Increasing Cannabinoid Levels By Pharmacological And Genetic Manipulation Delay Disease Progression In SOD1 MiceLynsey G. Bilsland, James R. T. Dick, Gareth Pryce, Stefania Petrosino, Vincenzo Di Marzo, David Baker and Linda Greensmith
The FASEB Journal - May 1, 2006
Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder characterized by the selective loss of motoneurons in the spinal cord, brain stem, and motor cortex. However, despite intensive research, an effective treatment for this disease remains elusive. In this study we show that treatment of postsymptomatic, 90-day-old SOD1G93A mice with a synthetic cannabinoid, WIN55,212—2, significantly delays disease progression. Furthermore, genetic ablation of the Faah enzyme, which results in raised levels of the endocannabinoid anandamide, prevented the appearance of disease signs in 90-day-old SOD1G93A mice. Surprisingly, elevation of cannabinoid levels with either WIN55,212—2 or Faah ablation had no effect on life span. Ablation of the CB1 receptor, in contrast, had no effect on disease onset in SOD1G93A mice but significantly extended life span. Together these results show that cannabinoids have significant neuroprotective effects in this model of ALS and suggest that these beneficial effects may be mediated by non-CB1 receptor mechanisms.– Bilsland, L. G., Dick, J. R. T., Pryce, G., Petrosino, S., Di Marzo, V., Baker, D., Greensmith, L. Increasing cannabinoid levels by pharmacological and genetic manipulation delay disease progression in SOD1 mice.
AMYOTROPHIC LATERAL SCLEROSIS (ALS) is a fatal neurodegenerative disorder characterized by the progressive loss of motoneurons in the spinal cord, brainstem, and motor cortex (1) . ALS is largely a sporadic disease, although around 10% of cases are familial and of these ∼20% are due to mutations in the Cu/Zn superoxide dismutase (SOD) (SOD1) enzyme (2) . Transgenic mice expressing SOD1 mutations (3) have been studied extensively, but the precise pathogenesis of motoneuron degeneration remains unclear. However, evidence implicates glutamate excitotoxicity in the selective vulnerability of motoneurons in ALS (1). Indeed Riluzole, an inhibitor of glutamate release, is currently the only licensed treatment for ALS, although its therapeutic effects are minimal (4) . Therefore, the identification of alternative and effective therapeutic compounds remains imperative.
The potential neuroprotective properties of cannabinoids have been explored in several models of neurodegeneration. Cannabinoids exert their actions by activation of CB1 and CB2 receptors and are neuroprotective under excitotoxic (56 7 8) , ischemic (9) , and inflammatory conditions (8 , 10 11 12 13 14) and following brain trauma (15) . The recent discovery of endocannabinoids, endogenous metabolites that activate cannabinoid receptors, and an increasing understanding of their synthesis, reuptake, and degradation have focused research into the therapeutic potential of agents that manipulate the endocannabinoid system.
Endocannabinoids are synthesized by neurons in an activity-dependent manner (16 , 17) . It is, therefore, possible that excitotoxicity-induced neuronal hyperexcitability can increase endocannabinoid levels. Indeed, injection of excitotoxins into the central nervous system (CNS) of mice increases levels of the endocannabinoids, anandamide (AEA), and 2-arachidonoylglycerol (2-AG; 7, 15, 18, 19). Activation of CB1 receptors in the CNS by elevated levels of cannabinoids has an inhibitory effect on neurotransmitter release and cellular calcium influx. Thus, CB1 receptor activation can potentially limit the release and postsynaptic effects of excitotoxic glutamate (5 6 7 8 , 20) , thereby providing endogenous protection against excitotoxicity. In support of this possibility, it has been shown that injection of kainate into the CNS of CB1 receptor knockout mice produces a more extreme behavioral reaction and a greater mortality than in wild-type (WT) mice (7 , 8) .
Because Bglutamate excitotoxicity is thought to play a significant role in motoneuron degeneration in ALS (1) , we examined the effects of manipulation of the cannabinoid system in a mouse model of ALS. We studied the effect of treatment with a cannabinoid receptor agonist on disease progression and life span in SODG93A mice. In addition, the effects of genetically manipulating levels of the endogenous cannabinoid system were also tested.
