PAPER Allelic Variants in the Gabaaa6 Receptor Subunit Gene (GABRA6) Is Associated with Abdominal Obesity and Cortisol Secretion

Allelic variants in the GABAAa6 receptor subunit gene (GABRA6) is associated with abdominal obesity and cortisol secretion

R Rosmond1,2*, C Bouchard2 and P Bjo¨rntorp3

1Department of Clinical Chemistry, Sahlgrenska University Hospital, Go¨teborg, Sweden; 2Pennington Biomedical Research Center, Baton Rouge, Lousiana, USA; and 3Department of Heart and Lung Diseases, Sahlgrenska University Hospital, Go¨teborg, Sweden

INTRODUCTION: Cortisol is involved in the regulation of adipose-tissue differentiation, function and distribution, and in excess causes abdominal obesity. At the level of the brain, cortisol secretion is partly controlled by g-aminobutyric acid (GABA), which is the major inhibitory neurotransmitter in the vertebrate brain, and acts by binding to GABAA receptors. METHOD: We examined the potential impact of a 1519T > C polymorphism in the GABAAa6 receptor subunit (GABRA6) gene on obesity and obesity-related phenotypes as well as circulating hormones, including salivary cortisol in 284 unrelated Swedish men born in 1944. The subjects were genotyped by using PCR amplification of the 3’ non-coding region of the GABRA6 gene followed by digestion with the restriction enzyme AlwNI. RESULTS: The frequency of allele T was 0.54 and 0.46 for allele C. Carriers for the T allele (n ¼ 211) had borderline significantly higher waist-to-hip ratio (P ¼ 0.094) and abdominal sagittal diameter (P ¼ 0.084) compared to homozygotes for the C allele (n ¼ 56). The homozygotes for the T allele had, in comparison to heterozygotes, significantly (P ¼ 0.004 – 0.024) higher mean cortisol levels at 11:45 am, and 30, 45 and 60 min after a standardized lunch and, finally, at 5:00 pm. In addition, T=T subjects had significantly (P ¼ 0.031) higher diurnal cortisol secretion compared to T=C subjects. Other hormones, glucose and serum lipids were not different across the genotype groups. CONCLUSION: These findings suggest a role of the 1519T > C polymorphism in GABRA6 in the predisposition to hypercorti- solism and perhaps abdominal obesity. The pathophysiology may involve various environmental factors, particularly stress, that destabilize the GABA – hypothalamic – pituitary – adrenal systems in those with genetic vulnerability. International Journal of Obesity (2002) 26, 938 – 941. doi:10.1038/sj.ijo.0802022

