Vascular effects of Siberian ginseng (Eleutherococcus senticosus): endothelium-dependent NO- and EDHF-mediated relaxation depending on vessel size.

Siberian ginseng (SG) has been widely and historically consumed as a health food product for the improvement of self well-being, but whether vascular relaxation may contribute to such a therapeutic health effect has not been studied. We therefore investigated the vasorelaxant effect of the aqueous extract of the roots of SG (Eleutherococcus senticosus Maxim) using several in vitro vascular rings prepared from dog carotid artery, rat aorta and rat mesenteric artery. SG extract (0.04-0.8 mg/ml) caused concentration-dependent relaxation in dog carotid arterial rings pre-contracted with 100 microM phenylephrine (PE), and the relaxation was primarily endothelium-dependent. Treatment with 100 microM L-NOARG (a nitric oxide synthase inhibitor) either prevented or totally reverted SG-induced relaxation, suggesting that the endothelium-dependent relaxation was mediated by NO. Similar endothelium-dependent vascular relaxant responses were also obtained with rat aortic and mesenteric arterial rings, except that it occurred over a relatively higher concentration range of SG (0.5-2.0 mg/ml). When tested in the presence of 300 microM L-NAME, the vasorelaxant effect of SG was inhibited totally in rat aorta but only partially in rat mesenteric artery. The relaxation to SG that was insensitive to L-NAME in rat mesenteric arterial rings was eliminated when the rings (both proximal and distal ends) were pre-treated with a combination of 300 microM L-NAME and 15 mM KCl indicating the involvement of endothelium-derived hyperpolarizing factor (EDHF). This vasorelaxant response of the SG extract was inhibited partially by atropine (1 microM), completely by TEA (5 mM), but not by indomethacin (1 microM) or propranolol (10 microM). SG up to 2 mg/ml had no effect on KCl-induced contraction in any of the vascular rings studied. When compared with carbachol-induced (CCh) relaxation, SG resembles CCh in that the sensitivity to L-NAME inhibition is dependent on vascular size, i.e. aorta >proximal end of mesenteric artery >distal end of mesenteric artery. However, SG exhibited different potencies to relaxation while CCh showed similar potency (EC(50) of about 0.2 microM) in all three vascular segments. In conclusion, we have demonstrated that the vascular effect of SG is endothelium-dependent and mediated by NO and/or EDHF depending on the vessel size. Other vasorelaxation pathways, such as inhibition of K(+)-channels and activation of muscarinic receptors, may also be involved.