Supersensitivity to anandamide and enhanced endogenous cannabinoid signaling in mice lacking fatty acid amide hydrolase

The FAAH gene was isolated from a 129SvJ genomic library, and a 2.5-kb region encompassing the first exon was mapped and sequenced. A PGK-Neo cassette (consisting of a phosphoglycerate kinase promoter driving the neomycin phosphotransferase gene) was inserted between EcoRI and EcoRV sites located 2.3 kb apart, replacing the first FAAH exon (encoding amino acids 1–65) and ≈2 kb of upstream sequence. Homologous recombinant 129SvJ embryonic stem cell clones were identified by Southern analysis, and two such clones were used to generate chimeric mice on a C57BL/6 background. Chimeras from both clones gave germline transmission of the mutated gene. All mice used in this study were second or third generation offspring from intercrosses of 129SvJ-C57BL/6 FAAH^+/− mice. Offspring from both clones were tested and provided indistinguishable results.

FAAH activity assays for oleamide and anandamide were measured by following the conversion of ¹⁴C-labeled substrates using a TLC assay as described (27), with the exception that enzyme assays were conducted at pH 7.2. FAAH Western blots were conducted by using polyclonal anti-FAAH antibodies as described (26). For immunofluorescence images, mice were killed under deep anesthesia and brain samples were harvested for cryopreservation. The tissue was embedded in O.C.T. compound (Sakura/TissueTek, Torrance, CA) and frozen on dry ice. Cryosections were cut in 5-μm steps, mounted on glass slides, and stored at −80°C until needed. Slides were fixed at 37°C in 4% paraformaldehyde, hydrated in PBS, and blocked with 1% BSA in PBS. Blocked slides were washed with PBS and then incubated with anti-FAAH polyclonal antibodies (26, 34). Stained slides were then washed in PBS and incubated with secondary antibody directed against rabbit IgG and conjugated to Alexa 488 (Molecular Probes). Specimens were washed once more in PBS, counterstained with propidium iodide, and mounted in SloFade preservative (Molecular Probes). Stained slides were then observed on a Zeiss Axiovert STV100 microscope with the Bio-Rad MRC100 confocal system. Imaging of both FAAH^+/+ and FAAH^−/− specimens was performed with identical laser power and signal amplification settings.

Locomotor activity was assessed by placing each mouse in a clear Plexiglas cage [18 × 10 × 8.5 inches (l × w × h)] that was marked in 7-cm-square grids on the floor of the cage. The number of grids that were traversed by the hind paws was counted from 15–20 min postinjection. Nociception was then assessed in the tail immersion assay, where each mouse was hand-held with ≈1 cm of the tip of the tail immersed into a water bath maintained at 56.0°C and the latency for the animal to withdraw its tail was scored. The cutoff was 15 s and the data are expressed as the percent maximum possible effect (%MPE), where %MPE = 100 ⋅ (postinjection latency − preinjection latency)/(15 − preinjection latency). Baseline thermal nociception was also analyzed, using the hot plate test, where the latency to jump or lick/shake a hind paw was scored. In the formalin assay, 20 μl of a 2.5% formalin solution were injected s.c. under the dorsal surface of the right hindpaw. The total licking time was recorded from 0–40 min after injection. The early phase of pain responses lasted from 0–5 min and the late phase lasted from 20–40 min. Catalepsy was evaluated by using the bar test, in which the front paws of each subject were placed on a rod (0.75 cm diameter) that was elevated 4.5 cm above the surface. Mice that remained motionless with their paws on the bar for 10 s (with the exception of respiratory movements) were scored as cataleptic. Rectal temperature was determined by inserting a thermocouple probe 1.2 cm into the rectum and temperature was obtained from a telethermometer. The preinjection rectal temperatures for FAAH^+/+ (35.6 ± 0.2°C, n = 28) and FAAH^−/− (35.5 ± 0.1°C, n = 30) mice were equivalent. Catalepsy and rectal temperature were assessed at 60 min postinjection unless otherwise stated. The data reported are from a combination of male and female mice (no significant sex differences were observed for either genotype). All drugs were administered i.p. in a mixture of 1:1:18 ethanol:Emulphor:saline (10 μl/g body weight), except naloxone, which was administered i.p. in saline (SR141716A and THC kindly provided by the National Institute on Drug Abuse).

