Serum Resistin Is Not Associated with Obesity Or Insulin Resistance in Humans

Home , Resistin

European Review for Medical and Pharmacological Sciences 2005; 9: 161-165 Serum resistin is not associated with obesity or insulin resistance in humans

N. IQBAL1,2, P. SESHADRI1,2, L. STERN1, J. LOH2, S. KUNDU3, T. JAFAR4, F. F. SAMAHA5

1 Department of Medicine, Section of Endocrinology Diabetes and Metabolism, Philadelphia VA Medical Center, Philadelphia, Pennsylvania (USA) 2 Department of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania (USA) 3 Department of Medicine, Pennsylvania Hospital, Philadelphia, PA (USA) 4 Department of Medicine, Section of Nephrology, Aga Khan University, Karachi (Pakistan) 5 Department of Medicine, Section of Cardiology, Philadelphia, VA Medical Center, Philadelphia, Pennsylvania (USA)

Abstract. – Background: Resistin has Introduction proposed link with obesity related insulin resistance and type 2 diabetes. The physiologic role of resistin in humans remains unknown. It is Resistin is a novel cysteine rich peptide suggested that circulating resistin levels are not hormone encoded at the RSTN gene secreted associated with obesity or insulin resistance in by the adipocyte. Since its discovery it has humans. However, the effects of weight loss on proposed link with obesity related insulin re- serum resistin concentration has not been stud- sistance and type 2 diabetes1. It is unclear ied. In order to better understand the physiolog- 1,2 ic role of resistin in human obesity, we mea- whether resistin production is increased or sured the serum resistin concentration in sub- decreased in obesity3. Expression of resistin jects with severe obesity (before and after 6- levels have been reported to be decreased in months of dietary intervention) to test the hy- white adipose tissue of several animal models pothesis that serum resistin concentrations are of obesity3,4. Additionally, Way et al reported elevated amongst individuals with severe obesi- that administration of PPAR-gamma ligands ty and weight loss would reduce these levels. Methods: Seventy-one obese subjects (de- actually increased resistin expression in both 3 fined as BMI > 35 kg/m2) who were randomized obese mice and Zucker diabetic fatty rats . to low fat (LF) vs low carbohydrates (LC) diets The physiologic role of resistin in humans and who completed the 6-month follow-up were remains unknown. Recently, it is suggested studied. Their baseline demographic information that circulating resistin levels are not associat- was collected and serum resistin, insulin, glu- ed with obesity or insulin resistance in hu- cose were measured at baseline and at 6- 5,6 months. mans . However, these are cross-sectional Results: Subjects in LC diet lost more weight studies and effects of weight loss were not than LF (-19.54 ± 7.87 lbs vs -7.83 ± 11.23 lbs., p studied. In order to better understand the = 0.001). Insulin sensitivity (HOMA) improved in physiologic role of resistin in human obesity, LC group compared with LF group [-3.72 ± 9.84 we measured the serum resistin concentra- (LC) vs +1.31 ± 7.31 (LF), p = 0.006]. Serum re- tions in patients with severe obesity to test the sistin levels did not decrease in either diet. hypothesis that serum resistin concentrations Conclusions: Our study found that despite a significant weight loss and improvement in in- are elevated amongst individuals with severe 2 sulin sensitivity there was no reduction in serum obesity (defined as BMI > 35 kg/m ), and resistin concentration in morbidly obese men weight loss would reduce these levels. Recent- with metabolic syndrome suggesting that re- ly we have published our findings of effects of sistin does not play a central role in obesity re- weight loss7 in patients with severe obesity by lated insulin resistance. using either low carbohydrate or low fat diets. Key Words: Here we are presenting our data on serum resistin levels in obese individuals before and Obesity, Resistin, Weight loss, Diet, Insulin. after 6-months of dietary intervention.

