2025 Volume 48 Issue 6 Pages 769-781
Cannabinoid receptor type 1 (CB1R) plays a key role in neuronal homeostasis, synaptic plasticity, and neuroprotection. CB1R antagonists typically protect against CB1R agonists-induced neurotoxicity. However, we previously found that the CB1R antagonists rimonabant and its analog AM251 can also be neurotoxic: under serum-free conditions, these compounds induce apoptosis in human neuroblastoma SH-SY5Y cells through mitochondrial damage and endoplasmic reticulum (ER) stress. To elucidate the mechanisms of this neurotoxicity, we examined the effects of CB1R agonists. We co-treated SH-SY5Y cells with rimonabant or AM251 in combination with either the CB1R agonist arachidonyl 2-chloroethylamide (ACEA) or WIN 55212-2 mesylate (WIN). ACEA, but not WIN, protected cells from rimonabant- and AM251-induced apoptosis. While ACEA had only a limited effect on mitochondrial damage, it significantly reduced phosphorylation of the eukaryotic initiation factor 2 alpha (eIF2α), a key marker of ER stress. Given that ACEA also functions as an agonist of transient receptor potential vanilloid 1 (TRPV1), we investigated its role in ACEA-mediated neuroprotection. The TRPV1 antagonist capsazepine blocked ACEA’s protective effects, suggesting that ACEA acts through TRPV1 rather than CB1R. ACEA also prevented apoptosis induced by camptothecin, a well-established apoptosis inducer, through a similar capsazepine-sensitive mechanism, demonstrating its broader protective effects against apoptosis. These findings indicate that rimonabant and AM251 induce neurotoxicity independently of CB1R under serum-free conditions and that ER stress is likely to be a key target of CB1R-independent neuroprotection by ACEA. Our study highlights the complexity of CB1R ligand-associated neurotoxicity and neuroprotection.
Cannabinoid receptor type 1 (CB1R) is a critical component of the endocannabinoid system, playing a fundamental role in maintaining neuronal homeostasis and regulates brain function.1,2) As one of the most abundant G-protein-coupled receptors in the central nervous system, CB1R is essential for synaptic plasticity, cognition, and neuroprotection.3,4)
CB1R signaling has been widely recognized for its neuroprotective effects against various forms of neurotoxicity.5–12) This protective role is evidenced by studies showing that CB1R-deficient mice exhibit increased sensitivity to neurotoxicity caused by ischemia or excitotoxicity.8,9) Furthermore, CB1R agonists demonstrate neuroprotective effects in both in vitro and in vivo models of neurotoxicity.5–7) CB1R activity also plays a critical role in mitigating age-related cognitive decline and in preserving neuronal physiology.10–12)
Recent findings indicate that CB1R is not only present on the plasma membrane but also localized on the outer membrane of mitochondria in neuronal cells.13,14) This mitochondrial CB1R (mtCB1R) has been shown to prevent mitochondrial damage in ischemic conditions,15–17) and to play a role in mitochondrial quality control, as CB1R deletion accelerates aging by disrupting the mitochondrial quality control system in hippocampal neurons.12) These studies highlight the importance of mtCB1R in CB1R-mediated neuroprotection.
While CB1R antagonists are generally considered protective against neurotoxicity, they can be neurotoxic under particular conditions.18–21) Our previous study revealed that CB1R antagonists rimonabant and AM251 induce caspase-dependent apoptosis in CB1R-expressing human neuroblastoma SH-SY5Y cells under serum-free conditions.21–24) This neurotoxicity was associated with mitochondrial damage and endoplasmic reticulum (ER) stress. Since environmental stress, including nutrient deprivation, alters cellular conditions and affects sensitivity to drugs,25,26) it is possible that adaptation to stress modifies CB1R signaling, leading to apoptosis when CB1R activity is inhibited. However, it remains unclear whether this neurotoxicity is mediated by CB1R antagonism or alternative mechanisms.
To elucidate the mechanisms underlying this neurotoxicity, we examined whether CB1R agonists could protect against CB1R antagonist-induced apoptosis. Our findings reveal that the CB1R agonist arachidonyl 2-chloroethylamide (ACEA) protected SH-SY5Y cells from apoptosis in a dose-dependent manner, while the CB1R agonist WIN 55212-2 mesylate (WIN) does not confer protection. Notably, ACEA appears to suppress ER stress rather than preventing mitochondrial damage. Furthermore, the protective effect of ACEA is antagonized by the transient receptor potential vanilloid 1 (TRPV1) antagonist capsazepine. ACEA also resisted camptothecin-induced apoptosis, and this effect is similarly by capsazepine. These results suggest that ACEA inhibits apoptosis at least via ER stress suppression in a CB1R-independent manner, implying that the neurotoxicity induced by rimonabant and AM251 is likely independent of CB1R inhibition.
Rimonabant, AM251, capsaicin, capsazepine, and erastin were purchased from Selleck Chemicals (TX, U.S.A.). ACEA was purchased from Tocris Bioscience (MN, U.S.A.). Carbonyl cyanide 3-chlorophenilhydrazone (CCCP), camptothecin, and WIN were purchased from FUJIFILM Wako (Osaka, Japan). Rimonabant, AM251, capsazepine, and WIN were dissolved in dimethyl sulfoxide (DMSO) to 10 mM. Erastin, and CCCP were dissolved in DMSO to 20, and 100 mM, respectively. ACEA was dissolved in EtOH to 10 mM.
