Cultured DIV8 hippocampal neurons co-expressing Flag-CB1R-eGFP and LifeAct-mCherry on (A–G) and LifeAct-mCherry only on (H and I). (A) Treatment with CB1R agonist WIN55,212-2 (WIN, 100 nM, added at 0 min) induces rapid retraction of the F-actin-rich domain (arrowheads). Open arrowheads: growth cone position at 0 min. (B) Progression of individual growth cones in control conditions. (C) WIN-induced retraction of individual growth cones. (D) Mean values of growth cone progression in control condition or after treatment with WIN with or without pre-treatment with the CB1R-specific antagonist AM281 (AM, 1 µM). WIN-induced growth cone retraction is effectively abolished by AM. (E) Amplitudes of growth cone retraction induced by different exo- and endocannabinoids, calculated as the net difference of mean growth cone position in the pre-treatment (PRE on D) and post-treatment (POST on D) time intervals from at least three independent experiments. (F) Concentration-response curve of WIN-induced retraction, 9 to 27 neurons per concentration from two independent experiments expressed as percentage of maximal retraction, Emax = 52.2 µm. (G) WIN-induced retraction (25 nM at 40 min) is fully reversible after WIN-washout (at 70 min), n = 9. (H) Mean values of growth cone retraction downstream of endogenous CB1R activation, from four pooled independent experiments, outliers were removed in accordance with the Grubb's test. (I) Amplitudes of growth cone retraction downstream of endogenous CB1R activation after treatment with WIN (100 nM), 2-AG (1 µM), or with WIN (100 nM) after pre-treatment with the CB1R-specific antagonist AM281 (AM, 1 µM). WIN-induced growth cone retraction is effectively abolished by AM. Values in D, F, G, and H are mean ± SEM; values in E and I are presented as boxplots; n.s = p > 0.05, ***p < 0.001, calculated using Kruskal–Wallis one-way ANOVA followed by Dunn's post-tests on (E and I) and paired t-test on (H). Scale bar: 20 µm.