MATERIALS AND METHODS
Transgenic mice carrying a human SOD1 gene with a G93A mutation (TgN[SOD1-G93A]1Gur) were obtained from Jackson Laboratories (Bar Harbor, ME). The colony was maintained by breeding male heterozygous carriers with female (C57BL/6 xSJL) F1 hybrids. Mice lacking the fatty acid amide hydrolase gene (Faah —/—) and mice lacking the CB1 receptor gene (Cnr1 —/—) were backcrossed for at least seven generations onto the ABH mouse background (21 , 22) . SOD1.Faah Â± and SOD1.Cnr1 Â± mice were obtained by breeding male heterozygous SOD1G93A carriers with female Faah or CB1 receptor knockout mice. Male heterozygous carriers for both genes were backcrossed again with female Faah or CB1 receptor knockouts to obtain a F2 generation of SOD1.Faah —/— and SOD1.Cnr1 —/— mice for use in this study. Individual animals were genotyped by polymerase chain reaction using DNA extracted from tail snips. The animals were housed in a controlled temperature and humidity environment and maintained on a 12-h light/dark cycle with access to food and water provided ad libitum. All experiments were performed under license from the U.K. Government in accordance with the Animals (Scientific Procedures) Act 1986 and following approval from the Institute of Neurology Ethical Review Committee. In this study, all experiments were carried out blind.
Analysis of endocannabinoid levels
The levels of endocannabinoids in spinal cords and brains of WT and SOD1G93A mice were assessed at 40, 90, or 120 d of age (n=at least 3 in each group). The mice were killed by overdose of pentobarbitone anesthetic (Fort Dodge Animal Health, Southampton, UK), and their spinal cords and brains were rapidly dissected and immediately frozen in liquid nitrogen. Tissues were dounce-homogenized with chloroform/methanol/Tris-HCl 50 mM, pH 7.4 (1/1/1 by vol), containing 5 pmol of octa-deuterated (d8)-anandamide and 50 pmol of d8—2-arachidonoylglycerol (Cayman Chemicals; Ann Arbor, MI) as internal standards. Lipid-containing organic phase was dried down, weighed, and prepurified by open-bed chromatography on silica gel and analyzed by liquid chromatography-atmospheric pressure chemical ionization-mass spectrometry using a Shimadzu HPLC apparatus (LC-10ADVP; Kyoto, Japan) coupled to a Shimadzu quadrupole mass spectrometer (LCMS-2010; Kyoto, Japan) via a Shimadzu APCI interface (Kyoto, Japan). Mass spectrometry analyses were performed in the selected ion-monitoring mode as described previously (23) . Temperature of the APCI source was 400Â°C; HPLC column was a Phenomenex (i.d.=3 Âµm, length 150 mM, section 4.5 mM) reverse-phase column, eluted as described (23) . AEA (retention time of 14.5 min) and 2-AG (retention time of 17.0 min) quasi-molecular ions were quantified by isotope dilution with the above-mentioned deuterated standards (23) , and their amounts in picomols and nanomols, respectively, normalized per milligram of lipid extract.
Drug preparation and administration
The cannabinoid receptor agonist R(+)-WIN55,212—2 (Tocris, Bristol, UK) was dissolved in a vehicle that consisted of ethanol, cremaphor (Fluka Biochemika, Buchs, Switzerland), DMSO (Sigma-Aldrich, Poole, UK), and PBS (PBS; Oxoid Ltd, Basingstoke, UK; ratio 1:1:1:17). WT and SOD1G93A expressing mice of both sexes were divided into three treatment groups: treatment with 0.1 ml WIN55,212—2 (5 mg/kg) injected intraperitoneally (i.p.) once daily (excluding weekends) from 90 d of age; treatment with 0.1 ml vehicle i.p.; no treatment to serve as untreated controls. At 90 d of age, SOD1G93A mice showed clear deficits in locomotor activity. Moreover, we have previously shown that by this stage they have already lost ∼40% of the motoneurons innervating their hind-limb muscles, and these muscles are considerably weaker than normal (27) . By 90 d of age, SOD1G93A mice were, therefore, considered to be "symptomatic".
Assessment of muscle force
The maximum force of the tibialis anterior (TA) and extensor digitorum longus (EDL) muscles of animals was assessed at 120 d of age for WIN55,212—2 treated mice, and 90 d of age for the Faah —/— and Cnr1 —/— mice. The animals were anesthetized with 4% chloral hydrate (1 ml/100g body wt, i.p.; Sigma-Aldrich, Poole, UK) and prepared for isometric tension recordings of muscle contraction (24) . The distal tendons of hind-limb TA and EDL muscles were exposed, dissected free from surrounding tissue, and cut. The sciatic nerve was exposed and sectioned, and all of its branches were cut, apart from the deep peroneal nerve, which innervates the TA and EDL muscles. The hind limbs of the animals were rigidly secured to the table with stainless steel pins, and the distal tendons of the TA and EDL muscles were attached to an isometric force transducer (Dynamometer UFI Devices, Welwyn Garden City, UK) via thread. Once attached, the length of each muscle was adjusted to obtain maximal twitch tension. Both muscles and nerves were kept moist with saline, and experiments were performed at room temperature.