Cell Lines and CultureHuman neuroblastoma SH-SY5Y cell (RRID: CVCL_0019) was obtained from KAC (Kyoto, Japan). SH-SY5Y cells were cultured in a growth medium (Dulbecco’s modified Eagle’s medium with Low Glucose (DMEM, FUJIFILM Wako)) with 10% fetal bovine serum (Cosmo Bio, Tokyo, Japan) and 1% penicillin–streptomycin (PS, FUJIFILM Wako) at 37 °C and 5% CO2.
Live/Dead AssaySH-SY5Y cells were seeded at the density of 1.0 × 104 cells/well in 100 µL of growth medium on 96-well clear plate. Three wells were used per condition in an independent experiment. Serum starvation was conducted overnight 24 h after seeding using serum-free medium (DMEM with 1% PS). Drug treatments were applied in a serum-free medium at 37 °C and 5% CO2 for 12 h. Cell viability was assessed using the LIVE/DEAD™ Viability/Cytotoxicity Kit (Thermo Fisher Scientific, MA, U.S.A.) as previously described.21) Fluorescent imaging and automated cell counting were performed using ImageJ Software (NIH, MD, U.S.A.) with the same macro configuration as described before.21)
Immunofluorescence Staining of Mitochondrial MorphologySH-SY5Y cells were seeded at 5.0 × 104 cells/well in 500 µL of growth medium on cover glasses. As described in the previous section, the cells underwent serum starvation and were treated with drugs in serum-free medium for 1 h. Immunofluorescence staining was conducted according to our previous work.21,27) Briefly, the adhered cells were fixed with 4% paraformaldehyde phosphate buffer solution (FUJIFILM Wako) for 20 min at room temperature. The fixed cells were treated with 50 mM ammonium chloride for 15 min to quench autofluorescence and then permeabilized with 0.2% Triton X-100 (Nacalai Tesque, Kyoto, Japan) in phosphate-buffered saline (PBS) for 3 min at room temperature. Subsequently, the cells were blocked with 3% bovine serum albumin (FUJIFILM Wako) in PBS for 1 h at room temperature. Immunostaining was performed using the primary antibody Tom20 FL-145 (1 : 500, Cat#sc-11415, Santa Cruz Biotechnology, TX, U.S.A.) and the Alexa Fluor™ 488 secondary antibody (1 : 1000, Cat#A-11008, Thermo Fisher Scientific,.). Images were acquired using an FV1200 laser scanning microscope (Olympus, Tokyo, Japan).
Mitochondrial Morphology Analysis in Immunofluorescence ImagesMitochondrial morphology was analyzed using ImageJ Software (NIH) according to our previous work,21) which followed established protocols.28,29) Briefly, pre-processing (Unsharp Mask; Enhance Local Contrast; Median) was conducted on images. The pre-processed images were binarized using Make Binary. Subsequently, post-processing (Despeckle; Remove Outliners) was conducted on the binarized images. Binarized mitochondria in the processed images were analyzed using Analyze Particles for mitochondrial circularity. Skeletonize and Analyze Skeleton (2D/3D) were applied to determine the average branch lengths of mitochondria. These processes were automated using a macro.
Preparation of Cell Culture Specimen for Scanning Electron Microscope (SEM)SH-SY5Y cells were seeded at 4.0 × 104 cells in 400 µL of growth medium on a µ-Dish 35 mm Grid-500 (Cat#81166, ibid.i GmbH, Gräfelfing, Germany). After serum starvation, the cells were treated with drugs in a serum-free medium for 1 h. Cells were then washed with PBS and fixed using pre-warmed fixative buffer (0.1 M cacodylate buffer, 100 mM NaCl, 2 mM CaCl2, pH 7.4, 2% paraformaldehyde (Nacalai Tesque), and 1% glutaraldehyde (TAAB, Berkshire, England)) for 15 min at room temperature. Fixed specimens were stored at 4 °C until postfixation, using the osmium tetroxide-tannic acid-osmium tetroxide (OTO) method, as described in our previous studies.21,30) Briefly, the specimens were rinsed five times with chilled 1 mM cacodylate buffer, followed by postfixation with 1% osmium tetroxide and 1.5% potassium ferrocyanide in 1 mM cacodylate buffer for 1 h on ice. After washing with distilled water, the specimens were treated with 1% thiocarbohydrazide for 1 h at room temperature, then washed and further treated with 1% osmium tetroxide for 30 min on ice. Following three washes with distilled water, the specimens were dehydrated in an ethanol gradient (50, 70, and 90% on ice, and 100% at room temperature). The specimens were infiltrated with epoxy resin EPON812 (TAAB) and polymerized at 65 °C for 2 d. The resin blocks were trimmed to 0.25 mm2, and serial sections were cut with an ultramicrotome (ARTOS 3D, Leica, Wetzlar, Germany) with a diamond knife (Diatome, Biel, Switzerland). Serial sections were mounted on a silicon wafer, dried at 60 °C, and stained with uranyl acetate and lead acetate.
