Phenylethanolamine N-Methyltransferase G-148A Genetic Variant and Weight Loss in Obese Women

Phenylethanolamine N-Methyltransferase G-148A Genetic Variant and Weight Loss in Obese Women Warren R. Peters,* James P. MacMurry, Jennifer Walker,* Russell J. Giese, Jr., and David E. Comings Abstract PETERS, WARREN R., JAMES P. MACMURRY, JENNIFER WALKER, RUSSELL J. GIESE, JR., AND DAVID E. COMINGS. Phenylethanolamine N-methyltransferase G-148A genetic variant and weight loss in obese women. Obes Res. 2003;11:415 419. Objective: To understand the impact of the phenylethanolamine N-methyltransferase (PNMT) G-148A gene and nutritional variables on weight loss in obese women. Research Methods and Procedures: One hundred fortynine women, ages 45 to 65 with a body mass index of 30 kg/m 2, participated in a 6-month, open-label intervention that included sibutramine (15 mg/d) and a monthly health-education class. Anthropometric measurements, vital signs, food frequency, exercise log, medication compliance, and psychological and sociological questionnaires were completed each month. Genetic polymorphisms of PNMT were determined. Results: Univariate analysis of G/G, G/A, and A/A genotypes against tertiles of percentage of weight loss were significant at 3 but not at 6 months (Pearson 2 : p 0.006; homozygous/heterozygosity: p 0.002, p 0.253, and p 0.122, respectively). A regression model that included the PNMT genetic variation and certain nutrition and exercise variables demonstrated that only the PNMT gene ( 0.360, SE 0.585, and p 0.003) was statistically significant at 6 months, and the total calories ( 0.925, SE 0.004, and p 0.009), fiber intake ( 0.621, SE 0.124, and p 0.000), and PNMT ( 0.262, SE 1.415, and p 0.024) were significant. Discussion: The homozygosity/heterozygosity of the Received for review April 16, 2002. Accepted for publication in final form December 9, 2002. *Loma Linda University, Center for Health Promotion, Loma Linda, California; City of Hope National Medical Center, Department of Medical Genetics, Los Angeles, California; and Department of Family Practice, University of Wisconsin, LaCrosse, Wisconsin. Address correspondence to Warren R. Peters, Loma Linda University, Center for Health Promotion, 24785 Stewart Street, Loma Linda, California 92350. E-mail: wpeters@univ.llu.edu Copyright 2003 NAASO PNMT gene was highly predictive of significant weight loss with sibutramine during the first 3 months, which highlights the need for specific pharmacotherapy. The early weightloss success of those subjects who were homozygous for PNMT may have motivated and selected those that would make further dietary changes, which then augmented their final weight loss. Key words: thermogenesis, epinephrine, lifestyle, gene transcription Introduction Obesity develops from a combination of low energy expenditure and energy intake beyond homeostatic metabolic requirements. This metabolic dysregulation may result from a dysfunction of the sympathetic nervous system. A recent intervention trial demonstrated a marked variation in response to epinephrine infusion between obese and nonobese subjects (1). There was a wide variation in the effect of epinephrine on leptin in those subjects who were obese, demonstrating a complex response that could lead to excess body weight. Phenylethanolamine N-methyltransferase (PNMT) 1 mediates the conversion of norepinephrine to epinephrine and is a rate-limiting enzyme in the catecholamine biosynthesis pathway and, thus, may participate in body weight regulation. It has been found that the expression of PNMT in the spleens of rats occurs through the pituitaryadrenocortical mechanism (2). This enzymatic energy regulation mechanism may be a target for pharmacological intervention for obesity. The noradrenaline and serotonin reuptake inhibitor sibutramine has emerged as a promising agent in the treatment of obesity (3). In animal studies, sibutramine causes a negative fat balance and weight loss by enhancing satiety through the dual-mechanism of noradrenergic and seroto- 1 Nonstandard abbreviations: PNMT, phenylethanolamine N-methyltransferase; BMI, body mass index; PCR, polymerase chain reaction. OBESITY RESEARCH Vol. 11 No. 3 March 2003 415

