Cannabinoid-induced autophagy regulates suppressor of cytokine signaling-3 in intestinal epithelium
ABSTRACT
Autophagy is a catabolic process involved in homeostatic and regulated cellular protein recycling and degradation via the lysosomal degradation pathway. Emerging data associate impaired autophagy, increased activity in the endocannabinoid system, and upregulation of suppressor of cytokine signaling-3 (SOCS3) protein expression during intestinal inflammation. We have investigated whether these three processes are linked. By assessing the impact of the phytocannabinoid cannabidiol (CBD), the synthetic cannabinoid arachidonyl-2′-chloroethylamide (ACEA), and the endocannabinoid N-arachidonoylethanolamine (AEA) on autophagosome formation, we explored whether these actions were responsible for cyclic SOCS3 protein levels. Our findings show that all three cannabinoids induce autophagy in a dose-dependent manner in fully differentiated Caco-2 cells, a model of mature intestinal epithelium. ACEA and AEA induced canonical autophagy, which was cannabinoid type 1 receptor-mediated. In contrast, CBD was able to bypass the cannabinoid type 1 receptor and the canonical pathway to induce autophagy, albeit to a lesser extent. Functionally, all three cannabinoids reduced SOCS3 protein expression, which was reversed by blocking early and late autophagy. In conclusion, the regulatory protein SOCS3 is regulated by autophagy, and cannabinoids play a role in this process, which could be important when therapeutic applications for the cannabinoids in inflammatory conditions are considered.
AUTOPHAGY EXHIBITS MANY physiological roles in the cellular process. Regulation and induction of autophagy correspond to an outcome for the cell: survival or death. During nutrient starving or growth factor deprivation, autophagy acts as the catabolic process to maintain homoeostasis in the cellular context. Stress-induced autophagy will recycle cellular content by transferring energy from a nonessential process to a more crucial cellular process (7, 27, 40). In contrast to the survival function, autophagy is utilized as a defense mechanism to eliminate the invasion of microbial content (18, 23, 30). In normal colonic cells, autophagy is required for renewal of the colonic epithelium. Autophagy is active at the lower part of the crypt of the colonic gland, where proliferation of stem cell populations is sustained (11). The importance of autophagy regulation in maintaining homeostasis has linked this system to various pathologies (22, 45). Studies have revealed polymorphisms of various autophagy-associated genes (ATG16L1 and IRGM) with increased susceptibility to Crohn’s disease (CD) (29, 33). Intestinal epithelial cells line the entire gastrointestinal (GI) tract, providing a barrier to microbes and, as such, encounter high exposure to inflammatory stimuli. Variation (T300A) of the ATG16L1 gene in CD resulted in an autophagy-associated defect in Paneth cells, which reside in the crypt of Lieberkühn within the small intestine (4, 19). This variation also increases processing by caspase-3, resulting in reduced autophagy and pathogen clearance (31).
During intestinal inflammation, upregulation of endocannabinoid levels and increased expression of the cannabinoid (CB) receptor will enhance the action of the endocannabinoid system (9). This is shown by an increase in CB1 receptor expression in the colon of intrarectal dinitrobenzene sulfonic acid-treated mice (28). Also, CB2 receptor expression was increased in the epithelium of inflammatory bowel disease patients compared with healthy human intestinal epithelium, where CB2 receptor expression is weak (48). Interestingly, cannabinoids such as Δ9-tetrahydrocannabidiol (THC) and cannabidiol (CBD) have been shown to induce autophagy in cancer cell lines (41, 42).
Both Δ9-THC and CBD decreased production and release of proinflammatory cytokines in BV-2 microglial cells (21). CBD, but not Δ9-THC, was shown to reduce nuclear factor-κ light chain enhancer of activated B cells (NF-κB) activity, upregulate the activation of STAT3, and decrease mRNA expression level of suppressor of cytokine signaling-3 (SOCS3) (21). Activation of SOCS3 limits transcription factor activation and nuclear translocation in response to stimulation from inflammatory cytokines (36). Both mRNA and protein expression of SOCS3 is upregulated in colon samples from ulcerative colitis (UC) and CD patients compared with healthy controls (44). Furthermore, in vivo SOCS3 limits proliferation of the inflammatory epithelial cells in damaged crypts. This leads to reduction of IL-6-mediated STAT3 and TNFα-mediated NF-κB activation (36). Cellular SOCS3 expression level is constantly oscillating, and there are no current data that identify how cyclic SOCS3 is regulated beyond the transcriptional level (25, 50).
