No preclinical experimental approach enables the study of voluntary oral consumption of high-concentration Δ⁹-tetrahydrocannabinol (THC) and its intoxicating effects, mainly owing to the aversive response of rodents to THC that limits intake. Here we developed a palatable THC formulation and an optimized access paradigm in mice to drive voluntary consumption. THC was formulated in chocolate gelatin (THC-E-gel). Adult male and female mice were allowed ad libitum access for 1 and 2 h. Cannabimimetic responses (hypolocomotion, analgesia, and hypothermia) were measured following access. Levels of THC and its metabolites were measured in blood and brain tissue. Acute acoustic startle responses were measured to investigate THC-induced psychotomimetic behavior. When allowed access for 2 h to THC-E-gel on the second day of a three-day exposure paradigm, adult mice consumed up to ≈30 mg/kg over 2 h which resulted in robust cannabimimetic behavioral responses (hypolocomotion, analgesia and hypothermia). Consumption of the same gelatin decreased on the following 3^rd day of exposure. Pharmacokinetic analysis show that THC-E-gel consumption led to parallel accumulation of THC and its psychoactive metabolite, 11-OH-THC, in brain, a profile that contrasts with the known rapid decline in brain 11-OH-THC levels following THC intraperitoneal (i.p.) injections. THC-E-gel consumption increased the acoustic startle response in males but not in females, demonstrating a sex-dependent effect of consumption. Thus, while voluntary consumption of THC-E-gel triggered equivalent cannabimimetic responses in male and female mice, it potentiated acoustic startle responses preferentially in males. We build a dose-prediction model that included cannabimimetic behavioral responses elicited by i.p. versus THC-E-gel to test the accuracy and generalizability of this experimental approach and found that it closely predicted the measured acoustic startle results in males and females. In summary, THC-E-gel offers a robust preclinical experimental approach to study cannabimimetic responses triggered by voluntary consumption in mice, including sex-dependent psychotomimetic responses.

Animal Studies

Animal studies followed the guidelines established by AAALAC and were approved by IACUC of the University of Washington. Male and female C57BL/6J mice ranging from 8-14 weeks of age were used. Animals were housed with sibling littermates and were provided with standard chow and water, ad libitum, and without any additional environmental enrichment. Investigators were not blinded to experimental exposure conditions throughout assays due to the noticeable behavioral effects measured in response to THC. Animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Washington and conform to the guidelines on the care and use of animals by the National Institutes of Health.

Pharmacological Agents

Animals received THC (0.1, 0.3, 1, 3, 5, 10, and 30 mg/kg) and SR141716 (SR1, 1 mg/kg) i.p. or were exposed to THC suspended in gelatin. THC and SR141716 (SR1) were provided by the National Institute of Drugs abuse Drug Supply Program (Bethesda, MD). THC in ethanol (50 mg/ml) was either added to gelatin mixtures (CTR or Ensure®) or prepared for i.p. injection. For i.p. injection, both THC (0.1, 0.3, 1, 3, 5, 10, and 30 mg/kg) and SR1 (1mg/kg) were dissolved in 95% ethanol and then vortexed thoroughly with equal volume Cremophor and finally dissolved in sterile saline to reach a final 1:1:18 solution consisting of ethanol:cremophor:saline.

Gelatin formulation

Control gelatin (CTR-gel): Deionized water (100 mL) was warmed to 40°C and stirred at a constant rate. 2.5 g of Polycal™ sugar and 3.85 g of Knox™ Gelatin were added and the mixture was maintained at a temperature below 43°C. The mixture was removed from heat, and THC (50 mg/ml in ethanol) was added to reach a concentration of 0.3, 1, 2, or 4 mg/15 ml. An equal volume of ethanol was added to vehicle gelatin (<1% total volume). Gelatin was poured into plastic cups ranging from 2-10 ml and set into a 4°C fridge to solidify overnight.

Ensure gelatin (E-gel): Chocolate-flavored Ensure™ (100 ml) was warmed to 40°C and stirred at a constant rate. 3.85 g of Knox™ Gelatin were added and the mixture was maintained at a temperature below 43°C. The mixture was removed from heat and THC (50 mg/15 ml ethanol) was added to reach a concentration of 1, 2, 5, or 10 mg/15 ml. At the 10 mg/15 ml concentration, ethanol was evaporated off to 50% volume before being added to the mixture to reduce total alcohol concentration below 1%. An equal volume of ethanol was added to vehicle gelatin (<1% total volume). Gelatin was poured into plastic cups ranging from 2-10 ml and set into a 4°C fridge to solidify overnight. Mice were always exposed to more gelatin than they could consume, smaller volumes were used to conserve THC.