Isometric contractions were elicited by stimulation of the sciatic nerve by using square-wave pulses of 0.02 ms duration at a supramaximal voltage (10 V) via platinum electrodes. Maximal tetanic contraction was assessed by stimulating the sciatic nerves with trains of stimuli at increasing frequencies of 40, 80, and 100 Hz for a total duration of 450 ms.
Motor unit survival
The number of functional motor units in each EDL muscle was determined by stimulating the sciatic nerve with stimuli of increasing intensity, resulting in stepwise increments in twitch tension due to successive recruitment of motor axons. The motor unit traces were recorded on a storage oscilloscope (Tektronix, Beaverton, OR) and photographed using a Tektronix oscilloscope camera (Beaverton, OR). The number of increments was counted to give an estimate of the number of motor axons present in the nerve.
Muscle fatigue resistance
At the end of the isometric tension recordings, the resistance of the EDL muscles to fatigue during repeated stimulation was tested. The EDL muscles were repeatedly stimulated at 40 Hz for 250 ms of every second for 3 min, and the tetanic contractions were recorded on a pen recorder (Lectromed Multitrace 2, Letchworth Garden City, UK).
At the end of each acute experiment, the EDL muscles were removed, weighed, and snap-frozen in isopentane cooled in liquid nitrogen and stored at —80Â°C until processing. Serial cross sections (10 Âµ m) of EDL muscle were cut on a cryostat and stained for succinate dehydrogenase (SDH) activity to determine the oxidative capacity of the muscle fibers, as described previously (24) .
Following each physiological experiment, and in the life span study, the mice were terminally anesthetized with 4% chloral hydrate (1 ml/100 g body wt, i.p.) and perfused transcardially with 4% paraformaldehyde (TAAB Ltd. Aldermaston, UK) in 0.1 M PBS. The lumbar spinal cords (L2-L6) were removed, postfixed for 4 h in the same fixative, and then cryoprotected in sucrose (30% in PBS). Serial sections (20 Âµ m) were cut on a cryostat and stained with gallocyanin, a Nissl stain (25) . The number of surviving motoneurons within the sciatic motor pool of each ventral horn was assessed by counting the number of Nissl-stained motoneurons in every third spinal cord section (n=60) using a light microscope (Leica DMR, Milton Keynes, UK). This method avoids counting the same cell twice in consecutive sections. In addition, morphological criteria were also used, so that only large polygonal neurons in which the nucleolus was clearly visible at high magnification were included in the counts. This method of counting has previously been used to assess motoneuron survival (24) . To assess the extent of motoneuron degeneration, mean motoneuron survival in the sciatic motor pool of SOD1G93A mice was compared with the number surviving in age-matched WT littermates. Furthermore, the effect of WIN55,212—2 treatment or ablation of Faah or the CB1 receptor on motoneuron survival was compared with both WT and untreated SOD1G93A littermates.
Statistical significance among the groups was assessed using either a Mann-Whitney U test or one-way ANOVA, incorporating a Student Neuman Keuls multiple comparisons test. Significance was set at P < 0.05.
Endocannabinoid levels in SOD1G93A mice
assess whether endocannabinoid levels varied during disease progression in SOD1G93A mice, we first examined levels of endocannabinoids in both WT and SOD1G93A mice at various ages. The levels of the endocannabinoids, AEA and 2-AG, were measured in the spinal cord of WT and SOD1G93A littermates at 40 (presymptomatic), 90 (early symptomatic), and 120 d of age (late stage), respectively. The results showed no significant difference in the levels of AEA and 2-AG between WT and SOD1G93A at a presymptomatic stage of disease (P >0.15). However, as shown in Fig. 1 , at 90 and 120 d of age levels of AEA in SOD1G93A mice were elevated more than 2-fold compared with WT littermates (P <0.001; 18.1 Â±2.3 pmol/g and 14.9 Â±2.4 pmol/g compared with 8.0 Â±2.3 pmol/g and 6.9 Â±1.7 pmol/g, respectively). Levels of 2-AG were similarly elevated in SOD1G93A mice at 120 d of age compared with WT (P =0.04; 6.7 Â±0.7 nmol/g compared with 4.4 Â±0.5 nmol/g), although they did not differ significantly at 90 d (P =0.09). Thus, as previously shown (26) , a clear increase was found in the levels of endocannabinoids in SOD1G93A mice during the symptomatic phase of disease in contrast with prior to symptom onset. Interestingly, an increase was also found in levels of 2-AG in the brain of symptomatic SOD1G93A mice (P <0.04), which might represent an attempt to protect descending cortical and bulbar neurons from cell death.