161 N. Iqbal, P. Seshadri, L. Stern, J. Loh, S. Kundu, T. Jafar, F.F. Samaha

Methods measured by radioimmunoassay (Laboratory Corporation of America, Raritan NJ). Insulin Subjects resistance was estimated by the by Home- The design and method of this study has ostasis Model Assessment [HOMA, fasting been reported before8. Briefly, the study was serum insulin (in micro units per liter) x fast- approved by the Institutional Review Board ing plasma glucose (in millimoles per at the Philadelphia Veterans Affairs Medical liter)/22.5]. The human resistin ELISA Center (PVAMC), and an approved consent (Biovendor, Inc) had an intra-assay cv of form was signed by each subject. Inclusion 2.4% and inter-assay cv of 5%. criteria were age ≥ 18 years and a body mass index (BMI) ≥ 35 kg/m2. Exclusion criteria Dietary Intake were a serum creatinine > 1.5 mg/dl, hepatic Dietary compliance was estimated by a disease, severe life-limiting medical illness, previously validated10 24-hour recall of di- inability of diabetic patients to self-monitor etary consumption obtained by interview at glucose levels, active participation in a di- baseline and 6 months, which was analyzed etary program, or use of weight loss medica- with Nutribase nutrition management soft- tions. Women, diabetic patients (either self ware (CyberSoft, Inc, Phoenix, AZ). As re- claimed, on hypoglycemic agents or having ported in our previous paper, our subjects fasting blood glucose ≥ 126 mg/dl on two sep- low carbohydrate dieters experienced a non- arate occassions), and severely obese (BMI ≥ significant reduction in calories and a signifi- 40 kg/m2) subjects were randomized separate- cant reduction in percentage of carbohy- ly (block randomization) to ensure approxi- drates and increase in percentage of protein mately equal numbers in the two groups. and fat compared with low fat dieters8. Subjects and investigators were not blinded during the study. Statistical Analysis For comparison of continuous variables be- Study Design tween the two groups, we calculated the The diet groups met separately in weekly 2- changes from baseline to 6 month in each hour group teaching sessions for the first 4 subject and compared the mean changes in weeks, followed by monthly one-hour ses- the two diet groups using an unpaired t-test11. sions, all led by individuals with expertise in We assessed the normality of the distribution nutritional counseling. Subjects received a diet of all variables before using the t-test. Insulin overview handout, instructional nutrition la- and glucose levels were skewed and were bels, sample menus and recipes, and a book on therefore log transformed for analysis. Di- counting calories and carbohydrates. No spe- chotomous variables were compared by chi- cific exercise program was recommended. The square analysis11. Linear regression and two- low carbohydrate (LC) group was instructed way analysis of covariance models were used to restrict carbohydrate intake to ≤ 30 grams to correct for potentially confounding vari- per day9. No instruction on restricting total fat ables and to identify interactions between intake was given. Vegetables and fruits with a variables and diet group assignment11. All p- high fiber-to-carbohydrate ratio were recom- values were two sided and a P-value of 0.05 mended. The Low Fat (LF) group received in- or less was considered to be statistically sig- struction in accordance with National Heart, nificant. Analysis was performed by using Lung and Blood Institute (NHLBI) obesity SPSS software (version 10.0). management guidelines9 including caloric re- striction to create a 500-calorie per day deficit, with ≤ 30 percent of these calories derived from fat. The quality of the counseling ses- Results sions was assessed by questionnaire adminis- tered to 15 subjects from each diet group. Baseline Characteristics Weights were measured initially and after 6 Thirty-nine subjects from LC and 32 from months on a single calibrated scale LF diet groups completed the study. The (SRScalesTM, SR Instruments Inc., Tanawan- baseline characteristics are given in Table I. da, NY. Serum insulin concentration was There were no differences between weights,

162 Serum resistin is not associated with obesity or insulin resistance in humans

Table I. Baseline characteristics.

Variable LF LC

No 32 39 Age (years) 54 ± 9 55 ± 9 Sex (n) 27 M 5 F 33 M 6 F AA* (n)(%) 18 (56) 19 (48) White (n) 12 (37) 19 (48) Resistin change (ng/ml) Latino (n) 2 (6) 1 (3) Dietary groups DM (n) 8 (25) 10 (25) Weight (lbs) 295 ± 52 294 ± 47 lbs Figure 2. Changes in Serum Resistin concentration. HOMA 7.45 ± 8.59 9.08 ± 11.91 Resistin (ng/ml) 11.1 ± 10.31 10.43 ± 6.10