Mitochondrial Morphology Analysis in SEM ImagesElectron micrographs of the cultured cells were obtained from ultrathin sections as backscattered electron images using a SEM (JSM-IT800; JEOL, Tokyo, Japan). Mitochondrial morphological characteristics, including circularity and cross-sectional area, were measured using ImageJ software (NIH).
Sample Preparation for Western BlottingSH-SY5Y cells were seeded at 2.0 × 106 cells/dish density in 4.0 mL of growth medium on a 60 mm dish. As described previously, the cells underwent serum starvation and were treated with reagents in serum-free medium for 3 h. After washing with PBS, the cells were lysed using RIPA buffer (FUJIFILM Wako) containing 1% protease inhibitor cocktail set I (FUJIFILM Wako), 1% phosphatase inhibitor cocktail 2 (Sigma-Aldrich, St. Louis, MO, U.S.A.), and 1% phosphatase inhibitor cocktail 3 (Sigma-Aldrich), following the protocol from our previous study.21) The lysates were sonicated and centrifuged at 16000 × g for 15 min at 4 °C. Supernatants were mixed with an equal volume of 2X SDS sample buffer (125 mM Tris–HCl, pH 6.8, 4% sodium dodecyl sulfate (SDS), 10% sucrose, 0.01% Bromophenol Blue, 1% 2-mercaptoethanol). Samples were reduced by heating at 95 °C for 5 min.
Western BlottingA total of 10 µg of each sample was loaded onto SDS polyacrylamide gels. Following electrophoresis, proteins were transferred to a PVDF membrane (FUJIFILM Wako). Blocking was performed using TBST (20 mM Tris–HCl, pH 7.6, 150 mM NaCl, 1% Tween 20 (FUJIFILM Wako)) containing 5% skim milk (FUJIFILM Wako) for 1 h at room temperature under gentle shaking. Membranes were incubated with eukaryotic translation initiation factor 2 α subunit (eIF2α) antibody (1 : 1500, Cat#11170-1-AP, Proteintech, IL, U.S.A.), phosho-eIF2α (peIF2α; Ser51) antibody (1 : 1000, Cat#28740-1-AP, Proteintech), activating transcription factor 4 (ATF4) antibody (1 : 1500, Cat#10835-1-AP, Proteintech), C/EBP homologous protein (CHOP) antibody (1 : 1500, Cat#15204-1-AP, Proteintech), or β actin antibody (1 : 10000, Cat#010-27841, FUJIFILM Wako) in Can Get Signal® Immunoreaction Enhancer Solution 1 (Toyobo, Osaka, Japan) for 1 h at room temperature. Afterward, membranes were incubated with anti-rabbit secondary HRP antibody (1 : 5000, Cat#7074, Cell Signaling Technology, MA, U.S.A.) or anti-mouse secondary HRP antibody (1 : 5000, Cat#7076, Cell Signaling Technology) in Can Get Signal® Immunoreaction Enhancer Solution 2 (Toyobo) under the same conditions. Chemiluminescence was detected using ImmunoStar LD (FUJIFILM Wako) for eIF2α, peIF2α, ATF4, and CHOP antibodies, and Western BLoT (TaKaRa Bio, Shiga, Japan) for β actin antibody. Images were acquired with the ChemiDoc Touch MP Imaging System (Bio-Rad, CA, U.S.A.). After blocking, and both primary and secondary antibody incubations, membranes were washed three times with TBST (10 min each) at room temperature. As described in previous studies,21,31) membranes were incubated with 10% acetic acid to inactivate HRP chemiluminescence, allowing sequential Western blotting with β actin antibody.
Data and Statistical AnalysisKruskal–Wallis ANOVA with Dunn’s multiple comparison test, paired t test, and two-way ANOVA with uncorrected Fisher’s LSD test were employed as statistical analysis. Statistical analysis was performed using GraphPad Prism 10.2.2–10.3.1 (CA, U.S.A.).
To assess the potential neuroprotective effects of CB1R agonists, we first evaluated the effects of CB1R agonists, ACEA and WIN, on cell viability in human neuroblastoma SH-SY5Y cells, which endogenously express CB1R.22–24) We treated SH-SY5Y cells with these agonists for 12 h in serum-free conditions, under the same experimental conditions previously used to demonstrate CB1R antagonist-induced neurotoxicity.21) ACEA treatment did not affect cell viability at any dose tested (Supplementary Fig. S1a). In contrast, WIN reduced cell viability by 45% at 10 µM (Supplementary Fig. S1b), suggesting that WIN, but not ACEA, induces cell death in SH-SY5Y cells under serum-free conditions.
We next investigated whether CB1R agonists could counteract the cytotoxicity induced by rimonabant and AM251. Treatment with 10 µM rimonabant resulted in significant cell death (Fig. 1a). Co-treatment with ACEA effectively prevented this cell death induced (Fig. 1a). Similarly, 5 and 10 µM AM251 induced significant cell death (Figs. 1b, 1c). One micromolar ACEA conferred protection against by 5 µM AM251-induced cell death (Fig. 1b), and 10 µM ACEA provided protection against 10 µM AM251-induced cell death (Fig. 1c). In contrast, WIN exhibited only a minimal protective effect against rimonabant-induced cell death (Fig. 1a) and did not improve viability in AM251-treated cells (Figs. 1b, 1c). In summary, ACEA showed a significantly greater neuroprotective effect against CB1 antagonists-induced apoptosis compared with WIN.