We used immunoblotting, confocal imaging, and small interfering RNA knockdown technology to examine whether cannabinoids regulate SOCS3 protein expression through autophagy and whether this action was dependent on the CB1 receptor.
MATERIALS AND METHODS
Unless otherwise stated, all cell culture reagents were obtained from Life Technologies (Paisley, UK) and all chemicals from Sigma-Aldrich (Dorset, UK). The cannabinoids were the phytocannabinoid (−)-cannabidiol (CBD), the synthetic CB1 receptor agonist arachidonyl-2′-chloroethylamide (ACEA), the endogenous cannabinoid N-arachidonoylethanolamine (AEA), and the synthetic CB1 receptor antagonist N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide (AM 251), all from R & D System Europe (Abingdon, UK). The autophagy inhibitors were the class III phosphatidylinositol 3-kinase (PI3KCIII) Vps34 inhibitor 3-methyladenine (3-MA; R & D System Europe) for early autophagy pathway inhibition and the endosomal acidification inhibitor bafilomycin A1 (BafA1; R & D System Europe) for late autophagy pathway inhibition.
Cell culture.
The human colonic epithelial cell line Caco-2 was cultured in MEM with 8% fetal bovine serum and 1% nonessential amino acids. Cells were maintained at 37°C in 5% CO2, and medium was changed every 2–3 days. When cells reached 70% confluency, they were passaged or plated for experiments. For experiments, cells were fully differentiated in culture for 14–17 days.
CB1 receptor gene knockdown.
Wild-type Caco-2 cells were seeded at a density of 2 × 105 cells per well in a six-well culture plate overnight before transfection with 29-mer short-hairpin RNA, corresponding specifically to the human CB1 receptor gene in retroviral HuSH pRS plasmid vector (gene ID no. 1268; OriGene product code TR316500; Insight Biotechnology, Middlesex, UK) for 48 h and termed CB1KD Caco-2 cells. Vector control (control Caco-2) cells were transfected with noneffective 29-mer scrambled short-hairpin RNA cassette in pRS plasmid vector. All procedures were performed according to the standard manufacturer’s protocol using TurboFectin 8.0 transfection reagent (OriGene product code TF81001, Insight Biotechnology). After 48 h, cells were transferred to 96-well plates and maintained in complete medium with puromycin (0.06 μl/ml) for selection. Within 2 wk, three to four clonal populations of cells were selected, passaged, and cultured for experiments with continued selection pressure.
Western immunoblotting.
Cells were washed with cold PBS and lysed in RIPA buffer (50 mM Tris·HCl, pH 8.0, with 150 mM sodium chloride, 1.0% Igepal CA-630, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate) supplemented with phosphatase inhibitor cocktail 2 and protease inhibitor cocktail. Insoluble cell debris was removed by centrifugation (12,000 rpm, 3 min), and protein concentration was determined by the Bradford assay (Bio-Rad Laboratories, Hertfordshire, UK). Samples were mixed with 4× NuPAGE lithium dodecyl sulfate sample buffer (Life Technologies, Paisley, UK) supplemented with 5% 2-mercaptoethanol and boiled at 95°C for 2 min, and 20- to 30-μg aliquots were resolved on Mini-PROTEAN TGX precast electrophoresis gels (Bio-Rad Laboratories). Proteins were transferred to nitrocellulose membranes using the Bio-Rad Trans-Blot SD system, and membranes were blocked in Pierce protein-free (Tris-buffered saline) blocking buffer (Fisher Scientific, Loughborough, UK). Immunodetection was performed with the primary antibodies LC3B (1:1,000 dilution; catalog no. L8918, Sigma-Aldrich), β-actin (1:5,000 dilution; catalog no. 4967, Cell Signaling Technologies), phosphorylated STAT3 (1:1,000 dilution; catalog no. sc-8001-R, Santa Cruz Biotechnology), total STAT3 (1:1,000 dilution; catalog no. sc-8019, Santa Cruz Biotechnology), SOCS3 (L210) antibody (1:1,000 dilution; catalog no. 2932, Cell Signaling Technologies), and CB1 receptor antibody (1:1,000 dilution; catalog no. 10006590, Cambridge Bioscience). Secondary horseradish peroxidase-conjugated anti-mouse (1:10,000 dilution; catalog no. 7076, Cell Signaling Technology) or anti-rabbit (1:10,000 dilution; catalog no. 7074, Cell Signaling Technology) monoclonal antibody was applied, the cells were exposed to Clarity Western enhanced chemiluminescence substrate (Bio-Rad Laboratories), and images were obtained using the Bio-Rad ChemiDoc XRS system.
