the General Clinical Research Center during the evening and received one pharmacological dose of r-metHuLeptin (0.3 mg/kg) the following morning at 0800 h. Fasting blood samples were collected in the morning before r-metHuLeptin administration (time 0) and again 5 h after the dose. During the study, subjects received an isocaloric diet, as described above.
Hormone measurements. Resistin levels were measured using an ELISA (Biovendor Laboratory Medicine Inc., Brno, Czech Republic; supplied by Alpco Diagnostics, Windham, NH) as previously described (28, 29). The limit of detection is 0.2 ng/ml, intraassay coefficient of variation (CV) 5.2% for low levels and 3.4% for high levels, and inter-assay CV 7.2% and 6.3%, respectively. The antibodies in the ELISA have no detectable cross-reactivity to mouse resistin or other cytokines in human serum. To validate the reliability of the resistin ELISA, we assayed serial dilutions of recombinant human resistin (PeproTech, Inc., Rocky Hill, NJ) between 3.13 ng/ml and 50.0 ng/ml using the Biovendor ELISA on two separate occasions. We found an excellent correlation between calculated and observed resistin concentrations each time (r2 = 0.99 and 0.98). To analyze the spiking recovery of the assay, serum from two subjects with baseline resistin levels of 5.52 ng/ml and 8.59 ng/ml were spiked with increasing amounts of recombinant human resistin (between 3.13 ng/ml and 50.0 ng/ml) and assayed. Similar to the spiking recovery data presented by the manufacturer, our analysis revealed high correlations (r2 = 0.9996 and 0.999). Moreover, the linearity data presented by the manufacturer show that the assay is highly linear (r2 = 0.999). Lastly, we studied the effect of repeated freezing and thawing of samples on resistin levels. Serum from four subjects, with resistin levels ranging from 6.77 ng/ml to 19.98 ng/ml, were each divided into two aliquots, one of which was frozen/thawed five times and the other that was thawed once and stored at 4 C for the 2 d before being assayed. No significant difference in resistin levels between the two groups was detected (12.2 ± 8.1 vs. 11.0 ± 6.1 ng/ml, P = 0.31, by paired t test).
Additional hormone concentrations were measured using commercially available RIAs as follows: leptin (Linco Research, St. Charles, MO; sensitivity 0.5 ng/ml; intraassay CV 8.3%); cortisol [Diagnostics Systems Laboratories (DSL), Webster, TX; sensitivity 0.5 µg/dl (13.8 nmol/liter); intraassay CV 5.3–8.4%]; insulin [DSL; sensitivity 1.3 µIU/ml (7.8 pmol/liter); intraassay CV 8.3%]; estradiol [Diagnostic Products Corp. (DPC), Los Angeles, CA; sensitivity 8.0 pg/ml (29.4 pmol/liter); intraassay CV 4.3–7.0%]; and free testosterone [DPC; sensitivity 0.15 pg/ml (0.52 pmol/liter); intraassay CV 8.0%]. For the Connecticut Children’s Medical Center adolescents, insulin was measured via ELISA (DSL; sensitivity 0.26 µIU/ml (1.56 pmol/liter); intraassay CV 2.6% for low levels and 1.3% for high levels; interassay CV 5.2% for low levels and 6.2% for high levels). The ELISA has no cross-reactivity with proinsulin or C peptide. Cross-validation with the DSL insulin RIA yielded a high correlation (r = 0.96). To minimize variability, hormone levels were measured in one assay for all subjects participating in each cross-sectional study and in one assay for each subject in the interventional studies. Glucose and lipid profiles were measured using a photometric method with liquid reactants (Hitachi 917, Indianapolis, IN). Insulin resistance (IR) was estimated using the homeostasis model assessment (HOMA) using the formula: IR = fasting insulin x fasting glucose/22.5 (30).
