This study assessed the effects of 4-weeks of treatment with the MOR antagonist GSK1521498 on hedonic evaluation and consumption of food, binge-eating severity and body weight in obese people with binge-eating symptoms. There were two key effects: GSK1521498 was associated with significant reductions in hedonic preference, specifically for higher concentrations of sugar and fat. GSK1521498 also markedly reduced calorie intake in an ad libitum
buffet and particularly for more palatable foods (high-fat desserts). These effects were dose dependent, occurring only in the 5
mg group, and were, to an extent, correlated with variability in plasma concentrations of GSK1521498. These observations are in keeping with predictions and offer proof of the drug's mechanistic on-target effects.
The effects of GSK1521498 were not entirely consistent across all measures of eating behaviour. Changes in hedonic ratings did not correlate with the effects on calorie intake, perhaps because the latter involved multiple foods of varying compositions, whereas the former used relatively homogeneous food samples. No significant effects on consumption were found for the ad libitum
snacking or the restaurant meal. The snacking paradigm may quantify a type of eating that is relatively difficult to modify: consumption of favourite foodstuffs following prolonged fasting (>12
h). The restaurant meal captures consumption in a context with limited food availability. It appears that the effects of mu-opioid antagonism become apparent only in settings that permit overconsumption. This distinction might be considered akin to eating what is in the fridge rather than eating what is on the plate.
Moreover, these effects on eating behaviour measured by in-unit assessments did not translate into reductions in weight or fat mass. There are several possible reasons for this. First, acute changes in hedonic and consummatory aspects of eating may take longer than 28 days to translate to changes in habitual decision making about food consumption, sufficient to cause significant weight loss. The duration of this study was limited by the toxicological data currently available for GSK1521498. Second, participants were treated as outpatients without dietary restriction or other active lifestyle management interventions. It is a limitation of the design that we do not have data on calorie consumption outside the controlled setting of in-unit assessments. Third, there may be significant heterogeneity between obese patients in their responsiveness to GSK1521498 or other MOR antagonists. For example, post-hoc pharmacogenetic analysis showed that a small sub-group of Caucasian participants carrying one or more copies of the 118G allele of the OPRM1
gene in the 5
mg treatment group showed a greater weight reduction compared with A118 homozygotes: indeed two of the three subjects who showed the greatest weight loss were G-carriers. These data must be treated with caution given the post-hoc, exploratory nature of the analysis and the very small sample size. Nonetheless, they suggest that allelic variation in the MOR gene may contribute to variable therapeutic response to MOR antagonists in the treatment of obesity. Finally, we note that the highest dose of GSK1521498 administered was not sufficient to cause sustained receptor occupancy >90% and high levels of MOR blockade have been considered necessary for optimal therapeutic effect of naltrexone and other MOR antagonists.54
The encouraging safety profile of GSK1521498 suggests that repeated dosing up to 10
may be tolerable in outpatients and higher doses might be expected to demonstrate stronger therapeutic effects in future studies of obesity.
Certain other limitations should be considered. BES scores diminished considerably across all groups even before randomisation, limiting the ability to discriminate any benefits of active treatment on this endpoint. The use of the self-reported BES is a limitation as high placebo response rates have been reported with subjective measures in clinical trials in binge-eating individuals.27
The findings are preliminary as the study was not powered to correct for multiple comparisons; however, we did deploy strategies designed to minimise these.
Multiple anti-obesity agents have been withdrawn due to central/psychiatric/cardiovascular safety concerns.26, 55, 56
GSK1521498 was generally well tolerated compared with placebo, with no deleterious effects detected on anxiety, mood or other aspects of hedonic function or on liver or other blood safety parameters. Akin to previous studies,31, 32
transient impairments on subjective ratings of alertness, contentedness and calmness, were seen with GSK1521498 on the first day of active treatment, largely driven by relative increases in the placebo group. Therefore, the only central effects of GSK1521498 identified were mild and not clinically significant.
Although the lack of effect on weight does not support immediate progression for obesity therapeutics, the effects on eating behaviour and the early pharmacogenetic data warrant further evaluation of GSK1521498 as a more genetically targeted treatment in obesity and binge eating. The similarities between the MOR dysregulation in binge eating and substance addiction motivate exploration of GSK1521498 as a treatment for drug addictions. This study highlights the value of using experimental medicine models to demonstrate clinical proof-of-mechanism for a novel compound targeting abnormal reward-driven eating behaviour.