Weight-loss response to naltrexone/bupropion is modulated by the Taq1A genetic variant near DRD2 (rs1800497): A pilot study

Abstract

Naltrexone/bupropion (NB) is a US Food and Drug Administration-approved antiobesity medication. Clinical trials have shown variable weight loss, with responders and non-responders. NB is believed to act on central dopaminergic pathways to suppress appetite. The Taq1A polymorphism near DRD2 (rs1800497) is associated with the density of striatal dopamine D2 receptors, with individuals carrying the A allele (AA or AG; termed A1+) having 30%−40% fewer dopamine binding sites than those who do not carry the A allele (GG; termed A1−). We performed a pilot study to assess the association of the rs1800497 ANKK1 c.2137G > A (p.Glu713Lys) variant with weight loss with NB treatment in 33 subjects. Mean (SD) weight loss was 5.9% (3.2%) for the A1+ genotype group (n = 15) and 4.2% (4.2%) for the A1− genotype group (n = 18). The mean weight loss for the A1+ genotype group was significantly greater than the predefined clinically significant 4% weight-loss target (one-sample t-test, P = .035), whereas the mean weight loss for the A1− genotype group was not (P = .85). Individuals with the A1+ genotype appear to respond better to NB than A1- individuals.

1 |. INTRODUCTION

Obesity is an epidemic, with more than 40% of US adults classified as obese.¹ Pharmacotherapy for obesity can be modestly effective, but there is a wide range of inter-individual weight-loss response. One such US Food and Drug Administration-approved antiobesity medication is Contrave, a combination of naltrexone 32 mg and bupropion 360 mg. Clinical trials of naltrexone/bupropion 32 mg/360 mg (NB) have shown a mean weight loss of 6.1% after 56 weeks of treatment but, notably, only 48% of patients had a reduction in body weight of 5% or more,² suggesting that there are responders and non-responders to the medication. The variability in response suggests obesity is a heterogeneous disease, for which we need a personalized approach accounting for individual differences in aetiological factors such as genetics.³

The mechanism of action of NB may be through action on pro-opiomelanocortin (POMC) neurons in the arcuate nucleus of the hypothalamus stimulating anorexigenic pathways and on the mesolimbic dopaminergic reward system.^2,4 Bupropion is a dopamine and norepinephrine reuptake inhibitor that stimulates POMC signalling.⁵ Naltrexone blocks the mu-opioid receptor, which results in reduced inhibition of POMC neurons by beta-endorphin.⁵ Given the probable importance of dopamine signalling for the action of NB, it is possible that the variability in response may be related in part to the availability of dopamine receptors, which is modulated by genetic variation in the DRD2 locus.

The Taq1A genetic variant (rs1800497) is located in the ankyrin repeat and protein-kinase domain-containing protein 1 (ANKK1) gene, adjacent to the dopamine D2 receptor (DRD2) gene. This polymorphism has been associated with addictive behaviours including alcoholism and smoking, as well as overeating.^6,7 Individuals with at least one A allele (AA or AG, termed A1+) have diminished striatal DRD2 density, with 30%−40% fewer receptors than those who do not carry the A allele (GG, termed A1−)⁸ and have an increased probability of obesity.⁷ The susceptibility to obesity is believed to be caused by the effect of this polymorphism on the neurophysiology of food reward and altered behavioural and brain responses to food.⁹ There are some data to suggest an association with the A1+ genotype for dietary preferences^7,8 as well as for less weight loss with an intervention involving caloric restriction and resistance training.¹⁰ Functional MRI studies have shown that individuals with the A1+ genotype had weaker striatal activation after eating or imagining a palatable food, and had a greater risk of weight gain than individuals who were A1−.¹¹ A study of the use of bupropion for smoking cessation found that the medication was effective in A1− individuals but not A1+ individuals.¹² However, bupropion may have a differential pharmacogenetic effect on weight as another study showed that the Taq1A AA genotype was associated with less weight gain at the end of successful smoking cessation treatment with bupropion.¹³ A recent study explored genetic mechanisms of the variation in antiobesity and metabolic effects of metformin in children with obesity,¹⁴ but otherwise there has been scarce investigation of the pharmacogenetics of obesity.

