Resistin Polymorphisms Are Associated with Muscle, Bone, and Fat Phenotypes in White Men and Women


Division of Exercise Physiology, 8313 Robert C. Byrd Health Science Center, West Virginia University School of Medicine, P.O. Box 9227, Morgantown, WV 26506-9227. E-mail:


Objective: The biological function of resistin (RST) is unknown, although it may have roles in obesity, diabetes, and insulin resistance. The objective of this study was to examine the effects of single nucleotide polymorphisms (SNPs) in the human RST gene on muscle, bone, and adipose tissue phenotypes and in response to resistance training (RT).

Research Methods and Procedures: Subjects were white and consisted of strength (n = 482) and size (n = 409) cohorts who had not performed RT in the previous year. Subjects completed 12 weeks of structured, unilateral upper arm RT aimed at increasing the size and strength of the non-dominant arm, using their dominant arm as an untrained control. Strength measurements were taken pre- and post-12-week RT and consisted of elbow flexor isometric strength and one-repetition maximum during a biceps curl using free weights. Whole muscle, subcutaneous fat, and cortical bone volumes were measured by magnetic resonance imaging. Six RST SNPs were identified. Analysis of covariance was used to test for effects of the SNPs on pre- and post-muscle strength and whole muscle, fat, and bone volumes independent of gender, age, and body weight.

Results: Five RST SNPs (−537 A>C, −420 C>G, 398 C>T, 540 G>A, 980 C>G) were associated with measured phenotypes among subjects when stratified by BMI (<25, ≥25 kg/m2). Several gender-specific associations were observed between RST SNPs and phenotypes among individuals with a BMI ≥ 25. Conversely, only two associations were observed among individuals with a BMI < 25.

Discussion: These data support previous identified associations of RST with adipose tissue and demonstrate additional associations with bone and skeletal muscle that warrant further investigation.


Adipose tissue has been shown to have endocrine-like properties. Secreted factors from adipose tissue, including adiponectin, leptin, and resistin (RST),1 are collectively known as adipokines and have roles in the regulation of energy balance (for review, see (1). Other factors now known to be secreted from adipose tissue include insulin-like growth factor-1, interleukin-1, interleukin-6, tumor necrosis factor-α, plasminogen activator inhibitor, and fasting-induced adipose factor (2). The physiological roles of these factors are under investigation.

RST was named for its observed ability to resist insulin-stimulated responses in mice. Serum levels of RST were increased in obese mice, neutralization of RST in these mice decreased insulin resistance, and recombinant RST protein administration impaired glucose action in vivo (3). In vitro data from the same study linked RST with obesity and insulin resistance because adipocytes incubated with recombinant RST showed impaired glucose action. The addition of peroxisome proliferator-activated receptor (PPAR)-γ decreased RST expression (3). RST also interferes with glucose metabolism in the skeletal muscle of animals (4, 5), a finding that reconciles with the endocrine role originally proposed.

In humans, the endocrine role(s) of RST and its relevance to insulin resistance in obesity remain uncertain. No relationship could be demonstrated between RST and the degree of insulin resistance in a variety of clinical patients (6), and although RST mRNA was detected in adipose tissue samples of severely obese subjects (7), levels were not associated with BMI. Because human and mouse RST genes share only 53% homology, physiology may differ between them (8), and other roles of RST may exist in humans.

RST may be related to adaptations in skeletal muscle and bone. Serum RST levels negatively correlated with lumbar spine bone mineral density in non-obese middle-aged men (9). Basal glucose uptake and fatty acid uptake were both reduced, and the effects of insulin were inhibited in L6 myotube cultures containing either recombinant RST protein or transfection of mouse RST cDNA (4), possibly through a mechanism involving decreased expression of CD36, fatty acid transport protein 1, and a decrease in adenosine monophosphate-activated protein kinase phosphorylation (10). Muscle preparations exposed to recombinant RST protein showed reductions in basal and in stimulated glucose uptake (5). These responses were independent of changes in glucose transporter 4 translocation. Collectively, these data suggest that RST may hamper the adaptability of skeletal muscle. Whether such responses are endocrine, paracrine, or autocrine functions of RST is not known.

