The Effect of the Triple Monoamine Reuptake Inhibitor Tesofensine on Energy Metabolism and Appetite in Overweight and Moderately Obese Men

International Journal of Obesity (2010) 34, 1634–1643 & 2010 Macmillan Publishers Limited All rights reserved 0307-0565/10 www.nature.com/ijo ORIGINAL ARTICLE The effect of the triple monoamine reuptake inhibitor tesofensine on energy metabolism and appetite in overweight and moderately obese men

A Sjo¨din1, C Gasteyger1, A-LH Nielsen1, A Raben2, JD Mikkelsen2, JKS Jensen3, D Meier2 and A Astrup1

1Department of Human Nutrition, Faculty of Life Sciences, University of Copenhagen, Frederiksberg, Denmark; 2Clinical Development, NeuroSearch A/S, Ballerup, Denmark and 3SlottStat, Bagsværd, Denmark

Background: Tesofensine (TE) is a new drug producing twice the weight loss in obese individuals as seen with currently marketed drugs. It inhibits the presynaptic reuptake of the neurotransmitters noradrenaline, dopamine and serotonin, and is thought to enhance the neurotransmission of all three monoamines. The mechanisms by which it produces weight loss in humans are unresolved. Objective: The aim of this study is to investigate the mechanism(s) behind weight reduction by measuring energy expenditure and appetite sensations in overweight and obese individuals. Design: Thirty-two healthy, overweight or moderately obese men were treated with 2.0 mg TE daily for 7 days followed by an additional 7 days with 1.0 mg TE daily or corresponding placebo (PL) in a randomized, controlled trial. They were instructed to maintain habitual food intake and physical activity throughout. Twenty-four-hour energy expenditure (24-h EE), fat oxidation and spontaneous physical activity were measured in a respiration chamber before and after treatment. Body composition was assessed by dual-energy X-ray absorption and appetite was evaluated by visual analogue scales in conjunction with a standardized dinner. Results: Despite efforts to keep body weight and composition constant, TE induced a 1.8 kg weight loss above PL after 2 weeks’ treatment (Po0.0001). TE also induced higher ratings of satiety and fullness and concomitantly lower prospective food intake than placebo. No significant effect of TE on total 24-h EE could be demonstrated compared with PL, but higher energy expenditure was observed during the night period (4.6%; Po0.05) when adjusted for changes in body composition. Furthermore, TE increased 24-h fat oxidation as compared with PL (18 g; Po0.001). Conclusion: TE has a pronounced effect on appetite sensations and a slight effect on energy expenditure at nightFboth effects can contribute to the strong weight-reducing effect of TE. International Journal of Obesity (2010) 34, 1634–1643; doi:10.1038/ijo.2010.87; published online 18 May 2010

Keywords: appetite; weight loss treatment; pharmacological treatment of obesity; monoamine reuptake inhibitor

Introduction and obesity are major risk factors for diseases such as type 2 diabetes, coronary heart disease, sleep apnoea and certain The prevalence of obesity in industrialized countries has cancers,1 extensive efforts have been made to improve their increased continuously over the past 30 years.1 Data from treatment and prevention. Unfortunately, with the excep- the National Health and Nutrition Examination Survey III tion of bariatric surgery, these treatments most often show that 32% of adults in the United States were obese produce only a modest and transient weight.3–5 As bariatric (body mass index X30.0 kg mÀ2) in 2003–04.2 As overweight surgery is indicated only in a minority of patients (classically those with a body mass index X40.0 kg mÀ2), the vast Correspondence: Dr A Sjo¨din, Department of Human Nutrition, Faculty of Life majority of obese adults need novel treatments, in addition Sciences, University of Copenhagen, Rolighedsvej 30, DK-1958 Frederiksberg to behavioural modifications including dietary and exercise C, Denmark. interventions, to help them lose weight and subsequently E-mail: [email protected] Received 20 October 2009; revised 16 March 2010; accepted 4 April 2010; maintain a healthy body weight. New chemical entities published online 18 May 2010 targeted at the increasing obese population are being Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1635 investigated for potential anti-obesity effects. They all aim, hypercholesterolaemia or hypertriglyceridaemia, thyroid through various mechanisms of action, to safely improve disease, malabsorptive intestinal disorders, no weight change weight loss and maintenance obtained through diet and of 43 kg within 2 months before screening, surgically exercise alone, by decreasing energy intake or uptake (for treated obesity, systemic infections, or inflammatory dis- example, of fat), sometimes in combination with stimulation eases, cardiovascular disease or significant abnormalities on of thermogenesis. the electrocardiogram, hypertension (X155 systolic and/or Tesofensine (TE) inhibits the presynaptic reuptake of the X95 mm Hg diastolic blood pressure), heart rate (HR) 490 neurotransmitters noradrenalin, dopamine and serotonin.6 beats per minute (b.p.m.), cancer within the past 5 years Selective serotonin reuptake inhibitors are used to treat (excluding treated basal cell carcinoma), human immuno- depression and eating disorders, but they have generally deficiency virus infection or serological evidence of active failed to produce sustained weight loss.7 Sibutramine, in hepatitis B and/or C, clinically significant eye disorder or addition to its effect on serotonin, also inhibits the reuptake medication with known ocular toxicity, not previously been of other monoamines and has been shown to enhance treated with TE or any other investigational drug within postprandial satiety, reduce total calorie intake and to 30 days or five half-lives (whichever was longer) preceding diminish the decline in energy expenditure usually asso- the first dose of study drug. ciated with a diet-induced negative energy balance.8–10 A meta-analysis of four, randomized, placebo (PL)- Treatment controlled, double-blind, phase-II trials over 14 weeks in a TE has a very long half-life of about 8 days in humans total of 968 patients, examining the effect of TE on body and the clearance after oral administration is low weight in subjects with Parkinson’s or Alzheimer’s disease, (30–40 ml minÀ1). After the first chamber visit, subjects were showed a PL-adjusted, unintended weight loss of up to 4% in randomly allocated to either PL or TE, 2 mg taken orally once an obese subgroup treated with TE.11 daily for 7 days followed by 1 mg the next 7 days. The up- Recently published data from the first randomized, titration with 2 mg was chosen to obtain a steady-state level double-blind, Pl-controlled phase-II trial in primarily on 1 mg TE within the 2-week study period. healthy, obese subjects showed that TE was able to produce B10% greater weight loss than PL after 24 weeks of 0.5 or 1.0 mg TE daily.12 This is about twice that of currently Study protocol approved drugs. Appetite was significantly suppressed after Each subject took part in the study for a total of 7–8 weeks, an overnight fast after treatment with TE in this study. divided into a run-in period of 7–14 days, 14 days of However, energy balance could also have been affected by a treatment and a 28 (±3) days follow-up period. Subjects simultaneous stimulation of thermogenesis, as has been seen were informed about the importance of weight stability. in animals.13 They were counselled by a dietician on how to maintain The aim of this study was to assess the potential effect of TE dietary habits and not to change their physical activity in on energy metabolism and on subjective appetite sensations. order to maintain energy balance and minimize changes in It was designed as a 14-day randomized, double-blind, body weight and composition. After the experimental PL-controlled, parallel-group, single-centre study including treatment period, subjects were invited to be included in a overweight or moderately obese, healthy male subjects. weight-reduction programme for 6 months, consisting of individual and group counselling sessions, aimed to help them to adopt a calorie-restricted, low-fat diet and to Subjects and methods increase their physical activity. After the run-in period, 24-h energy expenditure (24-h EE) Subjects, inclusion and exclusion criteria and related measures were conducted in a respiratory Screening of subjects took place at the investigational site. chamber calorimeter at baseline (Day À1 to 0) and at the Potential subjects were recruited by public advertisement in end of treatment (Day 13–14). The subjects were instructed daily newspapers. Thirty-two male subjects were recruited after not to perform any strenuous physical activity and to refrain having signed an informed consent form. They were randomly from alcohol consumption during the 2 days before the assigned to one of the two treatment groups at the baseline visit. chamber measurements. Moreover, for additional standardi- To comply with inclusion and exclusion criteria, subjects zation and in order to accustom them to the chamber and had to have a body mass index between 28 and 35 kg mÀ2, thereby reduce the effect of potential mental stress during age between 18 and 50 years, not undergoing a physical the subsequent 24-h measuring period, subjects slept in the training programme or planning major changes in physical chamber with the door open at the night before the actual activity during the study, not use tobacco, not taking any measurement (Day À2 and Day 12). The measurements concomitant medication including high-dose vitamins or started at Day À1 and Day 13 at 0900 hours and ended 24 h regular OTC preparations, not abuse or be dependent on any later (that is, Day 0 and Day 14 at 0900 hours). drug, including alcohol, have no endocrine, neurological The respiratory gas exchange was measured in one of the or psychiatric diseases, hepatic or renal dysfunction, department’s two duplicate respiration chambers.14,15 The