SH-SY5Y cells were treated with ACEA or WIN (100 nM, 1, or 10 µM) in the presence of 10 µM rimonabant (Rim; a) or 5 or 10 µM AM251 (b, c, respectively) in a serum-free medium for 12 h. Green, calcein-AM; red, ethidium homodimer-1. Calcein-AM positive cells are indicated as live cells and quantified by ImageJ software. Scale bar: 500 µm. The data of bar graph are shown as the mean ± standard deviation (S.D.) for n = 6 or 9 (2 or 3 per independent experiment). Normalization was conducted based on the mean value of control group per independent experiment. Data were analyzed by Kruskal–Wallis ANOVA with Dunn’s multiple comparison test. Each group was compared with the group of Rim or AM251 treatment. *,**, or **** indicate p < 0.05, 0.01, or 0.0001, respectively.
Our previous study demonstrated that the CB1R antagonists rimonabant and AM251 induce mitochondrial fragmentation and apoptosis.21) Since mitochondrial fragmentation is a stress response that allows damaged segments to be separated from healthy ones,32) we investigated whether ACEA’s protective effect against CB1R antagonist-induced apoptosis involves the prevention of mitochondrial damage.
First, we evaluated the effects of ACEA and WIN on mitochondrial morphology under serum-free conditions. Treatment with 10 µM ACEA for 1 h resulted in a slight increase in mitochondrial circularity and a reduction in branch length (Supplementary Fig. S2a). In contrast, WIN induced mitochondrial fragmentation in a dose-dependent manner, as evidenced by increased mitochondrial circularity and decreased branch length (Supplementary Fig. S2b). These results suggest that high doses of WIN promote mitochondrial fragmentation, whereas ACEA has only a minor effect on mitochondrial morphology, even at high doses.
Next, we assessed whether these agonists could counteract CB1R antagonist-induced mitochondrial fragmentation. SH-SY5Y cells were co-treated with either ACEA or WIN alongside 10 µM rimonabant or 10 µM AM251 for 1 h. ACEA partially mitigated rimonabant-induced mitochondrial fragmentation, as indicated by a slight reduction in mitochondrial circularity and an increase in branch length (Fig. 2a). However, ACEA failed to prevent 10 µM AM251-induced mitochondrial fragmentation at any dose tested (Fig. 2b). Furthermore, WIN did not exhibit any protective effect against mitochondrial fragmentation induced by either antagonist (Figs. 2a, 2b).
SH-SY5Y cells were treated with ACEA or WIN (100 nM, 1, or 10 µM) in the presence of 10 µM rimonabant (Rim; a) or 10 µM AM251 (b) in a serum-free medium for 12 h. Green, Tom20; blue, DAPI. Scale bar: 10 µm. Mitochondrial morphology was analyzed using ImageJ. Data were shown as box plots with 10–90 percentile. Rim: 54 cells (10 µM Rim), 47 cells (100 nM ACEA), 55 cells (1 µM ACEA), 48 cells (10 µM ACEA), 40 cells (100 nM WIN), 49 cells (1 µM WIN), and 48 cells (10 µM WIN); AM251: 52 cells (10 µM AM251), 50 cells (100 nM ACEA), 54 cells (1 µM ACEA), 54 cells (10 µM ACEA), 55 cells (100 nM WIN), 52 cells (1 µM WIN), and 53 cells (10 µM WIN). Data were analyzed by Kruskal–Wallis ANOVA with Dunn’s multiple comparison test. * or ** indicate p < 0.05 or 0.01, respectively. The total number of analyzed mitochondria (green particles) were indicated at the graph in each condition.
To further evaluate ACEA’s effect on mitochondrial fragmentation, SEM was employed. SEM analysis revealed that 10 µM ACEA slightly increased mitochondrial area in cells treated with either 10 µM rimonabant or 10 µM AM251 (Figs. 3a, 3b). However, ACEA did not prevent the increase in mitochondrial circularity induced by either antagonist (Figs. 3a, 3b). These findings suggest that while ACEA provides modest protection against mitochondrial fragmentation induced by CB1R antagonists, its effect is limited.
(a) SEM images of mitochondria. Cells were treated with ACEA in the presence of 10 µM Rim or 10 µM AM251 for 1 h, respectively. Scale bar: 1 µm. (b) Mitochondrial morphology analysis in co-treatment of Rim and ACEA. Non-treatment: 184 mitochondria; Rim: 163 mitochondria; Rim + ACEA: 150 mitochondria. Data were shown as box plots with 10–90 percentile. Data were analyzed by Kruskal–Wallis ANOVA with Dunn’s multiple comparison test. Each group was compared with the group of Rim or AM251 treatment. * or **** indicate p < 0.05 or 0.0001, respectively. (c) Mitochondrial morphology analysis in co-treatment of AM251 and ACEA. Non-treatment: 171 mitochondria; AM251: 153 mitochondria; AM251 + ACEA: 196 mitochondria. Data were shown as Box and whiskers with 10–90 percentile. Data were analyzed by Kruskal–Wallis ANOVA with Dunn’s multiple comparison test. * or **** indicate p < 0.05 or 0.0001, respectively.
