SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References

Objective: Epidemiological studies showing an association between the melanocortin-4-receptor (MC4R) 103I variant (rs2229616) and decreased BMI are complemented by functional studies; these suggest a mechanism for appetite regulation and a linkage signal for physical activity and dietary intake for the region encompassing the MC4R. This study aims to provide epidemiological evidence for showing the association of this polymorphism with features of the metabolic syndrome and with parameters related to energy expenditure and dietary habits as potential mediators.

Methods and Procedures: We analyzed this polymorphism in 7,888 adults of a population-based cross-sectional study applying regression-based statistical models.

Results: Carriers of the MC4R 103I (3.7%) exhibited a significantly decreased waist circumference (–1.46 cm, P = 0.020), decreased glycosylated hemoglobin (HbA1c) (–0.09%, P = 0.040), and increased HDL-cholesterol (HDL-C) (+1.76 mg/dl, P = 0.056), but no change in blood pressure. The odds of having three or more components of the metabolic syndrome were substantially reduced among carriers of MC4R 103I (odds ratio (OR) = 0.46, P = 0.003). Controlling for BMI reduced the HbA1c and HDL-C association. Mediator analyses revealed a borderline association of MC4R 103I with carbohydrate intake (OR = 1.26, P = 0.059) possibly mediating association with leanness.

Discussion: Our representative study of well-phenotyped Europeans is the first to describe the association of the MC4R V103I with the metabolic syndrome and with a nutrient-related phenotype. Our data support the idea that this polymorphism plays a role in appetite regulation that not only affects BMI, but also other features of the metabolic syndrome. It further establishes that the association of the MC4R V103I with obesity and related phenotypes is genuine.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References

The melanocortin-4 receptor (MC4R) is expressed in the hypothalamus and is part of the melanocortinergic pathway controlling energy homeostasis. The MC4R gene was suggested as a candidate gene for obesity (1,2,3). The MC4R variants, which are mostly rare mutations, account for up to 6% of morbid obesity in humans (4,5,6) and the role of this gene is supported by functional studies (7,8). The most common variant, the isoleucin allele (A allele) of the V103I polymorphism (rs2229616; G/A) with a frequency of 3–4%, was shown to be associated with decreased BMI in a meta-analysis (2) and in our own cohort study comprising a large representative sample of ∼8,000 whites from Southern Germany (1). Our study had shown a difference in BMI of −0.52 kg/m2 (P = 0.043) for A allele carriers compared to the wild type and an odds ratio (OR) of 0.75 (P = 0.017) for obesity defined as BMI ≥ 30 kg/m2.

Recent functional in vitro studies established a molecular mechanism potentially underlying this association of the MC4R 103I with enhanced MC4R function and thus possibly with appetite regulation leading to a lean phenotype (9). Furthermore, a linkage signal was described for physical activity and dietary intake for the region encompassing the MC4R (10). This together with the described role of the MC4R in modulating food intake in mice (6,7,8,9,10,11) leads to the hypothesis that this polymorphism affects BMI via appetite regulation or energy expenditure, a mechanism which would also have a potential influence on lipid and glucose metabolism, and therefore on parameters related to the metabolic syndrome.

The metabolic syndrome is defined as a cluster of risk factors that strongly predispose to type 2 diabetes mellitus (T2DM) (12,13), cardiovascular disease (14,15), kidney disease (16) as well as a higher mortality (17). A recent study confirmed again its clinical value (18). While the specific definition of the metabolic syndrome is broadly discussed (19,20,21) (reviewed in (22)), its importance for public health is undisputed.

We therefore investigated the association between the MC4R V103I polymorphism and features of the metabolic syndrome. We also adjusted for BMI in order to test for an association additional to the known influence of the V103I polymorphism on BMI. Furthermore, we explored the role of several parameters related to energy expenditure and dietary habits as mediators in this association.

Methods and procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References

The Study Subjects

This analysis is based on 9,111 participants from two population-based surveys conducted in the years 1994/95 (MONICA/KORA S3) and 1999–2001 (KORA S4) in Southern Germany, which formed the basis of the previously described association of the V103I polymorphism with obesity (1). Of these 9,111 subjects all with German citizenship and 99.5% being European whites, 7,888 with information on genotype, BMI, waist circumference, high-density lipoprotein cholesterol (HDL-C), HbA1c levels and blood pressure (BP) were included in this analysis. The potential of population stratification was reported to be small (23). Details of the surveys are reported elsewhere (24). All study participants gave informed written consent according to the ethics committee of the Bavarian Medical Association and every attempt was made to ensure anonymity of the participants.

Phenotyping

All study participants underwent a standardized face-to-face interview by certified medical staff and a standardized medical examination including blood draw, anthropometric measurements, systolic and diastolic BP, pulse rate, and triglyceride, HDL-C and glycosylated hemoglobin (HbA1c) concentrations were assessed. Diagnosis of T2DM was obtained by self-reporting or reporting of anti-diabetic medication. For a subgroup of 1,614 subjects in an overnight fasting status study, fasting triglyceride levels were available. Educational level was categorized as <10 years of schooling, 10–11 years, or >11 years. A smoker was defined as a participant who reported smoking currently at least one cigarette/day. From self-reported information, alcohol consumption was calculated in grams/day (g/day) (25) and dichotomized at 40 g/day for men and 20 g/day for women. A four-category seasonal physical activity score was assessed from questions on sports in summer and winter (1 = more than 2 h of sports weekly in leisure time, 2 = 1–2 h, 3 = <1 h, 4 = none) and dichotomized as 1 = high activity (sum of summer and winter scores = 2–4), 0 = low activity (sum of summer and winter scores = 5–8) (26). Scores of the frequency of consuming fat or carbohydrate containing foods were constructed based on a validated qualitative food frequency (27) and dichotomized at the sex-, age group, and survey-specific median (high/low carbohydrate or fat score).

