Table 1.
The dose ratio of the co-administration of tesofensine + 5-HTP (mg/kg).
Fig 1.
Tesofensine induces greater weight loss in HFD-fed rats compared to chow-fed rats.
A. Change in body weight across treatment days with the treatment groups receiving subcutaneous injections of tesofensine (2 mg/kg) and the control groups receiving saline. B. Total grams of food consumed per group (left panel) and total caloric intake (right panel). Both groups treated with tesofensine consumed fewer calories than saline control groups. C. Visceral fat content (%) relative to body weight in gonadal, perirenal, and mesenteric deposits. Data are presented as mean ± SEM. n = number of rats. Filled and open data points represent male rats (B-C). Repeated Measures ANOVA (A), One-way ANOVA (B-C). * p <0.05 compared with Chow-Saline, # p < 0.5 compared with Chow-Tesofensine and + p < 0.05 significantly different between groups.
Fig 2.
Tesofensine differentially modulates lateral hypothalamic neurons in lean and obese rats.
A. Left: A schematic representation for extracellular recording in a freely moving rat and pump for automatic drug delivery. No food was available during recordings. Right: histology, identifying recording sites in LH using a coronal section. Dil, a fluorescent and lipophilic red dye, is used to mark the tip of the electrode track. B. t-SNE and hierarchical cluster analysis of firing rates were used to group neurons with similar activity patterns into neuronal ensembles. This analysis uncovered four ensembles responding to tesofensine. Neurons were assigned into four ensembles (E1-4) for Chow-fed rats (left) and four ensembles for HFD-fed rats (right). Each dot represents a single neuron and the color of the ensemble to which it belongs in the t-SNE map. C. All neurons recorded were categorized into four ensembles (E1-E4) for Chow-fed rats (left) and HFD-fed rats (right). The normalized color-coded activity of each neuron over time is presented for chow-fed and HFD-fed rats, with red and blue indicating higher and lower z-score activity, respectively. Black vertical lines show the baseline (BL from 0–15 minutes), saline (Sal 15–30 minutes), tesofensine administered at 30 minutes and recordings that lasted up to 120 minutes. In three HFD rats, we recorded n = 127, 95, and 139 neurons, whereas in the other three Chow-fed rats, n = 220, 106, and 17 neurons. The color bar on the right indicates the rat in which the neuron was recorded. Peri-Stimulus Time Histograms (PSTHs) below demonstrate the average neuronal ensemble activity, with dashed lines dividing each cluster (ensemble). D. Pie charts depict the percentage of neurons in each ensemble for Chow-fed and HFD-fed rats. Chi-square analysis showed a significant difference compared to the same ensemble in Chow-fed rats (* p<0.05). E. The z-score normalized population activity of all lateral hypothalamic neurons recorded in HFD-fed (n = 361) and chow-fed rats (n = 343) is presented. The shadow represents the SEM.
Fig 3.
Tesofensine silences mice LH GABAergic neurons and attenuates their optogenetic activation.
A. Extracellular neuronal activity was recorded in two mice over a baseline period of 5 minutes before they received subcutaneous injections of saline or tesofensine (2 mg/kg). After 30 minutes, a laser was activated on an open loop protocol. It started with 5-minute blocks of inactive periods (no laser) and active (50 Hz; 2s on + 4s off). Employing a within-subjects design, the same two mice received injections in this order: tesofensine (day 1), saline (day 2), tesofensine (day 3), and saline (day 4). We combined the activity of neurons recorded on both days of tesofensine for further analysis, and the same was done for saline days. This design allowed each mouse to serve as its own control. While we cannot guarantee recordings were done in the same neurons across days due to technical limitations, however, we used fixed arrays to maintain the electrode position throughout the four recording days. B. Upper depicts responses of single GABAergic neurons inhibited after tesofensine administration. Bottom a non-GABAergic neuron. C. A heat map displaying neuronal activity during saline (left) and tesofensine (right) conditions. Blue horizontal lines indicate stimulation blocks. Below, the corresponding peri-stimulus time histogram (PSTH) shows the average activity of neurons during the stimulation blocks, with the dashed line representing a zero z-score. * GABAergic neuron activity was significantly reduced by tesofensine compared to saline treatment [RM ANOVA a significant treatment effect; F(1,35) = 6.012 p = 0.0193, and significant effect of block (3 stimulation blocks), F(1,2) = 10.6 p = <0.001, and no significant interaction F(2,70) = 0.02, p = 0.977]. Thus, tesofensine reduced neuronal activity even during optogenetic stimulation, highlighting its ability to silence GABAergic neurons in the LH. In contrast, non-GABAergic neuron activity showed no significant difference between saline and tesofensine groups (p > 0.05), as assessed by an RM ANOVA [treatment main effect, F(1,28) = 0.25, p = 0.8754].
Fig 4.
Tesofensine reduces the feeding behavior induced by the optogenetic activation of LH GABAergic neurons in lean Vgat-ChR2 mice.