Figure 1. Levels of endocannabinoids in spinal cords of WT and symptomatic SOD1G93A mice. Levels of the endocannabinoids (a) anandamide (AEA) and (b) 2-arachidonoylglycerol (2-AG) were analyzed in spinal cord extracts from untreated SOD1G93A mice at 90 and 120 d of age (n =at least 3 in each case). *P < 0.05, ***P < 0.005. Error bars represent SEM.
Effect of treatment with the cannabinoid agonist WIN55,212—2
WT and transgenic SOD1G93A mice of both sexes were injected daily with WIN55,212—2 (5 mg/kg, i.p.), a synthetic CB1 and CB2 receptor agonist, from 90 d of age, after symptom onset. The effect on disease progression, motoneuron survival, and life span was examined.
In vivo physiological assessment of muscle force of TA and EDL hind-limb muscles was performed at 120 d of age, and the results are summarized in Fig. 2 a. At this stage in untreated SOD1G93A mice, the maximum force of both TA and EDL muscles was significantly reduced compared with muscle force in their WT littermates (P<0.005). However, in SOD1G93A mice treated with WIN55,212—2, the hind-limb muscles were significantly stronger than in their untreated SOD1G93A littermates (P <0.05). We found no significant difference in the muscle force of untreated SOD1G93A mice and those treated with vehicle (P = 0.1).
Figure 2. Effect of treatment with WIN55,212—2 on muscle force and motor unit number in 120 d old SOD1G93A mice. a) Maximal tetanic force generated by the TA and EDL muscles of WT (n =10), untreated SOD1G93A (n =18) and WIN55,212—2 treated SOD1G93A mice (n =18). Force is expressed per 100 mg muscle wt to take into account the reduction in muscle wt in 120-day-old SOD1G93A mice. Examples of motor unit traces from (b) WT (n =10), (c) untreated SOD1G93A (n =18) and (d) WIN55,212—2 treated SOD1G93A (n =16) mice. e) Mean number of motor units surviving in each experimental group. *P < 0.05, ***P < 0.005. Error bars represent SEM.
Motor unit survival
The number of functional motor units in each EDL muscle was also established by stimulating the sciatic nerve with stimuli of increasing intensity, which resulted in stepwise increments in twitch tension due to successive recruitment of motor axons. Representative examples of traces from WT, untreated SOD1G93A and WIN55,212—2-treated SOD1G93A mice are shown in Fig. 2b-d , and the mean motor unit survival in each group is summarized in the bar chart (Fig. 2e ). In WT mice at 120 d of age, EDL muscles have on average 29 (Â±0.9) motor units compared with only 12 (Â±0.9; P<0.001) in untreated SOD1G93A mice. Treatment with WIN55,212—2 significantly improves motor unit survival in SOD1G93A mice and 19 (Â±0.9; P <0.001) motor units survive, even at 120 d of age.
Muscle fatigue characteristics
As disease progresses in SOD1G93A mice, the contractile characteristics of EDL muscles change dramatically (27) . We next examined the effect of WIN55,212—2 treatment on the fatigue characteristics of EDL muscles in 120-day-old SOD1G93A mice. The EDL muscles were stimulated repeatedly and the contractions were recorded, producing a fatigue trace. Examples of fatigue traces from EDL muscles in WT, untreated SOD1G93A and WIN55212—2 treated SOD1G93A mice are shown in Fig 3 a—c. EDL is normally a fast muscle that fatigues rapidly when repeatedly stimulated (Fig. 3a ). In contrast, in untreated SOD1G93A mice in the late stages of disease, EDL muscles become largely fatigue-resistant (Fig. 3b). However, in SOD1G93A mice treated with WIN55,212—2, EDL, muscles show fatigue characteristics that clearly resemble those of normal EDL muscles (Fig. 3c ).
Figure 3. Effect of treatment with WIN55,212—2 on the contractile characteristics of EDL muscles in SOD1G93A mice at 120 d of age. Representative fatigue traces obtained from repeated stimulation of EDL muscles from (a) WT (n=10), (b) untreated SOD1G93A (n=17), and (c) WIN55,212—2 treated SOD1G93A (n=16) mice. Scale bar = 50 s. Representative examples of EDL muscle sections stained for succinate dehydrogenase, an indicator of oxidative capacity, from (d) WT, (e) untreated SOD1G93A, and (f) WIN55,212—2 treated SOD1G93A mice. Scale bar = 200 Âµm.
These changes in the fatigue characteristics of EDL muscles of SOD1G93A mice were reflected in the histochemical properties of the muscle fibers. Staining of EDL muscle sections from 120-day-old WT mice for SDH, an indicator of oxidative capacity, revealed a mosaic pattern of lightly and darkly stained muscle fibers, with the majority of fibers staining lightly for SDH (Fig. 3d ). In 120-day-old SOD1G93A mice, however, a greater proportion of fibers was more intensely stained, indicating that these fibers have an increased oxidative capacity (Fig. 3e). In contrast, EDL muscles from 120-day-old SOD1G93A mice treated with WIN55,212—2 show a pattern of SDH staining that is more characteristic of WT EDL muscles, with a greater proportion of fibers lightly stained (Fig. 3f ) than in the untreated SOD1G93A EDL muscle.