Since most of the study participants were not fasting in a standardized way, we were not able to use triglyceride and glucose levels for the diagnosis of a metabolic syndrome. Instead, we defined a surrogate measure for “metabolic syndrome” using the following four components: (i) waist circumference >102 cm or >88 cm for men or women, respectively, (ii) HDL-C <40 mg/dl or <50 mg/dl for men or women, respectively, (iii) BP >130/85 mm Hg, and (iv) HbA1c >6.0% or diagnosis of T2DM. A subject was defined as having the “metabolic syndrome” when three or more of these components were present, in close relation to the ATPIII definition (19,20,21). Since HbA1c reflects mean glycemia over the preceding 2–3 months, it is not expected to be influenced by the fasting state (28,29). HDL-C is only marginally if at all influenced by the postprandial status (30,31,32).

Genotyping

Genotyping of the rs2229616G>A polymorphism was performed using a matrix-assisted laser desorption/ionization time-of-flight mass spectrometry system (Sequenom, Mass EXTEND, San Diego, CA) as described previously (2). The genotyping success rate was 98.5%.

Statistical methods

Subjects with one MC4R 103I allele (heterozygous subjects, genotype G/A) were compared with the subjects being homozygous for the MC4R V103 (genotype G/G). Hardy–Weinberg equilibrium was tested via the chi-square test for the two surveys separately. The one man with a BMI of 24 kg/m2 being homozygous for the 103I allele (genotype A/A) was collapsed into the heterozygous group.

In the primary analysis, the association of the polymorphism with the four quantitative features of the metabolic syndrome (waist circumference, HDL-C, HbA1c, BP) was tested using linear regression adjusting for age, sex, educational status, and survey. The outcomes HDL-C and HbA1c, which were not normally distributed, were additionally tested on the log-scale using the same adjusting variables and via the non-parametric Wilcoxon test, which allows no adjustment, as a sensitivity analysis. A simultaneous association of the polymorphism with all four features of the metabolic syndrome was assessed using the Hotelling–Lawley statistic, accounting for multiple testing in the four quantitative features: via OR for having the “metabolic syndrome” using logistic regression; via OR for each increase in the number of components via ordinal logistic regression, and by testing for a trend in the proportion of the G/A genotype among subjects with 0, 1, 2, 3, 4 applicable components of the “metabolic syndrome.” To explore the potential effect from population stratification, we performed sensitivity analyses including height or an indicator for being born in Germany as covariates into the linear regression models.

In a secondary analysis, we tested other quantitative measures of body size (hip circumference, percentage body fat, height) as well as fasting levels of triglycerides in 1,614 individuals in the fasting state via linear regression.

Furthermore, we explored the binary variables (high/low pulse rate, high/low physical activity, smoking yes/no, high/low alcohol intake, and high/low carbohydrate or fat score), for their potential as mediators in the association of the polymorphism with waist circumference, BMI, or the “metabolic syndrome” according to operational definitions (33) and guidelines for surrogacy analyses (34). This included the following models also illustrated in Figure 1: (i) the genotype is associated with the outcome; (ii) the mediator is associated with the outcome; (iii) the genotype is associated with the mediator, and (iv) including the mediator as an additional covariate into model (v) eradicates the association between genotype and outcome, while the mediator maintains its association with the outcome.

image

Figure 1. : The four models to test whether a variable MEDIATOR is in the pathway of the association of the GENOTYPE with the OUTCOME: If, firstly, the GENOTYPE is significantly associated with the OUTCOME (model (a)), secondly, the MEDIATOR is significantly associated with the OUTCOME (model (b)), and, thirdly, the GENOTYPE is significantly associated with the MEDIATOR (model (c)), then a variable is considered a mediator if the GENOTYPE–OUTCOME association (model (a)) is not significant any more when including the MEDIATOR as a covariate (model (d)).

Download figure to PowerPoint

All regression analyses were adjusted for sex, age, survey, and educational level, and the impact of additional adjustment for BMI was evaluated. All analyses were performed with SAS software, version 9.1.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References

Descriptive Statistics On Study Participants

The participants were 25–74 years of age and approximately half were male. The G/A genotype occurred in 3.68% of the participants (n = 291 out of 9,111), the A allele in 1.85% of the chromosomes. The P values for testing for HWE violation were 0.22 and 0.26 for the two surveys. Gender-stratified descriptive statistics of the phenotypes and their correlation coefficients are provided in Tables 1 and 2.

Table 1. . Descriptive statistics of the analyzed phenotypes for 7,888 subjects
 Mean ± s.d. or %
PhenotypeMen (n = 3,912)Women (n = 3,976)
  • a

    Out of: (i) waist >102 cm for men or >88 cm for women, (ii) high-density lipoprotein-cholesterol (HDL-C) <40 mg/dl for men or <50 mg/dl for women, (iii) blood pressure ≥130/85 mm Hg, (iv) HbA1c >6.0% or type 2 diabetes mellitus.