A. Schematic of the open-loop task, where VGAT-ChR2 mice (n = 4) with a fiber implanted in their LH, were stimulated at 50Hz for 2 seconds with 4 seconds off in blocks of 5 minutes (blue). Mice could lick a sucrose solution (10%) from a sipper. B. Cumulative licks during the open-loop stimulation with tesofensine (Teso) 2mg/kg or saline solution (Sal) administered by subcutaneous injection 30 minutes before the task. The tesofensine group (gray) was placed in the task, but the laser was not stimulated. Note that the teso+laser group exhibited an attenuated feeding response elicited by stimulation of LH GABAergic neurons. C. Total licks during a session, with blue bars representing the blocks that were laser stimulated in the task. D. Stimulus-bound feeding is the fraction of licks given during the stimulation window (spanning from the first laser pulse to 2.5 seconds after), with each circle representing a different mouse. Data are presented as mean ± SEM, and the results were statistically significant (*p-value < 0.05 based on paired t-test value).
Fig 5.
Tesofensine does not attenuate the optogenetic feeding elicited by LH GABAergic neurons in obese Vgat-IRES-cre mice.
A. Representative image of Vgat-IRES-Cre mouse expressing ChR2 neurons (green) in the LH. B. Behavioral protocol: Vgat-mice were fed an HFD or standard Chow diet for 12 weeks and were optostimulated in an open loop schedule. The mice had access to a sucrose solution (10%) in a sipper during the task. Mice had food and water ad libitum before the task. C-E. These six panels depict the cumulative licks (intake) during the open-loop task for lean versus obese mice administered with saline (C), tesofensine (Teso) 2 mg/kg (D), or 6 mg/kg (E) 30 minutes before the task. The left panels are control mice expressing EYFP, and the right panels express ChR2. Data are presented as mean ± SEM. The results show that compared to saline, tesofensine 2 mg/Kg reduced sucrose intake in lean mice expressing ChR2 (D right panel, see black line). In contrast, activating GABA neurons in obese mice significantly induced more sucrose intake than in lean mice.* p <0.05, Two-sample Kolmogorov-Smirnov test Lean vs. Obese. # p < 0.05 compared to its saline control group.
Fig 6.
Chemogenetic silencing of LH GABAergic neurons potentiates tesofensine’s anorexigenic effects.
A. Representative image of LH in mouse expressing hM4D(Gi), which inhibits GABAergic neurons chemogenetically. B. Schematic of a task where mice were placed in their home cages with an automated feeder (FED3) that delivered chocolate pellets for each nose poke. Recording started at 18:00 and lasted for 24 hours. At 18:30 h, the same five mice were injected with vehicle (Veh), Clozapine-N-Oxide (CNO), a ligand for hM4D(Gi), tesofensine (2 mg/kg), and CNO + tesofensine. The dark cycle lasted from 19:00 to 7:00 h. The moon symbol indicates when the ambient lights were turned off and the sun when the lights were turned on. C. Average cumulative nose pokes during the 24-hour recordings for each group. The syringe indicates the time of drug administration at 18:30. The teso+CNO group performed fewer cumulative nose pokes and thus obtained fewer pellets than the other groups. * Significantly different two-sample Kolmogorov-Smirnov test (for the first 19 hours after drug administration, see horizontal line); Veh vs. CNO p = 0.006, CNO vs. Teso; p = 0.01, and Teso vs. Teso+CNO p = 0.04.
Fig 7.
Effects of tesofensine and phentermine on locomotion, quiet-awake/sleep, and head weaving stereotypy in rats.
A. Total distance traveled (cm/240 min) during forward locomotion over seven days of treatment with saline (1 ml/kg, n = 6), phentermine (20 mg/kg, n = 6), tesofensine (2 mg/kg, n = 6), or tesofensine (6 mg/kg, n = 6). Each point depicts one rat, with the black line indicating the mean and the gray shaded area indicating the standard error of the mean distance traveled by all rats in the group. All treatments reduced locomotion B. The proportion of time that rats spend in a quiet-awake/sleep state is defined as when the rat is not moving but either awake or sleeping. Only the Vehicle and Tesofensine 2 mg/kg groups spent more than 60% of their time in these behavioral states. C. The distribution of the percentage of time that rats exhibited a head weaving stereotypic behavior over seven days following treatment injection. Each point in the plot represents one rat, and the width of the violin plot indicates the probability density distribution. The black line shows the median value for each day. For phentermine, the results showed that the percentage of time the rat exhibited head weaving stereotypy gradually increased across the seven days. Additionally, head weaving stereotypy was aggravated across days in all rats treated with phentermine. In Vehicle control rats, grooming behavior was mistakenly classified as head weaving stereotypy. This is because control rats do not express this behavior. Surprisingly, rats treated with tesofensine 2 mg/kg exhibited little stereotypy, thus apparently neither grooming. D. The onset of the first event of stereotypy was measured across days. Note that for phentermine, the onset of stereotypy decreases across days. E. For the first two days, the variables locomotion, quiet awake/sleep, onset, and stereotypy were analyzed using a clustering algorithm (t-SNE). This analysis uncovered two main clusters: the first corresponds to rats treated with vehicle and tesofensine 2 mg/kg. Note that rats treated with tesofensine 2 mg/kg were in a slightly different position than control rats. The second cluster mixed rats treated with phentermine and tesofensine 6 mg/kg. Hence, t-SNE seems to separate rats according to the overall motor profile effects induced by each drug. At therapeutic doses, tesofensine induced body weight loss without producing head weaving stereotypy.