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1636 same chamber was used on both occasions for any subject. diameter (reclining), and hip and waist circumference (in a The possible difference between the two chambers is standing position) were also measured to the nearest 100 g accounted for in the statistical model. Gas exchange in the and 0.5 cm, respectively. chamber was calculated from measured airflow and concen- Appetite was assessed in conjunction with the dinner

trations of O2 and CO2 at the outlet of the chamber as well as meal. Visual analogue scales (VAS), 100 mm in length with in the fresh air going in. Energy expenditure and substrate words expressing the most positive and the most negative partitioning were later calculated based on equations by Elia rating anchored at each end, were used to measure feelings and Livesey,16 including data on nitrogen losses in urine of hunger (‘how hungry do you feel?’), satiety (‘how satisfied from the same time period (analysed according to the do you feel?’), fullness (‘how full do you feel?’) and Kjeldahl method). prospective food consumption (‘how much do you think Subjects were instructed to be sedentary during their stay you can eat?’). Nausea was assessed simultaneously, using in the chamber, sitting (day) or lying (night) according to the the same technique. All questions were formulated in protocol, except for two short periods of defined light Danish.18 This procedure was performed 30 min before and physical activities in the morning and in the afternoon. again immediately before dinner (pre-meal), and repeated The physical activity comprised 15 min of cycling at 75 W every 30 min thereafter (post-meal) for a total of 300 min. A and a short period of standardized light walking (25 rounds composite appetite score was calculated for all time points in the chamber, corresponding to B5 min). They were according to the formula: further instructed to go to bed at 2200 hours, but were ½satiety þ fullness þð100 À prospective consumptionÞ allowed to read until 2300 hours when the light was þð100 À hungerÞ=4 switched off. On the next day subjects were awakened at At 0900 hours the registration of 24-h respiratory gas 0730 hours, but instructed to remain quietly in bed, apart exchange was terminated and the subjects were asked to from getting up for voiding, if needed. Registration of empty their bladder and allowed to leave the chamber for respiratory gas exchange between 0800 and 0900 hours was evaluation of body composition by dual-energy X-ray used for assessment of fasting resting metabolic rate, a proxy absorption. Blood pressure was then measured with subjects for basal metabolic rate (BMR). in a sitting position following 5 min of rest, and blood was Spontaneous physical activity (SPA) was assessed by two drawn and stored for later analysis. microwave radar detectors (Sisor Mini-Radar, Statistic Input Blood samples for analysis of TE pharmacokinetics were System SA, Lausanne, Switzerland). The radar detected when collected in the morning Day 7 (±1) and Day 14 before the subject was moving, and the generated signal was administration of the dose, and finally at follow-up visit Day received by the transceiver and electronically stored for later 42 (±3). Concentrations of TE (NS2330) and its active analyses. SPA measurements indicate the percentage of time metabolite (NS2360) were later analysed by Quintiles AB when the subject was active. A laboratory technician kept (Uppsala, Sweden). Additional blood samples were taken at the patient under surveillance during daytime and a trained each visit and analysed in a central laboratory (Capio medical student was in attendance during the night. Diagnostics A/S, Copenhagen, Denmark) for safety and The amount of energy provided during the chamber stay efficacy monitoring. was individually calculated to cover energy expenditure, HR was calculated as an average from a continuous based on age, gender, body composition and physical 17 registration throughout the 24-h chamber visit using a activity. The total energy content in the three meals portable HR monitor (Nonin Avant 4000, Nonin Medical consumed at both stays was distributed with 20% at breakfast, Inc, Plymouth, MN, USA). 33% at lunch and 47% at dinner. The contribution of energy Potential effects on mood state were assessed using the in the diet was 35% fat, 15% protein and 50% carbohydrate, Profile of Moods States (POMS Brief) questionnaire.19 resembling a typical Danish diet. Dietary calculations were Standard 12-lead electrocardiograms were recorded three made using a database from the Danish National Food Agency times with half-hour intervals at baseline (Day 0), and at the (‘Dankost 3000’, 2006). Any leftovers were recorded and end of treatment (Day 14) at 6, 6.5 and 7 h after the dose, and actual energy intake was re-calculated. Meals and beverages analysed for potential effects on PR, QT, QRS and QTc were served at 0900 hoursFbreakfast, 1300 hoursFlunch, intervals that could be attributed to TE. 1500 hoursFcaffeine-free coffee or tea, 1800 hoursFdinner Additional safety measures were treatment emergent adverse and at 2000 hoursFcaffeine-free coffee or tea. Water was events (AEs) (collected up to 28 days after the last treatment offered ad libitum during the first chamber stay and carefully dose), ophthalmoscopy and a standard physical examination. recorded. The same quantities were then served to the subject at corresponding time points during their second stay. To assess body composition, dual-energy X-ray absorption (GE Lunar Prodigy Advance, GE Medical Systems, Madison, Sample size and statistical analyses WI, USA) scanning was carried out according to a standar- Under the assumption that the s.d. of a 24-h EE measure- dized procedure, with subjects in a fasting state after leaving ment is 150 kJ, the study was powered to detect a treatment the chamber. At this time, body weight, abdominal sagittal difference of 300 kJ for the 24-h EE measurement using a