The effect of treatment with WIN55,212—2 on motoneuron survival was assessed in 120- day-old SOD1G93A mice, by counting the number of motoneurons in a segment of the sciatic motor pool of each spinal cord. Examples of Nissl-stained spinal cord sections from WT, SOD1G93A, and WIN 55,212—2-treated SOD1G93A mice are shown in Fig. 4 a—c. The improvement in motor unit survival observed in WIN55,212—2-treated SOD1G93A mice was reflected in an increase in motoneuron survival, and the results are summarized in Fig. 4d . At 120 d of age, a significant number of motoneurons in the sciatic pool had already died in untreated SOD1G93A mice and only 140 (Â±6.7) motoneurons survived compared with 369 (Â±12.9; P <0.001) in WT littermates. However, treatment with WIN55,212—2 rescues a significant proportion of motoneurons, so that 199 (Â±10.2; P <0.001) motoneurons survived. Thus, in WIN55,212—2 treated SOD1G93A mice, 42% more motoneurons survive even at 120 d compared with their untreated SOD1G93A littermates.
Figure 4. The neuroprotective effect of treatment with WIN55,212—2 in SOD1G93A mice at 120 d of age. Spinal cord sections stained for Nissl, showing motoneurons in the sciatic motor pool (dotted areas) of (a) WT (n =8), (b) untreated SOD1G93A (n =12), and (c) WIN55,212—2 treated SOD1G93A (n =16) mice. Scale bar = 200 Âµm. d) Mean motoneuron survival in each experimental group. e) Life span of untreated SOD1G93A mice (n =23) and those treated with WIN55,212—2 (n =10) from 90 d of age. ***P < 0.005. Error bars represent SEM.
Life span and end-stage motoneuron survival
We next examined the effect of treatment with WIN55,212—2 on the life span of SOD1G93A mice. Untreated SOD1G93A mice on the traditional (C57BL/6 xSJL) F1 background live on average for 131 d (Â±1.4; Fig. 4e ). End-stage in these experiments is determined as the age when the mice have lost ≥ 20% of their body wt or when they can no longer right themselves within 30 s. Surprisingly treatment with WIN55,212—2 had no effect on the life span, and treated SOD1G93A mice had an average life span of 134 d (Â±2.3; p =0.96; Fig. 4e ). The extent of motoneuron survival was also examined in these mice at end-stage. Although treatment with WIN55,212—2 significantly improved motoneuron survival when the SOD1G93A mice were assessed at 120 d of age, by end-stage disease at 134 d, no significant difference was found in the number of motoneurons surviving in the treated mice compared with their untreated SOD1G93A littermates (see Table 1A ).
Table 1. Extent of motoneuron survival
These results show that treatment of SOD1G93A mice with a cannabinoid agonist from 90 d of age results in a significant improvement in muscle function and motoneuron survival at 120 d of age, a relatively late stage of disease. However, this amelioration in disease does not result in a significant increase in life span and has no impact on motoneuron survival in the long term. Therefore, we next examined whether altering the action or endogenous levels of endocannabinoids by genetic manipulation would have a greater impact on disease than pharmacological treatment initiated at 90 d.
Manipulation of the endocannabinoid system by genetic ablation of either the Faah enzyme or the CB1 receptor
We next studied the effect of raising levels of the endocannabinoid AEA in SOD1G93A mice, by crossing SOD1G93A mice with knockout mice lacking the fatty acid amide hydrolase enzyme (Faah —/—). Faah is normally responsible for the hydrolysis of AEA, and in Faah —/— mice brain AEA levels are elevated up to 15-fold (21) . In addition, in a separate experimental set, in order to examine the effects of blocking the action of endogenous cannabinoids and to further elucidate the mechanism of action of WIN55,212—2 on disease progression in SOD1G93A mice, we crossed SOD1G93A mice with mice lacking the CB1 receptor gene (Cnr1 —/—). For these breeding experiments we crossed the control SOD1G93A mice onto the ABH background of the ABH.Faah —/— and ABH.Cnr1 —/— mice. Disease progression in SOD1.Faah —/— and SOD1.Cnr1 —/— mice was assessed at a symptomatic stage of disease, at 90 d of age, to enable us to observe any positive or negative effects of Faah or CB1 receptor ablation on disease progression.
