Δ⁹-tetrahydrocannabinol (Δ⁹-THC) exerts a direct neuroprotective effect in a human cell culture model of Parkinson's disease

Abstract

C. B. Carroll, M.-L. Zeissler, C. O. Hanemann and J. P. Zajicek (2012) Neuropathology and Applied Neurobiology38, 535–547

Δ⁹-tetrahydrocannabinol (Δ⁹-THC) exerts a direct neuroprotective effect in a human cell culture model of Parkinson's disease

Aims:Δ⁹-tetrahydrocannabinol (Δ⁹-THC) is neuroprotective in models of Parkinson's disease (PD). Although CB1 receptors are increased within the basal ganglia of PD patients and animal models, current evidence suggests a role for CB1 receptor-independent mechanisms. Here, we utilized a human neuronal cell culture PD model to further investigate the protective properties of Δ⁹-THC. Methods: Differentiated SH-SY5Y neuroblastoma cells were exposed to PD-relevant toxins: 1-methyl-4-phenylpyridinium (MPP⁺), lactacystin and paraquat. Changes in CB1 receptor level were determined by quantitative polymerase chain reaction and Western blotting. Cannabinoids and modulatory compounds were co-administered with toxins for 48 h and the effects on cell death, viability, apoptosis and oxidative stress assessed. Results: We found CB1 receptor up-regulation in response to MPP⁺, lactacystin and paraquat and a protective effect of Δ⁹-THC against all three toxins. This neuroprotective effect was not reproduced by the CB1 receptor agonist WIN55,212-2 or blocked by the CB1 antagonist AM251. Furthermore, the antioxidants α-tocopherol and butylhydroxytoluene as well as the antioxidant cannabinoids, nabilone and cannabidiol were unable to elicit the same neuroprotection as Δ⁹-THC. However, the peroxisome proliferator-activated receptor-gamma (PPARγ) antagonist T0070907 dose-dependently blocked the neuroprotective, antioxidant and anti-apoptotic effects of Δ⁹-THC, while the PPARγ agonist pioglitazone resulted in protection from MPP⁺-induced neurotoxicity. Furthermore, Δ⁹-THC increased PPARγ expression in MPP⁺-treated SH-SY5Y cells, another indicator of PPARγ activation. Conclusions: We have demonstrated up-regulation of the CB1 receptor in direct response to neuronal injury in a human PD cell culture model, and a direct neuronal protective effect of Δ⁹-THC that may be mediated through PPARγ activation.

Introduction

Cannabinoids are a group of compounds present in cannabis (Cannabis sativa) and include Δ⁹-tetrahydrocannabinol (Δ⁹-THC) and cannabidiol 1, 2. There are two main cannabinoid receptor subtypes: CB1 (which is found primarily in the brain, particularly in the basal ganglia and in the limbic system) 3 and CB2 (which is primarily localized to cells of the immune system) 4. Cannabinoid receptors are G protein-coupled receptors which inhibit adenylate cyclase 5.

Parkinson's disease (PD) is a neurodegenerative condition characterized by loss of dopaminergic neurones in the substantia nigra pars compacta with resulting neurochemical imbalance throughout the basal ganglia 6. Cannabinoids modulate neurotransmitter release within the basal ganglia 7, 8 and have been demonstrated to result in symptomatic benefit in animal models 9, most likely mediated by the CB1 receptor. An increase in CB1 receptor level and efficacy of activation has been demonstrated in the striatum of PD patients and MPTP (1-methyl-4-phenylpyridinium iodide)-treated marmosets 10, most likely reflecting loss of dopaminergic inhibitory influences 7, 10-13 and neurochemical compensatory mechanisms within the basal ganglia. Although these changes in humans may have been due to dopaminergic therapy, an in vivo positron emission tomography study of CB1 receptor binding in PD patients demonstrated increased binding in the putamen even in untreated patients 13. It is, however, not known to what extent the up-regulation of the CB1 receptor seen in PD may reflect a direct response to neuronal damage.

