Short Allele of Serotonin Transporter Gene Promoter Is a Risk Factor for Obesity in Adolescents

Authors


Instituto de Investigaciones Médicas A. Lanari, Cardiología Molecular, Combatientes de Malvinas 3150, Buenos Aires 1427, Argentina. E-mail: pirola.carlos@lanari.fmed.uba.ar

Abstract

Obesity and hypertension are increasing medical problems in adolescents. Serotonin transporter (5-HTT) is involved in mood and eating disturbances. Encoded by the gene SLC6A4, the promoter shows functional insertion/deletion alleles: long (L) and short (S). Because individuals who are carriers for the short version are known to be at risk for higher levels of anxiety, we hypothesized that this variant may be associated with overweight. Data and blood samples were collected from 172 adolescents out of a cross-sectional, population-based study of 934 high school students. To replicate the findings, we also included 119 outpatients from the Nutrition and Diabetes Section of the Children's County Hospital. We found that the S allele was associated with overweight (BMI > 85th percentile), being a risk factor for overweight independently of sex, age, and hypertension [odds ratio (OR): 1.85; 95% confidence interval (CI): 1.13, 3.05; p < 0.02]. Additionally, in the outpatient study, compared with the homozygous LL subjects, S allele carriers showed a higher BMI z-score (1.47 ± 1.09 vs. 0.51 ± 1.4; p < 0.002) and were more frequent in overweight children. In conclusion, the S allele of the SLC6A4 promoter variant is associated with overweight being an independent genetic risk factor for obesity.

Introduction

Serotonin (5-hydroxytryptamine) has been implicated in various central physiological functions including sleep, appetite, memory, sexual behavior, neuroendocrine function, and mood (1). A possible physiological role of endogenous serotonin in controlling natural patterns of eating and nutrient selection has also been reported (2). Furthermore, several lines of evidence implicate a role for the serotonergic system in body weight regulation and eating disorders (3). One important component of the serotonergic system, the serotonin transporter (5-HTT)1 recycles serotonin after its release, thereby determining the magnitude and duration of serotonergic responses (4). The human 5-HTT is encoded by a single gene (SLC6A4) on chromosome 17q11.1–17q12 (5), and a polymorphic region was identified: a 44-bp insertion (long: L)/deletion (short: S) in the promoter region [5-HTT gene-linked polymorphic region (5-HTTLPR)] (6). Transfection studies showed that the L and S variants of the promoter polymorphism differentially modulate transcription of SLC6A4, the S variant being less efficient (6), suggesting that 5-HTTLPR is associated with an altered functional response of the serotonin system.

Because individuals who are carriers for the S variant of the 5-HTT gene are known to be at risk for higher levels of anxiety (7), and rates of anxiety and depression are three to four times higher among obese individuals than among their leaner peers (8), we hypothesized that S variant may be associated with overweight. Thus, the aim of this study is to explore the effect of the S variant of the serotonin transporter gene promoter on obesity in Argentinean adolescents. Hence, we looked for the association of the S/L variant of the SLC6A4 with clinical and laboratory characteristics of adolescents harboring features of metabolic syndrome from a population-based study performed in a rural town. To replicate the findings, we also studied a group of outpatients from a Children's County Hospital located in a different geographic area (Buenos Aires, metropolitan area).

In the first population, genotypes for the SLC6A4 promoter were in Hardy-Weinberg equilibrium and were similar to that reported in other populations (7) (LL: 29%, LS: 45%, SS: 26%); the observed allele frequencies were 52% for the L allele and 48% for the S allele.

Because human SLC6A4 transcription is differentially modulated by the allelic variants of the SLC6A4 promoter, showing that LL homozygous have higher rate of 5-HTT mRNA transcription and higher 5-HT uptake than individuals carrying at least one copy of the S allele (7), we grouped LS and homozygous SS individuals in a dominant model as carriers of the S allele for further analysis.