Statistical analysis
SPSS 8.0 software (SPSS Inc., Chicago, IL) was used for statistical analysis, and P less than 0.05 (two-tailed) was considered statistically significant for all analyses. In the cross-sectional studies of middle-aged women and young subjects, we calculated Pearson’s correlation coefficients and then performed bivariate and multivariate regression analyses. We evaluated for associations between serum resistin levels and anthropometric (BMI, FM, WC, WHR), metabolic (fasting glucose, insulin, HOMA-IR, lipid profile), and hormonal (leptin, cortisol, estradiol, free testosterone) parameters, all expressed as continuous variables. Analyses were controlled for potential confounders, including age, menopausal status, FM, and estrogen levels for the middle-aged women, and gender, age, and FM for the young population. Several variables were logarithmically transformed to obtain a normal distribution. Subjects taking hormone replacement (n = 5) or oral contraceptive (n = 2) treatment at the time of the evaluation were included. In addition to adjustment for circulating estrogen levels, multivariate regression analysis models were also run to adjust for estrogen replacement using a dummy variable (yes/no) and after stratification for hormone replacement (i.e. by excluding women on hormone replacement). We used t tests to analyze for differences in body composition, Tanner stage, and hormonal parameters between the normal-weight and obese adolescents from the Connecticut Children’s Medical Center. We used one-way ANOVA with post hoc (least significant difference) analysis to assess for differences in resistin and insulin levels among the normal-weight, obese nondiabetic, and obese T2DM adolescents. For the interventional studies, because the data were not normally distributed, we used Wilcoxon signed ranks tests to evaluate for changes in resistin levels between d 1 and 3 of each fed or fasting admission and between time 0 and 5 h after administration of a pharmacological dose of leptin.
The size of each cross-sectional study of 123 middle-aged women and 120 young subjects provided more than 80% power, and the combination of both studies provided more than 96% power, to detect clinically-significant associations (r 0.22) at the conventional = 0.05 level. Comparing two groups of 19 subjects each (Connecticut Children’s Medical Center subjects) provided more than 80% power at the conventional = 0.05 level to detect a difference in mean resistin levels 80% or more of the corresponding SD.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
summarizes the demographic, anthropometric, metabolic, and hormonal parameters of the 130 middle-aged women enrolled in the first cross-sectional study. The mean WC was 88.6 ± 12.9 cm, FM 32.9 ± 9.8 kg, and %FM 41.3 ± 4.2%. Subjects had a mean resistin level of 9.3 ± 7.7 ng/ml, mean leptin level of 23.4 ± 13.7 ng/ml, and mean fasting insulin level of 7.9 ± 4.1 µIU/ml (47.4 ± 24.6 pmol/liter).
|
illustrates the positive association between serum leptin levels and BMI (Pearson’s correlation coefficient r = 0.69, P < 0.001) and FM (r = 0.72, P < 0.001), and the lack of association between resistin levels and BMI (r = -0.12, P = 0.18) or FM (r = -0.12, P = 0.17). To evaluate potential predictors of resistin, we performed bivariate regression analyses of resistin and several anthropometric, metabolic, and hormonal factors. We found no correlation of resistin with any markers of adiposity or insulin resistance, including BMI, WC, WHR, FM, insulin and HOMA-IR (Table 2). In addition, there was no association between resistin levels and lipid (total cholesterol, high-density lipoprotein, low-density lipoprotein, triglyceride) or hormone levels (leptin, cortisol, estradiol, free testosterone) (Table 2). To adjust for potential confounders, we then performed multivariate regression analyses, adjusting for successively introduced covariates, i.e. age, menopausal status, estradiol level (or use of estrogen replacement therapy), and FM. With these adjustments, we again found no association between resistin and markers of adiposity or insulin resistance, lipid levels, or hormone levels (Table 2). Similar multivariate models adjusting for age and menopausal status but excluding perimenopausal women or women on hormone replacement therapy yielded similar results (data not shown). Secondary analysis using bivariate and multivariate models assessing DEXA-derived measurements of FM and central obesity also revealed similar results (data not shown). Because of the impact of tobacco use on cardiovascular risk, we also performed multivariate analyses adjusting for age and smoking and did not find a correlation between resistin and any of the variables (data not shown). Furthermore, we explored nonlinear correlations between resistin and both FM and HOMA-IR and did not find significant associations (data not shown). Finally, we compared resistin levels between obese (BMI > 30 kg/m2, n = 59) and nonobese women (BMI < 30 kg/m2, n = 64) and found no significant difference between the two groups (8.7 ± 7.3 ng/ml vs. 9.9 ± 8.1 ng/ml, P = 0.41 by t test).