There is increasing evidence that, in addition to symptomatic effects, cannabinoids may have neuroprotective properties which could be exploited for the treatment of neurodegenerative conditions including PD 14-16. Although these protective effects may be receptor-independent 9, 17, studies in a range of in vivo and in vitro excitotoxicity models have suggested that the neuroprotective effect of cannabinoids may be mediated via the CB1 receptor 18-24. A protective effect mediated via the CB1 receptor and increased sensitivity to insult in CB1 receptor-deficient mice have been demonstrated in models of multiple sclerosis and closed head injury 18, 19. A similar CB1 receptor-mediated protective effect has also been demonstrated in rat models of cerebral ischaemia 22. One mechanism by which this protective effect could be exerted is by the presynaptic inhibition of glutamate release and suppression of excitotoxic mechanisms 20, 23, 25. However, a study of excitotoxicity in mouse spinal cord suggested that CB1 receptor agonists may have an effect by acting directly on postsynaptic neuronal CB1 receptors 21. Importantly, the spinal cord cultures used in these experiments also contained glial cells, and later studies have provided evidence of CB1-mediated modulation of glial activity as being neuroprotective 15, 26. Nevertheless, a direct neuronal CB1-mediated protective effect has been demonstrated in a model of diabetic neuropathy 27.

A neuroprotective effect of cannabinoids has been demonstrated in both in vivo and in vitro PD models, which has been attributed to their antioxidant properties 16, CB2 receptor activation 15, 28 or glial CB1 receptor activation 26. Additionally, a key role for the CB1 receptor has been demonstrated in a mouse model of PD 29. CB1-mediated neuroprotective effects may therefore be direct neuronal effects, the result of modulation of neurotransmitter release or modulation of microglial activation.

We utilized a human cell culture model of PD that enabled us to determine direct neuronal effects, namely SH-SY5Y neuroblastoma cells, differentiated with retinoic acid and exposed to toxins that mimic the various abnormalities that have been implicated in the pathogenesis of PD: inhibition of mitochondrial function [1-methyl-4-phenylpyridinium (MPP⁺) iodide]30, 31, free radical generation (paraquat) 32-34 and inhibition of the ubiquitin proteasome system (UPS) (lactacystin) 35. The use of differentiated SH-SY5Y cells is well established as a cell culture model of PD 36. In this paper, we investigated whether up-regulation of the CB1 receptor can be induced as a direct response to neuronal injury and investigated the neuroprotective mechanisms of Δ⁹-THC.

Materials and methods

Materials

Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich Chemicals (Dorset, UK). WIN55,212-2, AM251 and cannabidiol were purchased from Tocris (Bristol, UK). Nabilone was a kind gift from MEDA Pharmaceuticals UK Ltd (Bishop's Stortford, UK).

Culture of neuroblastoma cells

Human neuroblastoma cells (SH-SY5Y) were obtained from European Collection of Cell Cultures (ECACC), transferred to 75-cm² filter vent flasks (VWR, Leicestershire, UK), grown in Dulbecco's modified Eagle's medium (Invitrogen, Paisley, UK) containing 10% (v/v) foetal bovine serum (FBS) (PAA, Yeovil, UK), glutamine, 4.5 g/l glucose, 1 ml uridine (25 mg/ml), 5 ml pyruvate, 25 units/ml penicillin and 25 µg/ml streptomycin, and incubated at 37°C in a humidified atmosphere of 5% CO₂ and 95% air. For experiments, cells were seeded in six-well dishes (VWR) (200 000 cells/well) or 96-well plates (VWR) (10 000 cells/well) and treated with 10 µM retinoic acid for 5 days to promote differentiation to a neuronal phenotype. Medium was changed every 48 h.