Our previous study demonstrated that CB1R antagonists activate the PERK/eIF2α/ATF4/CHOP signaling pathway, ultimately leading to apoptosis.21) During ER stress-mediated apoptosis, phosphorylation of PERK increases the level of phosphorylated eIF2α (peIF2α), which in turn promotes the selective transcription of ATF4 and CHOP.33–35) To determine whether ACEA modulates this signaling cascade, we analyzed the expression of key signaling components peIF2α, ATF4, and CHOP. SH-SY5Y cells were co-treated with 10 µM ACEA and either 10 µM rimonabant or 10 µM AM251 for 3 h. ACEA treatment significantly suppressed the induction of peIF2α by both CB1R antagonists (Fig. 4). In rimonabant-treated cells, ACEA also reduced the expression of ATF4 but had no significant effect on CHOP (Fig. 4). In AM251-treated cells, ACEA tended to decrease ATF4 expression (p = 0.083) and reduced CHOP expression (Fig. 4). These findings suggest that ACEA may attenuate CB1R antagonist-induced activation of the ER stress signaling pathway by inhibiting the eIF2α/ATF4/CHOP cascade.
Lysates of cells co-treated with 10 µM ACEA and 10 µM Rim or 10 µM AM251 for 3 h were used. The data of bar graph are shown as the mean ± S.D. for n = 4–6 (4–6 per independent experiment). Normalization was conducted based on the mean value of DMSO group per independent experiment. Expression levels of peIF2α was relative to eIF2α per independent experiment, while ATF4 and CHOP were relative to β actin. Data were analyzed by paired t test. * indicates p < 0.05.
The differential effects of ACEA and WIN on CB1R antagonist-induced apoptosis (Fig. 1) suggest that ACEA’s neuroprotective effects may be CB1R-independent. CB1R agonists have been reported to interact with TRANSIENT RECEPTOR POTENTIAL VANILLOID 1 (TRPV1), a nonselective cation channel with high calcium permeability.36–40) ACEA functions as a TRPV1 agonist,41–44) exhibiting selective binding to CB1R at nanomolar concentrations,45) while engaging with both CB1R and TRPV1 at higher concentrations.42–44) Activation of TRPV1 by ACEA has been shown to enhance neurotransmission in nociceptive peripheral neurons by increasing calcium influx.42,43) In addition, TRPV1 is endogenously expressed in SH-SY5Y cells.23,24) Based on this, we investigated whether TRPV1 mediates ACEA’s protective effects in SH-SY5Y cells.
To assess TRPV1′s involvement, we examined the effect of the TRPV1 antagonist capsazepine on ACEA-induced neuroprotection. First, we confirmed that capsazepine alone exhibited cytotoxicity, reducing cell viability by 23% at 10 µM under serum-free conditions, suggesting that high doses of capsazepine induce cytotoxicity (Supplementary Fig. S3a). Next, we tested whether capsazepine antagonizes ACEA’s neuroprotective effects. SH-SY5Y cells were co-treated with 10 µM capsazepine and 10 µM ACEA in the presence of either 10 µM rimonabant or AM251 for 12 h.
A significant interaction between capsazepine and ACEA was observed in rimonabant-treated cells (F (1,32) = 6.665, p = 0.0146), where capsazepine significantly reduced ACEA’s protective effects (Fig. 5a). In contrast, no significant interaction was detected between capsazepine and ACEA in AM251-treated cells (F (1, 32) = 0.5353, p = 0.4697), although capsazepine still diminished ACEA’s protective effects without affecting vehicle-treated cells (Fig. 5b). These results suggest that capsazepine antagonizes ACEA’s neuroprotection against apoptosis induced by CB1R antagonists.
SH-SY5Y cells were treated with 10 µM CPZ and 10 µM ACEA in the presence of (a) 10 µM Rim or (b) 10 µM AM251 in a serum-free medium for 12 h. Green, calcein-AM; red, ethidium homodimer-1. Scale bar: 500 µm. The data of bar graph are shown as the mean ± S.D. for n = 9 (3 per independent experiment). Normalization was conducted based on the mean value of single DMSO treatment group per independent experiment. Data were analyzed by Two-way ANOVA with Uncorrected Fisher’s LSD test. *,**,***, or **** indicate p < 0.05, 0.01, 0.001, or 0.0001, respectively.
To further examine dose-dependent effects, we tested lower doses of capsazepine (2.5 and 5 µM) in the same experimental conditions. Significant interactions were observed between ACEA and both 2.5 µM (F(1,32) = 32.25, p < 0.0001) and 5 µM (F(1,32) = 4.257, p = 0.0475) capsazepine in rimonabant-treated cells, with capsazepine reducing ACEA’s protective effects (Supplementary Fig. S3b). Interestingly, 2.5 µM capsazepine slightly increased cell viability in rimonabant-treated cells (Supplementary Fig. S3b). These results indicate that capsazepine blocks ACEA-mediated neuroprotection against rimonabant-induced apoptosis in a dose-dependent manner. Conversely, no significant interaction was observed between 2.5 µM (F(1,32) = 0.03098, p = 0.8614) or 5 µM (F(1,32) = 0.009782, p = 0.9218) capsazepine and ACEA in AM251-treated cells (Supplementary Fig. S3c). Additionally, while 2.5 µM capsazepine had no effect on cell viability, 5 µM capsazepine exhibited cytotoxicity in both vehicle- and ACEA-treated cells (Supplementary Fig. S3c). These findings suggest that in AM251-treated cells, capsazepine’s cytotoxicity itself may contribute to the observed reduction in protected cells, rather than a specific interaction between capsazepine and ACEA.