Confocal imaging.
Cells were grown on coverslips at 106 cells/well until the cell monolayer was formed (7–10 days). The Cyto-ID autophagy detection kit (Enzo Life Sciences, Exeter, UK) was used to stain the treated cells with Cyto-ID green detection reagent and Hoechst 33342 nuclear stain; then the cells were fixed in 4% paraformaldehyde according to the standard manufacturer’s protocol. Images were captured with a Zeiss confocal microscope and analyzed with LSM Image software.
Statistical analysis.
Experiments were repeated multiple times, and data were pooled. To determine statistical significance, ANOVA was performed with two-sided Dunnett’s or Tukey’s post hoc test where appropriate.
RESULTS
Phytocannabinoid, synthetic cannabinoid, and endocannabinoid induce autophagosome formation.
To study the effect of cannabinoids (ACEA, AEA, and CBD) in the context of autophagy on fully differentiated Caco-2 cells, we examined autophagosome formation by monitoring conversion of microtubule-associated protein light chain 3 (LC3) from LC3-I to LC3-II by immunoblotting, which correlates with autophagosome formation (17). Dose- and time-course responses revealed that 10 μM CBD induced autophagosome formation within 4–6 h (Fig. 1A) and that this effect was dose-dependent from 0.1 to 25 μM (Fig. 1B). However, the low dose (0.1 μM) of CBD appeared to inhibit this effect. To further evaluate the inhibition of autophagosome formation, we examined whether 0.1 μM CBD could reverse this effect by treating the cells under low nutrient conditions. We found that LC3-II formation was not sustained. There was an initial reduction of LC3-II formation at 4 h, but LC3-II formation was increased by 8 h (Fig. 1C). The effect of endocannabinoid and synthetic cannabinoid on autophagy was further explored, and cells were treated with ACEA (10 and 100 nM) or AEA (1 and 10 μM). Both 100 nM ACEA (Fig. 1D) and 10 μM AEA (Fig. 1E) significantly increased LC3-II formation within the 4-h time frame.
Phytocannabinoids, synthetic cannabinoids, and endocannabinoids induced light chain 3 (LC3)-II formation. A: immunoblot analysis of LC3-II in fully differentiated Caco-2 cells in response to 10 μM cannabidiol (CBD) treatment within 24 h. B, D, and E: dose response to CBD (0.1–25 μM), arachidonyl-2′-chloroethylamide (ACEA, 10–100 nM), and N-arachidonoylethanolamine (AEA, 1–10 μM). Caco-2 cells were treated with cannabinoids for 4 h. Values (means ± SE) are expressed as relative fold increase in LC3-II protein expression (adjusted to β-actin). C: LC3-II formation in starvation-induced autophagy. Cells were starved in medium without serum for 48 h and then exposed to 0.1 μM CBD for 4–8 h. Values (means ± SD) are expressed as relative fold increase in LC3-II protein expression (adjusted to β-actin). *P < 0.05 vs. untreated control (by Dunnett’s 2-sided t-test).
Cannabinoids differentially affect the canonical autophagy pathway.
PI3KCIII has a significant role in initiation of canonical autophagy by recruiting autophagy-related gene (ATG) complexes to induce membrane phagophore formation (2). Involvement of PI3KCIII in autophagy can be assessed through the use of the PI3KCIII inhibitor 3-MA. To study the involvement of PI3KCIII in ACEA-, AEA-, or CBD-induced LC3-II formation, cells were treated with 3-MA (10 mM) for 1 h and then with cannabinoid for an additional 4 h. The ACEA and AEA response was completely abrogated in the presence of 3-MA (Fig. 2A), whereas the CBD response was not completely inhibited by 3-MA (Fig. 2A), suggesting an additional route to autophagy.