Keywords: cortisol; GABA; gene; obesity; polymorphism

Introduction becomes poorly regulated with disruption of central regula- The hypothalamic – pituitary – adrenal (HPA) axis plays a tory systems and increased cortisol secretion, the develop- central role in homeostatic processes.1 The presence of a ment of visceral obesity, insulin resistance, dyslipidemia and continuously changing and sometimes threatening external hypertension may occur.2,3 environment may, when the challenge exceeds a threshold, The hypothalamus has extensive and complex neural activate the HPA axis, with a resulting increase in plasma connections, and receives afferent regulatory signals from cortisol. Endocrine feedback regulation balances the central different parts of the brain. Stimulatory inputs arise from the afferent signals, which allow the HPA axis to keep cortisol suprachiasmatic nucleus, the amygdala and the raphe levels within an optimal range.1 However, if the HPA axis nuclei,4,5 while inhibitory inputs originate from the hippo- campus and the locus coeruleus.6,7 The inhibitory effects are exerted mainly by g-aminobutyric acid (GABA).8 GABA is the major inhibitory neurotransmitter in the vertebrate brain, *Correspondence: R Rosmond, Department of Clinical Chemistry, and acts by binding to the GABA receptors, which are Sahlgrenska University Hospital, S-413 45 Go¨teborg, Sweden. A Received 13 August 2001; revised 6 February 2002; hetero-oligomeric chloride channels that are modulated accepted 11 February 2002 among others by cortisol and benzodiazepines.9 To date, at GABRA6, abdominal obesity and cortisol R Rosmond et al 939 least 16 human GABAA receptor cDNA have been cloned, Genotyping 9 including the GABAAa6 receptor subunit. Genotyping of the 1519T > C SNP was carried out by PCR- Clinical evidence suggests that alprazolam, a benzodiaze- RFLP. Genomic leukocyte DNA (150 ng in a final volume of pine that activates GABAergic receptors, inhibits the activity 10 ml) was amplified by PCR using the following primers: 5’- of the HPA axis.10,11 In obese subjects, pretreatment with GGA GGC ACC AGT AAA ATA GAC CAG-3’ and 5’-AAT ACT alprazolam blunts the HPA axis response to pharmacological GAA CAA TGG AAG ACA AAA-3’.13 The primers amplified a challenge tests.12 Recently, a novel single nucleotide poly- product of 423 bp. The PCR conditions were: an initial morphism (SNP) in the 3’ non-coding region of the GABAAa6 denaturation step at 94 C for 3 min, followed by 35 cycles receptor subunit (GABRA6) gene has been described.13 DNA of denaturation at 94C for 30 s, annealing at 53C for 30 s, sequencing has shown that it is a T to C substitution result- and extension at 72C for 30 s, with a final extension of ing in the loss of an AlwNI restriction site at nucleotide 10 min at 72C. The PCR reaction product was digested at 1519.13 The GABRA6 gene is located on chromosome 5q31.1- 37C overnight with 5 U of the restriction enzyme AlwNI q35.14 The AlwNI SNP at the GABRA6 gene was found to be (New England Biolabs, Beverly, MA, USA), generation two associated with alcohol dependence and Korsakoff’s psycho- fragments of 257 and 166 bp when the T allele is present. The sis.13 Ethanol and benzodiazepines share many pharmacolo- fragments were separated on a 2% agarose gel. gical properties and the chronic consumption of alcohol, similar to that of benzodiazepines, is linked to GABAergic neurotransmission.15 Statistical analysis With this background, we addressed the hypothesis that All statistical analyses were performed using SPSS for Win- the 1519T > C variant of the GABRA6 gene may influence dows, release 10.0 (SPSS Inc., Chicago, IL, USA). P-values are obesity and obesity-related phenotypes as well as circulating two-sided throughout, and a P < 0.05 was considered signifi- hormones, including salivary cortisol. cant. The results are presented as mean and standard devia- tion (s.d.). Data comparisons were carried out with the General Linear Model, with genotype as independent factors Subjects and methods and BMI and WHR as covariates. All P-values were adjusted Subjects for multiple tests by using the Spjotvoll – Stoline post hoc In the present study, we recruited the subjects from an correction.20 ongoing cohort study of men born in 1944.16,17 The study was initiated in 1992. Based on self-measured waist-to-hip ratio (WHR), the following three subgroups, each of 150 Results men, were selected for further studies: the lowest ( 0.885) The frequency of allele T was 0.54 and that of the C allele and the highest values ( 1.01) as well as men around the 0.46. The genotype frequencies were 28.8, 50.2 and 21.0% arithmetic mean (0.94 – 0.96). We examined these men in for T=T, T=C and C=C, respectively. The observed genotype 1995 at the age of 51 y, and 284 (63%) volunteered to frequencies were in a Hardy – Weinberg equilibrium. participate.18,19 Nine subjects refused to be involved in the Table 1 presents the differences in obesity-related pheno- genetic studies, and in eight subjects the amount of blood types among the GABRA6 genotypes. Since these variables sample was not sufficient to conduct genetic analysis. All are highly influenced by obesity and body fat distribution, men gave written informed consent before participating in the results were adjusted for BMI and WHR whenever appro- the study, which was approved by the Go¨teborg University priate. Carriers for the T allele had borderline significantly Ethics Committee. higher WHR (P ¼ 0.094) and abdominal sagittal diameter (P ¼ 0.084) compared to homozygotes for the C allele. Hor- mones, glucose and serum lipids were not different across the genotype groups (all adjusted for BMI and WHR). Phenotypic measurements Table 2 shows the differences in salivary cortisol measure- Body mass index (BMI, kg=m2), WHR, and abdominal sagittal ments by genotype. The homozygotes for the T allele had, in diameter were measured as described before.18,19 Salivary comparison to heterozygotes, significantly (P ¼ 0.004 – cortisol was measured repeatedly over a random working 0.024) higher mean cortisol levels at 11:45 am, and 30, 45 day as described previously.18 Endocrine measurements, and 60 min after a standardized lunch and, finally, at besides cortisol, included plasma levels of testosterone, insu- 5:00 pm. In addition, T=T subjects had significantly lin-like growth factor I, and leptin as previously (P ¼ 0.031) higher diurnal cortisol secretion compared to described.18,19 Insulin, glucose, triglycerides, and total, T=C subjects. high- and low-density lipoprotein cholesterol were measured in an overnight fasting state as described.18,19 The serum lipids were determined by an enzymatic procedure with Discussion the Cobas Fara II analyzer (Roche Molecular Biochemicals, In developed societies, metabolic and circulatory disorders Mannheim, Germany). such as abdominal obesity, type 2 diabetes, dyslipidemia and