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1637 two-sided test, a significance level of 5%, a power of 90% The authors had full access to all data and had joint final with 13 subjects in each treatment arm. Expecting B20% responsibility for the decision to submit for publication. withdrawals, a total of 16 subjects was therefore included in each treatment arm. The basic model for the analysis of continuous end points at Day 14 was an analysis of variance with treatment as a Results class effect and subjects as random effects. The error variances were allowed to differ between treatment groups. Study population, protocol compliance and dropouts Baseline values were used as covariates. Treatment differ- A total of 32 Caucasian subjects were successfully screened ences, adjusted for covariates (for example, SPA, fat free and subsequently randomized to treatment with either TE or mass) in the final model, were shown with a two-sided 95% PL. The treatment groups had similar demographic and confidence interval (CI), as well as the corresponding baseline characteristics (Table 1). The reported drug com- P-value. Analyses of the appetite profile were carried out by pliance was high with no difference seen between groups. Of a repeated measures analysis of variance with treatment and a total of 21 prescribed tablets, 20.8±1.4 and 20.3±1.2 time points as fixed effects and a covariance structure tablets were reported taken in the TE and PL groups, reflecting the dependence among observations at different respectively. The short, but initially high, dosing of TE in time points for any subject. The covariance structure was this study produced plasma concentrations (mean±s.d.) at allowed to differ between treatments. days 7, 14 and 42±3 (follow-up) of TE of 14.95±4.37, À1 A level of significance of 5% was used to decide statistical 16.20±3.77, 1.72±1.24 ng ml and of its metabolite, significance for primary and secondary end points. The tests N-desmethyl-metabolite NS2360 of 1.12±0.45, 2.67±0.76 À1 were two sided. For the secondary end points, no correction and 1.11±0.69 ng ml , respectively. Plasma concentrations for multiplicity was used. Two respiration chambers were of these compounds were below detection levels À1 used. The same chamber was used on both occasions for a (0.01 ng ml ) in all subjects receiving PL. specific subject, but differed between subjects. The possible One subject in the PL group could not be persuaded to take difference between the two chambers was taken into account part in the follow-up assessment, but this subject was still in the model. Randomization data were kept strictly included in the efficacy analyses because complete data sets confidential, accessible only to authorized persons, until were obtained from both chamber visits. One subject in the the time of un-blinding. TE group withdrew consent during the intervention because of an adverse advent (abdominal pain or suspected peptic ulcer). All the remaining subjects completed all procedures Site, ethics and monitoring per protocol, with the exception that despite encourage- The study was carried out at the Department of Human ment, two subjects in the TE group could not eat all the food Nutrition, Copenhagen University, Denmark, between designated for their consumption during the second time in March and December 2007. All subjects gave written the chambers due to extreme fullness. On average, this led to informed consent before any study-related procedures. The project was registered at http://www.ClinicalTrials.Gov (Identifier NCT00428415), and was approved by the Ethics Committees for the Copenhagen and Frederiksberg Munici- Table 1 Subject characteristics at baseline palities (now Region Copenhagen), Denmark (J. no. Variable Tesofensine (n ¼ 16) Placebo (n ¼ 16) 02-321398), the Danish Medicines Agency (J. no. 2612-3271) and The Data Protection Agency (J. no. 2006-41-7000). The Age (years) 37.9±6.8 35.0±6.8 trial was conducted in accordance with the International Height (cm) 182.7±5.7 181.6±4.1 Weight (kg) 102.6±9.0 102.6±6.8 Conference on Harmonization/Good Clinical Practice (ICH/ BMI (kg mÀ2) 30.7±1.8 31.1±1.7 GCP) guidelines, and monitored by Cyncron Clinical FFM (kg) 68.5±6.0 69.4±4.9 Research Associates (Birkerød, Denmark). Waist circumference (cm) 105.1±7.1 107.6±6.6 Sagittal diameter (cm) 23.3±1.1 23.7±1.4 Waist/hip ratio 0.9±0.1 1.0±0.0 24-h EE (MJ per day) 11.1±0.8 11.1±0.8 Role of the funding source 24-h EE/FFM (kJ dayÀ1 kgÀ1) 163±10 160±11 The research relating to this paper was funded by Neuro- 24-h EI (MJ per day) 11.2±0.8 11.2±0.7 Search A/S (Ballerup, Denmark). SPA (%) 15.5±3.0 15.9±3.2 BP (mm Hg) 117.5±8.3 117.4±8.7 The study was designed by an external advisory board with systolic BPdiastolic (mm Hg) 82.1±6.1 82.4±6.0 representatives from the sponsor and with the principal HR (beats per minute) 61.9±6.6 64.1±11.7 investigator (AA) as chairman. Data were collected by the monitoring company (Cyncron AS) and were assessed jointly Abbreviations: BMI, body mass index; BP, blood pressure; FFM, fat free mass; HR, heart rate; SPA, 24-h EE, 24-h energy expenditure; 24-h EI, 24-h energy by the statisticians, the authors and the sponsor. The data intake. Results presented as mean and s.d. at baseline for each study group. were interpreted and the paper was written by the authors. No significant differences were found between groups.