*AA = African American. vs +1.31 ± 7.31 (LF), p = 0.006]. Resistin increased in both diets but the increase was not significant between diets [+0.29 ± 8.08 ng/ml HOMA and serum resistin concentrations at (LF) vs +2.74 ± 4.48 ng/ml (LC), p = 0.131]. baseline between these groups. . The subjects Within diet groups, changes were significantly were predominantly African American males different in LF for weight (p < 0.001) and in (52%) with the mean age of 54 ± 9 years. LC for weight, HOMA, and resistin (p < There were no differences between the 0.001) See (see Figure 1). baseline resistin and other anthropometric No correlation was found between changes parameters between subjects who completed in HOMA, weight or resistin in LF group. the study versus subjects who dropped out. Only in the LC diet were we able to demon- With regard to the baseline values per strate a correlation of weight and HOMA dietary group, for LF vs LC for weight (295 ± with no correlation of resistin to either of 52 lbs vs 294 ± 47 lbs), HOMA (7.45 ± 8.59 vs these variables. 9.08 ± 11.91) and resistin (11.1 ± 10.31 ng/ml vs 10.43 ± 6.10 ng/ml) were not significantly different between groups, respectively (Table I). Discussion Changes in Weights, Insulin Sensitivity and Serum Resistin Concentration We found that severely obese subjects with Subjects in LC lost more weight than LF a high prevalence of diabetes and metabolic (-19.54 ± 7.87 lbs vs -7.83 ± 11.23 lbs., p = syndrome who were randomized to LC diet 0.001). Insulin sensitivity (HOMA) increased lost more weight than individuals randomized in LC compared with LF [-3.72 ± 9.84 ± (LC) to LF calorie restricted diet in a 6 month pe- eight Change (Ibs) HOMA changes W

Dietary groups Dietary groups

Figure 1. Changes in weights. Figure 3. Changes in Insulin sensitivity (HOMA).

163 N. Iqbal, P. Seshadri, L. Stern, J. Loh, S. Kundu, T. Jafar, F.F. Samaha riod7. However, despite this weight loss and tween resistin and insulin sensitivity in non- an improvement in insulin sensitivity we did obese subjects. not find any reduction in serum resistin lev- There is also the issue of the actual ELISA els. This is the first study in humans that has used to measure serum resistin levels. While studied the effects of weight loss on serum re- the Biovendor assay is a highly specific sistin concentrations. Also, we did not find ELISA, with no cross reactivities to mouse any correlation between severity of obesity resistin or other human cytokines, it is a new- and serum resistin concentrations. ly designed assay and may still be in the Prior animal data have shown conflicting process of being fully developed. There is results regarding the role of resistin as a link some contradictory informations regarding between obesity and insulin resistance. While the exact normal values of serum resistin lev- Steppan and others1,2 have shown higher cir- els that need to be sorted further in the fu- culating resistin levels in genetic and diet in- ture. However, further studies by Pfutzer et duced models of obese mice and have shown al13 evaluated three commercially available worsening of insulin resistance with exoge- resistin ELISA with different epitopes nous resistin administration, others have not (Phoenix, Biovender, Immunodiagnostik), all found any such associations. In humans, Na- three passed successfully standardized techni- gaev et al12 found that resistin was expressed cal validation procedure with inter and intra only in very low levels in human fat cells with assay variability less than 10% and 15% re- no difference in expression in normal, insulin spectively. They proved to be different with resistant, and Type 2 diabetic individuals. regard to calibration and reference range Our findings are in agreement with other which may be linked to different antibody studies in humans that demonstrated no asso- specificity. No correlation was seen between ciation between mRNA and protein expres- any of the resistin assay and BMI and clinical sion and insulin resistance or obesity5,6. Also, measures for insulin resistance13. despite significant weight loss and an im- Fain et al14 found that resistin release in provement in insulin sensitivity in obese sub- human explants of adipose tissue was highly jects on low carbohydrate diet we did not see variable over 48 hours (3-158 ng/g). It is pos- any reduction in serum resistin levels. On the sible that serum resistin levels may fluctuate contrary, resistin levels appear to increase and sampling at various times during the 24 with weight loss though did not reach statisti- hour period may be necessary to further un- cal significance. derstand its secretion. However, Pfutzer et While our data supports the notion that re- al13 did not find any correlation between fast- sistin does not play a central role in the in- ing serum resistin level and clinical measures sulin resistance associated with obesity, it re- of obesity and insulin resistance. mains possible that it may have other au- In conclusion, our study found that despite tocrine and/or paracrine functions. a significant weight loss and improvement in It is also possible that inclusion of morbidly insulin sensitivity there was no reduction in obese men may be too restricted a study pop- serum resistin concentration in morbidly ulation to fully understand the role of resistin obese men with metabolic syndrome suggest- and insulin resistance. However, an initial re- ing that resistin does not play a central role in port by Lee et al5 reported a cross sectional obesity related insulin resistance. Given the study of 123 middle aged overweight women newness of the ELISA assay, the restricted (BMI of 30.9 ± 5.5 kg/m2 ) vs 118 lean healthy population, and the unknown kinetics of re- subjects that serum resistin levels did not cor- sistin, more studies are needed before the relate with markers of obesity or insulin resis- role of resistin in insulin sensitivity and obesi- tance. Also, a recent study6 showed resistin ty can fully be defined. concentrations were not different amongst non-obese (BMI: 25 ± 4.3 kg/m2), Obese (BMI: 33 ± 2.5 kg/m2) and obese individuals with type 2 diabetes (BMI: 34 ± 2.4 kg/m2). References