Collectively, these findings indicate that capsazepine antagonizes ACEA’s neuroprotection against apoptosis induced by CB1R antagonists, particularly in rimonabant-treated cells. The effects in AM251-treated cells appear to be influenced by the cytotoxicity of capsazepine rather than a direct interaction between capsazepine and ACEA. These findings suggest that ACEA’s neuroprotection may involve TRPV1 activation.
Antagonism by Capsazepine of ACEA-Induced Inactivation of ER Stress in SH-SY5Y CellsTo further investigate the role of TRPV1 in ACEA’s protective effects, we examined whether capsazepine influences ACEA-mediated suppression of the eIF2α/ATF4/CHOP pathway. SH-SY5Y cells were co-treated with 10 µM capsazepine and 10 µM ACEA in the presence of either 10 µM rimonabant or 10 µM AM251 for 3 h.
For peIF2α expression (Fig. 6), no significant interaction between ACEA and capsazepine was observed in rimonabant-treated cells (F(1,8) = 1.106, p = 0.3236). Neither ACEA (F(1,8) = 1.479, p = 0.2586) nor capsazepine (F(1,8) = 0.8274, p = 0.3896) significantly affected peIF2α expression. In contrast, in AM251-treated cells, while no significant interaction was detected (F(1,8) = 0.009506, p = 0.9247), ACEA significantly reduced peIF2α expression (F(1,8) = 5.986, p = 0.0401), and this effect was significantly reversed by capsazepine (F(1,8) = 11.51, p = 0.0095).
SH-SY5Y cells were treated with 10 µM capsazepine (CPZ) and 10 µM ACEA in the presence of 10 µM rimonabant (Rim) or 10 µM AM251 in a serum-free medium for 3 h. Expression of peIF2α/ATF4/CHOP from cell lysates were analyzed by Western blotting. The data of bar graph are shown as the mean ± S.D. for n = 3 (3 per independent experiment). Normalization was conducted based on the mean value of control group per independent experiment. Expression levels of peIF2α was relative to eIF2α per independent experiment, while ATF4 and CHOP were relative to β actin. Data were analyzed by Two-way ANOVA with Uncorrected Fisher’s LSD test. *,**, or *** indicate p < 0.05, 0.01, or 0.001, respectively.
For ATF4 expression (Fig. 6), a significant interaction between ACEA and capsazepine was observed in rimonabant-treated cells (F(1,8) = 5.602, p = 0.0455), where capsazepine significantly increased ATF4 expression (F(1,8) = 13.41, p = 0.0064), while ACEA alone had no significant effect (F(1,8) = 2.241, p = 0.1728). ACEA significantly reduced ATF4 expression compared with the rimonabant-only group, but this effect was reversed by capsazepine. Similarly, in AM251-treated cells, although there was no significant interaction (F(1,8) = 2.198, p = 0.1765), ACEA significantly decreased ATF4 expression (F(1,8) = 14.17, p = 0.0055), and capsazepine significantly reversed this suppression (F(1,8) = 25.51, p = 0.0010).
For CHOP expression (Fig. 6), in rimonabant-treated cells, no significant interaction was detected (F(1,8) = 1.799, p = 0.2167), and ACEA had no significant effect (F(1,8) = 0.5497, p = 0.4796), while capsazepine significantly increased CHOP expression (F(1,8) = 10.22, p = 0.0127). In AM251-treated cells, a significant interaction between ACEA and capsazepine was observed (F(1,8) = 5.452, p = 0.0478), with capsazepine significantly reversing ACEA-induced suppression of CHOP expression (F(1,8) = 39.87, p = 0.0002). ACEA alone did not show a significant effect (F(1,8) = 3.081, p = 0.1173), but it significantly reduced CHOP expression compared with the AM251-only group, an effect that was reversed by capsazepine co-treatment.
Collectively, these findings demonstrate that capsazepine antagonizes ACEA-induced suppression of ATF4 and CHOP expression, suggesting that ACEA’s protective effect against CB1R antagonist-induced ER stress is at least partially mediated through TRPV1 activation.
Protection by ACEA against Camptothecin-Induced Cell Death in SH-SY5Y CellsTo determine whether ACEA exerts protective effects beyond CB1R antagonists, we tested its impact against different cell death inducers: camptothecin, erastin, and carbonyl cyanide 3-chlorophenylhydrazone (CCCP). These compounds induce apoptosis,46) ferroptosis,47) and mitochondrial uncoupling,48) respectively. SH-SY5Y cells were treated with 10 µM ACEA or 10 µM WIN, in the presence of 10 µM camptothecin, 40 µM erastin, or 10 µM CCCP, in serum-free medium for 12 h.
Treatment with camptothecin, erastin, or CCCP significantly induced cell death in SH-SY5Y cells (Figs. 7a–7c). ACEA exhibited a trend toward attenuating camptothecin-induced cell death (p = 0.068), whereas WIN showed a tendency to exacerbate cytotoxicity (p = 0.068; Fig. 7a). In contrast, ACEA displayed a slight tendency to enhance erastin-induced cell death (p = 0.0925), while WIN had no significant effect (Fig. 7b). Neither ACEA nor WIN affected CCCP-induced cytotoxicity (Fig. 7c). These findings demonstrate that ACEA specifically protects SH-SY5Y cells against camptothecin-induced apoptosis.