|
|
summarizes the demographic, anthropometric, metabolic, and hormonal parameters of the 120 young subjects enrolled in the second cross-sectional study. The male:female ratio was 56:62. The population was lean (BMI 22.3 ± 3.6 kg/m2, WC 74.3 ± 10.5 cm, %FM 22.7 ± 7.9%), and none of the subjects had diabetes mellitus. Subjects had a mean resistin level of 4.5 ± 1.9 ng/ml, mean leptin level of 6.8 ± 5.9 ng/ml, and mean fasting insulin level of 5.6 ± 3.2 µIU/ml (33.6 ± 19.2 pmol/liter).
|
depicts the positive association between leptin and BMI (r = 0.30, P = 0.001) and FM (r = 0.65, P < 0.001), and lack of association between resistin and BMI (r = -0.02, P = 0.83) or FM (r = 0.15, P = 0.11). By bivariate and multivariate regression analysis, females had higher resistin levels, compared with males (ß = -0.29, P = 0.002), even after adjustment for age and FM (Table 4). Females had a mean resistin level of 4.9 ± 1.6 ng/ml, compared with 4.1 ± 2.1 ng/ml for males (P = 0.02 by t test). Although there were weak associations between resistin and sum of skinfolds (ß = 0.20, P = 0.03), %FM (ß = 0.25, P < 0.01), and free testosterone (ß = -0.28, P = 0.002) by bivariate analysis, these correlations decreased and lost significance after adjustment for gender, indicating that these associations are due to confounding by gender. Similarly, there was a significant positive association between resistin and leptin levels (ß = 0.32, P = 0.001) that achieved only borderline statistical significance after adjustment for gender (Table 4). There were also no significant associations between resistin levels and WC, WHR, FM, or insulin by bivariate and multivariate regression analysis (Table 4). Lastly, there was no association between resistin levels and either smoking or exercise status (data not shown).
|
|
To evaluate whether diabetes mellitus per se is associated with resistin levels, we compared the obese insulin resistant nondiabetic subjects with the T2DM subjects. Fasting insulin levels were 52.5 ± 40.5 µIU/ml (315.0 ± 243.0 pmol/liter) for the nondiabetic subjects and 37.6 ± 20.9 µIU/ml (225.6 ± 125.4 pmol/liter) for the T2DM subjects, compared with 9.0 ± 6.7 µIU/ml (54.0 ± 40.2 pmol/liter) for the normal-weight subjects (P < 0.001 by one-way ANOVA). By post hoc (least significant difference) analysis, there was a significant difference in insulin levels between the normal-weight subjects and both obese nondiabetic subjects (P < 0.001) and obese T2DM subjects (P = 0.005). Insulin levels and HOMA-IR were not different among patients on insulin, metformin, or no pharmacotherapy. There was no significant difference in resistin levels between the obese nondiabetic subjects (5.8 ± 1.6 ng/ml), obese T2DM subjects (7.8 ± 4.6 ng/ml), and normal-weight subjects (5.6 ± 1.9 ng/ml) (P = 0.13 by one-way ANOVA).
We then evaluated whether fasting and/or leptin administration regulates resistin levels in healthy subjects. In the fed state, leptin and resistin levels remained stable between d 1 and d 3 (leptin 1.6 ± 1.6 ng/ml vs. 1.7 ± 0.8 ng/ml, P = 0.31; resistin 9.5 ± 2.4 ng/ml vs. 10.4 ± 4.3 ng/ml, P = 0.53). After 48 h of fasting plus placebo administration, weight decreased from 75.7 ± 7.1 kg on d 1 to 73.4 ± 7.2 kg on d 3 (P = 0.03), and leptin levels declined by 67% (2.1 ± 1.8 ng/ml vs. 0.7 ± 0.9 ng/ml, P = 0.01). Leptin replacement during fasting restored the decrease in leptin levels to levels higher than baseline but within the normal physiological range for lean men (2.3 ± 1.9 ng/ml vs. 4.8 ± 2.1 ng/ml, P = 0.06). However, resistin levels remained unchanged after 48 h of fasting with placebo administration (11.2 ± 3.2 ng/ml vs. 10.6 ± 2.1 ng/ml, P = 0.34), and after 48 h of fasting plus physiologic-dose r-metHuLeptin administration (8.9 ± 2.4 ng/ml vs. 9.9 ± 2.3 ng/ml, P = 0.50).