We first examined the muscle characteristics of SOD1G93A mice bred onto the ABH background of the ABH.Faah —/— and ABH.Cnr1 —/— mice. In SOD1G93A mice on the ABH background, as observed in the F1 hybrid background, by 90 d of age maximal tetanic tension of TA and EDL muscles was already substantially reduced compared with their WT littermates (Fig. 5 a; p<0.001). However, both TA and EDL muscles were dramatically stronger in SOD1.Faah —/— mice at 90 d of age compared with their SOD1G93A littermates. Indeed, the maximal tetanic tension of EDL muscles in SOD1.Faah —/— mice was no different to that of EDL muscles from WT mice, and in the TA muscles only a slight reduction in force was found compared with WT (p>0.4 EDL and p>0.01 TA; Fig. 5a ). In contrast, no significant difference was found in the maximal force of either TA or EDL muscles between SOD1G93A mice and SOD1.Cnr1 —/— mice (p>0.7, TA; p>0.1 EDL), indicating that muscle force in SOD1G93A mice is not affected by ablation of the CB1 receptor.
Figure 5. The effect of Faah and CB1 receptor ablation on muscle force and motor unit survival in SOD1G93A at 90 d of age. a) Maximal tetanic force generated by TA and EDL muscles from WT (n=10), SOD1G93A (n=10), SOD1.Faah —/— (n=9), and SOD1.Cnr1 —/— mice (n=9). b) Mean motor unit survival. *P < 0.05, ***P < 0.005. Error bars represent SEM.
Motor unit survival
We also assessed the number of motor units in EDL muscles of WT, SOD1G93A, SOD1.Faah —/— and SOD1.Cnr1 —/— mice on the ABH background, and the results are summarized in Fig. 5b . At 90 d of age a significant loss of motor units in SOD1G93A mice was found, and only 19 (Â±1.4) motor units survived compared with 30 (Â±1.2; p <0.001) motor units in WT littermates. However, SOD1.Faah —/— mice had significantly more motor units than the SOD1G93A littermates of the same age. Thus 27 (Â±1.5; p<0.001) motor units survived at 90 d of age, which is not significantly different from the number of motor units present in the EDL muscles of WT littermates (p>0.09). Therefore, no motor unit loss has occurred in SOD1.Faah —/— mice at 90 d of age. In contrast, in SOD1.Cnr1 —/— mice the number of surviving motor units at 90 d of age was no different from SOD1G93A littermates, and only 19 (Â±1.4; p>0.7) motor units survived.
Motoneuron survival was also assessed both after the acute physiological experiments at 90 d and when the mice had reached end-stage, by examination of Nissl-stained spinal cord sections of mice from each experimental group (examples shown in Fig. 6 a—d). The results of motoneuron survival in each group of mice examined at 90 d are summarized in Fig. 6e , which shows that significant motoneuron death has already occurred by 90 d of age in SOD1G93A mice and only 237 (Â±19.5) motoneurons survive compared with 358 (Â±13.8; P<0.001) in their WT littermates. In SOD1G93A mice in which the Faah enzyme was ablated very little motoneuron death had occurred by 90 d of age, so that 305 (Â±18.1) motoneurons survived in SOD1.Faah —/— mice (P=0.035). Thus, although almost 35% of motoneurons have died by 90 d of age in SOD1G93A mice, far less motoneuron degeneration occurs in SOD1.Faah —/— mice, in which the Faah enzyme is ablated, and in which levels of the endocannabinoid AEA are significantly elevated. However, in SOD1.Cnr1 —/— mice the absence of the CB1 receptor did not affect motoneuron survival in SOD1G93A mice and at 90 d only 217 (Â±12.8; P=0.575) motoneurons survived, which is not significantly different from the motoneuron survival observed in their SOD1G93A littermates.
Figure 6. Motoneuron survival and life span in SOD1.Faah —/— and SOD1.Cnr1 —/— mice. Spinal cord sections showing Nissl-stained motoneurons in the sciatic motor pool (dotted areas) of (a) WT (n=10), (b) SOD1G93A (n=12), (c) SOD1.Faah —/— (n=10), and (d) SOD1.Cnr1 —/— (n=10) mice. Scale bar = 200 Âµ m. e) Mean motoneuron survival in each experimental group. f) Life span of untreated SOD1G93A mice on the ABH background (n=5) and SOD1G93A mice in which either the Faah enzyme (n=12) or the CB1 receptor was ablated (n=8). *P < 0.05, ***P < 0.005. Error bars represent SEM.