SH-SY5Y cells were treated with 10 µM ACEA or 10 µM WIN in the presence of (a) 10 µM CPT, (b) 40 µM erastin, or (c) 10 µM CCCP for 12 h. Calcein-AM positive cells are quantified as live cells by ImageJ software. The data of bar graph are shown as the mean ± S.D. for n = 9 or 15 (3 per independent experiment). Normalization was conducted based on the mean value of single DMSO treatment group per independent experiment. Data were analyzed by Kruskal–Wallis ANOVA with Dunn’s multiple comparison test. Data were analyzed by Kruskal–Wallis ANOVA with Dunn’s multiple comparison test. Each group was compared with the group of CPT, erastin or CCCP treatment. * or **** indicates p < 0.05 or 0.0001, respectively. Cells were treated with 10 µM CPZ and 10 µM ACEA in the presence of (d) 10 µM CPT for 12 h. Green, calcein-AM; red, ethidium homodimer-1. Scale bar: 500 µm. The data of bar graph are shown as the mean ± S.D. for n = 9 (3 per independent experiment). Normalization was conducted based on the mean value of single DMSO treatment group per independent experiment. Data were analyzed by Two-way ANOVA with Uncorrected Fisher’s LSD test. * or **** indicate p < 0.05 or 0.0001, respectively.
Next, we investigated whether the capsazepine antagonizes ACEA’s protective effect against camptothecin-induced apoptosis. SH-SY5Y cells were co-treated with 10 µM ACEA, 5 µM capsazepine, and 10 µM camptothecin in serum-free medium for 12 h. A significant interaction between capsazepine and ACEA was observed (F (1, 32) = 4.320, p = 0.0458), with capsazepine reducing cell viability in both ACEA- and vehicle-treated conditions (Fig. 7d). These findings indicate that ACEA may protect against camptothecin-induced neurotoxicity through a mechanism similar to its protection against CB1R antagonist-induced apoptosis, likely involving TRPV1 activation.
In this study, we uncovered a novel neuroprotective mechanism of the CB1R agonist ACEA against apoptosis induced by the CB1R antagonists rimonabant and AM251. While ACEA is known for its CB1R-dependent effects, our findings indicate that its neuroprotection was blocked by the TRPV1 antagonist capsazepine. The involvement of TRPV1 was further supported by ACEA’s ability to prevent camptothecin-induced apoptosis through a similar capsazepine-sensitive mechanism. These results suggest that ACEA exerts its neuroprotective effects through a previously unrecognized CB1R-independent pathway.
WIN treatment did not exhibit a dose-dependent protective effect against rimonabant- or AM251-induced apoptosis, whereas ACEA demonstrated significant neuroprotection. Both WIN and ACEA function as TRPV1 agonists,38,39) but, unlike ACEA, WIN also acts as an agonist at cannabinoid receptor type 2 (CB2R). Previous studies have reported CB2R expression in SH-SY5Y cells and the cytotoxic effects mediated by its activation.49,50) Thus, WIN may induce cytotoxicity via CB2R activation, potentially masking any protective effects mediated by TRPV1 activation. On the contrary, differences in neuroprotective effects between WIN and ACEA might be explained by CB1R desensitization. Both compounds induce CB1R desensitization via receptor internalization.51,52) However, WIN has been shown to promote CB1R internalization at significantly lower concentrations than ACEA.52) Since we evaluated neuroprotection under a 12-h treatment condition, WIN may have rapidly induced CB1R internalization and desensitization, thereby diminishing its potential protective effects. In contrast, ACEA, which requires higher concentrations for CB1R internalization, may have maintained its receptor signaling and protective effects throughout the treatment period. It would be necessary to conduct experiments with time-dependent drug treatment to clarify whether the differences in the effects of WIN and ACEA are the result of differences in pharmacological properties or time differences in CB1R internalization.
ACEA exerts complex neuroprotective effects through distinct mechanisms depending on the type of cellular stress. Previous studies have reported that ACEA provides neuroprotection against various neurotoxicities,15–17,53,54) while high doses of ACEA can induce neurotoxicity.55) In brain ischemia models, ACEA’s neuroprotective effects have been attributed to CB1R-dependent mitigation of mitochondrial damage.15–17) However, in the present study, ACEA showed only a minor protective effect against mitochondrial damage and instead appeared to inhibit ER stress in a capsazepine-sensitive manner. This observation aligns with recent evidence showing TRPV1-dependent protection against ER stress-mediated apoptosis in Neuro2a cells, another neuroblastoma cell line.56) These findings suggest that ACEA’s protective mechanisms vary depending on the type of neurotoxic insult.