General  
    Age (years)49 ± 1449 ± 14
    BMI (kg/m2)27.4 ± 3.926.7 ± 5.2
    Type 2 diabetes mellitus4.4%3.7%
Features of the metabolic syndrome (quantitative)  
    Waist (cm)96.5 ± 10.784.0 ± 12.5
    HDL-cholesterol (mg/dl)49.5 ± 14.062.1 ± 16.8
    HbA1c (%)5.37 ± 0.775.45 ± 0.69
    Systolic/diastolic blood pressure (mm Hg)134.2 ± 17.9/125.8 ± 19.8 /
    Fasting triglycerides (mg/dl) (n = 1,614)82.7 ± 10.978.0 ± 10.5 117 ± 55
 153 ± 125 
Number of components of the metabolic syndromea (categorical)  
    025.3%37.4%
    141.6%31.6%
    223.2%19.2%
    38.8%9.5%
    41.2%2.3%
    ≥3 components (“metabolic syndrome”)10.0%11.9%
Further parameters related to body size  
    Hip (cm)104.8 ± 7.0104.3 ± 10.4
    Percentage body fat (%)28.3 ± 6.336.2 ± 10.4
    Height (cm)174.6 ± 7.0161.7 ± 6.5
Parameters related to energy expenditure or energy intake  
    Pulse rate (min)72.6 ± 11.073.7 ± 10.0
    High physical activity44.6%44.6%
    Cigarette smokers26.9%18.8%
    High alcohol consumption24.6%16.7%
    High carbohydrate intake score52.9%52.9%
    High fat intake score56.1%56.4%
Table 2. . Spearman correlation coefficients of the features of the metabolic syndrome and body size measures for men (upper triangle) and women (lower triangle), respectively
 BMIWaist circumferencePercent body fatHDL-CHbA1cSystolic BPDiastolic BP
  • BP, blood pressure; HbA1c, glycosylated hemoglobin; HDL-C, high-density lipoprotein-cholesterol.

  • *

    P < 0.001.

BMI-0.86*0.74*−0.25*0.17*0.25*0.25*
Waist circumference0.90*-0.76*−0.21*0.22*0.28*0.24*
% Body fat0.95*0.90*-−0.14*0.26*0.37*0.21*
HDL-C−0.35*−0.35*−0.32*-0.000.01−0.01
HbA1c0.19*0.23*0.23*−0.07*-0.13*0.05*
Systolic BP0.39*0.40*0.43*−0.12*0.13*-0.63*
Diastolic BP0.31*0.31*0.30*−0.08*0.05*0.72*-

Association With Features Of The Metabolic Syndrome

Table 3 summarizes associations of the MC4R V103I polymorphism with the four quantitative features of the metabolic syndrome: waist circumference, HDL-C, HbA1c, and BP. All except BP were significantly or borderline significantly associated with MC4R V103I; thedirection of all estimates consistently indicated a reduction in risk for the metabolic syndrome for MC4R 103I carriers. The statistically significant difference between heterozygous subjects (G/A genotype) and subjects homozygous of the major allele (G/G genotype), was maintained under the simultaneous testing procedure accounting for multiple comparisons for the four features (P = 0.044). In addition, fasting triglyceride levels were decreased among G/A subjects, consistent with a reduction in risk for the metabolic syndrome; however, since the measurements were only available for 1,614 subjects the association was not statistically significant. Computation of the association for HbA1c and triglycerides on the log-scale yielded similar estimates and P values while that for the non-parametric Wilcoxon test yielded P = 0.030 for the association with HbA1c and P = 0.382 for triglycerides.

Table 3. . Association of MC4R V103I G/A vs. G/G genotype with features of the metabolic syndrome
PhenotypeAll (n = 7,888)Men (n = 3,912)Women (n = 3,976)Survey S3 (n = 4,048)Survey S4 (n = 3,840)
  • All analyses are adjusted for age, sex, educational level, and survey

  • a

    Linear regression.

  • b

    Logistic regression.

  • c

    ≥3 Components out of (i) waist >102 cm for men or >88 cm for women, (ii) high-density lipoprotein (HDL)-cholesterol <40 mg/dl for men or <50 mg/dl for women, (iii) blood pressure >130/85 mm Hg, (iv) HbA1c >6.0% or type 2 diabetes mellitus.

Quantitative outcomeDifference between mean values for subjects with G/A vs. G/G (P value)a 
    Waist (cm)−1.46 (P = 0.020)−1.18 (P = 0.135)−1.80 (P = 0.070)−1.33 (P = 0.112)−1.65 (P = 0.080)
    HDL-cholesterol (mg/dl)+1.76 (P = 0.056)+0.82 (P = 0.464)+2.81 (P = 0.059)+1.65 (P = 0.188)+1.87 (P = 0.166)
    HbA1c (%)−0.09 (P = 0.040)−0.06 (P = 0.273)−0.11 (P = 0.065)−0.07 (P = 0.297)−0.10 (P = 0.048)
    Systolic/diastolic blood pressure (mm HG)−0.46 (P = 0.647)/+0.76 (P = 0.567)/−1.78 (P = 0.236)/−1.12 (P = 0.419)/+0.40 (P = 0.784)/
    Fasting triglycerides (mg/dl) (n = 1614)−0.35 (P = 0.583)0.0 (P = 0.996)−0.71 (P = 0.437)−0.83 (P = 0.365)+0.19 (P = 0.826)
 −16.1 (P = 0.226)−27.3 (P = 0.240)−1.25 (P = 0.909)−26.4 (P = 0.562)−11.9 (P = 0.336)
Dichotomous outcome Odds ratio comparing subjects with G/A vs. G/G (P value)b 
    “Metabolic syndrome”c0.46 (P = 0.003)0.50 (P = 0.049)0.42 (P = 0.030)0.46 (P = 0.046)0.45 (P = 0.047)

After adjustment for BMI, associations between MC4R 103I and waist circumference, HDL-C, and HbA1c were no longer statistically significant: these were −0.30 cm (P = 0.303), +1.16 mg/dL (P = 0.182) and −0.074% (P = 0.071), respectively, for the subjects with the G/A genotype.

