The Cannabinoid Receptor CB1 Interacts with the WAVE1 Complex and Plays a Role in Actin Dynamics and Structural Plasticity in Neurons

Abstract

The molecular composition of the cannabinoid type 1 (CB1) receptor complex beyond the classical G-protein signaling components is not known. Using proteomics on mouse cortex in vivo, we pulled down proteins interacting with CB1 in neurons and show that the CB1 receptor assembles with multiple members of the WAVE1 complex and the RhoGTPase Rac1 and modulates their activity. Activation levels of CB1 receptor directly impacted on actin polymerization and stability via WAVE1 in growth cones of developing neurons, leading to their collapse, as well as in synaptic spines of mature neurons, leading to their retraction. In adult mice, CB1 receptor agonists attenuated activity-dependent remodeling of dendritic spines in spinal cord neurons in vivo and suppressed inflammatory pain by regulating the WAVE1 complex. This study reports novel signaling mechanisms for cannabinoidergic modulation of the nervous system and demonstrates a previously unreported role for the WAVE1 complex in therapeutic applications of cannabinoids.

Fig 1. CB1 physically interacts with members of the WAVE1 signaling complex in mouse brain and colocalizes with WAVE1 in neurons.

(A) Expression of EGFP-tagged CB1 (CB1-EGFP) in mouse brain 4 wk after cortical injection of rAAVs. Injection site is marked with a white arrow, and medial longitudinal fissure with a yellow asterisk. Scale bar represents 0.5 mm. (B) Validation of solubilization and pulldown of CB1 by immunoprecipitation using GFP-nanotrap on membrane fractions derived from cortex of mice expressing either CB1-EGFP or GFP alone (as control). Immunoblotting (IB) experiments show that anti-GFP antibody pulls down endogenous CB1 and CB1-EGFP from CB1-EGFP-expressing mice, but not from GFP control mice. The successful pulldown shows the high molecular weight form of CB1 (black arrows), the CB1-EGFP monomer (yellow arrow) and endogenous CB1 (blue arrow head). (C) Summary of MS analysis of GFP nanotrap-immunoprecipitates from the cortex of CB1-EGFP-expressing mice or GFP-expressing control mice (n = 5). rPQ is relative peptide query score. Values for rPQ > 4 indicates specific purification in comparison over negative control. (D) Schematic representation of activation of the WAVE1 complex by GTP-bound (activated) Rac1. (E) Immunoblotting on GFP-nanotrap-immunoprecipitates showing that cytoplasmic FMR1 interacting protein 2 (CYFIP2), NCK-associated protein 1 (NCKAP1), WAVE1 and Rac1 are coimmunoprecipitated with GB1-EGFP, but not with GFP, from the mouse cortex. (F) Immunoblotting on α-CB1-immunoprecipitates showing that WAVE1 is coimmunoprecipitated with CB1 from cortical lysates derived from wild-type mice, but not in lysates from CB1-deficient mice (CB1^-/-). (G) Colocalization of WAVE1 and EGFP-tagged CB1 in growth cones (magnified in inset) of developing cortical neurons with pyramidal morphology. The actin cytoskeleton is counterstained with Phalloidin (growth cone magnified in inset). Scale bars represent 10 μm.

Fig 2. Increased localization of colocalization of WAVE1 and CB1-EGFP at the cell membrane in heterologously-transfected COS7 cells upon ACEA treatment.

(A) Distribution of heterologously-transfected WAVE1 in cells cotransfected with either GFP (control, upper panels) or CB1-EGFP (lower panels) following treatment with vehicle (DMSO 1:30,000) or CB1 agonist (ACEA 100 nM) for 45 min. Scale bar represents 5 μm. (B) Quantitative summary of intensity of WAVE1-immunoreactivity relative to Phalloidin-stained actin at the plasma membrane in GFP- or CB1-EGFP-coexpressing COS7 cells treated with vehicle (white bars) or ACEA (red bars). (C) Quantitative summary of CB1-EGFP-WAVE1 colocalization in vehicle- or ACEA-treated cells calculated as a fraction of the total intensity of CB1-EGFP fluorescence. Values in panels B and C represent the mean ± SEM and are derived from analyses on at least 15 COS7 cells each over several independent culture experiments. *p < 0.05 one way ANOVA followed by posthoc Tukey’s test.