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1638 Table 2 Effects on body weight and composition

Parameter Tesofensine Placebo Difference a P-value *

Before b After b Before b After b

Body weight (kg) 102.6±9.0 100.3±10.0 102.6± 6.9 102.2± 7.1 À1.8 (À2.5; À1.1) o0.0001 FFM (kg) 68.5±6.0 67.0±4.9 69.4±4.7 69.4±4.9 À1.5 (À2.4; À0.6) 0.001 FM (kg) 34.6±5.6 33.3± 6.2 33.8±4.8 33.4±5.1 À1.0 (À1.6; À0.4) 0.002 Waist circumference (cm) 105.1±7.1 103.3±8.1 107.6±6.6 106.6±6.8 À0.7 (À1.7; 0.4) 0.23 Sagittal diameter (cm) 23.3±1.1 22.8±1.2 23.7±1.4 23.8±1.3 À0.7 (À1.2; À0.2) 0.01 Waist/hip ratio 0.95±0.05 0.94±0.06 0.97±0.05 0.97±0.05 À0.00 (À0.01; 0.00) 0.51

Abbreviation: FFM, fat free mass; FM, fat mass. aDifference after treatment between treatments from analysis of covariance with treatment as factor and baseline value as covariate expressed as mean with 95% confidential interval. bDenotes estimated means and s.d. before and at the end of treatment. *Level of significance of difference between treatments. Parameters significantly affected by tesofensine treatment are marked in bold text.

a 1% lower energy intake for the entire group (not significant). In contrast, none of the subjects receiving PL treatment had any problems eating the portions served to them.

Weight and body composition Despite the efforts taken to minimize alterations in body weight and composition between chamber visits, TE-treated subjects lost more body weight compared with subjects in the PL group. This was due to losses of both FM and FFM. Changes in sagittal abdominal diameter as well as waist circumference and waist/hip ratio did, however, not reach statistical significance within the time of observation (Table 2).

Energy metabolism The 24-h EE corrected for changes in FFM was not significantly affected by TE treatment ( þ 2.2 J kgÀ1 minÀ1 (95% CI: À1.3; 5.7) P ¼ 0.22) compared with PL. However, during the night period (2300–0800 hours; a proxy for sleeping metabolic rate), EE/FFM was higher after TE treatment ( þ 4.1 J kgÀ1 minÀ1 (95% CI: 0.1; 7.5) Po0.05) compared with PL, corresponding to an increase of 4.6% while resting metabolic rate (0800–0900 hours) was not Figure 1 Effect on energy expenditure and fat oxidation. Mean changes significantly different between treatments (P ¼ 0.23) from baseline in energy expenditure expressed per kg of FFM (a) and fat (Figure 1a) oxidation (b) for subjects treated with TE (filled bars) and PL (open bars) with error bars indicating s.d. *Denotes significant differences (Po0.05) between groups. Substrate oxidation In the TE group, total 24-h fat oxidation increased by 17.7 g (95% CI: 1.4; 48.9) compared with PL (Po0.001) (Figure 1b). Po0.01), satiety (20 vs À3 mm; Po0.05) and composite Concomitantly, a 1.8 g (95% CI: 0.1; 3.5) lower nitrogen appetite (12 vs À4 mm; Po0.01), whereas subjects scored excretion was found (Po0.05), indicative of a lower lower on prospective consumption (À17 vs 4 mm; Po0.01), protein oxidation while carbohydrate oxidation remained after TE compared with PL (Figure 2). PL-subtracted values unchanged. for hunger did, however, not reach statistical significance (À5 vs 5 mm; P ¼ 0.17). VAS scores were also significantly different for fullness and prospective consumption through- Appetite out the entire 5-h period when appetite was assessed Subjective appetite sensations evaluated by VAS scores before (Po0.05). The changes from pre-meal values in VAS scores the dinner (pre-meal) were higher for fullness (5 vs À5 mm; were, however, not different for any of the appetite-related

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1639 a 100 (5.6 vs 2.7 mm; Po0.05), and increased slightly, but without 90 80 reaching statistical significant differences, after the test meal. 70 Nausea did not, however, explain the effect on appetite seen 60 50 after TE treatment when adjusted for in the statistical model. mm 40 Fullness 30 20 10 0 SPA 17:00 18:00 19:00 20:00 21:00 22:00 23:00 The level of SPA over 24 h, measured by the radar system, was b 100 not different between groups. However, when looking 90 80 separately at the night period, when subjects were mostly 70 asleep and when a significant difference in EE was demon- 60 50 strated, SPA tended to be slightly higher (0.6% units (95% CI: mm

Satiety 40 30 0.0; 1.2) P ¼ 0.05) in TE compared with PL. An adjustment for 20 individual differences in SPA did not, however, change the 10 0 findings on EE. 17:00 18:00 19:00 20:00 21:00 22:00 23:00 c 100 90 80 HR and electrocardiogram 70 HR increased on average by 10 b.p.m. (95% CI: 7; 12.9) 60 50 throughout the 24 h of registration (Po0.0001) after TE mm