As shown in Table 1, in univariate analysis, we found that there was no difference between S allele carriers and noncarriers in most of the clinical characteristics of the metabolic syndrome. In contrast, using a two-way ANOVA, we found a significant (p < 0.03) higher age- and sex-adjusted BMI (BMI z-score) in the S allele carriers independently of the effect of hypertension. To make the difference clearer, absolute BMI values were compared between genotypes in both sexes using age as a covariate (Figure 1).

Table 1.  Clinical features, anthropometric variables, and laboratory findings of the subjects in the population-based study according to SLC6A4 promoter genotypes: homozygous LL was compared with S allele carriers (LS + SS)
FeaturesLLLS + SS
  • Results are expressed as mean ± SD. z-scores stand for sex- and age-adjusted values.

  • *

    p< 0.03 vs. LL.

  • After adjustment for age and sex, OR: 2.1; 95% CI: 1.05, 4.53; p < 0.04.

  • SABP, systolic arterial blood pressure; DABP, diastolic arterial blood pressure; HOMA, homeostasis model assessment of insulin resistance.

Number of subjects44128
Age (years)16 ± 216 ± 2
Sex (F/M)27/1777/51
SABP z-score0.34 ± 1.550.67 ± 1.69
DABP z-score−0.13 ± 0.81−0.07 ± 0.89
Homocysteine (μM)8.6 ± 4.97.5 ± 3.6
Total cholesterol (mg/dL)154.7 ± 24.5159.8 ± 28.1
Triacylglycerol (mg/dL)86.4 ± 38.991.8 ± 52
HDL-cholesterol (mg/dL)47.7 ± 7.849.3 ± 10.5
LDL-cholesterol (mg/dL)89.8 ± 20.492.5 ± 24
Uric acid (mg/dL)4.1 ± 1.14.1 ± 1
Glucose (mg/dL)88.6 ± 788 ± 6.9
Insulin (uU/mL)10.4 ± 3.812.4 ± 7.1
HOMA2.28 ± 0.862.73 ± 1.65
BMI z-score0.17 ± 0.910.49 ± 1.14*
Waist circunference (cm)73.0 ± 7.474.7 ± 10.2
Smoking (smokers/ nonsmokers)4/4017/111
Figure 1.

BMI of adolescents from the population-based study (top) and the outpatient-based study (bottom) according to sex and genotype of the long (L) and short (S) allele of the SLC6A4 promoter (homozygotes LL, white bars, vs. heterozygous LS plus homozygous SS, black bars). In both populations, BMIs of homozygous LL were significantly (p < 0.02) lower than those of the other group (analysis of covariance with age as a covariate). Results are expressed as mean ± SD.

In addition, the S allele was associated with overweight (BMI > 85th percentile; χ2, p < 0.02) because the S allele was more frequent in the overweight group (LL: 14.0%, n = 7; LS + SS: 86.0%, n = 43) in comparison with lean adolescents (LL: 30.3%, n = 37; LS + SS: 69.7%, n = 85). Logistic regression analysis indicated that the S allele is a risk factor for overweight independently of sex, age, and hypertension [odds ratio (OR): 1.85; 95% confidence interval (CI): 1.13, 3.05; p < 0.02]. Data concerning BMI z-score in the three genotypes are as follows: LL, 0.17 ± 0.91, n = 44; LS, 0.47 ± 1.18, n = 76; SS, 0.51 ± 1.08, n = 52. This indicates that LS and SS groups are phenotypically similar.