Fig 3. Cannabinoids directly modulate Rac1 activity and WAVE1 phosphorylation by Rac1 Activation and WAVE1 Phosphorylation via CB1.

(A) Schematic representation of the Raichu-Rac1 FRET-biosensor employed to measure Rac1 activity. (B) Real-time images representing changes in Rac1 activity in developing mouse cortical neurons following treatment with a CB1 agonist (ACEA; 100 nM) and an inverse agonist at CB1 (AM251; 600 nM). The FRET signal intensity is represented as a pseudocoloured heat map, and insets show magnified view of growth cones. Scale bars (white) represent 20 μm, and scale bars in the inset (yellow) represent 5 μm. (C) Quantitative summary of normalized FRET ratios over the growth cone area at various time points after addition of ACEA, AM251, NGF (100 ng/ml), or vehicle normalized to the average FRET ratio value over the same area prior to addition of pharmacological agents in developing neurons derived from wild-type mice. (D) Preserved effect of NGF and loss of effects of ACEA as well as AM251 on Rac1 activity in developing cortical neurons derived from CB1^-/- mice. Values in panels C and D represent the mean ± SEM and are derived from analyses on 10–16 neurons per group over at least three independent culture experiments. (E, F) Immunoblot analyses showing changes in phosphorylation state of Serine 397 (pSer397) in WAVE1 upon treatment with ACEA (100 nM) or AM251 (600 nM) as compared to vehicle treatment in cortical neurons derived from wild-type mice without pretreatment (E), with overnight pertussis toxin (PTX) (100 ng/ml) pretreatment or from CB1^-/- mice (F). (G) Quantitative summary of cannabinoid-induced modulation of pSer397 WAVE1 levels normalized to βIII-tubulin in the above groups (n = 5–6 independent culture experiments). All graphs represent mean values ± SEM *p < 0.05, two-way ANOVA for repeated measures (C, D) or one-way (G) ANOVA followed by posthoc Tukey’s test. N.s. stands for not significant.

Fig 4. Visualization of cannabinoidergic modulation of actin dynamics in neuronal growth cones mediated by CB1 and WAVE1.

(A) Real time images of wild-type developing cortical neurons transfected to express LifeAct-GFP, a biosensor for levels of F-actin prior to and at different time points following treatment with ACEA (100 nM), AM251 (600 nM), or vehicle. Scale bar represents 10 μm. (B, C) Analysis on the F-actin dynamics based on cannabinoid-induced changes in LifeAct-GFP intensity (B) and area (C) over time in growth cones of cultured cortical neurons derived from wild-type mice (n = 15–17 neurons per group from 5 independent culture experiments). (D, E) However, no changes in LifeAct intensity (D) nor area of the growth cones (E) were observed with cultured cortical neurons derived from CB1^-/- mice (n = at least 4 neurons per group from 2 independent culture experiments) after treatment. (F, G) Similarly, cultured cortical neurons with down-regulated WAVE1 showed no changes in LifeAct intensity (F) or area of growth cones (G) after cannabinoid treatment (n = at least 4 neurons per group). (H, I) Moreover, the abrogation of the effects after cannabinoid treatment can be contributed to the down-regulation state of WAVE1, since cultured cortical neurons nucleofected with scrambled siRNA, in turn show changes in LifeAct intensity (H) and area (I) of the growth cone following cannabinoid treatment (n = at least 4 neurons per group). (J, K) Immunoblot representation of WAVE1 down-regulation upon siRNA delivery in developing cortical neurons (J) and the corresponding quantification (K) (n = 6). All graphs represent mean values ± SEM *p < 0.05, two-way ANOVA (B-I) or one-way ANOVA (K) for random measures followed by posthoc Tukey’s test.