Life span and end-stage motoneuron survival
We next examined the effect of genetic manipulation of the endocannabinoid system on the life span of the SOD1G93A mice. In these breeding experiments, SOD1G93A mice were bred onto the ABH background of the Faah —/— and Cnr1 —/— mice and F2 generation animals were used in all experiments. In this colony the life span of the F2 generation SOD1G93A mice was reduced compared with that of SOD1G93A mice bred onto the traditional (C57BL/6xSJL) F1 background. Thus, ABH SOD1G93A mice lived for 118 d (Â±4.2; Fig. 6f ) compared with 131 d (Â±1.4) for SOD1G93A on the (C57BL/6xSJL) F1 background (P=0.011). Surprisingly, despite the dramatic improvement in disease signs observed at 90 d of age in SOD1.Faah —/— mice, ablation of the Faah enzyme had no significant effect on life span and SOD1. Faah —/— animals lived for 114 d (Â±1.5; P=0.399). Furthermore, motoneuron survival in the SOD1.Faah —/— mice at end-stage was not significantly different from that in their SOD1G93A littermates at end-stage (See Table 1B ; P>0.7). Conversely, although we observed no difference in disease signs between untreated SOD1G93Aand SOD1.Cnr1 —/— at 90 d of age, ablation of the CB1 receptor significantly increased life span and SOD1.Cnr1 —/— mice lived for 133 d (Â±5.9; P=0.03), by which time they had lost a similar proportion of their motoneurons as their SOD1G93A littermates (see Table 1B ; P>0.7).
These results show that treatment of symptomatic SOD1G93A mice with the exogenous cannabinoid WIN55,212—2 significantly delays disease progression. Moreover, our results show for the first time, that inhibition of the Faah enzyme and the consequent increase in endocannabinoid levels have clear beneficial effects in SOD1G93A mice. Genetic augmentation of endogenous levels of the endocannabinoid AEA, via ablation of the Faah enzyme, almost completely ameliorated disease symptoms in 90-day-old SOD1G93A mice. Surprisingly, however, the dramatic improvements in disease progression observed in SOD1G93A mice following both pharmacological and genetic enhancement of the cannabinoid system, were not reflected in an increased life span.
A number of antiexcitotoxic therapies have been shown to have beneficial effects in animal models of ALS (28 , 29) . However, in ALS patients, treatment with the anti-glutamatergic agent Riluzole has limited effects, with patient life span extended by only 2—4 mo (4) . Cannabinoids reduce excitotoxicity in models of neuronal death via a CB1 receptor-dependent mechanism (for example 5—8, 20). However, they also exhibit other potentially beneficial effects and have antiinflammatory (30) and antioxidant properties (31 , 32) , which might be particularly relevant to ALS. This combination of potential neuroprotective mechanisms against inflammation, oxidative stress, and excitotoxicity, suggests that cannabinoids might have a greater impact on disease progression in ALS than therapies targeting excitotoxicity alone.
Cannabinoids act via activation of CB1 and CB2 receptors, although in the CNS, CB1 receptors predominate (33) . In the present study, WIN55,212—2 or increased endogenous AEA may act via the CB1 receptor to inhibit glutamate neurotransmission and thus limit excitotoxicity, resulting in disease amelioration in SOD1G93A mice. Indeed, in WIN55,212—2-treated SOD1G93A mice, hind-limb muscles are significantly stronger than in untreated littermates. Moreover, Faah ablation, which raises AEA levels, resulted in an even greater improvement in muscle force. In addition, deleterious changes in muscle characteristics that occur in untreated SOD1G93A mice (27) do not occur and significantly more motor units and motoneurons survive in WIN55,212—2 treated and SOD1.Faah —/— mice compared with their untreated SOD1G93A littermates.
Other authors have previously shown that cannabinoid treatment might have therapeutic potential in ALS (34 , 35) . The present results are consistent with these studies, which show a significant delay in disease progression in SOD1G93A mice exposed to elevated levels of cannabinoids. Although Raman et al. (34)report a significant increase in survival of SOD1G93A mice following treatment with an exogenous cannabinoid, this increase was only 6.4 d. In the present study, treatment with WIN55,212—2 increased life span by 3—4 d, which was not found to be significant. However, the present study extends these observations to show that ablation of the CB1 receptor has no effect on disease progression at 90 d of age but significantly extends life span of SOD1G93A mice. These findings suggest that i) CB1 receptor activation might not play a significant role in the cellular defense mechanism of motoneurons in SOD1G93A mice; ii) the neuroprotective effects of cannabinoids are not necessarily mediated by the CB1 receptor; and iii) inhibition/blockade of the CB1 receptor might be neuroprotective. Indeed, previous results support such a possibility and for example treatment with a CB1 receptor antagonist protects against N-methyl-DL-aspartate toxicity (36) .