Numerous single nucleotide polymorphisms (SNPs) within both coding and non-coding regions of the human RST gene have been described (11, 12, 13, 14, 15, 16, 17, 18, 19, 20). However, associations among these SNPs and with obesity, diabetes, and BMI traits have been inconclusive. For example, the G allele of the −420 C>G SNP (also known as −180, −179) has been associated with high BMI in men (20) and the G/G genotype with the risk of insulin resistance (13) and type II diabetes (11). In contrast, the G allele of the same SNP was associated with low BMI and waist circumference in postmenopausal women of Portuguese descent (19). Furthermore, associations were found with weight-related traits and insulin sensitivity in Finnish diabetic and control patients and the frequency of the C allele of the −420 C>G SNP (15).

Given the intriguing nature of this adipokine and its potential links with adiposity, bone, and muscle tissue, we examined RST polymorphisms in association with muscle, bone, and fat volumes and with muscle strength before and in response to 12 weeks of structured resistance training (RT). Since RST is expressed by adipocytes (7), we hypothesized that expression is dependent on adiposity levels. Thus, we stratified our subjects based on a BMI ≥ 25, assuming that RST polymorphisms would be more likely to influence phenotypes when adiposity levels were higher. This idea was given further support based on recent findings indicating that overweight and obese individuals are limited in their strength response to a unilateral 12-week RT program as compared with normal-weight individuals (L. Pescatello, B. Kelsey, T. Price, R. Seip, T. Angelopoulos, P. Clarkson, P. Gordon, N. Moyna, P. Visich, R. Zoeller, H. Gordish-Dressman, S. Bilbie, P. Thompson, and E. Hoffman, unpublished data), suggesting that obesity-related mechanisms may impede muscle adaptability.

Research Methods and Procedures

Study Overview

This study is part of a multicenter study designed to uncover novel SNPs with associations to human muscle size and strength (functional SNPs associated with human muscle size and strength or FAMuSS). Detailed methodology of the study has been presented previously (21). Briefly, subjects were recruited to complete a 12-week progressive RT program aimed at increasing the strength and size of the elbow flexor and extensor muscles of the non-dominant arm only. Isometric and dynamic strength of the elbow flexors was determined before and after RT. Muscle size of the upper arm was measured by two methods: with a tape during muscle contraction and by magnetic resonance imagining (MRI) to determine muscle volume. DNA samples, obtained from subjects’ white blood cells, were collected before training and used to perform SNP analysis. Respective institutional review boards at each site approved the study protocol, and all subjects read and signed an informed consent document.

Subject Population

Men and women of all racial groups were included in this study if they were between the ages of 18 and 40 years and generally healthy. Additional exclusion criteria included the following: use of medications that affect skeletal muscle growth, restriction of activity for medical reasons, chronic medical conditions, metal implants that would prohibit MRI testing, RT or a job requiring repetitive use of the arms for at least 12 months prior, use of supplements designed to improve muscle size or strength (creatine, protein, etc.), habitual alcohol use, pregnancy, or use of the birth control shot.

Anthropometric Measurements

Body weight was measured before and after 12 weeks of RT. Stature was measured with a standard tape, and BMI was calculated based on pre- and post-training values (kilograms per meter squared). The waist circumference was measured with a standard tape as the narrowest point between the umbilicus and the xiphoid process of the sternum and subcutaneous fat measured using standard skinfold techniques over the biceps brachii and triceps brachii (22).

Maximal Isometric Contraction Test [Maximal Voluntary Contraction (MVC)]

MVC of the elbow flexors were tested before and after 12 weeks of RT on a specially designed and modified preacher bench using a strain gauge attached to a strength evaluation system (model 32628CTL; Lafayette Instrument Co., Lafayette, IN). Baseline isometric testing was performed over 3 days, separated by no more than 48 hours, and over 2 days at the completion of training. The average of the 2nd and 3rd testing days was used as the baseline criterion measurement. Subjects were positioned with their arms at a 90° angle at the medial epicondyle in line with the axis of rotation of the bench. Three tests lasting 3 seconds were performed on each arm, all separated by a 1-minute rest period. The three peak force values were averaged for each testing day. Strength gain was determined by calculating the percentage difference between post-training and pre-training maximal values.