A protective association of the G/A genotype with other summaries of the “metabolic syndrome” was found as a (i) statistically significant decrease in the odds for the “metabolic syndrome” (OR = 0.46, 95% confidence interval (CI) = (0.27, 0.76), P = 0.003, Table 3), (ii) borderline significant decrease in the odds for including a component of the “metabolic syndrome” (OR = 0.82, 95% CI = (0.66, 1.02), P = 0.071), and (iii) significantly decreasing trend in the proportion of G/A vs. G/G subjects by number of components (P = 0.024, Figure 2a). Note that the OR for high waist circumference (>102 cm for men, >88 cm for women) alone was 0.72, CI = (0.54, 0.96) (P = 0.027).

image

Figure 2. : G/A genotype frequencies (a) by number of components of the “metabolic syndrome” (out of (i) waist >102 cm for men or >88 cm for women, (ii) HDL-cholesterol <40 mg/dl for men or <50 mg/dl for women, (iii) blood pressure >130/85 mm Hg, (iv) HbA1c >6.0% or type 2 diabetes mellitus), and (b) by number of components of the “metabolic syndrome” disregarding the blood pressure criterion.

Download figure to PowerPoint

As there was no association of the genotype with BP in the quantitative analysis, we additionally analyzed components of the “metabolic syndrome” disregarding the BP criterion (waist circumference, HDL-C, and HbA1c or T2DM) yielding a significant decrease in odds of having (iv) ≥2 components (OR = 0.59, CI = (0.40–0.89), P = 0.011), and (v) an additional component (OR (95% CI) = 0.70 (0.55–0.89), P = 0.004) for G/A carriers and (vi) a statistically significant decreasing trend of the proportion of G/A subjects by increasing numbers of components (P = 0.003, Figure 2b).

All associations were consistent in both surveys constituting the full sample and in both genders, but were slightly more pronounced in women; see Table 3. Modifying the definition of the “metabolic syndrome” according to the extension of the ATPIII definition by adding the criterion of antihypertensive medication did not change the results. The adjustment for age, sex, educational status, or survey did not have a marked impact on the findings (data not shown). With regard to exploring potential population stratification, no remarkable effect, either for the quantitative analysis or for the analysis of the metabolic syndrome, was found from adding height or an indicator for being born in Germany as a covariate into the model.

Association With Further Measures Of Body Size

We found a decrease in hip circumference (−1.12 cm, P = 0.030) and percentage body fat (–0.54%, P = 0.083) for subjects with the G/A compared to the G/G genotype, which disappeared after adjustment for waist circumference or BMI, and no statistically significant difference in height for G/A compared to G/G genotype carriers (P = 0.652).

Mediator Analyses

The results from the mediator analyses are summarized in Table 4. No direct associations were found for the genotype with any of the potential mediators (high/low pulse rate, high/low physical activity, current smoking yes/no, high/low alcohol intake, high/low fat score (ORs between 0.95 and 1.05, all P values > 0.50)), except a borderline significantly higher carbohydrate score among subjects with the G/A genotype (OR = 1.26, P = 0.059), which remained after adjustment for BMI (OR = 1.24, P = 0.073). As there was a significant association between high carbohydrate score and decreased waist circumference (−1.17 cm, P < 0.0001), three criteria of Prentice et al. (34) for mediation of the genotype–waist circumference association by high carbohydrate score hold ((a), (b), and borderline significant (c) in Figure 1). However, the fourth criterion ((d) in Figure 1) did not hold, since adjustment for carbohydrate score only very slightly reduced the association between genotype and waist circumference from −1.46 cm (P = 0.020) to −1.39 cm (P = 0.026).

Table 4. . The role of life style parameters or pulse as a mediator in the MC4R V103I (comparing G/A vs. G/G) association with waist circumference, BMI, or the “metabolic syndrome”: column 2 provides the odds ratio (OR) estimates for the association of the GENOTYPE with the MEDIATOR (model (c) in Figure 1). The first row provides the original association of the GENOTYPE with the OUTCOME (model (a) in Figure 1). The following rows provide this association with additional adjustment for the respective mediator (model (d) in Figure 1)
  Association of OUTCOME with GENOTYPE adjusted for MEDIATORc OUTCOME
MEDIATOR (dichotomous)Direct association of MEDIATOR with GENOTYPEa OR comparing G/A vs. G/G (P value)Waist circumference (cm) difference in means comparing G/A vs. G/G (P value)BMI (kg/m2) difference in means comparing G/A vs. G/G (P values)“Metabolic syndrome” OR comparing G/A vs. G/G (P value)
  • a

    Via logistic regression, logit−1(MEDIATOR) = α + β1 GENOTYPE, and further adjusted for age, sex, educational level, and survey, stating OR = exp(β1).

  • b

    Via linear regression for waist circumference or BMI (g = Id) and via logistic regression for “metabolic syndrome” (g = logit)as OUTCOME, g−1(OUTCOME) = α + β1 GENOTYPE and further adjustment for age, sex, educational level, and survey stating β1 (for waist and BMI) or OR = exp(β1) (for “metabolic syndrome”). cVia linear regression for waist circumference or BMI (g = Id) and via logistic regression for “metabolic syndrome” (g = logit) as OUTCOME, g−1(OUTCOME) = α + β1 GENOTYPE + β2 MEDIATOR, and further adjustment for age, sex, educational level, and survey, stating β1 (for waist and BMI) and OR = exp(β1) (for “metabolic syndrome”).