Cell treatments

After differentiation, toxins [MPP⁺, lactacystin (Merck, Nottingham, UK) and paraquat] were added to retinoic acid-supplemented cell culture medium (total 100 µl). Concentrations of toxin were chosen that resulted in about two to threefold cell death at 48 h compared with vehicle: 5–7 mM MPP⁺, 20 µM lactacystin and 500 µM paraquat. For quantitative polymerase chain reaction (QPCR) and protein determination experiments, cells were harvested at 6, 12, 24 and 48 h. To investigate their effects, cannabinoids (Δ⁹-THC, WIN55,212-2, cannabidiol and nabilone) and modulators (AM251, T0070907) were co-administered with toxins in retinoic acid-supplemented medium. To determine the neuroprotective potential of antioxidant compounds, differentiated SH-SY5Y cells were treated with a range of concentrations of the antioxidants butylated hydroxytoluene (BHT) and α-tocopherol in retinoic acid-supplemented medium in combination with MPP⁺. For pretreatment, BHT was applied for 1 h prior to co-application of MPP⁺ and BHT. All plates contained positive (toxin or agent alone) and vehicle controls. Cell death was assessed after 48 h.

Lactate dehydrogenase (LDH)-release assay

To assess the cytotoxicity of the toxins under our experimental conditions, LDH release of cells grown and treated in 96-well plates was measured. Cell culture medium (50 µl) was used to analyse the LDH activity by measuring the oxidation of nicotinamide adenine dinucleotide hydride (NADH) at 450 nm as described in the manufacturer's protocol (Promega, Southampton, UK). The remaining cells were lysed and LDH activity similarly measured to allow the percentage of cell death relative to the total number of cells to be calculated.

MTT assay

Cells were treated in 96-well plates as described above. Ten microlitre of 5 mg/ml sterile-filtered 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Invitrogen, Paisley, UK) solution in serum-free medium was added to each well and incubated for 4 h. The medium was removed and 100 µl of dimethyl sulphoxide (DMSO) per well added to dissolve the formazan precipitate. Plates were incubated on a shaker for 10 min and the absorbance read at 562 nm using 650 nm as reference wavelength.

Measurement of reactive oxygen species

SH-SY5Y cells were seeded into 96-well plates and treated for 48 h as described above. The medium was removed and cells were incubated with 10 µM 2′,7′-Dichlorodihydrofluorescein diacetate (DCFDA) in serum-free medium for 30 min. Cells were then washed three times with phosphate-buffered saline (PBS) after which the fluorescence was measured at Ex: 485 nm and Em: 535 nm. To determine cell number, cells were lysed at −80°C and cell number estimated using the LDH assay. Experiments were carried out four times in triplicate.

Total RNA extraction and PCR

Total RNA was extracted using GenElute (Sigma-Aldrich) and treated with DNA-free (Ambion, Huntingdon, UK) according to the manufacturer's instructions and reverse transcribed in a 50-µl reaction using cDNA Archive Kit (Ambion). To determine whether differentiated SH-SY5Y cells expressed cannabinoid receptors, we used the following PCR primers: CB1 – forward 5′-aggggatgcgaagggatt-3′, reverse 5′-agtggtgatggtgcggaag-3′ giving an amplified fragment of 131 bp; CB2 – forward 5′-tcaaccctgtcatctatgctc-3′, reverse 5′-agtcagtcccaacactcatc-3′ giving an amplified fragment of 353 bp. Thermocycling was carried out as follows: 95°C for 10 min, 40 cycles of 95°C for 30 s, 54°C for 30 s, 72°C for 1 min, followed by 72°C for 5 min. QPCR was employed to detect changes of CB1 receptor mRNA. (FAM)-labelled CB1 receptor primers were purchased from Applied Biosystems (Warrington, Cheshire, UK) (Hs 01038522). Reactions were carried out on a Bio-Rad ICycler (Bio-Rad, Hemel Hempstead, UK). Each reaction was run in triplicate with 1-µl sample in a total volume of 20 µl. Amplification and detection were performed with the following conditions: an initial hold at 95°C for 10 s followed by 50 cycles at 95°C for 15 s and 60°C for 60 s. Gene expression was normalized to 18S expression run in triplicate concurrently. All samples were analysed in triplicate on three samples from four separate experiments. Statistical significance was determined using one-way analysis of variance (anova) in SPSS with Tukey honestly significant difference (HSD) post hoc test.