One-Repetition Maximum (1RM) Testing

A 1RM protocol modified from Baechle et al. (23) was used to measure dynamic strength of the elbow flexors on a standard preacher bench before and after 12 weeks of RT. Two warm-up sets were completed at 50% and 75% of the predicted 1RM for eight repetitions and five repetitions, respectively. Single attempts were performed until one single repetition with full-range of motion was completed. Dynamic strength gain was determined by calculating the percentage difference between post-training and pre-training 1RM values.

Volumetric Measurements

MRI was used to assess changes in whole muscle, subcutaneous fat, and cortical bone volume of each arm. Pre-training MRI scans were performed before or 48 hours after any isometric or 1RM testing, and post-training MRIs were performed 48 to 96 hours after the last training session to avoid any temporary water shifts that would skew measurements while also avoiding detraining. Before the MRI, the maximal circumference of the upper arm was assessed with the subjects’ arms abducted 90° at the shoulder, 90° at the elbow, and maximally contracted. The point of maximal circumference was marked with a radiographic bead (Beekley Spots; Beekley Corp., Bristol, CT).

Rapidia, a personal computer-based, three-dimensional, interactive system for viewing images from computerized tomography and MRI scans (3D Med Co. Ltd., Seoul, Republic of Korea), was used for all volumetric measurements. To ensure accurate and reliable measurements, six slices from each image were analyzed using the metaphyseal-diaphyseal junction landmark, making sure the same regions were measured from pre- and post-images. Cortical bone, subcutaneous fat, and muscle were isolated using image signal intensity differences between tissues, and once the region of interest was segmented, total volume was taken for the six evaluated slices. Repeatability and reliability of Rapidia volume measurements were verified using a phantom of known volume.

Exercise Training Program

A monitored, progressive 12-week unilateral upper arm RT program was initiated following baseline anthropometric and strength measurements. The 1RM values were used to assign appropriate weights in the following training progression: weeks 1 to 4, three sets at 12RM; weeks 5 to 9, three sets at 8RM; and weeks 10 to 12, three sets at 6RM. The order of exercises was as follows: preacher curl, seated overhead triceps extension, concentration curl, triceps kickback, and standing biceps curl. Each repetition was performed through a full range of motion using dumbbells (Power Blocks; Intellbell Inc., Owatonna, MN). The subject trained their non-dominant arm twice per week with a 24- to 48-hour rest period between training sessions. Although the primary focus of testing was the biceps brachii, the triceps brachii were trained to avoid creating a muscle imbalance across the joint.

Dietary Control Procedures

All subjects were instructed to maintain their normal diet for the duration of the study in an attempt to avoid significant weight gain or loss. In addition, data obtained from subjects that exhibited a significant weight loss over the course of the study were not included in the final analyses.

Standardization between Sites

Testing and training of subjects was performed at eight centers. To minimize variability in training-induced responses among various sites, the following measures were undertaken. First, identical testing and training equipment was purchased for use at each site. Second, all testing, training, and MRI methods were videotaped and made available to each site. All research personal reviewed the taped procedures before the start of each new cohort of subjects. Third, site representatives met twice per year to review the progress of the study and to review the standardization of measurements.


Blood samples were obtained from all individuals in EDTA anticoagulant, sent to the coordinating site in Washington, DC, without subject identification, and DNA was solated using QIAGEN kits (QIAGEN, Valencia, CA). Genotypes for all RST SNPs (Figure 1) in this paper were obtained with the use of a TaqMan allelic discrimination assay that employs the 5′ nuclease activity of Taq polymerase to detect a fluorescent reporter (VIC and FAM) signal-generated during polymerase chain reactions (PCRs). The PCR for the each SNP contained 20 ng of DNA, 900 nM primers, 200 nM probes, and TaqMan Universal PCR Master Mix, No AmpErase UNG (Applied Biosystems, Foster City, CA) in a final volume of 15 μL. PCR was performed on an MJ Research Tetrad thermal cycler (Waltham, MA). The primers and probes for each SNP are located in Table 1. The PCR profile was 10 minutes at 95 °C (denaturation), 44 cycles of 15 seconds at 92 °C, and 1 minute at an annealing temperature of 60 °C. All PCRs were analyzed using an ABI 7900 (Applied Biosystems).

Figure 1.

Structure of chromosome 19 and identified SNPs in the human RST gene. The start and stop regions (ATG, TGA) and the positions of the six polymorphisms are presented.