We performed a pilot study of patients treated with NB to determine if the Taq1A genetic variant is associated with a weight loss of 4% or more of body weight after short-term NB therapy. The 4% weight-loss cutoff to define clinical response was based on data showing that a weight loss of 4% or more prior to 16 weeks of treatment was the best predictor of 5% or more weight loss at 56 weeks,¹⁵ which is generally considered to be clinically significant. We hypothesized that individuals with the Taq1A A1+ genotype would be more probable to respond to NB as the higher dopaminergic activity would activate a hypofunctioning reward circuitry and result in less overeating.¹⁶ To our knowledge, this is the first investigation of the pharmacogenetics of NB. Our novel data add to the understanding of the individual variation in weight-loss response to NB, as well as to the drug’s mechanism of action. This pilot study provides the groundwork for further investigation, with the objective of improving the treatment of obesity through precision medicine.

2 |. METHODS

A retrospective chart review of patients seen at the Columbia University Irving Weight Control Center and the Weill Cornell Comprehensive Weight Management Center was undertaken to include men and women aged older than 18 years with a baseline body mass index (BMI) of 27–45 kg/m² who were treated with NB. Subjects with changes in medications that might affect body weight during the treatment period were excluded. Subjects provided a blood or saliva sample for genotyping of the rs1800497 ANKK1 c.2137G > A (p.Glu713Lys) allele near the DRD2 gene. Weight loss at 12–16 weeks was determined by chart review. If weight data were not available within this time frame, weight was used from the closest available office visit. Weight-loss percentage was calculated by dividing weight loss by baseline weight. Mean weight loss in each of the two genotype classes (A1+: AA or AG vs. A1−: GG) was separately compared with a threshold of 4% weight loss using a two-tailed one-sample t-test. One subject, who was treated for just 7 weeks prior to discontinuation of medication because of a clear lack of efficacy with only 0.9% (0.8 kg) weight loss, was included. One subject enrolled in the study was excluded from the analysis because of identification as an outlier for percentage weight loss (20.3%) as determined by Grubb’s Test. The power analysis was based on prior data showing that the standard deviation of percentage weight loss is about 5 (in percent units), and that ~45% of overweight/obese adults have the A1+ genotype.⁶ From power analysis for a one-sample t-test, we determined that, with 13 A1+ subjects, we would have 80% power to find the mean weight loss to be significantly greater than 4% if the true weight loss exceeded 8%, testing two-sided at P = .05. Data are presented as mean (SD).

Genomic DNA was extracted as previously described.¹⁷ Genotyping was performed at the New York Nutrition Obesity Research Center Molecular Genetics Core by polymerase chain reaction amplification of a 270 bp fragment and Sanger sequencing through the Taq1A variant (rs1800497). The AA and AG genotypes correspond to A1+ and GG to A1−. This study was approved by the Columbia University Institutional Review Board, conformed to standards set forth by the Declaration of Helsinki, and all subjects provided written informed consent.

3 |. RESULTS

Data were analysed from 33 subjects, 29 women and four men. The mean (SD) age was 49.8 (13.5) years and baseline weight was 99.0 (19.4) kg. For the entire cohort, mean weight loss was 5.0% (3.8)% and the mean follow-up period was 14.7 (5.6) weeks. The A1+ genotype was present in 15 subjects and 18 subjects had the A1− genotype. The characteristics of subjects by genotype are presented in Table 1. There were no significant differences in baseline age, weight, BMI or duration of therapy. Although of note, the A1+ group trended towards a higher baseline BMI (P = .10). Mean weight-loss percentage was 5.9% (3.2%) for the A1+ genotype group (P = .035 vs. 4%) and 4.2% (4.2%) for the A1− genotype group (P = .85 vs. 4%). In the A1+ genotype group, 86.7% were responders, defined as achieving a weight loss of 4% or more, whereas 61.1% reached the target weight loss in the A1- group (Figure 1).

TABLE 1.