Protein extraction and Western blot analysis

Cells were lysed for protein extraction at 6, 12, 24 and 48 h following exposure to the toxins. Cells were washed with ice-cold PBS and protein extracted with NET-Triton buffer [150 mM NaCl, 5 mM EDTA (ethylenediaminetetraacetic acid), 10 mM Tris (tris(hydroxymethyl)aminomethane), pH 7.4, 1% Triton X-100] supplemented with protease inhibitor cocktail (Sigma-Aldrich) according to the manufacturer's instructions.

The proteins were resolved by SDS/PAGE (10% gels) and blotted onto polyvinylidene fluoride (PVDF) membranes. Membranes were washed with tris-buffered saline (140 mM NaCl, 50 mM tris/HCl, pH 7.2) containing 0.1% Tween 20, 5% skimmed milk and 2% bovine serum albumin to block nonspecific protein binding. Membranes were incubated with primary antibody against CB1 receptor (1:200 CB1 1A, Alpha Diagnostic International, San Antonio, TX, USA) (approx. 60 kDa), CB2 receptor (1:200, #101550, Cayman Chemical, Ann Arbor, MI, US) (45 kDa) and cleaved caspase 3 (1:1000 #9664S, New England Biolabs) in tris-buffered saline (140 mM NaCl, 50 mM tris/HCl, pH 7.2) containing 0.1% Tween 20, 5% skimmed milk and 2% bovine serum albumin overnight at 4°C, washed three times and incubated with horseradish peroxidase-conjugated secondary antibody (Bio-Rad) for 1 h at room temperature. The protein bands were detected using the enhanced chemiluminescence method (Amersham Biosciences, Buckinghamshire, UK). Membranes were probed with β-actin (1:5000; Abcam, Cambridge, UK) to control for loading. Samples were analysed from at least four separate experiments. Statistical significance was determined using one-way anova in SPSS with Tukey HSD post hoc test.

Results

MPP⁺

MPP⁺ at 5 mM resulted in a greater than twofold increase in mRNA and protein levels for the CB1 receptor at 24-h and 48-h exposure (P < 0.05) (Figure 1a). A protective effect of 10 µM Δ⁹-THC against MPP⁺ was demonstrated with the LDH release assay (P < 0.001) and confirmed using the MTT assay (P < 0.005) (Figure 2b,c) and consequently 10 µM Δ⁹-THC was used for all future experiments. The LDH assay was chosen for the following experiments as calibration curves indicated a stronger linear relationship between optical density readings and cell number (Figure S1). Δ⁹-THC protected against the toxic effects of MPP⁺ (P < 0.01) (Figure 2d), an effect that was not blocked by co-administration of the selective CB1 receptor antagonist AM251 (1 µM). Additionally, the CB1 agonist, WIN55,212-2 (0.1–10 µM; data for 1 µM shown in Figure 2d) did not have the same protective properties. These results suggest that the protective effects of Δ⁹-THC against MPP⁺ are not mediated by the CB1 receptor. Δ⁹-THC is also an agonist at the CB2 cannabinoid receptor. However, we were unable to detect expression of CB2 in our differentiated SH-SY5Y cells, either by reverse transcription PCR (RT-PCR) or Western blotting (Figure 3a,b).

Figure 1

Cannabinoid CB1 receptor expression is increased in response to toxin exposure. Effect of exposure of differentiated SH-SY5Y cells to 5 mM 1-methyl-4-phenylpyridinium (MPP⁺) (a), 500 µM paraquat (b) and 20 µM lactacystin (c) on CB1 receptor mRNA and protein levels displayed graphically as mean ± SEM (*P < 0.05, **P < 0.001), with images of representative Western blots below (CB1 above and actin as loading control below). Protein data are normalized to vehicle-treated samples and corrected to the loading control actin. QPCR, quantitative polymerase chain reaction.