Hunger 40 treatment compared with PL. The difference was less 30 20 pronounced during the night (2300–800 hours: 6.6 b.p.m. 10 (95% CI: 3.6; 9.7), Po0.001) than during the day (0900–2300 0 P 17:00 18:00 19:00 20:00 21:00 22:00 23:00 hours: 12.2 b.p.m. (95% CI: 9.9; 15.4), o0.0001). No effects on electrocardiogram apart from HR were found and there d 100 were no correlations between changes in EE and changes in 90 80 HR (slightly negative slope and P ¼ 0.52 for TE-treated 70 subjects). 60 50 mm 40 30 20 Blood pressure Prosp. food intake 10 Diastolic blood pressure increased by 4.7 mm Hg (95% CI: 0 1.7; 7.8, P 0.01) after TE treatment, whereas systolic blood 17:00 18:00 19:00 20:00 21:00 22:00 23:00 o pressure remained unchanged (1.3 mm Hg (95% CI: À2.5; e 100 P 90 5.1), ¼ 0.49) compared with PL. 80 70 60 50 Laboratory assessments mm 40 Compared with PL, blood lipids were reduced on TE: 30 20 triglycerides by 23% (Po0.001), total cholesterol by 10% Comp. appetite score 10 (P 0.001) and low-density lipoprotein cholesterol by 10% 0 o (P 0.01), whereas high-density lipoprotein cholesterol was 17:00 18:00 19:00 20:00 21:00 22:00 23:00 o Time unchanged (Table 3). For free fatty acids (FFA), a significant increase was observed on TE compared with PL (by 33%, Figure 2 Effect on appetite. Mean and s.d. (error bars) for ratings of (a) perceived hunger (‘how hungry do you feel?’), (b) satiety (‘how satisfied Po0.01). No significant differences were, however, observed do you feel?’), (c) fullness (‘how full do you feel?’) and (d) prospective food in fasting glucose, glycosylated hemoglobin A1c (HbA1c), consumption (‘how much do you think you can eat?’) using visual analogue insulin, adiponectin, leptin, ghrelin, cholecystokinin (CCK), scales, 100 mm in length and with baseline values as covariates. A composite C-reactive protein or lactate dehydrogenase. appetite score (e) was also calculated for all time points according to the formula: [satiety þ fullness þ (100Àprospective consumption) þ (100Àhunger)]/4. Solid blue lines are results after TE treatment, while dotted red lines are after PL treatment. Profile of Moods States Overall Profile of Moods States has not been subjected to questions, indicating that the response to the meal was statistical analyses due to insufficient power. However, there similar for the two treatment groups. seems to be a tendency that subjects treated with TE Pre-meal nausea, scoring at the low end of the scale, was reported a general slight shift in parameters in a direction slightly but significantly higher in the TE-treated group associated with poorer mood compared with PL. This can be

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1640 Table 3 Effects on metabolic variables

Parameter Tesofensine Placebo Difference a P-level *

Before b After b Before b After b

Triglycerides (mmol lÀ1) 1.46±0.45 1.11±0.32 1.52±0.68 1.49±0.58 À0.34 (À0.52; À0.15) 0.0008 Total cholesterol (mmol lÀ1) 4.08±1.01 4.51±0.87 4.76±0.70 4.73±0.78 À0.51 (À0.77; À0.26) 0.0003 LDL-cholesterol (mmol lÀ1) 3.53±0.96 3.11±0.73 3.26±0.62 3.23±0.65 À0.32 (À0.54; À0.10) 0.005 HDL-cholesterol (mmol lÀ1 1.03±0.18 0.97±0.18 1.00±0.13 0.98±0.13 À0.03 (À0.08; 0.02) 0.26 FFA (mmol lÀ1) 0.44±0.10 0.53±0.15 0.41±0.09 0.38±0.09 0.31 (0.05; 0.21) 0.002 Glucose (mmol lÀ1) 5.74±0.46 5.64±0.42 5.61±0.35 5.56±0.31 À0.02 (À0.18; 0.14] 0.83 HbA1c (%) 5.30±0.26 5.25±0.24 5.28±0.29 5.25±0.33 À0.02 (À0.08; 0.04) 0.51 Insulin (pmol lÀ1)51.2±18.5 58.1±22.6 73.4±67.1 62.9±36.4 7.1 (À4.0; 18.2) 0.20 Adiponectin (mgmlÀ1) 7.94±2.64 6.81±2.54 5.63±2.63 5.31±2.33 0.51 (À1.17; 0.14) 0.12 Ghrelin (pg mlÀ1) 827±302 844±271 878±299 893±311 À8(À129; 113) 0.89 CCK (pmol lÀ1) 0.59±0.20 0.53±0.29 0.77±0.32 0.76±0.20 À0.17 (À0.36; 0.01) 0.07 C-reactive protein (mg lÀ1)1.9±2.3 1.2±1.4 1.8±1.6 1.9±1.4 À0.7 (À1.9; 0.5) 0.26 LDH (U lÀ1) 143±27 141±31 140±25 140±21 À2(À14; 10) 0.72

Abbreviations: FFA, fat free acids; HDL, high-density lipoprotein; LDH, lactate dehydrogenase; LDL, low-density lipoprotein. aDifference after treatment between treatments from analysis of covariance with treatment as factor and baseline value as covariate expressed as mean with 95% confidential interval. bDenotes estimated means and s.d. before and at the end of treatment. *Level of significance of difference between treatments. Parameters significantly affected by tesofensine treatment are marked in bold text.