nergic effects. Sibutramine increases thermogenesis and is accompanied by a significant increase in plasma epinephrine (4). Studies have confirmed that increases in urinary epinephrine and norepinephrine are associated with an increase of sympathetic and parasympathetic control (5). In view of the putative role of epinephrine in sibutramineinduced weight loss, the regulation of this enzyme becomes pertinent. The PNMT gene is known to map to chromosome 17q21-q22 (6). Studies in transgenic mice have shown that overexpression of PNMT produces an elevation of epinephrine levels and a marked suppression of circulating leptin levels that results in a significant increase in body fat (7). Thus, the established effect of sibutramine on epinephrine levels in obesity treatment, coupled with the mediation of leptin levels by the rate-limiting enzyme of epinephrine, show that PNMT is a candidate gene that may be implicated in sibutramine efficacy. Research Methods and Procedures Subjects One hundred forty-nine women [ages 45 to 65; body mass index (BMI) 30 kg/m 2 ] were recruited from a pool of 219 individuals who had previously been part of a casecontrol obesity study (NIH, National Institute of Diabetes and Digestive and Kidney Diseases Grant 1R03 DK53310-01). These subjects participated in a 6-month weight-loss trial that included the daily intake of sibutramine (15 mg) and a monthly 1-hour behavior modification seminar. We excluded those individuals with poorly controlled hypertension and those who were taking other weight-loss products and/or monoamine oxidase inhibitors or serotonin reuptake inhibitors, or who were allergic to sibutramine. Pregnant or nursing mothers and those with a previous diagnosis of myocardial infarction, arrhythmia, congestive heart failure, liver or kidney disease, and/or narrow angle glaucoma were also excluded. Eighty-two (55%) of the subjects were postmenopausal, 72 (48%) were on hormone replacement, and 29 (19%) had had hysterectomies. See Table 1 for other physical characteristics of the subjects. Behavioral seminars encouraged participants to shift their dietary intake to lower fat foods and to increase their consumption of vegetables and fruits. Each subject was shown exercises that would be appropriate to her physical abilities. Daily exercise and an active lifestyle were modeled and encouraged. Each participant completed questionnaires about her physical, psychological, and family health. At baseline, 3 months, and 6 months, they completed a food frequency questionnaire (Bringham and Women s Hospital 1988). Each month, they brought in their medication bottles for compliance assessment and were then issued the next month s supply of medications. Anthropometric measurements were taken on a monthly basis. Estimates of physical Table 1. Physical characteristics of subjects Mean SD Range Age 54 5.78 years 45 to 65 Weight 228 45.54 pounds 150 to 387 BMI 40.10 8.01 30 to 76 Waist 41.74 5.74 inches 23 to 65 Waist-to-hip ratio.83 1.11 1.06 to 0.48 Pulse 78.02 10.24 bpm 44.00 to 56.00 Systolic blood pressure 130.31 16.06 mm/hg 98 to 80 Diastolic blood pressure 80.34 9.06 mm/hg 60 to 100 activity were created from self-reported daily-activity records and were reported as calories. Genetic Methods A polymorphism in the promoter region of the PNMT gene was studied (8) consisting of a G-to-A transition at the 148 position with respect to the gene transcription start site. The region of the promoter corresponded to the polymorphism and was amplified by a polymerase chain reaction (PCR) kit (Qiagen, Santa Clarita, CA) and the following primer sets for 148: 5 -GTCTCCACCTCGAATCAG-3 (forward primer) and 5 -CCATCTCTCTTCTCCAGC-3 (reverse primer). The PCR reaction was performed using standard procedures in a 15- L volume containing 50 ng of genomic DNA, 200 nm of each primer, 200 M deoxynucleoside triphosphates, and 0.6 U of DNA Taq polymerase and buffer supplied by the manufacturer. Amplification for each of the markers was performed for 35 cycles with denaturation at 95 C for 40 seconds and an annealing temperature