Original analysisb-−1.46 (P = 0.02)−0.53 (P = 0.04)0.46 (P = 0.003)
Pulse (low/high)0.95 (P = 0.66)−1.44 (P = 0.02)−0.53 (P = 0.04)0.46 (P = 0.003)
Physical activity (low/high)1.01 (P = 0.94)−1.45 (P = 0.02)−0.54 (P = 0.04)0.45 (P = 0.003)
Smoking (no/yes)0.99 (P = 0.94)−1.46 (P = 0.02)−0.54 (P = 0.04)0.46 (P = 0.003)
Alcohol (low/high)1.05 (P = 0.75)−1.46 (P = 0.02)−0.54 (P = 0.04)0.46 (P = 0.003)
Carbo intake (low/high)1.26 (P = 0.06)−1.39 (P = 0.03)−0.52 (P = 0.05)0.47 (P = 0.004)
Fat intake (low/high)0.97 (P = 0.83)−1.46 (P = 0.02)−0.54 (P = 0.04)0.46 (P = 0.003)

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References

While epidemiological data from one large study (1) and from a meta-analysis of several smaller studies (2) has put forth evidence for an association of the MC4R 103I variant with decreased BMI, a functional study has recently indicated a role played by this polymorphism on MC4R function, which might imply a direct role in appetite regulation in humans. Furthermore, a linkage signal was just reported for physical activity and dietary intake (10). Both decreased physical activity and increased dietary intake enhance the development of the complex known as the “metabolic syndrome”. The data in this manuscript support this idea by showing that this variant is associated with lower HbA1c and higher HDL-C levels as well as a substantial 50% decreased risk of having the “metabolic syndrome” for subjects with the G/A genotype compared to those with G/G, although these findings are not completely independent of the BMI finding. Furthermore, we find an association with high carbohydrate intake frequency, which is borderline significant, but independent of BMI. There was no direct association or a mediating effect due to pulse rate, physical activity, smoking, alcohol consumption or frequency of fat intake.

Association With Features Of The “metabolic Syndrome”

It was intriguing to find that MC4R V103I was associated with all features of the “metabolic syndrome” except for BP. After controlling for BMI, association estimates for HDL-C and HbA1c were diminished and statistical significance was lost. The loss of significance, however, should be seen in the light of limited power—even in such a large sample—as the MC4R 103I is a rather rare variant (1.85% allele frequency). Furthermore, it is unlikely that the HDL-C and HbA1c association is induced purely by their correlation with BMI, as the correlation is about the same or even smaller than the correlation of BMI with BP, and BP shows no association. We therefore conclude that our data suggests an effect of the MC4R V103I on HDL-C and HbA1c independent from the known BMI association.

One way of jointly analyzing the four parameters involved in the metabolic syndrome: waist circumference, HDL-C, HbA1c, and BP, is by doing a multi-variate analysis by testing the difference in the four-dimensional outcome distribution (Hotelling–Lawley statistic), which showed significant association when avoiding the multiple testing. Another way of a joint analysis is by applying the definition of the metabolic syndrome. However, the current definitions of the metabolic syndrome could not be applied due to lack of fasting blood samples for the majority of the subjects. We have chosen a surrogate “metabolic syndrome,” which is closely related to the ATPIII definition (19), but differs in two aspects: it omits the criterion for fasting triglycerides, and it substitutes the criterion of fasting glucose >110 mg/dL for the criteria of HbA1c and T2DM. We chose an HbA1c value >6.0% as cutoff, as this was recently shown to be a good predictor for T2DM (35). The T2DM criterion has already been included in the revision of ATPIII (21). Due to the absence of fasting triglycerides, our definition requires the fulfillment of three out of four rather than three out of five conditions. Thus some subjects might not qualify for our definition of the “metabolic syndrome” but would qualify for the ATPIII definition, but most subjects with our “metabolic syndrome” would also qualify according to ATPIII or other definitions. Permitting this limitation, we found a strong and highly significant association of the MC4R V103I polymorphism with the “metabolic syndrome,” which was robust when utilizing several aspects of the “metabolic syndrome”; it did not matter whether the binary variable or the categories by number of components were analyzed. Furthermore, the association was consistent and significant in all analyzed subgroups (by survey and by gender, see Table 3). Our data thus provide strong evidence for the MC4R 103I being associated with a 50% reduction in the risk of having the “metabolic syndrome.”

This is a strong effect for a complex polygenic phenotype such as the metabolic syndrome. The population-based design of our study also provides the opportunity for describing the impact of this variant in the population: the metabolic syndrome prevalence was 5.52 % (p 1) among carriers and 11.13% (p 0) among non-carriers implying a difference of −5.61% (p 1p 0) and a relative risk of 0.50 (p 1/p 0). As the polymorphism is rather rare with 3.68% (f), the population attributable risk (PAR = (p 1p 0) × f) is small indicating a reduction of the prevalence by 0.21% attributable to this variant. However, for complex phenotypes, a substantially higher PAR due to one genetic variant is seldom the case. Furthermore, due to the high prevalence of the metabolic syndrome (of 10.93% (p)), the importance of this effect might also be emphasized by the fraction of the “metabolic syndrome” attributable to the variant, the population attributable fraction, of −1.92% (PAR/p), which also indicates that the prevalence of the “metabolic syndrome” would be 1.92% higher, if it was not for this variant.

Even with the current scepticism about the clear definition of the metabolic syndrome and about its usefulness for clinical diagnosis (36), it provides one way of summarizing the correlated phenotypic outcomes for a practical joint analysis. And despite the fact that this “metabolic syndrome” analysis is not completely independent of BMI, it still provides important insight into the great extent of the impact of MC4R V103I on this set of inter-related cardiovascular risk factors (summarized by the metabolic syndrome definition).