International Journal of Obesity GABRA6, abdominal obesity and cortisol R Rosmond et al 940 Table 1 GABAAa6 receptor subunit genotypes and variables of the study

Genotypes

T=TT=CC=C (n ¼ 77) (n ¼ 134) (n ¼ 56) Pa

Body mass index (kg=m2) 26.0 (3.6) 26.3 (4.0) 26.1 (4.2) > 0.20 Waist-to-hip ratio 0.94 (0.08)a 0.94 (0.06) 0.92 (0.08)b 0.094 Abdominal sagittal diameter (cm) 22.5 (3.7) 23.0 (3.5)a 22.2 (4.0)b 0.084 Testosterone (nmol=l) 19.6 (5.4) 19.8 (5.7) 19.6 (5.1) > 0.20 Insulin-like growth factor I (mg=l) 195.8 (66.2) 208.8 (61.3) 209.8 (71.6) > 0.20 Leptin (mg=l) 6.1 (4.4) 6.3 (4.1) 6.0 (4.6) > 0.20 Fasting insulin (mU=l) 11.7 (8.6) 13.4 (12.9) 12.2 (8.4) > 0.20 Fasting glucose (mmol=l) 4.4 (0.7) 4.6 (1.0) 4.7 (1.2) 0.197 Triglycerides (mmol=l) 1.8 (1.2) 1.8 (1.1) 1.7 (0.9) > 0.20 Total cholesterol (mmol=l) 6.2 (1.1) 6.2 (1.0) 6.1 (1.0) > 0.20 HDL cholesterol (mmol=l) 1.2 (0.3) 1.2 (0.3) 1.3 (0.4) > 0.20 LDL cholesterol (mmol=l) 4.2 (1.1) 4.1 (1.0) 4.0 (1.1) > 0.20

Values (meansÆ s.d.) with different letters are signiﬁcant at the 0.05 level. aAdjusted for BMI and WHR. HDL, High-density lipoprotein; LDL, low-density lipoprotein.

Table 2 Differences in salivary cortisol measurements by GABAAa6 receptor subunit genotypes

Genotypes

T=TT=CC=C (n ¼ 77) (n ¼ 134) (n ¼ 56) Pa

Cortisol level (nmol=l) in the morning 15.1 (8.4) 14.2 (6.1) 15.9 (8.4) > 0.20 Cortisol level (nmol=l) at 11:45 am 9.0 (8.8)* 6.3 (2.3){ 6.7 (2.8){ 0.004 Cortisol level (nmol=l) at 30 min after lunch 10.5 (13.4)* 7.2 (3.4){ 7.8 (3.8) 0.023 Cortisol level (nmol=l) at 45 min after lunch 9.0 (8.7)* 6.7 (2.7){ 7.3 (3.4) 0.023 Cortisol level (nmol=l) at 60 min after lunch 8.5 (8.9)* 6.0 (2.0){ 6.7 (3.0) 0.010 Cortisol level (nmol=l) at 5:00 pm 5.3 (3.1)* 4.3 (1.9){ 5.2 (2.4) 0.024 Cortisol level (nmol=l) before bedtime 4.1 (6.6) 3.0 (2.7) 3.2 (3.3) > 0.20 Diurnal cortisol secretion (nmol=l) 8.4 (5.9)* 6.8 (1.8){ 7.6 (2.5) 0.031

Values (meansÆ s.d.) with different symbols are signiﬁcant at the 0.05 level. aAdjusted for BMI and WHR.