of 60 C for G-148A. The amplified products were digested by the MspI restriction enzyme, and the resulting restriction fragments were separated on a 10% polyacrylamide gel in 1 Trisborate EDTA buffer and visualized by staining with 5 gof ethidium bromide. Cleavage of the 101-bp G-148A PCR product with MspI identified the A allele (65 bp) and the G allele (49 16 bp). There are two G-148A constant bands, which are 19 17 bp in length. Results One hundred forty-nine women, 68% of the eligible subjects, started in this study. During the next 3 months, 24 (16%) women stopped taking medication. By the end of the study, 42 (28%) subjects were no longer using medication. Of the 42 who stopped taking medication during the 6-month study, 7 continued in the lifestyle intervention part 416 OBESITY RESEARCH Vol. 11 No. 3 March 2003

Table 2. Weight loss of subjects during study Mean SD Range Weight loss at 3 months 4.925 5.99 lb 9 to 24 lb Percent of loss at 3 months 2.1 2.48% 4.92% to 9.74% Weight loss at 6 months 14.71 13.82 lb 9 to 58.5 lb Percent of loss at 6 months 6.36 5.58% 4.46% to 23.73% Table 3. Tertile of percentage of weight loss at 3 months with PNMT genotype Tertile % weight loss 1.00 2.00 3.00 Total PNMT G/G Count 9 13 15 37 % 24.3% 35.1% 40.5% 100% G/A Count 15 16 3 34 % 44.1% 47.1% 8.8% 100% A/A Count 6 5 12 23 % 26.1% 21.7% 52.2% 100% Total Count 30 34 30 94 % 31.9% 36.2% 31.9% 100% 100% 100% 100% 100% Pearson 2 p 0.006 by genotype and p 0.002 by homozygosity/heterozygosity. 1, Lowest; 2, intermediate; 3, highest. of the program. Of the 42 subjects that stopped taking the medication, 22% stopped for medical reasons, and the remainder stopped for personal and nonmedical issues. All data points were available for analysis on 94 subjects at 3 months and 74 subjects at 6 months. We compared the genetic distribution of the genetic variants for the PNMT gene between those who stopped taking medication at the 3- and the 6-month interval and those who continued medication. There was no statistically significant difference between these two groups. The effect of hormone status or replacement was evaluated, and there was no statistically significant difference in the percentage of weight loss at 3 and 6 months for those subjects who were postmenopausal, had had a hysterectomy, or were on hormone replacement. (Total weight loss leaving the study is recorded in Table 2.) The subjects who were taking sibutramine at the end of 3 months were divided into tertiles of percentage of weight loss that allowed a clear comparison of the highest vs. the lowest response individuals. These three groups were evaluated for the presence of the three PNMT gene variant groups. The homozygous variants G/G and A/A were also compared against the heterozygous variant G/A (see Table 3). In a similar way, the subjects who were taking sibutramine at the end of 6 months were divided into tertiles by the percentage of their weight loss and then compared with the presence of the three PNMT genotypes. The 2 model portrayed statistical significance at 3 months, but was not statistically significant at 6 months (see Table 4). Compared with the heterozygous PNMT variant, G/A, the presence of the homozygous PNMT variant, either G/G or A/A, was associated with a statistically significant weight loss or percentage of weight loss during the first 3 months and also during the entire 6-month period of treatment with sibutramine. Generally, homozygosity does not stratify phenotypically. Instead, the G/A variant would have an intermediate clinical level and A/A would have the most abnormal effect. This type of heterosis has been described, and potential mechanisms have been posed (9). A regression model was developed that included PNMT homozygosity vs. heterozygosity and lifestyle variables including dietary fiber, animal fat intake, and total calories consumed. This was performed in an effort to understand the impact of lifestyle changes and their interaction with the PNMT gene during the first 3 months and during the entire Table 4. Tertile of percentage of weight loss at 6 months with PNMT genotype Tertile % weight loss 1.00 2.00 3.00 Total PNMT G/G Count 8 10 10 28 % 28.8% 35.7% 35.7% 100% G/A Count 11 10 5 26 % 42.3% 38.5% 19.2% 100% A/A Count 4 8 10 20 % 20.0% 30.0% 50.0% 100% Total Count 23 26 25 74 % 31.1% 35.1% 33.9% 100% 100% 100% 100% 100% Pearson 2 p 0.253 by genotype and p 0.122 by homozygosity/heterozygosity. 1, Lowest; 2, intermediate; 3, highest. OBESITY RESEARCH Vol. 11 No. 3 March 2003 417