Table 1.  TaqMan primer sets for SNPs tested in resistin
SNPForward primerReverse primerWild-type allele probe (5′ VIC)

Statistical Analyses

Due to the low percentage of participants representing other ethnic groups, all analyses were performed in whites (strength, n = 482; size, n = 409), which accounted for 75% of the total study population. Hardy-Weinberg equilibrium was determined for RST SNPs using a χ2 test to compare the observed genotype frequencies to those expected under Hardy-Weinberg equilibrium (Table 2). All RST polymorphisms were within Hardy-Weinberg equilibrium (p > 0.5). Twelve size/strength phenotypes were analyzed as continuous quantitative traits (baseline measures of whole-muscle, cortical bone, and subcutaneous fat volume; 1RM strength and MVC strength; difference in whole-muscle, cortical bone, and subcutaneous fat volume; 1RM strength and MVC from baseline to post-exercise; and the percentage change in 1RM strength and MVC from baseline to post-exercise). Normality of each quantitative trait was confirmed using the Shapiro-Wilk normality test. Bivariate analyses of each quantitative measurement showed several significant correlations with both age and baseline body mass; therefore, associations between RST and size/strength phenotypes were assessed using analysis of covariance methods. Due to large gender differences in baseline values and the response to training, all analyses were performed separately for men and women, and all analyses of covariance included age and/or baseline body mass as covariates. Further analyses of covariance, with age as only covariate, were performed on gender-specific cohorts stratified by baseline BMI (BMI < 25 vs. BMI ≥ 25).

Table 2.  Genotype frequencies in RST gene SNPs
−537 A>CAA53490.3690.36
−420 C>GCC27446.2846.28
30 C>TCC57197.1197.11
398 C>TCC37263.1663.16
540 G>AAA6210.5310.53
980 C>GCC22037.2937.29

All significant associations from the main analysis of covariance model were subjected to pair-wise statistical tests between each of the three genotype groups for each RST SNP. Linear tests were performed between each of the genotype groups to determine which genotype groups were significantly different from one another. The resulting p values from these linear tests were adjusted for multiple comparisons using the Sidak post hoc multiple comparison test. Linear regression analysis, including likelihood ratio tests between full and constrained models, were performed to estimate the proportion of variance in muscle size/strength measurements attributable to RST genotype.


Subject Characteristics

Among those who completed the study, strength data were complete for 615 and size data for 529 subjects, respectively. However, due to the low percentage of participants representing other ethnic groups (Table 3), all analyses were performed in whites (strength, n = 482; size, n = 409), which accounted for 75% of the total study population. The characteristics of each cohort are presented in Table 3. There were no significant differences when comparing the age or BMI of the women and men, respectively, in either cohort (Table 3).

Table 3.  Subject characteristics and demographic characteristics
 Strength cohortSize cohort
  • *

    Significantly different mean BMI in strength cohort between white men and women (p < 0.001).

  • Significantly different mean age in size cohort between white men and women (p = 0.036).

Ethnic group  
 African-American27 (4.6%)16 (3.8%)
 Asian52 (8.8%)43 (10.3%)
 White465 (78.6%)323 (77.1%)
 Hispanic24 (4.0%)19 (4.5%)
 Other24 (4.0%)18 (4.3%)
White women286 (81.5%)203 (80.9%)
 Age [years (mean SD)]23.8 (5.9)23.6 (5.6)
 BMI [kg/m2 (mean SD)]23.9 (4.3)*23.8 (4.4)
 ≥25 (mean SD)29.2 (3.7)29.4 (3.9)
White men179 (74.3%)120 (71.4%)
 Age [years (mean SD)]24.7 (5.9)25.1 (6.0)
 BMI [kg/m (mean SD)]24.9 (4.3)*24.8 (4.1)
 ≥25 (mean SD)28.8 (3.9)28.3 (3.4)

RST SNP Associations when Stratified by BMI

Several associations were observed in those with a BMI ≥ 25. In addition to being BMI dependent, many RST SNP associations were also gender specific. The RST SNP associations with strength, muscle, bone, and fat phenotypes in women are presented in Table 4. Among female subjects, eight associations were observed between RST SNPs and measured phenotypes, and all were among those with a BMI ≥ 25. Five phenotypes were associated with the 398 C>T SNP. In addition, one promoter SNP (−537A>C) was associated with baseline subcutaneous fat volume (p < 0.01). The RST SNP associations with strength, muscle, bone, and fat phenotypes in men are presented in Table 5. Eleven associations with measured phenotypes were observed, nine among men with a BMI ≥ 25. Among those with a BMI ≥ 25, six phenotypes were associated with the −420 C>G SNP, and three phenotypes were associated with the 540 G>A SNP. In contrast, two RST SNPs were associated with measured phenotypes in men with a BMI < 25. Specifically, −537 A>C was associated with the difference in whole-muscle volume and 980 C>G with the difference in cortical bone volume in men only.