Fig 8.
Tesofensine prolongs body weight loss produced by serotonin precursor 5-HTP/CB on the chow control diet in rats.
A. Change in body weight (g) relative to pretreatment day 0. Tesofensine1 and tesofensine2 refer to doses of 1 and 2 mg/kg administered subcutaneously, respectively. 5-HTP/CB refers to doses of 31 and 75 mg/kg administered intraperitoneally. Note that the triple combination of tesofensine, 5-HTP, and CB led to significantly greater weight loss than any other group and did not show weight loss tolerance as seen in 5-HTP/CB group (see horizontal line at days 7–15). B. Changes in food intake for the groups shown in panel “A.” Note that 5-HTP/CB, tesofensine1, and 2 + 5-HTP/CB groups exhibited anorexigenic tolerance, as evidenced by their increased food intake from day 7 to 15. Data are presented as mean ± SEM. n = number of rats. Repeated Measures ANOVA (A-B). $ p < 0.05 tesofensine1 compared with tesofensine1 + 5-HTP/CB, and 5-HTP/CB groups. % p < 0.05 tesofensine2 compared to tesofensine2 + 5-HTP/CB, and 5-HTP/CB groups. * p < 0.05 compared with the control group.
Fig 9.
Isobolographic analysis of the interaction between tesofensine and 5-HTP/CB in a sucrose intake assay in rats.
The anorexigenic effect was measured as a reduction in sucrose intake (see text for details). Panels A-B depict the dose-dependent effect of tesofensine and 5-HTP/CB, respectively. Data are plotted as mean + SEM, open circles represent individual animals, and * corresponds to a statistically significant difference (p<0.05) with respect to the control group, as determined by one-way ANOVA. C. Depicts the dose-dependent effect of the 1:1 combination of tesofensine and 5-HTP/CB. Data are presented as for panels A and B. D. shows the isobolographic analysis of the 1:1 combination. The oblique line is the isobole corresponding to a purely additive interaction, and black circles represent the ED30 (+ SEM) values of the individual components and the theoretical ED30 value for a purely additive interaction. The red triangle indicates the experimental ED30 of the combination. A comparison of the theoretical and experimental ED30 was performed by the modified Student’s t-test, n.s. indicating a lack of statistically significant difference (p>0.05). E. depicts the dose-dependent effect of 3:1 combination of tesofensine and 5-HTP/CB. Data are presented as for panels A, B, and C. F. shows the isobolographic analysis of the 3:1 combination. Data are presented as in panel D.
Table 2.
Statistical analysis of tesofensine plus 5-HTP interaction.
Fig 10.
The appetite suppressant effects of tesofensine and 5-HTP/CB is not due to taste aversion.
A. The graph plots the average 1-hour sucrose intake during baseline sessions. On day 7 (D7), drugs were administered before giving access to sucrose. A dose-dependent intake suppression can be seen. Finally, on day 8 (D8), intake was observed the day after treatment. 5-HTP/CB suppressed acute sucrose intake on day 7 (D7), but consumption returned to baseline levels on day 8 (D8). Bar plots show acute sucrose intake (1 hour) for control rats (vehicle injection) and treated rats (different doses of 5-HTP/CB). Each dose was tested once in a new set of naïve rats (n = 6 rats per group). Note that 5-HTP/CB induced a dose-dependent decrease in sucrose intake. BL = baseline sucrose intake. B. Same convention as panel “A” but for tesofensine. n = number of rats, open data points represent each rat. One-way ANOVA (A-B).* p<0.05 significantly different from BL.
Fig 11.
Tesofensine did not affect performance of rats in a sucrose detection task.
A. Rats were trained to lick a central spout that dispensed the stimulus a drop of water or solutions of sucrose. To obtain a reward (3 drops of water), rats had to choose between two lateral spouts. B. Upper panel shows the number of trials, and the lower panel the correct performance across the baseline, tesofensine treatment, and post-tesofensine days. There were no significant differences in the percent correct, the trials per session, or the total volume consumed between these periods, except for an overall decrease in the number of trials during the baseline period as the rat re-learned the task. The gray rectangle depicts days of tesofensine administration C. Plots of high sucrose responses as a function of sucrose concentration. The psychometric curves for the sucrose detection task also did not differ significantly between the baseline, tesofensine, and post-tesofensine periods. These findings suggest that tesofensine does not affect performance in the sucrose detection task in rats. The x-axis is scaled logarithmically, and the data is represented as mean ± SEM.