Table 2 shows differences in anthropometric variables and laboratory findings in the outpatient sample of children according to SLC6A4 promoter genotypes. Compared with the homozygous LL allele carriers, the S allele carriers showed a greater BMI z-score and higher other estimates of obesity and fat distribution, such as waist circumference, subscapular, and triceps skinfold thickness (Figure 2). Differences in absolute BMI values are depicted in Figure 1. Again, the S allele was associated with overweight (BMI > 85th percentile; χ2, p < 0.01) because the S allele was more frequent in the overweight group (LL: 8.2%, n = 7; LS + SS: 91.2%, n = 75) in comparison with the lean group (LL: 26.3%, n = 10; LS + SS: 73.7%, n = 27). Logistic regression analysis indicated that the S allele is a risk factor for overweight independently of sex and age and arterial blood pressure (OR: 3.98; 95% CI: 1.31, 12.18; p < 0.02). Moreover, a similar result was found with obesity as defined by BMI >95th percentile (OR: 3.95; 95% CI: 1.18, 13.38; p < 0.03). Data concerning the three genotypes and BMI z-score were as follows: LL, 0.50 ± 1.39, n = 18; LS, 1.55 ± 1.02, n = 70; SS, 1.24 ± 1.21, n = 31.

Table 2.  Clinical features, anthropometric variables, and laboratory findings of the subjects in the outpatient hospital-based study according to SLC6A4 promoter genotypes: homozygous LL was compared with S allele carriers (LS + SS)
FeaturesLLLS + SS
  1. Results are expressed as mean ± SD. z-scores stand for sex- and age-adjusted values.

  2. SABP, systolic arterial blood pressure; DABP, diastolic arterial blood pressure; HOMA, homeostasis model assessment of insulin resistance.

Number of subjects18101
Age (years)12 ± 212 ± 2
Sex (F/M)5/1347/54
SABP z-score−0.32 ± 0.84−0.14 ± 1.01
DABP z-score0.5 ± 0.630.4 ± 0.75
Homocysteine (μM)5.9 ± 1.66.7 ± 2.7
Total cholesterol (mg/dL)171.3 ± 33.3168.8 ± 32.7
HDL-cholesterol (mg/dL)45 ± 10.646 ± 11.2
Triacylglycerol (mg/dL)86.9 ± 30.1101.2 ± 56.7
LDL-cholesterol (mg/dL)105.8 ± 29.2110.7 ± 29.6
Uric acid (mg/dL)4.6 ± 1.14.6 ± 1.2
Glucosa (mg/dL)87.5 ± 986.9 ± 7.1
Insulin (μU/mL)13.3 ± 7.815.1 ± 9
HOMA2.89 ± 1.723.22 ± 1.89
Figure 2.

Sex- and age-adjusted BMI (z-score), waist circumference, subscapular skinfold thickness (SSFT), and triceps skinfold thickness (TSFT) of the adolescents in the outpatient hospital-based study according to SLC6A4 promoter genotypes: homozygous LL was compared with S allele carriers (LS + SS). Results are expressed as mean ± SD (* p < 0.003, † p < 0.01, ‡ p < 0.004, and # p < 0.001).

By pooling the two samples by means of Mantel-Haenszel's fixed and random models, we observed that both groups were homogeneous, obtaining a very similar but more robust result regarding BMI differences between genotypes (OR, 3.36; 95% CI, 1.68 to 6.72; p < 0.001).

Consequently, the main finding of our study is a significant association between the short variant of the SLC6A4 promoter and overweight in a population-based sample of adolescents. Accordingly, S allele carriers are almost twice as likely to be overweight than homozygous LL, independently of sex, age, and hypertension. This result was replicated, even with a greater OR, in an independent sample of children outpatients of a hospital-based study from a different geographic location.