Figure 2

Δ⁹-THC exerts a protective effect against the toxins. (a) Microscopy image demonstrating protective effect of 10 µM Δ⁹-tetrahydrocannabinol (Δ⁹-THC) against 7 mM 1-methyl-4-phenylpyridinium (MPP⁺) in a representative field of view. (b) Lactate dehydrogenase (LDH) assay showing the effect of increasing concentrations of Δ⁹-THC on MPP⁺ toxicity [one-way anova with Tukey honestly significant difference (HSD) post hoc test: **P < 0.001 vs. MPP⁺]. (c) MTT assay showing the effect of increasing concentrations of Δ⁹-THC on MPP⁺ toxicity (one-way anova with Tukey HSD post hoc test: *P < 0.05, **P < 0.005 vs. MPP⁺). (d) Co-administration of 10 µM Δ⁹-THC to differentiated SH-SY5Y cells resulted in significantly reduced cell death in response to 5 mM MPP⁺, 500 µM paraquat (PQT) and 20 µM lactacystin (Lact). This protective effect could not be blocked by co-administration of 1 µM AM251, the specific CB1 antagonist, nor could it be reproduced by co-administration of the CB1 agonist 1 µM WIN55,212-2 (Win) with the toxins. Each bar represents the mean ± SEM of quadruplicate measurements from at least four independent experiments (one-way anova with Tukey HSD post hoc test compared with toxin alone: *P < 0.05, **P < 0.01).

Figure 3

Differentiated SH-SY5Y cells express CB1 but not CB2 receptors. Expression of CB1 (a) and CB2 (b) in differentiated SH-SY5Y cells. PCR on cDNA demonstrated CB1 expression (human placenta positive control) (a) but not CB2 (human testis positive control) (b). Western blot showed expression of CB1 (a) but not CB2 (b) with human glioma U373MG cell line as positive control.

Paraquat and lactacystin

Although MPP⁺ is widely employed to model PD both in vivo and in vitro, other cellular pathways are also implicated in the pathogenesis of the disease, including free radical damage and dysfunction of the UPS. We were interested to see whether cellular models involving these pathways also resulted in up-regulation of the CB1 receptor, and whether Δ⁹-THC would be protective.

The free radical-generating toxin, paraquat (500 µM), resulted in a modest 1.3-fold increase in CB1 receptor mRNA level and a significant twofold increase in protein level at 24 h (Figure 1b). Lactacystin (20 µM), which is a potent and specific inhibitor of the UPS, resulted in an almost eightfold increase in CB1 receptor mRNA level (P < 0.001) within 24 h, but this had returned to baseline by 48 h (Figure 1c). This was associated with a significant increase in CB1 protein level. Δ⁹-THC was protective against both toxins, and similar to the effect against MPP⁺; this protective effect was not blocked by AM251 or reproduced by WIN55,212-2 (Figure 2d).

It is possible that the protective effects of Δ⁹-THC are mediated by its antioxidant properties conferred by the presence of a phenolic ring. We therefore investigated the effect of the synthetic cannabinoid, nabilone, whose structure also contains a phenolic ring and which is known to have antioxidant properties 37, on the toxic effects of MPP⁺. Concentrations of 20 µM and higher were toxic to the cells (Figure 4a). However, lower concentrations failed to demonstrate a protective effect (Figure 4b) against MPP⁺, while 10 µM potentiated the toxicity of MPP⁺. Cannabidiol, another antioxidant phytocannabinoid, was similarly ineffective (Figure 4c). In addition, co-application of the antioxidants BHT or α-tocopherol failed to reproduce the protective effect of Δ⁹-THC (Figure 5). In contrast, pretreatment with BHT was protective.

Figure 4

The antioxidant cannabinoids, nabilone and cannabidiol, fail to exert a protective effect. Effect of nabilone on differentiated SH-SY5Y cells (a). Co-application of nabilone failed to protect against 7 mM 1-methyl-4-phenylpyridinium (MPP⁺) toxicity (b). Cannabidiol did not protect against 7 mM MPP⁺ (c). Each bar represents the mean ± SEM of quadruplicate measurements from four independent experiments (one-way anova with Tukey honestly significant difference post hoc test: ***P < 0.001).