NAE levels in the brains of FAAH^+/+ and FAAH^−/− mice were quantified by isotope dilution liquid chromatography mass spectrometry (LC-MS) (16). Briefly, mice were anesthetized by using CO₂/O₂ and killed by decapitation. Brains were removed and immediately homogenized in a 2:1:1 mixture of chloroform: methanol:50 mM Tris, pH 8.0 containing N-oleoyl-d₄-ethanolamine and d₄-anandamide standards (0.5 nmol each per brain). The organic layer was removed, dried under N₂ gas, washed with ethyl acetate, and the washes transferred to a fresh glass vial and dried. The remaining residue was solubilized in methanol and injected onto an Agilent 1100 series LC-MS. Levels of endogenous NAEs were quantified by comparing their mass ion peak heights to those of the corresponding isotopically labeled standards. Standard curves were generated to confirm a linear relationship between peak height and NAE concentration. The anandamide data are shown in Fig. 6A. Endogenous brain levels of N-oleoyl ethanolamine were found to be 710 ± 90 and 18 ± 12 pmol/g tissue for FAAH^−/− and FAAH^+/+ mice, respectively. Although a standard for N-palmitoyl ethanolamine was not included in the assay, a peak corresponding to the molecular mass of this lipid (m/z = 300.3) was greatly increased in FAAH^−/− samples (relative to the included NAE standards), indicating that brain levels of this NAE were also up-regulated in these animals.

To generate mice lacking FAAH, the first exon of the FAAH gene was removed by homologous recombination (Fig. 1 A and B). Western blotting with anti-FAAH antibodies confirmed the absence of FAAH protein in the brain and liver of FAAH knockout (FAAH^−/−) mice (Fig. 1C). Immunofluorescence analysis of cerebellar sections from FAAH^+/+ mice revealed intense staining in the cell bodies of Purkinje and granule neurons, as well as in the molecular layer, which contains the extensive Purkinje dendritic arbor (Fig. 1D Left). In contrast, no FAAH immunoreactivity was detected in cerebellar sections from FAAH^−/− mice (Fig. 1D Right). These data confirm that FAAH is predominantly localized to the somatodendritic compartment of neurons in the central nervous system (34, 35). Importantly, brain and liver tissues from FAAH^−/− mice also exhibited greatly reduced FAAH catalytic activity (Table 1). For example, brain extracts from FAAH^−/− mice hydrolyzed anandamide and oleamide 50–100-fold more slowly than brain extracts from FAAH^+/+ mice.

FAAH^−/− mice were born at the expected Mendelian frequency and were viable, fertile, and largely indistinguishable from wild-type littermates. No differences were observed in general appearance, body weight, locomotion, or overt behavior. The role of FAAH in regulating anandamide activity in vivo was investigated by comparing responses of FAAH^+/+ and FAAH^−/− mice to anandamide in neurobehavioral assays that measured spontaneous activity, thermal pain sensation, catalepsy, and rectal temperature. Over the dose range tested (6.25–50 mg/kg, i.p. administration), anandamide failed to produce any significant effects in FAAH^+/+ mice (Fig. 2 A–D). Considering that anandamide has been previously found to induce modest, but significant cannabinoid pharmacology in some strains of mice (10, 11), the absence of any activity for anandamide in the FAAH^+/+ mice studied here may be due to strain differences. In contrast to wild-type mice, FAAH^−/− mice exhibited robust, dose-dependent behavioral responses to anandamide, displaying hypomotility (Fig. 2A), analgesia (B), catalepsy (C), and hypothermia (D). At the highest dose tested (50 mg/kg), anandamide induced in FAAH^−/− mice an 84 ± 8% reduction in spontaneous activity, profound analgesia in the tail immersion test (89 ± 11% maximum possible effect), strong cataleptic behavior in the bar test (88% of the test group), and a 7.9 ± 0.3°C reduction in rectal temperature.

The striking impact of anandamide (12.5–50 mg/kg, i.p.) on the behavior of FAAH^−/− mice was readily detected within 5 min of treatment, at which time the animals adopted a flattened, rigid posture and remained completely motionless with their eyes open. If startled by sound or touch, these mice would react with brief fits of spastic movement before quickly reentering a flattened, immobile state. Anandamide-treated FAAH^−/− mice remained immobile for 2–4 hours (depending on dose), after which they gradually reinitiated normal cage activities (e.g., movement, rearing, grooming). The duration of anandamide's behavioral effects in FAAH^−/− mice was further examined by measuring rectal temperature and catalepsy at various times posttreatment. Coinciding closely with their overt cage behavior, anandamide-treated FAAH^−/− mice (50 mg/kg, i.p.) showed a robust drop in rectal temperature that peaked between 1 and 2 h posttreatment and began to return to wild-type values by 4 h (Fig. 3A). Likewise, catalepsy was most extreme at 1 h posttreatment and gradually dissipated by 4 h (Fig. 3B). FAAH^−/− mice were indistinguishable from FAAH^+/+ mice when analyzed 24 h after treatment with anandamide.