Our findings demonstrate that ACEA inhibits neurotoxicity induced by CB1R antagonists via a CB1R-independent mechanism, raising the possibility that CB1R antagonists may induce neurotoxicity independently of CB1R blockade. Previous studies have reported that rimonabant undergoes intracellular bioactivation to produce toxic reactive metabolites that induce ER stress.57–59) These metabolites can disrupt ER calcium homeostasis by interfering with calcium-regulating proteins, leading to calcium overload in the ER.60) The involvement of TRPV1 in ACEA’s protective effects, as evidenced by capsazepine’s ability to block this protection, provides mechanistic insight into this process. TRPV1, expressed on both the plasma membrane and the ER membrane,61,62) plays a crucial role in maintaining ER calcium homeostasis.63,64) Modulation of TRPV1 activity has been shown to prevent ER stress caused by disruption of calcium homeostasis. Given that capsazepine inhibits TRPV1, it may exacerbate ER stress by further disrupting calcium homeostasis, thereby counteracting ACEA’s neuroprotective effects. Collectively, ACEA may activate TRPV1 to prevent calcium overload in the ER by facilitating calcium release,62,64) thereby mitigating ER stress.
However, ACEA’s neuroprotective effects may not be exclusively TRPV1-dependent. When metabolized, ACEA is degraded to arachidonic acid and 2-chloroethylamine by fatty acid amide hydrolase.65) Previous studies have reported that arachidonic acid metabolites, such as epoxyeicosatrienoic acids, reduce ER stress by maintaining intracellular calcium homeostasis.66) Therefore, ACEA metabolites may also contribute to ER stress reduction. Future research using TRPV1 knockout models or other TRPV1 antagonists will be necessary to clarify the precise mechanism underlying ACEA’s protection against ER stress-mediated apoptosis.
To further explore the scope of ACEA’s protective effects, we examined its impact on three cytotoxic drugs with distinct cell death mechanisms. ACEA conferred protection against camptothecin-induced cytotoxicity. While camptothecin induces apoptosis through DNA damage by inhibiting topoisomerase I,46,67) it also induces ER stress and activates the PERK/eIF2α/ATF4/CHOP pathway.68) Thus, ACEA’s protective effect against camptothecin cytotoxicity may stem from its ability to inhibit this ER stress signaling pathway. In contrast, ACEA slightly exacerbated erastin-induced cytotoxicity. Erastin triggers ferroptosis by increasing the formation of lipid reactive oxygen species (ROS) and lipid peroxidation.47) Given that high doses of ACEA induce mitochondrial fragmentation, ACEA might enhance ROS production, thereby potentiating erastin’s ferroptotic effects. Furthermore, ACEA failed to protect against CCCP-induced cytotoxicity, which induces mitochondrial damage and subsequent cell death.48,69) This lack of protection is consistent with ACEA’s limited effects on mitochondrial morphology. These differential effects across various cell death mechanisms support ACEA’s specific role in protecting against ER stress-mediated apoptosis.
Our findings also suggest a differential effect of capsazepine on ACEA’s neuroprotection between rimonabant and AM251. Capsazepine blocked ACEA’s neuroprotective effects against rimonabant-induced apoptosis in a dose-dependent manner. However, in AM251-treated cells, capsazepine’s cytotoxicity itself may have contributed to the reduction in ACEA-mediated protection rather than a direct interaction between capsazepine and ACEA. This discrepancy may arise from a difference in TRPV1 activity between rimonabant and AM251. A previous study has shown that rimonabant, but not AM251, acts as a TRPV1 agonist.39) Thus, rimonabant may partially counteract capsazepine-induced cytotoxicity through TRPV1 activation, highlighting the interaction between capsazepine and ACEA.
In conclusion, we demonstrated that ACEA protects neuronal cells against apoptosis through TRPV1 activation rather than CB1R signaling. Moreover, ACEA’s protective mechanism extends beyond CB1R antagonist-induced toxicity to other ER stress-inducing agents such as camptothecin. This study provides new insights into the neuroprotective and neurotoxic mechanisms of CB1R ligands and may inform the therapeutic potential for neurological disorders.
Limitations of the StudyThis study demonstrates that ACEA exerts a CB1R-independent protective effect against apoptosis in SH-SY5Y cells. However, our experiments were limited to this single cell line. Whether this neuroprotective mechanism extends to other neuronal cell types or in vivo systems requires further investigation. Further investigations using primary neurons or animal models will be necessary to confirm the broader relevance of these findings. Additionally, the role of CB1R in ACEA-mediated neuroprotection was assessed solely through the use of the TRPV1 antagonist capsazepine. To definitively rule out CB1R involvement, future research should examine ACEA’s neuroprotective properties in CB1R knockdown neuronal cells or CB1R-deficient mice.
K.M. is supported by JSPS DC1 Research Fellowships for Young Scientists (22KJ2960). This work was supported by JSPS KAKENHI, Grant Nos: 19K20196 (K.K.), 19K24693 (C.N.), 20K06730 (K.O.), 22K17815 (K.K.), and 23K06239 (K.O.). Graphical abstract was created with Biorender.com.
Kazuaki Mori: Formal analysis, Investigation, Data Curation, Writing—original draft, Visualization. Akinobu Togo: Investigation, Supervision. Keisuke Ohta: Methodology, Resources, Writing—Review & Editing, Supervision, Funding acquisition. Toru Asahi: Resources, Project administration, Funding acquisition. Chihiro Nozaki: Conceptualization, Methodology, Resources, Writing—Review & Editing, Supervision, Project administration, Funding acquisition. Kosuke Kataoka: Conceptualization, Methodology, Resources, Writing—Review & Editing, Supervision, Project administration, Funding acquisition.
The authors declare no conflict of interest.
This article contains supplementary materials.