Table 4.  RST SNPs and muscle, bone, and fat phenotypes in women when stratified by BMI
SNPPhenotype measurementGender/BMIGenotypeNMean + standard errorpVariability of genotype (%)
  • *,†

    Denote significantly different means; means with like symbols are statistically different at the corresponding p value.

  • **

    Denotes p value from likelihood ratio test comparing the full model (phenotype, genotype, and covariates) to the constrained model (phenotype and covariates only).

−537 A>CBaseline subcutaneous fat volume (mm3)Women, BMI ≥ 25AA53372,453.0 ± 13,702.0*0.007*11.8, p = 0.005**
   AC451766.0 ± 49401*  
398 C>TPercentage change in 1RM strength (relative)Women, BMI ≥ 25CC6162.0 ± 4.6*0.041*7.5, p = 0.023**
   CT2755.2 ± 6.90.023 
   TT2129.0 ± 26.2*†  
398 C>TPercentage change in isometric strength (relative)Women, BMI ≥ 25CC5723.8 ± 2.6*0.038*7.5, p = 0.037**
   CT2511.8 ± 3.9*  
   TT114.9 ± 19.8  
398 C>TBaseline whole-muscle volume (mm3)Women, BMI ≥ 25CC41514,508.0 ± 32,521.0*0.024*8.9, p = 0.020**
   CT17652,857.0 ± 49,917.0*  
398 C>TBaseline cortical bone volume (mm3)Women, BMI ≥ 25CC4115,982.0 ± 404.0*0.013*10.7, p = 0.010**
   CT1814,117.0 ± 603.0  
398 C>TDifference in cortical bone volumeWomen, BMI ≥ 25CC41−456.5 ± 277.4*0.005*13.6, p = 0.003**
   CT181012.9 ± 413.6*  
980 C>GDifference in whole muscle volumeWomen, BMI ≥ 25CC1716,860.0 ± 10,300.7*0.031*11.5, p = 0.028**
   CG3151,233.3 ± 7642.3*  
   GG1035,016.9 ± 14,099.4  
980 C>GDifference in cortical bone volumeWomen, BMI ≥ 25CC17−571.3 ± 322.7*0.031*10.6, p = 0.038**
   CG31622.4 ± 352.4*  
   GG10−744.3 ± 1810.6  
Table 5.  RST SNPs and muscle, bone, and fat phenotypes in men when stratified by BMI
SNPPhenotype measurementGender/BMIGenotypeNMean + standard errorpVariability of genotype (%)
  • *,†

    Denote significantly different means; means with like symbols are statistically different at the corresponding p value.

  • **

    Denotes p value from likelihood ratio test comparing the full model (phenotype, genotype, and covariates) to the constrained model (phenotype and covariates only).