exemplified by a slight numerical increase in total mood with PL, complicating the interpretation of data on EE. disturbance for TE (9.8±20) vs PL (1.8±9.8). Thermogenesis (resting metabolic rate) normally decreases during weight reduction, mainly due to loss of FFM and a AEs reduced sympathetic nervous system activity, and to a lesser extent, due to loss of fat mass.20–23 Changes in FFM should No serious AEs occurred in either treatment group. The therefore be part of an analysis when assessing a potential majority of the AEs were mild (42 of 57 events), whereas the effect on thermogenesis. Accordingly, the loss of FFM in the number of moderate to severe AEs was higher in the TE TE group should theoretically have caused a decrease of 2% group (seven events vs two events). A total of 39 events (P 0.05) in resting metabolic rate.24 However, although occurred in the TE group (in 93.8% of subjects in the group), o there was a 1.9% drop in 24-h EE/FFM in the PL group whereas 18 events occurred in the PL group (in 62.5% of (P 0.05), EE/FFM was clearly maintained in the TE-treated subjects in the group), with insomnia (TE 5; PL 0); headache o group ( þ 1.0%; P ¼ 0.31) (Figure 1a). This numerical differ- (TE 5; PL 2), loss of appetite (TE 5; PL 3) and dry mouth (TE 3; ence between groups, although not statistically significant PL 0), seemingly more frequently reported in the actively (P ¼ 0.22), corresponds to an approximately 250 kJ increase treated group. Causality showed a similar number (7 vs 7) of in energy expenditure over 24 h, and could be considered a events deemed ‘Not related’ between the two treatments, biologically meaningful effect in relation to weight whereas the number of events deemed ‘Probable relation’ regulation. It should also be acknowledged that the lack of was higher for the TE-treated group compared with controls effect of TE on the primary outcome in this trial may have (TE 22 vs PL 6). AE outcomes were however, similar in the been due to a statistical type II error arising from the limited two treatment groups. sample size. Although stimulation of serotonergic and dopaminergic Discussion systems does not lead to increased thermogenesis in humans, the sympathetic nervous system driven by nora- The main findings of this study, apart from the favourable drenaline-dependent neurotransmission is a key pathway effects of TE on appetite, is that treatment with TE results in known to increase thermogenesis in man.15,22 It is therefore decreased body weight and increased fat oxidation, although likely that the observed effect on sleeping EE is due to a the primary outcomeF24-h EE corrected for changes in higher sympathetic tone that is easier to detect during sleep, FFMFwas not significantly affected. However, it should be when the variability of EE is lowest. However, the effect is noted that during the night (2300–0800 hours), when rather small and additional research is required to fully assess subjects were generally sleeping, a significant increase in the potential effects of TE on EE. EE/FFM ( þ 4.6%; Po0.05) was observed in TE-treated Energy expenditure can obviously be affected differently compared with PL-treated subjects. under less controlled, free-living conditions, in which Despite efforts to keep subjects’ food intake and physical subjects are not confined to a small room, instructed to be activity unchanged, a significant weight loss of almost 2 kg sedentary and closely monitored for adherence to the was found already after 2 weeks of treatment with TE but not protocol. Several lines of evidence suggest that increased

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1641 central monoaminergic neurotransmission affects energy support for this hypothesis, as increases in HR were clearly expenditure, either directly by increasing thermogenesis at not linked to changes in EE. rest, or indirectly by increasing locomotor activity.25,26 An Fat oxidation was increased in the TE-treated subjects. For animal study showing that high doses of sibutramine 24 h, this corresponded to 18 g compared with PL (Po0.05). increase locomotor activity has suggested that the latter The effect on fat utilization was also accompanied by an might be the main mechanism by which this drug leads to increased availability of FFA (elevated plasma concentration) increased EE27 and subsequent weight loss. Whether this is a and a reduction in serum triglyceride (s-TG) concentration. pharmacologically relevant mode of action of the drug when Both are likely to be effected by a period with negative given in more moderate doses, relevant for clinical use, is energy balance per se. Although subjects during the 24-h however unknown. evaluation of substrate utilization at the end of treatment Effects on locomotion, antagonized by the dopamine were in a neutral energy balance similar to baseline

D1-receptor antagonist SCH23390, have also been shown conditions, a negative energy balance before this, as after administration of TE in rats.28 It is possible that indicated by the weight loss seen in the TE-treated group, treatment with TE may also result in increased motor may have had a persistent effect on these parameters. When activity in man. However, this was not found to be the case corrected for weight changes, the effects on s-TG, total and in this study, in which an increase should have been detected low-density lipoprotein cholesterol disappeared, while the as increased SPA by the radar system used in the chambers. A enhancing effect on fat oxidation no longer reached minor effect of potentially disturbed sleep leading to statistical significance (P ¼ 0.13). increased EE could also be suspected, as SPA tended to be As could be expected based on the proposed effects of TE slightly higher during the night after TE treatment in this on a number of monoaminergic postsynaptic receptors, study. However, EE during the night remained significantly appetite was found to be robustly affected in this study, higher in this group even after correction for changes in SPA which is also in line with results from the previous phase-II during this period. Furthermore, there was no indication trial in obese subject.12,30 The elevated feelings of fullness that SPA for the entire 24-h period was affected (a drop of 1.6 and satisfaction, in combination with a higher composite vs 0.9% units for TE and PL, respectively). appetite score and a simultaneously lower anticipated food Considering the almost equally low nanomolar efficacy on consumption at a following meal, would probably result in in vitro inhibition of the suggested receptors involved,29 it is lower energy intake in an ad libitum meal situation, and likely that TE could show a direct effect on EE through would most likely result in a lower habitual energy intake. positive modulation of several monoaminergic pathways However, this has still to be demonstrated in future studies. involved in metabolic activity. Animal studies with sibutra- It is plausible that the effect on appetite is the main mine support this assertion, as they have found a dose- mechanism leading to the negative energy balance and dependent stimulation of EE, of which brown adipose tissue subsequent weight loss seen with treatment with TE. metabolism accounts for the major thermogenic response.26 Reduced appetite may be due to nausea. The pre-meal VAS Postprandial EE has also been shown to be increased when score for nausea was also slightly higher for TE compared human subjects were treated with higher doses of sibutra- with PL. However, both groups reported scores in the lower mine.8 An effect of TE on EE was also found in a study using range and VAS scores for appetite were not affected when an inbred rat model of diet-induced obesity (DIO), in which nausea was included as a covariate in the analysis. It should animals treated with TE not only lost weight as a result of also be noted that only one subject (in the TE-treated group) lowered energy intake, but interestingly also lost consider- reported mild nausea as an AE. ably more weight (17.9 vs 6.7%) compared with pair-fed The mechanisms by which TE might affect appetite are not vehicle-treated DIO animals.13 This clearly suggests that TE, fully understood. The multiple targets of TE inhibiting apart from its effect on appetite, also resulted in increased noradrenaline, serotonin and dopamine reuptake function, EE. It should be noted, however, that neither EE nor and thereby potentially influencing a great number of locomotor activity was assessed in this study, thus it still monoaminergic postsynaptic receptors, make any prediction remains to be determined whether thermogenesis, loco- of specific anatomical and pharmacological targets particu- motor activity, or both, contributed to this effect. larly difficult. However, very recent animal studies have TE shares the presynaptic reuptake inhibition of noradre- helped us in our understanding of the receptors involved in naline and serotonin, and potentially also of dopamine, with the immediate appetite-suppressing effect of TE.28 Acute sibutramine. It also shows some of the same hemodynamic administration of TE to diet-induced obese rats was shown to effects, for example, increased HR. Sibutramine has been immediately suppress food intake. The hypophagic response found to exert a small thermogenic effect in humans at was almost completely reversed by co-administration of 8 higher doses, and it has been suggested that the thermo- prazosin (a1-adrenoreceptor antagonist) and partially antag- genic effect of sibutramine may be indirectly responsible onized by co-administration of SCH23390 (a dopamine for the increase in HR, that is, a hemodynamic effect of D1-receptor antagonist). Other monoamine antagonists increased sympathetic nervous system activity and increased tested were ineffective. On the basis of these findings, it is tissue heat production. However, in this study, we found no most likely that the acute effect on appetite is mainly