Fig 5. Cannabinoids regulate growth cone morphology via CB1-Gi-Rac1-WAVE1 signaling.

(A) Characterization of changes in growth cone morphology in developing cultured cortical neurons immunostained with anti-pGAP43 to recognize normal or enlarging growth cones and Phalloidin to label actin. Insets show normal, enlarged, or collapsed morphology in developing cortical neurons; scale bar represents 10 μm in all images and insets. (B) Bidirectional modulation of the proportion of growth cones with collapsed or enlarged morphology upon treatment with ACEA (100 nM), AM251 (600 nM) or vehicle (DMSO 1:30,000) for 1 h in neurons derived from wild-type mice (n = 10 independent culture experiments), but not in neurons from CB1^-/- mice (n = 4). (C–F) Lack of modulation of cannabinoidergic modulation of growth cone morphology upon treatment in the presence of PTX (100 ng/ml), Rac1 inhibitor, CAS 1090893-12-1 (50 μM) or in neurons with siRNA-mediated downregulation of WAVE1; control siRNA-transfected neurons showed significant cannabinoidergic modulation (n = 4). All graphs represent mean values ± SEM. *p < 0.05, one-way ANOVA followed by posthoc Tukey’s test. N.s. stands for not significant.

Fig 6. CB1 receptor is expressed in axonal as well as dendritic compartments of primary neurons matured over 4 wk in vitro.

(A) Coimmunostaining against CB1 receptor and dendritic marker MAP2 on 28 days in vitro (DIV) cortical neurons derived from wild-type mouse embryos (upper panels) shows the distribution of CB1 receptor in the axonal compartments (MAP2-negative) as well as dendritic compartments (MAP2-positive). The rabbit anti-CB1 antibody used throughout this experiment did not yield substantial staining in cortical neurons cultured from CB1-deficient mouse embryos at 28 DIV (lower panels). (B) Magnified images represent white boxed areas from main images. (C) CB1 receptor colocalizes with CamKII-GFP and the synaptic protein PSD95 in dendritic compartments, including several spine heads (white arrow heads), indicating postsynaptic localization. (D) Heterologously-expressed CB1-EGFP in mature cultured cortical neurons shows axonal and dentritic localization upon costaining with MAP2. Phalloidin is used to counterstain the entire actin cytoskeleton. Phall. stands for Phalloidin. (E) Dendritic (arrows) ad axonal (arrowhead) localization of endogenous CB1 in cortical neurons cultured from wild-type mice and matured at 28 DIV. White scale bars represent 20 μm and yellow scale bars represent 5 μm.

Fig 7. Visualization of disassembly of F-actin by cannabinoid agonists in in dendritic spines on mature cortical neurons, leading to reduction in the density of mature spines.

(A) Representative real time images of FRAP experiments. Localized photobleaching over single, individual postsynaptic spines (red circles), led to loss of LifeAct-mCherry labeling of F-actin and progressive recovery of fluorescence with actin polymerization, which was inhibited by ACEA as compared to vehicle control. (B & C) Summary of FRAP values as a function of time expressed as one phase exponential function curves fitted to the respective data sets from neurons treated with ACEA or vehicle either from wild-type cultured cortical neurons (B) or with siRNA-mediated WAVE1 knockdown (C) (13–15 dendritic spines/group from 4 independent culture experiments). (D) Immunoblot example of WAVE1 down-regulation 3 d following siRNA delivery in mature cultured cortical neurons and the corresponding quantification (seven independent culture experiments). (E) Quantitative summary of the magnitude of F-actin recovery in the dendritic spine after photobleaching in neurons from the diverse treatment groups (five independent culture experiments). (F) Representative confocal images of mature cortical neurons transduced with rAAV-CamK-II-GFP virions and treated with ACEA (100 nM) or vehicle for 24 h. GFP-labelled dendritic spines were morphologically classified (lower panels) and quantified. Yellow bars represent 50 μm and white bars represent 5 μm. (G) Quantitative summary of average density of dendritic spines of varying morphology in cortical neurons treated with ACEA (100 nM) or vehicle (n = 20–21 dendrites analyzed from three cultures/group). All graphs represent mean values ± SEM *p < 0.05 as compared to control group and †p < 0.05 as compared to WT neurons, two-way ANOVA (E) or one-way ANOVA (D & G) followed by posthoc Tukey’s test. N.s. stands for “not significant.”

