Cannabis Sativa targets mediobasal hypothalamic neurons to stimulate appetite

Cannabis plant

In accordance with federal law, the Cannabis sativa plant matter used in this study was obtained under federal guidelines with a Drug Enforcement Agency Schedule I drug license from the National Institute on Drug Abuse (NIDA) via Research Triangle. The ground whole plant Cannabis sativa used contained 7.8% THC and 0.5% CBD.

General procedure

Experiment 1: cannabis-induced feeding pattern analysis in free feeding rats

To measure the effects of cannabis on meal patterns, one cohort of rats were habituated to BioDAQ chambers (n = 8/group) and a second cohort of rats were habituated to the Sable Promethion metabolic (n = 6 experimental and n = 4 control) chambers for 10 days prior to any experimental manipulation. During this process, rats were introduced to the vapor chambers without any cannabis exposure to eliminate the effects of stress due to handling. On test days, rats received an 800 mg dose of vaporized cannabis plant matter and were immediately returned to the BioDAQ chambers or Sable system chamber. Subsequently, feeding measurements were taken continuously for 4 h to examine temporal changes in meal patterns. Meals were defined as 0.2 g or larger separated by 10 min.

Experiment 2: cannabis effects on locomotor activity

An open-field arena was used to quantify locomotor activity following cannabis exposure. The field measured 45 cm by 45 cm and was constructed with a white floor and transparent plexiglass walls around the perimeter, leaving the top of the field exposed to the room. A camera was stationed directly above the field to capture locomotor activity. Rats were habituated to the field environment for 10 min on three separate days. Subsequently, rats received 800 mg of vaporized cannabis and were placed in the open-field arena 30 min later (n = 4/group). Locomotor activity was evaluated by measuring total movement and total distance traveled over the 30 min period with Ethovision XT 11.5 software. Separately, the locomotor activity of a second cohort of rats living in Sable Promethion metabolic chambers (n = 6 experimental and n = 4 control) was monitored following 800 mg of vaporized cannabis. Locomotor activity was evaluated by measuring total movement and total distance traveled over a period of 24 h and later segmented to the second 30 min following vapor cannabis exposure. Metabolic measurements of Energy Expenditure (EE), Respiratory Exchange Rates (RER), rate of oxygen consumption (VO₂), and rate of carbon dioxide emission (VCO₂) were also gathered from rats in the Sable Promethion chambers. RER was calculated using the Weir Equation [Energy Expenditure (kcal/hr) = 3.941(VO₂) + 1.106(VCO₂)]²¹. Measurements were grouped by rat and combined into three minute bins for the 30 min to 1 h post-exposure timeframe for analysis.

Experiment 3: operant responding following cannabis exposure

The effects of cannabis on food-reinforced behavior were examined using a rodent touch screen operant system. In this experiment, rats were trained to respond for 45 mg sucrose pellets in a Bussey Touch Screen System (89547-R-Lafayette Instrument Company, Lafayette, IN), as described previously¹⁸. Rats maintained on ad libitum chow were trained using this protocol until they reached a sufficient response (15 correct touches) over a 30-min period (5 days). Following acquisition of food-reinforced behavior, one cohort (n = 6/group) of rats was placed into the chambers 1 h after air or cannabis vapor exposure (800 mg). To determine if the effects on food-seeking were time-dependent, on a second day, rats was placed into the chambers 2 h after air or cannabis exposure. During each experiment, food was removed from the home cages during drug assimilation (1 h or 2-h period) and subsequently, correct touches, blank touches, correct touch latency, and blank touch latency were recorded over a 30-min period for each condition.