hypertension rarely occur in isolation, but are typically part In the present study, we found a new polymorphism of a complex phenotype of metabolic abnormalities usually potentially associated with abdominal obesity. Carriers for referred to as the Metabolic Syndrome.2 The Metabolic the T allele tended to have both elevated WHR and abdom- Syndrome is usually deﬁned as a set of risk factors that inal sagittal diameter (Table 1). In animal models, the obese cluster in some people.21,22 The risk factors include elevated Zucker rat possesses altered brain GABAergic mechanisms insulin levels (insulin resistance), abdominal obesity, high that contribute to their overeating.25,26 In subjects with levels of both LDL cholesterol and triglycerides, and hyper- Prader – Willi syndrome, the mean level of plasma GABA tension. In addition to insulin resistance, abdominal obesity was found to be two to three times higher than in retarded in adults is a key contributor to the development of the nonobese controls.27 Although a role of GABA in human Metabolic Syndrome. There are currently several important obesity has not, to our knowledge, been previously theories that attempt to explain the causes of abdominal described, these studies suggest that GABAergic tonic inhibi- obesity. Studies of twins and families suggest that about 40 – tion may be potentially involved. 80% of the variation in human obesity can be ascribed to In addition to GABA, complex interactions among various genetic factors.23 Most variation in human obesity-related neurotransmitters and neuromodulators involved in the phenotypes is due to a limited number of common genetic regulation of energy balance, such as leptin, neuropeptide variations that interact with the environment to produce the Y and several hormones, including cortisol, may directly or ﬁnal phenotype.23 The number of genes, markers and chro- indirectly participate in the GABAergic regulation. Interest- mosomal regions that have been associated or linked with ingly, carriers for the T allele had also elevated diurnal human obesity-related phenotypes is now well above 200 cortisol secretion (Table 2). These results indicate that and continues to increase.24 the 1519T > C SNP in GABRA6 might be involved in the

