In order to determine whether the faster

In order to determine whether the faster Rh2-induced apoptosis is specific of cholesterol depletion, we determined the cytotoxic effect of Rh2 in 46619 sale depleted or not in sphingomyelin, another abundant plasma membrane lipid exhibiting enrichment in lipid rafts. We showed that, in contrast to cholesterol depletion, sphingomyelin decrease reduced Rh2-induced apoptosis, suggesting the essential role of sphingomyelin in the cytotoxic activity of Rh2. To corroborate this idea, it has been shown that protopanaxadiol (aglycon of ginsenoside Rh2) mediates cytotoxic effects through the activation of neutral sphingomyelinase 2 leading to the hydrolysis of membrane sphingomyelins into pro-apoptotic intracellular ceramides (Park et al., 2013). Altogether, our results highlight the importance of membrane lipid composition for the ginsenoside Rh2-induced apoptosis. To further define the interaction of Rh2 with interfacial and hydrophobic domains of the membrane, fluorescence anisotropy measurements using TMA-DPH and DPH were carried out. Both probes are located within the bilayer, with a shallower depth for TMA-DPH (do Canto et al., 2016). The steady state fluorescence anisotropy is associated to their rotational diffusion, which is sensitive to the order in the membrane. TMA-DPH fluorescence anisotropy values were not affected by cholesterol depletion in accordance with the literature, whereas cholesterol-depleted cells showed higher rigidity in the lipid core region as compared to non-depleted cells (Prasad et al., 2009; Goodwin et al., 2005). Like cholesterol, we report here that Rh2 compacted the bilayer hydrophobic core. In addition, Rh2 relaxed the interfacial packaging of the phospholipid polar heads. These features were altered faster upon cholesterol depletion. To investigate whether changes of the plasma membrane fluidity could potentially elicit cellular responses, we measured the effect of Rh2 on the lipid raft-associated Akt signaling and on the apoptotic pathway. We observed decreased levels of phosphorylated Akt and the activation of the intrinsic apoptotic pathway faster upon cholesterol depletion. Apoptosis markers manifested rapidly in Rh2 treated cells and only 20 to 90 min were necessary to induce the mitochondrial membrane depolarization and the activation of caspase-9 leading to apoptosis. These findings suggest that the primary action of Rh2 involved rapid pathways including phenomena such as dephosphorylation or cleavage of proteins which take less time than slower events like protein expression. Our results do not exclude the involvement of other pathways in the cytotoxic activity of Rh2 as suggest by the partial inhibition of Rh2-induced apoptosis by caspase-9 inhibitor. It has been reported that Rh2 also mediates apoptosis via the death receptor signaling leading to the activation of caspase-8 in Hela cells (Guo et al., 2014) and in A549 cells (Cheng et al., 2005). In addition with the effect of cholesterol to delay the apoptosis induced by Rh2, two other cell effects of ginsenoside Rh2 are attributed to cholesterol-dependent mechanisms: (i) the induction of dendrite formation by Rh2 (18 μM, 2 h) is suppressed by depletion of cholesterol upon MβCD in B16 melanoma cells (Jiang et al., 2010); and (ii) a protective effect on the Aβ-induced amyloid pathology in primary neurons by Rh2 (3 μM, 12 h) is induced by reduction of cholesterol and lipid rafts (Qiu et al., 2014). These two studies could be at first glance in disagreement with the data we reported here. However, in the context of our study and the anticancerous effect of Rh2 we worked at higher concentrations (60 μM). Therefore, a differential ratio between cholesterol and Rh2 in the membrane could reconcile the different observations in distinct cell models treated with differential Rh2 concentrations.
Acknowledgements We thank M.C. Cambier and V. Mohymont for technical assistance and A.S. Cloos (UCL, DDUV) for guiding us in Thin Layer Chromatography experiments. S. Verstraeten is doctoral fellow of the Télévie. This work was supported by the Belgian Fonds de la Recherche Scientifique (F.S.R-FNRS).