Figure 5

Percentage protection afforded by antioxidants against 7 mM 1-methyl-4-phenylpyridinium (MPP⁺) in differentiated SH-SY5Y cells compared with that of Δ⁹-tetrahydrocannabinol (Δ⁹-THC). Each bar represents the mean ± SEM of quadruplicate measurements from at least six independent experiments (one-way anova with Tukey honestly significant difference post hoc test, comparison with protective effect of Δ⁹-THC: **P < 0.01, ***P < 0.001).

It is known that cannabinoids including Δ⁹-THC are able to bind to receptors other than CB1 and CB2. To investigate whether Δ⁹-THC mediates its neuroprotective effect through peroxisome proliferator-activated receptor-gamma (PPARγ) activation, we examined the effects of the specific PPARγ antagonist, T0070907. T0070907 was not toxic to SH-SY5Y cells at concentrations between 1 µM and 10 µM (Figure 6a); this range of concentrations also had no significant effect on MPP⁺-induced neurotoxicity (Figure 6b). We therefore selected this concentration range for further investigation and found that T0070907 was able to dose-dependently inhibit the protective effect of Δ⁹-THC (Figure 6c). Furthermore, the specific PPARγ agonist pioglitazone resulted in significant neuroprotection against MPP⁺ toxicity (Figure 7a) which was completely reversed by 5 µM T0070907 (Figure 7b). A protective effect of Δ⁹-THC blocked by T0070907 was further confirmed by Western blotting of activated caspase 3 cleaved at Asp 175, a marker of apoptosis. Addition of 10 µM Δ⁹-THC or 5 µM pioglitazone significantly reduced cleaved caspase 3 levels compared to MPP⁺ (P < 0.005), and this reduction was inhibited by 10 µM and 5 µM of the PPARγ antagonist, T0070907 for Δ⁹-THC and pioglitazone respectively (Figure 8). We also found that Δ⁹-THC treatment led to a significant reduction (P < 0.0001) in reactive oxygen species (ROS) production, an effect that was blocked by 10 µM T0070907 (Figure 9). In addition, we found that Δ⁹-THC treatment induced up-regulation in PPARγ expression at the protein level (Figure 10). Thus, we show that the neuroprotective properties of Δ⁹-THC may be mediated via PPARγ through which it may exert its antioxidant effect.

Figure 6

The effect of the peroxisome proliferator-activated receptor-gamma (PPARγ) antagonist, T0070907. T0070907 was toxic at concentrations higher than 20 µM (a) [one-way anova with Tukey honestly significant difference (HSD) post hoc test: *P < 0.05, **P < 0.001 vs. vehicle]. T0070907 had no effect on 7 mM 1-methyl-4-phenylpyridinium (MPP⁺) toxicity (b) (one-way anova with Tukey HSD post hoc test: non-significant vs. MPP⁺). T0070907 dose-dependently inhibited Δ⁹-tetrahydrocannabinol (Δ⁹-THC)-mediated neuroprotection against 7 mM MPP⁺ (c) (one-way anova with Tukey HSD post hoc test: *P < 0.05, **P < 0.001 vs. MPP⁺). Each bar represents the mean ± SEM of quadruplicate measurements from at least three independent experiments.

Figure 7

Pioglitazone was protective against MPP+. The peroxisome proliferator-activated receptor-gamma (PPARγ) agonist pioglitazone (Pio) was significantly protective against 7 mM 1-methyl-4-phenylpyridinium (MPP⁺)-induced neurotoxicity at a concentration of 5 µM (a) [one-way anova with Tukey honestly significant difference (HSD) post hoc test: **P < 0.001 vs. MPP⁺]. This protective effect could be blocked dose-dependently by T0070907 (b) (one-way anova with Tukey HSD post hoc test: **P < 0.001 vs. MPP⁺). Each bar represents the mean ± SEM of quadruplicate measurements from at least three independent experiments.