Table 5. Regression model for percentage of weight loss at 3 months Variable SE p Value PNMT gene 0.360 0.585 0.003 Age 0.062 0.106 NS Calories 0.509 0.003 NS Total fat 0.293 0.054 NS Fiber 0.374 0.089 0.067 Animal fat 0.134 0.060 NS Exercise level 0.071 0.000 NS 6-month study. All subjects actively participated in the lifestyle-education program. It would be expected that variation in the compliance to diet and exercise restriction could explain or impact the effect of genotypes. The dietary intake of fiber was important at 3 months, but did not reach statistical significance. By 6 months, dietary-fiber intake was statistically important, as was the total caloric intake. Total dietary fat intake was close to statistical significance (p 0.058). By controlling for other important lifestyle variables in the regression model, the gene variant remained statistically significant at 3 and 6 months (see Tables 5 7).There was no statistically significant relationship between the PNMT genetic variants and variations in the waist-to-hip ratio at baseline, 3 months, and 6 months. Discussion To our knowledge, this is the first evaluation of the effect of the PNMT gene on weight loss and, more specifically, weight loss with the pharmaceutical preparation sibutramine. There seems to be a striking genotype-dependent variation in the effective weight loss of these subjects, who were all taking sibutramine. During the first 3 months, the only statistically important variant was genetic. By months 4 through 6, lifestyle variables such as fiber intake and total fat intake became statistically significant. Previous studies with the PNMT genetic polymorphism (10) have found that increased risk for Alzheimer s disease was associated with homozygosity at the 148 position (either GG or AA genotypes), supporting our present finding that it is heterozygosity vs. homozygosity at this genetic locus that may mediate the sibutramine response. Molecular heterosis has challenged the common understanding that in a two-allele polymorphism, the presence of the 1 allele would be associated with decreased gene expression, and the 11 genotype would show the greatest effect. A recent review of heterosis, negative or positive, was performed to better understand what seems to run counter to conventional wisdom. Three explanations were proposed. They were an inverted U-shaped response curve, an independent third factor causing a hidden stratification, and a greater fitness in 12 heterozygotes because they show a broader range of gene expression than 11 or 22 homozygotes (9). These data reinforce the view that variations in epinephrine metabolism may contribute significantly to the efficacy of sibutramine, and such variations may be linked to PNMT genotype. These findings also suggest the possibility that the efficacy of sibutramine may be related to the PNMTmediated variations in leptin concentrations (or resistance) alluded to in the findings of the study by Bottner et al. (7). Indeed, our findings suggest the need for a follow-up study that would examine the interaction between PNMT genotypes and leptin concentrations in determining the efficacy of sibutramine. There is evidence that some obese individuals are subject to autonomic nervous system dysfunction, which seems to be associated with chronic hyperinsulinemia, leading to a high-output, low-resistance hemodynamic state, postprandial sympathetic dominance, and a persistent baroreflex down-regulation (11). However, the obese person has a lower-than-normal thermogenic response to food intake (12). This may represent an autonomic nervous system dysregulation that responds to the drug sibutramine with its known central nervous system effects. Obese rats on sibutramine have increased sympathetic