Mechanism involved in cannabinoid-induced LC3-II formation. A: immunoblot of cannabinoid-induced LC3-II protein expression in the presence of the class III phosphatidylinositol 3-kinase (PI3KCIII) inhibitor 3-methyladenine (3-MA). Fully differentiated Caco-2 cells were pretreated with or without 3-MA (10 mM) for 1 h prior to addition of CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for an additional 4 h. Vehicle control (VC) was treated for a total duration of 5 h. B: relative fold change in fluorescence intensity in relation to the untreated sample, as generated from the pseudo-3-dimensional data in C. C: confocal microscopy projection images of fully differentiated Caco-2 cells stained with Cyto-ID and Hoechst 33342 dyes following treatment. Caco-2 cells were grown to maturity on glass coverslips and treated with or without 3-MA for 1 h prior to addition of CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for an additional 4 h. Images are composites of all the z-stack images from the same field of view for each treatment. Pseudo-3-dimensional graph was generated using LSM Image software and plotted as fluorescence intensity vs. number of pixels detected from each level of intensity. Dand E: immunoblot of cannabinoid-induced LC3-II protein expression in the presence of the endosomal acidification inhibitor bafilomycin A1 (BafA1). Fully differentiated Caco-2 cells were pretreated with or without BafA1 (100 nM) for 20 h prior to addition of CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for an additional 4 h. Vehicle control (VC) was treated for a total duration of 24 h. Because of nonlinearity of enhanced chemiluminescence with the use of BafA1, BafA1 data are presented in two forms: overexposure (D) and normal exposure (E). Values (means ± SE) represent relative fold increase in LC3-II protein expression (adjusted to β-actin). *P < 0.05 vs. untreated control; ‡P < 0.05 vs. paired treatment control (by Tukey’s post hoc test).
To further verify the significance of these data, we stained cells using the Cyto-ID autophagy detection kit as a selective marker of autolysosomes and earlier autophagic compartments (6). Analysis of fluorescence using confocal microscopy indicated an increase in autolysosomes and earlier autophagic compartments in response to ACEA, AEA, and CBD treatment (Fig. 2B) that was inhibited in the presence of 3-MA (Fig. 2C). Again, the reduction was less pronounced with CBD.
Cannabinoids inhibit autophagosome degradation.
To further evaluate the impact of ACEA, AEA, and CBD on autophagosome synthesis, we performed the “autophagy flux assay” by treating the cells with the endosomal acidification inhibitor BafA1 (20). ACEA, AEA, CBD, and BafA1alone significantly increased LC3-II formation in this experimental setting (Fig. 2D), but the cannabinoids did not further enhance BafA1-induced LC3-II formation (Fig. 2E).
Cannabinoids differentially engage the CB1 receptor.
ACEA and AEA are well-studied agonists for the CB1 receptor, whereas previous studies showed that CBD is an antagonist for the CB1 receptor agonist in CB1 receptor-expressing cells (34). First, to determine whether ACEA-, AEA-, and CBD-induced LC3-II formation was CB1 receptor-mediated, we pretreated Caco-2 cells with the CB1 receptor antagonist AM 251 (Fig. 3A). Despite AM 251-induced blockage of the CB1 receptor in Caco-2 cells, 10 μM CBD was able to increase LC3-II conversion. In contrast, ACEA and AEA required the CB1 receptor. Next, to further verify these data, we generated CB1 receptor knockdown in Caco-2 (CB1KD Caco-2) cells. Successful knockdown of the CB1 receptor gene in Caco-2 cells was verified at the translational level (Fig. 3B), and vector control Caco-2 (control Caco-2) cells were transfected with plasmid vector that contained scrambled DNA sequences and were used as the control for CB1KD Caco-2 cells. We successfully knocked down the CB1 receptor in CB1KD Caco-2 cells by ∼70% compared with control Caco-2 cells (Fig. 3B). CBD increased LC3-II formation in CB1KD and control Caco-2 cells, albeit to different maxima, whereas ACEA and AEA increased LC3-II formation in control, but not CB1KD, Caco-2 cells (Fig. 3C).
Cannabinoids differentially engage the CB1 receptor. A: immunoblot of cannabinoid-induced LC3-II protein expression in the presence of AM 251. Fully differentiated Caco-2 cells were pretreated with or without AM 251 (100 nM) for 10 min prior to addition of CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for an additional 4 h. Values (means ± SD) represent relative fold increase in LC3-II protein expression (adjusted to β-actin). B: immunoblot analysis of CB1 receptor expression in untreated wild-type, vector control, and CB1 receptor knockdown (CB1KD) Caco-2 cells. C: effect of cannabinoid treatments on LC3-II protein expression in vector control and CB1KD Caco-2 cells. Cells were treated with CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for 4 h. Values (means ± SE) represent relative fold increase in LC3-II protein expression (adjusted to β-actin). *P < 0.05 vs. untreated control (by Dunnett’s 2-sided t-test).