Figure 8

1-Methyl-4-phenylpyridinium (MPP⁺) (7 mM)-induced apoptosis was inhibited by both Δ⁹-tetrahydrocannabinol (Δ⁹-THC) and pioglitazone (Pio), an effect that was reversed by T0070907 (T) (10 µM and 5 µM respectively) as indicated by protein levels of cleaved caspase 3 (Asp175) (5A1E). Protein data are corrected to the loading control. Each bar represents the mean ± SEM of triplicate measurements from at least four independent experiments (one-way anova with Tukey honestly significant difference post hoc test: *P < 0.005 vs. MPP⁺).

Figure 9

Δ⁹-tetrahydrocannabinol (Δ⁹-THC) significantly reduced reactive oxygen species (ROS) production following 7 mM 1-methyl-4-phenylpyridinium (MPP⁺) administration; this effect was inhibited by the peroxisome proliferator-activated receptor-gamma (PPARγ) inhibitor, T0070907 (10 µM) (one-way anova with Tukey honestly significant difference post hoc: ***P < 0.0001 vs. MPP⁺). Each bar represents the mean ± SEM of four independent experiments in triplicate.

Figure 10

Expression of peroxisome proliferator-activated receptor-gamma (PPARγ) in differentiated SH-SY5Y cells. Cells were treated with 7 mM 1-methyl-4-phenylpyridinium (MPP⁺) and 7 mM MPP⁺ + 10 µM tetrahydrocannabinol (THC) for 48 h after differentiation (one-way anova with Tukey honestly significant difference post hoc test: *P < 0.05). Each bar represents the mean ± SEM of three independent experiments.

Discussion

We have demonstrated up-regulation of CB1 receptor mRNA in a human neuronal cell culture model of PD in direct response to an inhibitor of mitochondrial complex I (MPP⁺) and an inhibitor of the UPS (lactacystin), both mechanisms relevant to PD pathogenesis. There was a modest increase following 24-h exposure to paraquat, a free-radical generator, but this failed to reach significance. These changes were accompanied by significant increases in CB1 protein expression in response to all three toxins. There have been no previous reports of alteration in neuronal CB1 mRNA and protein level in direct response to toxin administration in cell culture models of PD. Interestingly, similar changes in CB1 receptor level have not been consistently found in the substantia nigra of PD patients or animal models 10, 12, 13, although one study has demonstrated increased nigral CB1 receptor activity 10. The failure to demonstrate an increase in CB1 mRNA in the nigra may reflect loss of dopaminergic neurones. Nevertheless, our finding suggests that up-regulation of the CB1 receptor may occur as a direct response to neuronal injury caused by the disease process. The mechanisms underlying this increased transcription remain unclear. It is known that sequences flanking exon 1 of the CB1 gene have marked promoter activity 38, 39, and it is possible that transcription factors produced in response to toxin exposure result in increased gene transcription.

Given the up-regulation of the CB1 receptor in our cell culture model of PD, we hypothesized that agonists at the CB1 receptor would have neuroprotective effects. We found that 10 µM Δ⁹-THC was protective against all three toxins tested – MPP⁺, lactacystin and paraquat – showing a direct neuroprotective effect of Δ⁹-THC in a human cell culture PD model. However, this protective effect was not blocked by the CB1 receptor antagonist AM251 and was not reproduced by the CB1 receptor agonist WIN55,212-2, suggesting that the protective effect of Δ⁹-THC is unlikely to be mediated by the CB1 receptor.

It has previously been shown that undifferentiated SH-SY5Y cells express the CB2 cannabinoid receptor 40 which may be an alternative site of action of Δ⁹-THC. However, in our differentiated cell culture model, we were unable to detect the presence of CB2 receptors either by Western blotting or RT-PCR. The lack of effect of WIN55,212-2, which is also a CB2 receptor agonist, makes it further unlikely that the CB2 receptor is the site of action mediating the protective effect, although future work may include specifically investigating the effects of CB2 receptor antagonists. We therefore investigated the PPARγ as a potential site of action of Δ⁹-THC in our model. PPARγ is a nuclear receptor which, upon ligand binding, heterodimerizes with the retinoic X receptor to initiate transcription of genes featuring a specific PPAR response element within their promoter region 41. There are three PPARγ variants all of which can be found in the brain and peripheral nervous system 42, 43. Studies in vascular endothelial cells showed that Δ⁹-THC is able to initiate PPARγ transcriptional activity, making the receptor a potential target for Δ⁹-THC 44-47. Whether Δ⁹-THC is able to directly bind the receptor or modulate the concentrations of endogenous ligands through indirect pathways is to date unclear 44.