International Journal of Obesity GABRA6, abdominal obesity and cortisol R Rosmond et al 941 dysfunction of the hypothalamic cortisol homeostasis- 9 Chebib M, Johnston GA. The ‘ABC’ of GABA receptors: a brief review. Clin Exp Pharmac Physiol 1999; 26: 937 – 940. controlling mechanisms within the central nervous system. 10 Curtis GC, Abelson JL, Gold PW. Adrenocorticotropic hormone It is well documented that cortisol contributes to the and cortisol responses to corticotropin-releasing hormone: regulation of adipose-tissue differentiation, function and dis- changes in panic disorder and effects of alprazolam treatment. tribution and, in excess, causes abdominal obesity.3 Inade- Biol Psychiat 1997; 41:76– 85. 11 Arvat E, Maccagno B, Ramunni J, Di Vito L, Giordano R, Gianotti quate cortisol secretion is associated with abnormalities in L et al. The inhibitory effect of alprazolam, a benzodiazepine, glucose, insulin and lipid metabolism, bringing the impor- overrides the stimulatory effect of metyrapone-induced lack of tance of the HPA axis in the development of abdominal negative cortisol feedback on corticotroph secretion in humans. obesity to the forefront.3 J Clin Endocrinol Metab 1999; 84: 2611 – 2615. 12 Grottoli S, Arvat E, Gauna C, Maccagno B, Ramunni J, Giordano R > In summary, we found that a 1519T C polymorphism of et al. Alprazolam, a benzodiazepine, blunts but does not abolish the GABRA6 gene is associated with elevated salivary cortisol the ACTH and cortisol response to hexarelin, a GHRP, in obese levels and potentially abdominal obesity. As this polymorph- patients. Int J Obes Relat Metab Disord 2000; 24: S136 – 137. 13 Loh EW, Smith I, Murray R, McLaughlin M, McNulty S, Ball D. ism is not located in a coding region of the GABAAa6 Association between variants at the GABAAb2, GABAAa6and receptor subunit gene, its functional role, if any, is uncertain. GABAAg2 gene cluster and alcohol dependence in a Scottish Further studies are required to replicate these prelim- population. Mol Psychiat 1999; 4: 539 – 544. inary findings and to define the biological effects of this 14 Hicks AA, Bailey ME, Riley BP, Kamphuis W, Siciliano MJ, John- polymorphism. son KJ et al. Further evidence for clustering of human GABAA receptor subunit genes: localization of the alpha 6-subunit gene (GABRA6) to distal chromosome 5q by linkage analysis. Genomics 1994; 20: 285 – 288. 15 Brailowsky S, Garcia O. Ethanol, GABA and epilepsy. Arch Med Res Acknowledgements 1999; 30:3– 9. This study was supported by grants from the Swedish Med- 16 Rosmond R, Lapidus L, Ma˚rin P, Bjo¨rntorp P. Mental distress, ical Research Council (K97-19X-00251-35A) and the obesity and body fat distribution in middle-aged men. Obes Res Pennington Biomedical Research Center. R Rosmond 1996; 4: 245 – 252. 17 Rosmond R, Lapidus L, Bjo¨rntorp P. The influence of occupa- would also like to acknowledge the Henning and Johan tional and social factors on obesity and body fat distribution in Throne-Holst Foundation for the support of a postdoctoral middle-aged men. Int J Obes Relat Metab Disord 1996; 20: 599 – fellowship at the Pennington Biomedical Research Center. C 607. Bouchard is partially supported by the George A Bray Chair 18 Rosmond R, Dallman MF, Bjo¨rntorp P. Stress-related cortisol secretion in men: relationships with abdominal obesity and in Nutrition. endocrine, metabolic and hemodynamic abnormalities. J Clin Endocrinol Metab 1998; 83: 1853 – 1859. 19 Rosmond R, Chagnon YC, Holm G, Chagnon M, Perusse L, References Lindell K et al. Hypertension in obesity and the leptin receptor 1 Chrousos GP. Stressors, stress, and neuroendocrine integration of gene locus. J Clin Endocrinol Metab 2000; 85: 3126 – 3131. the adaptive response. The 1997 Hans Selye Memorial Lecture. 20 Spjotvoll E, Stoline M. An extension of the T-method of multiple Ann NY Acad Sci 1998; 851: 311 – 335. comparison to include the cases with unequal sample sizes. JAm 2Bjo¨rntorp P, Rosmond R. Hypothalamic origin of the metabolic Stat Assoc 1973; 68: 976 – 978. syndrome X. Ann NY Acad Sci 1999; 892: 297 – 307. 21 Bjo¨rntorp P. The associations between obesity, adipose tissue 3Bjo¨rntorp P, Rosmond R. Obesity and cortisol. Nutrition 2000; 16: distribution and disease. Acta Med Scand 1988; 723(Suppl): 924 – 936. 121 – 134. 4 Dornhorst A, Carlson DE, Seif SM, Robinson AG, Zimmerman EA, 22 Reaven GM. Banting lecture 1988. Role of insulin resistance in Gann DS. Control of release of adrenocorticotropin and vasopres- human disease. Diabetes 1988; 37: 1595 – 1607. sin by the supraoptic and paraventricular nuclei. Endocrinology 23 Bouchard C. Genetics of obesity: an update on molecular mar- 1981; 108: 1420 – 1424. kers. Int J Obes Relat Metab Disord 1995; 19:S10– 13. 5 Grossman A, Costa A. The regulation of hypothalamic CRH: 24 Perusse L, Chagnon YC, Weisnagel SJ, Rankinen T, Synder E, impact of in vitro studies on the central control of the stress Sands J et al. The human obesity gene map: the 2000 update. response. Funct Neurol 1993; 8: 325 – 334. Obes Res 2001; 9: 135 – 169. 6 Ward DG, Bolton MG, Gann DS. Inhibitory and facilitatory areas 25 Coscina DV, Castonguay TW, Stern JS. Effects of increasing brain of the ventral midbrain mediating release of corticotropin in the GABA on the meal patterns of genetically obese vs. lean Zucker cat. Endocrinology 1978; 102: 1147 – 1154. rats. Int J Obes Relat Metab Disord 1992; 16:425– 433. 7 Carlson DE, Dornhorst A, Gann DS. Organization of the lateral 26 Orosco M, Jacquot C, Wepierre J, Cohen Y. GABA levels and hypothalamus for control of adrenocorticotropin release in the synthesis index in different brain areas of two obese rat models. cat. Endocrinology 1980; 107: 961 – 969. Gen Pharmac 1981; 12: 369 – 371. 8 Calogero AE, Gallucci WT, Chrousos GP, Gold PW. Interaction 27 Ebert MH, Schmidt DE, Thompson T, Butler MG. Elevated plasma between GABAergic neurotransmission and rat hypothalamic gamma-aminobutyric acid (GABA) levels in individuals with corticotropin-releasing hormone secretion in vitro. Brain Res either Prader – Willi syndrome or Angelman syndrome. J Neurop- 1988; 463:28– 36. sychiat Clin Neurosci 1997; 9:75– 80.

International Journal of Obesity