Cannabinoids modulate SOCS3.
Previous research showed an interaction of cannabinoids with the STAT3-SOCS3 pathway, revealing its anti-inflammatory potential (16, 21). We sought to establish whether SOCS3 protein levels were altered by cannabinoids in our system and whether any changes were related to autophagy. All three cannabinoids reduced endogenous SOCS3 protein level at the 4-h time point (Fig. 4, A and B). Blocking the early stage of autophagy with 3-MA reversed this effect in the presence of AEA and CBD (Fig. 4A), whereas blocking the late stage of autophagy with BafA1 reversed only the ACEA effect (Fig. 4B). We previously showed that cannabinoid-induced SOCS3 reduction was negated in the presence of the CB1 receptor antagonist AM 251 in wild-type Caco-2 cells (data not shown). To clarify these data, we performed the same experiment by using the generated CB1KD Caco-2 cells. Interestingly, treatment of the CB1KD cells with cannabinoids did not affect SOCS3 expression (Fig. 4C).
Cannabinoids modulate suppressor of cytokine signaling 3 (SOCS3). A: immunoblot of cannabinoid effect on SOCS3 protein in the presence of 3-MA. Fully differentiated Caco-2 cells were pretreated with or without 3-MA (10 mM) for 1 h prior to addition of CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for an additional 4 h. Values (means ± SE) represent relative fold increase in SOCS3 protein expression (adjusted to β-actin). *P < 0.05 vs. untreated control; ‡P < 0.05 vs. paired treatment control (by Tukey’s post hoc test). B: immunoblot of cannabinoid effect on SOCS3 protein in the presence of BafA1. Fully differentiated Caco-2 cells were pretreated with or without BafA1 (100 nM) for 20 h prior to addition of CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for an additional 4 h. Vehicle control (VC) was treated for a total duration of 24 h. Values (means ± SE) represent relative fold increase in SOCS3 protein expression (adjusted to β-actin). *P < 0.05 vs. untreated control; ‡P < 0.05 vs. paired treatment control (by Tukey’s post hoc test). C: immunoblot of cannabinoid effect on SOCS3 protein in vector control and CB1KD Caco-2 cells. Cells were treated with CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for 4 h. Values (means ± SE) represent relative fold increase in SOCS3 protein expression (adjusted to β-actin). *P < 0.05 vs. untreated control (by Dunnett’s 2-sided t-test). D: immunoblot of cannabinoid-induced phosphorylated (Ser727) STAT3 [p-STAT3 (Ser727)] protein in fully differentiated Caco-2 cells. Cells were treated with cannabinoids for 4 h. Values (means ± SE) represent relative fold increase in phosphorylated (Ser727) STAT3 protein expression (adjusted to total STAT3). *P < 0.05 vs. untreated control (by Dunnett’s 2-sided t-test). E: immunoblot of cannabinoid-induced phosphorylated (Ser727) STAT3 protein in the presence of BafA1. Cells were pretreated with BafA1 (100 nM) for 20 h prior to addition of CBD (10 μM), ACEA (100 nM), or AEA (10 μM) for an additional 4 h. Values (means ± SE) represent relative fold increase in phosphorylated (Ser727) STAT3 protein expression normalized against BafA1 treatment (adjusted to total STAT3).
A known role for SOCS3 is inhibition of STAT3 activation (8, 50). Preliminary experiments revealed no significant change in STAT3 phosphorylated at Tyr705 in response to cannabinoid treatment (data not shown), whereas only CBD revealed phosphorylated Ser727 at 4 h (Fig. 4D), which was reversed by blocking autophagy (Fig. 4E).
DISCUSSION
First, it is essential to establish an appropriate in vitro cell culture model to obtain reliable data that can be translated into clinical development. Cells have been treated with medium in which the concentration of supplemented serum is reduced from 10% to 1%, with no changes in nonessential amino acid concentration. Reduced serum conditions change the sensitivity of the cellular response. In accordance with our findings, treatment with tamoxifen + Δ9-THC, CBD, or AEA in 10% FBS had no impact on C6 glioma cell viability, whereas 0.4% FBS reduced cell viability (14). Consistent with this finding, we determined that nutritional status in cells is important for the basal level of LC3-II (data not shown). Treated cells were not synchronized to the G0 stage of the cell cycle prior to administration of cannabinoids. The rationale for this was to mimic the physiological setting of the intestinal epithelium, in that all epithelial cells in the GI tract are not at the G0 stage in vivo.