Because there was no effect of physiological levels of leptin on resistin levels, we studied whether resistin levels are affected by pharmacological leptin levels by measuring leptin and resistin levels immediately before and 5 h after a pharmacological dose of r-metHuLeptin was administered to 15 healthy subjects, i.e. five normal-weight men (age 22.2 ± 2.0 yr, BMI 22.0 ± 1.0 kg/m2), five obese men (age 23.4 ± 3.4 yr, BMI 32.0 ± 2.3 kg/m2), and five normal-weight women (age 20.4 ± 1.5 yr, BMI 21.9 ± 1.5 kg/m2). Leptin levels increased more than 20-fold (10.6 ± 7.6 ng/ml vs. 243.7 ± 114.2 ng/ml, P = 0.001), but resistin levels did not change significantly (6.2 ± 2.4 ng/ml vs. 5.8 ± 1.8 ng/ml, P = 0.21) over this period of elevated leptin levels. Separate analyses within each subgroup (lean men, obese men, and lean women) also showed no significant difference in resistin levels at baseline and 5 h after r-metHuLeptin administration (data not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
20% higher), but unlike leptin, circulating resistin levels are not affected by acute fasting and/or by r-metHuLeptin administration at either physiological or pharmacological doses.
Prior studies in mice have shown conflicting results regarding the link between resistin and insulin resistance and/or obesity. Initial studies supported a role for resistin in mediating insulin resistance. Circulating resistin levels were found to be elevated in genetic (ob/ob and db/db) and diet-induced models of obesity (1). Exogenous resistin administration to healthy mice worsened glucose tolerance and induced insulin resistance, but neutralizing antiresistin antibodies in obese, insulin-resistant mice improved insulin sensitivity and glucose tolerance (1, 5). Interestingly, infusion of either resistin or RELMß induced severe hepatic but not peripheral insulin resistance, including inhibition of glucose metabolism and increased glucose production (31). TZDs, which lower insulin resistance by binding to peroxisome proliferator activated receptor-
(PPAR) receptors, decreased resistin levels in ob/ob mice and diabetic fatty rats (1, 5, 11, 12). Furthermore, tissue-cultured murine adipocytes treated with TZDs decreased resistin mRNA expression (1, 11, 13, 32). In contrast, resistin expression in 3T3-L1 adipocytes and adipose tissue was increased by high glucose concentrations and dexamethasone (13, 33). Finally, serum resistin levels in mice as measured by immunoblot decreased after a 48-h fast, with similar changes in resistin mRNA and protein expression in adipocytes (1, 2, 33).
Other studies, however, have not supported a role for resistin in mediating insulin resistance in rodents. TZD treatment increased resistin expression in both lean and obese rodents, including ob/ob mice and Zucker diabetic fatty rats (6, 14), and resistin protein expression was increased by 66% in db/db mice treated with metformin (34). Others have shown that adipose tissue of obese mice has decreased expression of resistin mRNA (6, 7, 8, 10, 33) and that resistin gene expression in the adipose tissue of insulin-resistant Fischer 344 rats is similar to that of age-matched, insulin-sensitive Sprague Dawley rats (35). Additionally, resistin gene expression in 3T3-L1 adipocytes was inhibited by insulin and TNF, which are increased in obesity (13, 32, 33, 36). The inconsistencies between animal studies may be related to methodological limitations, including different methods used to measure resistin, and/or variation among serum concentrations, mRNA, and protein levels, or may indicate that resistin does not play a significant role in the pathophysiology of insulin resistance (37).