Our results show that ablation of CB1 receptors has beneficial effects in SOD1G93A mice, although elevation of endocannabinoid levels also has neuroprotective effects. Together, these results suggest that the beneficial effects of Faah ablation and WIN55,212—2 treatment might not be mediated via CB1 receptors, suggesting a role for CB2 receptors. Therefore, selective targeting of CB2 receptors might have therapeutic potential in ALS. Neuroprotective effects of cannabinoids in the presence of CB1 receptor antagonism have previously been reported (9 , 13 , 37) , indicative of a non-CB1 receptor mediated effect. Interestingly, CB2 receptors are expressed on microglia (38) and in ALS the role of inflammation is becoming increasingly apparent (39) . Inflammation occurs in the spinal cord of ALS patients, with significant proliferation and accumulation of activated microglia, reactive astrocytes, and CD4+ and CD8+ cells in areas of degeneration (39) . Furthermore, increased levels of inflammatory markers such as cyclo-oxygenase 2 (COX2) and prostaglandin E2 have also been reported in the spinal cords of ALS patients (40 , 41) and in the SOD1G93A mouse, where a correlation between the intensity of inflammation and disease progression has been observed (42 , 43) . Moreover, minocycline, an inhibitor of microglial activation, delays disease onset and extends survival in SOD1 mice (44) . Activated microglia release various neurotoxic mediators, including pro-inflammatory cytokines, reactive oxygen species, and glutamate (10 , 38) , all of which would contribute to motoneuronal injury. Cannabinoids have been shown to reduce microglial activation (45) and reduce expression and release of pro-inflammatory cytokines from microglia and astrocytes (10 , 11) via a CB2 receptor-mediated mechanism (13) . An inhibitory effect of cannabinoids on microglia would also limit their glutamate release, thus lowering the potential contribution of excitotoxicity to disease pathogenesis (46) . Indeed, inhibitors of inflammation have already been shown to have beneficial effects in ALS models. Selective COX2 inhibitors delay onset (47) and extend life span by 25% in the SOD1G93A mouse model (48) . However, a recent clinical trial using celebrex (a selective COX2 inhibitor) failed to have any significant effect in ALS patients, although it is unclear whether this result is solely related to celebrex or to all COX2 inhibitors (49) .
Another possible target mediating the effect of WIN55,212—2 in nervous tissue is the as-yet-unidentified G-protein-coupled receptor characterized pharmacologically in the brain of CB1-knockout mice (23) . Given the protective actions of elevation of AEA levels following Faah ablation shown here, such a receptor, which is equally activated by AEA and WIN55,212—2, but not by other synthetic cannabinoids, or by Δ (9) -tetrahydrocannabinol, might contribute to the protective effects observed here. Indeed, it has been suggested that this receptor is selectively coupled to inhibition of glutamate release in the mouse hippocampus (50) .
The results of the present study identify for the first time the Faah enzyme as a novel therapeutic target in ALS. We have shown that in mice in which the Faah enzyme is absent and in which levels of the endocannabinoid AEA are significantly increased (21) , a dramatic neuroprotective effect on motoneurons occurs, at least at 90 d of age. In view of this finding, the increase in endocannabinoid levels in the spinal cord and brain of SOD1G93A mice observed here and by previous authors (26) might be regarded as an adaptive response aimed at counteracting excitotoxicity, similar to that observed in the hippocampus of mice treated with kainate (7) , or in the spinal cord of mice with chronic relapsing experimental allergic encephalomyelitis during the spasticity phase (12) . Such an endogenous response is clearly not sufficient to counteract the progress of disease, whereas genetic deletion of the Faah gene, which is known to bring about up to a 15-fold elevation of brain AEA levels (21) , results in a much stronger neuroprotective action. Pharmacological inhibition of the Faah enzyme has similarly been shown to have beneficial effects in autoimmune models of inflammation (12 , 51) , and recently a class of highly selective and reversible Faah inhibitors have been developed, which will further aid research in this field (52) . In the present study, despite the dramatic effects on disease signs observed in 90-day-old SOD1.Faah —/— mice, we found no increase in life span, which suggests that endocannabinoids might be therapeutic but only in the early stages of disease. However, ablation of Faah and the subsequent increase in endogenous AEA dramatically delay disease onset. Therefore, pharmacological inactivation of Faah might be particularly effective if used in combination with agents that act on mechanisms active during later stages of the disease such as hsp coinducers (53) or glutamate inhibitors (28 , 29) .
In conclusion, our results show that an increased tone of the endocannabinoid system, obtained either by treatment with synthetic cannabinoids such as WIN55,212—2 or by ablation of the Faah enzyme, ameliorates disease, at least in the short term. However, despite the significant amelioration in disease signs observed in SOD1G93A mice following elevation of cannabinoid levels, we found no effect on life span. Our results also suggest that the neuroprotective effects might be mediated by a non-CB1 receptor-dependent mechanism. The results of this study, therefore, identify the Faah enzyme and possibly the CB2 receptor as potential therapeutic targets in ALS.
Source with Visuals: http://www.fasebj.org/content/20/7/1003.full