Acute gelatin access: Animals were first habituated to gelatin by receiving an excess of gelatin in their home cage the day before the first timed access. On the first day of access, mice were placed into a home cage-like environment equipped with a vehicle gelatin cup that was stabilized to the cage. Behavior was recorded during the consumption window via an overhead camera. On the second day of access, animals were placed into the same gelatin access cage with either a vehicle or THC gelatin cup. Animals experienced either a triad of behaviors (open field, tail flick, and body temperature) measured immediately preceding and following the consumption window or an acoustic startle trial immediately following consumption. On the third and final day of access animals were placed into the same cage with a vehicle gelatin cup. For all gelatin access days, gelatin cups and animals were weighed before and after the consumption window. Access to gelatin during the consumption window was limited to either 1 or 2 h after which the animals were removed and returned to their home cage.

Triad of Cannabimimetic behaviors

Hypolocomotion, hypothermia, and analgesia were measured 1 h post-i.p. injection or immediately following gelatin exposure. Pre-tests were collected immediately prior to injection or gelatin exposure.

Open Field: A 50 cm x 50 cm chamber (25 LUX) was equipped with an overhead camera to record movement. Animals were placed in the chamber for 15 minutes and then returned to their home cage. Total distance traveled (cm) was measured using Noldus Ethovision behavioral tracking software. Locomotion behavior was measured immediately before and after gelatin access to calculate a gelatin dependent difference score (post-pre). Pre-test measurements for CTR-gel were not collected and pre-test values were instead normalized to vehicle post-tests to produce a difference score as Post-VEH.

Tail Flick Analgesia: A hot water bath was set to 52.5°C. Mice were securely held upright in the air with their tail hanging downward. A timer was started as 75% of their tail was submerged into the water. Time was measured once a painful response was presented, marked as a latency to flick their tail out of the hot water. Tail flick responses were measured immediately before and after gelatin access to calculate a gelatin-dependent difference score (post-pre).

Measuring body temperature: Animals were placed on a stable surface with their tails lifted. A rectal thermometer probe (RET-3 Kent Scientific) was inserted into the anus for 10-20 s until the temperature recording stabilized. This test was always performed prior to the Tail Flick test to reduce any potential temperature contamination effects. Body temperatures were measured immediately before and after gelatin access to calculate a gelatin dependent difference score (post-pre).

Blood and brain tissue collection and quantification

Animals underwent the same gelatin access paradigm for day 1 and 2 described in acute gelatin access. After 2 h of gelatin access, blood was collected by cardiac puncture with a 23-gauge needle and placed on ice. Immediately following, brain tissue was collected and flash frozen in liquid nitrogen. Blood samples were spun in a 4°C centrifuge at 1450 x g for 15 min. Plasma was transferred to another tube and stored alongside brain samples in -80°C until being shipped on dry ice to the Piomelli Lab at UCI for sample analysis. Samples were collected immediately following 1 and 2 h gelatin access and 30 min and 24 h after 2 h gelatin access. THC and its first-pass metabolites 11-OH-THC and 11-COOH-THC were quantified in plasma and brain tissue using a selective isotope-dilution liquid chromatography/tandem mass spectrometry assay (26). Concentration data after E-gel consumption were compared to blood and brain tissue concentrations after i.p. administration from previous publications from the Piomelli Lab [35]. 

Acoustic startle

Acoustic Startle behaviors were measured after 1 or 2 h THC-E-gel exposure (10 mg/15 ml) and THC-i.p. (0.1, 1, 5, 10 mg/kg) injection. Sound-buffered startle chambers (SR-Lab, San Diego Instruments) were used to measure acoustic startle responses, equipped with a holding tube and an accelerometer. Background sound was maintained at 65 dB from a high-frequency speaker producing white noise. Startle tests were conducted 1 h post-THC-i.p. injection or immediately following THC-E-gel exposure. Animals were set in the holding tube for 5 min to habituate prior to a series of seven trials presenting escalating sound levels of null, 80, 90, 100, 105, 110, and 120 dB. Tones were presented for 40 ms with an inter-trial interval of 30 s. Animals were only ever exposed to the acoustic startle paradigm once, immediately after gelatin access, to avoid auditory habituation.

Data/Statistical analysis

All data were analyzed using GraphPad Prism 10-11. For all statistical analyses (unpaired t test, one- and two-way ANOVA, and post hoc analyses), alpha level was set to 0.05. For all ANOVA analyses, Sidak’s post-test was performed for increased power and repeated measures analysis was performed for time-dependent consumption data. All behavioral locomotor tracking was analyzed using Noldus Ethovision software and statistical analyses performed through GraphPad Prism 10-11. Nonlinear regressions (Figure 4B-d and Figure 1-figure supplemental 3B-G) were performed with a 3-parameter nonlinear, least squares, regression.