Although association does not necessarily mean a causal relationship, in this case, several lines of evidence support the association between the S/L SLC6A4 promoter variant and overweight: the biological plausibility of this relationship, the proven effect of the L/S SLC6A4 variant on serotonin transporter activity (7), and our findings were replicated in two different samples of adolescents from two distant geographic regions of our country. In this regard, we wish to note that we selected the 5-HTTLPR variant because it was previously reported that serotonin uptake is genetically controlled, and the polymorphism mentioned contributes to the interindividual differences in 5-HTT expression and regulation. Moreover, it was shown that the polymorphism influences a constellation of traits related to anxiety and depression, features that were the background we used to formulate the hypothesis about the putative relationship between 5-HTT and obesity (6,7). Furthermore, it was recently published that, using high-resolution structural images and automated processes to test for brain volume and gray matter density, there are not only functional differences in the 5HTT genotypes but also structural ones (9). Finally, most of the studies that look for genetic relationships between the gene variants at the SLC6A4 and quantitative traits show significant associations only with the 5-HTTLPR but not the 5-HTTVNTR or the other single nucleotide polymorphisms. These results are shown, for instance, in the study of Curran et al. (10) that searches for differences at 5-HTTLPR, 5-HTTVNTR in the second intron, a single nucleotide polymorphism in the 3′-untranslated region, and 10 single nucleotide polymorphisms spread across the gene in a large epidemiological study of attention-deficit/hyperactivity disorder.

Although novel, these results are consistent with several independent observations. On one hand, the above-mentioned cumulative body of experimental evidence on the role of serotonin as a key regulator of energy homeostasis includes an important interaction with leptin-responsive central neuropeptide systems (11).

On the other hand, Rosmond et al. (12) found that a polymorphism in the 5-HT2a receptor gene promoter is strongly related with increased BMI and abdominal fat distribution, suggesting a relative deficiency of serotonergic effects in the carriers of the mutant allele.

More recently, Muldoon et al. (13) revealed an association between reduced central serotonergic response and the metabolic syndrome because the prolactin response evoked by a serotonin-releasing agent was blunted in overweight, insulin-resistant, and dyslipidemic individuals.

Conversely, Hinney et al. (14) did not find any association between the 5-HTT variant and body weight, although the study was focused on the allele distribution in relation to anorexia nervosa patients and extremely obese individuals in whom other factors and candidate genes seem to be involved.

Otherwise, obesity is a biologically heterogeneous disorder. The increasing prevalence of obesity in adolescents justifies a widespread effort and attention of physicians looking for depressive symptoms, anxiety, and loss of self-esteem in their patients.

In the light of the association between obesity and the SLC6A4 short variant we found, along with previous reports suggesting a strong association between the variant and anxiety and major depression, one may speculate that individuals who carry the S allele may be at risk of being overweight/obese as a result of a genetic predisposition, leading them to eat more.

For instance, Pine et al. (15) showed that childhood depression was associated with an increased BMI in adulthood, a finding that was recently replicated by a large community-based cohort study (16). Also, Goodman and Whitaker (17) reported that depressed adolescents are at increased risk for the development of obesity during adolescence, reinforcing the previous-mentioned evidence of common neurobiological mechanisms between obesity and depression.

Last, considering that dysfunction of serotonergic neurotransmission may result in a clinical picture of carbohydrate-craving obesity (18), it is reasonable to speculate that this may be an alternative explanation to our results as a potential mechanism affecting food intake. In this regard, obesity can be seen as a food addiction. In fact, it has been proposed to treat obese patients with similar therapeutic schemes than those used in patients who are addicted to drugs (19). It is tempting to speculate that 5-HTT plays a similar role in obesity as in drug and alcohol consumers or smokers (20,21).

To summarize, in Argentinean children and adolescents, the S allele of the SLC6A4 promoter is associated with overweight being an independent genetic risk factor for overweight/obesity.

In other words, by using the BMI cut-off values reported by Cole et al. (22), it can be predicted that, after reaching 18 years of age, 1.6-fold more S allele carriers than LL homozygous persons would be overweight or obese in our adolescent population.

This is, to our knowledge, the first report about the association between an SLC6A4 promoter variant and obesity in adolescents harboring features of the metabolic syndrome. We hope our study can serve as a primer, because further research is needed to confirm and extend these findings, improving the power of the study by increasing the number of subjects and revealing the intimate mechanism by which the putative decrease in the 5-HTT activity related to the SLC6A4 S variant may lead to a central serotonergic system dysfunction.