−537 A>CDifference in whole-muscle volumeMen, BMI < 25AA5867,350.8 ± 4741.2*0.015*8.9, p = 0.013**
   AC999,967.6 ± 12,122.1*  
−420 C>GDifference in 1RM strength (absolute)Men, BMI ≥ 25CC3110.4 ± 0.8*0.026*8.1, p = 0.026**
   CG419.2 ± 0.7  
   GG105.9 ± 1.5*  
−420 C>GPercentage change in 1RM strength (relative)Men, BMI ≥ 25CC3137.2 ± 3.0*0.031*7.7, p = 0.028**
   CG4131.7 ± 2.7  
   GG1021.3 ± 5.3*  
−420 C>GPercentage change in isometric strength (relative)Men, BMI ≥ 25CC2916.7 ± 2.90.030*9.3, p = 0.029**
   CG3420.7 ± 2.7*  
   GG105.9 ± 4.9*  
−420 C>GBaseline whole-muscle volume (mm3)Men, BMI ≥ 25CC19564,588.0 ± 36,694.0*0.030*12.6, p = 0.026**
   CG29691,608.0 ± 29,523.0*  
   GG6630,775.0 ± 64,698.0  
−420 C>GBaseline cortical bone volume (mm3)Men, BMI ≥ 25CC1921,742.0 ± 652.0*0.036*11.8, p = 0.032**
   CG2920,848.0 ± 524.0  
   GG618,292.0 ± 1149.0*  
540 G>ADifference in 1RM strength (absolute)Men, BMI ≥ 25AA85.8 ± 1.6*0.043*7.4, p = 0.037**
   GA358.9 ± 0.8  
   GG3910.2 ± 0.7*  
540 G>APercentage change in 1RM strength (relative)Men, BMI ≥ 25AA820.7 ± 5.9*0.047*7.4, p = 0.026**
   GA3530.6 ± 2.9  
   GG3936.6 ± 2.7*  
540 G>ADifference in whole-muscle volumeMen, BMI ≥ 25AA6111,651.1 ± 17,477.3*0.042*10.1, p = 0.035**
   GA2562,366.3 ± 8486.7*  
   GG2267,103.2 ± 9087.2  
980 C>GBaseline whole-muscle volume (mm3)Men, BMI ≥ 25CC18724,162.0 ± 37,781.0*0.035*12.7, p = 0.025**
   CG8596,695.0 ± 30,104.0*  
   GG8603,259.0 ± 55,915.0  
980 C>GDifference in cortical bone volumeMen, BMI < 25CC2110.0 ± 502.8*0.049*14.3, p = 0.006**
   CG37−517.6 ± 376.90.006 
   GG82389.7 ± 819.6*†  


Polymorphisms within the human RST gene have previously been associated with measures of obesity (19, 20), insulin resistance (16, 18), and diabetes (14). Our findings now extend these associations to skeletal muscle strength and muscle, bone, and adipose tissue volumes before and after 12 weeks of structured unilateral RT. Moreover, these novel observations were dependent on BMI and gender.

Numerous associations were observed when subjects were stratified by BMI. Given that RST mRNA is expressed by adipocytes (7), we used BMI ≥ 25 as a surrogate measure of adiposity, which is the NIH-accepted cut-off for overweight (24). We hypothesized that individuals with a higher BMI would have more adiposity and that RST SNP associations with the measured phenotypes were likely adiposity dependent. As such, we observed numerous associations with strength, muscle, bone, and adipose measures among individuals with a BMI ≥ 25. Moreover, associations between RST genotype and the various phenotypes among individuals with a BMI ≥ 25 were strong. The percentage of the phenotypic variance that can be explained by each RST genotype for individuals with a BMI ≥ 25 ranged between 7% and 12%. Although this measure may be slightly inflated due to the homogeneity of the study population when stratified by BMI, these ranges seem robust enough to indicate a strong physiological association. Although the biological role of RST remains to be fully elucidated, the associations of RST SNPs and various phenotypes observed in this study identify potential areas for future research.

Numerous associations were observed in the current study between three SNPs (−420C>G, 398 C>T, 540G>A) and phenotypic measures of fat, bone, and muscle that were also gender dependent. Among white men, numerous associations with the −420 C>G and 540 A>G RST SNPs and the measured phenotypes were observed. Specifically, men with a BMI ≥ 25 that were homozygous for the mutant allele −420 (G/G) had lower increases in absolute and relative strength and decreased cortical bone volume when compared with those that were homozygous C/C. In addition, overweight white men that were homozygous for 540 A/A had lower strength and muscle volume when compared with those that were homozygous G/G. These data suggest that overweight white men with allele-specific RST polymorphisms may be less likely to adapt to an RT program. The mechanisms for these associations are not currently known.

The observed associations between 540 A>G SNP and strength and muscle phenotypes, to our knowledge, is the first report linking this polymorphism to any phenotypes. However, numerous studies have linked the −420 RST SNP with various obesity phenotypes (19, 20), although the present study is the first to identify associations in strength, muscle, and bone phenotypes. Recent data have demonstrated that the −420 G allele is a major determinant of increased promoter activity (25) of the RST gene and serum RST levels (26), although other studies have not found associations between RST SNPs and serum RST levels. Mice heterozygous for the RST gene were shown to have normal levels of circulating RST when compared with homozygous normal mice (27). This was explained by a compensatory up-regulation in the heterozygous mice that suggests that serum RST levels may be independent of genotype in this model. Nevertheless, whether differences in RST mRNA activity and/or circulating protein concentrations were responsible for the present findings is unknown. To date, the relations between serum RST concentrations and various phenotypes have been inconsistent, suggesting that RST polymorphisms may be linked to phenotypes by a mechanism other than serum protein levels.