Fig 8. Intrathecal delivery of ACEA inhibits mechanical nociceptive hypersensitivity and nociceptive activity-induced structural plasticity of dendrites on spinal dorsal horn neurons in inflammatory pain conditions in vivo.

(A) Frequency of paw withdrawal responses to mechanical force via plantar application of graded von Frey hairs recorded prior to and 24 h after hind paw intraplantar injection of CFA. Leftward shift in response curves following CFA (indicative of hypersenstivity) is diminished in mice treated intrathecally with ACEA (2 pmol) over 24 h as compared to mice intrathecally receiving vehicle (n = at least 8 per group). (B) Traced images of Golgi-stained large pyramidal-like neurons in laminae II and V of the spinal dorsal horn and dendritic spines (insets) in mice represented in (A). Scale bars represent 10 μm in all panels. (C) Quantitative summary of dendritic spine density in spinal neurons from mice represented in (A) and (B); CFA-induced enhancement in spine density does not occur in mice receiving intrathecal ACEA (n = 12–16 neurons counted from over 4 mice per group). (D, E) Mice with hind paw inflammation show reduced levels of pSer397 WAVE1, quantified over βIII-tubulin as loading control, in spinal lumbar segments L3-L5 at 24 h post-CFA, which is partially and significantly reversed by intrathecal ACEA as compared to vehicle. An example of western blot analysis and quantitative summary from seven mice per group is shown in (D) and (E), respectively. All graphs represent mean values ± SEM *p < 0.05 as compared to basal values within the group and †p < 0.05 as compared to corresponding values in the vehicle group, two-way (A) and one-way (C, E) ANOVA for repeated measures followed by posthoc Tukey’s test.

Fig 9. Role of spinally-expressed WAVE1 in nociceptive activity-induced structural plasticity, inflammatory pain, and cannabinoidergic analgesia in vivo.

(A, B) An immunoblot example and quantitative data from western blot analysis showing down-regulation of WAVE1 in spinal cord segments L3–L5 after intrathecal in vivo application of WAVE1 siRNA as compared to control siRNA. (C) Normal basal spine density but lack of CFA-induced spine remodeling in lamina II/V lumbar spinal neurons at 24 h post-CFA in mice intrathecally injected with WAVE1 siRNA as compared to control siRNA (n = 16–17 spines counted over four mice per group). (D) Frequency of paw withdrawal responses to mechanical force via plantar application of graded von Frey hairs recorded prior to (dashed lines) and 24 h after hind paw intraplantar injection of CFA (solid lines). Leftward shift in response curves following CFA (indicative of hypersensitivity) is diminished with intrathecal siRNA-mediated WAVE1 knockdown (red symbols) as compared to mice intrathecally injected with control siRNA (black symbols). (E) Mechanical hypersensitivity at 24 h post-CFA is significantly reduced by intrathecal ACEA injection over 24 h in mice intrathecally treated with control siRNA (black squares in E, upper graph), but not in mice intrathecally treated with WAVE1 siRNA (red squares in E, lower graph). (F) Quantitative summary of dendritic spine density in laminae II or V neurons in L3–L5 segments of mice which were tested behaviorally in panels animals in (E). Intrathecal ACEA reduces spine density in control siRNA-injected CFA-inflamed mice, but not in WAVE1 siRNA-treated CFA-inflamed mice (16–18 neurons counted from over four mice per group). All graphs represent mean values ± SEM, n = 7–9 mice per group in panels D & E. *p < 0.05 as compared to basal values within the group and †p < 0.05 as compared to corresponding control, two-way (D, E) and one-way (B, C and F) ANOVA for repeated measures followed by posthoc Tukey’s test.