Association With Other Obesity Phenotypes

The association with waist circumference was slightly stronger than the previously reported association with BMI (1) in terms of estimate precision (P = 0.021 for waist; P = 0.047 for BMI). However, due to the high correlation between these two obesity measures, it may not be feasible to differentiate between a rather general obesity or a central obesity effect. The MC4R V103I association was also apparent with hip circumference, but less pronounced with percentage body fat; the latter one might be due to a higher measurement error when assessing the percentage of body fat, which resulted in reduced estimates and reduced precision. As there was no association with height, neither with nor without adjustment for BMI, (and with a statistical power of 73% required in order to find an association of 1 cm), our data suggests that the MC4R 103I association with decreased BMI is not due to increased height.

Mediator Analyses In The Light With Functional Evidence

Because of the difficulties in measuring food intake in epidemiological studies (that can result in a measurement error of up to 75% and its deflating effect on associations (37)), the borderline significant association of the MC4R V103I polymorphism with carbohydrate intake frequency observed in the mediator analyses is noteworthy. That a carbohydrate-rich diet is associated with a leaner phenotype has already been reported by other studies (38,39). The fact that there was no association of the polymorphism with the pulse rate or physical activity, both related to energy consumption, might also rather point towards an energy-intake related mechanism. However, the lack of an association with physical activity could also be due to the variable being assessed via interview and therefore possibly error prone. Based on our data, it could be speculated that MC4R V103I might be involved in appetite regulation resulting in more frequent intake of carbohydrate-rich food. This would be supported by a recent linkage signal reported for dietary intake in Hispanic children encompassing the region of the MC4R (10). Therefore, the “latent variable,” which is more directly associated with the polymorphism and affects obesity as well as lipid and glucose metabolism, could well be a nutrient-related phenotype.

An elegant physiological system with the clear role of MC4R in energy homeostasis has previously been revealed (40,41,42,43,44,45,46): in a starving state, leptin levels drop, reducing the activity of pro-opiomelanocortin neurons and increasing the activity of the agouti-related protein, both of them powerfully inhibiting MC4R signaling, and thus increasing the demand for food intake. In the converse situation, leptin levels increase after a period of feeding, leading to increased signaling through MC4R and to suppression of appetite. The relevance of the MC4R in appetite regulation and obesity was proposed by knockout mice (11), and studies in human families unveiled MC4R as the most common single-gene causing obesity (4,5,47) based on the gene's rare mutations. Functional data indicate that nonsense mutations in this gene result in impaired inhibition of MC4R signaling (7,8), thus disturbing the suppression of appetite after feeding, which can lead to severe obesity.

The most common MC4R polymorphism, the V103I shows a normal endogenous agonist ligand affinity profile and normal cell surface expression levels (48). It has thus been difficult to pinpoint a molecular mechanism explaining the weight-lowering effect until recently, when it has been shown that the 103I allele possesses a twofold decrease in antagonist potency of the human agouti-related protein, which might decrease the orexigenic effect of the antagonist (9). Additionally, the effect of β-MSH, a potent agonist at the MC4R (49) seemed to be increased for the 103I-allele (9). Hence, both, the decreased antagonist and the increased agonist potencies are compatible with an enhanced MC4R function, which could explain the weight-lowering effect of the variant. Based upon our data, it could be speculated that this polymorphism has a role in appetite regulation in humans by reducing the quantity or modifying the quality of food intake, which modulates central obesity, as well as lipid and glucose metabolism.

Strengths And Limitations Of The Study

It is among the strengths of our study (i) that we analyzed a large representative sample with 7,888 participants of a rather homogeneous population with high-quality phenotyping, including extensive anthropometric measures and a detailed questionnaire on nutrition, physical activity and other behavioral factors, (ii) that we elucidated the association of the polymorphism with features of the “metabolic syndrome” by various methods of analysis for many different aspects of the complex phenotype (showing consistent results for all methods and all subgroups), and (iii) that we performed a systematic mediator analysis that could be a model for further genetic association studies, (iv) that we found some indication of this MC4R polymorphism being associated with a nutrient-related variable for the first time in epidemiological data, (v) that hereby the MC4R V103I association could be shown to hold not only for BMI in several studies, but also for other obesity-related parameters, which further strengthens the evidence that this association is genuine. It is a limitation of our phenotyping that only a fifth of the subjects were in a fasting state and thus the usual definitions of the metabolic syndrome could not be applied. However, our data indicate a slight decrease in triglycerides for subjects with the 103I allele, which would be in line with the reported decreased triglycerides for G/A subjects (50).

In summary, our study is the first to show an association of the MC4R 103I with parameters of the metabolic syndrome yielding up to a 50% metabolic syndrome risk reduction, which illustrated a strong role of this polymorphism for cardiovascular risk factors. Furthermore, it is the first study giving some indication of the association between this polymorphism and a nutrient-related variable found in an epidemiological study in humans.