Studies of resistin mRNA and protein expression in adipose tissue of humans have provided additional support that resistin may not play a role in insulin resistance or obesity in humans. Resistin mRNA and protein expression are increased in abdominal fat depots (23, 38) and whole adipose tissue of obese individuals, but isolated human adipocytes have very low resistin mRNA levels (21). Furthermore, previous studies have reported no differences in resistin expression in human fat and muscle cells when healthy, insulin-resistant, and T2DM subjects were compared (22), no correlation between resistin gene expression in human adipocytes and insulin resistance (39) or BMI (21), and, in contrast to mice studies (1, 5, 6, 11, 12, 13, 14, 32), no change in resistin expression in human mononuclear cells in response to peroxisome proliferator activated receptor- agonists (21). Finally, studies on the potential association between genetic variations of the resistin gene and obesity or insulin resistance have revealed inconsistent results (15, 16, 17, 18, 19, 20). Thus, human studies have not provided consistent evidence that tissue resistin expression plays a key role in the development of insulin resistance (37).
The cross-sectional and interventional studies we report herein do not support a major role for circulating resistin in mediating obesity or insulin resistance and/or linking obesity to diabetes in humans. Given the incomplete homology between human and mouse resistin (4) and the differences in energy metabolism between mice and humans, the physiology of resistin in mice may differ from that in humans (37, 40). Furthermore, in humans, resistin lies on chromosome 19p13.3, a region that has not been linked with susceptibility to obesity or insulin resistance (40). Our findings are consistent with prior studies that demonstrated no association between human resistin mRNA and protein expression and insulin resistance or obesity (21, 22, 39). Additionally, we previously found no association between serum resistin levels and either total energy or macronutrient intake (28), and a recent study, using the same resistin ELISA as in our studies, found no difference in serum resistin levels between insulin-resistant nondiabetic partial lipodystrophy patients and healthy controls (29).
Although there is evidence for higher resistin mRNA in visceral, compared with peripheral, fat (23, 38, 41), our data do not reveal a significant association between circulating resistin and central obesity in humans and thus do not support a role for resistin in mediating the effect of central obesity on insulin resistance or diabetes. Similar to leptin, resistin levels in mice decrease in response to fasting and increase with feeding or insulin administration (1, 2). In contrast to mice, however, we detected no changes in circulating resistin levels after 2 d of fasting in healthy humans. Furthermore, we found no significant correlation between resistin and leptin after adjustment for gender. Lastly, neither restoring the fasting-induced decrease in leptin levels to physiological levels nor achieving pharmacological leptin levels with r-metHuLeptin administration altered resistin levels.
This study was adequately powered to demonstrate clinically significant associations and differences between groups and used state-of-the-art methodology. We evaluated not only linear correlations of resistin with the outcomes of interest but also the possibility that nonlinear associations may exist between resistin and its putative effects, as previously suggested (42). We have studied different groups with varying degrees of insulin resistance and obesity and included a nonobese group of young subjects to evaluate the hypothesis that the impact of resistin may be more important in the absence of severe obesity (42) and/or that the effect of resistin on insulin resistance may be delayed, reflecting a low-penetration phenomenon.
Although our data do not support the notion that circulating resistin plays a critical role in insulin resistance and/or energy homeostasis in humans, it remains possible that resistin acting in a paracrine or autocrine manner may play a relatively more important role. Future research on the biology of circulating resistin in humans and the role of adipose tissue expression of resistin and RELMs in adipogenesis, obesity, and insulin resistance would shed more light on resistin physiology in humans. Interventional studies including recombinant human resistin administration in the context of well-designed clinical trials or the identification and detailed study of humans with congenital resistin deficiency would be invaluable. Finally, identification of the receptor system for resistin, relevant signaling pathway(s), and their sensitivity state are needed for a complete evaluation of the role of the resistin system in humans (42).
In summary, we present cross-sectional and interventional studies on the physiology of circulating resistin in humans and demonstrate that resistin levels are not associated with markers of insulin resistance and/or obesity and are not regulated by fasting or leptin administration. We conclude that circulating resistin is unlikely to play a major role in obesity, insulin resistance, or energy homeostasis in humans. Further detailed studies are needed to conclusively address these topics.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: BIA, Bioelectrical impedance analysis; BMI, body mass index; CV, coefficient of variation; DEXA, dual-energy x-ray absorptiometry; FM, fat mass; %FM, percent fat mass; HOMA, homeostasis model assessment; IR, insulin resistance; RELM, resistin-like molecule; r-metHuLeptin, recombinant methionyl human leptin; T2DM, type 2 diabetes; TZD, thiazolidinedione; WC, waist circumference; WHR, waist-hip ratio.
Received March 26, 2003.