activity demonstrated through increased levels of urinary epinephrine levels and attenuated arcuate nucleus neuropeptide Y. Chronic use of sibutramine may reset the function of one or both of these pathways for weight maintenance after weight loss (13). This is a very preliminary study that may, however, point to other innovative pharmaceutical answers to the difficult problem of obesity in the clinical setting. Currently, our neurochemical, pharmaceutical, and thermogenic tools are Table 6. Regression model for percentage of weight loss at 6 months Variable SE p Value PNMT gene 0.262 1.415 0.024 Age 0.121 0.108 NS Calories 0.925 0.004 0.009 Total fat intake 0.622 0.078 0.058 Fiber intake 0.621 0.124 0.000 Animal fat intake 0.027 0.088 NS Exercise level 0.068 0.000 NS 418 OBESITY RESEARCH Vol. 11 No. 3 March 2003

Table 7. Univariate analysis for PNMT gene heterozygosity vs. homozygosity Variable PNMT G/A PNMT GG/AA p Value Age (years) 54.7 52.9 NS BMI 38.4 39.9 NS Calories 3 months (kcal) 1602 1687 NS Calories 6 months (kcal) 1365 1461 NS Total fat intake 3 months (Gm) 55 58 NS Total fat intake 6 months (Gm) 49 51 NS Animal fat intake 3 months (Gm) 31 32 NS Animal fat intake 6 months (Gm) 27 29 NS Fiber intake 3 months (Gm) 21 19 NS Fiber intake 6 months (Gm) 17 17 NS Exercise 3 months (cal) 5296 4856 NS Exercise 6 months (cal) 5083 5588 NS Weight loss 3 months (pounds) 3.0 8.2 0.002 Weight loss 6 months (pounds) 10.0 19.0 0.010 Percentage of weight loss 3 months (%) 1.4 3.4 0.003 Percentage of weight loss 6 months (%) 4.8 8.2 0.018 severely limited when compared with what is available for the treatment of hyperlipidemia, hypertension, or diabetes. This study must be expanded to evaluate leptin and insulin levels during weight loss, in addition to direct measurement of sympathetic nervous system hormones and function. Acknowledgments This work was supported by Knoll, MERIT 13 (Mount Olive, NJ). References 1. Couillard C, Mauriege P, Prud homme D, et al. Plasma leptin response to an epinephrine infusion in lean and obese women. Obes Res. 2002;10:6 13 2. Jelokova J, Rusnak M, Kubovckova L, et al. Stress increases gene expression of phenylethanolamine N-methyltransferase in spleen of rats via pituitary-adrenocortical mechanism. Psychoneroendocrinology. 2002;27:619 33. 3. Reasoner CA. Promising new approaches. Diab Obes Metab. 1999;1(Suppl 1):S41 8. 4. Hansen DL, Toubro S, Stock MJ, MacDonald IA, Astrup A. Thermogenic effects of sibutramine in humans. Am J Clin Nutr. 1998;68:1180 6. 5. Hirsch J, Mackintosh RM, Aronne LJ. The effects of drugs used to treat obesity on the autonomic nervous system. Obes Res. 2000;8:227 33. 6. Hoehe MR, Plaetke R, Otterud B. Genetic Linkage of the human gene for phenylethanolamine N-methyltransferase (PNMT), the adrenaline-synthesizing enzyme, to DNA markers on chromosome 17q21 22. Hum Mol Genet. 1992;1: 175 8. 7. Bottner A, Haidan A, Eisenhofer G. Increased body fat mass and suppression of circulating leptin levels in response to hypersecretion of epinephrine in phenylethanolamine N-methyltransferase (PNMT)-overexpressing mice. Endocrinology. 2000;141:4239 46. 8. Wu S, Comings DE. Two single nucleotide polymorphisms in the promotor region of the human phenylethanolamine N-methyltransferase PNMT gene. Psychiatr Genet. 999; 9:187 8. 9. Commings DE, MacMurray JP. Molecular heterosis: a review. Mol Genet Metab. 2000;71:19 31. 10. Mann MG, Wu S, Rostamkani M. Phenylethanolamine N- methethyltransferase PNMT gene and early-onset Alzheimer disease. Am J Med Genet. 2001;105:312 6. 11. Emdin M, Gastaldelli A, Muscelli E, et al. Hyperinsulinemia and autonomic nervous system dysfunction in obesity: effects of weight loss. Circulation. 2001;103:513 9. 12. Matsumoto T, Miyawake C, Ue H, Kanda T, Yoshitake Y, Moritani T. Comparison of thermogenic sympathetic response to food intake between obese and non-obese young women. Obes Res. 2001 9:78 85. 13. Levin BE, Dunn-Meynell AA. Sibutramine alters the central mechanisms regulating the defined body weight in diet-induced obese rats. Am J Physiol Regul Integr Comp Physiol. 2000;279:R2222 8. OBESITY RESEARCH Vol. 11 No. 3 March 2003 419