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1642

mediated by a´1-adrenoceptors, presumably located in the 14 days’ treatment with TE. We therefore propose that the medial hypothalamus, and perhaps by a minor contribution weight-reducing effect of TE is potentially caused by a dual

from dopamine D1-receptors, likely located in the ventral mechanism, mainly by decreasing appetite, but possibly also striatum. However, these observations are restricted to the to some extent by stimulating thermogenesis. acute effects of TE, and do not exclude that other targets might be involved in other aspects of the potentially multiple chronic effects of TE on metabolism and appetite, Conflict of interest or whether the same targets may have several roles. Administration of TE to obese subjects in doses from AA and JDM receive an honorarium as a consultant and AA is 0.25 mg and upwards has been shown to induce a small but a member of the Tesofensine Advisory Board for NeuroSearch significant increase in HR at rest.12 A mild increase in supine A/S. AR and DM are employees of NeuroSearch A/S and diastolic and systolic blood pressure of 1–3 mm Hg was also own shares in the company. J-KSJ works as a consultant for seen at 0.25 and 0.5 mg doses, which are the doses to be used NeuroSearch A/S. CG is presently employed by Merck for further clinical development. Only the highest dose of TE Serono, Switzerland, while AS and ALHN have no conflict (1 mg) induced a significant increase in blood pressure in of interest. obese patients. These changes in vital signs were generally less pronounced in an elderly population with Parkinson’s and Alzheimer’s disease.11 Acknowledgements Although the types of AEs reported in this study were in line with those in longer-term intervention studies,11,12 the We thank John Lind (technician) and Nathalie Bendsen (PhD incidence of some AEs tended to be higher, and the effects student) for their help in collecting the data. The study was on blood pressure and HR were more pronounced. However, financed by NeuroSearch AS, Denmark. this is probably explained by the forced titration scheme used in this study (2 mg per day for 7 days) in order to References achieve steady state concentrations equivalent to 1 mg per day. The effects on vital signs found in this short-term study 1 Ogden CL, Yanovski SZ, Carroll MD, Flegal KM. The epidemiology have therefore to be thoroughly evaluated under relevant of obesity. Gastroenterology 2007; 132: 2087–2102. clinical conditions in future trials. 2 Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal Several weight-reduction studies have found that the KM. Prevalence of overweight and obesity in the United States, JAMA relative reduction in BMR is larger than the relative 1999–2004. 2006; 295: 1549–1555. 31,32 3 Curioni CC, Lourenco PM. Long-term weight loss after diet reduction in body weight. Other studies have indicated and exercise: a systematic review. Int J Obes (Lond) 2005; 29: that a lower BMR for a given body size and composition may 1168–1174. be involved in some cases of obesity. In a prospective study, 4 Shaw K, Gennat H, O’Rourke P, Del MC. Exercise for overweight Cochrane Database Systemic Reviews Ravussin et al.33 reported that subjects with a low BMR, or obesity. 2006; Issue No. 4, Art. no.: CD003817. adjusted for body composition, were at higher risk of 5 Wing RR, Hill JO. Successful weight loss maintenance. Annu Rev subsequent weight gain than subjects with a high adjusted Nutr 2001; 21: 323–341. BMR. Furthermore, a meta-analysis of BMR has shown that 6 Thatte U. NS-2330 (Neurosearch). Curr Opin Investig Drugs 2001; 2: formerly obese subjects have a 3–5% lower weight-adjusted 1592–1594. 34 7 Ayyad C, Andersen T. Long-term efficacy of dietary treatment of BMR than matched never-obese controls. If TE has a obesity: a systematic review of studies published between 1931 sufficient thermogenic effect to attenuate the decline in and 1999. Obes Rev 2000; 1: 113–119. metabolic rate seen after weight reduction, without causing 8 Hansen DL, Toubro S, Stock MJ, Macdonald IA, Astrup A. Thermogenic effects of sibutramine in humans. Am J Clin Nutr unacceptable adverse effects, this could help maximize the 1998; 68: 1180–1186. impact on weight of its demonstrated effect on food intake. 9 Hansen DL, Toubro S, Stock MJ, Macdonald IA, Astrup A. The A combination of stimulation of thermogenesis and suppres- effect of sibutramine on energy expenditure and appetite during sion of food intake may thus have the potential to produce chronic treatment without dietary restriction. Int J Obes 1999; 23: 1016–1024. better long-term weight loss than energy restriction alone. 10 Li ZP, Maglione M, Tu WL, Mojica W, Arterburn D, Shugarman LR B Even a slight thermogenic effect of 250 kJ per day, as et al. Meta-analysis: pharmacologic treatment of obesity. indicated in this study, would then allow body weight to Ann Intern Med 2005; 142: 532–546. become stabilized at a slightly higher and more acceptable 11 Astrup A, Meier DH, Mikkelsen BO, Villumsen JS, Larsen TM. Weight loss produced by tesofensine in patients with Parkinson’s level of energy intake. or Alzheimer’s disease. Obesity 2008; 16: 1363–1369. 12 Astrup A, Madsbad S, Breum L, Jensen TJ, Kroustrup JP, Larsen TM. Effect of tesofensine on bodyweight loss, body composition, Conclusion and quality of life in obese patients: a randomised, double-blind, placebo-controlled trial. Lancet 2008; 372: 1906–1913. 13 Hansen H, Hansen G, Tang-Christensen M, Larsen PJ, Axel AM, This study found a clear effect on appetite sensation and a Raben A et al. The novel triple monoamine reuptake inhibitor slightly reduced energy expenditure during the night after tesofensine induces sustained weight loss and improves glycemic