Accepted July 7, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
in adipose tissue. Mol Endocrinol 16:1920–1930 is a negative regulator of resistin gene expression and secretion in 3T3-L1 adipocytes. Biochem Biophys Res Commun 288:1027–1031[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  C. Heidemann, Q. Sun, R. M. van Dam, J. B. Meigs, C. Zhang, S. S. Tworoger, C. S. Mantzoros, and F. B. Hu Total and High-Molecular-Weight Adiponectin and Resistin in Relation to the Risk for Type 2 Diabetes in Women Ann Intern Med, September 2, 2008; 149(5): 307 - 316. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-F. Hivert, L. M. Sullivan, C. S. Fox, D. M. Nathan, R. B. D'Agostino Sr., P. W. F. Wilson, and J. B. Meigs Associations of Adiponectin, Resistin, and Tumor Necrosis Factor-{alpha} with Insulin Resistance J. Clin. Endocrinol. Metab., August 1, 2008; 93(8): 3165 - 3172. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Weikert, S. Westphal, K. Berger, J. Dierkes, M. Mohlig, J. Spranger, E. B. Rimm, S. N. Willich, H. Boeing, and T. Pischon Plasma Resistin Levels and Risk of Myocardial Infarction and Ischemic Stroke J. Clin. Endocrinol. Metab., July 1, 2008; 93(7): 2647 - 2653. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Jansson, A. Nilsfelt, M. Gellerstedt, M. Wennergren, L. Rossander-Hulthen, T. L Powell, and T. Jansson Maternal hormones linking maternal body mass index and dietary intake to birth weight Am. J. Clinical Nutrition, June 1, 2008; 87(6): 1743 - 1749. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. S. Ahima and M. A. Lazar Adipokines and the Peripheral and Neural Control of Energy Balance Mol. Endocrinol., May 1, 2008; 22(5): 1023 - 1031. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Saito, N. Yamaguchi, Y. Shimazaki, H. Hayashida, K. Yonemoto, Y. Doi, Y. Kiyohara, M. Iida, and Y. Yamashita Serum Levels of Resistin and Adiponectin in Women with Periodontitis: the Hisayama Study J. Dent. Res., April 1, 2008; 87(4): 319 - 322. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. F. Chehab Minireview: Obesity and LipOdystrophy--Where Do the Circles Intersect? Endocrinology, March 1, 2008; 149(3): 925 - 934. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. Gentile, C. Vecchione, G. Marino, A. Aretini, A. Di Pardo, G. Antenucci, A. Maffei, G. Cifelli, L. Iorio, A. Landolfi, et al. Resistin Impairs Insulin-Evoked Vasodilation Diabetes, March 1, 2008; 57(3): 577 - 583. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Korbonits, D. Blaine, M. Elia, and J. Powell-Tuck Metabolic and hormonal changes during the refeeding period of prolonged fasting Eur. J. Endocrinol., August 1, 2007; 157(2): 157 - 166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Knudson, G. M. Dick, and J. D. Tune Adipokines and Coronary Vasomotor Dysfunction Experimental Biology and Medicine, June 1, 2007; 232(6): 727 - 736. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Osawa, Y. Tabara, R. Kawamoto, J. Ohashi, M. Ochi, H. Onuma, W. Nishida, K. Yamada, J. Nakura, K. Kohara, et al. Plasma Resistin, Associated With Single Nucleotide Polymorphism -420, Is Correlated With Insulin Resistance, Lower HDL Cholesterol, and High-Sensitivity C-Reactive Protein in the Japanese General Population Diabetes Care, June 1, 2007; 30(6): 1501 - 1506. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Fantuzzi and T. Mazzone Adipose Tissue and Atherosclerosis: Exploring the Connection Arterioscler. Thromb. Vasc. Biol., May 1, 2007; 27(5): 996 - 1003. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L Senolt, D Housa, Z Vernerova, T Jirasek, R Svobodova, D Veigl, K Anderlova, U Muller-Ladner, K Pavelka, and M Haluzik Resistin in rheumatoid arthritis synovial tissue, synovial fluid and serum Ann Rheum Dis, April 1, 2007; 66(4): 458 - 463. [Abstract] [Full Text] [PDF]
|