LC3-II is the hallmark for autophagy. In this study, we found that all three cannabinoids, phytocannabinoid (CBD), endocannabinoid (AEA), and synthetic cannabinoid (ACEA), significantly enhanced LC3-II formation in a dose-dependent manner. CBD-enhanced LC3-II was sustained up to 24 h. CBD has been shown to be cytotoxic in a number of cell lines, in particular, breast cancer cells (42). CBD was shown to decrease viability of a breast cancer cell line in a dose-dependent manner but maintained a higher survival rate in normal cells (42). It is important to note that fully differentiated Caco-2 cells are thought to more closely resemble normal enterocytes, and we found that CBD did not affect the viability of fully differentiated Caco-2 cells (27a). Previous study showed that CBD-induced autophagy mediated cell death in a breast cancer cell line (42), whereas our present data show that CBD induces autophagy, but this does not lead to cell death. This suggests a cell type-dependent effect. Administration of a low concentration (0.1 μM) of CBD led to inhibition of LC3-II formation, but this effect was transient. There are no data showing the exact dose of cannabinoid delivered to the GI tract and absorbed by the GI cells after CBD administration. Therefore, our data show that CBD-induced autophagy is dose-dependent, and this could be an important finding for its therapeutic use in disease with disordered autophagy, for instance, CD.
A snapshot of LC3-II by immunoblotting may not accurately reflect the effect of the cannabinoids on autophagosome formation, as the increased LC3-II may well correlate with increased autophagic flux with a reduction of autophagosome degradation (39). To further evaluate the impact of cannabinoids on autophagosome synthesis, we performed an autophagy flux assay by treating the cells with the endosomal acidification inhibitor BafA1. BafA1 neutralizes lysosomal pH and affects LC3-II degradation (20, 49). We found that BafA1 alone induces a significant increase in LC3-II formation as a result of accumulation of LC3-II, corresponding to blocking of the autophagolysosomal degradation step. Cannabinoids do not further enhance or inhibit the BafA1-induced accumulation of LC3-II, which suggests that the cannabinoid effect on autophagy may be related to reduced autophagosome degradation as opposed to increased autophagosome synthesis. Since cannabinoids induced two- to threefold increases in LC3-II formation and BafA1 induced four- to fivefold increases, it is perhaps not surprising that we did not see a further enhancement of the BafA1 effect. The sensitivity of the method does not allow us to confirm or refute the notion that cannabinoids induce autophagosome synthesis but, rather, indicates an inhibitory role in late autophagy. However, in light of the SOCS3 data, in which early and late autophagy seem to be important, the tools available to study this phenomenon are quite crude.
As all the cannabinoids delivered significant LC3-II formation, we questioned whether the cannabinoid-induced effect was CB1 receptor-mediated. Here, we focused on the CB1 receptor, as we previously showed that the CB2 receptor is not expressed in our Caco-2 cell line (data not shown). ACEA- and AEA-induced effects were CB1 receptor-mediated. Unlike other G protein-coupled receptors, the CB1 receptor can be intracellularly activated by endogenous cannabinoids (37). Remarkably, this study showed localization of the CB1 receptor at lysosomal compartments (38). Intracellular injection of AEA activated functional CB1 receptor, and subsequent mobilization of intracellular calcium concentration was reduced by administration of BafA1(3). Such results illustrate involvement of the CB1 receptor in lysosomal degradation. Taken together, our data support the ACEA- and AEA-induced effect on late degradation. In contrast, the pharmacological block of the CB1 receptor showed some variation, but statistical evaluation confirmed that CBD did not require the receptor for its action. However, the knockdown data would suggest some requirement for the CB1 receptor, because, while autophagy induction by CBD was convincing, it did not reach the same maximum as control cells, which may be related to incomplete CB1 receptor knockdown or an unknown receptor-independent target. Δ9-THC has been shown to induce autophagy in human astrocytoma cells, and the Δ9-THC-induced effect was CB1 receptor-mediated (41). Although the studies were performed in different cell lines, these results suggest that Δ9-THC and CBD may act through both common and distinct mechanisms during autophagy.