International Journal of Obesity Effect of tesofensine on energy metabolism and appetite ASjo¨din et al 1643 control in the diet-induced obese rat: comparison to sibutramine 24 Westerterp KR, Donkers JHHL, Fredrix EWHM, Boekhoudt P. and rimonabant. Eur J Pharmacol 2010; e-pub ahead of print 10 Energy-intake, physical-activity and body-weightFa simulation- April 2010. model. Br J Nutr 1995; 73: 337–347. 14 Astrup A, Thorbek G, Lind J, Isaksson B. Prediction of 24-H 25 Nelson DL, Gehlert DR. Central nervous system biogenic amine energy-expenditure and its components from physical character- targets for control of appetite and energy expenditure. Endocrine istics and body-composition in normal-weight humans. Am J Clin 2006; 29: 49–60. Nutr 1990; 52: 777–783. 26 Connoley IP, Liu YL, Frost I, Reckless IP, Heal DJ, Stock MJ. 15 Toubro S, Christensen NJ, Astrup A. Reproducibility of 24-h Thermogenic effects of sibutramine and its metabolites. Br J energy expenditure, substrate utilization and spontaneous phy- Pharmacol 1999; 126: 1487–1495. sical activity in obesity measured in a respiration chamber. Int J 27 Golozoubova V, Strauss F, Malmlof K. Locomotion is the major Obes Relat Metab Disord 1995; 19: 544–549. determinant of sibutramine-induced increase in energy expendi- 16 Elia M, Livesey G. Energy expenditure and fuel selection in ture. Pharmacol Biochem Behav 2006; 83: 517–527. biological systems: the theory and practice of calculations based 28 Axel AM, Mikkelsen JD, Hansen HH. Tesofensine, a novel on indirect calorimetry and tracer methods. World Rev Nutr Diet triple monoamine reuptake inhibitor, induces appetite 1992; 70: 68–131. suppression by indirect stimulation of alpha(1) adrenoceptor 17 Klausen B, Toubro S, Astrup A. Age and sex effects on energy and dopamine D(1) receptor pathways in the diet-induced obese expenditure. Am J Clin Nutr 1997; 65: 895–907. rat. Neuropsychopharmacology 2010; e-pub ahead of print 3 March 18 Flint A, Raben A, Blundell JE, Astrup A. Reproducibility, power 2010. and validity of visual analogue scares in assessment of appetite 29 Lehr T, Staab A, Tillmann C, Nielsen EO, Trommeshauser D, sensations in single test meal studies. Int J Obes 2000; 24: 38–48. Schaefer HG et al. Contribution of the active metabolite M1 to 19 Buffenstein R, Karklin A, Driver HS. Beneficial physiological and the pharmacological activity of tesofensine in vivo: a pharmaco- performance responses to a month of restricted energy intake in kinetic-pharmacodynamic modelling approach. Br J Pharmacol healthy overweight women. Physiol Behav 2000; 68: 439–444. 2008; 153: 164–174. 20 Toubro S, Christensen NJ, Astrup A. Reproducibility of 24-H 30 Gasteyger C, Sjo¨din A, Meier DH, Astrup A. Effects of tesofensine energy-expenditure, substrate utilization and spontaneous on appetite in obese subjects. Obesity 2008; 16: S138–S139 physical-activity in obesity measured in a respiration chamber. (abstract). Int J Obes 1995; 19: 544–549. 31 Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expendi- 21 Toubro S, Sorensen TI, Ronn B, Christensen NJ, Astrup A. Twenty- ture resulting from altered body weight. N Engl J Med 1995; 332: four-hour energy expenditure: the role of body composition, 621–628. thyroid status, sympathetic activity, and family membership. 32 Prentice AM, Goldberg GR, Jebb SA, Black AE, Murgatroyd PR. J Clin Endocrinol Metab 1996; 81: 2670–2674. Physiological-responses to slimming. Proc Nutr Soc 1991; 50: 22 Astrup A, Andersen T, Henriksen O, Christensen NJ, Bulow J, 441–458. Madsen J et al. Impaired glucose-induced thermogenesis in 33 Ravussin E, Lillioja S, Knowler WC, Christin L, Freymond D, skeletal muscle in obesity. The role of the sympathoadrenal Abbott WG et al. Reduced rate of energy expenditure system. Int J Obes 1987; 11: 51–66. as a risk factor for body-weight gain. N Engl J Med 1988; 318: 23 Astrup A, Buemann B, Christensen NJ, Madsen J. 24-hour energy 467–472. expenditure and sympathetic activity in postobese women 34 Astrup A, Gotzsche PC, van de Werken K, Ranneries C, Toubro S, consuming a high-carbohydrate diet. Am J Physiol 1992; 262: Raben A et al. Meta-analysis of resting metabolic rate in formerly E282–E288. obese subjects. Am J Clin Nutr 1999; 69: 1117–1122.

International Journal of Obesity