Serum resistin concentration was not related 1) STEPPAN C, BAILEY S, BHAT S, et al. The hormone re- to the percent of body fat, BMI or fat cell size sistin links obesity to diabetes. Nature 2001; 409: and demonstrated a weak relationship be- 307-312.

164 Serum resistin is not associated with obesity or insulin resistance in humans

2) STEPPAN CM, LAZAR MA. Resistin and obesity asso- 9) NHLBI, NIDDK. Obesity education initiative. In: ciated insulin resistance. Trends Endocrinol Clinical Guidelines on the Identification, Educa- Metab 2002; 13: 18-23. tion, and Treatment of Overweight and Obesity in Adults: The Evidence Report. NIH Publication no. 3) AY ORGUN ONG Adipose tis- W JM, G CZ, T Q, et al. 98-4083. Bethedsa, MD: NHBLI in cooperation sue resistin expression is severely suppressed in with the NIDDK; 1998. obesity and stimulated by peroxisome proliferator-activated receptor gamma agonists. J Biol 10) KARVETTI RL, KNITS LR. Validity of the 24-hour diet Chem 2001; 276: 25651-25653. recall. Am J Diet Assoc 1985; 85: 1437-1442.

4) FUKUI Y, M OTOJIMA K. Expression of resistin in adi- 11) DAWSON-SAUNDERS B, TRAPP RG. Basic and clinical pose tissue is modulated by various factors in- biostatistics. Norwalk, Conn.: Appleton and cluding peroxisome proliferator-activated receptor Lange, 1990. alpha. Diabetes Obes Metab 2002; 4: 342-345. 12) NAGAEV I, SMITH U. Insulin resistance and type 2 5) LEE JH, CHAN JL, YIANNAKOURIS N, et al. Circulating diabetes are not related to resistin expression in resistin levels are not associated with obesity or human fat cells or skeletal muscle. Biochem Bio- insulin resistance in humans and are not regulat- phys Res Commun 2001; 285: 561-564. ed by fasting or leptin administration: cross sec- 13) PFUTZER A, LAANGENFELD, M, KUNT T, L OBIG M, FORST tional and interventional studies in normal, insulin T. Evaluation of human resistin assays with resistant, and diabetic subjects. J Clin Endocrinol serum from patients with type 2 diabetes and dif- Metab 2003; 88: 4848-4856. ferent degrees of insulin resistance; Evaluation 6) HEILBRONN LK, ROAD J, JANDEROVA L, et al. Relation- studies. Clinical Lab 2003; 49: 571-576. ship between serum resistin concentration and in- 14) FAIN JN, CHEEMA PS, BAHOUTH SW, et al. Resistin sulin resistance in nonobese, obese, and obese release by human adipose tissue explants in pri- diabetic subjects. J Clin Endocrinol Metab 2004; mary culture. Biochem Biophys Res Commun 84: 1844-1848. 2003; 300: 674-678. 7) STERN L, IQBAL N, SESHADRI P, et al. The effects of low carbohydrate versus conventional weight loss diets in severely obese adults: One year follow- up of a randomized trial. Ann Intern Med 2004; ______140: 778-785. Acknowledgments 8) SAMAHA FF, IQBAL N, SESHADRI P, et al. A low-carbohydrate as compared with a low-fat diet in severe Supported by Funding from the Veterans Affairs obesity. N Engl J Med 2003; 348: 2074-2081. Healthcare Network Competitive Pilot Project Grant.

165