Among white women, the 398 intronic SNP was associated with numerous phenotypes. Women heterozygous (CT) for the 398 C>T SNP having a BMI ≥ 25 demonstrated larger baseline muscle volumes and smaller baseline cortical bone volumes than women homozygous (CC) for this SNP. However, after training, the heterozygous women increased their cortical bone volumes but showed less strength improvement when compared with the homozygous women. The physiological mechanisms to account for these findings are unclear, but given the robust nature of this data set and the proportionately large percentage of the variance that is attributable to these polymorphisms, further investigation is warranted. Indeed, previous reports have also demonstrated gender-specific associations between the human RST gene and phenotypes (19, 28). It is conceivable that the biological role of RST differs between genders. Serum RST concentrations seem to be higher in women than in men (29, 30, 31), although animal experiments have suggested that estrogen is a negative regulator of RST mRNA in adipocytes (32).

Although the link between RST SNPs and skeletal muscle strength and muscle and cortical bone volumes before and after structured RT are, indeed, novel findings, recent data may have begun to shed some light on the biological role of RST in humans and, perhaps, provide clues that may link it to these measured phenotypes. RST expression in both adipose tissue and skeletal myogenic cultures has been shown to be affected by two specific transcription factors, namely E/EBP (CCAAT/enhancer-binding protein) and PPAR (33, 34) and several signaling pathways such as phosphoinositide 3-kinase/Akt and mitogen-activated protein kinase (35), all of which may alter RST gene expression. For instance, a PPARγ-specific response element upstream of the RST coding sequence has been identified (34) that inhibits RST expression in adipocytes (33) and plays an important role in differentiation of mesenchyme (36). This has been a suggested link between RST and bone (9) because bone marrow cells are a common precursor to osteoblasts and adipocytes. It is plausible that RST SNP associations with cortical bone may be mediated along this pathway.

Although the RST receptor has not been identified, several signaling pathways are affected by the RST peptide, which may link to skeletal muscle. Smooth muscle cells proliferate in the presence of RST, and this was associated with activation of the mitogen-activated protein kinase and phosphoinositide 3-kinase/Akt pathways (35). Furthermore, previous research has shown that recombinant RST inhibited basal glucose uptake in cultured myotubes while dose-dependently inhibiting the insulin-mediated effects in skeletal muscle (4). Thus, it is clear that RST may alter the metabolic properties of muscle, possibly though inhibition of insulin-mediated or other signaling pathways. However, autocrine and/or paracrine roles may also be plausible. As such, the specific role RST polymorphisms play in skeletal muscle remains to be determined.

This study is not without its limitations. Some of the phenotypes are related to each other, and an association with one phenotype may be due to the same genetic effect that is observed in another phenotype (i.e., absolute difference 1RM and relative difference 1RM). Nevertheless, given that the phenotypes are physiologically different, we felt it was important to examine these associations individually. Also, we cannot completely eliminate the possibility of some false positives in these data. Yet, the polymorphisms in the human RST gene were associated with physiologically different phenotypes (i.e., variation in fat, bone, and muscle and with the muscle response to structured RT). These novel associations were gender-specific and adiposity-dependent, as evidenced when subjects were stratified by BMI. As such, we believe the novelty of the associations and the strong genetic effects observed warrant further investigation as the biology of RST in humans remains to be elucidated.


This work was supported by NIH Grant RO1 NS40606, with co-support by the National Institute of Neurological Disorders and Stroke, the National Institute of Arthritis and Musculoskeletal and Skin Diseases, and the National Institute on Aging.


  • 1

    Nonstandard abbreviations: RST, resistin; PPAR, peroxisome proliferator-activated receptor; SNP, single nucleotide polymorphism; RT, resistance training; MRI, magnetic resonance imagining; MVC, maximal voluntary contraction; 1RM, one-repetition maximum; PCR, polymerase chain reaction.

  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.