Activation of PI3KCIII is required for recruitment of the ATG protein complex, leading to the induction and expansion of the phagophore membrane (26). PI3KCIII activation is obligatory for canonical formation of autophagosomes (15). We found that blocking PI3KCIII activation inhibited ACEA- and AEA-induced effects but not the CBD-induced autophagy effect, which was only partially inhibited. This suggests that the noncanonical autophagy pathway may be induced in part by CBD, as opposed to ACEA and AEA. Genome-wide association studies revealed single-nucleotide polymorphisms in the ATG16L1 gene in CD patients (35). ATG16L1 localized the ATG5/12/16L1 protein complex to the isolation membrane, followed by the formation of an autophagosome double-membrane vesicle (10, 12). PI3KCIII activity is required for recruitment of the ATG16L1 protein complex (32). In addition, a recent study showed that CB1 receptor polymorphism modulated disease susceptibility in UC patients and the phenotype of CD patients (43). Taken together, these findings indicate that CBD may bypass these pathways, at least in part, to activate autophagy with a functional ATG16L1 gene and polymorphism of the CB1 receptor compromised.
In our model, all three cannabinoids were able to reduce endogenous SOCS3 protein through the CB1 receptor, albeit at different stages of the autophagic process. The dependence on the CB1 receptor for the CBD response was unexpected, as we predicted that CBD-induced autophagy, which was largely independent of the receptor, would still be able to reduce SOCS3 protein. However, CBD has multiple targets, both receptor-dependent and -independent (1, 5, 13, 46), and it may be that an unknown receptor-dependent target feeds into this system. It may also be a reflection of the time point chosen, in that CB1 receptor knockdown may cause unpredictable changes in the rates of protein turnover. We anticipated that reduced SOCS3 would lead to an altered STAT3 phosphorylation status and that blocking autophagy might modulate this effect. However, only CBD appeared to bring about the predicted functional consequence of reduced SOCS3, i.e., increased STAT3 phosphorylation, suggesting that CBD does not directly impact JAK/STAT signaling but, rather, has an effect at the “regulator” (i.e., SOCS3) level. When late autophagy was blocked, CBD no longer had an effect on STAT3 phosphorylation status, confirming a feedback mechanism provided by SOCS3. These findings support previous studies where reduced SOCS3 corresponded to increased STAT3 phosphorylation (50) and CBD reduced SOCS3 mRNA in microglial cells with concomitant phosphorylated STAT3 (16). SOCS3 is expressed in an oscillatory manner in nontransformed cells (50). Unusually, for a cancer cell line, basal SOCS3 levels are consistently high in Caco-2 cells, and our finding that cannabinoids reduce SOCS3 may indicate that they act to increase autophagic degradation of SOCS3 protein, representing a possible mechanism by which cyclic protein expression could be regulated. Furthermore, elevated intestinal SOCS3 in UC and CD (24) could be due to compromised autophagosomal degradation of the protein. We propose that CBD-induced autophagy represents a mechanism contributing to the cyclic expression of proteins, such as SOCS3, in the homeostatic setting. Such cyclic patterns are disrupted in inflammation.
In conclusion, the cannabinoids induce autophagy in a model of intestinal epithelium, an effect that leads to the reduction of SOCS3 (Fig. 5). Overall, our study supports an important role for autophagy in homeostatic regulation of cyclic proteins, as well as therapeutic potential for cannabinoids, where the CB1 receptor and/or autophagy is comprised.
Proposed mechanism for cannabinoid-induced autophagy. The CB1 receptor mediates canonical autophagy, which leads to reduced SOCS3 protein. In addition, CBD can induce receptor-independent and noncanonical autophagy.
GRANTS
K. L. Wright is partly supported by the Dowager Countess Eleanor Peel Trustand R. J. Rigby is partly funded by Medical Research Council New Investigator Research Grant G1100211.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the authors.
AUTHOR CONTRIBUTIONS
L.C.K. performed the experiments; L.C.K. and K.L.W. analyzed the data; L.C.K., R.J.R., and K.L.W. interpreted the results of the experiments; L.C.K. prepared the figures; L.C.K. drafted the manuscript; L.C.K., R.J.R., and K.L.W. approved the final version of the manuscript; R.J.R. and K.L.W. are responsible for conception and design of the research; R.J.R. and K.L.W. edited and revised the manuscript.
ACKNOWLEDGMENTS
The authors thank Dr. Jane Andre for technical expertise with the confocal microscope.
- Copyright © 2014 the American Physiological Society