DISCLOSURE

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References

We certify that none of the authors have in the past five years accepted any reimbursement for attending a symposium, any fee for speaking or for organizing education, any funds for research or a member of staff, or any fee for consulting from an organization that may in any way gain or lose financially from the results of our study or the conclusions of our manuscript. None of the authors have in the past five years been employed by or hold any stocks or shares in such an organization. None of the authors has acted as an expert witness on the subject of our study. None of the authors has any other competing financial interests. The corresponding author has the right to grant on behalf of all authors and does grant on behalf of all authors, an exclusive licence on a worldwide basis to the BMJ Publishing Group Ltd and its Licensees to permit this article to be published in JMG and any other BMJPGL products to exploit all subsidiary rights, as set out in our licence.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References

We thank all study participants and M. Wimmer, A. Rosin, G. Fischer, A. Wulff and A. Schneider for technical support and data management. The KORA research platform and the MONICA Augsburg studies were initiated and financed by the GSF-National Research Center for Environment and Health, by the German Federal Ministry of Education and Research and the State of Bavaria. Genotyping was performed in the GSF Genome Analysis Center (GAC) chaired by J. Adamski. This genetic investigation was funded by the National Genome Research Net of the German Ministry of Education and Research (OI-GSF0482, UWS15T03, N2NV S31T10, 01GS0482, and 01GS0482), the “Sonderforschungsbereich-SFB-386,” the Munic Center of Health of the Ludwig-Maximilians-Universität München, Germany, the EU FP6 LSHMCT-2003-503041, and by the “Genomics of Lipid-associated Disorders—GOLD” of the Austrian Genome Research Program GEN-AU.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and procedures
  5. Results
  6. Discussion
  7. DISCLOSURE
  8. Acknowledgments
  9. References
  • 1
    Heid IM, Vollmert C., Hinney A. et al. Association of the 103I MC4R allele with decreased body mass in 7937 participants of two population-based surveys. J Med Genet 2005; 42: e21.
  • 2
    Geller F., Reichwald K., Dempfle A. et al. Melanocortin-4 receptor gene variant I103 is negatively associated with obesity. Am J Hum Genet 2004; 74: 572581.
  • 3
    Dempfle A., Hinney A., Heinzel-Gutenbrunner M. et al. Large quantitative effect of melanocortin-4 receptor gene mutations on body mass index. J Med Genet 2004; 41: 795800.
  • 4
    Vaisse C., Clement K., Durand E. et al. Melanocortin-4 receptor mutations are a frequent and heterogeneous cause of morbid obesity. J Clin Invest 2000; 106: 253262.
  • 5
    Farooqi IS, Keogh JM, Yeo GS et al. Clinical spectrum of obesity and mutations in the melanocortin 4 receptor gene. N Engl J Med 2003; 348: 10851095.
  • 6
    O'Rahilly S., Yeo GS, Farooqi IS. Melanocortin receptors weigh in. Nat Med 2004; 10: 351352.
  • 7
    Yeo GS, Lank EJ, Farooqi IS et al. Mutations in the human melanocortin-4 receptor gene associated with severe familial obesity disrupts receptor function through multiple molecular mechanisms. Hum Mol Genet 2003; 12: 561574.
  • 8
    Hinney A., Bettecken T., Tarnow P. et al. Prevalence, spectrum, and functional characterization of melanocortin-4 receptor gene mutations in a representative population-based sample and obese adults from Germany. J Clin Endocrinol Metab 2006; 91: 17611769.
  • 9
    Xiang Z., Litherland SA, Sorensen NB et al. Pharmacological characterization of 40 human melanocortin-4 receptor polymorphisms with the endogenous proopiomelanocortin-derived agonists and the agouti-related protein (AGRP) antagonist. Biochemistry 2006; 45: 72777288.
  • 10
    Cai G., Cole SA, Butte N. et al. A quantitative trait locus on chromosome 18q for physical activity and dietary intake in Hispanic children. Obesity (Silver Spring) 2006; 14: 15961604.
  • 11
    Huszar D., Lynch CA, Fairchild-Huntress V. et al. Targeted disruption of the melanocortin-4 receptor results in obesity in mice. Cell 1997; 88: 131141.
  • 12
    Haffner SM, Valdez RA, Hazuda HP et al. Prospective analysis of the insulin-resistance syndrome (syndrome X). Diabetes 1992; 41: 715722.
  • 13
    Resnick HE, Jones K., Ruotolo G. et al. Insulin resistance, the metabolic syndrome, and risk of incident cardiovascular disease in nondiabetic american indians: the Strong Heart Study. Diabetes Care 2003; 26: 861867.
  • 14
    Isomaa B., Almgren P., Tuomi T. et al. Cardiovascular morbidity and mortality associated with the metabolic syndrome. Diabetes Care 2001; 24: 683689.
  • 15
    Bonora E., Targher G., Formentini G. et al. The metabolic syndrome is an independent predictor of cardiovascular disease in type 2 diabetic subjects. Prospective data from the Verona Diabetes Complications Study. Diabet Med 2004; 21: 5258.
  • 16
    Chen J., Muntner P., Hamm LL et al. The metabolic syndrome and chronic kidney disease in U.S. adults. Ann Intern Med 2004; 140: 167174.
  • 17
    Lakka HM, Laaksonen DE, Lakka TA et al. The metabolic syndrome and total and cardiovascular disease mortality in middle-aged men. JAMA 2002; 288: 27092716.
  • 18
    Lorenzo C., Williams K., Hunt KJ, Haffner SM. Trend in the prevalence of the metabolic syndrome and its impact on cardiovascular disease incidence: the San Antonio Heart Study. Diabetes Care 2006; 29: 625630.