Interestingly, activation of PPARγ by rosiglitazone and pioglitazone has been found protective in both animal and cell culture models of PD 48-51. Most importantly, however, a protective effect of the PPARγ agonist rosiglitazone against MPP⁺-induced neurotoxicity has been demonstrated in SH-SY5Y cells 49. However, reasons for this neuroprotective effect are not known. Here, we present the first evidence of a PPARγ-mediated protective effect of Δ⁹-THC in MPP⁺-treated SHSY-5Y cells, as the protective effects of Δ⁹-THC on cell death, apoptosis and ROS production could be blocked by the specific PPARγ antagonist, T0070907, and reproduced by the PPARγ agonist, pioglitazone. Furthermore, we detected significant up-regulation of PPARγ expression in response to Δ⁹-THC treatment, a finding indicative of PPARγ activation as the receptor has previously been shown to induce its own expression upon activation in a positive feedback response 52. These data support PPARγ as a potential site of action of Δ⁹-THC. The synthetic CB1/CB2 receptor agonist WIN55,212-2 has also been reported to induce PPARγ expression in hepatoma HepG2 cells leading to induction of apoptosis 53. However, in that study the increase in PPARγ expression could partly be blocked by the CB2 antagonist AM630 and the cholesterol depletor methyl-b-cyclodextrin. CB1 receptor-mediated stimulation of PPARγ expression by WIN55,212-2 has also been demonstrated in early, but not fully, differentiated adipocytes 54, indicating that this response may be highly cell type-specific, which may explain why WIN55,212-2 was not protective in our model.

Previous evidence supports antioxidant properties underlying the neuroprotective effects of Δ⁹-THC and other cannabinoids, possibly conferred by the presence of a phenolic ring 15, 16, 37. Although this hypothesis would be supported by our finding of a lack of protective effect of WIN55,212-2, which does not contain a phenolic ring, the phytocannabinoid cannabidiol and the synthetic cannabinoid and Δ⁹-THC analogue, nabilone, which do contain a phenol ring, were not protective, despite previous reports of their antioxidant properties 37. In addition, we were unable to reproduce the same protective effect with co-application of the antioxidants α-tocopherol or BHT. We therefore measured ROS production within our cells and confirmed the antioxidant capacity of Δ⁹-THC. However, rather than being mediated through its structural properties, we provide evidence that the antioxidant effects of Δ⁹-THC are mediated by its interaction with PPARγ. Possible PPARγ-mediated antioxidant effects include down-regulation of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase expression 55 and enhanced activity of superoxide dismutase (SOD) and catalase 49 as well as increased expression of peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC1-α) 56, 57 which up-regulates oxidative stress response pathways such as nuclear respiratory factor 1 (NRF-1) and nuclear factor erythroid 2-related factor 2 (NRF-2) and has been implicated in mitochondrial biogenesis 58. This is the subject of ongoing investigation.

Conclusions

In conclusion, we have demonstrated up-regulation of the CB1 receptor in a human cell culture model of PD, as well as a direct neuroprotective effect of the phytocannabinoid, Δ⁹-THC, not mediated by the CB2 receptor. Although a CB1 receptor-mediated effect cannot totally be excluded, we propose that activation of PPARγ leading to antioxidant effects is highly relevant in mediating the neuroprotection afforded by Δ⁹-THC in our model.

Acknowledgements

Dr Carroll was an MRC post-doctoral fellow at the time of undertaking this work.

Conflict of interest

The authors have no conflict of interest.