Characteristics by genotype at rs1800497 ANKK1 c.2137G > A (p.Glu713Lys)

Genotype	A1+ (AA or AG)	A1− (GG)	P
N (M/F)	15 (0/15)	18 (4/14)
Genotype	13 GA, 2 AA	18 GG
White/Black/Asian/Hispanic (n)	10/2/1/2	12/0/0/6
Age (y)	50.5 (9.1)	49.3 (13.9)	.80
Weight (kg)	102.8 (22.3)	95.8 (16.7)	.31
BMI (kg/m2)	39.3 (9.1)	35.1 (4.7)	.10
Duration of therapy (wk)	15.2 (6.8)	14.3 (4.6)	.63
Weight loss (%)	5.9 (3.2)	4.2 (4.2)	.20
Weight loss ≥4% (n/%)	13 (86.7)	11 (61.1)	A+ .035 A1− .853

Note: Data are presented as mean (SD) except where indicated. The data satisfied a normality assumption so mean values are presented. P-value is for comparison between groups using two-sided unpaired t-test for between-group comparison, or one-sample t-test for within-group weight change ≥4%.

FIGURE 1.

A, Weight-loss response of individual subjects with mean (SD) by genotype. B, Responders and non-responders by genotype at rs1800497 ANKK1 c.2137G > A (p.Glu713Lys) (A1 + is AA or AG; A1− is GG)

4 |. DISCUSSION

Our findings suggest that the Taq1A variant near DRD2 (rs1800497: ANKK1 c.2137G > A) may play a role in modulating the weight-loss response to the antiobesity medication NB. It is notable that mean weight loss was significantly greater than the threshold for a clinical response of 4% or more in the A1+ genotype (AA or AG) group but not the A1− (GG) group. Moreover, 87% of A1+ individuals were NB responders compared with only 61% of A1− subjects. In fact, while all A1+ subjects lost some weight, one subject in the A1− group lost no weight, and three experienced weight gain while on NB. Given its location proximal to the DRD2 gene,¹⁹ the Taq1A polymorphism is believed to influence the neural response to food⁹ and the probability of obesity,⁶ with A1+ individuals having diminished striatal DRD2 receptor density.⁸ Individuals with the A1+ genotype have an increased probability of obesity⁶ and, consistent with prior data, there was a trend even in this small cohort towards higher BMI in this group,^7,18 which in our study appears to respond better to NB than A1− individuals. We hypothesize that NB helps these individuals overcome a hypofunctioning dopamine reward circuit. It should be noted that the association we identified with rs1800497 does not mean that this specific variant is causally related to the differential response in weight loss. Other genetic variants in linkage disequilibrium such as rs6276 or other variants in or near DRD2 or ANKK2 should be investigated in future studies to better define the causative variant.

Given the wide variability in response to obesity medications, the Taq1A variant in DRD2 has the potential to provide some degree of precision medicine for patients with obesity. Identification of patient characteristics that are associated with response to a particular medication remains elusive, and a large number of patients go through the expense and potential side effects of a medication that may ultimately prove ineffective.³ Our preliminary results appear promising and may provide the first step to identify a marker to help determine weight-loss response to NB. If an individual was found to have the Taq1A A1 + genotype, NB could be considered first-line therapy, and timely and effective care would be more probable. Clearly, we are just starting to explore precision medicine for obesity, but given the current obesity epidemic, it is of the utmost importance to expand upon our findings and explore other possible genetic variants influencing response to various obesity interventions.

This study has important limitations. The retrospective nature of the study did not allow for evaluation of weight loss at a standardized predefined time point. The variability in follow-up time, however, was not different between genotypes or between responders and non-responders. Furthermore, patients were treated by several different practitioners and did not have uniform recommendations for diet and behavioural modifications. Changes in physical activity and adherence to medication use were also not monitored. The sample size was small, but the distribution of the A1+ and A1− genotypes was consistent with findings in published literature.⁶ The study was too small to check for the effects of sex, race or ethnicity. Despite these limitations, these results add important preliminary data and provide the groundwork for a prospective trial of NB therapy to confirm our findings, as well as further explore the mechanism of action of this medication.