  • 19
    National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 2006; 106: 31433421.
  • 20
    Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications part 1: diagnosis and classification of diabetes mellitus—provisional report of a WHO consultation. Diabet Med 1998; 15: 539553.
  • 21
    Grundy SM. Metabolic syndrome scientific statement by the American Heart Association and the National Heart, Lung, and Blood Institute. Arterioscler Thromb Vasc Biol 2005; 25: 22432244.
  • 22
    Reynolds K., He J.. Epidemiology of the metabolic syndrome. Am J Med Sci 2005; 330: 273279.
  • 23
    Steffens M., Lamina C., Illig T. et al. SNP-based analysis of genetic substructure in the German population. Hum Hered 2006; 62: 2029.
  • 24
    Loewel H., Doring A., Schneider A. et al. The MONICA Augsburg surveys—basis for prospective cohort studies. Gesundheitswesen 2005; 1: S13S18.
  • 25
    Wellmann J., Heidrich J., Berger K. et al. Changes in alcohol intake and risk of coronary heart disease and all-cause mortality in the MONICA/KORA-Augsburg cohort 1987–97. Eur J Cardiovasc Prev Rehabil 2004; 11: 4855.
  • 26
    Meisinger C., Löwel H., Thorand B., Döring A.. Leisure time physical activity and the risk of type 2 diabetes in men and women from the general population. The MONICA/KORA Augsburg Cohort Study. Diabetologia 2005; 48: 2734.
  • 27
    Winkler G., Döring A.. Validation of a short qualitative food frequency list used in several German large scale surveys. Z Ernahrungswiss 1998; 37: 234241.
  • 28
    American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2007 30 (Suppl 1): S42S47.
  • 29
    Derr R., Garrett E., Stacy GA, Saudek CD. Is HbA(1c) affected by glycemic instability? Diabetes Care 2003; 26: 27282733.
  • 30
    Igarashi M., Hirata A., Yamauchi T. et al. Clinical utility and approach to estimate postprandial hypertriglycemia by a newly designed oral fat-loading test. J Atheroscler Thromb 2003; 10: 314320.
  • 31
    Sharman MJ, Gomez AL, Kraemer WJ, Volek JS. Very low-carbohydrate and low-fat diets affect fasting lipids and postprandial lipemia differently in overweight men. J Nutr 2004; 134: 880885.
  • 32
    Chung BH, Cho BH, Liang P. et al. Contribution of postprandial lipemia to the dietary fat-mediated changes in endogenous lipoprotein-cholesterol concentrations in humans. Am J Clin Nutr 2004; 80: 11451158.
  • 33
    Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986; 51: 11731182.
  • 34
    Prentice RL. Surrogate endpoints in clinical trials: definition and operational criteria. Stat Med 1989; 8: 431440.
  • 35
    Greci LS, Kailasam M., Malkani S. et al. Utility of HbA(1c) levels for diabetes case finding in hospitalized patients with hyperglycemia. Diabetes Care 2003; 26: 10641068.
  • 36
    Kahn R., Buse J., Ferrannini E., Stern M.. The metabolic syndrome: time for a critical appraisal. Joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetologia 2005; 48: 16841699.
  • 37
    Carroll RJ, Ruppert D., Stefanski LA. Measurement Error in Nonlinear Models. Chapman & Hall: London, 1995.
  • 38
    Siggaard R., Raben A., Astrup A.. Weight loss during 12 week's ad libitum carbohydrate-rich diet in overweight and normal-weight subjects at a Danish work site. Obes Res 1996; 4: 347356.
  • 39
    de la Puebla RA Fernandez, Fuentes F., Perez-Martinez P. et al. A reduction in dietary saturated fat decreases body fat content in overweight, hypercholesterolemic males. Nutr Metab Cardiovasc Dis 2003; 13: 273277.
  • 40
    Schwartz MW, Seeley RJ, Woods SC et al. Leptin increases hypothalamic pro-opiomelanocortin mRNA expression in the rostral arcuate nucleus. Diabetes 1997; 46: 21192123.
  • 41
    Mizuno TM, Mobbs CV. Hypothalamic agouti-related protein messenger ribonucleic acid is inhibited by leptin and stimulated by fasting. Endocrinology 1999; 140: 814817.
  • 42
    Cowley MA, Smart JL, Rubinstein M. et al. Leptin activates anorexigenic POMC neurons through a neural network in the arcuate nucleus. Nature 2001; 411: 480484.
  • 43
    Seeley RJ, Yagaloff KA, Fisher SL et al. Melanocortin receptors in leptin effects. Nature 1997; 390: 349.
  • 44
    Thornton JE, Cheung CC, Clifton DK, Steiner RA. Regulation of hypothalamic proopiomelanocortin mRNA by leptin in ob/ob mice. Endocrinology 1997; 138: 50635066.
  • 45
    Ollmann MM, Wilson BD, Yang YK et al. Antagonism of central melanocortin receptors in vitro and in vivo by agouti-related protein. Science 1997; 278: 135138.
  • 46
    Graham M., Shutter JR, Sarmiento U., Sarosi I., Stark KL. Overexpression of Agrt leads to obesity in transgenic mice. Nat Genet 1997; 17: 273274.
  • 47
    Yeo GS, Farooqi IS, Challis BG, Jackson RS, O'Rahilly S.. The role of melanocortin signalling in the control of body weight: evidence from human and murine genetic models. QJM 2000; 93: 714.
  • 48
    Hinney A., Hohmann S., Geller F. et al. Melanocortin-4 receptor gene: case-control study and transmission disequilibrium test confirm that functionally relevant mutations are compatible with a major gene effect for extreme obesity. J Clin Endocrinol Metab 2003; 88: 42584267.
  • 49
    Biebermann H., Castaneda TR, Van LF et al. A role for beta-melanocyte-stimulating hormone in human body-weight regulation. Cell Metab 2006; 3: 141146.
  • 50
    Bronner G., Sattler AM, Hinney A. et al. The 103I variant of the melanocortin 4 receptor is associated with low serum triglyceride levels. J Clin Endocrinol Metab 2006; 91: 535538.