The intense behavioral effects induced by anandamide in FAAH^−/− mice were reminiscent of those traditionally observed in rodents treated with THC (11), suggesting that anandamide may indeed act as an endogenous CB₁ ligand. To determine the contribution of CB₁ receptor pathways to the behavioral pharmacology elicited by anandamide in FAAH^−/− mice, these animals were administered the CB₁ antagonist SR141716A (10 mg/kg, i.p.) 10 min before treatment with anandamide (50 mg/kg, i.p.). Remarkably, all of anandamide's behavioral effects in FAAH^−/− mice were completely blocked by pretreatment with SR141716A (Fig. 4 A–D). In contrast, pretreatment with the opioid receptor antagonist naloxone (2 mg/kg, i.p.) failed to reduce any of anandamide's activities in FAAH^−/− mice (data not shown). The ED₅₀ values (mg/kg) for SR141716A-based antagonism of the behavioral effects of anandamide in FAAH^−/− mice were: for hypomotility, 1.2 [0.3–4.9 (95% confidence limits)]; for thermal analgesia, 1.2 (0.8–2.0); for catalepsy, 0.8 (0.5–1.5); and for hypothermia, 1.3 (0.6–3.0). These values are nearly identical to those reported previously for the antagonism of the behavioral effects of exogenous cannabinoid ligands by SR141716A (36), strongly supporting the notion that the observed behavioral effects of anandamide in FAAH^−/− mice occur predominantly, if not exclusively through the CB₁ receptor. The profound CB₁-dependent pharmacological effects induced by anandamide in FAAH^−/− mice initially suggested that a sensitized or up-regulated CB₁ receptor system might exist in these animals. However, FAAH^−/− and FAAH^+/+ mice exhibited similar THC dose-response profiles for all of the behavioral assays tested (Table 2), indicating that their CB₁ receptors were functionally equivalent. Collectively, these data indicate that in the absence of FAAH, anandamide acts as potent and selective CB₁ agonist in vivo.

During the course of characterizing the pharmacological effects of anandamide in FAAH^+/+ and FAAH^−/− mice, baseline measures of spontaneous activity, rectal temperature, and pain sensitivity were recorded. FAAH^−/− mice did not differ from FAAH^+/+ mice in terms of their respective locomotor activities or rectal temperatures (Fig. 2 A and D). However, FAAH^−/− mice were found to display altered thermal pain sensation, exhibiting prolonged response latencies in both the tail immersion and hot plate tests when compared with both FAAH^+/+ and FAAH^+/− mice (Fig. 5A Left and Fig. 6B). In the hot plate test, FAAH^−/− mice differed significantly from FAAH^+/+ mice over a restricted temperature range (Fig. 5A Right). Interestingly, a similar phenotype was observed in the substance P/neurokinin A knockout mouse and interpreted to implicate these peptides in neural pathways that communicate moderate to intense pain stimuli (37). To examine whether other forms of pain sensation were also altered in FAAH^−/− mice, these animals were compared with FAAH^+/+ and FAAH^+/− mice in the formalin test for chemical pain sensation. FAAH^−/− mice exhibited a significant reduction in pain behavior during the first phase of the formalin test relative to both FAAH^+/+ and FAAH^+/− mice (Fig. 5B). A similar trend of reduced pain responses in FAAH^−/− mice was also observed in the second phase of the formalin test, but these data just failed to reach significance (P = 0.055 for FAAH^−/− versus FAAH^+/+ mice, planned comparison). Collectively, these results implicate FAAH in the regulation of mammalian behavioral responses to multiple forms of pain stimuli.

The reduced pain perception exhibited by FAAH^−/− mice suggested that these animals might possess enhanced endogenous cannabinoid activity. To explore this notion further, brain levels of anandamide and related NAEs were measured by isotope dilution liquid chromatography mass spectrometry (LC-MS). Strikingly, brains from FAAH^−/− mice possessed 15-fold higher levels of anandamide than brains from FAAH^+/+ mice (Fig. 6A). Other NAEs, including N-oleoyl and N-palmitoyl ethanolamine, were similarly up-regulated (see Materials and Methods), consistent with shared biosynthetic and degradative pathways for these lipids (38). To test whether enhanced levels of endogenous anandamide might be responsible for the analgesia observed in FAAH^−/− mice, the pain responses of these animals, FAAH^+/+, and FAAH^+/− mice were tested in the hot plate test both before and after treatment with either vehicle or SR141716A (Fig. 6B). No significant changes from baseline were observed for the hot plate response latencies of vehicle-treated FAAH^+/+, FAAH^+/−, and FAAH^−/− mice, or in SR141617A-treated FAAH^+/+ and FAAH^+/− mice (Fig. 6B Right). In sharp contrast to all of these test groups, FAAH^−/− mice treated with SR141716A showed a dramatic reduction in their pain response latencies (Fig. 6B Right), indicating that a substantial fraction of their altered pain sensation was mediated by CB₁ receptor pathways.