Research Methods and Procedures

The study design encompasses two populations groups: a cross-sectional, population-based study and an outpatient hospital-based study.

For the first population, data and blood samples were collected from 122 lean adolescents and 50 overweight ones that satisfied the criteria described below out of a cross-sectional, population-based study of 934 high school students of self-reported European ancestry in whom we have previously addressed the prevalence of obesity and hypertension (23). In our previous study, we showed a high correlation between systolic arterial blood pressure z-score and BMI z-score (Spearman correlation coefficient, 0.56; p = 0.000000), showing that, in our adolescents, hypertension is associated with an increased degree of obesity among other characteristics of the metabolic syndrome.

The second study was carried out in a total of 119 consecutive children and adolescents of self-reported European ancestry that were randomly recruited as outpatients from the Nutrition and Diabetes Section of the Children's County Hospital JM Gutierrez at Buenos Aires during a month period, excluding those who were diabetic patients.

All of the studies performed in this study were conducted in accordance with the guidelines of The Declaration of Helsinki. Written consent from children's parents was granted, in accordance with the procedures approved by the Ethical Committee of our institution.

Plasma glucose, insulin, uric acid, total cholesterol, high-density lipoprotein- and low-density lipoprotein-cholesterol, and triglycerides were measured by standard clinical laboratory techniques. We used the homeostasis model assessment index (fasting insulin in microunits per milliliter multiplied by fasting glucose in millimoles per liter divided by 22.5) as an estimation of insulin resistance. A commercial enzyme-linked immunosorbent assay kit was used to measure plasma concentrations of homocysteine by immunochemoluminiscence according to the manufacturer's instructions (Immulite 1000; DPC, Los Angeles, CA) in blood samples collected with sodium EDTA.

Genotyping for the 44-bp deletion-insertion polymorphism in the SLC6A4 promoter was performed by hot-start polymerase chain reaction using molecular biology grade reagents unless indicated and a Robocycler 96 thermal cycler (Stratagene, La Jolla, CA) from genomic DNA extracted from white blood cells.

The details of polymerase chain reaction protocol have been described previously (7) but primers to detect the variant were different—5′-CGTTGCCGCTCTGAATGC-3′ and 5′-TGGTAGGGTGCAAGGAGAATG-3′—amplifying a 384-bp (L allele) and a 340-bp (S allele) fragment.

Quantitative data were expressed as mean ± SD. Because for most of the variables we observed a significant difference of the SD between groups, differences between groups were assessed by the Mann-Whitney test except for comparison of BMI z-score means, for which we used ANOVA, and for comparison of BMI means between genotypes and sexes using age as covariate, in which case we used analysis of covariance. Genotypes frequencies were analyzed by means of the χ2 test. Logistic and multiple regressions were used for testing of the multivariate association between variables. We used the CSS/Statistica program package StatSoft V 6.0 (Tulsa, OK) to perform these analyses.

We performed a combined analysis of ORs using both a Mantel-Haenzsel's fixed effect and random effect procedures as described (24).

Acknowledgments

This study was supported partially by Grants B119 (Universidad de Buenos Aires), PICT 05–25920 (Agencia Nacional de Promoción Científica y Tecnológica), and PIP 5195 (Consejo Nacional de Investigaciones Científicas y Técnicas). S.S., S.G., and C.J.P. belong to Consejo Nacional de Investigaciones Científicas y Técnicas, and S.S. is a recipient of a Health Ministry Fellowship (Beca Ramón Carrillo-Arturo Oñativia Ministerio de Salud y Ambiente de la Nación) Convocatoria 2006.

Footnotes

  • 1

    Nonstandard abbreviations: 5-HTT, serotonin transporter; L, long; S, short; 5-HTTLPR, 5-HTT gene-linked polymorphic region; OR, odds ratio; CI, confidence interval.

  • 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.

Ancillary