Mouse Studies

Vapor cannabis regulates general locomotor activity in a food specific fashion

Tetrahydrocannabinol (THC), the primary psychoactive constituent of the cannabis plant, produces sedative effects in patients and rodents²³. As a result, THC-based therapies are often not well tolerated in the clinical population^7,8,24,25. Thus, to determine if the observed appetite stimulation induced by cannabis vapor was accompanied by potentially adverse sedative effects, we conducted an open field assay and collected data from rats housed in metabolic chambers to examine potential changes in locomotor activity (Fig. 2A–C). Rats exposed to cannabis vapor and placed in the open field arena without access to food did not display differences in total distance traveled (Fig. 2A; right), cumulative movement (Fig. 2B; right), or time spent still (Fig. 2C; right) relative to air-treated control rats. In contrast, when rats were exposed to cannabis in the presence of food in metabolic chambers, they traveled a greater total distance (Fig. 2A; left), displayed increased cumulative movement (Fig. 2B; left), spent less time still (Fig. 2C; left) and more time around the food hopper (Fig. 2E–H) relative to air-treated control rats. Notably, when compared to rats with no food present, both distance traveled and time spent still were significantly decreased in rats with food present, while time spent moving was significantly increased, indicating that in the presence of food, cannabis treated rats spent more time around the food hopper as opposed to other areas of the environment (Fig. 2A–C). Importantly, rats continued to display increased total distance traveled for up to 4 h after exposure (Fig. 2D). Together these data suggest that appetite stimulation with this dose (800 mg) of cannabis vapor does not induce noticeable sedative effects and instead, motivates behavior towards food.

Cannabis augments metabolic activity

Previous studies suggest that inhaled cannabis leads to increases in resting metabolic and ventilation rate²⁶. In the present study, rats exposed to an acute dose of vaporized cannabis displayed no significant difference in respiratory quotient (RQ), the ratio of carbon dioxide emission and oxygen consumption, compared to air controls (Fig. 2K). However, the rate of oxygen consumption was significantly increased in the cannabis exposed rats (Fig. 2L) while the rate of carbon dioxide emission did not reach statistical significance (Fig. 2M). Importantly, this ratio is often used as an indicator of the macronutrient utilization, a ratio of 0.7 indicates fat metabolism, whereas 1.0 indicates carbohydrates, and 0.8 suggests a mixed diet²¹. Both cannabis and air exposed rats had average ratios around 0.88 suggesting they are metabolizing both carbohydrates and fats (Fig. 2K). Interestingly, and in support of the current theory that cannabis use increases metabolism²⁶, our data indicate that cannabis exposed rats had a significantly higher rate of energy expenditure (Fig. 2J), an observation that occurred during the same timeframe when rats were more active (Fig. 2I).

Cannabis augments food-motivated behavior

To determine if cannabis sativa enhances the motivational properties of food, we evaluated operant performance for sucrose in rats exposed to cannabis vapor (Fig. 3A). Vapor cannabis exposure caused a significant increase in operant responding for sucrose 1 h after exposure (Fig. 3B; left) without affecting the total number of blank touches, or the latencies to perform correct or blank touches (Fig. 3B, C). In contrast, no differences in operant responding for sucrose were detected when rats were tested 2 h following vapor cannabis exposure (Fig. 3B, C). However, cannabis exposed rats did have significantly less correct and blank touches 2 h after exposure relative to 1 h after, indicating that cannabis augments food-motivated behavior in a time-dependent fashion.

Vapor cannabis stimulates food intake in mice

To conduct in vivo imaging experiments, we first needed to establish a dose response for cannabis-induced feeding behavior in mice (Fig. 4A). Our data indicate that food intake in C57BL6J mice exposed to the 100 mg dose did not differ significantly from air. However, mice displayed a significant increase in food intake when exposed to the 200 mg cannabis dose whereas a significant decrease was detected at the 400 mg dose (Fig. 4B). It is important to note here that this dose (200 mg), was 4× less than that required for cannabis-induced appetite stimulation in rats and followed a similar temporal pattern as previously published in this model¹⁸. Subsequently we used this dose in all calcium imaging experiments and chemogenetic manipulations.

In vivo Ca²⁺ imaging in mixed populations of MBH neurons

The MBH, regulates appetite and contains neurons that express the cannabinoid-1 receptor (CB1R; 17). Thus, we reasoned neurons in this region may participate in cannabis-induced appetite. To understand how MBH neuronal activity was modified by cannabis, we applied in vivo microendoscopic imaging to visualize somatic intracellular calcium [Ca²⁺] dynamics in mixed neuronal ensembles in freely behaving mice. Specifically, we expressed the calcium sensor GCamp6s in the MBH, specifically the ARC, of male C57BL/6J mice and then utilized microendoscopes to visualize changes in [Ca²⁺] in neurons when mice approached or consumed palatable food (Fig. 5A). Mice exposed to this paradigm rapidly learned to approach the food containing zone (FZ), where a palatable high fat diet (HFD) was consistently offered (Fig. 5B). Using this approach, our first experiment involved exposing mice to air only in the vapor chambers (control) then visualizing GCamp6s fluorescence, which reflects changes in intracellular Ca²⁺ dynamics (Fig. 5C, G, H). Cannabis-exposed mice displayed increased [Ca²⁺] in numerous MBH neurons that became activated when mice entered the FZ (Fig. 5I). In agreement with previous reports^27,28, a portion of FZ activated neurons remained active when HFD was consumed whereas others were inhibited. In addition, a separate population of neurons displayed increased [Ca²⁺] only when HFD was consumed (Fig. 5J).

Vapor cannabis stimulates food anticipatory responses in MBH neurons

We next exposed the same mice to the behaviorally characterized dose of cannabis shown to increase feeding in mice (200 mg, Fig. 4B) prior to placement in the test arena. Importantly, the number of active MBH neurons remained nearly constant during both the anticipatory and consummatory timeframes following cannabis exposure (Fig. 5I). Under this condition, we observed a significant increase in the number of MBH neurons that displayed [Ca²⁺] when mice entered the FZ relative to air FZ active neurons (Fig. 5F, J). A similar effect was observed once HFD was delivered (Fig. 5F, J). Remarkably, we observed increased [Ca²⁺] in a novel population of FZ and HFD stimulated MBH neurons in mice exposed to cannabis (Fig. 5D, E, J, supplemental video file ^S1). Overall, these observations indicate that cannabis exposure augments activity of MBH neurons during functionally diverse aspects of feeding behavior.

CB1R activation disinhibits AgRP neurons

Within the MBH, AgRP neurons are instrumental in promoting hunger and motivation to obtain food¹⁹. In addition, recent reports indicate that AgRP neurons receive presumed GABAergic afferent inputs that express CB1Rs²⁰. With this in mind, we employed an ex vivo slice electrophysiology approach to determine if activation of CB1Rs affected GABAergic synaptic activity onto AgRP neurons. To accomplish this, we injected Agrp^tm1(cre)Low1/J mice with an AAV expressing a GFP reporter to allow visualization and selection of AgRP neurons for patch-clamp recording (Fig. 6A, B). Recordings were performed in the presence of a glutamate receptor antagonist (Kynurenate, 1 mM) to isolate GABA_A receptor (GABA_AR)-mediated synaptic currents. Voltage clamped (Vh = − 60 mV, with E_Cl- =  ~ 0 mV, making GABAergic currents inward) AgRP neurons exhibited robust spontaneous inhibitory post synaptic currents (sIPSCs), mediated by GABA_ARs, as evidenced by their elimination following treatment with the GABA_AR antagonist GABAzine (10 mM; Fig. 6C). Bath application of the selective CB1R agonist, WIN 55, 212-2 significantly reduced the frequency of sIPSCs onto AgRP neurons (Figs. 6C, D) without altering peak amplitude (Fig. 6E). These data confirm that functional CB1Rs are indeed expressed on GABAergic synapses onto MBH neurons, and that their activation suppresses vesicular release of GABA, resulting in disinhibition of AgRP neurons.

We would like to thank Mr. Dave Savage for his invaluable skills in constructing the cannabis vapor chambers and Mr. Than Hutson for his technical assistance.